Specific Cu2+-catalyzed oxidative cleavage of Na,K-ATPase at the extracellular surface.

This paper describes specific Cu2+-catalyzed oxidative cleavage of alpha and beta subunits of Na,K-ATPase at the extracellular surface. Incubation of right side-out renal microsomal vesicles with Cu2+ ions, ascorbate, and H2O2 produces two major cleavages of the alpha subunit within the extracellular loop between trans-membrane segments M7 and M8 and L7/8. Minor cleavages are also detected in loops L9/10 and L5/6. In the beta subunit two cleavages are detected, one before the first S-S bridge and the other between the second and third S-S bridges. Na,K-ATPase and Rb+ occlusion are inactivated after incubation with Cu2+/ascorbate/H2O2. These observations are suggestive of a site-specific mechanism involving cleavage of peptide bonds close to a bound Cu2+ ion. This mechanism allows several inferences on subunit interactions and spatial organization. The two cleavage sites in L7/8 of the alpha subunit and two cleavage sites of the beta subunit identify interacting segments of the subunits. L7/8 is also close to L9/10 and to cation occlusion sites. Comparison of the locations of Cu2+-catalyzed cleavages with Fe2+-catalyzed cleavages (Goldshleger, R., and Karlish, S. J. D. (1997) Proc. Natl. Acad. Sci. U. S. A. 94, 9596-9601) suggests division of the membrane sector into two domains comprising M1-M6 and M7-M10/Mbeta, respectively.

A full understanding of the working of P-type cation pumps will require knowledge of molecular structure. The recent cryoelectron microscopy studies of Ca-ATPase and H-ATPase demonstrate the overall shape of these proteins at 8-Å resolution and the presence of 10 trans-membrane ␣-helical rods, most of which are tilted at an angle to the membrane (2,3). Previous work, utilizing a variety of biochemical and molecular techniques, has also indicated the existence of 10 trans-membrane segments in catalytic subunits of the higher type II P-type pumps (4). In addition, unexpected features such thermal instability of the C-terminal domain has been described (5,6), and a re-entrant loop between M7 and M8 (L7/8) has been postulated (7). Spatial organization of trans-membrane helices is unknown. Although an attempt has been made to predict the identity of the 10 helices of Ca-ATPase (2), direct determination is lacking. Covalent cross-linking experiments are beginning to provide information on proximities between transmembrane helices (8 -10), 1 but additional approaches are required.
The ␤ subunit is an important component of Na,K-ATPase and H,K-ATPase. The ␤ subunit stabilizes the ␣ subunit and is required for functional expression of the pump at the plasma membrane (12,13). In addition, ␣-␤ interactions affect kinetic properties of the pump (e.g. Refs. 14 -16). The ␣ subunit interacts with the ␤ subunit strongly in the extracellular loop L7/8 (17), and studies using the yeast two-hybrid system show that only four residues 894 SYGQ are required (18). Site-directed mutagenesis indicates an important role of several extracellular regions or residues of the ␤ subunit (19 -21). However, these experiments do not distinguish between direct participation in subunit interactions or indirect effects. Thus, again, direct information on interacting regions of subunits is necessary. Metal-catalyzed oxidative cleavage is a popular technique for footprinting nucleic acids (22) and is now being used for mapping proteins (23)(24)(25). Metal-catalyzed cleavage of proteins often involves site-specific mechanisms in which only peptide bonds close to bound metals are cleaved (26 -28). Recently we have used this approach to demonstrate specific Fe 2ϩ -catalyzed cleavage of the ␣ subunit of Na,K-ATPase (1). The experiments indicate that peptide bonds are cleaved with a probability depending on their proximity to a Fe 3ϩ or Fe 2ϩ ion bound at the cytoplasmic surface and provide information on the organization of cytoplasmic domains in E 1 and E 2 conformations. This paper describes a different application of the technique, namely Cu 2ϩ /ascorbate/H 2 O 2 2 -catalyzed oxidative cleavage of Na,K-ATPase at the extracellular surface. These cleavages provide information on interacting regions of ␣ and ␤ subunits, organization of trans-membrane segments, and functional effects. EXPERIMENTAL PROCEDURES Na,K-ATPase, with specific activities 12-17 units/mg of protein, was prepared from pig kidney (29) and was stored at Ϫ70°C in a solution of 250 mM sucrose, 25 mM histidine, pH 7.2, and 1 mM EDTA (Tris). Right side-out renal microsomes were prepared as described (5). Before use, the sealed vesicles were dialyzed for 5 h at 20°C and overnight at 4°C against a solution containing 250 mM sucrose, 10 mM Tris-HCl, pH 7.2, in order to remove trapped K ϩ ions, histidine, and EDTA Cleavage Reactions-Suspensions of renal Na,K-ATPase or microsomes (0.1-1 mg/ml) were incubated at 20°C with freshly prepared solutions of 4 -16 mM ascorbate (Tris) plus 4 -16 mM H 2 O 2 , without or with added CuCl 2 or other metals. To arrest the reaction, 5-10 mM EDTA or 5-fold concentrated gel sample buffer with 5 mM EDTA was added. Samples were assayed for Na,K-ATPase activity, Rb ϩ occlusion, or applied to gels, respectively.
Deglycosylation of the ␤ Subunit-Control or cleaved microsomes were washed twice in a medium containing 10 mM Tris, pH 8.0, 1 mM EDTA. 0.1-1 units of PNGase F per g of microsomal protein was added, and the mixture incubated for 24 h at 37°C. Sample buffer was then added.
Gel Electrophoresis, Blotting to PVDF, and Immunoblots-Procedures for running of 10% Tricine, SDS-PAGE (25-45 g of protein per lane), electroblotting to polyvinylidene difluoride (PVDF) paper, and immunoblots have been described (5,30). For quantification of the ␣ subunit, the band was cut out of the gel, and optical density of Coomassie stain was extracted into 1% SDS solution and was measured at 595 nm (1).
Functional Assays-Na,K-ATPase and Rb ϩ occlusion activity of microsomes was assayed after demasking with deoxycholate (5,29). The medium contained 250 mM sucrose, 25 mM histidine, pH 7.2, and 1 mM EDTA (Tris), 0.25-2 mM RbCl (plus 86 Rb), about 100 g of vesicle protein treated or not treated with 1 mM ouabain, 1 mM MgCl 2 , and 2 mM P i (Tris). Rb ϩ occlusion represents the ouabain-inhibited fraction of 86 Rb binding (about 50% at 0.5 mM RbCl). Materials-For SDS-PAGE all reagents were electrophoresis-grade from Bio-Rad. Tris (ultra pure) was from Bio-Labs, Jerusalem. L-(ϩ)-Ascorbic acid (catalog number 100127) and 30% H 2 O 2 (catalog number 822287) were from Merck. PNGase F recombinant was obtained from Boehringer Mannheim. All other reagents were of analytical grade. Fig. 1 presents the basic finding that micromolar concentrations of Cu 2ϩ ions, together with ascorbate and H 2 O 2 , catalyze specific cleavages of the ␣ subunit at the extracellular surface. The ␣ subunit and its fragments were visualized in immunoblots to detect C-terminal residues (anti-KETYY). As we have described recently (1), incubation of renal Na,K-ATPase with ascorbate plus H 2 O 2 produces five specific cleavages of the ␣ subunit, catalyzed by contaminant Fe 2ϩ or Fe 3ϩ ions (Fig. 1A, A/HP, asterisks). Addition of Fe 2ϩ ions amplifies these cleavages (1), but addition of many other metals, including Cu 2ϩ ions, does not affect these cleavages during short incubations. However, we now find (Fig. 1A, A/HP, Cu 2ϩ ) that longer incubations of renal Na,K-ATPase with ascorbate/H 2 O 2 and added Cu 2ϩ ions (10 M) produce specific cleavage of the ␣ subunit to two major fragments, 13.6 and 12.3 kDa, and also reduces intensity of the Fe 2ϩ -catalyzed fragments (see "Discussion"). Prolonged incubation with ascorbate/H 2 O 2 /Cu 2ϩ led to the appearance of a variety of additional minor fragments and general degradation of the protein (see also Fig. 1B, ϩDOC). The 13.6-and 12.3-kDa fragments were also detected by Coomassie stain, in addition to the Fe 2ϩ -catalyzed cleavage fragments, but attempts to sequence them after transfer to PVDF were unsuccessful (not shown). By comparison with known chymotryptic or tryptic fragments (see Fig. 2A, Chy, and see Ref. 5), the N termini of the 13.6-and 12.3-kDa fragments were predicted to lie at the extracellular surface. Therefore, the sidedness of the cleavages was tested directly using right side-out oriented renal microsomes (5) Incubation of permeabilized vesicles with ascorbate/H 2 O 2 plus Cu 2ϩ produced the 13.6-kDa fragment but also extensive degradation of the protein due presumably to access of the Cu 2ϩ ions to the cytoplasmic surface and nonspecific Cu 2ϩ -catalyzed cleavages. Fig. 2 presents further characterization of specific Cu 2ϩcatalyzed cleavages using intact microsomes and anti-KETYY (Fig. 2, A-D) or other antibodies (Fig. 2B). Incubation with ascorbate/H 2 O 2 and increasing Cu 2ϩ concentrations led to a roughly parallel increase in the major 13.6-and 12.3-kDa fragments and also to several minor fragments (6.5, 7.8, and 21 kDa and higher molecular masses, see asterisks) and a smear of antibody staining, see Fig. 2A. By comparison with known chymotryptic fragments (N termini, Arg 880 , Glu 902 , Gln 939 , and Arg 972 ) produced by cleavage at the extracellular surface (Chy, see Ref. 5) the N termini of the 13.6-and 12.3-kDa fragments can be assigned quite closely. Their N termini are located just before and just after that of the chymotryptic fragment with N terminus Glu 902 , i.e. they lie within a stretch of about 10 residues, Tyr 895 -Lys 905 , located within the extracellular loop L7/8. The minor 7.8-and 6.5-kDa fragments are slightly larger or smaller, respectively, than the known chymotryptic fragment (N terminus Arg 972 ) produced by cleavage of the extracellular loop L9/10 (5), and thus they too must be the products of cleavage within L9/10. The minor 21-kDa fragment is a product of an extracellular cleavage in L5/6 (see below). The other minor higher molecular weight fragments (asterisks) and the general degradation of the protein are the results of nonselective Cu 2ϩ -catalyzed cleavages at the cytoplasmic surface in open vesicles. This inference is based on similar observations of digestion of purified (non-sided) Na,K-ATPase with ascorbate/H 2 O 2 and 10 M Cu 2ϩ mentioned above. Use of other antibodies complements the information provided by anti-KE-TYY (Fig. 2B). Anti-Leu 815 -Gln 828 , which recognizes a sequence thought to lie within M6 and the cytoplasmic loop L6/7, stains three fragments, the 21-kDa and also 91-and 7.2-kDa fragments. Anti-Leu 815 -Gln 828 does not stain the 13.6-, 12.3-, 7.8-, and 6.5-kDa fragments recognized by anti-KETYY, because their N termini lie downstream of this epitope. Staining of the 21-kDa fragment by anti-Leu 815 -Gln 828 shows that the cleavage site lies upstream of the epitope and, together with the apparent M r value, provides a strong indication that this fragment is a product of an extracellular cleavage between M5 and M6 (L5/6). The 91-kDa fragment(s) stained by anti-Leu 815 -Gln 828 complements the major 13.6-and 12.3-kDa fragments from the N terminus of the ␣ subunit (see "Discussion" for a comment on the 7.2-kDa fragment). Anti-Asn 889 -Glu 902 , which recognizes a sequence in L7/8 located just before M8, stains the 21-kDa and major 13.6-and 12.3-kDa fragments, detected with anti-KETYY, but not the 7.8-and 6.5-kDa fragments because their N termini lie downstream of the epitope. Anti-Asn 889 -Glu 902 does not stain the 91-kDa fragment(s), recognized by anti-Leu 815 -Gln 828 , because the C terminus lies upstream of the epitope. Fig. 2C presents a time course with parallel appearance of 13.6-and 12.3-kDa fragments and also detection of the minor 21-, 7.5-, and 6.5 -kDa fragments. Fig. 2D shows that, of the heavy metals ions Cu 2ϩ , Fe 2ϩ , Mn 2ϩ , or Ni 2ϩ added to the medium, only Cu 2ϩ ions were able to catalyze the specific extracellular cleavages.

RESULTS
Other experiments showed that the Cu 2ϩ -catalyzed cleavages were not significantly affected by OH radical scavengers (formate, t-butyl alcohol, or mannitol at 20 mM). Also no significant differences were found in the presence of Na ϩ , Rb ϩ , P i ϩ Mg, or P i ϩ Mg ϩ ouabain (not shown).
The findings in Figs. 1 and 2 lead to the hypothesis that Cu 2ϩ ions catalyze cleavage of the ␣ subunit from a specific binding site. As a test of this hypothesis we have examined whether Cu 2ϩ ions induce cleavage of the ␤ subunit, which is known to interact strongly with the ␣ subunit in L7/8 (17,18). The experiment in Fig. 3 used antibodies that recognize either the N-or C-terminal halves of the ␤ subunit (Fig. 3). Fragments of the ␤ subunit were visualized better after deglycosylation. Large amounts of PNGase were used, but deglycosylation, although extensive, was usually incomplete, and three bands with none, one, and two remaining sugar chains were usually observed (lanes A/HP). After treatment with Cu 2ϩ /ascorbate/ H 2 O 2 two fragments were clearly recognized by both sets of antibodies (A/HP, ϩCu 2ϩ ). The Cu 2ϩ concentration depend-ence and time course of appearance of both fragments were similar to those for cleavages of the ␣ subunit (not shown). M r values of the intact deglycosylated ␤ subunit and deglycosylated fragments are presented in Table I. (No significant difference in M r values of fragments was detected in experiments with incomplete deglycosylation (Fig. 3) or complete deglycosylation (Fig. 4).) The 22.3-kDa (N-terminal) plus 11.5-kDa (Cterminal) peptides represent one set and the 10.2-(N-terminal) plus 21.8-kDa (C-terminal) peptides a second set of complementary fragments. Since the apparent M r of the deglycosylated ␤ subunit based on SDS-PAGE (34.7 Ϯ 1.13 kDa) is close to the true value (34.97 kDa), the same is likely to be true for the cleaved fragments. In support of this hypothesis, a tryptic fragment of the ␤ subunit, Ala 5 -Arg 142 , with true M r 16,053, runs with an apparent mass of 16 kDa (30). Thus, using both sets of values in Table I and sequence data, the first cleavage site can be located between residues 90 and 115, before the first S-S bridge (Cys 125 -Cys 148 ), and the second cleavage site between residues 194 and 205, between the second (Cys 158 - Cys 174 ) and third (Cys 212 -Cys 275 ) S-S bridges. The larger fragment has the same electrophoretic mobility in reducing and non-reducing conditions (not shown). This observation implies that the more C-terminal cleavage is not located within an S-S bridge, and confirms its location between the second and third S-S bridge.
Another test for a site-specific mechanism exploited thermal denaturation (Fig. 4). Heating of microsomes to 55°C inactivates Na,K-ATPase and exposes the cytoplasmic loop L8/9 and also the ␤ subunit to proteases at the extracellular surface (5).
Site-specific cleavage should be prevented by thermal disorganization, whereas cleavages dependent on random interactions of OH radicals or several sites for Cu 2ϩ ions might be amplified. The experiment utilized anti-KETYY (Fig. 4A) and the anti-␤ (C-terminal) antibodies (Fig. 4B). The experiment using anti-KETYY is complicated by the fact that heating itself exposes L7/8 to an endogenous protease and appearance of a fragment of 14.5 kDa (described in Fig. 9 of Ref. 5, Denatured ϪCu). Nevertheless it is clear that, after heating, incubation with ascorbate/H 2 O 2 /Cu 2ϩ produced neither the 12.3-kDa nor the 7.8-and 6.5-kDa fragments (Denatured ϩCu). The experiment in Fig. 4B using anti-␤ is clear cut because the endogenous protease does not digest the ␤ subunit. In this case, after heating to 55°C, Cu 2ϩ -catalyzed cleavage of the ␤ subunit did not occur (compare Native and Denatured ϩCu). In short, disorganization of structure prevents specific Cu 2ϩ -catalyzed cleavages in L7/8, L9/10, and the ␤ subunit. Fig. 5A shows that incubation of microsomes with Cu 2ϩ / ascorbate/H 2 O 2 inactivated Na,K-ATPase slightly faster than cleavage of the ␣ subunit. A similar phenomenon has been  4. Effect of thermal denaturation on Cu 2؉ -catalyzed cleavage of ␣ and ␤ subunits. Native microsomes or microsomes heated to 55°C for 30 min (denatured) were incubated with ascorbate/ H 2 O 2 without or with 5 M CuCl 2 for 5 min. The blots were probed with anti-KETYY or anti-␤ (C-terminal) antibodies.

FIG. 5. Functional effects of incubation with Cu 2؉ -/ascorbate/ H 2 O 2 .
A, microsomes, 1 mg/ml, were incubated with ascorbate/H 2 O 2 without or with 10 M CuCl 2 for the indicated times. Samples were then taken for assay of Na,K-ATPase activity or applied to a gel. B, microsomes, 0.1 mg/ml, were incubated with ascorbate/H 2 O 2 for 5 min without or with the indicated concentrations of CuCl 2 . Samples were then taken for Na,K-ATPase activity and Rb ϩ occlusion (see "Experimental Procedures"). Control refers to samples incubated with ascorbate/H 2 O 2 but without CuCl 2 .
observed for Fe 2ϩ -catalyzed cleavage (1). Fig. 5B shows that Rb ϩ occlusion was also inactivated by ascorbate/H 2 O 2 /Cu 2ϩ . Inactivation of Rb ϩ occlusion appeared to be a little less sensitive than inactivation of Na,K-ATPase over this range of Cu 2ϩ concentrations but the difference is within the experimental error. From Fig. 5B and similar experiments, we estimate K 0.5 for Cu 2ϩ of 1-2 M for inactivating Na,K-ATPase and Rb ϩ occlusion.

DISCUSSION
A Site-specific Mechanism of Cu 2ϩ -catalyzed Cleavage- Fig.  6 presents a topological model of ␣ and ␤ subunits, positions of Cu 2ϩ -catalyzed extracellular cleavages (arrows), and Fe 2ϩ -catalyzed cytoplasmic cleavages (asterisks). The proposal of a sitespecific mechanism of Cu 2ϩ -catalyzed cleavage, which implies proximity of L7/8, L9/10, and the ␤ subunit, is based on the following: (a) the specificity of cleavage of both ␣ and ␤ subunits (Figs. 1-3); (b) sideness of action of Cu 2ϩ ions (Figs. 1 and 2); (c) similar time course and Cu 2ϩ concentration dependence of appearance of the various fragments (Fig. 2); (d) suppression of cleavages in ␣ and ␤ subunits after thermal inactivation (Fig.  4); and (e) insensitivity to OH radical scavengers. This mechanism assumes that the bound Cu 2ϩ ion catalyzes oxidative cleavages of peptide bonds with a probability depending on their spatial arrangement and proximity to the Cu 2ϩ ion. Since the major cleavage sites are located in L7/8, presumably the Cu 2ϩ ion binds primarily within this loop, makes contact with the two interacting regions of the ␤ subunit, and to a lesser extent with L9/10, Cu 2ϩ , or Cu ϩ ions are usually ligated by cysteine, methionine, or histidine residues. Candidate residues at the extracellular surface of the ␣ subunit include His 875 (L7/8), Cys 911 (M8), His 912 (M8), Cys 964 (M9), and Cys 983 (M10).
The evidence on L5/6 (open arrow) is less clear cut. The minor 7.2-kDa fragment recognized by anti-Leu 815 -Gln 828 (Fig.  2B) must be the product of two splits, one in L5/6 and the other in L7/8. Therefore, it cannot be excluded that the split in L5/6 is mediated by a different Cu 2ϩ ion from that which catalyzes all other splits.
Site-specific Cu 2ϩ -mediated cleavage of peptide bonds may involve a Cu 2ϩ -peroxyl intermediate (24) or locally generated OH radicals (27,28). Cleavage can lead to fragments with C-terminal amide and N-terminal ketoacyl derivatives (␣-amidation pathway) (28). Blocked N termini preclude sequencing by Edman degradation. Alternatively (diamide pathway), cleavage can produce fragments, with C-terminal diamides and N-terminal isocyanate derivatives that are hydrolyzed to generate free N termini, thus permitting sequencing (23,27). The fact that the 13.6-and 12.3-kDa fragments gave no sequence is compatible with the ␣-amidation pathway. Cu 2ϩ ions also catalyze site-specific oxidative modification of side chains without cleavage (26,28). This phenomenon could explain faster rates of enzyme inactivation compared with chain cleavage (see Fig.  5A and see Ref. 1). Sites of oxidative modification and chain cleavage are presumed to be close to each other since both are catalyzed by the bound metal.
Regions of ␣ and ␤ Subunit Interactions-The two major Cu 2ϩ cleavage sites in L7/8 of the ␣ subunit, within the sequence approximately 895-905, are very close to the residues 894 SYGQ shown to be crucial for ␣-␤ interactions (18). Identification of the two Cu 2ϩ cleavage sites in the ␤ subunit should thus indicate which residues interact with the ␣ subunit. These two points are located within residues approximately 90 -115, before the S-S bridge (Cys 125 -Cys 148 ), and residues approximately 194 -205, between the second (Cys 158 -Cys 174 ) and third (Cys 212 -Cys 275 ) S-S bridges. The yeast two-hybrid work (18) also pointed to residues 63-125 of the ␤ subunit as a region of ␣-␤ interaction. Recently, similar experiments for H,K-ATPase (33) have pointed to two regions of interaction of the ␤ subunit (64 -126, and 156 -188), similar to those found in this work. Note, however, that unlike the two-hybrid system, Cu 2ϩ -catalyzed cleavages reveal interacting segments of the native protein.
Implications for Topological and Spatial Arrangement of Trans-membrane Segments-Cleavages in L5/6 and L9/10 (21or 7.8-and 6.5-kDa fragments, respectively) demonstrate that these loops reach the extracellular surface. Previously an extracellular location of L9/10 has been demonstrated only for denatured protein (5). For Na,K-ATPase the location of L5/6 has not been demonstrated, although that of H,K-ATPase has been demonstrated by omeprazole labeling of Cys 813 and Cys 822 (34).
Comparison of the location of Cu 2ϩ -and Fe 2ϩ -catalyzed cleavages suggests that the two sets of cleavages demarcate separate domains. One domain includes M1-M6 and major cytoplasmic loops, whereas the other includes M7-M10 of the ␣ subunit and the ␤ subunit. (Reduction of anti-KETYY staining of the Fe 2ϩ -catalyzed cleavage fragments by Cu 2ϩ -catalyzed cleavages near the C-terminal (Fig. 1) implies that Fe 2ϩ -and Cu 2ϩ -catalyzed cleavages occur on the same molecule and emphasizes the separateness of the two domains.) A division of structure into two domains has also been proposed on the basis of sequence comparisons of type I heavy metal pumps and the type II higher P-type pumps. Type I pumps contain the equivalent of the first six segments and the conserved cytoplasmic regions (and two extra N-terminal trans-membrane segments), whereas type II pumps contain the extra four C-terminal segments (4,35). Thus, M1-M6 with the cytoplasmic loops may constitute a core structure for ion transport (M4 -M6) and energy transduction, whereas M7/M10 of higher pumps evolved to serve additional functions (4,35). A further indication for compact folding of the M7-M10 domain comes from the observation of high resistance to proteolytic digestion of the C-terminal fragment (M7-M10) of Na,K-ATPase (30,37) and H,K-ATPase (36). Finally, the cryoelectron microscope studies of Ca-ATPase and H-ATPase show directly that the organization of the trans-membrane segments is non-uniform (2, 3).
These indications for a domain organization provide an important constraint on spatial arrangement of trans-membrane segments. For example, compact folding of M7/M10 segments makes it unlikely that M7 and M8 are widely separated, although the relatively long connecting loop (Ϸ40 residues) could, in principle, allow such a separation. Based on this constraint of domain organization, as well as specific evidence on proximity of trans-membrane segments (8 -10), 1 we have recently proposed a model for the approximate spatial organization of the trans-membrane segments of ␣ and ␤ subunits (41).
Functional Effects of Cu 2ϩ -catalyzed Cleavage-Since L7/8 is by far the major site of Cu 2ϩ -catalyzed cleavage (and presumably Cu 2ϩ -catalyzed oxidative modification of side chains), inactivation of Rb ϩ occlusion and Na,K-ATPase can be attributed to structural perturbation in this loop. Cation occlusion sites are located within the membrane, particularly within M4, M5, and M6 (37)(38)(39). Thus our findings imply a close connection between L7/8 and M4, M5, or M6. The interaction could occur at the membrane surface or within the membrane, if L7/8 is a re-entrant loop (7). The present findings also supply an economical explanation of inactivation of Na,K-ATPase or Rb ϩ occlusion by reduction of S-S bridges in the ␤ subunit (11,40) or effects of different ␤ subunits on K ϩ affinity (14 -16). These effects are probably mediated indirectly via perturbation of the ␣-␤ interaction in L7/8 and the interaction of L7/8 with M4, M5, and M6. Thus an important role of the ␤ subunit may be to maintain L7/8 in the correct disposition with respect to M4, M5, and M6.