HOME HELP FEEDBACK SUBSCRIPTIONS ARCHIVE SEARCH TABLE OF CONTENTS		QUICK SEARCH:   [advanced] Author: Keyword(s): Year:  Vol:  Page:

J. Biol. Chem., Vol. 275, Issue 25, 18623-18637, June 23, 2000

This Article Abstract Full Text (PDF) All Versions of this Article: 275/25/18623    most recent M000470200v1 Alert me when this article is cited Alert me if a correction is posted Citation Map Services Email this article to a friend Similar articles in this journal Similar articles in PubMed Alert me to new issues of the journal Download to citation manager Citing Articles Citing Articles via HighWire Citing Articles via Google Scholar Google Scholar Articles by Champeil, P. Articles by le Maire, M. Search for Related Content PubMed PubMed Citation Articles by Champeil, P. Articles by le Maire, M. Social Bookmarking                      What's this?

Interaction of Amphipols with Sarcoplasmic Reticulum Ca²⁺-ATPase^*

Philippe Champeil^§, Thierry Menguy, Christophe Tribet^¶, Jean-Luc Popot, and Marc le Maire

From the  Unité de Recherche Associée 2096 (CNRS et CEA) and Section de Biophysique des Protéines et des Membranes, Département de Biologie Cellulaire et Moléculaire, Commissariat à l'Energie Atomique Saclay, 91191 Gif-sur-Yvette Cedex, the ^¶ Unité Mixte de Recherche 7615 (CNRS, Université Paris VI & Ecole Supérieure de Physique et Chimie Industrielles de la Ville de Paris), 10 rue Vauquelin, 75231 Paris Cedex 05, and  Unité Propre de Recherche 9052 (CNRS), Institut de Biologie Physico-Chimique, 11 rue Pierre et Marie Curie, 75005 Paris, France

Received for publication, January 21, 2000, and in revised form, February 28, 2000

	ABSTRACT

TOP ABSTRACT INTRODUCTION EXPERIMENTAL PROCEDURES RESULTS DISCUSSION REFERENCES

Amphipols are short-chain amphipathic polymers designed to keep membrane proteins soluble in aqueous solutions. We have evaluated the effects of the interaction of amphipols with sarcoplasmic reticulum Ca²⁺-ATPase either in a membrane-bound or a soluble form. If the addition of amphipols to detergent-solubilized ATPase was followed by removal of detergent, soluble complexes formed, but these complexes retained poor ATPase activity, were not very stable upon long incubation periods, and at high concentrations they experienced aggregation. Nevertheless, adding excess detergent to diluted detergent-free ATPase-amphipol complexes incubated for short periods immediately restored full activity to these complexes, showing that amphipols had protected solubilized ATPase from the rapid and irreversible inactivation that otherwise follows detergent removal. Amphipols also protected solubilized ATPase from the rapid and irreversible inactivation observed in detergent solutions if the ATPase Ca²⁺ binding sites remain vacant. Moreover, in the presence of Ca²⁺, amphipol/detergent mixtures stabilized concentrated ATPase against inactivation and aggregation, whether in the presence or absence of lipids, for much longer periods of time (days) than detergent alone. Our observations suggest that mixtures of amphipols and detergents are promising media for handling solubilized Ca²⁺-ATPase under conditions that would otherwise lead to its irreversible denaturation and/or aggregation.

	INTRODUCTION

TOP ABSTRACT INTRODUCTION EXPERIMENTAL PROCEDURES RESULTS DISCUSSION REFERENCES

The transmembrane surface of integral membrane proteins is highly hydrophobic. Thus, detergents are generally used to facilitate the handling of these proteins in aqueous media (1). However, detergents do not always maintain solubilized membrane proteins in an active and stable state (see Refs. 2 and 3, and references therein). The use of short amphipathic polymers ("amphipols"), comprising a hydrophilic backbone with hydrophobic chains, has been proposed as an alternative approach (4). Amphipols can interact with hydrophobic surfaces at many points along their length. They should therefore form a noncovalent but stable layer at the interface between the transmembrane region of the protein and the aqueous solution. Initial studies with four integral membrane proteins showed that complexes of these proteins with amphipols could indeed be handled in surfactant-free aqueous solutions as if they were soluble proteins (4-6); back-exchange between amphipols and detergents was also demonstrated (5), but its kinetics was not studied. Amphipols have a wide range of potential applications in biochemistry, biophysics, and biotechnology. Little is known, however, about their effects on membrane protein function. Published data are limited to cytochrome b₆f, which was shown to retain its ability to catalyze electron transfer following injection of cytochrome b₆f-amphipol complexes into a detergent-containing reaction medium (4). Various degrees of inhibition were nevertheless observed, depending on the charge of the amphipols. This inhibition was hypothesized to be due to the electrostatic repulsion of one of the substrates of the reaction, but other effects of amphipols on function, an essential determinant of their usefulness in membrane biology, remain largely unknown.

In this study, we investigated the interaction of amphipols (mainly amphipol A8-35, a polyacrylate backbone derivatized with octyl and isopropyl chains) with the sarcoplasmic reticulum (SR)¹ Ca²⁺-ATPase from rabbit skeletal muscle. This enzyme, a P-type ATPase, is responsible for active calcium transport from the cytosol to the lumen of the sarcoplasmic reticulum (7-9). In addition to its transmembrane region, it has a bulky cytosolic region that is critical for catalytic activity. The interactions of Ca²⁺-ATPase with detergents have been extensively characterized (2, 10, 11). The ATPase is highly sensitive to the nature of its hydrophobic environment; it may aggregate, and changes in its aggregation state may be associated with a reversible or irreversible loss of protein activity. Thus, Ca²⁺-ATPase is an appropriate model for studying the functional consequences of protein dispersion using amphipols. We first briefly characterized the interaction of amphipols with intact SR membranes. We then studied the effects of the association of amphipols with solubilized ATPase, with special emphasis on the complexes formed upon subsequent removal of detergent. In some respects, functional and structural properties of these complexes were not up to our expectations, particularly as regards monodispersity of the complexes. However, we also observed that, in the presence of both amphipol and detergent, the ATPase was much more stable than when it was exposed to either of the surfactants alone (in the presence or absence of calcium), and it remained monomeric for longer periods of time. These synergistic effects provide evidence for the rapid formation of a mixed detergent/amphipol layer at the hydrophobic surface of the protein. The remarkable protective effect of amphipols may be of particular value in studies of membrane protein states prone to rapid denaturation.

	EXPERIMENTAL PROCEDURES

TOP ABSTRACT INTRODUCTION EXPERIMENTAL PROCEDURES RESULTS DISCUSSION REFERENCES

A general description of amphipol synthesis and use has already been published (4, 12). Most experiments in this work were performed using amphipol A8-35 (4). A8-35 (Scheme 1) is an anionic polymer with a mean apparent molecular mass of 8 kDa, prepared by forming amide bonds between hydrophobic amines and the carboxyl groups of polyacrylic acid precursors, with about 35% (mol/mol) of the carboxyl groups in the sodium carboxylate form. On average, 24 carboxylic groups in the polymer molecule are free, 18 have reacted with octylamine, and the remaining 28 have reacted with isopropylamine (Scheme 1). These values are means for a polydisperse population of molecules, some of which have chains with more or fewer acrylic acid units than the mean value of 70 determined by gel filtration of the polyacrylic precursor (4). The A8-35 stock solution was 50 mg/ml in water. The nonionic detergents C₁₂E₈ and DM were obtained from Nikko and Calbiochem (or Anatrace), respectively. Their critical micellar concentrations and approximate aggregation numbers are about 0.05 mg/ml (0.09 mM) and 100-120, and 0.09 mg/ml (0.18 mM) and 110-140, respectively (23). The dispersion state of amphipol polymers in buffer has not yet been described.

View larger version (12K): [in this window] [in a new window]   Scheme 1.   Amphipol A8-35.

SR vesicles were prepared as described previously (13). Deoxycholate (DOC)-purified membrane-bound ATPase was a gift from Prof. J. V. Møller and was obtained by treating SR vesicles with a low concentration of deoxycholate (16). Unilamellar liposomes were a gift from A. Bluzat and were prepared from egg yolk phosphatidylcholine and phosphatidic acid (9/1, w/w) by reverse phase evaporation (14). For several of the experiments described below, the aqueous medium contained 100 mM KCl, 1 mM Mg²⁺, 0.1 mM Ca²⁺, and 50 mM TES-Tris at pH 7.5 (20 °C). This medium is referred to as the "TES 7.5 medium." In particular, it was used to assess the solubilization of SR membranes and liposomes, from light-scattering measurements (Refs. 11, 14, and 15, and Fig. 1C), and the hydrolytic activity of the SR Ca²⁺-ATPase. The elution buffer for HPLC was slightly different, and contained 100 mM KCl, 1 mM Mg²⁺, 0.5 mM Ca²⁺, and 20 mM TES-NaOH at pH 7. It is referred to as the "TES 7.0" medium. This medium was used as the initial solubilization medium in all experiments in which the ATPase was subsequently delipidated.

The hydrolytic activity of the SR Ca²⁺-ATPase was determined at 20 °C with a coupled enzyme assay (2, 17), in 2 ml of TES 7.5 medium supplemented with 5 mM MgATP, 0.05 mg/ml pyruvate kinase, 1 mM PEP, 0.05 mg/ml lactate dehydrogenase, and an initial concentration of about 0.3 mM NADH. The hydrolytic activity of membrane-bound ATPase (Fig. 1A) was assayed in the presence or absence of the ionophore A23187 (Calbiochem). The hydrolytic activity of solubilized ATPase was assayed in the presence of various concentrations of C₁₂E₈ (Fig. 2) or DM (Fig. 4). Excess C₁₂E₈ (1 mg/ml final concentration) was also used to determine the proportion of ATPase molecules that had escaped irreversible inactivation during preliminary incubation (Figs. 3, 11, 12, and 14; see also Ref. 2). In all cases, the final SR protein concentration during the ATPase assay was 0.002 mg/ml. At this low concentration of protein, the final concentrations of DM and amphipol resulting from dilution of the initial samples are very low, and should not themselves affect ATPase activity significantly, either in detergent-free assay medium or in the presence of 1 mg/ml C₁₂E₈. For all figures, the low (about 0.2 µmol·mg¹·min¹) Ca²⁺-independent ATPase activity observed in the presence of excess EGTA was subtracted.

Ca²⁺ uptake into SR vesicles and its release after the exhaustion of ATP was measured with the calcium-sensitive dye Arsenazo III, essentially as described previously (18). The medium used contained 100 mM KCl, 10 mM Mg²⁺, 5 mM P_i, 0.1 mg/ml pyruvate kinase, and 50 mM MOPS-Tris at pH 6.7. Ca²⁺ (three sequential additions of 20 µM), PEP (0.04 or 0.2 mM), and MgATP (0.04 mM) were added at the beginning of each experiment (see Fig. 1B). The experiment was performed at 25 °C, and SR vesicles were present at 0.4 mg/ml. Similar experiments were performed using the standard TES 7.5 buffer. They gave similar results, but less Ca²⁺ was taken up by SR vesicles at the lower Mg²⁺ concentration and higher pH of the TES 7.5 buffer. Fig. 1B therefore shows the data obtained at pH 6.7 and in the presence of 10 mM Mg²⁺. These ionic conditions correspond to those of the original transport assay (18), except that we used MOPS-Tris rather than Tris-maleate, to eliminate possible maleate-specific effects (19, 20), and we included 5 mM phosphate to provide a membrane-permeant calcium-complexing agent. A sharp transition between the end of the Ca²⁺ uptake phase and the beginning of the Ca²⁺ release phase (after ATP and PEP exhaustion) was observed even in the absence of an adenylate kinase inhibitor, suggesting that contamination with adenylate kinase (18) was negligible for our membranes.

Light scattering and fluorescence measurements were performed with a SPEX fluorolog instrument. Light scattering was measured before and after the addition of amphipol or detergent, at 90°, with excitation and emission wavelengths both set at 546 nm. Intrinsic (tryptophan) fluorescence was measured with excitation and emission wavelengths of 290 and 330 nm, respectively (bandwidths were 5 and 10 nm, respectively). A few measurements were also made with an extrinsic probe, fluorescein 5'-isothiocyanate covalently bound to the ATPase (see Refs. 21, 52, and 53). Absorbance measurements (either at a single wavelength, for NADH, or at multiple wavelengths, for Arsenazo III and Murexide) were made with a diode array HP 8453 spectrophotometer. In some cases, turbidity changes were recorded with this spectrophotometer (e.g. Fig. 10). In all cases, samples were stirred continuously in the thermostat-regulated cuvette. Arsenazo III and Murexide were obtained from Sigma.

In many experiments, amphipol was incubated with solubilized ATPase. The detergent-solubilized ATPase was prepared either together with its endogenous lipids, by simple detergent-induced solubilization of SR or DOC-purified membranes, or in an essentially delipidated form, by preparative size exclusion chromatography in the presence of DM of a solubilized concentrated sample (23).

For experiments performed in the presence of endogenous SR lipids, the protocol generally involved preliminary solubilization of SR vesicles with dodecylmaltoside (DM) or C₁₂E₈ (2 mg/ml SR protein and 10 mg/ml detergent in TES 7.5 buffer at room temperature) followed by 2-fold dilution in the presence or absence of amphipol, so that the final samples contained 1 mg/ml SR protein and 5 mg/ml detergent, with or without 5 mg/ml of amphipol. The resulting SR/detergent/amphipol mixture was used after equilibration for 15-30 min. For the experiment illustrated in Fig. 5 (A and B), DOC-purified membrane-bound ATPase was used as the starting material, rather than SR vesicles (although SR vesicles gave similar results). This purified ATPase was treated with DM (2 mg/ml protein and 10 mg/ml DM) and was then centrifuged (Beckman TL100, 200-µl aliquots in a TLA100 rotor, 55,000 rpm for 6 min at 8 °C) to remove nonsolubilized material (which accounted for a very small proportion of the original protein content).

For size exclusion HPLC in the presence or absence of detergent, 200-µl samples were loaded onto a TSK 3,000 SW silica gel column (7.5 mm × 30 cm, i.e. about 11.5 ml total volume), and subjected to chromatography (22, 23) using a Beckman Gold system with diode array detection (1-cm optical path length). The elution buffer was TES 7.0 medium, to protect the gel phase, containing 0.5 mM Ca²⁺ (to stabilize the ATPase), in the presence or absence of 1 mg/ml DM (20 °C). For calibration (K_d versus R_S, Ref. 24), we used Bio-Rad (catalog no. 151-1901) or Amersham Pharmacia Biotech (catalog no. 17-0441-01) gel filtration standard proteins, to which DM and amphipol were added if required. In some cases, we also used an elution buffer of greater ionic strength (250 mM NaCl, 0.1 mM EDTA, and 20 mM TES-Tris at pH 7), with similar results both for ATPase complexes and for aggregates of pure A8-35. We used two different columns of the same type in this work, one for the early experiments illustrated in Fig. 5, and the other for all other experiments.

For experiments performed in the absence of most of the endogenous SR lipids, delipidated ATPase was prepared by preliminary preparative chromatography (see Fig. 1A of Ref. 23 for description of a similar experiment), starting from a highly concentrated detergent-solubilized SR sample, and using size exclusion chromatography in the presence of detergent to separate mixed lipid/detergent micelles from an essentially delipidated (and monomeric) ATPase·DM complex (note, however, that this HPLC-"delipidated" ATPase is known to retain some of the original SR lipids: 10-12 mol of lipid (see Ref. 23) out of the 100 mol of lipid/mol of ATPase in native SR membranes). In this case, SR (10 mg/ml protein) was solubilized using 100 mg/ml DM in TES 7.0 medium. The mixture was centrifuged (as described above for DOC-ATPase) to remove nonsolubilized material. 200-µl aliquots of supernatant were applied to the TSK column and eluted with 1 mg/ml DM. The peak fractions of eluted ATPase (data not shown) were pooled, generally resulting in a total of about 0.6 ml (per 200 µl of injected sample) of a solution of essentially delipidated and monomeric ATPase, with a protein concentration of 1.2-1.8 mg/ml. As delipidated monomer binds 0.9 g of DM/g of protein (23), the total DM concentration in the pooled sample was therefore (1 + 0.9[protein]) mg/ml. Aliquots of this solution were incubated at the desired protein and detergent concentrations, with or without amphipol A8-35, before subsequent treatment (ATPase assay, detergent removal, or a second chromatography step with or without dilution).

During HPLC experiments like the one illustrated in Fig. 10A, we incidentally found that if amphipol in buffer was injected onto a column that had been extensively washed with detergent before reequilibration with buffer, a significant amount of the injected amphipol was retained by the column (second and third injections gave slightly higher amplitudes than the first injection; data not shown). TSK columns thus bind a small amount of amphipol (about 0.1 mg of A8-35/7.5 mm × 30-cm column), as they do with DM and C₁₂E₈ (23). This amphipol can be washed off with excess detergent. In experiments with detergent-free buffer, these amphipol-binding sites were therefore first saturated by preliminary chromatography of amphipol alone in buffer.

Centricon-100 filters (nominal cut-off, 100 kDa; Amicon) were used to reconcentrate ATPase·A8-35 samples that had been diluted: thus, 2-ml aliquots of dilute samples (containing 0.02 mg/ml protein) were centrifuged in these devices (at 2400 rpm for 40 min in a Beckman JA12 rotor), and the resulting concentrated droplets (about 40 µl at 1 mg/ml protein, as shown by UV absorption) were finally rediluted to a concentration of 0.4 mg/ml protein. We checked that Centricon-100 filters did not retain DM monomers and retained only a small proportion of pure DM micelles (data not shown; the protocol was similar to that described in Ref. 25 for Centricon-10 filters).

Detergent was removed from samples with high and constant protein concentrations by treatment with 100 mg/ml of methanol-washed (26) Bio-Beads SM-2 (Bio-Rad). The rate and extent of this removal was followed with [¹⁴C]DM (27) (1.5-2 µCi/ml) by counting 4-µl bead-free aliquots. We checked that protein losses during this Bio-Bead treatment were no more than a few percent.

Analytical sedimentation (velocity and equilibrium) experiments were performed at 20 °C with a Beckman Optima XL-A analytical ultracentrifuge (AN60 Ti rotor). 100-µl samples (about 0.5 mg/ml protein) in TES 7.0 buffer, with or without DM and A8-35, were placed in double-sector cells (1.2-cm optical path length), and absorption was scanned thoughout the cell (generally at 280 nm, against the solvent). For sedimentation velocity experiments, the rotor speed was 50,000 rpm and scans were taken every 6 min. Data were analyzed by two methods, assuming either a discrete number of independent species (28), or a continuum of species with different sedimentation coefficients (29). The results of the two types of analysis were consistent, and thus only the results for the latter are given in Fig. 9. For sedimentation equilibrium experiments, the rotor speed was 7,000 or 6,000 rpm; the concentration gradient was assessed after various periods, from 16 to 43 h, to check for equilibrium.

Protein-protein contacts in ATPase·DM and ATPase·A8-35 complexes were detected by incubating samples with 40 mM glutaraldehyde for 4 min at 25 °C (30). The reaction was quenched by adding an equal volume of 200 mM hydrazine, and covalent oligomers were detected either by SDS-PAGE in a 7% Laemmli gel or by HPLC in the presence of 1 mg/ml DM. Cross-linking with dithiobis(succinimidyl propionate) (31) was also attempted, and similar results were obtained (data not shown).

A few experiments were performed with soluble polypeptides, either standard proteins (e.g. bovine serum albumin, Sigma A6003, and lysozyme, Sigma L6876) or a 29-kDa compact subdomain of the ATPase cytosolic "head" (32) kindly provided by P. Falson.

	RESULTS

TOP ABSTRACT INTRODUCTION EXPERIMENTAL PROCEDURES RESULTS DISCUSSION REFERENCES

Effects of Amphipols on ATPase Function: Apparent Competition with Detergents

Amphipol A8-35 Permeabilizes SR Vesicles and Affects the Function of Membrane-bound Ca²⁺-ATPase, but Without Inducing Membrane Solubilization-- Before studying in detail the interaction of amphipols with solubilized ATPase, we first characterized the interaction of amphipol A8-35 with Ca²⁺-ATPase in its native membrane environment. If intact SR vesicles were used, low concentrations of amphipol A8-35 were found to stimulate steady-state ATPase activity, whereas higher concentrations were inhibitory. In contrast, if the inhibitory constraints imposed upon Ca²⁺-ATPase by accumulated Ca²⁺ were released by making vesicles permeable to Ca²⁺, amphipol had only an inhibitory effect (Fig. 1A). In both cases, the ATPase regained significant activity if subsequently solubilized with 1 mg/ml C₁₂E₈ (data not shown). These features are reminiscent of the perturbation exerted by various detergents under nonsolubilizing conditions (11, 33-35). We also found that the rate of hydrolysis of paranitrophenyl phosphate (another well known substrate of Ca²⁺-ATPase) by Ca²⁺-ATPase in leaky SR vesicles was reduced in the presence of amphipol A8-35 (data not shown), as observed for detergents (35-37).

View larger version (27K): [in this window] [in a new window]   Fig. 1.   Interaction of amphipol A8-35 with SR membranes: perturbation of ATPase activity, permeabilization to Ca2+, and lack of solubilization. Panel A, the hydrolytic activity of SR Ca2+-ATPase (0.002 mg/ml protein) was monitored after sequential additions of amphipol A8-35 in the absence (closed circles) or presence (closed squares) of 0.001 mg/ml ionophore A23187. Panel B, the active transport of calcium into SR vesicles (0.4 mg/ml protein) and its spontaneous release following the exhaustion of high energy supply were monitored by differential absorbance in the presence of 30 µM Arsenazo III, a Ca2+-sensitive dye. The signal was first calibrated with three sequential additions of Ca2+ (each 20 µM), PEP was added as a substrate for the pyruvate kinase also present in the medium (dashed line, 0.04 mM; solid line, 0.2 mM), and 0.04 mM ATP was then added. After the exhaustion of ATP and PEP, resulting in the initiation of spontaneous release, either 0.002 mg/ml ionophore A23187 (dashed line) or 0.025 mg/ml amphipol A8-35 (solid line) was added. Panel C, the extent of amphipol-induced membrane solubilization, if present, was evaluated by measuring 90° light scattering by SR vesicles (closed circles) or unilamellar liposomes (open circles) after the sequential addition of various concentrations of amphipol A8-35 and equilibration for 1-2 min. The membrane concentration was low: 0.02 mg/ml protein and 0.01 mg/ml lipid, or 0.04 mg/ml lipid alone. Amphipol concentration was increased to 0.5 mg/ml, and solubilizing concentrations of C12E8 were finally added.

We monitored the accumulation of Ca²⁺ by intact SR vesicles and its spontaneous release after exhaustion of the high energy PEP and ATP supply (18). We found that adding amphipol A8-35 (0.025 mg/ml) accelerated this release almost as efficiently as the classical Ca²⁺ ionophore A23187 (Fig. 1B). Thus, stimulation of ATPase activity observed upon addition of low concentrations of amphipols to intact SR is due to this ionophore-like effect of amphipols, which releases inhibitory constraints; detergents act in a similar way when they partition into SR membranes (11, 33, 34).

The permeabilization observed was due to amphipol partitioning into the membranes rather than rapid membrane solubilization, as shown by light scattering measurements; high concentrations of A8-35, although in great weight/weight excess over the membranes, did not significantly reduce the intensity of light scattered by SR membranes or pure lipid vesicles. In contrast, the subsequent addition of C₁₂E₈ reduced the intensity of scattered light to almost zero within seconds (Fig. 1C).² The small decrease in scattered light intensity observed at low amphipol concentrations (0.02 mg/ml) in unlikely to indicate partial solubilization. Instead, it probably reflects changes in membrane refraction index, as observed with SR membranes treated with concentrations of detergent too low to trigger membrane solubilization but high enough to result in significant detergent partitioning into the membrane (11, 15, 33-35). Thus, amphipol A8-35 does not solubilize membranes over this time scale (minutes).

Amphipol A8-35 Reversibly Inhibits the Activity of C₁₂E₈- or DM-solubilized Ca²⁺-ATPase, and Apparently Competes with Detergent, Irrespective of pH-- We then investigated the effect of amphipol A8-35 on the hydrolytic activity of C₁₂E₈-solubilized SR ATPase. The addition of C₁₂E₈ alone to intact SR vesicles stimulated ATPase activity, concomitant to membrane permeabilization and solubilization (11, 33, 34), and the subsequent addition of amphipols to the solubilized ATPase resulted in the immediate inhibition of ATPase activity, in a concentration-dependent manner; this inhibition was partially overcome by increasing C₁₂E₈ concentration (Fig. 2A). If SR membranes were first treated with 1.1 instead of 0.1 mg/ml C₁₂E₈, the inhibition resulting from the addition of amphipols was less marked than at the lower C₁₂E₈ concentration (Fig. 2B). The concentration of SR protein present during these activity measurements was very low (0.002 mg/ml), implying that SR membranes were fully solubilized at both concentrations of C₁₂E₈ (whose CMC is 0.05 mg/ml). Presumably, at the higher concentration of C₁₂E₈, amphipol polymers were diluted in protein-free detergent micelles. Amphipol and detergent seem to compete in their interactions with the solubilized ATPase, amphipol at the surface of the protein being more inhibitory than C₁₂E₈, which is known to be one of the best detergents at maintaining the activity of solubilized ATPase (2, 11). Qualitatively similar results were obtained after solubilization of SR ATPase with 0.1 versus 1 mg/ml DM (data not shown). The hydrolysis of para-nitrophenyl phosphate by detergent-solubilized ATPase was also inhibited by amphipol A8-35 (data not shown).

View larger version (24K): [in this window] [in a new window]   Fig. 2.   Reversible inhibition by amphipol A8-35 of the ATPase activity of C12E8-solubilized SR membranes. SR vesicles were diluted to 0.002 mg/ml in the ATPase assay medium, and hydrolytic activity was monitored continuously with a coupled enzyme system after various additions. Panel A, a typical recording of NADH oxidation, showing the stimulatory effect of 0.1 mg/ml C12E8, the immediate inhibition of ATPase activity on addition of amphipol A8-35 in the presence of 0.1 mg/ml C12E8, and the rapid partial reversal of this inhibition on addition of a higher concentration of C12E8. Panel B, ATPase activity was measured after addition of either 0.1 mg/ml C12E8 followed by various concentrations of A8-35 (open triangles), or 1.1 mg/ml C12E8 (causing slight delipidation-induced inhibition, see Refs. 2 and 11) followed by A8-35 (open diamonds). Panel C, as for panels A and B, but adjusting the pH of the medium as indicated. Open triangles, 0.1 mg/ml C12E8 alone; closed triangles, 0.1 mg/ml C12E8 followed by 0.3 mg/ml A8-35; small closed squares, 0.3 mg/ml A8-35 first, followed by 0.1 mg/ml C12E8.

In the same series of experiments and at various pH values, we compared ATPase activity following the addition of amphipols to C₁₂E₈-solubilized membranes (as previously) and following the addition of C₁₂E₈ to amphipol-treated SR membranes. ATPase activity, if tested immediately, was more strongly inhibited if amphipol was added to detergent-solubilized membranes than if detergent was added to amphipol-perturbed membranes (Fig. 2C). This suggests that the addition of amphipols to intact membranes partially protects the ATPase from complete solubilization on the subsequent addition of detergent (with a concomitant smaller decrease in activity), at least over a time scale of minutes. Thus, the initial state of ATPase dispersion (in membranes or solubilized by detergent micelles) is of key importance for the ATPase·amphipol interaction. To prevent kinetic effects in subsequent experiments, we therefore always added the polymer to detergent-solubilized ATPase.

If plotted as a function of pH, ATPase activity displayed a similar bell shape in the presence and absence of amphipols (Fig. 2C). This suggests that the inhibitory effect of the negatively charged amphipols is not due to a change in local pH at critical sites in the ATPase. In similar studies at various ATP concentrations (data not shown), we obtained no clear-cut evidence for an electrostatic repulsion between ATP and the amphipols being the main cause of ATPase inhibition. We also investigated whether the negatively-charged amphipols could inhibit ATPase activity by chelating Ca²⁺ from the solution. By titrating the competition between A8-35 and Murexide (a classical low affinity Ca²⁺-sensitive dye) for Ca²⁺ binding, we found that Ca²⁺ binding to amphipols was not a major cause of ATPase inhibition in our experiments.²

Dilution of an ATPase·DM·A8-35 Sample with Detergent-free Medium Leaves the ATPase with a Low Level of Activity, but Full Activity Is Recovered by Adding Back Detergent: A8-35 Prevents Short Term Irreversible Inactivation by Detergent Removal-- Amphipols were originally designed to make it possible, after the initial solubilization of membrane proteins by detergent, to reduce the detergent concentration below the CMC while keeping membrane proteins soluble. We assessed the activity of Ca²⁺-ATPase treated in this way. The nonionic detergent DM was used for the initial solubilization of the SR, because it maintains the ATPase in a stable, monomeric form, prior to amphipol treatment, more efficiently than C₁₂E₈ (2, 23). DM-solubilized ATPase was incubated in the presence or absence of amphipols, diluted with detergent-free assay medium (or C₁₂E₈-containing medium, for control), and the hydrolytic activity of the diluted sample was measured.

If DM-solubilized SR, in the absence of amphipol, was diluted with a medium containing 1 mg/ml C₁₂E₈, a detergent that supports ATPase turnover (2, 11, 34), its activity was high and sensitive to Ca²⁺ chelation by EGTA and to adding back Ca²⁺. If a similar DM-solubilized sample was diluted with detergent-free assay medium (such that the final free DM concentration dropped well below the CMC for DM), the resulting residual activity was low (about 20%) and was not increased by the subsequent addition of C₁₂E₈ (Fig. 3A). This low residual activity is presumably due to preservation of a small fraction of ATPase molecules in an active and probably oligomeric state after dilution, whereas most of the solubilized and initially monomeric Ca²⁺-ATPase have been irreversibly inactivated and start to aggregate, as has been demonstrated for C₁₂E₈-solubilized ATPase diluted with C₁₂E₈-free medium (2). In contrast, if detergent-solubilized SR was supplemented with amphipol A8-35 before dilution with detergent-free assay medium, ATPase activity after dilution was low but sensitive to chelation and re-addition of Ca²⁺. In addition, full activity was immediately recovered by adding excess C₁₂E₈ (Fig. 3B). Thus, amphipols prevented irreversible inactivation of ATPase monomers upon dilution with a detergent-free medium.²

View larger version (20K): [in this window] [in a new window]   Fig. 3.   Protection conferred by amphipol A8-35 against the irreversible inactivation of solubilized ATPase induced by its dilution with detergent-free medium. Panels A and B, DM-solubilized SR proteins were incubated in the presence or absence of amphipol A8-35 (1 mg/ml SR protein, 5 mg/ml DM, and either 0 (SR·DM) or 5 (SR·DM·A8-35) mg/ml A8-35). Aliquots (4 µl) of these samples were diluted in 2 ml of ATPase assay medium, which either was detergent-free (right traces) or contained 1 mg/ml C12E8 from the start (left traces). NADH absorbance was recorded to monitor ATPase activity. Various additions were also made, as indicated (arrows): 1 mM EGTA and 1 mM extra Ca2+ (for panels A and B), and 1 mg/ml C12E8 at the end for the traces on the right. Panel C, HPLC-delipidated ATPase was used for a similar experiment. The initial sample contained 0.2 mg/ml delipidated ATPase, 0.4 mg/ml DM, and 1 mg/ml A8-35, and was diluted for ATPase assay (20 µl in 2 ml).

In complementary experiments, we investigated whether lipids were involved in the amphipol-dependent protection of ATPase activity, using a classical protocol to obtain delipidated ATPase (23). The residual activity observed upon dilution of a delipidated ATPase·DM·A8-35 complex with detergent-free medium was even lower than that in the presence of endogenous SR lipids, but ATPase activity again was fully restored by adding excess C₁₂E₈ (Fig. 3C). Thus, upon dilution with a detergent-free medium, the ATPase·A8-35 complex was resistant to irreversible denaturation irrespective of the presence or absence of lipids. These experiments were repeated using various protein/detergent/amphipol ratios before dilution, with and without endogenous lipids. We found that the initial amphipol to detergent ratio had to be high enough (about 1/1, w/w) to permit full recovery of ATPase activity upon subsequent addition of detergent (Table I).

We found that relatively low concentrations of detergent were sufficient for the diluted ATPase·amphipol complexes to recover activity immediately. This was true not only for C₁₂E₈ (full activity recovered at a total concentration of 44 µg/ml, i.e. 90% of the CMC in water; data not shown), but also for DM. In the latter case, full activity recovered at a total DM concentration of 0.1 mg/ml, 110% of the CMC in water (Fig. 4). In a control experiment, the ATPase in the absence of amphipol had a similar high activity in the presence of 0.1 mg/ml DM (data not shown). This stimulating effect of low concentrations of DM on A8-35-inhibited ATPase is particularly remarkable because DM (like C₁₂E₈, but to a much greater extent) added to permeabilized SR membranes actually inhibits ATPase activity if used at concentrations too low to cause membrane solubilization (Fig. 4, and Ref. 11). Another clear result was that the inhibition of activity observed at concentrations of DM higher than 0.1 mg/ml was similar for ATPase·amphipol complexes and for solubilized SR membranes in the absence of amphipol (Fig. 4). This is probably mainly due in both cases to removal of lipids: either the endogenous SR lipids, in the latter case or, in the former, the few residual lipids (23) present in this "delipidated" ATPase preparation (obtained in the presence of DM, a detergent that is known to be a poor solubilizer of lipid alone; Ref. 75).²

View larger version (17K): [in this window] [in a new window]   Fig. 4.   Diluted ATPase-amphipol complexes recover activity with low concentrations of DM, whereas very high concentrations are inhibitory. HPLC-delipidated ATPase was incubated in the presence of DM and amphipol A8-35 (1 mg/ml ATPase, 2 mg/ml DM, and 5 mg/ml A8-35, open squares) and subsequently diluted with ATPase assay medium (4 µl in 2 ml). Various concentrations of DM were then sequentially added, with NADH oxidation recorded continuously to monitor ATPase activity. In a control experiment (open circles), permeabilized SR vesicles were used instead of ATPase·DM·A8-35 complexes (in this case, the diluted sample consisted of 0.002 mg/ml SR protein (without DM or A8-35) to which 1 µg/ml A23187 had been added).

Dispersion State of ATPase·A8-35 Complexes

The Addition of Amphipol A8-35 to DM-solubilized ATPase Keeps Most of the ATPase Soluble after Extensive Dilution in the Absence of Detergent-- The molecular dimensions of the diluted inactivation-resistant ATPase·A8-35 complex in detergent-free medium were estimated by size exclusion chromatography (22, 23). If, as a preliminary control, ATPase solubilized by DM in the absence of amphipol was diluted 1:25 in the presence of 1 mg/ml DM and subjected to chromatography with 1 mg/ml DM as the eluant, the ATPase eluted, as expected, as a main peak (Fig. 5A), the elution volume of which is known to correspond to the ATPase monomer (Refs. 22 and 23; see also the sedimentation experiments below). A minor peak was also detectable, corresponding to mixed lipid/detergent micelles (23). If, as a second control, DM-solubilized ATPase was similarly diluted 1:25 and subjected to chromatography but now in the absence of DM, the ATPase eluted close to the column void volume (V₀ is approximated here by the elution volume for the first marker of the calibration kit, thyroglobulin, with M_r = 670,000 and R_S = 86 Å), consistent with the expected formation of membrane-like very large aggregates (Fig. 5B). In contrast, if DM-solubilized ATPase was incubated with amphipol A8-35 before dilution and chromatography in the absence of DM, most of the ATPase penetrated the gel and eluted as a main peak (solid thick line in Fig. 5B).

View larger version (38K): [in this window] [in a new window]   Fig. 5.   Size exclusion chromatography of a DM-free diluted ATPase·A8-35 complex, compared with DM-solubilized ATPase in the absence of A8-35. Panel A, dashed thick line, as a reference, DM-solubilized DOC-purified ATPase (1 mg/ml protein and 5 mg/ml DM) was diluted (1:25) with a buffer containing 1 mg/ml DM and chromatographed with the same medium. MM indicates the elution of mixed lipid/detergent micelles. Bio-Rad gel filtration standard proteins as well as two other proteins not included in the Bio-Rad kit (ferritin and catalase) were also chromatographed in the same way (thin lower line and arrows). The two traces have been arbitrarily shifted apart by a few milliunits of absorbance, to improve clarity. Panel B, dashed thick line, another aliquot of the same DM-solubilized ATPase was diluted (1:25) with detergent-free buffer and chromatographed in the absence of detergent. Solid thick line, DM-solubilized ATPase was incubated with amphipol (1 mg/ml protein, 5 mg/ml DM, and 5 mg/ml A8-35) before dilution and elution with detergent-free buffer. The thin dashed line and the arrows correspond to gel filtration standard proteins. Panel C, same experiment as in panel B, except that HPLC-delipidated ATPase was used as the starting material, at 0.2 mg/ml protein and 0.4 mg/ml DM, in the absence (dashed thick line) or presence (solid thick line) of 1 mg/ml A8-35. These samples were then diluted (1:20), applied to the column, and eluted with detergent-free buffer. Protein standards are also shown, after appropriate normalization (thin lines).

The ATPase·A8-35 complexes in detergent-free buffer eluted earlier than the monomeric ATPase·DM complex in DM (compare Fig. 5, B and A). Similar results were obtained if a buffer of higher ionic strength (250 mM NaCl, 0.1 mM EDTA, and 20 mM TES-Tris, pH 7) was used for column equilibration, ATPase dilution, and elution (data not shown). Experiments were also performed with HPLC-delipidated ATPase (cf. Refs. 38 and 39 for effects of lipid on HPLC elution properties). Unlike amphipol-free control samples, which eluted as highly aggregated species, most delipidated ATPase·amphipol complexes again deeply penetrated the gel (Fig. 5C), and again eluted earlier than ATPase monomers in DM alone.

Unexpectedly, however, the elution profile for the diluted delipidated ATPase·A8-35 complexes was broader after dilution in the absence of detergent than that for monomeric ATPase in DM (Fig. 5, compare C with A). Moreover, although delipidated ATPase, obtained from the peak fraction of a preparative HPLC run, contained virtually no aggregates (data not shown, but see Fig. 6A below), the elution profile for the diluted delipidated ATPase·A8-35 complex revealed a small but significant amount of material eluting close to the column void volume (V₀ in Fig. 5C). We thus tested whether aggregation of ATPase·A8-35 complexes was occurring, using more concentrated samples.

View larger version (37K): [in this window] [in a new window]   Fig. 6.   Size exclusion chromatography of ATPase samples in the presence of A8-35 alone or of DM alone, before and after reconcentration using Centricon-100 filters. Panel A, solid thick line, chromatography in the absence of detergent of delipidated ATPase·A8-35 complexes, obtained by 1:20 dilution in the absence of detergent of an initial sample containing 0.4 mg/ml protein, 0.8 mg/ml DM, and 2 mg/ml A8-35 (thus, final concentrations were 0.02, 0.04 (i.e. below CMC), and 0.1 mg/ml, respectively); dashed thick line, chromatography in the presence of 1 mg/ml DM of a delipidated ATPase·DM complex without amphipol, obtained after 1:20 dilution in the presence of DM of a similar but amphipol-free initial sample (thus, final concentrations were 0.02, 1.04 and zero mg/ml, respectively). For this and the following HPLC experiments, we used a new TSK column (but of the same type as previously). The thin lines correspond to gel filtration standard proteins, eluted in the absence or presence of detergent; these traces have been shifted for clarity. Panel B, solid thick line, elution profile in the absence of detergent of a concentrated ATPase·A8-35 complex, obtained after reconcentration using Centricon-100 filters of the diluted DM-free samples in panel A; dashed thick line, elution profile in the presence of 1 mg/ml DM of a concentrated ATPase·DM complex in DM, obtained after similar reconcentration of the diluted amphipol-free sample in panel A. The protein concentration in both reconcentrated samples was 0.4 mg/ml. On the chromatogram illustrated by the solid line, hatched boxes indicate fractions collected for subsequent rechromatography in the absence of detergent. Panel C shows the result of this second chromatography.

If the Dilute Delipidated ATPase·A8-35 Sample Is Reconcentrated with Centricon-100 in the Absence of Detergent, Large Aggregates Are Apparently Formed-- We first reconcentrated freshly diluted samples by ultrafiltration through Centricon-100 filters. We compared delipidated ATPase incubated with A8-35 in DM, subsequently diluted in the absence of detergent, and finally reconcentrated, with control delipidated ATPase incubated in DM without A8-35, subsequently diluted in the presence of detergent and finally reconcentrated. The diluted samples (Fig. 6A) behaved as described above. However, after concentration (Fig. 6B), the detergent-free ATPase·A8-35 complexes had a very broad elution profile suggestive of a significant proportion of aggregates. The apparent heterogeneity of the concentrated ATPase·A8-35 complexes was confirmed by a second chromatography for two different fractions, initially eluting early or late, and which again eluted early and late, respectively (Fig. 6C). In contrast, reconcentration of A8-35-free ATPase in DM gave homogeneous complexes (Fig. 6B).

Slowly Removing Detergent from ATPase·DM·A8-35 Complexes by a Bio-Bead Treatment at High Protein Concentration Also Creates a Heterogeneous Population-- We then used an alternative procedure for removing detergent from ATPase·DM·A8-35 complexes while keeping protein concentrations high: polystyrene bead treatment (Bio-Bead SM2; Refs. 26 and 40-44). Delipidated ATPase·DM samples supplemented with various amounts of A8-35 were treated with Bio-Beads, and the elimination of detergent was followed using [¹⁴C]DM (Fig. 7A). Once the DM concentration had fallen below the CMC, samples were subjected to chromatography in the absence of detergent. A8-35-containing Bio-Bead-treated samples eluted with broad and composite profiles (Fig. 7B). The longer the incubation period with Bio-Beads (and thus the lower the final DM concentration), the more aggregated the final sample was (data not shown). The peak corresponding to the smallest complexes (late elution) almost disappeared from the chromatogram if the amount of A8-35 added to the sample before Bio-Bead treatment was much smaller than that of DM, as expected from Fig. 5 and the likely reformation, as a result of detergent removal, of large membrane-like aggregates. However, the proportion of small complexes was also low if the A8-35 concentration was much higher than that of detergent (Fig. 7B).² Fractions eluting early or late essentially retained their specific behavior in subsequent chromatography steps, as for the samples shown in Fig. 6, again ruling out the possibility that the apparent heterogeneity of ATPase·A8-35 samples could result from interaction with the gel matrix (data not shown). Thus, excess amphipol in conditions of moderately high protein concentration (0.5 mg/ml) does not favor ATPase dispersion in the absence of detergent.

View larger version (36K): [in this window] [in a new window]   Fig. 7.   Removal of detergent from ATPase·DM·A8-35 samples by treatment with Bio-Bead polystyrene beads, and the resulting HPLC elution profiles. Panel A, delipidated ATPase was incubated with DM and A8-35 at the concentrations indicated (in mg/ml), in TES 7.0 buffer. Tracer [14C]DM was added to the mixture, and samples were then treated with 100 mg/ml Bio-Bead SM-2. Residual detergent was determined after various periods. Protein losses, as deduced from UV absorbance measurements, were insignificant (data not shown). Panel B, similar samples were treated for 2 h 40 min and loaded onto the HPLC column for chromatography in the absence of detergent. In this series of experiments, thyroglobulin, -globulin, ovalbumin, myoglobin, and vitamin B-12 from the standard Bio-Rad kit eluted at 5.68, 8.00, 8.52, 10.27, and 11.50 ml, respectively. The various traces have been shifted arbitrarily along the y axis to separate them so as to improve clarity. Hatched boxes indicate fractions collected for subsequent rechromatography (see text).

Ultracentrifugation Analysis-- The apparent heterogeneity of ATPase·A8-35 samples was further characterized by analytical ultracentrifugation. Sedimentation velocity experiments confirmed that, even at an almost optimal initial ATPase·DM·A8-35 w/w/w ratio (0.5/0.75/1), delipidated ATPase·A8-35 complexes after Bio-Bead-treatment were less homogeneous than the initial ATPase in DM alone. The predominant species in the detergent-free ATPase·A8-35 complexes accounted for 70% of the total sample, with an estimated sedimentation coefficient of 6.2 ± 0.05 S (Fig. 8, B and D), but a significant component was detected at 8.8 ± 0.5 S, together with components sedimenting even more rapidly, consistent with the small, intermediate, and large particles on the chromatograms in Fig. 7B. In contrast, the ATPase·DM sample sedimented as an essentially homogeneous species at 7.44 ± 0.02 S, with minor components at 9.5 S (Fig. 8, A and C).²

View larger version (43K): [in this window] [in a new window]   Fig. 8.   Sedimentation velocity analysis of DM-solubilized delipidated ATPase (A and C) or of DM-free ATPase·A8-35 complexes obtained by treatment with Bio-Beads (B and D), and sedimentation equilibrium analysis of similar samples (E and G, and F and H). Aliquots of amphipol-free delipidated ATPase·DM complexes (0.5/0.8, in mg/ml; A and C) or of Bio-Bead-treated delipidated ATPase·DM·A8-35 complexes (initially 0.5/0.8/1, in mg/ml; after treatment with 100 mg/ml of Bio-Beads for 50 min, the final DM concentration was less than 0.06 mg/ml; B and D) were centrifuged at 50,000 rpm in the analytical ultracentrifuge at 20 °C, together with protein-free controls in blank cells. Panels A and B, sequential scans, taken at 6-min intervals, as a function of radial distance. Panels C and D, sedimentation coefficient distributions, g*(s) (thick lines), and their decomposition into individual components (thin lines) (29). Panels E and F, sedimentation equilibrium analysis, after 19 h of equilibration at 6,000 rpm, of similar samples (concentrations, in mg/ml, were 0.3/0.6/0 (E) and 0.3/0.007/0.6 (F); for F, the original sample before Bio-Bead treatment had concentrations of 0.3/0.6/0.6). The thin line represents a tentative fit (which, in F, is poor) of a single species model to the data; residuals are given in panels G and H, respectively.

Sedimentation equilibrium experiments were also performed. The ATPase sample in DM alone proved much more homogeneous than the ATPase·A8-35 complexes (Fig. 8, E and G versus F and H). For the ATPase sample in DM alone, an apparent "buoyant" molar mass (45) of 48,600 ± 800 was obtained for the most prominent (98%) component (panels E and G). This value, together with the known amount of bound DM (0.9 g/g) and its partial specific volume (0.82 cm³/g), enabled us to calculate a molecular mass for the ATPase protein moiety of 117,000, demonstrating that delipidated ATPase in DM alone was monomeric (23). We also used this buoyant molar mass, together with the measured sedimentation coefficient 7.44 ± 0.02 S (i.e. s_20,w = 7.62 S), to calculate a Stokes radius of 5.5 nm for ATPase in DM. This value, which can also be deduced from the HPLC elution pattern after calibration with standard proteins of known Stokes radii (Figs. 5A and 6A, 5.5 ± 0.3 nm), is identical to that for ATPase in C₁₂E₈ (22) (the sedimentation coefficient for Ca²⁺-ATPase monomers in DM is larger than that for Ca²⁺-ATPase monomers in C₁₂E₈, 7.62 S (this work) versus 5.1 S (45), mainly because DM is more dense than C₁₂E₈ (DM has a partial specific volume smaller than that of C₁₂E₈, 0.82 (Ref. 46) versus 0.97 cm³/g (Ref. 23)), and also because the protein binds slightly more (23) DM than C₁₂E₈).

For the detergent-free ATPase·A8-35 complexes, equilibrium sedimentation gave a much more heterogeneous distribution of components (Fig. 8, F and H), which was not analyzed in detail. However, the smallest of these complexes are probably complexes of A8-35 with monomeric ATPase, as judged from their low sedimentation coefficient. Note that heterogeneity of the A8-35-containing samples was not due to the mere presence of amphipol, but was dependent on removal of detergent; the sedimentation velocity pattern of an amphipol-containing sample before Bio-Bead treatment (ATPase·DM·A8-35 = 0.5/0.5/1) proved nicely homogeneous, with a sedimentation coefficient about 5% lower than that of ATPase in DM alone. The same sample, before Bio-Bead treatment, also appeared homogeneous in equilibrium measurements, with an apparent buoyant molar mass again slightly lower than that of ATPase in DM (data not shown).

View larger version (34K): [in this window] [in a new window]   Fig. 9.   Analysis of various samples after reaction with glutaraldehyde to reveal protein-protein contacts, as detected by HPLC in the presence of DM. Delipidated ATPase was incubated with DM and A8-35 (concentrations were 0.5, 1, 2.5, in mg/ml) in TES 7.0 buffer; it was treated with Bio-Beads for 2 h 40 min, and subsequently subjected to chromatography in the absence of DM. This resulted in the top chromatogram in Fig. 7B. Late and early fractions, corresponding to small and large particles, respectively, were collected. These fractions were concentrated using Centricon-100 filters, to about 0.25 and 0.5 mg/ml protein, respectively. They were then reacted with glutaraldehyde (+glut, panels A and B) while controls were left unreacted (control, panels C and D), diluted 1:1 with hydrazine, and analyzed by HPLC in the presence of 1 mg/ml DM after dilution 1:4 with the DM-containing buffer. Arrows indicate the elution volumes of thyroglobulin, -globulin, ovalbumin, myoglobin, and vitamin B-12. The presence of glutaraldehyde for panels A and B (and of hydrazine for all four panels) accounts for the large increase in optical density at about 12 ml.

Protein-Protein Contacts in Bio-Bead-treated ATPase·A8-35 Samples Can Be Detected by Cross-linking, but Only for a Fraction of the Large Particles-- We investigated whether the large particles in Bio-Bead-treated ATPase·DM·A8-35 samples corresponded to oligomeric complexes by using nonspecific cross-linking with high concentrations of glutaraldehyde to detect protein-protein contacts (30, 47-49). We first checked for covalent protein aggregates by SDS-PAGE. With this technique, it is possible to distinguish ATPase monomers from ATPase oligomers made covalent by glutaraldehyde (dimers or higher order aggregates, which may not even enter the gel). In parallel experiments, cross-linked complexes were diluted and analyzed by HPLC in the presence of DM, to quantify the extent of ATPase cross-linking after its reaction with glutaraldehyde more easily. Cross-linking experiments were carried out on two different fractions of a Bio-Bead-treated ATPase·A8-35 sample (shown by the hatched boxes in Fig. 7B): an early-eluting fraction, corresponding to large particles, and a late-eluting fraction, corresponding to small particles. HPLC analysis (Fig. 9) and SDS-PAGE² gave consistent results; ATPase molecules in small particles did not undergo significant cross-linking during reaction with glutaraldehyde, as expected (Fig. 9A; the elution position of glutaraldehyde-reacted monomers was slightly altered compared with unreacted samples because of the internal cross-links). In contrast, most of the glutaraldehyde-reacted ATPase molecules in large particles eluted close to the void volume, although a significant proportion still eluted at the position of the cross-linked monomer (Fig. 9B). Thus, in the early-eluting fraction of these detergent-free ATPase·A8-35 samples, particles with protein chains in close proximity with each other coexist with particles with protein chains that cannot be cross-linked by glutaraldehyde to any neighbor, even though these particles elute close to the column void volume.

Further information about the nature of the aggregates came from control experiments in which small and large particles from Bio-Bead-treated ATPase·A8-35 preparations were analyzed in the same manner except for omitting glutaraldehyde cross-linking. Under these conditions, the non-cross-linked small particles migrated as monomers upon HPLC in the presence of DM (Fig. 9C), whereas a significant proportion of the large ones again eluted close to the void volume, despite the presence of DM (circled in Fig. 9D). Upon SDS-PAGE, on the other hand, large particles were effectively resolved into monomers.² This indicates the presence of DM-resistant non-covalent protein-protein contacts in the large aggregates, and will be discussed below in relation to irreversible denaturation.

Possible Reasons for Formation of Large Particles in DM-free ATPase·A8-35 Samples: Ability of A8-35 to Form Aggregates, Some of Which Are Very Large, and Possible Interaction of Amphipol with the ATPase Domain Protruding from the Membrane-- In the absence of detergent, the formation of amphipol aggregates has not been described before, but is certainly not unexpected from what is known of amphiphilic molecules or polymers (74). We assessed the apparent size of these aggregates by size exclusion chromatography, using the weak but unambiguous absorption of amphipols in the far UV region to detect their elution. If amphipol A8-35 alone was loaded onto an HPLC column and subjected to chromatography in the absence of detergent, part of the injected polymer eluted at a position similar to that for -globulin (i.e. with an apparent R_S of about 5 nm), while a significant fraction eluted close to the void volume (Fig. 10A).² We checked that the apparent polydispersity of A8-35 aggregates in buffer was not an artifact of HPLC analysis, by subjecting early and late fractions to a second chromatography run: this again generated early and late fractions, respectively (Fig. 10, B and C), suggesting that, in the absence of detergent, these large and small particles are in only slow equilibrium, as expected from such polymers. In contrast, if A8-35 was first dispersed into an equivalent amount of DM and subjected to chromatography in the presence of DM, it partitioned into the gel phase, as expected for newly formed A8-35·DM mixed micelles (Fig. 10D), larger than pure DM micelles (which normally elute close to 9.3-9.4 ml; data not shown). We also monitored aggregation of A8-35 in buffer through analytical sedimentation experiments, estimating the velocity of A8-35 sedimentation (at 60,000 rpm) from absorption measurements in the far UV region (between 280 and 235 nm; data not shown). The presence of very large particles was shown by the sedimentation in the first few minutes of the run at 60,000 rpm of a significant proportion of the material contributing to the original optical density in the centrifugation cell (as measured after preliminary acceleration to 3,000 rpm). Most of the remaining A8-35 aggregates had sedimentation coefficients of 1.6-1.7, and a small fraction of aggregates sedimented with a coefficient of about 3 (data not shown). Thus, from all these data, it is clear that under ordinary ionic conditions, amphipol A8-35 itself, in the absence of detergent, may form micelle-like aggregates, some of which are large enough to be excluded from a column or sediment rapidly. The existence in certain conditions of large aggregates for A8-35 alone in buffer makes it possible that some of the ATPase·A8-35 complexes in the absence of detergent contain monomeric ATPases surrounded by an extended A8-35 moiety.

View larger version (44K): [in this window] [in a new window]   Fig. 10.   Chromatography of A8-35 in the absence of DM (A-C) or of mixed A8-35·DM micelles in the presence of DM (D), and A8-35-induced aggregation of lysozyme (E and F). Panels A-D, 5 mg/ml A8-35 in TES 7.0 medium was injected onto the column and eluted in the absence of detergent (panel A). Two fractions were separately collected, as indicated by the hatched areas, and subjected to a second round of chromatography on the same column (panels B and C). Amphipol elution was detected at 257 nm, because amphipols have only little absorption at 280 nm. For panel D, mixed A8-35·DM micelles (5 mg/ml each) were injected onto a column equilibrated with 1 mg/ml DM, and were eluted with 1 mg/ml DM. The arrows in panels A and D indicate the elution positions of thyroglobulin, -globulin, and ovalbumin. Panels E and F, lysozyme was suspended at 0.5 mg/ml in TES 7.0 buffer, and A8-35 was sequentially added to various final concentrations (indicated in mg/ml). Absorbance (or turbidity) over time was monitored at various wavelengths (E); the resulting spectra at various times are also shown (F).

To account for the formation of large ATPase·A8-35 complexes with relatively few protein-protein contacts (Fig. 9B), we also investigated the possibility that amphipol A8-35 might associate not only with the protein transmembrane region but also with the water-soluble globular domain of the ATPase that protrudes from the membrane. It might thus link several ATPase molecules via these domains. We first monitored, in HPLC experiments, the elution in the absence of detergent of various soluble proteins, previously incubated in the absence or presence of amphipol A8-35 (0.5 mg/ml protein and 0 or 2.5 mg/ml A8-35). For some of these proteins, the elution volume was completely unaffected by prior incubation with amphipols (elution volumes identical to ±0.03 ml). This was the case for -lactalbumin, transferrin, ovalbumin, and -globulin. However, minor interactions were detected for myoglobin and -lactoglobulin, and clear interactions were detected for bovine serum albumin, with a shift of more than 0.25 ml toward earlier elution and a greater tendency to aggregate (data not shown). Similar shifts and tendencies were also obtained in experiments performed with a water-soluble 29-kDa compact fragment of the cytosolic globular domain of the Ca²⁺-ATPase in the presence of A8-35 (data not shown)²; this 29-kDa fragment had been produced by expression in E. coli, by P. Falson, and designed based on a protease-resistant subdomain of the Ca²⁺-ATPase (32).We then investigated whether such interactions were due solely to the interaction of amphipols with hydrophobic pockets or regions in water-soluble domains, or whether ionic interactions between the negatively charged amphipols and positively charged patches on globular proteins could also occur, as documented for hydrophilic polymers (71). Lysozyme was chosen as a model system. Lysozyme aggregated in the presence of amphipol A8-35 (Fig. 10, E and F), demonstrating the possibility of long range interactions between amphipols and protein water-soluble domains. Thus, due to the presence of carboxylate groups, amphipols may be considered to be weak cation exchangers to which certain protein domains may bind.

Finally, to account for the formation (Fig. 9D) of large ATPase·A8-35 complexes with non-covalent protein-protein contacts resistant to nonionic detergent (30, 72, 73), we investigated the irreversible inactivation of ATPase in more detail.

Denaturation of ATPase·A8-35 Complexes in the Absence of DM, as Opposed to Long Term Stability of ATPase in the Simultaneous Presence of A8-35 and DM

In the Absence of DM, Incubation of ATPase with A8-35 Leads to Irreversible Denaturation of ATPase, Whereas Low Concentrations of DM Stabilize A8-35-complexed ATPase-- Irreversible denaturation of the ATPase·A8-35 complex probably does contribute significantly to the heterogeneities detected both in HPLC experiments at high protein and amphipol concentrations (Figs. 6B and 7B) and in equilibrium sedimentation experiments (performed over 30-40 h) at medium amphipol concentrations (Fig. 8, F and H). In fact, Bio-Bead-treated ATPase·A8-35 complexes with a high A8-35 to protein ratio (ATPase·DM·A8-35 concentrations were 0.5/1/2.5 in mg/ml) had lost most of their activity irreversibly (as tested in the presence of excess C₁₂E₈) after incubation with Bio-Beads, and even the "optimal" initial ATPase·DM·A8-35 ratio (based on the HPLC elution patterns in Fig. 7B) was far from being optimal in terms of the long term stability of Bio-Bead-treated ATPase molecules. As judged from C₁₂E₈-induced recovery of activity, Bio-Bead-treated samples prepared from initial mixtures of ATPase and DM with equivalent weights of A8-35 were more unstable than those prepared from initial mixtures with much less A8-35 (Fig. 11A). Moreover, for diluted ATPase-amphipol complexes, the ATPase environment turned out to be less favorable for long term ATPase stability if the detergent concentration was left well below the CMC after dilution of the ATPase·DM·amphipol complex than if a small amount of detergent was added to the diluted sample (compare rightside-up and upside-down triangles in Fig. 11B). Upon addition of a much higher concentration of DM (0.5 mg/ml) to a diluted ATPase·DM·A8-35 sample, we also found that the ATPase was destabilized, as expected, but less so than a control sample to which no A8-35 had been added (Fig. 11B). Thus, we started investigating the effects of the simultaneous presence of detergent and amphipol, and first assessed the long term stability of detergent-solubilized ATPase at high protein concentration, in the presence or absence of amphipol used in combination with detergent rather than as a substitute for it.

View larger version (32K): [in this window] [in a new window]   Fig. 11.   C12E8-induced activity recovery for Bio-Bead-treated (A) or diluted (B) ATPase·A8-35 complexes, after incubation for various periods in the presence of different concentrations of A8-35 or DM. Panel A, delipidated ATPase (0.5 mg/ml) was incubated with DM (0.75 mg/ml) in TES 7.0 medium along with various concentrations of A8-35 (from 0.125 to 1 mg/ml, as indicated) and the mixture was then incubated with Bio-Beads for 1 h 45 min. This resulted in low DM concentrations, here noted (BB). The samples were then incubated at 20 °C and aliquots (8 µl) were taken after various periods during the course of several days, to measure residual ATPase activity after dilution with 2 ml of medium containing 1 mg/ml C12E8. Delipidated ATPase (0.5 mg/ml) was also incubated with DM (1 mg/ml) and A8-35 (2.5 mg/ml), treated with Bio-Beads for 3 h 20 min, and residual ATPase activity was measured 3 h later in the presence of 1 mg/ml C12E8 (square and dotted arbitrary line). Panel B, instead of being treated with Bio-Beads, an ATPase·DM·A8-35 sample (0.5/1/1) was diluted 1:50 in detergent-free TES 7.0 medium, such that the final concentration of DM dropped to 0.02 mg/ml (and that of ATPase to 0.01 mg/ml). This sample was incubated for several days at 20 °C, either in this form (rightside-up triangles) or in the presence of 0.06 or 0.48 mg/ml of additional DM, resulting in final concentrations of 0.08 or 0.5 mg/ml DM during incubation (upside-down triangles and open circles, respectively). The ability to recover activity after various periods was measured by 1:5 dilution into a C12E8-containing assay medium. A control was also included, in which solubilized ATPase in the absence of amphipol was diluted and incubated in the presence of 0.5 mg/ml DM (closed circles).

Moderate Amounts of A8-35 Added to Detergent-solubilized ATPase Optimize ATPase Stability over Days, Whether in the Presence and Absence of Lipids-- These experiments were conducted with either SR or delipidated ATPase, at either 4 °C or 20 °C. In all cases, the presence of amphipols in addition to DM afforded very significant protection against the loss of ATPase activity that occurs during long term incubation in the presence of detergent alone. Relative protection was even greater if the experiment was performed with delipidated ATPase instead of SR membranes (Fig. 12, A and B), since the rate of irreversible denaturation in the presence of amphipols was similar in the presence and absence of lipids, whereas in the absence of amphipol, delipidated ATPase is more