Chymotryptic digestion of the cytoplasmic domain of the beta subunit of Na/K-ATPase alters kinetics of occlusion of Rb+ ions.

This paper demonstrates that specific chymotryptic digestion of the cytoplasmic domain of the beta subunit of Na/K-ATPase leads to changes in the kinetics of occlusion of Rb+ ions. The experiments utilize extensively trypsinized Na/K-ATPase, "19-kDa membranes," which lack cytoplasmic loops of the alpha subunit, whereas membrane-embedded fragments (a COOH-terminal 19 kDa and three fragments of 8.1-11.7 kDa) containing transmembrane segments and extracellular loops are intact. The beta subunit is partially split into NH2- and COOH-terminal fragments of 16 and approximately 50 kDa, respectively. Cation occlusion and ouabain binding are preserved. The 19-kDa membranes were incubated, at 37 degrees C, with a selection of proteases, in the presence of Rb+ ions. In these conditions, only alpha-chymotrypsin destroyed the ability to occlude Rb+ ions. This process was associated with truncation of the 16-kDa fragment of the beta subunit in two stages. In the first stage, chymotrypsin removed 10 residues from the 16-kDa fragment to form a 15-kDa fragment (NH2-terminal Ile15) and 4 or 6 residues from the NH2 terminus of the alpha subunit fragment beginning at Asp68. In these membranes Rb+ occlusion was still intact at 37 degrees C. Strikingly, however, deocclusion of two Rb+ ions, which is characteristically biphasic in 19-kDa membranes, displayed deocclusion kinetic with mainly one fast phase. These membranes also showed a much lower affinity for Rb+ ions compared with 19-kDa membranes; and, consistent with the lower Rb+ affinity, Rb+ ions, at nonsaturating concentrations, protected less well against thermal inactivation of Rb+ occlusion. In the second stage, the 15-kDa fragment was truncated further to a 14-kDa fragment (NH2-terminal Leu24), followed by thermal destabilization of Rb+ occlusion even at high concentrations of Rb+ ions. Eventually, the thermally inactivated complex of fragments of alpha and beta subunits was digested to the limit peptides. The results suggest that the cytoplasmic domain of the beta subunit interacts with that of the alpha subunit, possibly with residues leading into the first transmembrane segment, and controls access of Rb+ ions into or out of the occlusion sites.

of gastric H/K-ATPase shows the highest degree of homology with the ␣ subunit of Na/K-ATPase, and H/K-ATPase is the only other P-type pump known to have a ␤ subunit. The ␣ subunits of Na/K-and H/K-ATPase are thought to contain 10 transmembrane segments (for recent evidence, see Goldshleger et al. (1995)) and the functional sites for ATP and cations and sites for the cardiac glycosides on Na/K-ATPase or K ϩ -competitive inhibitors of H/K-ATPase (Glynn and Karlish, 1990;Lingrel and Kuntzweiler, 1994;Sachs et al., 1995).
The ␤ subunit with a single transmembrane segment, three conserved S-S bridges, and three glycosylation sites (seven for H/K-ATPase ␤), is an essential component of the functional pump unit, as shown by expression of ␣ and ␤ subunits in oocytes, yeast, insect, and animal cells (Noguchi et al., 1987;Horowitz et al., 1990;McDonough et al., 1990;Blanco et al., 1994; for a recent review see Chow and Forte (1995)). One major role of the ␤ subunit is stabilization of the ␣ subunit in a functional conformation and facilitation of transfer from the endoplasmic reticulum to the plasma membrane (Geering, 1991). An important functional role has been inferred from observations that different isoforms of the ␤ subunit (␤1, ␤2, ␤3), or the H/K-ATPase ␤ subunit, are partially or wholly interchangeable and have significant effects on the apparent affinity of extracellular potassium or cytoplasmic sodium for activating ATPase activity (Jaisser et al., , 1994Blanco et al., 1994Blanco et al., , 1995 or modulating ouabain binding (Eakle et al., 1992(Eakle et al., , 1994(Eakle et al., , 1995. Further evidence comes from the finding that DTT 1 inactivates cation occlusion of renal Na/K-ATPase presumably due to reduction of the extracellular S-S bridge(s) (Lutsenko and Kaplan, 1993).
Segments involved in subunit assembly and functional effects include 26 or fewer residues in the extracellular loop between M7 and M8 of the ␣ subunit (Lemas et al., 1992Fambrough et al., 1994;Shin and Sachs, 1994). In the extracellular domain of the ␤ subunit the region near the first S-S bridge is thought to interact with the ␣ subunit (Renaud et al., 1991;Hamrick et al., 1993;Fambrough et al., 1994;Noguchi et al., 1994;Lutsenko and Kaplan, 1994;Goldshleger et al., 1995), but hydrophobic residues at the COOH-terminal domain of the ␤ subunit  or a conserved proline residue in the loop of the third S-S bridge  are also important, possibly via indirect conformational effects. Two studies show clearly that the cytoplasmic domain and/or the transmembrane segment of the ␤ subunit is important for assembly and stability of functional complexes (Eakle et al., 1994;Jaunin et al., 1993). Obviously, several regions of subunit interactions are likely to exist and remain to be identified. This paper describes an application of proteolytic digestion to investigate a role of the ␤ subunit in modifying the kinetics of cation occlusion. The experiments utilize "19-kDa membranes" produced by extensive tryptic digestion of pig kidney Na/K-ATPase, in the presence of Rb ϩ or Na ϩ ions and absence of Ca 2ϩ ions Capasso et al., 1992). 19-kDa membranes lack most of the cytoplasmic domain of the protein, whereas the transmembrane segments and extracellular loops are intact Capasso et al., 1992). 19-kDa membranes consist of a 19-kDa fragment, Asn 831 -Tyr 1016 , corresponding to transmembrane segments M7-M10, and five smaller peptides containing pairs of the other transmembrane segments of the ␣ subunit (11.7, 11.3, and 10.9 kDa, NH 2terminal Asp 68 , M1/M2; 8.5 kDa, NH 2 -terminal Ile 263 , M3/M4; 8.1 kDa, NH 2 -terminal Gln 737 , M5/M6). The ␤ subunit is intact or is partially cut to a fragment of 16 kDa (NH 2 -terminal Ala 5 ), extending from the cytoplasmic domain via the transmembrane segment to the loop between the first extracellular S-S bridge, and a glycosylated fragment of about 50 kDa (NH 2terminal Gly 143 ) corresponding to the remaining COOH-terminal domain. In 19-kDa membranes, Rb ϩ and Na ϩ occlusion and ouabain binding are fully preserved, but ATP-dependent functions are lost Capasso et al., 1992;Schwappach et al., 1994).
Direct evidence for a role of the ␤ subunit in cation occlusion would be obtained if it could be demonstrated that a defined proteolytic split in the ␤ subunit causes a defined change in kinetics of cation occlusion. However, such evidence is hard to obtain because 19-kDa membranes are highly protected against proteases in the presence of Rb ϩ ions (Capasso et al., 1992), whereas the occlusion is rapidly thermally inactivated in the absence of occluded Rb ϩ Shainskaya and Karlish, 1994). Previous, indirect evidence, includes the finding that potassium or sodium ions protect the 16-kDa fragment of the ␤ subunit against proteases. This result led Capasso et al., (1992) to discuss a possible role of the ␤ subunit in cation occlusion, probably via conformational interactions with the ␣ subunit. More recently, we have demonstrated that dissociation of occluded Rb ϩ ions leads to digestion of all fragments of ␣ and ␤ subunits to limit membrane-embedded peptides (Shainskaya and Karlish, 1994). The cytoplasmic sequence Ala 5 -Arg 26 and the extracellular sequence Asp 113 -Arg 142 , removed upon further digestion of the 16-kDa fragment of the ␤ subunit, were suggested to be candidates for interactions with the ␣ subunit. We now report that in the presence of Rb ϩ ions, ␣-chymotrypsin digests 19-kDa membranes and destroys Rb ϩ occlusion, as a result of selective splits of the 16-kDa fragment of the ␤ subunit. The experiments suggest that interactions of the cytoplasmic domains of the ␤ and ␣ subunit control access of Rb ϩ ions into and out of the sites and so affect the observed kinetics of cation occlusion and deocclusion. EXPERIMENTAL PROCEDURES Na/K-ATPase was prepared from fresh pig kidney red outer medulla by the rapid procedure described by Jørgensen (1974). Protein, by the method of Lowry, and ATPase activity were determined as described by Jørgensen (1974). Specific activity was in the range of 13-20 units/mg of protein. Before use, the enzyme was dialyzed at 4°C against 1,000 volumes of a solution containing 25 mM histidine, pH 7.0, and 1 mM EDTA (Tris). Standard conditions for preparation of tryptic 19-kDa membranes were described by Capasso et al. (1992). After digestion, membranes were washed, suspended, and stored in a standard medium containing 25 mM imidazole, pH 7.5, 1 mM EDTA, to which 2 mM RbCl was added.
Rb ϩ Occlusion Assays-The Rb ϩ occlusion assays were performed as described by Shani et al. (1987). The medium contained in a volume of 20 -50 l, 5 mM RbCl plus Ϸ5⅐10 6 cpm 86 Rb ϩ , 12.5 mM imidazole, pH 7.5, 0.5 mM EDTA, and 10 -20 g of 19-kDa membranes or digested membranes. For measurement of deocclusion rates, membranes (20 g of protein) previously equilibrated in the standard medium, containing 50 M 86 Rb ϩ , were mixed with 20 l of the medium, containing 3 mM RbCl, at the desired temperature. At times from 0 to 10 min, 500 l of ice-cold sucrose was added, and samples were transferred to Dowex 50 columns for measurement of occluded 86 Rb ϩ .
Treatment with Trypsin, Thermolysin, Clostripain, or Carboxypeptidase B-Standard conditions for digestion with trypsin, thermolysin, clostripain, and carboxypeptidase B were as follows. 19-kDa membranes, 1.5 mg/ml, were suspended in a medium containing 25 mM imidazole, pH 8.0, adjusted with Tris base, 1.0 mM EDTA (Tris), 10 mM RbCl and were incubated at 37°C for 1 h with TPCK-treated trypsin, thermolysin, clostripain, or carboxypeptidase B, 1:5 w/w. Clostripain was dissolved in 10 mM Tris, pH 7.0, and was activated by preincubation in 2.5 mM DTT. A mixture of inhibitors containing 1 mM PMSF, soybean trypsin inhibitor 5:1 (w/w), 2 mM iodoacetamide, 5 mM EDTA, 2 mM EGTA was added to arrest proteolysis. The suspensions were diluted to about 25 ml with a solution of 25 mM imidazole, pH 7.5, 1.0 mM EDTA, 2 mM RbCl, warmed at 37°C for 10 min, and the membranes were collected by centrifugation at 250,000 ϫ g for 1 h. The warming procedure removes traces of proteases adsorbed to the membranes. The membranes were hand homogenized in the latter solution, and the procedure of dilution, warming, and centrifugation was repeated twice. Pellets were resuspended in a standard medium of 25 mM imidazole, pH 7.5, 1 mM EDTA, and 2 mM RbCl.
Digestion with ␣-Chymotrypsin-19-kDa membranes (1-2 mg/ml) were suspended in the standard medium containing 10 or 20 mM RbCl, with the pH adjusted to 8.0 with Tris base, and were incubated with ␣-chymotrypsin (1:1, 1:5, or 1:40 w/w) at 37 or 20°C for different times. 0.2 mM TPCK, 1 mM PMSF, and 150 mM KCl were added sequentially, and the membranes were incubated at room temperature for 10 min upon each addition. The membranes were diluted 15-fold with a solution of the standard medium containing also 150 mM KCl, 1 mM PMSF, 0.2 mM TPCK, centrifuged at 250,000 ϫ g for 1 h, and the pellet was resuspended in standard medium and incubated again with TPCK and PMSF for 10 min at room temperature. The suspension was then centrifuged again and then washed and suspended in standard medium. These procedures completely inactivate chymotrypsin and remove traces of chymotrypsin adsorbed to the membranes.
Prior to SDS-polyacrylamide gel electrophoresis, pellets were resuspended in standard medium, dissolved with 4% SDS, and protein was precipitated by the addition of ice-cold methanol (4:1 v/v) and stored overnight at Ϫ20°C. The delipidated protein was collected by centrifugation for 30 min at 10,000 rpm in a Sorvall centrifuge, dried in a stream of nitrogen, and dissolved in 10% SDS or the gel buffer.
Sequencing-An Applied Biosystems model 475A protein sequenator with an on-line model 120A phenylthiohydantoin analyzer was used. Strips of polyvinylidene difluoride with fragments from four or five lanes of gels run with 50 -100 g of protein/lane were processed together, yielding 10 -70 pmol of amino acids/sequenator cycle.
Calculations-Linear and nonlinear regression analyses were done using the program Enzfitter (Elsevier Bio-Soft) or Matlab.
Materials-86 RbCl was obtained from DuPont NEN. Dowex 50W-X8 (100 mesh) H-form (converted to Tris-form before use) was obtained either from Sigma or Fluka. Trypsin inhibitor (type I-S from soybean), carboxypeptidase Y, carboxypeptidase B, bovine serum albumin (fraction V), TPCK, PMSF, MES, iodoacetamide, thioglycolate, and molecular mass markers (2.5-16.9 kDa) were from Sigma. Choline chloride (recrystallized from hot ethanol) was obtained from Fluka. ␣-Chymotrypsin, thermolysin, and clostripain were obtained from Merck. For SDS-polyacrylamide gel electrophoresis all reagents were of electrophoresis grade from Bio-Rad. Polyvinylidene difluoride paper was from Millipore. Fig. 1 shows effects of treatment of 19-kDa membranes, at 37°C in the presence of Rb ϩ ions, with a selection of proteolytic enzymes (arrows designate cleaved fragments). Clostripain (lane 2) and thermolysin (lane 4) and also the specific endoproteinases Arg-C and Lys-C (not shown) produced no detectable effects on the electrophoretic profile and did not inactivate occlusion. More extensive digestion with trypsin (lane 3) or carboxypeptidase B (lane 5) revealed that certain fragments could be shortened. Trypsin shortened the 19-kDa fragment to 18.5 kDa and also clipped the 11.7-and 11.3-kDa fragments to a 10.9-kDa fragment (containing M1/M2 transmembrane segments) and clipped the 8.1-kDa fragment (containing M5/M6) and the 9.5-kDa peptide, which is the ␥ subunit of Na,K-ATPase (Capasso et al., 1992). In 19-kDa membranes, the three fragments, 11.7, 11.3, and 10.9 kDa, all have the same NH 2 -terminal Asp 68 and so differ in length at the COOH terminus (Capasso et al., 1992). The NH 2 terminus of the 10.9-kDa tryptic fragment in Fig. 1 produced by further tryptic digestion of 19-kDa membranes was also found to be Asp 68 . Carboxypeptidase B (Fig. 1, lane 5) shortened the 8.1-kDa fragment slightly from the COOH terminus. Neither trypsin nor carboxypeptidase B affected Rb ϩ occlusion. ␣-Chymotrypsin (lane 6) was the only protease tested which was able to inactivate Rb ϩ occlusion and digest all components of the 19-kDa membranes in these conditions. Therefore, chymotryptic digestion was looked at in more detail. Fig. 2 depicts the time course of disappearance of Rb ϩ occlusion upon chymotryptic treatment at two concentrations of the protease (chymotrypsin:19-kDa membranes 1:1 and 1:5 w/w). At the lower ratio of chymotrypsin to 19-kDa membranes there was a distinct time lag before inactivation of Rb ϩ occlusion. This biphasic kinetics indicates that cleavage of more than one bond is required to inactivate Rb ϩ occlusion. An increase in Rb ϩ concentration or reduction in the amount of chymotrypsin lengthened the lag period (not shown), whereas at the higher ratio of chymotrypsin to 19-kDa membranes, no lag was observed (Fig. 2).

Structural and Functional Consequences of Chymotryptic Digestion of 19-kDa Membranes-
A characteristic time course of the fragmentation pattern upon chymotryptic digestion is presented in Fig. 3A. Eventually the 19-kDa fragment and other smaller fragments are digested, and Rb ϩ occlusion disappears (Fig. 3A, 90 min; see also Fig. 1, lane 6, and Fig. 6A). The chymotryptic destruction of occlusion is correlated with disappearance of the 19-kDa fragment, as was found previously for destruction of occlusion by trypsin in the presence of Ca 2ϩ . How- . 19-kDa membranes (2 mg/ml) were incubated with chymotrypsin (1:5 w/w) in the standard medium, pH 8.0, and 10 mM RbCl. At times from 1 min to 1.5 h, as indicated, aliquots were withdrawn, the reaction was stopped, and membranes were washed twice with 150 mM KCl as described under "Experimental Procedures." Equal amounts of delipidated protein (ϳ100 g) were applied per lane of a 16.5% Tricine gel. Incubation times with chymotrypsin and Rb ϩ occlusion (% of control) were as follows.  19-kDa membranes (2 mg/ ml) were incubated at 37°C in the standard medium, pH 8.0, containing also 10 mM RbCl ϩ 86 Rb ϩ , and chymotrypsin 1:1 (q) or 1:5 (E) (w/w). At times from 1 min to 1.5 h samples were transferred to the Dowex 50 columns. ever, the fragments appearing prior to loss of occlusion provided more information on the nature of events leading to destruction of occlusion. Identification of intermediate fragments by NH 2 -terminal sequencing is presented in Table I, and these are also depicted visually in Fig. 9A. The striking finding in Fig. 3A is that the 16-kDa fragment of the ␤ subunit (NH 2terminal Ala 5 ; Capasso et al. (1992)) was progressively digested to a fragment a, of 15 kDa with NH 2 -terminal Ile 15 (Table I) i.e. by removal of 10 residues from the NH 2 terminus. At the end of the lag period (Fig. 3A, 20 min), when Rb ϩ occlusion was still intact, most of the 16-kDa fragment had been truncated to the 15-kDa fragment. Upon further digestion, and loss of Rb ϩ occlusion (30 and 90 min), the 15-kDa fragment was truncated further to a fragment of 14 kDa (NH 2 -terminal Leu 24 ), and then the 19-kDa and all other fragments of the ␣ subunit were digested. Note that the band at 11.7 kDa was also progressively digested to the smaller fragments of 11.3 and 10.9 kDa, fragment f, in parallel with truncation of the 16-to the 15-kDa fragment of the ␤ subunit. The 11.7-kDa fragment has NH 2terminal Asp 68 and contains the M1/M2 transmembrane segments (see Capasso et al. (1992)), whereas fragment f (seen also in Fig. 3B) was found to be a mixture of two peptides, in roughly equal amounts, having NH 2 -terminal Ala 72 and Thr 74 (see Table I). Of the other smaller peptides (8 -12 kDa), the fragment of 8.1 kDa disappeared within 1 min, and two fragments, marked d and e, of 7.8 and 7.3 kDa, respectively, appeared. Fragment d, NH 2 -terminal Ala 749 , is a truncated form of the 8.1-kDa fragment (NH 2 -terminal Gln 737 ) containing the M5/M6 transmembrane segments of the ␣ subunit (Capasso et al., 1992). Fragment e is a fragment of a water channel protein from erythrocytes and proximal tubule (Preston and Agre, 1991) and represents a contaminant in the preparation. However, these very early splits are unrelated to inactivation of occlusion (see "Discussion"). The fragments of 9.5 kDa (␥ subunit) and 8.7 kDa (M3/M4) were not cut. Fragment c of 13.5 kDa is a contaminant fragment of chymotrypsin and was not seen in all experiments (see Fig. 6). Fig. 3B shows that digestion at 20°C with a high ratio of chymotrypsin to 19-kDa membranes (1:5 or even 1:1 w/w) led to quantitative conversion of the 16-kDa to the 15-kDa fragment and also conversion of the M1/M2-containing 11.7-kDa fragment to the smaller 10.9-kDa fragment, peptide f; i.e. the electrophoretic profile is similar to that at the end of the lag phase at 37°C (Fig. 3A, 20 min). In this condition there was no loss of occlusion, and at 20°C the 15-kDa fragment was stable for hours upon application of chymotrypsin, and occlusion was maintained (data not shown). A similar pattern was obtained at 37°C using a low ratio of chymotrypsin to 19-kDa membranes (1:40 w/w) for 1 h (not shown).
The availability of a stable form (Fig. 3B), referred to as the chymotryptic intermediate, has allowed us to compare some properties of the Rb ϩ occlusion in these membranes with those of the 19-kDa membranes. Since digestion experiments indicated a difference in behavior at 20 and 37°C, it was of interest to investigate possible temperature-dependent differences in properties of Rb ϩ occlusion.
Measurements of Rb ϩ occlusion with varying Rb ϩ concentrations at 37°C showed a large difference in apparent K m between 19-kDa membranes (24 -35 M in different experiments) and the chymotryptic intermediate (211.1 Ϯ 62.9 M), whereas no difference was observed at 20°C (31.6 Ϯ 3.9 and 35.0 Ϯ 4.3 M) (experiments not shown). A possible complication with such experiments at 37°C is that occlusion is partially thermally inactivated, at nonsaturating concentrations of Rb ϩ (Or et al., 1993 and see below). An alternative approach, which is free of this complication, was to compare dissociation of occluded 86 Rb ϩ ions, at 37 and 20°C, into media containing nonradioactive Rb ϩ ions (Fig. 4, A and B). At 20°C (Fig. 4A) deocclusion is markedly biphasic as reported previously , and again, there was essentially no difference between 19-kDa membranes and the chymotryptic intermediate. At 37°C (Fig. 4B), dissociation from 19-kDa membranes remains strongly biphasic, with fitted rates of 0.52 s Ϫ1 and 0.017 s Ϫ1 and equal amplitudes of the two phases (0.52 and 0.48) respectively. For the chymotryptic intermediate, at 37°C, dissociation of 86 Rb ϩ was fitted to a biphasic curve with a rate constant of 1.14 s Ϫ1 and 0.014 s Ϫ1 and amplitudes of 0.86 and 0.14 respectively. Thus, 86 Rb ϩ deocclusion from the chymotryptic intermediate is characterized by a large increase in the amplitude and rate of the first phase, so that the slow phase is much reduced in amplitude (see "Discussion" and Fig. 9A). This phenomenon is consistent with a large decrease in Rb ϩ affinity at 37°C for the chymotryptic intermediate (see "Discussion").
Another way of demonstrating the decrease in Rb ϩ affinity of the chymotryptic intermediate utilized a phenomenon of thermal inactivation of Rb ϩ occlusion in 19-kDa membranes and protection by occluded cations such as Rb ϩ or Na ϩ ions Shainskaya and Karlish, 1994). The experiment in Fig. 5  A more detailed time course of the profile of chymotryptic fragments identified later steps in the digestion (Fig. 6). Since the 15-kDa fragment is stable at 20°C but not at 37°C membranes digested at 20°C as in Fig. 3B were subjected to a second chymotryptic digestion at 37°C (Fig. 6A). Note that although eventually, the 15-kDa fragment was digested to the 14-kDa fragment, and occlusion was destroyed (Fig. 6A, 60 min), after 1 and 2 min the 15-kDa fragment appeared to undergo partial conversion to the 14-kDa fragment, while occlusion was fully intact. Quantification of the ratios of 15-to 14-kDa fragments by scanning the gel (Fig. 6B) showed that conversion of the 15-to the 14-kDa fragment preceded destruction of occlusion, about 30% of the 15-kDa fragment being converted to the 14-kDa fragment before occlusion began to drop.
A strong hint as to order and mechanism of later steps in the destruction of occlusion was obtained in the experiment of Fig. 7. We compared thermal stability, at 5 mM RbCl and 37°C, of the chymotryptic intermediate with that of a preparation redigested for 2 min (as in Fig. 6A), so that about 30% of the 15-kDa fragment is converted to the 14-kDa fragment, but Rb ϩ occlusion is still 100% intact or redigested for 10 min when 58% of occlusion remained. In the latter two preparations, a fraction of the occlusion is distinctly more thermolabile than that of the chymotryptic intermediate. The remaining fraction was inactivated at about the same rate as the control. The result in Fig.  7 could be expected if the more labile fraction represented those molecules in which the 15-kDa fragment has been truncated further to the 14-kDa fragment (see "Discussion" and Fig. 9B).
Interactions between the ␤ and ␣ Subunits?-Under "Discus- sion" we shall consider the possibility of an interaction between the ␤ subunit and the M1/M2-containing fragment, based on the digestion and functional data presented above.
An additional indication for interactions between the ␤ subunit and the ␣ subunit has come from an observation of effects of ouabain. Ouabain is known to decrease greatly the rate of deocclusion of Rb ϩ ions from native Na/K-ATPase (Forbush, 1983) and as shown recently also from 19-kDa membranes. 2 Fig. 8A demonstrates a similar effect of ouabain on Rb ϩ deocclusion from the chymotryptic intermediate at 37°C. Deocclusion in the presence of ouabain is a double exponential with rate constants of 0.23 and 0.025 s Ϫ1 , with equal amplitudes for the two phases compared with the fit without ouabain to rate constants of 0.81 s Ϫ1 and 0.016 s Ϫ1 and amplitudes 0.76 and 0.2, respectively (seen also in Fig. 4B). An alternative way of demonstrating that ouabain greatly decreased the rate of Rb ϩ deocclusion comes from a finding that 2 mM ouabain stabilized the chymotryptic intermediate to thermal inactivation of Rb ϩ occlusion, at 2 mM Rb ϩ (Fig. 8B). Thus, ouabain, which is thought to be bound to the ␣ subunit, appears to be able largely to reverse the functional changes characteristic of the chymotryptic intermediate.

Stages in the Destruction of Rb ϩ Occlusion by Chymotrypsin-
The primary indication that digestion of the ␤ subunit is a crucial event in the destruction of Rb ϩ occlusion was the finding that ␣-chymotrypsin is the only one of a selection of proteases tested which is able to inactivate Rb ϩ occlusion in the presence of occluded Rb ϩ ions, and this process is associated with specific splits of the 16-kDa fragment of the ␤ subunit.
Other proteases tested include trypsin, carboxypeptidase B, clostripain, thermolysin (Fig. 1), endoproteinase Arg-C, endoproteinase Lys-C, and as shown previously (Capasso et al., 1992) the nonspecific proteases, Pronase and proteinase K. The reason for this remarkable selectivity is unknown. Stages in the chymotryptic digestion and associated changes in Rb ϩ occlusion kinetics, depicted figuratively in Fig. 9, are now discussed in turn.
The earliest event is truncation of the 8.1-kDa fragment (NH 2 -terminal Gln 737 , containing M5/M6) to 7.8 kDa (NH 2terminal Ala 749 ) (Fig. 3, A and B, 1 min; see Fig. 9A). However, this split is not correlated with loss of occlusion or the end of the lag period. Furthermore, a control experiment showed that after 1 min of chymotryptic digestion, the thermal stability of Rb ϩ occlusion at 37°C and nonsaturating Rb ϩ is unchanged (result not shown). The 7.8-kDa fragment is similar to a fragment (NH 2 -terminal Ser 750 ) described earlier (Karlish et al., 1991), which was produced by digestion of renal Na/K-ATPase with trypsin untreated with TPCK, i.e. it was generated by traces of chymotrypsin in the trypsin. This truncation also had no effect on Rb ϩ occlusion .
The first stage, prior to loss of occlusion, is associated with truncation of the 16-kDa fragment of the ␤ subunit to the 15-kDa fragment by removal of 10 residues, AKEEGSWKKF, and also by truncation of the 11.9-and 11.3-kDa fragments of the ␣ subunit (M1/M2) by removal of 4 or 6 residues, DGPN or DGPNAL, to 2 E. Or and S. J. D. Karlish, unpublished data.

FIG. 7. Comparison of thermal stability of chymotryptic intermediate and membranes redigested with chymotrypsin. Chymotryptic intermediate (E) membranes
(1 mg/ml) were preequilibrated at room temperature for 30 min in the standard medium, pH 7.5, containing 20 mM RbCl. Aliquots were transferred to 37°C, and chymotrypsin (1:5, w/w) was added. After 2 (q) and 10 (f) min the reaction was stopped by the addition of ice-cold reaction medium, containing 1.5 mM RbCl ϩ 86 Rb ϩ , 100 mM Tris⅐HCl, pH 7.5, 1 mM PMSF, and 0.2 mM TPCK. After 1 h at 0°C, samples were incubated at room temperature for 20 min and then transferred to 37°C. At a time from 5 min to 1 h, 20-l aliquots were transferred to Dowex 50 columns for measurement of Rb ϩ occlusion. was preequilibrated for 40 min at 20°C in a standard medium, containing 50 M of 86 Rb ϩ , with (Ⅺ) and without (f) 1 mM ouabain. 20 l of the medium, containing 3 mM RbCl at 37°C, was added. At times from 0 to 10 min samples were transferred to Dowex 50 columns for measurement of occluded Rb ϩ . The lines represent the theoretical curves for double exponential decays, using the best fit parameters quoted under "Results." Panel B, chymotryptic intermediate membranes (1 mg/ml) were preequilibrated at 20°C for 40 min in the standard medium, pH 7.5, and 2 mM RbCl ϩ 86 Rb ϩ with (Ⅺ) and without (f) 2 mM ouabain. Tubes were then incubated at 37°C. At times from 3 min to 3 h samples were transferred to Dowex 50 columns. produce overlapping 10.9-kDa fragments (Fig. 9A). Which of these splits is necessary for inactivation of occlusion, and are the two splits related? The specific chymotryptic split of the ␤ subunit between Phe 13 and Ile 14 appears to be an essential first step, for there is prior evidence that a proteolytic split at the Asn 71 -Ala 72 bond is not a prerequisite for inactivation of Rb ϩ occlusion. In the presence of magnesium, P i , and ouabain, trypsin truncates the 11.7-kDa fragment to a fragment with NH 2 -terminal Ala 72 , without inactivating Rb ϩ occlusion (Lutsenko and Kaplan, 1994). In Rb ϩ -containing media with added Ca 2ϩ ions, trypsin gains access and splits the Asn 71 -Ala 72 bond only when Rb ϩ occlusion is already thermally inactivated (Shainskaya and Karlish, 1994). There are two hints that the chymotryptic splits of the 16-and 11.7-kDa fragments are connected. First, these two fragments are digested with a similar time course (Fig. 3A). Second, a comparison of the tryptic and chymotryptic digestion is suggestive of the connection. As seen in Fig. 1, trypsin truncated the 11.7-and 11.3-fragments to a 10.9-kDa fragment, with the same NH 2 -terminal residue, Asp 68 , i.e. it cut at the COOH-terminal side, in the presence of Rb ϩ ions. By contrast to the result with chymotrypsin, the 16-kDa fragment of the ␤ subunit was not cut by trypsin, and Rb ϩ occlusion was not affected. The specific chymotryptic split between Phe 13 and Ile 14 of the ␤ subunit may facilitate the truncation of the M1/M2 fragment between Asn 71 and Ala 72 , which is a nonclassical chymotryptic split site as well as between Leu 73 and Thr 74 . (The converse hypothesis, that chymotrypsin first truncates M1/M2 and as a result the Phe 13 -Ile 14 bond of the ␤ subunit becomes accessible, is less likely because the split between Asn 71 and Ala 72 is unspecific and could be expected to occur with trypsin or other proteases in the same conditions.) The chymotryptic intermediate showed a decrease in apparent affinity for Rb ϩ and faster deocclusion rate, particularly from the second site, at 37°C but not at 20°C (Fig. 4). Biphasic deocclusion kinetics reflects an ordered release of 86 Rb ϩ from the two sites in a narrow pocket (Glynn et al., 1985) or one that is intermittently open because of the presence of a flickering gate (Forbush, 1987). The result is that the nonradioactive Rb ϩ in the medium is able to block dissociation of the deeper of the two 86 Rb ϩ ions. For the chymotryptic intermediate, at 37°C, the medium Rb ϩ is largely unable to block dissociation of the inner Rb ϩ ion, and thus both Rb ϩ ions dissociate rapidly and randomly. This phenomenon is depicted in Fig. 9B, step 1, as a result of a wider pocket, but it might also be depicted as the result of more frequent openings of a gate. The lower Rb ϩ affinity of the chymotryptic intermediate is also reflected in the faster rate of thermal inactivation of occlusion at nonsaturating concentrations of Rb ϩ ions (Fig. 5). Thus, Rb ϩ ions protect the chymotryptic intermediate less well than 19-kDa membranes against thermal inactivation at 37°C, even though the intrinsic thermal stability of the two membranes in the absence of Rb ϩ is quite similar (Fig. 5). Protection by ouabain against thermal inactivation, associated with a reduced dissociation of occluded Rb ϩ (Fig. 8A), provides strong support for this mechanism. At a high Rb ϩ concentration (30 mM in Fig. 5 and also 10 mM) the chymotryptic intermediate is stable at 37°C, i.e. thermal instability is not the immediate cause of destruction of occlusion. At least one additional chymotryptic split is required to destroy Rb ϩ occlusion at 37°C.
The second stage of digestion appears to be further truncation of the 15-kDa to the 14-kDa fragment. The Phe 23 -Leu 24 bond may become accessible to chymotrypsin at 37°C but not at 20°C, because of loss of a stabilizing interaction between the 15-kDa fragment and other domains of the protein.
The key observation is that at 37°C, the membranes with the 15-kDa fragment partially truncated to 14 kDa, but with intact Rb ϩ occlusion, display a fraction of thermally unstable molecules compared with the chymotryptic intermediate, even at a high Rb ϩ concentration (Fig. 7). Presumably, if it were possible to produce an intermediate consisting of only 14-kDa fragments, a much higher fraction of thermally unstable occlusion might be observed. The Rb ϩ affinity in those molecules may be even lower than in the chymotryptic intermediate (Fig. 9B, step 2).
A third and final stage appears to involve irreversible thermal inactivation of Rb ϩ occlusion and then digestion of all fragments in the membranes. The process is similar to that we have described for digestion of 19-kDa membranes by trypsin following displacement of occluded Rb ϩ by Ca 2ϩ ions (Shainskaya and . Displacement of occluded Rb ϩ ions leads to thermal inactivation of Rb ϩ occlusion, loss of Rb ϩ -dependent stabilizing interactions, and disaggregation of the complex of fragments, and exposure of all the extramembrane tails and loops of ␣ and ␤ subunits to trypsin. The peptide products of the chymotryptic digestion were not analyzed in detail, but the overall picture (Fig. 3A, 90 min, or Fig. 6A, 60 min) is quite similar to that observed for the Ca 2ϩ /trypsin digestion. Eventually the 19-kDa fragment is digested to small peptides of 3-4 kDa containing single transmembrane segments, while the peptides of 8 -11 kDa containing M1/M2, M3/M4, and M5/M6 are truncated at either or both NH 2 -and COOH-terminal tails to peptides of 6 -8 kDa, still containing the pairs of transmembrane segments. The present results are in line with our previous conclusions that maintenance of occlusion requires an intact 19-kDa fragment, and the fragments in 19-kDa membranes exist in a Rb ϩ -stabilized complex Shainskaya and Karlish, 1994).
The proposed mechanism whereby a change in Rb ϩ affinity causes dissociation of Rb ϩ ions and then thermal inactivation due to loss of stabilizing interactions may be applicable to other structural modifications. A case in point is inactivation of Rb ϩ occlusion in 19-kDa membranes by DTT (Lutsenko and Kaplan, 1993). In the experiments of Lutsenko and Kaplan (1993)  intermediate displays a similar sensitivity to DTT, and Rb ϩ ions partially protect (experiment not shown).) Rb ϩ occlusion in 19-kDa membranes is saturated at 2 mM Rb ϩ (K D Ϸ 40 M; Karlish et al. (1990)), and hence the requirement for 25 mM Rb ϩ for protection implies that reduction of the S-S bridge(s) decreased the affinity for Rb ϩ but did not itself inactivate occlusion. However, dissociation of Rb ϩ ions at 2 mM should then lead to irreversible thermal inactivation. This mechanism is consistent with the proposal that reduction of S-S bonds in the ␤ subunit inactivates the enzyme by causing delocalized denaturation of the ␣ subunit (Kirley, 1990). It is less easily reconciled with Lutsenko and Kaplan's proposal (1993) that the extracellular portion of the ␤ subunit directly gates cation occlusion, closing access to the sites by interacting with the intramembrane portion.
Interactions between the Cytoplasmic Domains of the ␤ and ␣ Subunits-Alteration of properties of cation occlusion by truncating the cytoplasmic domain of the ␤ subunit suggests the existence of interactions between cytoplasmic domains of the ␤ and ␣ subunits. These interactions appear to be disrupted progressively as chymotryptic digestion proceeds from 16-to 15-kDa fragments and then to 14-kDa fragments. Candidate regions of the ␣ subunit include the cytoplasmic tails of M1/M2, M3/M4, and M5/M6 pairs, and within the 19-kDa fragment, the cytoplasmic tail of M7, loop between M8/M9 and the COOHterminal tail of M10. As discussed above, the cytoplasmic domains of the ␤ subunit and the M1/M2-containing fragment might be interacting. However, firm identification of interacting domains remains to be established.
How might a structural interaction between the cytoplasmic domains of the ␤ and ␣ subunits affect gating of the occlusion domain, so that relatively minor structural modification associated with the chymotryptic digestion produces the subtle functional changes seen in Figs. 4, 5, 7? One possibility is that the cytoplasmic domains of the ␤ and ␣ subunits interact directly within the occlusion domain at the flickering gate (Forbush, 1987), i.e. the NH 2 -terminal domain of the ␤ subunit constitutes a component of the gate. This is reminiscent of the ball-and-chain mechanism suggested for inactivation of potassium channels (see Jan and Jan (1992)). Alternatively, the gate may lie deeper within the transmembrane segments, so that disruption of interactions of cytoplasmic domains of the ␤ and ␣ subunits affects deocclusion less directly, by loosening the interactions between transmembrane segments. The indirect mechanism appears to be somewhat favored by the finding that ouabain decreases the rate of Rb ϩ deocclusion on the chymotryptic intermediate and restores the ordered release phenomenon (Fig. 8). Ouabain is bound in the ␣ subunit and is thought to interact with residues in several extracellular loops of the ␣ subunit (Price and Lingrel, 1988;Lingrel et al., 1991;Canessa et al., 1993;Shultheis et al., 1993;Blostein et al., 1993) and also with residues within the M1 transmembrane segment (Arystarkhova et al., 1992;Canessa et al., 1992;Shultheis et al., 1993;Antolovic et al., 1995).
Independent evidence for a functional role of the cytoplasmic domain of the ␤ subunit has been obtained recently (Jaisser et al., 1996). Xenopus ␤1 subunits truncated at the NH 2 terminus were expressed with Bufo ␣1 subunit in Xenopus oocytes. Truncation of 34 residues, at a position equivalent to Trp 31 in the pig ␤ sequence, induced a large decrease of the apparent affinity for extracellular potassium. Truncation at positions equivalent to Gly 25 or Asn 17 in the pig ␤ sequence had no effect. Chow and Forte (1993) have reported that an inhibitory monoclonal antibody, recognizing epitopes on the cytoplasmic surface of the ␤ subunit of H/K-ATPase, affects potassium activation of p-nitrophenol phosphatase, suggesting again a role for this domain in modulating potassium activation of H/K-ATPase.