Báo cáo khoa học: The molecular chaperone a-crystallin incorporated into red cell ghosts protects membrane Na/K-ATPase against glycation and oxidative stress ppt

The molecular chaperone a-crystallin incorporated into red cell ghosts protects membrane Na/K-ATPase against glycation and oxidative stress Barry K. Derham 1 , J. Clive Ellory 2 , Anthony J. Bron 1 and John J. Harding 1 1 Nuffield Laboratory of Ophthalmology, University of Oxford, UK; 2 Laboratory of Physiology, University of Oxford, UK a-Crystallin, a molecular chaperone and lens structural protein protects soluble enzymes against heat-induced aggregation and inactivation by a variety of molecules. In this study we investigated the chaperone function of a-crys- tallin in a more physiological system in which a-crystallin was incorporated into red cell ÔghostsÕ. Its ability to protect the intrinsic membrane protein Na/K-ATPase from external stresses was studied. Red cell ghosts were created by lysing the red cells and removing cytoplasmic contents by size- exclusion chromatography. The resulting ghost cells retain Na/K-ATPase activity. a-Crystallin was incorporated in the cells on resealing and the activity of Na/K-ATPase assessed by ouabain-sensitive 86 Rb uptake. Incubation with fructose, hydrogen peroxide and methylglyoxal (compounds that have been implicated in diabetes and cataract formation) were used to test inactivation of the Na/K pump. Intracellular a-crystallin protected against the decrease in ouabain sensi- tive 86 Rb uptake, and therefore against inactivation induced by all external modifiers, in a dose-dependent manner. Keywords: a-crystallin; ghost cells; glycation; Na/K-ATPase; oxidation. Na/K-ATPase is a highly conserved, ubiquitous membrane protein. The enzyme is composed of three subunits; the alpha subunit ( 113 kDa) binds ATP and sodium and potassium ions, and contains the phosphorylation site. The smaller beta subunit ( 35 kDa glycoprotein) is necessary for activity of the complex and the gamma subunit ( 10 kDa) is involved with modulation of Na/K-ATPase. Several isoforms of both alpha and beta subunits have been identified [1]. Red blood cell membranes contain Na/K-ATPase. Removal of the erythrocyte cytoplasm by lysis followed by size-exclusion chromatography produces white ghost cells showing Na/K-ATPase activity. Erythrocyte ghost Na/K-ATPase activity can be determined by measuring the ouabain-sensitive uptake of 86 Rb (as a congener for potassium). Tissue proteins existing within an environment containing reactive small molecules such as sugars, cyanate, methylglyoxal and other reactive metabolites are vulnerable to nonenzymatic modification that may affect their physio- logical function [2]. Such post-translational modifications contribute to systemic and ocular disease including cataract and the complications of diabetes. Glycation, a process that is pertinent to the aetiology of diabetes, is initiated by the reaction between the carbonyl group of a sugar with an amino group (usually a lysine or the N-terminal amino group) of a protein, to form a Schiff base. This may undergo a further Amadori rearrangement, to produce a ketoamine. There is evidence from experimen- tal diabetes that glycation may play a central role in the impairment of Na/K-ATPase activity in this disorder and contribute to the pathophysiology of diabetic complications [3]. Glycation has far-reaching consequences including the production of increased amounts of the reactive metabolite methylglyoxal, especially in the lens [4]. Methylglyoxal is a reactive a-dicarbonyl with 100% open chain that modifies proteins more rapidly than glucose by an interaction with arginine and cysteine, in addition to lysine [2]. It has been shown to cross-link proteins during glycation or Maillard reactions resulting in protein-bound fluorescent molecules or advanced glycation end products [5]. At physiological concentration (1 l M ) methylglyoxal binds to proteins in blood plasma [6]. Reactive oxygen species such as hydrogen peroxide (H 2 O 2 ) are continually produced in biological systems as unwanted by-products of normal oxidative metabolism. Antioxidant defences detoxify these reactive oxygen species, but increased production by various biological and envi- ronmental factors can lead to oxidative damage to key molecules such as lipid, protein, DNA, etc. Previous experiments in our laboratory have demon- strated inactivation of enzymes by fructose, cyanate and prednisolone-21-hemisuccinate. Fructation causes a decrease in activity of a range of enzymes in vitro [7–9], and the inactivation was prevented by a-crystallin. a-Crystallin, a lens structural protein, comprising of aAandaB subunits is a ubiquitous molecular chaperone, which has been shown to protect many enzymes from inactivation and heat-induced aggregation [10]. Ingolia and Craig [11] discovered an approximate 55% sequence homology Correspondence to J. J. Harding, Nuffield Laboratory of Ophthal- mology, University of Oxford, Walton Street, Oxford, OX2 6AW, UK. Fax: + 44 1865 794508, Tel.: + 44 1865 248996, E-mail: john.harding@eye.ox.ac.uk Enzymes: creatine kinase (EC 2.7.3.2, type 1 from rabbit muscle). (Received 10 February 2003, revised 7 April 2003, accepted 23 April 2003) Eur. J. Biochem. 270, 2605–2611 (2003) Ó FEBS 2003 doi:10.1046/j.1432-1033.2003.03631.x between small heat shock proteins from Drosophila melanogaster and bovine a-crystallin. Horwitz [12] first characterized a-crystallin as a molecular chaperone in vitro, based on its ability to prevent heat-induced aggregation of lens proteins and enzymes. These protective capabilities have been demonstrated with other, in vitro systems, including prevention of aggregation of insulin B chain following reduction of disulphide bonds [13], refolding of guanidine hydrochloride (or urea)-denatured proteins [12,14] and prevention of inactivation of enzymes by small molecules [7,8]. a-Crystallin has also been shown to decrease the degree of thiol oxidation of other lens crystal- lins under conditions of oxidative stress [15]. Characteri- zation of a-crystallin using these assays has indicated similar mechanisms of protection. However, the molecular mechanism of the interaction between a-crystallin and substrates remains enigmatic. Recently, we have shown that a-crystallin incorporated into ghost cells protects soluble enzymes such as catalase, malate dehydrogenase and glutathione reductase from inactivation by fructose [16]. Enzymes were resealed within ghost cells and inacti- vated by fructose. When a-crystallin was resealed with the enzyme, activity was retained. In the present study we demonstrate the protection of the membrane enzyme Na/K-ATPase from inactivation by the heat shock protein a-crystallin. Na/K-ATPase activity decreased upon incubation with fructose, H 2 O 2 and methylglyoxal. However, Na/K-ATPase activity was preserved when the heat shock protein a-crystallin was sealed within the ghost cells. In this situation a-crystallin was able to protect against each form of modification. Methods Materials 86 Rb was purchased from NEN Life Sciences. All other chemicals and enzymes, including luciferin–luciferase firefly lantern extracts were obtained from Sigma. Sepharose 2B and Sephacryl S300 H were obtained from Pharmacia Ltd. a-Crystallin was isolated from bovine lenses by Sephacryl S300 H size-exclusion chromatography as described by Derham and Harding [17]. Preparation of the ghost cells Freshly drawn human blood (30 mL) was collected from volunteers with consent and stored at 4 °C with heparin for a maximum of 3 days. Erythrocyte ghosts, free of haemo- globin were prepared by gel filtration chromatography [18]. Blood (5–6 mL) was centrifuged (1000 g) for 10 min, the plasma and white cells aspirated and the red cells resus- pended at 0 °C with isotonic Hepes buffer (20 m M Hepes, 146 m M NaCl, pH 7.4). This procedure was repeated four times. After the final wash the supernatant was aspirated and the packed red cells ( 3 mL) were lysed with hypotonic buffer (15 m M Pipes, 0.1 m M EDTA pH 6, and approx. 50 mOsm) at a 10% haematocrit. The suspension was shaken gently and cooled in an ice bath for 5 min before loading onto a Sepharose 2B size-exclusion column (5 · 28 cm) pre-equilibrated with the hypotonic Pipes buffer and maintained at 0 °C by a cooling jacket with circulating antifreeze. The column was eluted with Hepes buffer at a constant flow rate of 30 mLÆh )1 and fractions collected in tubes in an ice bath to prevent resealing. The ghost cells eluted in the void volume (70 mL) while the main haemoglobin band followed about 130 mL later (Fig. 1). The lysed cells were white and therefore practically haemo- globin-free. The lysed cells were collected by centrifugation (11 000 g,10min,0°C), the supernatant aspirated and the pellet re-suspended in isotonic Hepes buffer at 0 °Cto prevent resealing. This washing procedure was repeated four times, to remove the hypotonic buffer. The low temperature prevents the ghost cells resealing. Re-sealing ghost cells After the final wash the supernatant was aspirated and the packed ghost cells were suspended in 5 mL of resealing buffer at 0 °C. Resealing buffer contained NaCl (10 m M ), KCl (140 m M ), Mops (10 m M ), dithiothreitol (2 m M ), EGTA (0.1 m M ), potassium phosphate (1 m M )andMgCl 2 (0.15 m M ) at pH 7.4. Potassium ATP (2 m M ), sodium phosphocreatine (5 m M ) and creatine kinase (EC 2.7.3.2, type 1 from rabbit muscle) 5 UÆlL )1 were added as an ATP-regenerating system to maintain membrane integrity [19]. The lysed cell suspension was divided equally into two tubes and to one tube a-crystallin (1 mgÆmL )1 ) was added and incorporated on resealing. For controls, BSA (1 mgÆmL )1 ) and lysozyme (1 mgÆmL )1 )wereusedinstead of a-crystallin. The tubes containing the suspensions were placed on ice for 10 min then at 37 °C for 30 min so that the lysed ghost cells would reseal. After resealing the ghost cells were washed three times in Mops buffer (10 m M Mops, 146 m M sodium nitrate) and successively centrifuged (10 000 g, 5 min) and re-suspended to achieve chloride replacement (to eliminate K-Cl cotransporter activity [20,21] and remove resealing solution. Fig. 1. Elution profile of a haemolysed erythrocyte suspension loaded onto a Sepharose 2B size-exclusion column (5 cm · 28 cm) pre-equili- brated with hypotonic Pipes buffer pH 6 that was maintained at 0 °C. The void volume at  70 mL contains the ghosts. Haemoglobin eluted at 200 mL. 2606 B. K. Derham et al. (Eur. J. Biochem. 270) Ó FEBS 2003 Measurement of 86 Rb flux while modifying Na/K-ATPase The activity of Na/K-ATPase was assessed by ouabain- sensitive 86 Rb uptake. Ghost cells were incubated at 37 °C in Mops buffer and assayed at time zero and after 6 h. Resealed ghost cells, with and without incorporated a-crystallin, were suspended in Mops buffer to give a final haematocrit of  10%. Resealed cells were aliquotted into 1.5-mL Eppendorf tubes (triplicate) and were assigned to the addition of: (a) buffer (control); (b) ouabain (0.1 m M ); (c) modifier; and (d) modifier and ouabain. The modifiers included sucrose (50 m M ), fructose (50 m M ), methylglyoxal (0.1 m M ,1m M and 10 m M )andH 2 O 2 (0.5 m M ). All experiments were performed in triplicate. The Na/K-ATPase flux was initiated by the addition 86 Rb and sufficient potassium nitrate to yield a final concentration of 7.5 m M (20 lCi 86 Rb per ml buffer). The tubes were shaken and then incubated at 37 °C for 30 min, transferred onto ice for 1.5 min, centrifuged (2 min, 10 000 g), and the supernatant removed by aspiration. The ghost cells were washed free of 86 Rb by four successive re-suspensions and centrifugation in ice-cold wash solution (0.1 M MgCl 2 ,10m M Mops pH 7.4). The cell pellet was lysed by addition of 0.5 mL Triton 0.1% (v/v) (TX-100), precipitated by the addition of 0.5 mL 5% w/v trichloro- acetic acid and centrifuged (5 min, 10 000 g). The super- natant was transferred into scintillation vials for Cerenkov counting in a b-scintillation spectrometer. Flux was expressed as percentage control (mean of a minimum of five experiments). The data were analysed using a Student’s paired t-test, with P <0.05 (*), P < 0.01 (**) and P < 0.001 (***) considered statistically significant. ATP determination ATP levels were measured using a luciferin–luciferase enzyme system [22] using a BioOrbit 1253 luminometer. One vial of freeze-dried firefly lantern extract was reconsti- tuted with 5 mL distilled water 1–2 h before assay and stored on ice. To 1 mL buffer (100 m M Tris/HCl pH 6.8, 5m M MgSO 4 ,0.5m M EDTA, 0.5 m M dithiothreitol, 0.1 mgÆmL )1 human serum albumin) was added 50 lL luciferin–luciferase solution and 10 lL sample and gently stirred. After 30 s the light emitted from the ATP-dependent firefly extract was determined and result subtracted from the background value. A calibration curve was set up. All readings were in triplicate. Na and K photometry determination An IL943 flame photometer (Instrumentation Laboratory, Lexington, MA, USA) was used for the determination of sodium and potassium within the erythrocyte ghost cells. Values of Na + and K + were measured and expressed in m M ÆL )1 . All readings were taken in triplicate. SDS/PAGE SDS/PAGE (12.5% w/v gel) was performed as described by Laemmli [23] under reducing conditions with a Bio-Rad system. Ghost cells (20 lL) were dissolved in the sample buffer containing 5% (v/v) 2-mercaptoethanol. Coomassie brilliant blue G was used to detect the polypeptide bands. The relative abundance of sample band was determined by quantitative analysis of digital photographs of gels on a computer (Labworks, UVP Products, Upland, CA, USA). Resealing assay To check that ghosts sealed tightly in our procedure, 14 C sucrose (50 lCiÆmL )1 ) (unlabelled sucrose acted as control) was resealed inside control ghost cells. After resealing the ghost cells were sequentially washed three times in Mops buffer (10 m M Mops, 146 m M sodium nitrate) and centri- fuged (10 000 g, 5 min). The ghost cells were then incubated for 2 h at 37 °C. Samples from the ghost cells and the supernatants were taken for scintillation counting. Creatine kinase determination Creatine kinase activity was assayed following the method of Bernt and Bergmeyer [24], in the presence of fructose (50 m M ) to determine if the inhibitory effect of the modifier was through inhibition of the ATP-regenerating system. Results The ghost cells eluted through the Sepharose 2B size exclusion column in the void volume while haemoglobin and cellular enzymes eluted distinctly later (Fig. 1), com- parable to a previously published gel filtration method of ghost cell preparation [18]. The eluted ghost cells are in their open form, allowing various proteins to be incorporated before resealing. SDS/PAGE analyses of ghost cells alone showed the usual pattern for red cell membrane proteins (Fig. 2, lane 2). Added a-crystallin ( 800 kDa) was incorporated inside the ghost cells. Ghosts resealed in the presence of a- crystallin showed the extra two a-crystallin subunit bands at around 20 kDa confirming the efficient sealing of the protein within the cells (Fig. 2, lane 3). a-Crystallin alone showed clear bands at the expected subunit weight of 20 kDa (lane 4). a-Crystallin was loaded at 5 mgÆmL )1 onto theSDS/PAGEtohighlightthefactthatitcanberesealed, and at high concentrations. These results are in accordance with previous resealing experiments [16]. BSA reseals inside the ghost cells with equal efficiency (result not shown). To confirm that the lysing and resealing procedure produced effectively sealed ghost cells, 14 C sucrose was resealed into ghost cells. The cells were washed three times to remove radioactivity in the supernatant and incubated for 2 h at 37 °C. After this time no radioactivity was detectable in the supernatant. At the same time, significant radioactivity of 14 C sucrose in the ghost cells was measured whichwasthesameatt ¼ 0 and 2 h. This indicates that the 14 C sucrose was trapped by loading and resealing and did not leak out of the ghost cells. The ghost cells were incubated with and without modifiers over 6 h at 37 °C, after which time the activity of Na/K-ATPase was measured over a 30-min period by 86 Rb uptake. A steady-state of Na, 86 Rb exchange is achieved during a 30 min assay. The concentration levels of the modifiers were selected so that they would not interfere with the chaperone function of a-crystallin reported previously [10]. Ó FEBS 2003 a-Crystallin protects Na/K-ATPase in red cell ghosts (Eur. J. Biochem. 270) 2607 The effects of ouabain, a-crystallin and sugars on the 86 Rb uptake of the ghost cells are shown in Fig. 3. Ouabain (0.1 m M ), a specific inhibitor of Na/K-ATPase, caused a 35% decrease in the 86 Rb uptake within the ghost cells. This demonstrates that the ghosts have functional Na/K- ATPase. The presence of a-crystallin within the ghost cell did not cause any decrease in 86 Rb uptake. The presence of 50 m M sucrose, a nonreducing sugar, on 86 Rb uptake after a 6 h incubation caused no significant decrease in 86 Rb uptake. When the ghost cells were incubated with 50 m M fructose for 6 h the 86 Rb uptake was inhibited by  45%, which is  10% greater than that induced by ouabain. When the experiment was repeated with ouabain and 50 m M fructose, the 86 Rb uptake was inhibited by the same amount as before (fructose alone). This suggests that fructose inhibited all the Na/K-ATPase activity. When the ghost cells were incubated with 50 m M fructose for 6 h with a-crystallin (1 mgÆmL )1 ) resealed inside the ghost cells, the 86 Rb uptake was maintained at  90% of control (P < 0.001 compared to inhibition by fructose). To show that the protection that a-crystallin provided against Na/K-ATPase inactivation was not due to the removal of fructose by binding to a-crystallin, BSA was resealed (in the same manner and concentration as that of a-crystallin) and incubated with 50 m M fructose for 6 h. BSA (1 mgÆmL )1 ) was used as a control protein because it has a greater lysine content than a-crystallin and therefore a greater ability to bind fructose. The resealed BSA did not display any protective activity and the 86 Rb uptake was similar to that with fructose alone, i.e.  45% inhibition (Fig. 3). When the ghost cells were incubated with 0.1 m M methylglyoxal for 6 h the 86 Rb uptake was inhibited by  20% (Fig. 4). When the experiment was repeated but Fig. 2. SDS/PAGE of red cell ghosts with and without a-crystallin resealed. Lanes 1 and 5, molecular mass markers; lane 2, ghost cells alone; lane 3, ghost cells resealed with a-crystallin present, 5 mgÆmL )1 (the double band around 20 kDa indicates presence of a-crystallin); lane 4, a-crystallin alone, 5 mgÆmL )1 (double band around 20 kDa). Fig. 3. Effects of 0.1 m M ouabain, 50 m M sucrose and 50 m M fructose on the 86 Rb uptake into red cell ghosts after 6 h incubation at 37 °C, and the effect of the molecular chaperone a-crystallin (1 mgÆmL -1 )andBSA (1mgÆmL -1 ) separately resealed inside red cell ghosts upon the rubidium uptake of those modifiers. Error bars represent standard deviation. Fig. 4. Effects of 0.1, 1 and 10 m M methylglyoxal and 0.1 m M ouabain on the 86 Rb uptake into red cell ghosts after 6 h incubation at 37 °C, and the effect of the molecular chaperone a-crystallin (1 mgÆmL -1 ) resealed inside red cell ghosts upon the rubidium uptake of those modifiers. Error bars represent standard deviation. 2608 B. K. Derham et al. (Eur. J. Biochem. 270) Ó FEBS 2003 with the addition of ouabain, the 86 Rb uptake was inhibited by  35%, indicating that 0.1 m M methylglyoxal did not inhibit ouabain-sensitive 86 Rb uptake completely. When the ghost cells were incubated with 0.1 m M methylglyoxal for 6hwitha-crystallin (1 mgÆmL )1 ) resealed inside the ghost cells, the 86 Rb uptake was maintained at 100% of control (P < 0.05 compared to inhibition by methylglyoxal). Increasing concentrations of methylglyoxal caused a dose- dependent decrease in 86 Rb uptake. Incubation with 1 m M methylglyoxal inhibited 86 Rb uptake by  40%, the pres- ence of ouabain however, did not change the amount of inhibition suggesting that most of the Na/K-ATPase had been modified. When a-crystallin (1 mgÆmL )1 ) was resealed inside the ghost cells that were incubated with 1 m M methylglyoxal 86 Rb uptake was restored to  90% of the control (P < 0.05 compared to inhibition by methylgly- oxal). At 10 m M methylglyoxal the 86 Rb uptake of the ghost cells was inhibited by  50%, and the presence of ouabain did not inhibit it further. This suggests that 10 m M methylglyoxal was inhibiting other K + permeability path- ways, in addition to Na/K-ATPase. When the experiment was repeated with a-crystallin (1 mgÆmL )1 ) resealed on the inside 86 Rb uptake was maintained at  85% of the control values (P < 0.01 compared to inhibition by methylglyoxal alone). Ghost cells were subjected to oxidative stress in the form of H 2 O 2 . When the ghost cells were incubated with 0.5 m M H 2 O 2 for 6 h the 86 Rb uptake was inhibited by  40%, the additional presence of ouabain did not change the degree of inhibition significantly (Fig. 5). When the ghost cells were incubated with 0.5 m M H 2 O 2 for 6 h with a-crystallin (1 mgÆmL )1 ) resealed inside the ghost cells, the 86 Rb uptake was restored to  85% of control (P < 0.01 compared to inhibition by H 2 O 2 ). When the ghost cells were incubated with 0.5 m M H 2 O 2 , ouabain and a-crystallin, the 86 Rb uptake was approximately the same, as that with ouabain alone, indicating a selective effect of H 2 O 2 on the Na, K pump. To ensure that changes in 86 Rb flux were not due to alterations in ATP concentrations or Na + and K + levels three control experiments were performed. The efficiency of the ATP regenerating system was checked: ATP levels over 6 h were measured, as were concentrations of Na + and K + . Without the ATP regen- erating system active transport via Na/K-ATPase measured as 86 Rb flux is greatly reduced (results not shown). The activity of creatine kinase did not decrease after 6 h incubation with fructose (results not shown). ATP levels within resealed ghost cells at time zero and at 6 h incubations were measured using a luciferin–luciferase enzyme system (Fig. 6). Ghost cells were prepared and incubated with modifiers as described. The ATP levels within the treated ghost cells at time zero were approxi- mately equal to those of the control ghost cells. The control value at 6 h had decreased by  30%, and the modified ghost cells showed similar decreases in ATP levels (Fig. 6). Thus, the differences in 86 Rb flux are not caused by a lowering of ATP. Fig. 5. Effects of 0.5 m M H 2 O 2 and 0.1 m M ouabain on the 86 Rb uptake into red cell ghosts after 6 h incubation at 37 °C, and the effect of the molecular chaperone a-crystallin (1 mgÆmL -1 ) resealed inside red cell ghosts upon the rubidium uptake of those modifiers. Error bars represent standard deviation. Fig. 6. ATP levels in ghost cells subjected to various challenges. ATP levels were measured using a luciferin–luciferase enzyme system using a BioOrbit 1253 luminometer. Ghost cells were prepared as described in the methods section and incubated at 37 °C, 10 lLsamplestaken at t ¼ 0andt ¼ 6 h and assayed. Error bars represent standard deviation. Ó FEBS 2003 a-Crystallin protects Na/K-ATPase in red cell ghosts (Eur. J. Biochem. 270) 2609 Flame photometry was used for determining the levels of sodium and K + within the ghost cells at time zero and 6 h. At time zero the control ghost cells had a Na + concentra- tion of approximately 110 m M , which did not decrease over 6 h. All modifiers had zero time Na + concentrations of  100 m M that did not decrease over 6 h. The K + concentration in all ghost cells, control and modified, were  13 m M at time zero and  9m M at time 6 h. Thus the changes in 86 Rb flux were not due to changes in Na + or K + . Discussion The membrane protein Na/K-ATPase of red blood cell ghosts was stressed using externally applied fructose, methylglyoxal and H 2 O 2 and activity measured by 86 Rb uptake. We report for the first time the ability of the chaperone protein a-crystallin, to prevent the inhibition of the membrane-bound enzyme by fructose, methylglyoxal and H 2 O 2 . a-Crystallin was able to protect Na/K-ATPase from inactivation by all the modifiers. a-Crystallin has previously been shown to protect soluble unfolding proteins by forming stable high molecular weight complexes that retain their functional state, but does not refold the proteins back into the native state [7,8,17]. This study implies that a-crystallin protects Na/K-ATPase in a similar manner from the cytosolic side of the ghost cell. The means by which it affords such protection is unknown but presumably inactivation of the enzyme by the modifier is due to the targeting of a domain of the enzyme that is accessible to both the modifier and to intracellular a-crystallin. The cytoplasmic loop of Na/K-ATPase might provide such a target for protection. This would be in keeping with the failure of a-crystallin to reverse ouabain-induced inhibition of the enzyme. It is thought that a-crystallin may act through dynamic interactions, such that the chaperone may prevent further unfolding but not bind to the target protein. Binding of a-crystallin to ghost cell membranes was not seen under experimental conditions (results not shown). This has been previously observed by [25] looking at soluble proteins. It is possible that more severe conditions are necessary for complex formation. The process of ghost cell preparation from red blood cells by lysis followed by size exclusion chromatography pro- duced very pure intact membranes with fully operational ion transporters and an intact cytoskeleton [18]. Production of red blood cell ghosts by hypotonic lysis results in the formation of a large number of pores in the red cell membrane, estimated to be 30 nm in diameter [26]. a-Crystallin was shown to reseal within ghosts (Fig. 2). Other molecules that have been resealed include albumin (70 kDa) [27], ferritin (474 kDa; diameter 8 nm) and gold particles (10–15 nm) [26]. The amounts of a-crystallin to Na/K-ATPase were determined by densitometry of a SDS/PAGE gel (a-crystallin 1 mgÆmL )1 resealed inside a ghost cells), using the b-subunit (35 kDa) of the Na/K- ATPase as reference. The ratio of a-crystallin to the b-subunit was  2 : 1 by mass, a ratio previously observed between target enzymes and a-crystallin when a-crystallin protected the enzyme activity [9]. The decreases in 86 Rb flux caused by fructose, methyl- glyoxal and H 2 O 2 were not caused by impairment of the ATP regeneration system, nor by loss of ATP or by changes in concentrations of Na + and K + levels. These experiments were performed in the absence of a-crystallin, suggesting that the chaperone function of a-crystallin is committed to protecting primarily membrane-bound proteins. The uptake of 86 Rb in the ghost cells with ouabain (0.1 m M ) decreased by  35%. This is the amount of 86 Rb flux across the ghost cell membrane contributed by Na/K- ATPase. The other 65% of K influx presumably reflects an increased K leak pathway in the resealed ghosts. The protection is not a result of a-crystallin reacting with free inactivators as a-crystallin does not simply compete for fructose [7]. Incorporation of radiolabelled fructose with proteins such as lysozyme, which has a similar lysine content, and BSA, which has a greater lysine content, all bind fructose at a similar rate and displayed no chaperone protective ability [7]. Previous experiments have demon- strated that increased incorporation of radiolabelled fruc- tose mirrored a decline in activity of glucose-6-phosphate dehydrogenase [8]. Protection may be via transient dynamic complex formation that would allow enzymes, soluble and membrane bound, to retain their functional state. All modifiers studied here can cross the ghost cell membrane easily; they diffuse through the membrane because they are not charged. There is no active transport for methylglyoxal or H 2 O 2 in the erythrocyte membrane. There is an active glucose-transporter but it is specific for glucose and not for other hexoses; fructose can move through but 15 times slower then glucose. It is thought that this would not be significant to the overall amount of fructose in the cell. The modifiers can all pass through the membrane so a steady state would be achieved at the molarity of the modifying agent. The site of modification differs slightly between modifiers. Fructose reacts with lysine residues, whereas methylglyoxal reacts principally with arginine residues although modifica- tion of lysine and cysteine also occurs [2,28]. H 2 O 2 oxidizes methionine and cysteine residues as well as lipids [29]. Effective defence systems exist intracellularly to reduce these modifications such as catalase, glutathione peroxidase and glutathione. The reduced glutathione-dependent glyoxalase converts methylglyoxal to D -lactate [30]. Abnormalities in Na/K-ATPase activity are thought to be involved in several pathologic states, in particular heart disease, hypertension and cataract. Altered Na + and K + concentrations are observed in many forms of human cataract and correlate with increasing lens colour and with cortical opacification [31]. The change in monovalent cation concentrations may in part be attributed to decreased efficiency of the Na/K-ATPase. Resealing of the molecular chaperone a-crystallin within a red cell ghost, followed by stress from post-translational modifications protected the ghost cell Na/K-ATPase. This type of assay provides additional evidence of the important role of the small heat shock proteins in cell protection. Also, the protection from modification of ghost cell Na/K- ATPase by a-crystallin highlights the diverse nature of molecular chaperones and suggests that protection of Na/K-ATPase is ubiquitous to all cells, not just red cell membranes. Heat stress in NIH3T3 cells causes a transient decrease of aB-crystallin levels from the cytosol as it is translocated reversibly to the membrane providing protein 2610 B. K. Derham et al. (Eur. J. Biochem. 270) Ó FEBS 2003 synthesis is not inhibited [32]. Heat shock of Reuber H35 hepatoma cells did not cause decrease in ouabain-sensitive 86 Rb influx [33], possibly because of transient protection from heat shock proteins. As far as we are aware this is the first report of the protection of a membrane enzyme by a molecular chaperone. Acknowledgements We are grateful to the Wellcome Trust and to the Knoop Trust for a Junior Research Fellowship. We are grateful to Dr Steve Ashcroft, at the Diabetes Research Laboratories, University of Oxford for the use of his luminometer; and to Dr Simon Golding at the Department of Physiology, University of Oxford for the use of the flame photometer. References 1. Kaplan, J.H. (2002) Biochemistry of Na,K-ATPase. Annu. Rev. Biochem. 71, 511–535. 2. Riley, M.L. & Harding, J.J. (1995) The reaction of methylglyoxal with human and bovine lens proteins. Biochim. Biophys. Acta 1270, 36–43. 3. Garner, M.H. & Spector, A. (1985) Glucose-6-phosphate modi- fication of bovine renal Na/K-ATPase: a model for changes occurring in the human renal medulla in diabetes. Biochem. Bio- phys. Res. Commun. 131, 1206–1211. 4. Haik, G.M., Jr., Lo, T.W.C. & Thornalley. P.J. (1994) Methyl- glyoxal concentration and glyoxalase activities in the human lens. Exp. Eye Res. 59, 497–500. 5. VanderJagt,D.L.,Robinson,B.,Taylor,K.K.&Hunsaker,L.A. (1992) Reduction of trioses by NADPH-dependent aldo-keto reductases. Aldose reductase, methylglyoxal, and diabetic com- plications. J. Biol. Chem. 267, 4364–4369. 6. Lo, T.W.C., Selwood, T. & Thornalley, P.J. (1994) The reaction of methylglyoxal with aminoguanidine under physiological condi- tions and prevention of methylglyoxal binding to plasma proteins. Biochem. Pharmacol. 48, 1865–1870. 7. Heath, M.M., Rixon, K.C. & Harding, J.J. (1996) Glycation- induced inactivation of malate dehydrogenase protection by aspirin and a lens molecular chaperone, a-crystallin. Biochim. Biophys. Acta. 1315, 176–184. 8. Ganea, E. & Harding, J.J. (1995) Molecular chaperones protect against glycation-induced inactivation of glucose 6-phosphate dehydrogenase. Eur. J. Biochem. 231, 181–185. 9. Hook, D.W. & Harding, J.J. (1998) Protection of enzymes by a-crystallin acting as a molecular chaperone. Int. J. Biol. Macro- mol. 22, 295–306. 10. Derham, B.K. & Harding, J.J. (1999) Alpha-crystallin as a molecular chaperone. Prog. Retin. Eye Res. 18, 463–509. 11. Ingolia, T.D. & Craig, E.A. (1982) Four small Drosophila heat shockproteinsarerelatedtoeachotherandtomammalian a-crystallin. Proc. Natl Acad. Sci. USA 79, 2360–2364. 12. Horwitz, J. (1992) a-Crystallin can function as a molecular chaperone. Proc. Natl Acad. Sci. USA 89, 10449–10453. 13. Farahbakhsh, Z.T., Huang, Q.L., Ding, L.L., Altenbach, C., Steinhoff, H.J., Horwitz, J. & Hubbell, W.L. (1995) Interaction of a-crystallin with spin-labelled peptides. Biochemistry 34, 509–516. 14. Raman, B., Ramakrishna, T. & Rao, C.M. (1995) Rapid refolding studies on the chaperone-like a-crystallin. Effect of a-crystallin on refolding of b-andc-crystallins. J. Biol. Chem. 270, 19888–19892. 15. Wang, K. & Spector, A. (1995) a-Crystallin can act as a chaperone under conditions of oxidative stress. Invest. Ophthalmol. Vis. Sci. 36, 311–321. 16. Derham, B. & Harding, J.J. (2002) Enzyme activity after resealing within ghost erythrocyte cells, and protection by a-crystallin against fructose-induced inactivation. Biochem. J. 368, 865–874. 17. Derham, B.K. & Harding, J.J. (1997) Effect of aging on the cha- perone-like function of human alpha-crystallin assessed by three methods. Biochem. J. 328, 763–768. 18. Wood, P.G. & Passow, H. (1981) Techniques for the modification of the intracellular composition of red blood cells. In Techniques in Cellular Physiology (Baker, P.F., ed.), Vol. P1/I. (112), pp. 1–43. Elsevier/North-Holland, County Clare, Ireland. 19. Glynn, I.M. & Hoffman, J.F. (1971) Nucleotide requirements for sodium-sodium exchange catalysed by the sodium pump in human red cells. J. Physiol. 218, 239–256. 20. O’Neill, W.C. (1989) Volume-sensitive, Cl-dependent K trans- port in resealed human erthrocyte ghosts. Am.J.Physiol.256, C81–C88. 21. Brugnara, C., Van-Ha, T. & Tosteson, D.C. (1988) Properties of K + transport in resealed human erythrocyte ghosts. Am. J. Physiol. 255, C346–C356. 22. Stanley, P.E. & Williams, S.G. (1969) Use of the liquid scintillation spectrometer for determining adenosine triphosphate by the luci- ferase enzyme. Anal. Biochem. 29, 381–392. 23. Laemmli, U.K. (1970) Cleavage of structural proteins during the assembly of the head of bacteriophage T4. Nature 227, 680– 685. 24. Bernt, E. & Bergmeyer, H.U. (1963) Creatine phosphokinase. In Methods of Enzymatic Analysis (Bergmeyer, H.U., ed.), pp. 859–862. VCH, Weinheim. 25. Carver,J.A.,Lindner,R.A.,Lyon,C.,Canet,D.,Hernandez,H., Dobson, C.M. & Redfield, C.S.O. (2002) The interaction of the molecular chaperone alpha-crystallin with unfolding alpha-lac- talbumin: a structural and kinetic spectroscopic study. J. Mol. Biol. 318, 815–827. 26. Seeman, P. (1967) Transient holes in the erythrocyte membrane during hypotonic hemolysis and stable holes in the membrane after lysis by saponin and lysolecithin. J. Cell Biol. 32, 55–70. 27. Colclasure, G.C. & Parker, J.C. (1991) Cytosolic protein con- centration is the primary volume signal in dog red cells. J. Gen. Physiol. 98, 881–892. 28. Derham, B.K. & Harding, J.J. (2002b) The effects of modifications of a-crystallin on its chaperone and other properties. Biochem. J. 364, 711–717. 29. Garner, M.H. & Spector, A. (1980) Selective oxidation of cysteine and methionine in normal and senile cataractous lenses. Proc. Natl Acad. Sci. USA 77, 1274–1277. 30. Thornalley, P.J. (1990) The glyoxalase system: new developments towards functional characterisation of a metabolic pathway fundamental to biological life. Biochem. J. 269, 1–11. 31. Marcantonio, J.M., Duncan, G., Davies, P.D. & Bushell, A.R. (1980) Classification of human senile cataracts by nuclear colour and sodium content. Exp. Eye Res. 31, 227–237. 32. Klemenz, R., Frohli, E., Steiger, R.H., Schafer, R. & Aoyama, A. (1991) aB-Crystallin is a small heat shock protein. Proc. Natl Acad. Sci. USA 88, 3652–3656. 33. Boonstra,J.,Schamhart,D.H.,deLaat,S.W.&vanWijk,R. (1984) Analysis of K + and Na + transport and intracellular con- tents during and after heat shock and their role in protein synthesis in rat hepatoma cells. Cancer Res. 44, 955–960. Ó FEBS 2003 a-Crystallin protects Na/K-ATPase in red cell ghosts (Eur. J. Biochem. 270) 2611 . The molecular chaperone a-crystallin incorporated into red cell ghosts protects membrane Na/K-ATPase against glycation and oxidative stress Barry. m M sucrose and 50 m M fructose on the 86 Rb uptake into red cell ghosts after 6 h incubation at 37 °C, and the effect of the molecular chaperone a-crystallin