Differential Exposure of Surface Epitopes in the β-Strand Region of LOOP1 of the Yeast H+-ATPase during Catalysis*

The plasma membrane H+-ATPase of yeast assumes distinct conformational states during its catalytic cycle. To better understand structural changes in the LOOP1 domain, a catalytically important cytoplasmic loop segment linking transmembrane segments 2 and 3, surface epitopes were examined at different stages of catalysis. A polyclonal rabbit antibody was prepared to a fusion protein consisting of LOOP1 and the maltose binding protein. This antibody was affinity-purified to produce a LOOP1-specific fraction that could be used in competition enzyme-linked immunosorbent assays to assess surface exposure of the LOOP1 epitopes. It was found that in an E1 conformation stabilized with either adenosine 5′-(β,γ-imino)triphosphate (AMP-PNP) or ADP, less than 10% of the LOOP1 epitopes were accessible on native enzyme. However, when the enzyme was stabilized in an E2-state with ATP plus vanadate, approximately 40% of the surface epitopes on LOOP1 became accessible to antibody. The remaining 60% of the LOOP1 epitopes were fully occluded in the native enzyme and never showed surface exposure. Enzyme-linked immunosorbent assays utilizing fusion proteins consisting of LOOP1 subdomains demonstrated that all of the available epitopes were contained in the β-strand region (Glu-195—Val-267) of LOOP1. The epitopes that were differentially exposed during catalysis were included in regions upstream and downstream of the highly conserved TGES sequence. Our results suggest that during catalysis either the β-strand region of LOOP1 or an interacting domain undergoes substantial structural rearrangement that facilitates epitope exposure.

The plasma membrane H ϩ -ATPase from Saccharomyces cerevisiae is an electrogenic proton pump that couples ATP hydrolysis to the transport of H ϩ . It is essential to cell physiology through its action in regulating intracellular pH and in establishing the large electrochemical proton gradient needed for nutrient uptake. The H ϩ -ATPase belongs to the type II subclass of P-type ion translocating enzymes, which are found in bacteria, fungi, plants, and animals (1). All members share a number of characteristic catalytic features including the formation of a phosphorylated intermediate and the cycling between at least two distinct conformational states. A wide range of sequence diversity is observed between the various family members, although there is conservation of certain essential sequences and motifs (2). More importantly, recent structural studies suggest an extremely high degree of structural conservation, even between enzymes with relatively low sequence similarity such as the fungal H ϩ -ATPase and sarcoplasmic reticulum Ca 2ϩ -ATPase (3,4).
The H ϩ -ATPase is composed of a membrane transport domain consisting of 10 transmembrane segments, a cytoplasmically located ATP hydrolysis domain that composes ϳ60% of the total enzyme mass, and a slender stalk domain that links both domains. The catalytic ATP hydrolysis domain undergoes distinct conformational changes during catalysis that are critical to the hydrolytic and coupling reactions. The ATP binding domain and site of phosphorylation (Asp-378) are located on a large protein segment extending approximately 340 residues between TM4 and TM5. However, this segment alone is insufficient for catalysis (5) and requires interaction with addition enzyme components, including the cytoplasmic segment termed LOOP1, which extends ϳ135 residues (E162-L297) between TM2 and TM3 (6,7). How these domains are organized and how LOOP1 influences the large central phosphorylation domain during catalysis are not understood.
LOOP1 is predicted to consist of a central ␤-strand domain flanked at the N and C termini by ␣-helical stretches believed to form part of the stalk domain (8,9). Mutations in the ␤-strand region are known to alter conformational transitions (10,11), induce vanadate insensitivity (12,13), and alter phosphate binding (14). Mutations G183A and H285Q in the N-and C-terminal portions, respectively, of the ␣-helical stretches of LOOP1 are known to induce partial uncoupling of ATP hydrolysis from proton transport in the yeast H ϩ -ATPase (7,15). Finally, this region shows ion-dependent differential sensitivity to proteolytic cleavage in the Na ϩ ,K ϩ -ATPase (16).
Dynamic protein movements related to conformational rearrangement are an integral function of nearly all enzymes, but dissecting these movements is a difficult task. The reaction cycle of the P-type enzymes produce distinct conformational states, E1 and E2, that reflect ion binding or release and enzyme phosphorylation and dephosphorylation (17). In the fungal H ϩ -ATPases, these states can generally be assessed by examining the overall tryptic digestion pattern of the enzyme locked in a particular conformational state (18 -20). However, little is known about how specific regions of the catalytic domain change dynamically during catalysis and contribute to these distinct conformational states.
In general, it has been difficult to prepare antibodies to the LOOP1 domain (21). In this report, we use a fusion protein consisting of the maltose-binding protein and LOOP1 to generate polyclonal antibodies in rabbits with diverse epitopes on LOOP1. The anti-LOOP1 antibodies were then used in competition ELISA 1 assays to map changes in surface exposure of epitopes in the ␤-strand subdomain of LOOP1 at different stages of the catalytic cycle.

EXPERIMENTAL PROCEDURES
Yeast Strains and Cultures-The yeast strains utilized in this study were wild type Y55 (HO gal3 MAL1 SUC1) and a pma1 mutant strain pmal1-S368F (22,23). All yeast cultures were grown in YEPD medium (1% yeast extract, 2% peptone, and 2% dextrose, pH 5.7) at 22°C to mid-log phase (A 590 nm ϳ 3.) Construction of Maltose-binding Protein (MBP)-LOOP1 Fusion Proteins- Fig. 1 shows a schematic map of the construction of different fusion proteins between the MBP and LOOP1. The LOOP1 segment of PMA1 corresponding to amino acids Glu-162-Leu-97 (135 amino acid residues) was excised from plasmid pGW201 (7) with restriction endonucleases EcoRI and SpeI. The resulting 0.4-kilobase fragment was subcloned into the EcoRI and XbaI sites of commercial vector pMAL-c2 (New England BioLabs). The resulting vector, pMB100, contained the malE gene (encoding MBP) fused with LOOP1. Vector pMB100 was treated with endonuclease SalI to liberate the C-terminal half of LOOP1, and the linearized vector was ligated to produce a new vector, pMH100, which contained malE fused with the first half of LOOP1 (Glu-162-Val-237). Two SmaI sites were introduced into the LOOP1 region of pMB100 at positions 1549 bp (Glu-193) and 1758 bp (Val-263). The intervening fragment containing the ␤-strand region was excised after digestion with SmaI. The linearized vector was ligated to yield a new vector, pDSY100, that contained malE fused to the linked ␣-helical flanking ends of LOOP1. Finally, new restriction enzyme sites for EcoRI and HindIII were introduced into the LOOP1 region of pMB100 at nucleotide positions 1556 bp (Glu-195) and 1771 bp (Val-267), respectively. The ␤-strand-containing fragment was excised by digestion with EcoRI and HindIII and cloned into the equivalent site of pMAL-c2 to yield new vector pDSY101.
Expression and Isolation of the Fusion Proteins-The various fusion vector constructs were transformed into Escherichia coli strain XL2-Blue ultracompetent cells (Stratagene). Transformants were grown to early log phase in LB medium containing 0.2% glucose and 150 g/ml ampicillin at 37°C. Isopropyl-1-thio-␤-D-galactopyranoside (3 mM) was added to the medium to induce fusion protein expression, and the cells were grown for an additional 6 -7 h at 37°C (early stationary phase). Cells were harvested by centrifugation at 4000 ϫ g and washed by resuspension in column buffer consisting of 20 mM Tris-HCl, pH 7.4, 0.2 M NaCl, 1 mM EDTA, 10 mM mercaptoethanol, and 1 mM NaN 3 , and centrifuged as above. The washed cells were resuspended in 40 ml of column buffer containing 1 mM phenylmethylsulfonyl fluoride and lysed in a French pressure cell at 20,000 p.s.i. The lysate was centrifuged for 10 min at 4500 ϫ g to remove the cell debris. NH 4 SO 4 at 0.35 g/ml was added to the supernatant, and the suspension was stirred for 1 h at 4°C. The precipitate was centrifuged at 4,500 ϫ g for 10 min, and the pellets were resuspended with 5 ml of column buffer. Amylose resin (3 mg of MBP/ml of bed volume of amylose resin) was added to adsorb the MBP fusion protein by incubating at 22°C for 1 h. The bound amylose resin with MBP fusion protein was washed three times by resuspension with excess column buffer and brief centrifugation, as above. The bound fusion proteins were eluted twice in a 3-ml suspension with 100 mM maltose (final concentration). The resin suspension was then centrifuged at 4,500 ϫ g for 10 min. Supernatant containing the MBP fusion protein was centrifuged again to remove traces of residual amylose resin.
Preparation of LOOP1 Antibodies-The MBP-LOOP1 fusion proteins were evaluated for purity on a 10% SDS-polyacrylamide gel. A band corresponding to the MBP-LOOP1 fusion protein (M r ϭ 56,000) was excised from the gel and used to inoculate New Zealand White rabbits at Pocono Rabbit Farms. The resulting whole serum contained antibodies to both MBP, as the predominant antigen, and LOOP1. The contaminating antibodies to MBP were removed by affinity chromatography. A MBP affinity resin was prepared by coupling purified MBP (5 mg/ml) to Affi-Gel-10 resin, as described by the manufacturer (Bio-Rad) The coupled Affi-Gel complex was conditioned with 10 ml of 100 mM sodium acetate, pH 4.9, followed by three washes with 10 ml of antibody binding buffer (Pierce IgG Ab Purification Kit). Whole antiserum (4.5 ml) was added to excess resin along with 10 ml of antibody binding buffer, and the suspension was incubated 1 h at 37°C. The remaining serum containing unbound LOOP1 antibody was collected, and the IgG fraction was concentrated on a protein A column (Pierce). Antibodies to LOOP1 were further affinity-purified, as described by Seto-Young et al. (24). The LOOP1 antibodies were shown by Western blot analysis to bind to both purified LOOP1 and whole H ϩ -ATPase but not to purified maltose binding protein.
ELISA Competition Assay-Purified LOOP1 fusion antigens (1 pmol) in 50 l of 94 mM NaCO 3 buffer, pH 9.8, and 1 mM NaN 3 were added to wells of a microtiter plate (Nunc-Immuno Plate-MaxiSorp Tm ) and incubated at 4°C for 16 h. The plate was washed several times with TTBS buffer consisting of 10 mM Tris, pH 7.5, 0.05% Tween 20, and 150 mM NaCl. A blocking solution consisting of 100 l of TTBS plus 3% bovine serum albumin was added to each well, and the plate was incubated for 1 h at 22°C. Antigen competition was performed in a 100-l volume consisting of 10 mM Tris-HCl, pH 7, 40 mM NaCl, 1 mM NH 4 Cl, 0.5 mM NaN 3 , and 3% bovine serum albumin. Additions of deoxycholate-enriched H ϩ -ATPase (0 -37.5 pmol), affinity-purified LOOP1 antibody, and different hydrolysis substrates were made as indicated in the text. The reaction was incubated at the 22°C for 1 h, and the wells were washed three times with TTBS. A 100-l suspension of TTBS and 3% bovine serum albumin containing goat anti-rabbit IgG conjugated with alkaline-phosphatase (1/300 dilution) was added to each well and incubated for 1 h at 22°C. The wells were washed, as described above. LOOP1 affinity antibody bound to the wells was detected by colorimetric assay using Sigma 104-phosphatase substrate (Sigma), as described by the manufacturer. The -nitrophosphate formation was monitored in a SLT Spectra ELISA microplate reader at 405 nm.
Other Procedures-Deoxycholate-enriched H ϩ -ATPase was prepared by sequential extraction of plasma membranes with deoxycholate and washing with KCl medium, as described previously (24). SDS-polyacrylamide gel electrophoresis and semi-dry electroblotting of LOOP1 and the sub-domain fusion proteins were performed as described previously (24). Protein concentration was determined by a modified Lowry method (25).

RESULTS
A fusion protein consisting of the LOOP1 region (Glu-162-Leu-297) of PMA1 fused to the C-terminal end of the MBP was purified and used to raise polyclonal antibodies in rabbits. Antibodies specific to the LOOP1 domain were purified by affinity chromatography, as described under "Experimental Procedures." The purified anti-LOOP1 antibodies were shown by Western blotting to recognize only the LOOP1 domain of the fusion protein; they did not cross-react with MBP. The antibodies cross-reacted strongly with intact H ϩ -ATPase in both Western blots and ELISA assays, which enabled them to be used as probes of the LOOP1 domain in the native enzyme.
The catalytic reaction cycle of the typical P-type enzymes is characterized by the formation of distinct conformational states, which are known to be manifested by changes in the large cytoplasmic domain (16, 26 -28). A competition ELISA assay was developed to examine epitope exposure of the LOOP1 domain at distinct stages of catalysis. In this assay, LOOP1 fusion protein is bound to the well of a microtiter plate. Antibody is added at fixed amount in free solution along with increasing amounts of a challenge protein. Exposed epitopes on the challenge protein compete with epitopes on the bound LOOP1 protein for antibody. The advantage of this system is that it is extremely sensitive and allows the native, catalytically active H ϩ -ATPase to be evaluated as a challenge protein.
Most importantly, since the enzyme can be locked into defined conformations by altering substrate availability or by the addition of inhibitors, it is possible to correlate changes in epitope exposure within LOOP1 with distinct stages of the catalytic reaction cycle. Finally, by altering the LOOP1 portion of the fusion protein bound in the competition assay, it is also possible to dissect the regions of the LOOP1 that express the changing epitopes. Fig. 2 shows a typical competition assay in which bound LOOP1 was challenged with increasing amounts of itself. It can be seen that Ͻ1 pmol of LOOP1 fusion protein was required to give ϳ90% competition with bound LOOP1. Complete competition was observed at higher levels of challenge protein (not shown). When purified H ϩ -ATPase was used as the challenge protein, it was found that the enzyme competed with bound LOOP1, but the extent of competition was highly dependent on the conformational state of the enzyme. In the presence of 2 mM AMP-PNP and 2 mM EDTA, the enzyme is known to adopt an E 1 conformational state (19). Fig. 3 indicates that under these conditions the H ϩ -ATPase competed poorly with bound LOOP1, showing a saturatable competition of ϳ10% at the highest level. Comparable results were obtained in the presence of 7.5 mM ADP (data not shown). In contrast, when the enzyme was locked in an E 2 conformational state by incubation with 7.5 MgATP and 200 M VO 4 , a classical transition state inhibitor of P-type enzymes, the H ϩ -ATPase competed much more effectively with bound LOOP1, typically showing about 40% competition. MgATP alone also promoted competition but not as effectively as that observed in the presence of vanadate (not shown). This latter result was complicated by the fact that extensive hydrolysis of substrate occurred during the incubation period. The differential effects observed were absolutely dependent on active enzyme, since there was no difference in competition observed when LOOP1 was challenged against itself in the presence of the various conditions that favor either E 1 or E 2 . In addition, a S368F mutant enzyme, which is ϳ500fold less sensitive to vanadate (29), weakly competed for epitopes on bound LOOP1 relative to wild type enzyme, since it assumes a steady-state conformation that disfavors VO 4 interaction. However, in the presence of acetyl phosphate, which has been shown to promote an E 2 state (18) and to increase the sensitivity of the S368F enzyme to VO 4 (18), the enzyme showed enhanced competition in the presence of VO 4 relative to that observed in ATP-containing medium (Fig. 4). These results support the notion that the E 2 conformational state of the enzyme exposes more epitopes in the LOOP1 than the E 1 conformational state. These experiments further indicate that only ϳ40% of the epitopes present on the LOOP1 fusion protein are available for interaction in native enzyme. The remaining 60% are most likely occluded within the 3-dimensional structure of LOOP1. The behavior observed for enzyme in the E 1 and E 2 states indicates that a substantial fraction of the surfaceaccessible epitopes are differentially exposed during the catalytic cycle.
The LOOP1 segment could be dissected to explore more precisely the location of the epitopes involved in conformationspecific binding. The fusion malE gene constructs with intact LOOP1, and its sub-domains are shown in Fig. 1 (see "Experimental Procedures"). The four constructs include the full LOOP1 (Glu-162-Leu-297), the N-terminal half of LOOP1 (Glu-162-Ala-235), the central ␤-strand domain (Glu-195-Val-267), and the ␣-helical flanking region (Glu-163-Asn-193/Gly-263-Leu-297). All of the fusion proteins except the ␤-strand domain-containing fragment were expressed at high levels and could be purified to near homogeneity. Special expression conditions involving early log phase cells and short induction times were required to yield a suitable product for the ␤-strand protein, which even under the best circumstances showed approx- imately 50% proteolysis with numerous fragments in Western blots. Fig. 5 shows interaction of the whole anti-LOOP1 antibody with the various fusion proteins in a standard ELISA-type assay. It can be seen that antibody bound strongly to the full LOOP1 and the N-terminal half-LOOP1. The ␤-strand-containing fragment showed high affinity binding at low protein levels, comparable with the other peptides but rapidly saturated at about 35% of the full LOOP1 level due to proteolysis. In contrast, the ␣-helical flanking region-containing peptide show negligible interaction with the antibody, making it unlikely that significant anti-LOOP1 epitopes were present in this region of the protein. This result was confirmed by Western blot analysis, which clearly showed a lack of binding of the anti-LOOP1 antibodies to this peptide. These results suggest that all of the epitopes present in the anti-LOOP1 antibody are present in the 72-amino acid ␤-strand domain. The fact that this region was difficult to express and purify in intact form may indicate that the flanking ␣-helical regions stabilize the overall peptide structure. A destabilized protein is likely to be more susceptible to endogenous cell proteases.
ELISA competition assays were performed using fusion proteins containing the N-terminal half-LOOP1, ␤-strand domain, and ␣-helical regions. Fig. 6 shows that the N-terminal half-LOOP1 showed the same type of differential effect with AMP-PNP and MgATP plus vanadate as the full LOOP1. It should be noted, however, that the extent of competition was about onehalf of the full LOOP1, suggesting that epitope accessibility on LOOP1 may be evenly distributed along the ␤-strand domain. As expected, the partially degraded ␤-strand protein showed only a weak competition effect (not shown). The ␣-helical-containing peptide showed no competition, since antibody binding was barely detectable above base line. DISCUSSION Dynamic movements within P-type ATPases are an integral part of the overall catalytic reaction of these enzymes. Numerous studies have defined global changes in enzyme form that are associated with defined conformational states (28,30). Yet, relatively few studies have dealt with localized changes in protein structure that occur during catalysis (16, 27, 31). An understanding of the partial reactions for nucleotide binding, phosphorylation-dephosphorylation, ion translocation and energy coupling requires an accounting of local dynamic events. The change in electron density observed in the apparent ATP binding groove on the Ca 2ϩ -ATPase in the presence and absence of nucleotide pointedly illustrates the need to understand such local changes (32). In this study, we have used antibodies FIG. 4. Differential exposure of epitopes in a VO 4 -insensitive mutant enzyme from pma1-S368F. A competition ELISA assay identical to that shown in Fig. 3 was used to evaluate the differential exposure of LOOP1 epitopes in enzyme from wild type and pma1-S368F in the presence of either ACP or ATP. In this assay, competition with free H ϩ -ATPase, either wild type (‚, Ⅺ) or mutant (E, छ) was performed in the presence of either 2 mM AMP-PNP ϩ 2 mM EDTA (Ⅺ), 7. to the LOOP1 domain, which is bounded by TM2 and TM3, in a competition ELISA assay to examine surface-exposed epitopes that change during catalysis. It was found that epitopes in a 72-amino acid region from Glu-195 to Val-267, comprising the ␤-strand domain and including the highly conserved signature sequences VPGDIL and TGES (1), were differentially exposed at distinct stages of the catalytic cycle. In an E 1 conformation stabilized with AMP-PNP and EDTA or ADP, approximately 10% of the epitopes were exposed, indicating that this region was not highly surface-exposed. However, in an E 2 state formed with MgATP and vanadate, approximately 40% of the total sites became available, suggesting a major change in this region resulting in surface exposure of epitopes. These results are consistent with a report by Serrano et al. (33) that a monoclonal antibody recognizing an epitope in this region of the yeast H ϩ -ATPase could only react with the native enzyme (in an E 1 state) after detergent treatment. Furthermore, they are qualitatively similar to the tryptic proteolysis study of Lutsenko and Kaplan (16) on the Na ϩ ,K ϩ -ATPase in which the LOOP1 segment was involved in structural rearrangements upon phosphorylation or ion binding.
Movements in the ␤-strand region during catalysis are supported by several studies. In the Ca 2ϩ -ATPase, it is known that there is conformation-specific exposure of  and  in LOOP1, as revealed by proteolysis with trypsin and V8 protease, respectively. The role of the LOOP1 domain in catalysis has largely been inferred from the affects of perturbations in this region. It is known that mutations in the LOOP1 ␤-strand region can alter the progression of catalysis by blocking the interconversion of catalytic intermediates during E 1 P-E 2 -P transition. In the yeast H ϩ -ATPase, scanning alanine mutagenesis of conserved motifs in this domain revealed that overall enzyme function was highly sensitive to mutations in this region. Mutations in LOOP1 of the H ϩ -ATPase also yield vanadate insensitivity (7,13,15), which has been linked to changes in phosphate binding (14). In the flanking ␣-helical regions of LOOP1, mutations I183A and H285Q resulted in partially uncoupled enzymes, suggesting that the LOOP1 participates in the overall coupling reactions (7,15). These regions have been proposed to form a bundle-type organization, largely because targeted factor Xa proteolysis of a site around residue 275 had no effect on coupled proton transport, suggesting strong intramolecular interactions in this region (13).
In general, it has been difficult to prepare antibodies to the LOOP1 region (34). Surface exposure of an epitope around Arg-198 (T2 site) in the Ca 2ϩ -ATPase has been reported, and this site shows some latent exposure upon denaturation of the enzyme (21). The flexible proline-rich region containing the T2 site in Ca 2ϩ -ATPases is absent in PMA1-type enzymes (2). However, the TGES sequence just upstream is conserved. We show in competition ELISA assays that in an E 2 state approximately 40% of the epitopes present in the ␤-strand region of LOOP1, which includes the TGES sequence, are surface-exposed (Fig. 3). However, surface exposure of the ␤-strand domain is not localized to the region around TGES, because there was only partially differential epitope exposure observed with a fusion protein containing the N-terminal portion of the ␤-strand up to Val-237 (Fig. 6). Thus, a region downstream of Val-237 must also contain differentially exposed epitopes. It appears likely that during catalysis, there is a substantial exposure of the ␤-strand portion of LOOP1, and this is not restricted to a small region around TGES.
Our data suggest that in the E 1 state, with the nucleotide bound, the enzyme assumes a state in which most of the ␤-strand domain is shielded from surface exposure. However, after proton binding and phosphorylation, the LOOP1 struc-ture is altered, and up to 40% of the epitopes in the ␤-strand domain are exposed. This notion for localized conformational breathing of LOOP1 is fully consistent with a recent model proposed by Lutsenko and Kaplan (16) for the Na ϩ ,K ϩ -ATPase, based on a limited tryptic digestion study. It was found that in the presence of ADP or ATP the enzyme forms a compact, protease-resistant form, but when the enzyme is phosphorylated, it forms a less compact protease-sensitive form. It was proposed that in the presence of bound ATP, there is a tight interaction between the ATP binding domain and the membrane resulting in a compact structure. When transported ions bind, these interactions are relaxed, and upon phosphorylation, the cytoplasmic domain moves away from the membrane, resulting in a less compact molecule. After de-phosphorylation and ion release, the interactions increase, and the enzyme again forms a more compact structure.
Interestingly, an alternative model was recently proposed after an analysis of metal-catalyzed cleavage of the Na ϩ ,K ϩ -ATPase at different stages of the catalytic cycle. It was suggested that there is an interaction between LOOP1 and the large central domain, and these domains interact strongly in an E 2 state but move apart in an E 1 state. The implication is that the E 2 -stabilized enzyme forms a tightly packed structure. Conformation-dependent cleavage at and around the conserved TGES sequence in this study and in previous studies using limited tryptic digestion (26,35), is consistent with such a suggestion.
Apparent differences between these models may reflect limitations of the probing mechanisms. In our case, only surface exposure of the ␤-strand domain can be inferred given the large size of the antibody probe. A similar conclusion can also be drawn about limited proteolysis studies with trypsin, given its relative size. On the other hand, Fe 2ϩ -catalyzed peptide bond cleavage around a metal binding center should provide excellent information about closely opposed protein structure elements, provided that the reactive chemical species remain localized. To further reconcile the various models, it is possible that movements that expose one portion of the molecule could actually bring other portions closer together in a scissors-like mechanism. As shown in Fig. 7, the LOOP1 and major loop domains could be viewed as a splayed scissors close to the membrane. In the E 1 state, the ␤-strand sector would be occluded because of close contact with the membrane surface, and the metal binding center would be loosely organized. In the E 2 state, the domains would interact like a closing scissors, moving away from the membrane and thus exposing the epitopes on LOOP1. The metal binding center would become more organized, as the two domains interact strongly. Thus, surface sites in the ␤-strand domain would become more exposed in the more organized but less compact E 2 state. In contrast, the E 1 state would appear more compact but less organized around the metal binding center.
In conclusion, our data suggest that the region around the highly conserved TGES sequence in the ␤-strand region of LOOP1 undergoes significant changes in solvent exposure during catalysis. This finding is consistent with recent models for interactions within this domain in which an interacting domain, most likely the large central domain, changes its position relative to LOOP1.