Effects of Bilirubin Molecular Species on Membrane Dynamic Properties of Human Erythrocyte Membranes: A Spin Label Electron Paramagnetic Resonance Spectroscopy Study

Archives of Biochemistry and Biophysics Vol. 387, No. 1, March 1, pp. 57– 65, 2001 doi:10.1006/abbi.2000.2210, available online at http://www.idealibrary.com on

Effects of Bilirubin Molecular Species on Membrane Dynamic Properties of Human Erythrocyte Membranes: A Spin Label Electron Paramagnetic Resonance Spectroscopy Study Maria Alexandra Brito,* Carlos D. Brondino,† ,‡ Jose´ J. G. Moura,‡ and Dora Brites* ,1 *Centro de Patoge´nese Molecular, Faculdade de Farma´cia, Universidade de Lisboa, Avenida das Forc¸as Armadas, 28, 1600-083 Lisboa, Portugal; †Facultad de Bioquı´mica Y Cs. Biolo´gicas, Universidad Nacional del Litoral, Santa Fe, Argentina; and ‡Departamento de Quı´mica (Centro de Quı´mica Fina e Biotecnologia), Faculdade de Cieˆncias e Tecnologia, Universidade Nova de Lisboa, Monte da Caparica, Portugal

Received August 2, 2000, and in revised form November 13, 2000; published online January 26, 2001

Unconjugated bilirubin is a neurotoxic pigment that interacts with membrane lipids. In this study we used electron paramagnetic resonance and the spin labels 5-, 7-, 12-, and 16-doxyl-stearic acid (DSA) to evaluate the depth of the hydrocarbon chain at which interaction of bilirubin preferentially occurs. In addition, we used different pH values to determine the molecular species involved. Resealed right-side-out ghosts were incubated (1– 60 min) with bilirubin (3.4 – 42.8 ␮M) at pH 7.0, 7.4, and 8.0. Alterations of membrane dynamic properties were maximum after 15 min of incubation with 8.6 ␮M bilirubin at pH 7.4 and were accompanied by a significant release of phospholipids. Interestingly, concentrations of bilirubin up to 42.8 ␮M and longer incubations resulted in the elution of cholesterol and further increased that of phospholipids while inducing less structural alterations. Variation of the pH values from 8.0 to 7.4 and 7.0, under conditions of maximum perturbation, led to a change from an increased to a diminished polarity sensed by 5-DSA. Conversely, a progressive enhancement in fluidity was reported by 7-DSA, followed by 12- and 16-DSA. These results indicate that bilirubin while enhancing membrane lipid order at C-5 simultaneously has disordering effects at C-7. Furthermore, recovery of membrane dynamics after 15 min of bilirubin exposure along with the release of lipids is compatible with a membrane adaptive response to the insult. In addition, our data provide evidence that uncharged diacid is the species primarily interacting with the membrane as

1 To whom correspondence should be addressed. Fax: 351 217946491. E-mail: [email protected].

0003-9861/01 $35.00 Copyright © 2001 by Academic Press All rights of reproduction in any form reserved.

perturbation is favored by acidosis, a condition frequently associated with hyperbilirubinemia in premature and severely ill infants. © 2001 Academic Press Key Words: bilirubin; electron paramagnetic resonance; human erythrocytes; lipids; membrane fluidity; membrane order; spin labels.

Unconjugated bilirubin (UCB) 2 is a tetrapyrrole pigment formed by the catabolism of hemoglobin that becomes clinically important when the hepatic excretory processes are impaired in relation to the rate of UCB formation, resulting in jaundice (1). Deposition of UCB in the central nervous system is the major factor causing bilirubin encephalopathy during severe neonatal hyperbilirubinemia (2). The risk of neurological damage in newborns with hyperbilirubinemia is enhanced by acidosis, which is known to reduce the solubility of UCB while favoring its cellular binding (3–5). Evidence indicates that changes in pH from 8.0 to 7.0 induces an increase from 47 to 93% in uncharged diacid proportion and a decrease from 38 to 7% in monoanionic UCB (6). Interaction of UCB with cells appears to involve two distinct mechanisms, one more aggressive, resulting in increased cell death and most likely responsible for the irreversible lesions, and the other more mild, affecting cell function and probably resulting in transitory or 2 Abbreviations used: UCB, unconjugated bilirubin; EPR, electron paramagnetic resonance; SL, spin labels; DSA, doxyl-stearic acid; DTNB, 5,5⬘-dithiobis(2-nitrobenzoic acid); RO, right-side-out; Pl, phospholipids; Ch, cholesterol.

57

58

BRITO ET AL.

reversible damage (7–10). In erythrocytes, whereas low concentrations of UCB stabilize the membrane against hypotonic hemolysis, an induced release of hemoglobin was observed at high concentrations (11). Different levels of toxic manifestations also include alterations of erythrocyte morphology, loss of the normal bilayer asymmetry with exposure of phosphatidylserine, and changes in membrane lipid composition (8, 11–13). It has been proposed that the toxicity of UCB is determined by its association with membrane lipids (14 –16) and studies using infrared spectroscopy have shown that UCB intercalates into the polymethylene chain region of the bilayer and is located between the carbonyl region and the methylene group two carbons from the methyl terminus (17). However, kinetics and thermodynamic parameters of UCB transfer between small unilamellar phosphatidylcholine vesicles indicate a more superficial interaction, and a UCB penetration into the bilayer equivalent to only 5-carbon depth was suggested (18). Therefore, the avidity of UCB molecules for different regions of the membrane and its precise distribution within the lipid regions of the membrane, e.g., the lipid–water interface, the carbonyl region, the methylene chain region, or, eventually, the middle of the bilayer, are still unclear. The electron paramagnetic resonance (EPR) technique of spin labeling has been used as a highly sensitive method to study the interaction of several compounds with different domains of human erythrocyte membranes (19). Spin labels (SL) specific for the lipid bilayer, with the nitroxide group at different positions of the hydrocarbon chain, have been used to establish the depth at which perturbation occurs (20 –22). Although the spin labeling EPR method was previously used to evaluate the disturbance induced by UCB on membrane organization (23), the depth of the hydrocarbon chain at which interaction of UCB preferentially occurs was never investigated. In the current study we employed EPR spectroscopy using 5-, 7-, 12-, and 16-doxyl–stearic acids (DSA) as phospholipid markers to evaluate perturbations of UCB on membrane dynamic properties and to compare effects at different depths of the leaflet. Incubations at pH values of 7.0, 7.4, and 8.0 were also performed to determine which UCB species is primarily involved in erythrocyte membrane structural alterations. Our results allowed us to conclude that alterations of membrane dynamic properties by UCB are maximum at the carbon number 7 region and that binding is favored by acidosis, a condition frequently associated with hyperbilirubinemia in premature and severely ill infants. MATERIALS AND METHODS Chemicals. UCB, 5-DSA, 12-DSA, 16-DSA, 5,5⬘-dithiobis(2-nitrobenzoic acid) (DTNB), acetylthiocholine chloride, and phosphorus standard solution were purchased from Sigma Chemical Co. (St.

Louis, MO). 7-DSA was from Aldrich Chemical Company (Milwaukee, WI). Triton X-100 and cholesterol oxidase enzymatic assay kit were from Boehringer Mannheim (Mannheim, Germany). All other chemicals were of analytical grade and were purchased from Merck (Darmstadt, Germany). Preparation of right-side-out (RO) vesicles. Human blood was collected from healthy adult donors into 50 U/ml sodium heparin. Erythrocytes were separated from plasma and buffy coats by centrifugation at 600g for 10 min at 4°C and washed three times with cold 145 mM NaCl, 5 mM sodium phosphate, pH 7.4 (isotonic saline buffer). Right-side-out vesicles were obtained from freshly prepared erythrocyte membranes by an adaptation of the method of Deziel and Girotti (24). Membranes were isolated by centrifugation at 29,000g for 30 min at 4°C, after hypotonic lysis with 5 mM sodium phosphates buffer, pH 8.0, followed by three washings with the same buffer (25). Aliquots of the ghost suspension (2 ml) were immediately added dropwise, under continuous shaking, to 35 ml of 10 mM sodium phosphate, 100 mM NaCl, 4 mM MgSO 4, 0.01 mM CaCl 2, 5 mM glucose, pH 7.4 (vesiculation buffer), previously warmed at 37°C. Vesiculation was achieved by 45 min of incubation at 37°C under shaking. RO vesicles were isolated by centrifugation at 27,000g for 15 min at 4°C and washed twice with isotonic saline buffer (pH 7.4). Vesicles were further washed with isotonic saline buffer at pH values of 7.0, 7.4, or 8.0 whenever the effect of the UCB ionic species was to be analyzed. Membrane protein content was evaluated by the method of Lowry et al. (26). Characterization of vesicles sidedness by the acethylcholinesterase activity. Acethylcholinesterase is an outer surface membranebound enzyme usually used as a marker of vesicle sidedness and sealing. The determination of the acethylcholinesterase activity was performed in freshly prepared RO vesicles in either the absence or the presence of Triton X-100 (27). Duplicate aliquots (10 ␮l) of freshly prepared vesicles (diluted 1:10 in isotonic saline buffer, pH 7.4) were pipetted to the bottom of 1-cm-path-length semimicrocuvettes and 40 ␮l of isotonic saline buffer was added. Isotonic saline buffer (pH 7.4) with or without 0.2% (v/v) Triton X-100 (50 ␮l) was mixed well with the vesicles and the volume completed to 0.7 ml with Na phosphate (100 mM, pH 7.5). After a 5-min preincubation at 25°C, 50 ␮l of both DTNB (10 mM DTNB, 100 mM Na phosphate, pH 7.0, and 3 mg NaHCO 3 per 8 mg of DTNB) and acetylthiocholine chloride (12.5 mM) solutions were added to the mixture and mixed by inversion, and the increase in absorbance was recorded at 412 nm. A 17.0 increase in optical density per minute corresponds to the formation of 1 ␮mol of product. The acethylcholinesterase activity levels in the preparations of sealed vesicles (n ⫽ 30) were 1.54 ⫾ 0.12 and 1.39 ⫾ 0.05 ␮mol/min/mg protein according to the presence or absence of Triton, respectively, thus indicating that the preparations contained 96.9 ⫾ 0.70% of RO vesicles. Spin labeling. The 5-, 7-, 12-, and 16-DSA were used to examine the characteristics of the membrane at varying depths based on the attachment of the nitroxide group at different positions along the hydrocarbon chain of the stearic acid (28). RO vesicles were separately probed with each marker, prior to UCB interaction. Membrane vesicles were added to a dried film of the SL, previously prepared from a stock solution of the SL at 2.60 mM in chloroform. The solution was evaporated under a nitrogen stream and left for no less than 2 h under vacuum in order to yield systems containing 2 mol% probe (SL/membrane lipids), corresponding to a 25 ␮M SL final concentration in the assay. The SL incorporation was achieved by incubation at 37°C for 30 min for 5-DSA, 60 min for 7-DSA, and 90 min for 12- and 16-DSA, under shaking. These different incubation periods likely led to the same incorporation of each SL into the membranes from experiment to experiment. In addition, distortions of the spectra due to spin–spin interactions were not noticed, as spectra were coincident with others originated by lower concentrations of the marker (1 mol%), having the advantage of a better signal/noise relationship. Moreover, control of the level of SL incor-

EPR STUDIES ON INTERACTION OF BILIRUBIN SPECIES WITH ERYTHROCYTE MEMBRANES poration by double integration of the spectra guaranteed independence of the results from SL concentration. Incubation of UCB with spin-labeled vesicles. UCB was purified according to McDonagh and Assisi (29) and a 8.55 mM UCB stock solution in 100 mM NaOH was prepared just before use. All the experiments with UCB were performed under light protection (tin foil wrapping of the vials and dim light) to avoid photodegradation. For spectroscopic analysis, spin-labeled RO vesicles (1 mg protein/ ml) were incubated at 37°C in the absence (control) or in the presence of UCB under gentle shaking. Three different sets of experiments were performed to evaluate the effects of: (i) different UCB concentrations (3.4, 8.6, 17.1, and 42.8 ␮M) using 15 min of incubation and the pH 7.4; (ii) different times of exposure (1, 5, 10, 15, 30, and 60 min) using 8.6 ␮M UCB and the pH 7.4; and (iii) different pH values (7.0, 7.4, and 8.0) using 8.6 ␮M UCB concentration and 15 min of incubation. Following incubation, RO pellets obtained by centrifugation (11,000g, 15 min) were used for EPR analysis and supernatants from experiments (i) and (ii) for analysis of membrane lipid release. We cannot exclude that self-aggregation of UCB to multimers could have occurred at high concentrations, since we had to exclude albumin from the solutions to preserve the EPR signal in membranes. Nevertheless, in this range of concentrations it is expected to have either monomers and dimers or multimers of UCB interacting with the membrane. EPR spectra. EPR measurements were performed at 9.8 GHz (X-band) on a Bruker EMX EPR spectrometer using a rectangular cavity (Model ER 4102ST) and 100 KHz field modulation frequency. Membrane pellets were sucked into 100-␮l glass capillaries that were sealed at both ends, and the sample capillaries were introduced within standard 4-mm quartz EPR tubes containing silicone oil for thermal stability. The spectra were acquired at 22°C, with modulation amplitude of 0.5 G and microwave power of 20 mW. Spectral analysis. The alterations in the membrane dynamic properties of SL showing restricted motion in the membranes (5-, 7-, and 12-DSA) were evaluated by measuring the outer half-width at half-height of the low-field extremum (⌬l ), as indicated in Fig. 1A. According to Mason et al. (30), the larger ⌬l, the more motion and less order in the local microenvironment reported by the nitroxide group. These authors have proposed that this parameter is more sensitive for detecting small changes in membrane motion than the separation measurements employed in the classical parameter S (31). Our data corroborated this finding since we also obtained higher percentage changes for ⌬l than for S (8.6, 5.0, and 2.4 times for 5-, 7-, and 12-DSA, respectively), which is the reason why only results of the former are reported. In addition, another advantage of the ⌬l index of fluidity over S is its independence from changes in the polarity of the environment around the nitroxide, rendering corrections unnecessary (30). For 16-DSA, which showed a higher degree of motional freedom, the ratio of the heights of the low-field line (h ⫹1 ) and the center line (h 0 ) was taken (Fig. 1B) as an empirical measurement of the organization of membrane lipids (32, 33). An increase in this ratio means that a decrease in membrane organization has occurred. Hereafter, we designate both ⌬l and h ⫹1 /h 0 as motion parameters. Variations in the micropolarity of the label environment were evaluated for all the SL by estimating the isotropic splitting factor a 0 following the expression (34) a 0 ⫽ 1/3共 A // ⫹ 2 A ⬜ 兲, where A // and A ⬜ are the parallel and perpendicular components of the hyperfine tensor of the spin label probe (S ⫽ 21, I ⫽ 1), respectively. Values of A // and A ⬜ were taken from the experimental spectra for all the SL (Fig. 1C), except for 7-DSA, at pH 8.0 since the peak for measuring A ⬜ at the m I ⫽ ⫺1 resonance was not properly defined.

59

FIG. 1. Parameters evaluated in the spectra of nitroxide-labeled stearic acids incorporated in right-side-out vesicles from erythrocyte membranes. (A) Outer half-width at half-height of the low-field extremum (⌬l ), determined for 5-, 7-, and 12-DSA (h, height; h/ 2, half-height). (B) Ratio of the heights of the low-field line (h ⫹1 ) and of the center line (h 0 ) calculated for 16-DSA. (C) Measurements of the maximum ( A//) and minimum ( A⬜) hyperfine splittings, for determination of a 0 , performed for all the spin labels. The spin-labeled stearic acids were intercalated into the lipid phase of the membranes at a concentration of 2 mol%, as described under Materials and Methods. Spectra were recorded at 22°C, 9.8 GHz frequency, 100 KHz field modulation; 20 mW microwave power, and 0.5 Gauss modulation amplitude.

Spectra of both 12- and 16-DSA in RO vesicles showed overlapped SL spectra identical to those observed for free SL in buffer solutions. This small component was subtracted from all the spectra before evaluation of EPR parameters. Extraction and quantification of lipids released from the membrane by UCB. Extraction of lipids from the supernatants collected as indicated above and quantification of both cholesterol (Ch) and phospholipids (Pl) were performed according to previously described methods (8). Results of phospholipids are expressed as lipid phosphorus. Statistical analysis. All data are expressed as means ⫾ standard error of the mean (SEM) from at least five separate experiments.

60

BRITO ET AL.

FIG. 2. Effect of different concentrations of unconjugated bilirubin on polarity and phospholipid motion at different depths of the membrane leaflet. (A) Values of isotropic hyperfine splitting constant (a 0 ), outer half-width at half-height of the low-field extremum (⌬l ) and ratio of the heights of the low and of the center lines (h ⫹1 /h 0 ), as a function of bilirubin concentration. (B) Percentage of altered a 0 and ⌬l as a function of the nitroxide position down the lipid chain of the stearic acid spin labels, obtained with 8.6 ␮M bilirubin. Right-side-out vesicles were prepared from erythrocyte membranes, spin labeled, and incubated with bilirubin for 15 min at pH 7.4, as described under Materials and Methods. Results are means ⫾ SEM of at least five separate experiments. Values of a 0 , ⌬l, and h ⫹1 /h 0 were calculated as indicated in Fig. 1. # P ⬍ 0.10, *P ⬍ 0.05 and **P ⬍ 0.01 from control using ⌬l or h ⫹1 /h 0 ; § P ⬍ 0.05 and §§ P ⬍ 0.01 from control using a 0 ; ‡ P ⬍ 0.05 from labeled carbon atom number 16; ƒ P ⬍ 0.10 from labeled carbon atom number 12.

Comparisons were made using the unpaired two-tailed Student’s t test performed on the basis of equal or unequal variance as appropriate. Differences were considered statistically significant when P values were lower than 0.05 and a statistical trend was considered to be present when P ⬍ 0.10.

RESULTS

Effects of different concentrations of unconjugated bilirubin on polarity and phospholipid motion at different depths of the membrane leaflet. Figure 2A shows the variations obtained for a 0 , ⌬l, and h ⫹1 /h 0 , induced by several concentrations of UCB during 15 min of incubation at pH 7.4 with different SL. The

results indicate that the decrease in a 0 and ⌬l at the 5-DSA level was followed by an increase of the former and of the motion parameters at 7-, 12-, and 16-DSA. These two types of alterations induced by UCB reflect a more ordered bilayer organization (lower ⌬l ) and a reduction in membrane polarity (lower a 0 ) at regions closer to the membrane–water interface simultaneously with a decreased order and increased polarity of the membrane leaflet at more hydrophobic regions. Maximum membrane perturbation was achieved with 8.6 ␮M UCB (0.5 mg/dl), independently of the SL, from which less pronounced changes were produced. To

EPR STUDIES ON INTERACTION OF BILIRUBIN SPECIES WITH ERYTHROCYTE MEMBRANES

61

FIG. 3. Time-dependent effect of unconjugated bilirubin on the polarity of membrane vesicles at different depths of the membrane leaflet. (A) Values of isotropic hyperfine splitting constant (a 0 ). (B) Comparison between spin labels. Right-side-out vesicles were prepared from erythrocyte membranes, spin labeled, and incubated with 8.6 ␮M bilirubin at pH 7.4, as described under Materials and Methods. Results are means ⫾ SEM of at least five separate experiments. Values of a 0 were calculated as indicated in Fig. 1. # P ⬍ 0.10 and *P ⬍ 0.05 from control; § P ⬍ 0.05 from 16-DSA and from 1, 5, 30, and 60 min of incubation.

evaluate the membrane region more perturbed by UCB interaction, data on percentage changes of a 0 and ⌬l relative to control were compared (Fig. 2B). Once more, increased polarity and fluidity at carbon numbers 7, 12, and 16 were in contrast with the decrease at C-5, but enhancement could now be identified as occurring more markedly at C-7 and progressively diminishing toward C-12 and C-16. Time-dependent effect of unconjugated bilirubin on the polarity of membrane vesicles at different depths of the membrane leaflet. Having verified that the highest membrane perturbation occurred with 8.6 ␮M UCB we selected this concentration to evaluate the alterations of membrane dynamic properties as a function of time exposure to UCB. Despite the fluctuations in control data

along the time of incubation, only a single significance was observed between 1 and 60 min for 7-DSA (P ⬍ 0.05). However, significant effects on SL microenvironment polarity were obtained in all the experiments with UCB after 15 min of incubation, although 10 min of interaction had already produced marked alterations (Fig. 3A). Values of a 0 obtained for longer incubations tend to approach control values. This time-course profile was observed for all the SL tested and corroborated by the motion parameters (data not shown). As before, when the results were expressed as percentage changes from control values (Fig. 3B), a reduction of a 0 was noticed for 5-DSA, conversely to the increase obtained for 7-, 12-, and 16-DSA. Again, maximum a 0 variation was presented by 7-DSA and occurred within 10 to 15 min.

62

BRITO ET AL.

FIG. 4. Concentration- and time-dependent release of membrane phospholipids and cholesterol by unconjugated bilirubin. (A) Vesicles incubated with different concentrations of bilirubin for 15 min at pH 7.4. (B) Vesicles incubated with 8.6 ␮M bilirubin at pH 7.4 for 1 to 60 min. Right-side-out vesicles were prepared from erythrocyte membranes as described under Materials and Methods. Phospholipids are expressed in terms of lipid phosphorus. Results are means ⫾ SEM of at least five separate experiments. *P ⬍ 0.05 and **P ⬍ 0.01 from control.

Concentration- and time-dependent release of membrane phospholipids and cholesterol by unconjugated bilirubin. We subsequently examined whether different concentrations of UCB and time of exposure influence the rate of Pl and Ch release from RO vesicles. While Pl significantly increased in the incubation medium by interaction with all concentrations of UCB, apart from 4.3 ␮M, significant elution of Ch only occurred with 42.8 ␮M UCB (Fig. 4A). Nevertheless, concentrations of both Pl and Ch in supernatants correlated with those of UCB (r ⫽ 0.959 and r ⫽ 0.993, respectively, P ⬍ 0.01). As before, time of exposure also influenced the release of Pl more than it did that of Ch (Fig. 4B). In fact, and in contrast with the earlier elution of Pl to the supernatants at 15 min of incubation, only longer periods (60 min) produced equivalent statistical significance for the appearance of Ch in the medium (P ⬍ 0.05). Effect of pH on the alterations of phospholipid polarity induced by unconjugated bilirubin at different depths of the membrane leaflet. To differentiate the effects produced by uncharged diacid from those of anionic UCB species, erythrocytes were incubated at pH values of 7.0, 7.4, and 8.0 under conditions where maximal effects on membrane properties were produced (8.6 ␮M UCB, 15 min). In the presence of UCB,

SL microenvironment polarity increased at pH 7.0 (P ⬍ 0.01) and 7.4 (P ⬍ 0.05) at both C-12 and C-16 regions, whereas it only slightly increased at pH 8.0 (Fig. 5A). In regions closer to the polar–apolar interface of the membrane, as reported with 5-DSA, the decreased polarity environment at pH values of 7.0 and 7.4 (P ⬍ 0.05), changed to an inverse behavior at pH 8.0 (P ⬍ 0.05). Analysis of percentage changes in a 0 by UCB as a function of pH (Fig. 5B) indicated that acidosis was particularly felt by 12-DSA and that polarity presented a significant enhancement as compared to pH 7.4 and 8.0 (P ⬍ 0.05). In contrast, less important changes were obtained for 16-DSA and only differences between pH 7.0 and 8.0 tended to be significant (P ⬍ 0.10). Interestingly, elevation of acidic UCB representation by decreasing the pH from 7.4 to 7.0 did not further decrease the membrane polarity reported by 5-DSA. Evaluation of the influence of pH on the interaction of UCB at C-7 was performed by using ⌬l instead of a 0 analysis, an estimation that was impracticable at pH 8.0. The results indicated that the increased membrane fluidity at pH 7.0 (10.7 ⫾ 1.7%, P ⬍ 0.01) and 7.4 (5.9 ⫾ 1.5%, P ⬍ 0.01) was not reproduced at pH 8.0 (3.0 ⫾ 1.4%, N.S.). To establish whether these effects were also more marked at C-7 than at C-12, in

EPR STUDIES ON INTERACTION OF BILIRUBIN SPECIES WITH ERYTHROCYTE MEMBRANES

63

accordance with data previously obtained for UCB concentration and time of exposure, we further measured variations given by ⌬l for 12-DSA. The lower variations of 8.0 ⫾ 1.2, 4.0 ⫾ 0.8, and 1.6 ⫾ 0.5%, respectively, obtained for the same pH values, are consistent with the concept that C-7 is the region where the most increased fluidity changes occur by uncharged diacid UCB species. DISCUSSION

FIG. 5. Effect of pH on the alterations of phospholipid polarity induced by unconjugated bilirubin at different depths of the membrane leaflet. (A) Values of isotropic hyperfine splitting constant (a 0 ). (B) Percentage of altered a 0 . Right-side-out vesicles were prepared from erythrocyte membranes, spin labeled, and incubated with 8.6 ␮M bilirubin for 15 min at pH values of 7.0, 7.4, and 8.0. Results are means ⫾ SEM of at least five separate experiments. Values of a 0 were calculated as indicated in Fig. 1. *P ⬍ 0.05 and **P ⬍ 0.01 from control; # P ⬍ 0.10, ƒ P ⬍ 0.05 and ƒƒ P ⬍ 0.01 from pH 8.0; § P ⬍ 0.05 from pH 7.4.

Our results showed that UCB while enhancing membrane lipid order at C-5 simultaneously has disordering effects at C-7 and that uncharged diacid is the species involved in the phenomena. While concentrations of 8.6 ␮M UCB at pH 7.4 decreased the amplitude of chain motion of Pl at C-5 (P ⬍ 0.10), they increased the fluidity from C-7 to C-16, with the former reflecting the most intense alteration in phospholipid packing (Fig. 2A). These opposite effects reflected by ⌬l and h ⫹1 /h 0 were paralleled by similar changes in the polarity profile given by a 0 . This was further corroborated even when analysis of percentage changes from control was performed (Fig. 2B) in order to surpass the influence of polarity and order gradients along the depth of the membrane leaflet (31, 34 –36). Therefore, it is conceivable that accommodation of UCB will occur at C-5, which, as a second-hand effect, renders the inner regions of the leaflet more fluid and more permeable to water diffusion. The decrease in membrane perturbing effects from C-7 to C-16 may be understood as a wave progressively attenuating as the distance from the primary interaction target increases. This is in line with the creation of “free volume” in the hydrophobic part of the membrane proposed to occur upon incorporation of amphiphiles (20). Superficial hydrophobic interactions of UCB with the membrane bilayer was previously proposed by others (17, 18, 37) and sustains the appearance of echinocytic forms when erythrocytes are incubated with greater than normal levels of UCB (12, 38). The lack of effect at C-16 as compared with the perturbation sensed at C-7 is also consistent with the proposal of Zakim and Wong (17) that UCB does not interact with the methyl terminal end of the acyl chain. Thus, from our data we may conclude that at least a portion of UCB molecule incorporates into the hydrocarbon core at the 5th carbon depth, as postulated by Zucker et al. (18) based on the free energy of activation for UCB extrapolated from studies of Pownall et al. (39). The increase in membrane perturbation by concentrations of UCB up to 8.6 ␮M, decreasing thereafter and approaching control values at 42.8 ␮M (Fig. 2), raises the possibility that the accommodation of high concentrations of UCB within the lipid bilayer follows the saturation of the polymethylene chains at lower

64

BRITO ET AL.

levels. This is corroborated by comparable profiles when increasing the time of exposure (Fig. 3), further indicating that membrane adaptive responses to UCB interaction may be implicated in the observed phenomena. According to Ali and Zakim (16), a maximum perturbation of the thermal properties of phosphatidylcholine bilayers is observed at UCB concentrations of 0.5 mol%, above which no further disturbance is noticed. Similarly, in the present study we also found 0.7 mol% as the concentration of UCB at which the shift toward control values was obtained. This slight difference may reside in the presence of proteins in our system that are known to increase the partitioning of UCB into bilayers as compared with bilayers of identical lipid composition but lacking those proteins (40). Based on the assumption that maximum UCB solubility at pH 7.4 is close to 100 nM (41) it is conceivable that self-aggregation of UCB over the concentration of 8.6 ␮M should also contribute to the absence of linearity between UCB concentrations and membrane perturbing effects. Thus, if precipitation of UCB occurs, less free molecules are expected to be available to interact with the membranes. In parallel experiments we observed that when addition of UCB changed from 8.6 to 17.1 ␮M, the concentration of unbound UCB almost tripled (from 2.6 ⫾ 0.4 to 6.7 ⫾ 1.4 ␮M, P ⬍ 0.05). Under the same conditions, membrane-bound UCB less than doubled (from 6.0 ⫾ 0.9 to 10.3 ⫾ 0.8 nmol/mg protein, P ⬍ 0.05). Therefore, although UCB binding to membrane did not increase with the same relevance as unbound UCB, the formation of some colloidal UCB particles cannot explain the reversal of the trend observed in Fig. 2A. Moreover, it is worthwhile to mention that no alterations of the physical state of UCB were noticed during the time of incubation as suggested by the identical concentrations of unbound and membrane-associated UCB obtained during the 60-min incubation. In addition, also accounting for the observed phenomenon is the influence that the release of Pl, but overall that of Ch (Fig. 4), might have had in favoring the inclusion of UCB into the membrane (42, 43). While no significant loss of membrane Ch was achieved with either 8.6 ␮M UCB or 15 min of incubation, a marked exfoliation was noticed at the highest concentration (⬃2.7-fold, P ⬍ 0.01) and time of exposure (⬃1.5-fold, P ⬍ 0.05). Then, a maintenance or even a rise in the perturbation by increasing time of exposure and UCB concentration would have been observed rather than the decrease obtained. An explanation for these apparently contradictory findings may be that the entrance of new UCB molecules replaced and further forced the preexisting ones to migrate deeper within the bilayer, counteracting the perturbation effects previously noticed with the formers. This is in line with the hindering of crenation observed when longer than 15–30 min

of incubation with UCB was used (11, 44) and with the requirement of chloroform over albumin to fully extract UCB from erythrocytes incubated with UCB (8). This deeply located or aggregated UCB within the membrane showed, in the present study, to vary along with the concentration of UCB (r ⫽ 0.973, P ⬍ 0.01) and the incubation time (r ⫽ 0.848, P ⬍ 0.10). Hence, more and more UCB molecules would be presumably accommodated within the membrane, a finding that is consistent with increased membrane lipid exfoliation, fusion-like phenomena, and hemolysis observed as a response to UCB concentration (8, 11, 44). It has been proposed that the dianion form of UCB binds to the polar heads of Pl protruding from the outer leaflet, while acid UCB is inserted into the lipophilic region of the bilayers (37). Several lines of evidence indicate that both electrostatic and hydrophobic interactions are involved (42, 45). To confirm these hypotheses we have further evaluated the degree and sort of perturbation that occur in membranes exposed to UCB at pH values of 7.0, 7.4, and 8.0, indicated as favoring acid UCB to 93% or increasing anionic species to 17 and 53%, respectively (6). The slight increase in the polarity environment sensed by all SL at pH 8.0 (Fig. 5) suggests that anionic species interact with the membrane leaflet at a very superficial level, presumably at the polar–apolar or even at the membrane–water interface. These data support the reversibility of toxic effects by albumin washing, accompanied by the removal of the pigment bound to erythrocytes when a molar ratio of UCB to human serum albumin below 1 was used (7, 8). In contrast, the reduction in the polarity environment felt by 5-DSA and the elevation in membrane permeability reported by 7-, 12-, and 16-DSA as pH decreases from 8.0 to 7.4 and 7.0 indicate that protonization of UCB promotes membrane perturbation. In addition, increased effects at C-7 and C-12 by acidic pH indicate that uncharged diacid UCB is the species that gains access into the polymethylene chain region. However, we did not find relevant changes from 7.4 to 7.0 as we were expecting to occur at C-5, probably reflecting the partitioning of the pigment to other regions of the membrane as discussed above. Increased UCB binding to circulating cells or tissues at low pH (4, 5, 46, 47) and a multistep interaction mechanism of UCB with liposomes, erythrocytes, and synaptosomes, culminating in membrane aggregation of acid UCB, were proposed by others (5, 42, 48 –51). These findings support the existence of a firmly UCB-bound fraction (37) only removed by chloroform (8). Collectively, our results show that UCB while enhancing membrane lipid order at C-5 simultaneously has disordering effects at C-7. Recovery of membrane dynamics after 15 min of UCB exposure along with the release of lipids is compatible with a membrane adap-

EPR STUDIES ON INTERACTION OF BILIRUBIN SPECIES WITH ERYTHROCYTE MEMBRANES

tive response to the insult. Moreover, the data support uncharged diacid as the species involved since membrane perturbation by UCB is aggravated by acidosis, a condition frequently associated with hyperbilirubinemia in premature and severely ill infants. ACKNOWLEDGMENTS The authors thank Sofia Marques and Sofia Ferreira for their help during the experimental procedures. This work was supported by Grant PRAXIS/PSAU/C/SAU/127/96 from Fundac¸a˜o para a Cieˆncia e a Tecnologia, Portugal.

REFERENCES 1. Ostrow, J. D. (1986) in Bile Pigments and Jaundice: Molecular, Metabolic and Medical Aspects (Ostrow, J. D., Ed.), pp. 1– 6, Dekker, New York. 2. Blanckaert, N., and Fevery, J. (1990) in Hepatology. A Textbook of Liver Disease (Zakim, D., and Boyce, T. D., Eds.), pp. 254 –302, W. B. Saunders, Philadelphia. 3. Nelson, T., Jacobsen, J., and Wennberg, R. P. (1974) Pediatr. Res. 8, 963–967. 4. Kashiwamata, S., Suzuki, F., and Semba, K. (1980) Jpn. J. Exp. Med. 50, 303–311. 5. Brodersen, R., and Stern, L. (1990) Acta Paediatr. Scand. 79, 12–19. 6. Ostrow, J. D., Mukerjee, P., and Tiribelli, C. (1994) J. Lipid Res. 35, 1715–1737. 7. Silva, R., Mata, L. R., Gulbenkian, S., Brito, M. A., Tiribelli, C., and Brites, D. (1999) Biochem. Biophys. Res. Commun. 265, 67–72, doi:10.1006/bbrc.1999.1646. 8. Brito, M. A., Silva, R., Tiribelli, C., and Brites, D. (2000) Eur. J. Clin. Invest. 30, 239 –247. 9. Silva, R. F. M., Rodrigues, C. M. P., and Brites, D. (2001) J. Hepatol, in press. 10. Silva, R., Rodrigues, C. M. P., and Brites, D. (2000) Proceedings, Europaediatrics 2000, Rome, Italy. 11. Brites, D., Silva, R., and Brito, A. (1997) Scand. J. Clin. Lab. Invest. 57, 337–350. 12. Brito, M. A., Silva, R. M., Matos, D. C., da Silva, A. T., and Brites, D. T. (1996) Clin. Chim. Acta 249, 149 –165. 13. Brito, M. A., Marques, S., Almeida, C., and Brites, D. (2000) Proceedings, Europaediatrics 2000, Rome, Italy. 14. Mustafa, M. G., and King, T. E. (1970) J. Biol. Chem. 245, 1084 –1089. 15. Hayward, D., Schiff, D., Fedunec, S., Chan, G., Davis, P. J., and Poznansky, M. J. (1986) Biochim. Biophys. Acta 8600, 149 –153. 16. Ali, S., and Zakim, D. (1993) Biophys. J. 65, 101–105. 17. Zakim, D., and Wong, P. T. T. (1990) Biochemistry 29, 2003– 2007. 18. Zucker, S. D. J., Storch, J., Zeidel, M. L., and Gollan, J. L. (1992) Biochemistry 31, 3184 –3192. 19. Lee, J., Trad, C. H., and Butterfield, D. A. (1993) Toxicology 83, 131–148. 20. Ondrias, K., Stasˇko, A., and Balgavy´, P. (1987) Biochem. Pharmacol. 36, 3999 – 4005. 21. Gu¨ldu¨tuna, S., Zimmer, G., Imhof, M., Bhatti, S., You, T., and Leuschner, U. Gastroenterology 104, 1736 –1744.

65

¨ ., Korkmaz, M., and Akso¨z, E. (1997) Z. Naturforsch. 22. Korkmaz, O 52, 522–529. 23. Chen, L. X., Lu, J. F., and Wang, K. (1992) Cell Biol. Int. Rep. 16, 567–573. 24. Deziel, M. R., and Girotti, A. W. (1980) J. Biol. Chem. 255, 8192– 8198. 25. Dodge, J. T., Mitchell, C., and Hanahan, D. J. (1963) Arch. Biochem. Biophys. 100, 119 –129. 26. Lowry, O. H., Rosebrough, N. J., Farr, A. L., and Randall, R. J. (1951) J. Biol. Chem. 193, 265–275. 27. Steck, T. L. (1974) Methods Membr. Biol. 2, 245–281. 28. Jost, P., Libertini, L. J., Hebert, V. C., and Griffith, O. H. (1971) J. Mol. Biol. 59, 77–98. 29. McDonagh, A. F., and Assisi, F. (1972) Biochem. J. 129, 797– 800. 30. Mason, R. P., Giavedoni, E. B., and Dalmasso, A. P. (1977) Biochemistry 16, 1196 –1201. 31. Hubbell, W. L., and McConnell, M. (1971) J. Am. Chem. Soc. 93, 314 –326. 32. Schreier, S., Frezzatti, W. A., Jr., Araujo, P. S., Chaimovich, H., and Cuccovia, I. M. (1984) Biochim. Biophys. Acta 769, 231–237. 33. Bianconi, M. L., Amaral, A. T., and Schreier, S. (1988) Biochem. Biophys. Res. Commun. 152, 344 –350. 34. Harris, J., Power, T. J., Bieber, A. L., and Watts, A. (1983) Eur. J. Biochem. 131, 559 –565. 35. Gaffney, B. J., and McConnell, H. M. (1974) J. Magn. Reson. 16, 1–28. 36. Fretten, P., Morris, S. J., Watts, A., and Marsh, D. (1980) Biochim. Biophys. Acta 598, 247–259. 37. Cestaro, B., Cervato, G., Ferrari, S., Di Silvestro, G., Monti, D., and Manitto, P. (1983) Ital. J. Biochem. 32, 318 –329. 38. Kaul, R., Bajpai, U. K., Shipstone, A. C., Kaul, A. H., and Murti, C. R. K. (1981) Exp. Mol. Pathol. 34, 290 –298. 39. Pownall, H. J., Hickson, D. L., and Smith, L. C. (1983) J. Am. Chem. Soc. 105, 2440 –2445. 40. Leonard, M., Noy, N., and Zakim, D. (1989) J. Biol. Chem. 264, 5648 –5656. 41. Hahm, J. S., Ostrow, J. D., Mukerjee, P., and Celic, L. (1992) J. Lipid Res. 33, 1123–1137. 42. Va´zquez, J., Garcı´a-Calvo, M., Valdivieso, F., Mayor, F., and Mayor, F., Jr. (1988) J. Biol. Chem. 263, 1255–1265. 43. Zucker, S. D., Goessling, W., Zeidel, M. L., and Gollan, J. L. (1994) J. Biol. Chem. 269, 19,262–19,270. 44. Brites, D. (1988) Bilirrubina: Contribuic¸a˜o para o estudo do mecanismo da sua acc¸a˜o to´xica com particular relevaˆncia para o perı´odo neonatal precoce, Ph.D. thesis, University of Lisbon. 45. Carey, M. C., and Koretsky, A. P. (1979) Biochem. J. 179, 675– 689. 46. Bratlid, D. (1972) Scand. J. Clin. Lab. Invest. 29, 453– 459. 47. Wennberg, R. P. (1988) Pediatr. Res. 23, 443– 447. 48. Brodersen, R. (1979) J. Biol. Chem. 254, 2364 –2369. 49. Eriksen, E. F., Danielsen, H., and Brodersen, R. (1981) J. Biol. Chem. 256, 4269 – 4274. 50. Glushko, V., Thaler, M., and Ross, M. (1982) Biochim. Biophys. Acta 719, 65–73. 51. Amit, Y., Fedunec, S., Thomas, P. D., Poznansky, M. J., and Schiff, D. (1990) Biochim. Biophys. Acta 1055, 36 – 42.

The author has requested enhancement of the downloaded file. All in-text references underlined in blue are linked to publications on ResearchGate.