Color profile: Generic CMYK printer profile Composite Default screen

51

Amyloid insulin interaction with erythrocytes J. Murali, D. Koteeswari, J.M. Rifkind, and R. Jayakumar

Abstract: Erythrocyte membrane interactions with insulin fibrils (amyloid) have been investigated using centrifugation, fluorescence spectroscopy, light scattering, and flow cytometric techniques. The results indicate that insulin fibrils are having moderate affinity to erythrocyte membrane. However, analysis of the apparent dissociation constants of human erythrocyte membranes (leaky and resealed vesicles) with amyloid insulin reveal that the insulin binding is drastically reduced on attaining the fibrillar state compared with native insulin. To understand the role of insulin receptors on erythrocytes binding to amyloid, we have studied the interaction of biotinylated forms of denatured and amyloidic insulin with erythrocytes. FITC-streptavidin was used as a counter staining in flow cytometry measurements. We found that insulin fibrils bind 10 times more with erythrocyte membranes than with amylin and denatured insulin. Key words: insulin amyloid, erythrocyte membrane, amyloid binding, flow cytometry, dissociation constant. Résumé : Les interactions de fibrilles amyloïdes d’insuline avec la membrane érythrocytaire ont été étudiées à l’aide de techniques de centrifugation, de spectroscopie de fluorescence, de diffusion de la lumière et de cytométrie en flux. Les résultats montrent que les fibrilles d’insuline ont une affinité moyenne envers la membrane érythrocytaire. Cependant, l’analyse des constantes de dissociation apparentes de fibrilles amyloïdes d’insuline de membranes érythrocytaires humaines (vésicules perméabilisées et refermées) montre que la liaison de l’insuline est grandement réduite lorsque l’insuline forme des fibrilles comparativement à l’insuline non dénaturée. Afin de comprendre le rôle des récepteurs de l’insuline dans la liaison des fibrilles amyloïdes d’insuline aux érythrocytes, nous avons étudié l’interaction d’insuline dénaturée biotinylée et de fibrilles amyloïdes d’insuline biotinylée avec les érythrocytes. La FITC-streptavidine a été utilisée pour la contre-coloration dans les mesures de cytométrie en flux. Nous avons noté que les fibrilles amyloïdes d’insuline se lient 10 fois plus aux membranes érythrocytaires que l’insuline dénaturée ou l’amyline. Mots clés : insuline, fibrilles amyloïdes, liaison, membrane érythrocytaire, cytométrie en flux, constante de dissociation. [Traduit par la Rédaction]

Murali et al.

59

Introduction Insulin is one among the 22 proteins that are known to form amyloid and also found to be toxic to mammalian cells (Storkel et al. 1983; Brange et al. 1977a, 1977b; Li et al. 2000). The tendency of insulin to undergo structural transformation leading to aggregation and formation of insoluble fibrils has been an intriguing but ill-understood phenomena (Burke and Rougvie 1972). Insulin monomers tend to adopt and maintain three-dimensional structure in which its hydrophobic surfaces are buried by folding and assembly of individual molecules. However, aggregation of insulin with subsequent formation of fibril structures is encountered durReceived 2 July 2002. Revision received 12 November 2002. Accepted 15 January 2003. Published on the NRC Research Press Web site at http://bcb.nrc.ca on 21 February 2003. Abbreviations: FITC, fluorescein isothiocyanate; 12-AS, 12(9-anthroyloxy) stearic acid; βAP, β-amyloid peptide; SAA, serum amyloid-A; sulfo-NHS–biotin, sulfo biotin N-hydroxy succinimide ester. J. Murali, D. Koteeswari, and R. Jayakumar.1 Bio organic and Neurochemistry laboratory, Central Leather Research Institute, Adyar, Chennai 600 020, India. J.M. Rifkind. Section of Molecular Dynamics, National Institute of Aging, Baltimore, U.S.A. 1

Corresponding author (e-mail: [email protected]).

Biochem. Cell Biol. 81: 51–59 (2003)

I:\bcb\bcb8101\O03-009.vp Tuesday, February 18, 2003 10:34:35 AM

ing purification, production, storage, and particularly during the use of infusion pumps (Brange et al. 1987b). During these processes the amyloid formation is initiated by exposure of certain hydrophobic residues, which are normally buried in the interior of the peptide fold, to the surface of the insulin monomer (Brange and Langkjoer 1993; Waugh et al. 1953; Brange et al. 1987a; Sluzky et al. 1991; Glenner et al. 1974). The exposed hydrophobic moieties of the monomeric insulin have the tendency to associate, leading to insoluble fibril formation (Dische et al. 1988). Insulin amyloids are made up of long, unbranched fibrils that look similar to other amyloids, such as βAP, amylin etc. (Perry et al. 1981). Recently, the molecular basis of fibril formation has been studied with 20 different human insulin mutants (Nielsen et al. 2001). In type I diabetes treatment, insulin is injected into the musculature where it typically forms a deposit (Sipe 1992). It then dissolves and travels to the blood where it circulates and acts on its cellular receptors (Brange et al. 1997b). As the insulin injection results in a high concentration at the site of application, amyloidosis during such administration can not totally be ruled out. The ultra-structural morphology of amyloid deposits is characteristically fibrillar, with an unbranched appearance, consisting of numerous fibers of 75– 100 Å in diameter and with variable length (Teplow 1998). Even though insulin-mediated glucose transport is not present in erythrocytes, the occurrence of insulin receptors are observed (Gambhir et al. 1978). The interaction could be

doi: 10.1139/O03-009

© 2003 NRC Canada

Color profile: Generic CMYK printer profile Composite Default screen

52

biologically relevant, since insulin is a primary messenger and exerts its function in a membranous environment (Lee et al. 1997; Klingbeil et al. 1995; Olsen et al. 1996; Zhi-Wen Yu et al. 1996). It should be noted that physiological responses elicited by insulin at target tissues depend on the concentration of insulin near insulin responsive tissues as well as the nature of interaction at target cells (Blundell et al. 1972). Native insulin binding to erythrocytes has been studied in diabetic patients (Okada et al. 1981; Kappy et al. 1980). Development of localized amyloidosis at the site of the porcine insulin in insulin-dependent diabetes mellitus has been observed (Storkel et al. 1983; Sluzky et al. 1991). Few groups have shown that the β-amyloid fibrils have profound deleterious effects on RBC membranes, including increased lipid peroxidation and volume changes (Hajimohammadreza et al. 1990, Mark et al. 1997, Ajmani et al. 2000). Erythrocytes sequester the βAP peptides, showing a preference for binding βAP 1–42 compared with βAP 1–40 (Yu-Min Kuo et al. 2000). The amphoteric and amphipathic characteristics of βAP peptides endow these molecules with a capacity to interact with a large number of plasma proteins such as albumin, α2-microglobulin, α1antichymotrypsin, amyloid P component, complement proteins, transthyretin, apoferritin, apolipoprotein, and lipoproteins (Yu-Min Kuo et al. 2000). As insulin amyloid has similar morphological features to beta amyloid fibrils, it is worth studying the insulin amyloid interaction with erythrocytes. It has been suggested that the ability of amyloid structures to disrupt and (or) aggregate phospholipid vesicles is mediated predominantly through electrostatic interactions involving phospholipid head group (Martinez-Senac et al. 1999), and the nonspecific physicochemical interaction of amyloid peptides with negatively charged membrane. The electrostatic associations therefore may be anticipated to be dependent on the charge of the membrane surface (Van Veen and O’Shea. 1995). The purpose of the present study was to characterize the binding of synthetic insulin fibrils to erythrocytes to contribute to the comprehension of the nature of the amyloid–erythrocyte interaction and also to establish a model system to study. Further, erythrocytes constitute an ideal model membrane system for amyloid interaction studies because of their size and availability. Recently, flow cytometry has become a preferred method for assessment and characterization of protein binding to the cell lines (Berger et al. 1994; Shaw et al. 2000; Ernst et al. 1998). In the present study, we have estimated the binding constant of insulin amyloid fibril with the red cell membranes (leaky and resealed erythrocyte membrane) (Gambhir et al. 1978; Ivarsson et al. 1984). We have used flow cytometry to characterize the amyloid fibril binding to the erythrocytes. Strong binding of acute phase protein serum amyloid-A (SAA) to human neutrophils has been reported using flow cytometry (Linke et al. 1991). By using this analysis, a relationship between protein amyloidogenic activity and amyloid membrane association can be established. Taken together, these observations allow us to describe briefly the nature of interaction and also provide evidence to support the existence of the membrane associated amyloid forms in erythrocyte.

Biochem. Cell Biol. Vol. 81, 2003

Materials and methods Materials Bovine insulin and FITC-streptavidin (Sigma Chemical Co., St. Louis, Mo.), rat amylin (Bachem, Torrance, Calif.), Aβ25–35 (Biosource international, Camarillo, Calif.), anthroyloxy stearic acid (Molecular Probes Inc., Eugene, Oreg.), sulfo- NHS–biotin (Pierce Biotechnology, Rockford, Ill.), streptavidin (SRL, Sisco Research Laboratory, India) were used in this investigation. All other reagent grade chemicals used in this investigation were obtained from Fischer Scientific Company (India). Insulin fibril preparation Bovine insulin was dissolved at 10 mg·mL–1 in milli-Q water and the pH was adjusted to 2 (using 0.1 M HCl). An aliquot of 10 mL of this solution was placed in a glass test tube and sealed. The tube was heated in a water bath maintained between 80 and 100°C until a clear gel was formed. The sample was then cooled and frozen rapidly by immersing in liquid nitrogen and reheated (80–100°C) for approximately 2 min. The process of freezing and reheating was repeated four times until a firm gel was formed. After five freeze–thaw cycles, the fibrils were sedimented by centrifugation at 15 000 × g for 30 min in a microcentrifuge and the pellet was resuspended in water. Fibril preparation procedure was followed as described earlier (Burke and Rougvie 1972; Cardoso et al. 2000). The fibril formation was confirmed by CD measurements in the drying film. ThT fluorescence and congored bireferingence techniques were also carried out to check the fibril formation. In the fibril binding data, the molarity refers to insulin concentration before fibril formation. Human erythrocyte membrane preparation Human venous blood was collected from a healthy donor using trisodium citrate (3.2%) as an anticoagulant. Erythrocytes in the human blood were isolated by the reported procedure (Dodge et al. 1963) and the isolated cells were washed four times with phosphate buffered saline and lysed using 5 mM phosphate buffer (pH 7.4). Leaky membrane was resealed by incubating for 30 min at 37°C in isotonic 5 mM phosphate buffered saline (pH 7.4). The protein concentration ~1 mg/mL of red cell membrane was maintained throughout the experiments. The purity of the cell membrane was confirmed by SDS–PAGE method (Lammeli 1970). Erythrocyte ghosts labeling by 12-(9-anthroyloxy) stearic acid 12-(9-anthroyloxy) stearic acid (12-AS) in 1:20 weight ratio to the estimated protein in the erythrocyte ghosts (Lowry et al. 1951) was added and incubated for 2 h. The cells were then washed twice with 5 mM phosphate buffer (pH 7.4) and the 12-AS labeled membrane samples were used to study the fluorescence and light scattering measurements (Shaklai et al. 1977b; Salhany et al. 1998; Salhany et al. 1980). Centrifugation measurements Centrifugation measurements of insulin amyloid were performed using Remi centrifuge (India) with rotor SS-34 as © 2003 NRC Canada

I:\bcb\bcb8101\O03-009.vp Tuesday, February 18, 2003 10:34:36 AM

Color profile: Generic CMYK printer profile Composite Default screen

Murali et al. Fig. 1. Insulin fibril binding studies using centrifugation method. Centrifugation measurements of insulin fibril binding to human erythrocyte membranes: leaky membrane (×) and resealed (䊊). The absorbance at 275 nm for free insulin fibril present in the supernatant at pH 7.4.

described earlier (Salhany et al. 1998; Salhany et al. 1980; Jeong Hyokð et al. 1984). Erythrocyte membrane [~1.5 × 108/mL] were incubated with varying concentrations of insulin amyloid for 45 min. Erythrocyte membranes (leaky and resealed) were sedimented by centrifuging at 5000 rpm (2400 × g) and ensured that the sediment is free from insulin amyloid. Each supernatant was aspirated and the concentration of unbound insulin amyloid was determined by measuring the absorbance at 275 nm using UV–Vis spectrophotometer (model Shimadzu-160A). Since insulin adheres to the surface of the plastic centrifuge tubes, we have used glass tubes in all the experiments (Jeong Hyokð et al. 1984). Fluorescence spectroscopy Fluorescence spectroscopy was measured using Hitachi model 650-40 fluorescence spectrophotometer. A 150 W Xenon lamp was used as the excitation light source. Excitation and emission slit widths were set at 5 nm. For each measurement, membrane sample was incubated in different concentration of insulin amyloid at 21°C for 1 h and then measured. The samples were excited at 360 nm and the fluorescence intensity at 450 nm was recorded (Shaklai et al. 1977b; Salhany et al. 1998; Salhany et al. 1980; Jeong Hyokð et al. 1984; Lakowicz 1983; Thulborn et al. 1979). Light scattering measurements Light scattering measurements were obtained on a CD spectropolarimeter (model Jasco 715). The erythrocyte membrane was taken in phosphate buffered saline at pH 7.4. About 1.5 × 108 erythrocytes per millilitre was incubated with 10–100 µM insulin fibril. The results were expressed in terms of ∆LS. The data presented have is an average of three scans per sample with a time constant of 2 s bandwidth of 1 nm and sampling at the interval of 0.5 nm starting from 350 to 500 nm. The sample was equilibrated at room temperature, 1 h before data acquisition (Berne and Pecora 1976; Shaklai et al. 1977a; Salhany et al. 1998; Dathe et al. 1990).

53 Fig. 2. Comparison of insulin fibril binding to resealed and leaky membranes by Scatchard plot analysis. Scatchard plot of insulin fibril binding to erythrocyte membrane under stoichiometric conditions. Data from centrifugation analysis were plotted which shows both low and high affinity binding sites for both leaky (×) and resealed (䊊) membranes.

Biotinylation of insulin Insulin was biotinylated with sulfo-NHS–biotin according to the manufacturer’s specifications. Briefly, 1 mg of insulin was dissolved in 1 mL of 50 mM sodium bicarbonate buffer, pH 6.0, containing 30% acetonitrile and allowed to react with 0.32 mg of sulfo-NHS–biotin for 2 h at 4°C. Unreacted biotin was removed from the mixture by centrifugation at 1000 × g for 30 min and subsequently washed three times with phosphate buffer. The final concentration of the sample used was 100 µM. Biotinylated insulin was later converted into amyloidic fibril as described in the procedure (Cuatrecases and Parikh 1972; and Koudinov et al. 1994). Analysis of insulin amyloid binding by flow cytometry Freshly washed human erythrocytes (200 µL; 106 cells/mL) were incubated with biotinylated denatured and amyloidic insulin (each 10, 20 µM, respectively) for 30 min at room temperature. They were washed three times with 500 µL of phosphate buffered saline, pH 7.4 and resuspended in 500 µL of the same buffer. The non-specific binding of biotinylated, denatured insulin to erythrocytes was studied by counter-labeling with FITC-streptavidin. After incubation for 15 to 30 min in the dark at 4°C, aliquot of 20 µL of the sample was used for analysis. The total FITCstreptavidin on intact erythrocytes was determined by flow cytometry in a FACScan equipped with 488 nm argon ion laser and a 530 nm band pass filter for FITC emission light scatter and fluorescence signals from 10 000 cells were recorded using linear and logarithmic amplifications. Results were expressed in arbitrary units as the mean fluorescence intensity (Berger et al. 1994; Shaw et al. 2000; Ernst et al. 1998; Linke et al. 1991). The rat amylin was used as a negative control because of its non-amyloidogenic and non-toxic property, whereas, Aβ25–35 was used as a positve control (Lorenzo et al. 1994). © 2003 NRC Canada

I:\bcb\bcb8101\O03-009.vp Tuesday, February 18, 2003 10:34:36 AM

Color profile: Generic CMYK printer profile Composite Default screen

54

Biochem. Cell Biol. Vol. 81, 2003 Table 1. Erythrocyte membrane binding properties of insulin fibrils by different methods. Experiment

Temperature (K)

Centrifugation High affinity Low affinity Fluorescence method Light scattering method

277

298 298

Leaky membrane Kd (µM)

Resealed membrane Kd (µM)

85±20 344±12 300±18 407±40

65±15 169±10 240±20 339±50

Note: Values shown are the mean of three experiments ± standard deviations.

Results Native insulin binding to the erythrocyte membrane has been characterized previously (Jeong Hyokð 1984). In the investigation, we have employed centrifugation, fluorescence spectroscopy (Salhany et al. 1998), light scattering, and flow cytometric methods to characterize the amyloid binding to erythrocytes. Centrifugation measurements The interaction of insulin fibril to leaky and resealed membrane is shown in Fig. 1. The linear plot indicates a moderate binding to the membrane indicating the fibril formation doesn’t totally prevent the binding. Further analysis of the insulin binding data, according to Scatchard plot, resulted in a curve with an upward concave configuration (Fig. 2). Insulin fibrils have both high and lower affinity binding similar to native insulin. The dissociation constants (Kd) determined for leaky and resealed membrane are given in Table 1. Binding studies were also carried out with leaky membranes to determine the sidedness of the fibril binding site. Dissociation constant does not vary much with leaky membrane when compared with resealed membrane (Table 1) indicating that the binding of fibrils with leaky and resealed membranes is the same. Fluorescence measurements The fluorescence spectrum of 12-AS is shown in Fig. 3. The addition of insulin fibril to the 12-AS treated membranes resulted in a decrease in fluorescence intensity. The Stern–Volmer equation was applied in this investigation to calculate the binding constant (Kq) of insulin fibril (Shaklai et al. 1977b). [1]

I0/I – 1 = Kq[Q]

Where, Io is the fluorescence intensity before the addition of quencher, I is the fluorescence intensity after the addition of quencher, and Q is the concentration of the insulin amyloid (quencher) (Fig. 4). The number of binding sites were calculated by the procedure reported in the literature (Lehrer. 1971). Light scattering measurements Light scattering changes associated with insulin fibril binding to leaky and resealed membranes were studied. The signal change (∆LS) is directly proportional to the stoichiometric binding of insulin amyloid to membranes, es-

tablishes the validity of this method for measuring the insulin fibril binding process (Shaklai et al. 1977a). Construction of saturation curves for the high-affinity component in insulin fibril binding shows that the maximal light scattering change (∆LS) is significantly reduced in resealed membrane (0.163) compared with leaky membrane (0.31) (Fig. 5a). Double reciprocal plots of the data indicate that the apparent Kd for binding (Table 1) did not change significantly for leaky and resealed membrane (x axis intercept in Fig. 5b) as observed in centrifugation method. By fitting a simple hyperbolic function to the data the Kd values were calculated for insulin fibril binding, [2]

∆LS = [∆LSmax(insulin fibril)] Kd + (insulin fibril)

The Kd values obtained for both types of membranes are given in Table 1. Flow cytometric measurements Flow cytometric technique was used to measure the changes in mean fluorescence intensity with control, denatured, and amyloidic insulin incubated intact red cells. The difference in the mean fluorescence intensity (FLm) and the percentage of insulin amyloid binding to erythrocytes are shown in Table 2. Insulin amyloid, denatured insulin and Aβ25–35 peptide incubated red cells showed positive fluorescence, while the amylin (rat) incubation did not result in significant fluorescence change. The experiment adopted in this investigation clearly indicates that FITC-streptavidin binding is significantly higher in insulin amyloid incubated (bound) cells when compared with cells interacted with denatured insulin. The mean fluorescence intensity in the denatured insulin bound cells is only 14.5% of the insulin amyloid bound cells. The observed signal change is directly proportional to the binding of the amyloid to erythrocytes establishes the validity of this method. Other than the fluorescence changes, the forward and side scatter is also observed on binding with amyloid fibrils (data not shown) indicating changes in the volume and granulation (Berger et al. 1994; Shaw et al. 2000; Ernst et al. 1998; Linke et al. 1991).

Discussion The present study indicates that insulin in a fibrous amyloidic form have moderate binding activity. Our results indicate that insulin amyloid shares several commonalties with reference to βAP interaction with erythrocytes (Yu-Min Kuo et al. 2000). Other than the structure of insulin amyloid, © 2003 NRC Canada

I:\bcb\bcb8101\O03-009.vp Tuesday, February 18, 2003 10:34:36 AM

Color profile: Generic CMYK printer profile Composite Default screen

Murali et al.

55 Fig. 3. Fluorescence titration of insulin fibril with resealed (a) and leaky (b) erythrocyte membrane. Fluorescence spectra of bril. 12AS labeled erythrocyte membrane: resealed (a) and leaky (b) membranes} with different concentrations of insulin fibrils. The emission Spectra of 12-AS labeled erythrocyte membrane (~1 × 108 cells/mL) obtained at excitation of 360 nm following the addition of insulin fibril (10–100 µM). Spectra 1–7 represents decrease in fluorescence as increment with 10 µM of insulin fibril.

Fig. 4. Stern–Volmer plot of insulin fibril binding to erythrocyte. Membrane. Stern-Volmer plot of insulin fibril binding to erythrocyte membrane (leaky (×) and resealed (䊊) membranes). From the curve, half saturation concentration (Kd) value for the insulin fibril-erythrocyte interaction was calculated.

which also has cross β-structure as in the case of βAP, insulin amyloid binds to the membrane with a low number of binding sites like that of βAP. Monomeric βAP has about 12 binding sites for each intact erythrocyte (Yu-Min Kuo et al. 2000). This is comparable to the present AS fluorescence

data suggesting ~16 ± 2 and ~9 ± 2 binding sites for the leaky and resealed and membranes, respectively (Mendel et al. 1985). The approximate doubling of binding sites in the case of leaky membrane indicate that the number of binding sites for amyloid interaction is the same on both the sides of the membrane. However, the dissociation constant is lower in the case of βAP with erythrocyte when compared with insulin amyloid indicating that monomeric βAP binds more strongly to erythrocyte membrane than insulin amyloid (Lee et al. 1997). The higher binding of βAP to the erythrocyte can be correlated to its hydrophobicity when compared with the insulin. The hydrophobic to hydrophilic amino acid ratio for Aβ1–40 and Aβ1–42 are 0.74 and 0.83, respectively, and the binding constant linearly increased with this ratio (White and Wimley 1998). In the present investigation the hydrophobic to hydrophilic ratio for insulin comes to 0.54, which accounts for lower binding capacity of insulin amyloid to the membrane. It should be noted in this context that insulin like any other amyloid forming peptides, shows a cross βstructure, and therefore it may be interacting with the hydrophobic core of the membrane. The fluorescence of 12-AS by the insulin amyloid indicates either moderate penetration of insulin amyloid or amyloid binding indirectly triggering the change in the microenvironment of the probe by which it is quenched. The N-(12-anthroyloxy) stearic acids are largely useful in quantitative studies of a variety of membrane interactions (Thulborn et al. 1979). The observation of quenching © 2003 NRC Canada

I:\bcb\bcb8101\O03-009.vp Tuesday, February 18, 2003 10:34:37 AM

Color profile: Generic CMYK printer profile Composite Default screen

56 Fig. 5. Light scattering measurements of insulin fibril binding to leaky and resealed membranes. The change in light scattering signal (∆LS) showed saturation behavior for the high affinity binding component (a), double reciprocðl plots (b) were linear, indicating the involvement of one class of binding sites over this insulin fibril concentration range. The lines through the data in both “a” and “b” represent fits to equation of the text with constants. Resealed membrane (䊊), ∆LSmax = 0.163, Kd = 339 ± 50 µM; leaky membrane (×), ∆LSmax 0.31, Kd = 407 ± 40 µM.

by fibrils indicate the fibril-binding site is in proximal to the fluorophor. Fluorescence quenching by insulin to 12-AS is attributed to disulfide groups of insulin (White and Wimley 1998). The quenching behavior in terms of fluorescence intensity depends on the concentration of 12-AS excited species and their location in the insulin–membrane complex. As insulin fibrils fail to quench the 12-AS fluorescence in the absence of the membrane, the membrane insulin complex formation is an essential requirement for the quenching of the 12-AS (Undefriend. 1969; Lehrer 1971; Pinto et al. 1991). Thus the fluorescence results indicate that the insulin fibrils are having access site nearer to the 12-AS rather than in the extra cellular aqueous phase. Thus the structure of insulin fibrils changes on binding with the membrane. Light scattering from protein assemblies whose dimension

Biochem. Cell Biol. Vol. 81, 2003

exceeds ~0.05 the wavelength of light will be dominated by optical interference effects. The interference effect will be proportional to the assembly shape and size in the membrane (Salhany et al. 1998). This dependence is used to study cellular size and shape. The light scattering technique has been previously applied to the intact erythrocyte volume changes. This technique is also used to determine the structure of erythrocyte ghosts (Salhany et al. 1998). The association of insulin fibril with AS-labeled membrane is evidenced by increase in the absorption values ranging between 0.09 and 0.31 in leaky membrane, between 0.01 and 0.163 in resealed membrane reveals efficient binding (Salhany and Shaklai 1979). Insulin B-chain is mainly composed of highly hydrophobic patches (residues 25–30), which are involved in amyloid fibril formation. In case of native insulin residues 25–30 are freely available for the more avid binding to its erythrocyte receptors (Brange and Langkjoer 1993). To determine cytotoxicity of insulin fibrils, a shared characteristic of human amyloids like βAP (Schubert et al. 1995), cells were also examined for changes in side scatter using flow cytometry. Earlier reports suggest that in addition to Aβ, human amylin and β2-microglobulin, which accumulate as amyloid in patients with non-insulin-dependent diabetes mellitus and insulin-dependent diabetes mellitus and kidney diseases, respectively, are toxic to nerve cells. Rat amylin, which differs in several amino acids from the human form, is not toxic (Lorenzo et al. 1994). The binding of biotinylated insulin amyloid, rat amylin, and denatured insulin to erythrocyte membrane was analyzed by flow cytometry. The binding of insulin amyloid with erythrocytes is indicated in Table 2. Cellular mean fluorescence intensities were measured after incubation with insulin amyloid for 30 min at 30°C, it was observed that binding of amyloid insulin to erythrocytes was significantly higher when compared with denatured insulin and amylin. Flow cytometric studies were carried out to study white blood cell interaction with another amyloid forming proteinSAA (monomer). The binding of SAA protein to white blood cells with flow cytometry has been reported (Linke et al. 1991). This binding of SAA was specific as demonstrated by the inhibition of unlabelled SAA or its proteolytic derivative AA. The inhibition of SAA-binding by the AA indicated that the N-terminal portion of SAA-binding contained the ligand for binding to the cell surface. It was reported that the SAA bound to neutrophils and monocytes, but not to erythrocytes and lymphocytes. Using flow cytometric analysis, we found that an increase in the mean fluorescence intensity and side scatter showing the binding of insulin amyloid with erythrocytes. In contrast, we have detected little (or no) binding of amylin with erythrocytes. Flow cytometric methods were employed to rule out that the amyloid toxicity is exclusively mediated via specific cell surface receptors. It was reported with FACS analysis that the protein components of other human amyloidosis, including amylin, calcitonin, and atrial natriuretic peptide, were toxic to clonal and primary cells (Schubert et al. 1995). However, the FACS analysis reveals that the binding of insulin amyloid to erythrocytes is avid than in amyloid forms as do in the denatured form. © 2003 NRC Canada

I:\bcb\bcb8101\O03-009.vp Tuesday, February 18, 2003 10:34:37 AM

Color profile: Generic CMYK printer profile Composite Default screen

Murali et al.

57 Table 2. Flow cytometric analysis. Erythrocyte

Concentration

Mean fluorescence (FLm) (a.u)

Erythrocyte surface binding (%)*

Amylin (rat)

10 20 10 20 10 20 5 10

6.93 6.94 10.85 10.85 53.77 70.77 136.05 480.11

1 1.01 1.56 1.56 7.76 10.21 19.62 69.22

Denatured insulin Insulin amyloid Beta amyloid peptide

*Because quantitative measurements are subject to error owing to the nonratiometric emission of FITC, the data are expressed as a percentage change in relation to control cells, which is arbitrarily set as 1% (Schubert et al. 1995).

In summary, our findings suggest that amyloidic conversion of insulin leads to impaired receptor binding. In addition, nonspecific moderate binding on both sides of the membranes was also observed in insulin fibrils. The altered interactions of these aggregates may influence in the delocalization of insulin amyloid to various vital toxic targeting of insulin amyloid. This higher concentration of insulin fibril in erythrocyte receptor–membrane binding process appears to be less efficient than that for native insulin. Insulin fibril formation, therefore, may results in aberrant biochemical or physiological process caused by insulin. Experimental modulation of insulin amyloid action in cellular system in vitro or in vivo will initiate one or more of these mechanisms. To understand the potency of amyloid forming proteins (insulin), it is necessary to examine in a model system like the erythrocyte, where the receptor exist.

Acknowledgements We are gratefully acknowledging Dr. T. Ramasami, Director, CLRI, Chennai for his kind permission to publish this work. Authors J.M. and D.K. thank CSIR, New Delhi, India for the financial support in the form of senior research fellowship.

References Ajmani, R.S., Metter, E.J., Jayakumar, R., Ingram, D.K., Spangler, E.L., Abugo, O.O., Rifkind, J.M. 2000. Hemodynamic changes during aging associated with cerebral blood flow and impaired cognitive function. Neurobiol Aging, 21: 257–269. Berger, S., Karamanos, Y., Schoentgen, F., and Julien, R. 1994. Characterization and use of biotinylated Escherichia coli K99 Lectin. Biochem. Biophys. Acta. 1206: 197–202. Berne, B.J., and Pecora, R. 1976. Dynamic Light Scattering. Wiley and Sons, New York. pp. 1–206. Blundell, T., Dodson, G., Hodgkin, D., and Mercola, D. 1972. Adv. Protein Chem. 26: 279–402. Brange, J., and Langkjoer, L. 1993. Insulin structure and stability. Pharm. Biotechnol. 5: 315–350. Brange, J., Dodson, G.G., Edwards, D.J., Holden, P.H., and Whittingham, J.L. 1997a. A model of insulin fibrils derived from the X-ray crystal structure of a monomeric insulin despentapeptide insulin). Proteins, 27: 507–516. Brange, J., Hansen, J.F., Havelund, S., and Melberg, S.G. 1987a. In

Advanced models for therapy of insulin-dependent diabetes. Edited by P. Brunetti and W.K. Waldhausl. Raven Press, New York. pp. 85–90. Brange, J., Whittingham, J., Edwards, D., You- Shang, Z., Wollmer, A., Brandenburg, D., Dodson,G., and Finch, J. 1997b. Insulin structure and diabetes treatment. Curr. Sci. 72(7): 470– 476. Brange, J., Skelbaek-Pedersen, B., Langjaer, L., Damgaard, U., Ege, H., Havelund, S., Heding, L. G., Jorgensen, K.H., Lykkeberg, J., Markussen, J., Pingel, M., and Rasmussen, E. 1987b. In Galenics of insulin: the physicochemical and pharmaceutical aspects of insulin and insulin preparations. SpringerVerlag, Berlin, Heidelberg, New york, London, Paris, Tokyo. Burke, M.J., and Rougvie, M.A. 1972. Cross – β protein structures. I. Insulin fibrils. Biochem. 11: 2435–2439. Cardoso, I., Barbosa Pereira, P.J., Damas, A.M., and Saraiva, M.J.M. 2000. Aprotinin binding to amyloid fibrils. Eur. J. Biochem. 267: 2307–2311. Cuatrecases, P., and Parikh, I. 1972. Adsorbents for affinity chromatography. Use of N-hydroxysuccinimide esters of agarose. Biochemistry, 11: 2291. Dathe, M., Gast, K., Zirwer, D., Welfle, H., and Mehlis, B. 1990. Insulin aggregation in solution. Int. J. Peptide. Protein Res. 36: 344–349. Dische, F.E., Wernstedt, C., Westermark, G.T., Westermark, P., Pepys, M.B., Rennie, J.A., Gilbey, S.G., and Watkins, P.J. 1988. Insulin as an amyloid fibril protein at sites of repeated insulin injections in a diabetic patient. Diabetologia, 31: 158–161. Dodge, J.T., Mitchell, C., and Hanahan, D.J. 1963. The preparation and chemical characteristics of Hemoglobin free ghosts of human erythrocytes. Arch. Biochem. Biophys. 100: 119–130. Ernst, J.D., Yang, L., Rosales, J.L. and Broaddus, C. 1998. Preparation and characterization of an endogenously Fluorescent Annexin for Detection of Apoptic cells. Anal. Biochem. 260: 18–23. Gambhir, K.K., Archer, J.A., and Bradley, C.J. 1978. Characteristics of Human erythrocyte Insulin Receptor. Diabetes, 27: 701– 708. Glenner, G.G., Eanes, E.D., Bladen, H.A., Linke, R.P., Termine, J.D. 1974. β-pleated sheet fibrils. A comparison of native amyloid with synthetic protein fibrils. J. Histochem. Cytochem. 22: 1141–1158. Hajimohammadreza, I., Brammer, M.J., Eagger, S., Burns, A. and Levy, R. 1990. Platelet and erythrocyte membrane changes in Alzheimer’s disease. Biochem. Biophys. Acta. 1025: 208–214. Ivarsson, S.A., Nilsson,K.O., Sjoblad, S., and Thorell, J.I. 1984. Insulin binding to erythrocytes in children with type I diabetes © 2003 NRC Canada

I:\bcb\bcb8101\O03-009.vp Tuesday, February 18, 2003 10:34:37 AM

Color profile: Generic CMYK printer profile Composite Default screen

58 mellitus before, and after insulin treatment. Horm. Metab. Res, 16: 171–174. Jeong Hyokð, I.M., Cuppoletti, J., Meezan, E., Rackley, C.E., and Djukim, H. 1984. Distribution of insulin receptors in human erythrocyte membranes insulin binding to sealed right side out and inside out human erythrocyte vesicles. Biochem. Biophys. Acta. 775: 260–264. Kappy, M.S., Plotnick, L.P., Findlay, J.C., Kayne, R.D. 1980. Erythrocyte insulin binding in insulin-dependent diabetes mellitus: lack of relationship to duration and control of diabetes in children and adolescents. Pediatrics, 66: 385–90. Klingbeil, C.K., Gill, G.N., and Cadena, D.L. 1995. Analysis od substrate recognition determinants in a synthetic peptide containing the Tyr 1173 autophosphorylation site of the epidermal growth factor receptor. Arch. Biochem. Biophys. 316: 745–750. Koudinov, A., Matsubra, E., Frangione,B., and Ghiso, J. 1994. The soluble form of Alzheimer’s Amyloid beta protein is complexed to high density lipoprotein 3 and very high density lipoprotein in normal human plasma. Biochem. Biophys. Res. Commun. 205: 1164–1171. Lakowicz, J.R. 1983. In Principles of fluorescent spectroscopy. Plenum Press, New York. Lammeli, V.K. 1970. Cleavage of structural proteins during the assembling of the head of bacteriophage T4. Nature, 227: 680– 685. Lee, J., Pilch, P.F., Shoelson, S.E. and Scarlata S.F. 1997. Conformational changes of the insulin receptor upon insulin binding and activation as monitored by fluorescence spectroscopy. Biochemistry 36: 2701–2708. Lehrer, S.S. 1971. Solute perturbation of protein fluorescence. The quenching of the tryptophyl fluorescence of model compounds and of lysozyme by iodide ion. Biochem. 10: 3254–3263. Li, Z., Joyal, J.L., and Sacks, D.B. 2000. Binding of IRS proteins to calmodulin is enhanced in insulin resistance. Biochem. 39: 5089–5096. Linke, R.P., Bock, V., Valet, G., and Rothe, G. 1991. Inhibition of the oxidative burst response of N-formyl peptide-stimulated neutrophils by serum amyloid A-Protein. Biochem. Biophys. Res. Commun. 176: 1100–1105. Lorenzo, A., Rapzaboni, B., Weir,G. C., and Yankner, B.A. 1994. Pancreatic islet cell toxicity of amylin associated with type-2 diabetes mellitus. Nature, 368:756–760. Lowry, O.H., Rosebrough, N.J., Farr, A.L., and Randall. R.J. 1951. Protein measurements with folin phenol reagent. J. Biol. Chem. 193: 265–275. Mark. R.J., J.N. Keller. I. Kruman, and Mattson, M.P. 1997. Abeta25–35 induces rapid lysis of red blood cells: contrast with Abeta1–42 and examination of underlying mechanisms. Brain Res. 756: 205–214. Martinez-Senac. M., Villalain. J., and Gomez-Fernandez, J. C. 1999. Structure of the Alzheimer β-Amyloid peptide (25–35) and its interactions with negatively charged phospholipid vesicles. Eur. J. Biochem. 265: 744–753. Mendel, C.M., Licko, V., and Kane, J.P. 1985. The effect of ligand heterogeneity on the scatchard plot. J. Biol. Chem. 260: 3451– 3455. Nielsen, L., Frokjaer, S., Brange, J., Uversky, V.N., and Fink, A.L. 2001. Probing the mechanism of insulin fibril formation with insulin mutants. Biochem, 40: 8397–8409. Okada, Y., Arima, T., Okazaki, S., Nakata,V., Yamabuki, T., and Nagashima, H. 1981. Insulin binding to erythrocytes in diabetes mellitus. Acta. Med.Okayama. 35: 273–7. Olsen, H.B., Ludvigsen, S., and Kaarsholm, N.C. 1996. Solution

Biochem. Cell Biol. Vol. 81, 2003 structure of an engineered insulin monomer at neutral pH. Biochem. 35: 8836–8845. Perry, E,K., Oakley, A.E., Candy, J.M., and Perry, R.H. 1981. Properties and possible significance of substance of substance P and insulin fibrils. J. Neurosci Lett. 25, 321–325. Pinto, V.D,.Jamal, J.A.A., Benz, R., Gench, G. and Palmieri, F. 1991. Characterization of SH groups in polin of bovine heart mitochondria. Eur. J. Biochem. 202: 903–911. Rhoades, E., Agarwal, J., and Gafni, A., 2000. Aggregation of an amyloidogenic fragment of human islet amyloid polypeptide. Biochem. Biophys. Acta. 1476: 230–238. Salhany, J.M., and Shaklai, N. 1979. Functional properties of human hemoglobin bound to the erythrocyte membrane. Biochem. 18: 893. Salhany, J.M., Cordes, K.A., and Sloan, R.L. 1998. Characterization of the pH dependence of hemoglobin binding to band 3 evidence for a pH-dependent conformational change within the hemoglobin-band 3 complex. Biochem. Biophys. Acta. 1371: 107–113. Salhany,J.M., Cordes, K.A., and Gaines, E.D. 1980. LightScattering measurements of hemoglobin binding to the erythrocyte membrane. Evidence for transmembrane effects related to a disulfonic stilbene binding to Band 3. Biochem. 19: 1447–1454. Schubert, D., Behl,C., Lesly, R., Brack, A., Dargusch, R., Sagara, Y., and Kimura, H. 1995. Amyloid peptides are toxic via common oxidative mechanism. Proc. Natl. Acad. Sci. U.S.A. 92: 1989–1993. Shaklai, N., Yguerabide, J., Ranney, H.M. 1977a. Classification and localization of hemoglobin binding sites on the red blood cell membrane. Biochemistry, 16: 5593–5597. Shaklai, N., Yguerabide,V., and Ranney, H.M. 1977b. Interaction of hemoglobin with red blood cell membranes as shown by a fluorescent chromophore. Biochemistry, 16: 5585–5592. Shaw, D.W., Embleton, M.J., Westwater, C., Ryan, M.G., Myers,K.A., Kingsman,S.M., Carroll,M.W., and Stern, P.L. 2000. Isolation of a high affinity scFv from a monoclonal antibody recognising the oncofetal antigen 5T4. Biochem. Biophys. Acta 1524: 238–246. Sipe, J.D. 1992. Amyloidosis. Annu. Rev. Biochem. 61: 947–975. Sluzky, V., Tamada, J.A., Klibanov, A.M., and Langer, R. 1991. Kinetics of insulin aggregation in aqueous solutions upon agitation in the presence of hydrophobic surfaces. Proc. Natl. Acad. Sci, U.S.A. 88: 9377–9381. Storkel, S., Schneider, H-M., Muntefering, H. and Kashiwagi, S. 1983. Iatrogenic, insulin-dependent, local amyloidosis. Lab. Invest. 48: 108–111. Teplow, D.B. 1998. Structural and kinetic features of amyloid βprotein fibrillogenesis. Amyloid: Int. J. Exp. Clin. Invest. 5: 121–142. Thulborn, K.R., Tilley, L.M., Sawyer, W.H., and Treloar, F.E. 1979. The use of n-(9-anthroyloxy) fattyacids to determine fluidity and polarity gradients in phospholipid bilayers. Biochem. Biophys. Acta, 558: 166–178. Undefriend, S. 1969. Proteins and peptides. In Fluorescence assay in biology and medicine. Edited by B. Horecker, N.O. Kaplan, J. Marmur, and H.A. Scheraga. Academic press, London, New York. pp. 248–283. Van Veen, M., and O’Shea, P. 1995. The interaction of neurodegenerative fragment of the β-amyloid peptides with phospholipid membranes. Eur. J. Biochem. 245: 355–363. Waugh, D.F., Whilhelmson, D.F., Commerford, S. L. and Sackler, M.L. 1953. Studies of the nucleation and growth reactions of selected types of insulin fibrils. J. Am. Chem. Soc. 75: 2592– 2600. © 2003 NRC Canada

I:\bcb\bcb8101\O03-009.vp Tuesday, February 18, 2003 10:34:38 AM

Color profile: Generic CMYK printer profile Composite Default screen

Murali et al. White, S.H. and Wimely, W.C. 1998. Hydrophobic interactions of peptides with membrane interfaces. Biochem. Biophys. Acta. 1376: 339–352. Yu-Min Kuo., Kokjohn, T.A., Kalback, W., Luehrs, D., Galasko, D. R., Chevallier, N., Koo, E.H., Emmerling, M.R., and Roher, A.E. 2000. Amyloid-β peptides interact with plasma proteins

59 and eryhtrocytes: Implications for their quantitation in plasma. Biochem. Biophys. Res. Commun, 268: 750–756. Zhi-Wen Yu, Wickman, A. and Eriksson, J. W. 1996. Cryptic receptors for insulin-like growth factor II in the plasma membrane of rat adipocytes — a possible link to cellular insulin resistance. Biochem. Biophys. Acta, 1282: 57–62.

© 2003 NRC Canada

I:\bcb\bcb8101\O03-009.vp Tuesday, February 18, 2003 10:34:38 AM