Role of the S3 Stalk Segment in the Thapsigargin Concentration Dependence of Sarco-endoplasmic Reticulum Ca2+ ATPase Inhibition*

The sarco-endoplasmic reticulum Ca2+ ATPase (SERCA) is specifically inhibited by thapsigargin (TG), whereas the Na+,K+-ATPase is not. Large chimeric exchanges between Ca2+ and Na+,K+-ATPases (Norregaard, A., Vilsen, B., and Andersen, J. P. (1994) J. Biol. Chem. 269, 26598–26601), as well as photolabeling with a TG azido derivative (Hua, S., and Inesi, G. (1997) Biochemistry 36, 11865–11872), suggest that the S3-M3 (stalk and membrane-bound) region of the Ca2+ ATPase is involved in TG binding. We produced small site-directed changes in the S3 stalk segment of the Ca2+ ATPase and found that mutation of five amino acids to the corresponding Na+,K+-ATPase residues increases by 3 orders of magnitude the TG concentration required for inhibition of Ca2+ ATPase and coupled Ca2+transport. A single mutation in the S3 stalk segment (Gly257 → Ile) is sufficient to increase by 1 order of magnitude the TG concentration required to produce 50% inhibition. By comparison, mutations yielding a nine-amino acid homology in the M3 transmembrane segment, or a 25-amino acid homology in the S4 stalk segment, do not affect the ATPase sensitivity to TG. We suggest that specific binding of TG to the S3 stalk segment, in addition to stacking of the TG ring structure at the membrane interface, determines the high affinity of the ATPase for the inhibitor.

consisting of defined parts of SERCA and Na ϩ ,K ϩ -ATPase, because TG interacts specifically with the former and not with the latter. Previous studies with large chimeric exchanges, however, produced strong inhibition of catalytic turnover and transport. Nevertheless, Ca 2ϩ -dependent formation of phosphorylated intermediate was preserved, and a reduced sensitivity of this parameter to TG was obtained upon chimerization of the entire S3-M3 (stalk and membrane-bound) region (13). In contrast, the TG sensitivity was not altered significantly if other large regions were exchanged (13)(14)(15). We describe here the construction and functional characterization of more discrete chimeric changes, involving stepwise mutations of one and up to several amino acids in the S3 or S4 stalk segments of the SERCA, to match the corresponding residues of the Na ϩ ,K ϩ -ATPase. A similar strategy was previously used in studies of ouabain binding by the Na ϩ ,K ϩ -ATPase, taking advantage of ouabain-sensitive and ouabain-insensitive isoforms of the enzyme (16).

EXPERIMENTAL PROCEDURES
PCR Mutagenesis and Protein Expression-The chicken fast muscle SERCA-1 cDNA (17), containing 11 artificial and unique restriction sites spaced at approximately 300-bp intervals to facilitate cassette exchange (18), was subcloned into pUC19 vector for site-directed mutagenesis. For this purpose, the cassette delimited by BssHII and Bsu36I restriction sites (579 bp including S3 and M3 coding sequences), and the cassette delimited by BamHI and Bsu36I (318 bp including M4 and S4 coding sequences), were amplified by PCR using oligonucleotide "flanking" primers. Furthermore, complementary mutagenic oligonucleotides of 23-35 bp length were synthesized for each individual mutation. These primers were utilized to hybridize DNA sequences internal to the flanking primers and were used for PCR mutagenesis by the overlap extension method as described by Ho et al. (19). Briefly, two overlapping fragments containing the mismatched base(s) of the targeted sequence were amplified in separate PCR reactions. The PCR mixtures contained 1 M each of flanking and mutagenic primers, 800 M dNTPs, 20 ng of SERCA-1 cDNA, 2.5 units of Pfu (Pyrococcus furiosus) DNA polymerase, and Pfu buffer (Stratagene, Menasha, WI) in a final volume of 100 l. The reaction products were separated by electrophoresis on a 3% low melting agarose gel (FMC, Rockland, ME), and the appropriate M r band was excised and melted at 72°C for 5 min. The eluted fragments were fused, and the entire cassette was amplified using both flanking primers. The mutant cassette was then exchanged with the corresponding cassette of wild-type cDNA in pUC19 vector, and sequenced by the dideoxy chain-termination method using Sequenase (U. S. Biochemical Corp.). Additive chimeric mutations were introduced by sequential PCR mutagenesis using mutant DNA as template. Finally, the mutated cDNA was subcloned into COS-1 expression vector pCDL-SR␣296 (20) for transfection and overexpression of protein under control of the SV40 promoter. COS-1 cell cultures and transfections were carried out as described by Sumbilla et al. (15).
Microsomal Preparation and Immunodetection of Expressed Protein-The microsomal fraction of transfected COS-1 cells was obtained by differential centrifugation of homogenized cells (15). Immunodetection of expressed ATPase in the microsomal fraction was obtained by Western blotting, using the CaF-5C3 monoclonal antibody to SERCA-1 (17), as described by Sumbilla et al. (15).
Functional Studies-ATP-dependent Ca 2ϩ transport was measured by following the uptake of radioactive calcium tracer by microsomal vesicles. The reaction mixture contained 20 mM MOPS, pH 7, 80 mM KCl, 5 mM MgCl 2 , 0.2 mM CaCl 2 , 0.2 mM EGTA, variable concentrations of TG, 5 g of microsomal protein/ml, 5 mM potassium oxalate, and 3 mM ATP. The reaction was started (37°C) by the addition of oxalate and ATP, and was terminated at sequential times by vacuum filtration (0.45-m Millipore filters). The filters containing the calcium loaded vesicles were washed with 2 mM LaCl 3 and 10 mM MOPS, pH 7.0, and were then processed for determination of radioactivity by scintillation counting. The observed rates of Ca 2ϩ transport were corrected to reflect the level of expressed ATPase in each microsomal preparation, as revealed by immunoreactivity and with reference to microsomes obtained from COS-1 cells transfected with wild-type SERCA-1 cDNA.
ATPase activity was assayed in a reaction mixture containing 20 mM MOPS, pH 7.0, 80 mM KCl, 3 mM MgCl 2 , 0.2 mM CaCl 2 , 5 mM sodium azide, 20 g of microsomal protein/ml, 3 M ionophore A23187, 3 mM ATP, and 0 -1000 nM TG. Ca 2ϩ -independent ATPase activity was assayed in the presence of 2 mM EGTA and no added Ca 2ϩ . The reaction was started (37°C) by the addition of ATP, and samples were taken at serial times for determination of P i by the method of Lanzetta et al. (21). The Ca 2ϩ -dependent activity was calculated by subtracting the Ca 2ϩindependent ATPase from the total ATPase and was corrected to account for the level of expressed protein in each microsomal preparation as revealed by immunoreactivity, and with reference to microsomes obtained from COS-1 cells transfected with wild-type SERCA-1 cDNA.

RESULTS
Description of Mutants-The chimerization scheme for analysis of the S3 stalk segment is shown in Table I. In part A (top and bottom lines), nine amino acids of the SERCA S3 segment and of the corresponding Na ϩ ,K ϩ -ATPase segment are aligned according to Norregaard et al. (1). In the intervening lines, it is then shown how eight amino acids of the Ca 2ϩ ATPase sequence were stepwise mutated to complement the conserved Phe 256 , and yield a nine-amino acid chimeric sequence identical

(A) and M3 (B) segments
The native sequences of the S3 (stalk) and M3 (transmembrane) segments of SERCA-1 (17,25) and of the corresponding segments of the rat kidney Na ϩ ,K ϩ -ATPase ␣1 isoform (26,27) are aligned (top and bottom, respectively) according to Norregaard et al. (1). In the S3 segment (A), one out of nine amino acids is identical in the two ATPases, and eight are subjected to stepwise mutation from the Ca 2ϩ to the Na ϩ ,K ϩ -ATPase sequence to yield a nine amino acid segment of homology (Mutant S3,8). In the M3 segment (B), four out of nine amino acids are homologous, and five are mutated from the Ca 2ϩ to the Na ϩ ,K ϩ -ATPase sequence to yield a nine amino acid segment of homology (Mutant M3,5).  (17,25) and of the rat kidney Na ϩ ,K ϩ -ATPase (26,27) are aligned (top to bottom, respectively) by matching the aspartyl residues undergoing phosphorylation. Sixteen amino acids are homologous, and four are conservative replacements in the native sequences. Five residues were mutated from the Ca 2ϩ to the Na ϩ ,K ϩ -ATPase to yield, therefore, a 25-amino acid homologous segment (Mutant S4,5). mutations on the S3,5 chimera Single mutations of the WT SERCA-1 to the corresponding Na ϩ ,K ϩ -ATPase residues, as well as single or double mutations in the S3,5 chimera (to more conservative replacements as compared to the corresponding Na ϩ ,K ϩ -ATPase residues) were produced, and their effects on the sensitivity of the enzyme to TG were studied as shown in Figs. 4 and 5. The K I values are the numbers used to fit the inhibition curves (three experiments per each mutant), and correspond to the TG producing 50% inhibition. to that of the Na ϩ ,K ϩ -ATPase. Seven mutants were derived from this stepwise procedure and processed for transgenic expression in COS-1 cells. Table 1, part B, shows a similar chimerization scheme for analysis of the M3 transmembrane segment. Five mutations were produced to yield a nine-amino acid chimeric sequence (including four residues of native homology).
Another chimerization scheme was directed to the Ca 2ϩ ATPase S4 segment, and involved mutation of five amino acids to complement 16 homologous amino acids and four conservative replacements in the native sequence, yielding a 25-residue chimeric sequence that is homologous to the corresponding segment of the Na ϩ ,K ϩ -ATPase (Table II).
In addition to the chimeric exchanges described above, more discrete mutations were produced in the S3 segment (as explained in Table III and in the text below) to test the effects of limited perturbations in this region.
Levels of Expression-Approximately 10% of the COS-1 cells transfected under our conditions overexpress and target the Ca 2ϩ ATPase to the endoplasmic reticulum, as shown by in situ microscopic visualization following immunofluorescent staining (18). In the experiments reported here, Western blot analysis of microsomal fractions obtained from the harvested cells revealed similar levels of expression for the wild-type ATPase and ATPase mutants (Fig. 1). Minor variations of expression levels were generally related to the efficiency of transfection rather than the presence of mutations. At any rate, the expression levels were quantitated by densitometry of Western blots, and the resulting values were used to correct the functional parameters to be described below, with reference to the wildtype enzyme.
Ca 2ϩ Uptake and ATP Hydrolysis-As originally reported by Maruyama and MacLennan (22), microsomal vesicles obtained from transfected COS-1 cells sustain ATP-dependent Ca 2ϩ uptake and related ATPase activity. These are specific and useful functional signals, which, as shown in Fig. 2, proceed at constant rates for several minutes. Ca 2ϩ uptake is a highly specific functional parameter, which is totally inhibited by TG ( Fig.  2A). On the other hand, the observed ATPase activity includes Ca 2ϩ -independent and TG-insensitive components that must be subtracted from the total in order to obtain the specific Ca 2ϩ -dependent and TG-sensitive ATPase activity (Fig. 2B).
When we compare the Ca 2ϩ transport activities of wild-type and mutant proteins (Fig. 3), we find that the transport rates are unaffected by mutations of up to four amino acids in the S3 segment (mutants S3,2 and S3,3; Fig. 3), but undergo a progressive reduction as the number of mutated amino acids in the S3 segment is increased (mutants S3,4 to S3,8; Fig. 3). On the other hand, the nine-amino acid chimeric homology with the Na ϩ ,K ϩ -ATPase in M3 (M3,5) and the 25-amino acid homology in S4 (S4,5), produce only 60% and 30% reduction of the transport rates, respectively (see Fig. 3). Similar effects of mutations were observed on the ATPase hydrolytic rates (data not shown). We then tested the TG concentration sensitivity of wild-type enzyme and of mutants retaining at least 50% activity.
It is shown in Fig. 4A that the K I for Ca 2ϩ transport inhibition by TG is gradually shifted by 3 orders of magnitude to higher concentrations (1.7 ϫ 10 Ϫ10 M to 1.25 ϫ 10 Ϫ7 M) as the number of mutated amino acids in the M3 segment is increased FIG. 1. Quantitation of protein expression. Wild-type and mutated constructs were expressed in COS-1 cells, and identical protein aliquots of solubilized cell homogenates were analyzed by Western blotting as described under "Experimental Procedures." The bands were quantitated by densitometry, and the so-derived values were utilized to estimate the content of transgenic protein per unit of total protein.

FIG. 2. Examples of ATP-dependent Ca 2؉ uptake (A) and
ATPase activity (B). These steady state measurements were performed as explained under "Experimental Procedures," using wild-type SERCA-1 enzyme as the catalyst. Note the total inhibition of Ca 2ϩ uptake by TG, whereas the ATPase activity displays a component that is Ca 2ϩ -independent and TG-insensitive. The symbols refer to: no TG to five or six, to yield a six-or seven-amino acid homologous segment between Ca 2ϩ -ATPase and Na ϩ ,K ϩ -ATPase. Similar results were observed for Ca 2ϩ -dependent ATPase activity (Fig. 4B).
Because in the work of Norregaard et al. (1), the ATPase sensitivity to TG was lost following an extensive chimeric replacement involving the entire S3-M3 segment, we produced a nine-amino acid chimeric exchange within the M3 transmembrane segment, to be compared with the analogous chimeric exchange produced in the S3 stalk segment. We found that the chimeric exchange in the M3 segment produced approximately 60% transport inhibition (Fig. 3), but had no effect on the ATPase sensitivity to TG (Fig. 5).
To test the specificity of mutations in the S3 segment with regard to TG sensitivity, we then studied a five-amino acid chimeric mutation in the S4 segment of the ATPase (S4,5 in Table II). It should be pointed out that the segment chosen for our studies contains already 16 homologous, and four conservatively replaced, amino acids, when compared with the corresponding segment of the Na ϩ ,K ϩ -ATPase (Table II). Therefore, mutation of five heterologous amino acids results in a 21-amino acid chimeric segment. It is of interest that this S4 mutant produces only 30% inhibition of function (Fig. 4), and its TG sensitivity is identical to that of the wild-type enzyme (Fig. 5).
Considering that a five-amino acid mutation in the S3 stalk segment reduces the ATPase sensitivity to TG by 3 orders of magnitude, we then produced more discrete mutations to test the effects of limited perturbations in this region. We found that a single mutation of Gly 257 to the corresponding Na ϩ ,K ϩ -ATPase residue (Ile) is sufficient, by itself, to reduce the ATPase sensitivity to TG by 1 order of magnitude (Table III). The important role of Gly 257 is also revealed by the significant reversal of TG sensitivity reduction observed when the Gly 257 3 Ile mutation in the S3,5 chimera is changed to the more conservative Gly 257 3 Ala mutation. Additional reversal to WT behavior is produced by a further change involving the Gln 259 3 Leu mutation to the more conservative Gly 259 3 Asn mutation (Table III). This further reversal is not obtained if Glu 259 is mutated to Gly (Table III). DISCUSSION TG produces global inhibition (i.e. phosphoenzyme formation, hydrolytic activity, and Ca 2ϩ transport) of ATPase, whereas no inhibition is produced by TG on the Na ϩ ,K ϩ -ATPase (5). This specificity has motivated chimeric studies to obtain information on the TG binding site, assuming that replacement of a critical SERCA sequence with the corresponding Na ϩ ,K ϩ -ATPase sequence would interfere with inhibition. Large chimeric exchanges produce nearly total inhibition of hydrolytic activity and Ca 2ϩ transport. Nevertheless, forma- FIG. 4. Sensitivity of wild-type SERCA-1 and S3 mutants to TG. Ca 2ϩ uptake (A) and ATPase (B) measurements were performed as described under "Experimental Procedures." The symbols refer to wildtype SERCA-1 enzyme (‚) and to chimeras S3,2 (f), S3,3 (), S3,4 (ࡗ), S3,5 (q), and S3,6 ("). Each point is the average of three steady state velocities obtained as in Fig. 2, using two or three different protein preparations. The experimental points were computer-fitted, using a single site binding equation with dissociation constants yielding the best fit for each set of data. Note that the theoretical fits obtained for the Ca 2ϩ uptake data are better than those obtained for the ATPase data, perhaps due to a greater specificity of the Ca 2ϩ uptake measurements, relative to the ATPase measurements. tion of Ca 2ϩ -dependent phosphoenzyme is retained by such chimeric proteins and, based on phosphoenzyme measurements, it was shown that large chimeric exchanges in the cytosolic SERCA region do not interfere with inhibition by TG (15). In contrast, chimeric exchange of a 30-amino acid (Leu 253 -Val 283 ) S3-M3 segment (1) reduces the inhibitory effect of TG concentrations as high as 2 M TG ( Fig. 6 and Table IV).
In our experiments, we produced small chimeric changes by stepwise site-directed mutations in the S3 stalk segment and studied the so-derived SERCA mutants with regard to their ability to sustain Ca 2ϩ transport and hydrolytic activity, and their sensitivity to TG. Mutations of up to six amino acids in the S3 stalk segment yield an enzyme retaining ample Ca 2ϩ transport and ATPase activities for studies of inhibition by TG. We then found that small chimeric mutations of the S3 segment ( Fig. 6 and Table IV) produce a marked reduction of the ATPase sensitivity to TG. Single mutation of Gly 257 3 Ile is, by itself, very effective. It should be noted that, given higher TG concentrations, the full inhibitory effect is obtained in all cases. This indicates that chimeric mutations in the S3 segment reduce the ATPase affinity for TG, but do not interfere with the inhibitory mechanism.
It is of interest that the SERCA S4 segment includes already, in its native structure, a very high and unique degree of structural homology with the Na ϩ ,K ϩ -ATPase as well as other cation ATPases. Such a localized homology suggests that this segment may serve as a common structural device for long range functional linkage of the catalytic site in the extramembranous region, and the cation binding site in the membranebound region (18,23). On the other hand, such a high degree of homology between the TG-sensitive SERCA and other TGinsensitive ATPases suggests that S4 segment is not involved in the enzyme interaction with TG. In fact, we found that a five-amino acid mutation, which (due to additional native ho-mology) yields a 25-amino acid segment identical to that of the Na ϩ ,K ϩ -ATPase, does not interfere at all with the concentration dependence of TG inhibition (Figs. 5 and 6). Therefore, the lack of effect of larger chimeric mutations in the S4 segment underlines the topological specificity of the rather limited S3 chimeric mutations that interfere with ATPase sensitivity to TG. Mutations in the transmembrane M3 segment are also ineffective.
Our experiments demonstrate that the S3 stalk segment is specifically involved in determining the ATPase sensitivity to TG and suggest that the S3 segment is involved in TG binding. That the reduced sensitivity may be due to long range effects of the S3 mutations is unlikely, because extensive chimeric exchanges in other regions ( Fig. 6 and Table IV) do not affect the ATPase sensitivity to TG (1,15). It is possible that specific interaction of TG with the S3 segment, in addition to stacking of the TG ring structure at the membrane interface (24), determines the high affinity of the inhibitor for the ATPase.   Fig. 6 Characterization of chimera A was obtained from Sumbilla et al. (15), B from Norregaard et al. (1), and C, D, and E from the studies reported here. Diagram F was derived from Hua and Inesi (2