Carbamylation of erythrocyte membrane aminophospholipids: an in vitro and in vivo study

Clinical Biochemistry, Vol. 29, No. 4, pp. 333-345, 1996 Copyright © 1996 The Canadian Society of Clinical Chemists Printed in the USA. All rights reserved 0009-9120/96 $15.00 + .00 ELSEVIER

S0009-9120(96)00018-5

Carbamylation of Erythrocyte Membrane Aminophospholipids: an in Vitro and in Vivo Study DANIEL J. TREPANIER and ROGER J. THIBERT* Department of Chemistry and Biochemistry, University of Windsor, Windsor, Ontario, N9B 3P4 Canada Objectives: To study the binding of cyanate to erythrocyte membrane aminophospholipids in vitro, and to investigate whether carbamylated aminophospholipidscan be detected in the plasma membrane of native erythrocytes. Design and Methods: For in vitro studies, the lipid components of 14C-carbamylated erythrocyte membranes were resolved by thinlayer chromatography (TLC). The covalent incorporation of cyanate was visualized by autoradiography and quantitated by phosphorus analysis. For the in vivo studies, phospholipid headgroups were enzymatically hydrolyzed by phospholipase D and subsequently reacted with diacetyl monoxime. Results: Both phosphatidylethanolamine(PE) and phosphatidylserine (PS) were covalently modified by [14C] cyanate; incorporating 15.76 ± 0.09 and 13.34 ± 0.81 mol%, respectively, following a 15-h incubation. Carbamylated PE (carb-PE) was resolved with PE by TLC in a solvent system consisting of chloroform~methanol~ammonia (65/35/5, v/v/v). Treatment of native erythrocyte membrane lipid micelles with phospholipase D, followed by reaction with diacetyl monoxime, suggests the presence of intrinsic carb-PE (2.85 ± 0.65 percent of the total PE). Conclusions: Carbamylation of erythrocyte aminophospholipidmay be involved in some of the hematological consequences of uremia on the erythrocyte.

KEY WORDS: carbamylation; erythrocyte ghosts; phosphatidylethanolamine; phosphatidylserine; phospholipase D; uremic; diacetyl monoxime; cyanate. Introduction

s a consequence of chronic renal failure (uremia),

A the plasma levels of urea are greatly increased

(30-50 mM) above normal (3-7 mM). In the uremic state, urea is not considered a toxic agent. However, isocyanic acid, the reactive form of cyanate, is in

equilibrium with urea at physiological pH and temperature (1), and has been shown to form a stable adduct with protein amino groups (carbamylation) (2). Carbamylated hemoglobin (3), plasma proteins (4), and erythrocyte membrane protein (5) have been shown to be elevated in uremic individuals. The reaction between cyanate and protein is, clearly, amino group-specific rather than protein-specific. It is, therefore, considered a possibility that carbamylation may extend beyond the level of protein to other nonprotein primary amino group-containing compounds. One such potential target is the plasma membrane of the erythrocyte. The erythrocyte membrane contains nearly equivalent amounts (w/w) of lipid and protein and, together, phosphatidylserine (PS) and phosphatidylethanolamine (PE) make up approximately 35 mole% of the total erythrocyte membrane phospholipid (6). Both of these phospholipids contain a primary amino group in their polar headgroup, and are potentially accessible to derivatization because they are relatively easy to label with exogenous amino-specific chemical reagents such as trinitrobenzene sulfonic acid (TNBS) (7). We decided, therefore, to study the binding of cyanate to erythrocyte membrane aminophospholipids - - PS and PE in vitro, and to investigate whether or not carbamylated aminophospholipids can be detected in the plasma membrane of native erythrocytes. Materials and methods MATERIALS

*Correspondence: Dr. R.J. Thibert, Department of Chemistry and Biochemistry, University of Windsor, Windsor, Ontario, N9B 3P4, Canada. Manuscript received December 1, 1995; revised and accepted February 7, 1996. Abbreviations: TLC, thin-layer chromatography; Chol, cholesterol; PA, phosphatidic acid; PE, phosphatidylethanolamine; PC, phosphatidylcholine; PS, phosphatidylserine; SM, sphingomyelin; Carb-PE, carbamylated phosphatidylethanolamine; Carb-PS, carbamylated phosphatidylserine; PLD, phospholipase D; Cer, ceramide. C L I N I C A L B I O C H E M I S T R Y , V O L U M E 29, A U G U S T 1996

Whole blood (in tubes containing ethylenediaminetetraacetic acid (EDTA) as anticoagulant) was provided by the hematology departments at the Hotel Dieu Hospital (Windsor, ON, Canada) and the Salvation Army - - Grace Hospital (Windsor, ON, Canada). The following materials were purchased from BDH Chemicals (Toronto, ON, Canada): sodium phosphate (mono and diphosphates); sodium chloride; Tris-[hydroxymethyl]-aminomethane hydro333

TREPANIERAND THIBERT chloride (Tris-HC1); chloroform (OmniSolv®); methanol (OmniSolv®); glacial acetic acid; perchloric acid; hydrochloric acid; diacetyl monoxime; manganese chloride; ethanolamine; mercury; phenol reagent (Folin and Ciocalteu reagent); sulfuric and orthophosphoric acids. The following materials were purchased from the Sigma Chemical Company (St. Louis, MO, U.S.A.): phenylmethysulfonylfluoride (PMSF); choline chloride; L-serine; urea; ethylenediaminetetraacetic acid (EDTA); ascorbic acid; calcium chloride; Triton X-100; hydrogen peroxide (30%); Sigma-FluorTM scintillation fluid; phospholipase D (PLD) from Streptomyces chromofuscus; lipids (egg yolk L-~-phosphatidylethanolamine (Type III); L-~-phosphatidylcholine (Type III-E); sphingomyelin; bovine brain L-~-phosphatidylserine; anhydrous cholesterol; ammonium molybdate tetrahydrate; and bovine serum albumin (BSA). Sodium cyanate was from Aldrich Chemical Company (Milwaukee, WI, U.S.A.). Potassium [14C] cyanate (1 mCi) was purchased from Amersham International (Buckinghamshire, U.K.). Sodium azide, potassium (mono and diphosphates), sodium carbonate, copper (II) sulfate pentahydrate and sodium potassium tartrate were obtained from Fisher Scientific Co. (Toronto, ON, Canada). Ultrapure electrophoresis reagents (acrylamide, ammonium persulfate, Bromophenol Blue, 2-mercaptoethanol, N,N ' - m e t h y l e n e - b i s - a c r y l a m i d e ) , and Coomassie Brilliant Blue R-250 were from Bio-Rad Laboratories (Richmond, CA, U.S.A.). Sodium dodecyl sulphate (SDS) was obtained from United States Biochemical Corporation (Cleveland, OH, U.S.A.). Cadmium sulfate was from Mallinckrodt Chemical Works (St. Louis, MO, U.S.A.). Cadmium chloride was obtained from Anachemia Chemicals (Montreal, QUE, Canada). Thiosemicarbazide was from Eastman Organic Chemicals (Rochester, NY, U.S.A.). Ammonium hydroxide was obtained from McArthur Chemical Co. (Montreal, QUE, Canada). L-lysine was from Fluka Chemika-Biochemika (New York, NY, U.S.A.). Kodak GBX developer and replenisher, and fixer were obtained from Picker International Canada (Brampton, ON, Canada). Hydroxyethylurea was purchased from TCI America (Portland, OR, U.S.A.). Thin-layer chromatography (TLC) plates (glassbacked Silica Gel 60, 0.25 mm) were purchased from BDH Chemicals (Toronto, ON, Canada). Kodak X-Omat AR5 film was obtained from Picker International Canada (Brampton, ON, Canada). METHODS

Preparation of erythrocyte ghosts Unsealed erythrocyte ghosts were prepared by the method of hypotonic hemolysis (8) in the presence of 0.1 mM phenylmethylsulfonyl fluoride (PMSF) and stored in 10 mM sodium phosphate, pH 7.4. Long 334

storage periods (>24 h) included 0.2 mg/mL sodium azide as antibacterial.

Phospholipid and protein quantitation Phospholipid was quantitated as phosphorus (9) using ascorbic acid as the reducing agent (10). Potassium monophosphate was used as the standard and the samples read at 826 nm. Protein determinations were conducted using a modified Lowry procedure (11). Human serum albumin (HSA) was used as the standard and samples read at 660 nm.

Liquid scintillation counting All liquid scintillation counting was performed on a LS 6500 Liquid Scintillation System (Beckman Instruments Inc., Fullerton, CA, U.S.A.). Unless otherwise indicated, polyethylene scintillation vials and 10 mL of scintillation cocktail were used.

IN VITROSTUDIES 14C-carbamylation of erythrocyte ghosts for lipid studies Erythrocyte ghosts (1.5 mg/mL membrane protein) were incubated in 10 mM sodium phosphate, 25 mM sodium cyanate, and 0.5 mM potassium [14C] cyanate, pH 7.4, for 15 h at 37 °C. The sample was subsequently centrifuged (12,000 x g for 30 min at 5 °C) and washed to remove unbound cyanate. The carbamylated membranes were brought up to approximately 50% hematocrit with 10 mM sodium phosphate, pH 7.4, and generally contained 1-2 mg/ mL membrane protein.

Extraction of erythrocyte membrane lipids Native and/or 14C-carbamylated ghost cells were centrifuged (12,000 x g for 20 min at 10 °C, Sorvall Instruments RC2-B and SS-34 fLxed angle rotor) and the supernatant discarded. The lipid components of the erythrocyte membrane pellet (1-2 mL) were extracted into chloroform by the method of Folch et al. (12) and the solvent evaporated (Brinkmann Rotavapor RE 120, Brinkmann Instruments Inc., Westbury, NY) to approximately 0.5 mL and stored in small (1.5 mL) glass vials at -20 °C.

Quantitation of [14C] cyanate binding to membrane protein and lipid fractions Aliquots (100 ~L) of washed 14C-carbamylated ghosts (1-2 mg/mL) were removed and incorporation of cyanate (DPM) into the total cell, protein [resolved by SDS-PAGE and excised from the gel as previously described (5)], and lipid (by solvent extraction, as above) fractions were measured by liquid scintillation counting (Beckman LS 6500) in 10 mL Sigma-FluorTM, with a counting efficiency of 96%. CLINICALBIOCHEMISTRY,VOLUME29, AUGUST1996

CARBAMYLATION OF ERYTHROCYTE MEMBRANE AMINOPHOSPHOLIPIDS

Cyanate incorporation (nmol) into each fraction was calculated as follows: DPM/fraction (52 Ci mole-1) (22.2 x 1012 DPM Ci-1)

x (cyanate/[14C]cyanate)

where (cyanate x [14C] cyanate) is the ratio of unlabeled to labelled cyanate in the reaction mixture. The final concentration of [14C] cyanate was calculated as [(raM stock [14(7] cyanate) (volume of stock [14C] cyanate)/total assay volume]. Stock [14C] cyanate was prepared by the addition of 1.89 mL water to 1 mCi of crystalline potassium [14C] cyanate and the DPM of an aliquot determined and converted to mM stock [14C] cyanate according to [(DPM of aliquot)/(52 Ci mole-1) (2.22 x 1012 DPM Ci-1)]. Greater than 85% of the total cell counts were recovered in the combined lipid and protein fractions.

Preparation and 14C-carbamylation of phosphatidylethanolamine / phosphatidylcholine liposomes Phosphatidylethanolamine (PE)/phosphatidylcholine (PC) liposomes were prepared by adding 5 mg of egg yolk PC (50 ~L of a 100 mg/mL stock solution of egg yolk PC in chloroform) and 4 mg of egg yolk PE (400 ~L of a 10 mg/mL stock solution of PE in chloroform) to 2 mL chloroform. The phospholipid mixture was rotary-evaporated to dryness and 2 mL of buffer A (5 mM sodium phosphate, 100 mM sodium chloride, pH 8.0) added. The suspension was sonicated (Branson 1200, Branson Cleaning Equipment Company, Shelton, CT, U.S.A.) for 10 min at 10 °C and 1 mL of buffer B (5 mM sodium phosphate, 100 mM sodium cyanate, pH 8.0) and 200 ~L of stock [14C] cyanate i:a water (prepared as above) were added. The liposomes were incubated at 30 °C for 16 h and, then, exhaustively dialyzed against buffer A to remove unbound cyanate. The washed liposomes were, then, solvent extracted into chloroform as above.

Thin-layer chromatography, autoradiography, and imaging densitometry Thin-layer chromatography (TLC) of solvent extracts of native and [14C] cyanate-treated erythrocyte membrane lipid, and egg yolk PE/PC liposomes was performed. Briefly, TLC plates (20 × 20 cm) were spotted with sample (typically 5-10 ~L) using a 5-~L Hamilton syringe and developed in a solvent system (appropriate to the lipid being investigated) to within 0.5 cm from t:he top of the plate. The TLC plate was air-dried and stained for either phospholipid or total lipid visualization. Phospholipids were visualized by spraying the plates with a phospholipid specific stain, according to the procedure of Vaskovsky and Kostevsky (13). Total lipids were visualized by dipping the plate in a manganese chloCLINICAL BIOCHEMISTRY, VOLUME 29, AUGUST 1996

ride-sulfuric acid derivatizing reagent followed by heat activation at 110 °C for 40 min (14). Stained plates were scanned by imaging densitometry and, in some cases, the phospholipids were quantitated by use of the volume analysis program (Molecular Analyst, Bio-Rad Laboratories). Chemical quantitation of the individual phospholipids and cyanate incorporation were conducted by phosphorus analysis (9) and scintillation counting, respectively, of spots scraped from the TLC plate. Autoradiography was performed on the stained plates by exposure to Kodak X-Omat AR5 film at -80 °C for 3 days.

Time-course of erythrocyte phosphatidylethanolamine carbamylation Fresh (<12 h) whole blood from nonuremic individuals was obtained in tubes containing EDTA as anticoagulant. Ghost cells (2 mg/mL membrane protein) were prepared and incubated with 25 mM sodium cyanate in 10 mM phosphate buffer, pH 7.4, at 37 °C. One volume of reaction mixture was removed at various time intervals (1-10 h) and added to an equivalent volume of 2 M lysine in water to terminate any further reaction between cyanate and erythrocyte membrane lipid. The samples were then stored at 5 °C until all time points were collected. All samples were then centrifuged (12,000 × g for 30 min at 5 °C, Sorvall Instruments RC2-B and SS-34 fixed-angle rotor) and washed to remove unbound cyanate. All samples were made up to an equivalent concentration (1.5 mg/mL membrane protein) and the lipids extracted, applied to TLC plates, and resolved in chloroform/methanol/ammonia (65/35/5, v/v/v). IN VIVO STUDIES

Preparation of erythrocyte membrane lipid / Triton X-I O0 micelles and treatment with phospholipase D Erythrocyte ghost membrane pellets (1-2 mL containing 2-3 mg protein) in 10 mM sodium phosphate, pH 7.4, were added to 25-30 mL of chloroform/methanol (2:1, v/v) and the lipids extracted, as above. The chloroform solvent was removed by rotary evaporation and the lipids left to vacuum dry for 15 min. Subsequently, 500 ~L of Buffer C (50 mM Tris-HC1, 10 mM CaC12, 1% Triton X-100, pH 8.0) were added and the suspension sonicated for 10 min at 10 °C (15). The resulting phospholipid/Triton X-100 micelle suspension was incubated with 80 ~L of freshly prepared phospholipase D (PLD, 1 mg/mL) in water for 2 h at 37 °C.

Preparation and treatment of egg yolk phospholipid / Triton X-IO0 micelles, cholesterol / Triton X- I O0 micelles, and phospholipid headgroup alcohols with phospholipase D All phospholipid and cholesterol/Triton X-100 micelles were prepared by rotary evaporation of com335

TREPANIER AND THIBERT

mercially purified egg yolk phospholipid (in chloroform suspension) to dryness followed by sonication (10 min) in 1.0 mL of buffer C. Phosphatidylcholine (PC)/Triton X-100 micelles were prepared using 17 ~L of a 100 mg/mL PC stock in chloroform to give a final concentration of 2.3 ~mol PC/1 mL buffer C. Phosphatidylethanolamine (PE)/Triton X-100 micelles were prepared using 137 ~L of a 10 mg/mL PE stock in chloroform to give a final concentration of 1.8 ~mol PE/1 mL buffer C. Phosphatidylserine (PS)/Triton X-100 micelles were prepared using 68 ~L of a 10 mg/mL PS stock in chloroform to give a final concentration of 0.91 ~mol PS/1 mL buffer C. Sphingomyelin (SM)/Triton X-100 micelles were prepared using 27 ~L of a 50 mg/mL SM stock in chloroform to give a final concentration of 1.8 ~mol SM/1 mL buffer C. Cholesterol/Triton X- 100 micelles were prepared using 132 ~L of a 10 mg/mL cholesterol stock in chloroform to give a final concentration of 2.6 ~mol cholesterol/1 mL buffer C. Triton X-100 micelles containing more than one phospholipid were prepared in an identical manner using the same concentrations of each phospholipid indicated above. Ethanolamine (1.82 ~mol/500 ~L buffer C) was prepared by adding 110 ~L of ethanolamine to 390 ~L of buffer C. Choline (4.1 ~mol/500 ~L buffer C) was prepared by adding 100 ~L of a 5.72 mg/mL choline stock in water to 400 ~L of buffer C. Serine (0.91 ~tmol/500 ~L buffer C) was prepared by adding 100 ~L of a 0.957 mg/mL serine stock in water to 400 ~L of buffer C. All of the above solutions (500 ~L) were incubated in the absence (80 ~L of buffer C) and presence of phospholipase D (80 ~L of a 220 U/mL stock in water) for 2 h at 37 °C.

Measurement of phospholipase D-catalyzed phospholipid hydrolysis by thin-layer chromatography and imaging densitometry Erythrocyte membrane lipid/Triton X-100 micelles and egg yolk phospholipid/Triton X-100 micelles, prepared as described above, were incubated in the absence (80 ~L of buffer C) and presence of phospholipase D (80 ~L of a 220 U/mL stock in water) at 37 °C. For the hydrolysis of native and PLDtreated egg yolk phospholipidfrriton X-100 micelles, samples were incubated for 2 h and subsequently added directly to 12 mL of chloroform/methanol (2:1, v/v). The chloroform layer was rotary evaporated to less than 500 btL and 10- to 20-~L aliquots were applied to TLC plates. For the time-course of erythrocyte membrane lipid/Triton X-100 hydrolysis by PLD, 100-~L aliquots were removed at various time intervals and added directly to chloroform/methanol (2:1, v/v) to stop the reaction. Samples were treated as above and 10- to 20-~L aliquots were applied to TLC plates and the phospholipids resolved using a solvent system of chloroform/methanol/acetic acid/ water; 90/40/12/2 (9) and stained for total lipid. The dried TLC plates were scanned by imaging densi336

tometry and analyzed for relative phospholipid content by use of the volume analysis program.

Reaction of native and phospholipase D-treated lipid~Triton X- I O0 micelles with diacetyl monoxime The method of Wybenga et al. (16) for determination of urea and substituted ureas (carbamyl derivatives) was modified and employed for measurement of carbamylated erythrocyte membrane lipid. Briefly, the native and PLD treated phospholipid/ Triton X-100 micelle suspensions (580 ~L) were added directly to 2.5 mL of urea-nitrogen reagent (0.83 M sulfuric acid, 1.13 M orthophosphoric acid, 0.55 mM thiosemicarbazide, and 2.6 mM cadmium sulfate) and 500 ~L of 3% diacetyl monoxime in water. A solution of PLD (80 ~L of a 1 mg/mL solution in water) in buffer C was used as a control. All samples were incubated for 30 min at 100 °C in a dry heater block and cooled for 5 min in a beaker of cold water. To clear the sample turbidity, sodium dodecyl sulfate (SDS) was added (100 ~L of a 20% SDS solution in water), prior to recording the absorption spectrum (400-650 nm). Studies with homocitrulline demonstrated that the addition of SDS does not alter the final absorbance (data not shown).

Reaction of urea, carbamyl ethanolamine, and carbamyl serine with diacetyl Aliquots (100 ~L) of 400 ~M stock solutions of urea (24.02 ~g/mL), carbamyl ethanolamine (hydroxyethylurea; 41.G4 ~g/mL), and carbamyl serine (8.16 ~g/mL) in water were added to 400 ~L of buffer (50 mM Tris-HC1, 10 mM CaC12, 1% Triton X-100, pH 8.0). All samples were reacted with diacetyl, as above, except that SDS was not added.

Quantitation of carbamylated phosphatidylethanolamine in erythrocytes of normal individuals Fresh (<10 h) pooled whole blood from nonuremic individuals (urea = 4.46 _+0.51 mM) was obtained in tubes containing EDTA as anticoagulant. Ghost cells were prepared; the last two washes were in buffer containing 50 mM Tris-HC1, 1% Triton X-100, pH 8.0, and 1.25-mL aliquots (containing 2.75 mg/ mL membrane protein) were added to 25 mL of chloroform/methanol (2:1, v/v). Erythrocyte lipid/Triton X-100 micelles (in 500 ~L buffer C) were prepared from the lipid extracts and aliquots were analyzed for phospholipid and protein content. The erythrocyte lipid/Triton X-100 micelles were, then, treated in the absence and presence of PLD, as above, and subsequently reacted with diacetyl. The carbamyl ethanolamine (hydroxyethylurea) released was quantitated from a standard curve (0-40 nmol) prepared using hydroxyethylurea. CLINICAL BIOCHEMISTRY, VOLUME 29, AUGUST 1996

CARBAMYLATION OF ERYTHROCYTE MEMBRANE AMINOPHOSPHOLIPIDS

5.0 i u,.

.5 4.0.

3.5-

~ 3.0-

~ 2.5-

"1-Membrane Lipid FrmactbMe i°nra(e6nO%pr )otein

//_,,o%, /

~-

~ "1.5~ 1.0~ 0.5U

0.0

o

,t

1'o

1'2

1'4

1'6

1'8

Time (h)

Figure 1 - - 14C-carbamylated erythrocyte ghosts (1.5 mg/ mL membrane protein) were prepared. Aliquots of 100 tLL were removed and separated into protein and lipid fractions by SDS-PAGE and ,~olvent extraction, respectively. The amount of cyanate (nmol) incorporated into each fraction was measured by liquid scintillation counting. The lipid and protein fractions of the membrane incorporated 60% and 40%, respectively, of the total counts (DPM) recovered. The values represent the mean of duplicate determinations.

Results Erythrocyte ghosts "were carbamylated in vitro with [14C] cyanate for ].5 h and the membrane components separated into total protein and total lipid fractions. Greater t h a n 85% of the total ghost cell membrane [14C] cyanate counts were recovered. Figure 1 demonstrates exl~ensive incorporation of cyanate into both the protein and lipid fractions. In fact, a greater degree of the total membrane cyanate incorporated is associated with the lipid components (60%) of the membrane relative to the protein components (40%). To establish covalent binding of cyanate to the phospholipids, thin-layer chromatography (TLC) of the extracted membrane lipid fraction was performed in a solvent system of chloroform/ methanol/acetic acid/water (60/50/1/4, v/v/v/v) followed by autoradiography of the plate stained for phospholipid (Figure 2). This solvent system has been used to completely resolve all the major classes ofphospholipids (9). Good TLC separation of the major erythrocyte phospholipids is clearly illustrated and an identical phospholipid staining pattern is observed for both natiw~ and [14C] cyanate-treated CLINICAL BIOCHEMISTRY, VOLUME 29, AUGUST 1996

ghost cells. The autoradiograph of the TLC plate (Figure 2), however, clearly illustrates the presence of two distinct areas of radioactivity co-migrating with phosphatidylethanolamine (PE) and phosphatidiylserine (PS), respectively. All of the phospholipids separated by TLC were scraped from the plate and analyzed for both [14C] cyanate incorporation and phospholipid content (Table 1). The results confirm the covalent incorporation of cyanate into both PE and PS (15.76 _+ 0.09 and 13.34 +_ 0.81 mole% cyanate/phospholipid, respectively). Under the same conditions, no incorporation of [14C] cyanate into cholesterol (Chol), phosphatidylcholine (PC), or sphingomyelin (SM) was detectable (Table 1 and Figure 2). Because the degree of carbamylation of PE and PS following a 15-h incubation with cyanate are essentially equivalent (Table 1), it would appear that both phospholipids are equally reactive towards cyanate. Thin-layer chromatography of the in vitro carbamylated membrane aminophospholipids in chloroform/methanol/acetic acid/water (60/50/1/4, v/v/v/v) does not, however, separate carbamylated PE (carbPE) or carbamylated PS (carb-PS) from PE or PS (Figure 2). It was considered unlikely that one solvent system would accomplish separation of both species on one TLC plate, and so research efforts were focused upon separation of carb-PE from PE. To accomplish this task without the potential complexity of interference from other erythrocyte phospholipids, egg yolk PE/PC liposomes were constructed from commercially purified PE and PC. PE/ PC liposomes were used because liposomes of PE alone are known to adopt a hexagonal (Hn) phase (i.e., inverted micelle, in which the PE amino headgroups are not exposed to the solvent) (17). The liposomes were carbamylated with [14C] cyanate and a number of solvent systems investigated. Complete resolution of egg yolk carb-PE from PE was accomplished using a chloroform/methanol/ammonia (65/ 35/5, v/v/v) solvent system (Figure 3). Under these conditions, the Rfvalue of PE shifts from 0.61 _+0.01 (Table 2, row 1) to 0.46 _+0.02 (Table 2, row 2), and carb-PE remains at 0.61 _+0.01 (Table 2, row 3). This may be partially due to the different acid/base characteristics of the solvents. Because amines and amides have pKas in the region of 9 to 10, and -1 to 0, respectively (18), a change of solvent system from chloroform/methanol/acetic acid/water (60/50/1/4, v/v/v/v) (pH, 3.2) to chloroform/methanol/ammonia (65/35/5, v/v/v) is associated with a PE amine charge change from (+1) to (0), which potentially contributes to the observed shift in Rf value (Figure 3). In contrast, the newly formed headgroup amide of carbPE is uncharged in both solvent systems and its Rf value is unchanged (Figure 3). The chloroform/methanol/ammonia (65/35/5, v/v/ v) solvent system was next applied to the separation of 1 4 C-carbamylated erythrocyte membrane lipids. Comparison of native with 1 4 C-carbamylated lipids (Figure 4) shows the appearance of a new band that 337

TREPANIER AND THIBERT

Thin- Layer Chromatograph

Autoradiograph

:I

Figure 2 - - Thin-layer chromatography (TLC) (left) and autoradiography (right) of native and 14C-carbamylated membrane lipids resolved in chloroform/methanol/acetic acid/water (60/50/1/4). Under these conditions, cholesterol migrates just slightly behind the solvent front (9) and does not stain. Chol, cholesterol; PE, phosphatidylethanolamine; PS, phosphatidylserine; PC, phosphatidylcholine; SM, sphingomyelin. PE, PS, and PC standards (10 ~g of each) were applied. Phospholipids were visualized on the TLC plate using the procedure of Vaskovsky and Kostevsky (13). CNO, unlabelled cyanate; laCNO, labelled cyanate. The autoradiogram is for lanes reacted with [14C] cyanate. m i g r a t e s a h e a d of PE. The new b a n d can be identified as carb-PE (Table 2, row 5) from t h e following: 1. It i n c o r p o r a t e s identical [14C] c y a n a t e counts Table 2, row 5) as e r y t h r o c y t e carb-PE u n r e solved from P E (Table 2, row 4); 2. It h a s t h e s a m e Rf v a l u e as egg yolk c a r b - P E (Table 2, row 3), resolved from PE, in the s a m e solvent system; a n d 3. T h e combined nmol of s e p a r a t e d carb-PE and

P E (Table 2, rows 5 a n d 6, respectively) are e q u i v a l e n t to u n r e s o l v e d (PE + carb-PE) (Table 2, row 4). T h e a u t o r a d i o g r a p h of F i g u r e 4 also clearly shows t h a t t h e c h l o r o f o r m / m e t h a n o l / a m m o n i a (65/35/5, v/v/v) s o l v e n t s y s t e m does n o t s e p a r a t e c a r b a m y l a t e d - P S f r o m PS. No f u r t h e r a t t e m p t s w e r e m a d e to s e p a r a t e t h e s e species. The s e p a r a t i o n of carb-PE from P E by TLC en-

TABLE 1 Quantitation of Phospholipid and Cyanate Binding to 14C-Carbamylated Ghost Membrane* Lipid Extracts Resolved by TLCt Phospholipid$ (nmol) SM PC PS PE Chol

72.79 88.15 30.21 63.3 NA

_+1.7 +_3.2 ± 2.1 _+0.15

Phospholipid (% of Total)

Phospholipid (% of Total)§

Cyanate II (nmol)

Cyanate/Phospholipid (mol %)

28.61 ± 0.59 34.65 ± 1.36 11.88 ± 0.78 24.88 ± 0.01 NA

25-30 25-30 10-15 25-30 NA

0 0 4.02 ± 0.11 9.97 ± 0.06 ND

0 0 13.34 ± 0.81 15.76 ± 0.09 0

SM, sphingomyelin; PC, phosphatidylcholine; PS, phosphatidylserine; PE, phosphatidylethanolamine; Chol, cholesterol; NA, not applicable; ND, not determined. *Erythrocyte ghosts (1.5 mg/mL membrane protein) incubated with 25 mM sodium cyanate and 0.5 mM potassium [14C] cyanate for 15 h at 37 °C, pH 7.4. tThe solvent system was chloroform/ methanol/acetic acid/water (60/50/1/4, v/v/v/v). SQuantitated as phosphorus (9) subsequent to scraping of band from TLC plate. §Actual mol percentages (27). IfQuantitated by scintillation counting of TLC scraping. 338

CLINICAL BIOCHEMISTRY, VOLUME 29, AUGUST 1996

CARBAMYLATION OF ERYTHROCYTE MEMBRANE AMINOPHOSPHOLIPIDS

Chllmethlaalwater

(60/50/1/4)

system of chloroform/methanol/ammonia (65/35/5, v/v/v) would allow the detection of in vivo carb-PE. This is because any band in the absence of cyanate incubation with an Rf value equivalent to carb-PE might well be intrinsically carbamylated PE. However, due to the diffuse nature of lipid content in this region (Figure 5A), phosphate analysis alone would not necessarily be indicative of the presence of in vivo carb-PE. It was also clear, from previous work conducted in our laboratory, that the direct incubation of ghost cell membranes with diacetyl does not result in the formation of a detectable chromogen (5). Consequently, an alternate route to investigate whether or not carbamylated aminophospholipids can be detected in the native erythrocyte membrane was explored. The general concept (Figure 6) was to treat erythrocyte membrane lipid micelles with phospholipase D (PLD) (EC 3.1.4.4) from Streptomyces chromofuscus and to react any released carbamylated aminophospholipid headgroups with diacetyl. PLD catalyzes cleavage of the phosphodiester bond of a phospholipid (or lysophospholipid) molecule to produce phosphatidic acid (PA) and an alcohol, such as choline from PC and SM, ethanolamine from PE, serine from PS, and inositol from phosphatidylinositol (PI) (19). The relative activity of PLD on PC, SM, PE, and PI/Triton X-100 micelles has been measured to be 100%, 25%, 31%, and 15%, respectively (20). The enzyme has a pH optimum range of 7.0-8.5 and is activated by Ca 2÷ and detergents such as Triton X-100 (20). If carb-PE and carb-PS are present in the native membrane, their hydrolysis would be expected to liberate carbamyl serine and carbamyl ethanolamine (hydroxyethylurea). Both carbamyl serine and carbamyl ethanolamine would, then, be available for reaction with diacetyl. The absorbances, relative to urea, for equivalent concentrations of ethanolamine and carbamyl serine reacted with diacetyl are shown in Figure 7. Although carbamyl ethanolamine and urea have essentially equivalent molar absorption coefficients

Chllmethlammonia (65/35/5)

Figure 3 - - Thin-layer Chromatographs and autoradiographs of 14C-carbamylated egg yolk phosphatidylcholine (PC)/phosphatidylethanolamine (PE) liposomes resolved in different solvent systems. In both cases, the developed autoradiogram is overlayed onto the stained TLC plate and scanned by imaging densitometry. The dark spots represent areas of radioactivity. Phospholipids were visualized using the procedure of Vaskovsky and Kostevsky (13). Chl, chloroform; meth, methanol; aa, acetic acid. ables one to follow the carbamylation of erythrocyte m e m b r a n e PE (and potentially any s u b c e l l u l a r membrane source of PE) in vitro without the use of [14C] cyanate and autoradiography. Figure 5A clearly demonstrates the increasing carbamylation of erythrocyte PE in vitro as a function of incubation time with cyanate. Imaging densitometry of the phospholipid separated by TLC shows that, by 48 h, greater that 40% of the PE is carbamylated (Figure 5B). The in vitro results indicate that, in theory, TLC of native erythrocyte membrane lipid in a solvent

TABLE 2

Phospholipid Quantitation, DPM, and Rf Values of 14Carbamylated and Native Phosphatidylethanolamine Resolved by Thin-Layer Chromatography in Different Solvent Systems Phospholipid band

Rf Value

[14C] Counts$ (DPM)

Cyanate (nmol)

Phospholipid§ Cyanate/phospholipid (nmol) (mol/mol)

ND ND ND

NA 0 NA

ND ND ND

NA NA NA

3 1 9 8 3 _+ 191 31504 ± 242

NA 12.28 ± 0.14

58.9 ± 0.93 16.21 ± 2.1

NA 0.77 + 0.12

0

0

43.41 _+1.6

NA

1) E g g Y o l k ( P E + C a r b - P E ) *

0.61±0.01 (Figure 3) 0.46±0.02 2) Egg Yolk (PE)t (Figure 3) 0.61±0.01 3) Egg Yolk (Carb-PE)t (Figure 3) 4) Erythrocyte (PE + Carb-PE)* 0.61 _+ 0.01 (Figure 2) 5) Erythrocyte (Carb-PE)t (Figure 4) 0.62 ± 0.01 0.45 ± 0.01 6) Erythrocyte (PE)t (Fi~.~re 4)

PE-, phosphatidylethanolamine; Carb-PE, carbamylated phosphatidylethanolamine; NA, not applicable; ND, not determined. *Resolved in Chloroform/Methanol/Acetic Acid/Water (60/50/1/4, v/v/v/v). ~Resolved in Chloroform/Methanol/ Ammonia (65/35/5, v/v/v). SBands were scraped from TLC plate and scintillation counted in 10 mL Sigma-FluoF M. §Quantitated as phospho:cus (9). All values represent the mean of duplicate determinations. CLINICAL BIOCHEMISTRY, VOLUME 29, AUGUST 1996

339

TREPANIER AND THIBERT Autoradiograph Memb.... Lipids I J SZd [ Na.ve In,+r.,,] Stdl

Thin-Layer Chromatograph

I

Chol

Carb-PE

Carb-PE

! PE

PE PC

PS

PC

l

i

''!

PS

Figure 4 - - Thin-layer chromatographs (left) and autoradiogaphs (right) of native and 14C-carbamylated erythrocyte membrane lipids resolved in chloroform/methanol/ ammonia (65/35/5). The phospholipids were visualized using the procedure of Vaskovsky and Kostevsky (13). PE, phosphatidylethanolamine; PC, phosphatidylcholine; PS, phosphatidylserine; Carb-PE, carbamylated phosphatidylethanolamine; Std, standard; O14CN, [14C] cyanate. upon reaction with diacetyl, carbamyl serine does not produce any detectable chromogen with diacetyl. This indicates that either there is no reaction of diacetyl with carbamyl serine or that the reaction product is either nonchromogenic or unstable under the conditions of the diacetyl monoxime assay. The significance of this observation to the present work is that the liberation of carbamyl serine from erythrocyte membrane lipid micelles treated with PLD would not be expected to produce a measureable chromogen upon reaction with diacetyl. Any chromogen, therefore, generated in the presence of PLD would be suggestive specifically of carbamyl ethanolamine. Figure 8 demonstrates t h a t PLD efficiently catalyzes the hydrolysis of purified egg yolk PC/Triton X-100, PE/Triton X-100, PS/Triton X-100, PC/PS/Triton X-100, and SM-Triton X-100 micelles. The solvent system utilized (chloroform/methanol/ acetic acid/water, 90/40/12/2, v/v/v/v) completely resolves PA from the native phospholipids (9). Interestingly, ceramide, resulting from the hydrolysis of SM, has essentially the same RWvalue as native PE. The time-course treatment of erythrocyte memb r a n e lipid/Triton X-100 micelles with PLD, as monitored by TLC, is shown in Figure 9. By visual inspection, it is evident that PLD catalyzes the hydrolysis of erythrocyte membrane PC, PS, SM, and PE into PA. The degree of hydrolysis of PE would appear to be less than the other phospholipids (lane 4); however, the lipid present in this region is most likely due to ceramide (Figure 8). Imaging densitometry of the TLC plate in Figure 9 allows an estimate of the degree of hydrolysis of the various erythrocyte phospholipids. The results (Figure 10) show that the 340

B

4O 3O. '~ O

20.

'~ u

10-

o

1'o

2'o

"

3'o

"

,'o

s'o

Time (h)

Figure 5 -- A: Time course of carbamylation of erythrocyte phosphatidylethanolamine (PE) monitored by thin-layer chromatography (TLC). Membrane lipids were resolved in chloroform/methanol/ammonia (65/ 35/5) and visualized by dipping the TLC plate in a manganese chloride-sulfuric acid derivatizing reagent (14). B: Percent PE carbamylation as a function of incubation time with cyanate, where percent PE carbamylation was calculated by imaging densitometry of the scanned TLC plate. hydrolysis is complete within a 30-min incubation with PLD. Consistent with this observation, the levels of PA rise rapidly and are essentially unchanged after 30 min. Control values of cholesterol, as expected, r e m a i n u n c h a n g e d over the time period studied. The less than complete hydrolysis of the phospholipids (Figure 10) does not result from PLD instability because subsequent additions of fresh PLD do not induce further hydrolysis (data not shown). The results demonstrate the nonspecific nature of phospholipid hydrolysis catalyzed by PLD, and suggest that carbamylated aminophospholipids would, if present, be expected to undergo a similar hydrolysis. The reaction of native and PLD-hydrolyzed erythrocyte lipid/Triton X-100 micelles with diacetyl was next investigated (Figure 11). Erythrocyte lipidTIYiton X-100 micelles untreated with PLD are shown (Figure 11) to generate a detectable absorbance at 530 nm upon reaction with diacetyl. The absorbance does not result from residual carbamylated memCLINICAL BIOCHEMISTRY,VOLUME 29, AUGUST 1996

CARBAMYLATION OF ERYTHROCYTE MEMBRANE AMINOPHOSPHOLIPIDS

NODETECTABLE CHROMOGEN

Diacetyl

GH(~TCELL TrX

SOLVENT EXTRACTION

PA

P~-"W~ 1 SONICATE (TRITON X-100) ERYTHROCYTE LIPID MICELt.E

(PE, PS, PC, SM, CHOL)

~1~

PE

Cer

PhospholipaseD Ca2.

PA

Cholesterol + HO-CHz-CHrN-(CH=)= (Choline)

Ceramide

HO-CHrCHrNH2 HO-CHrCH-NH=

(Ethanolamine) (Serine)

I

C0ff

D~acetyl

I~ CHROMOGEN ~

PS

SM

HO-CH=-C H-NH-CO-NH=(Carb-Serine) COO" HO-CH=-CH=-NH-CO-NI'I= (CarbJEthanolamlne)

Figure 6 -- Reaction scheme for the measurement of carbamylated erythrocyte aminophospholipids. PE, phosphatidylethanolamine; PS, pbosphatidylserine; PC, phosphatidylcholine; SM, sphingomyelin; Chol, cholesterol; PA, phosphatidic acid; Carb, carbamylated. brahe protein because no detectable protein was found upon protein analysis of the lipid micelles (data not shown). Treatment of erythrocyte lipid/ Triton X-100 micelles with PLD results in an approximately 4-fold increase in absorbance (Figure 11). The increased absorbance does not arise from the reaction of diacetyl with PLD because a buffer blank containing PLD does not produce any absorbance at 530 nm above; that of buffer alone (Figure 11). Clearly, therefore, the treatment of erythrocyte lipid/Triton X-100 micelles with PLD results in the production of a nonprotein compound(s) that is(are) reactive with diacetyl. The incubation of PLD with the erythrocyte membrane lipid/Triton X-100 micelles clearly produces a number of reaction products (Figure 6) not present in the native micelle, which may potentially react

Urea CarbamylLysine (Homocitrulline) CarbamylEthanolamine (Hydroxyethylurea)

II CarbamylSerine oo

0'2

0:4

o:6

0:8

PC

1'.o

£2

1:4

Relative Absorbance (530 nm) 1 40 nmol F i g u r e 7 - - R e l a t i v e a b s o r b a n c e s o f u r e a , c a r b a m y ] ]ysine, c a r b a m y ] e t h a n o l a m i n e , a n d c a r b a m y ] s e r i n e w i t h

diacetyl. CLINICAL BIOCHEMISTRY, VOLUME 29, AUGUST 1996

Figure 8 -- Thin-layer chromatography of native and phospholipase D-treated egg yolk phospholipid/Triton X-100 micelles. PC, phosphatidylcholine; PE, phosphatidylethanolamine; PS, phosphatidylserine; SM, sphingomyelin; TrX, Triton X-100; Cer, ceramide; PA, phosphatidic acid. with diacetyl. These include: PA; ceramide; ethanolamine; choline; serine; and, finally, if PE is intrinsically carbamylated, carbamyl ethanolamine. Carbamyl serine is not considered as a potential reactant because it is nonchromogenic in the presence of diacetyl (Figure 7). A study was, consequently, u n d e r t a k e n to ascertain, by process of elimination, whether or not the chromogen produced upon reaction of erythrocyte membrane lipid/Triton X-100 micelles with PLD could be the reaction product of diacetyl with carbamyl ethanolamine. The results are recorded in Figure 12. No chromogenic development with diacetyl is observed for native or PLD-hydrolyzed PC, PS, or SM/Triton X-100 micelles. Likewise, ethanolamine, choline, and serine in the absence or presence of PLD are nonreactive with diacetyl. Interestingly, however, native PE/ Triton X-100 micelles generate a detectable chromogen (km~ = 530 nm) when reacted with diacetyl, and PLD treatment of the PE/Triton X-100 micelles results in an approximately 4-fold increase in the absorbance. This is analogous to the chromogenic development of erythrocyte lipid in the absence and presence of PLD (Figure 11). The reactivity of diacetyl specifically with purified egg yolk PE is also demonstrated by the observation that PE/PC/PS/SM/ cholesterol/Triton X-100 micelles (simulating the composition of erythrocyte membrane lipid/Triton X-100 micelles) treated with PLD are chromogenic; however, PC/PS/SM/cholesterol/Triton X-100 micelles (i.e., without PE) are not chromogenic (Figure 12). Because the hydrolysis of PE produces only PA and ethanolamine (Figure 6), neither of which are reactive with diacetyl (Figure 12), this suggests that 341

TREPANIER AND THIBERT 0.20

Micelle (+ PLD) 0.15

Chol

PA

O O P 0.10 o {D J~ <

~E

icelle (- PLD)

3er

0.05.

~S

0.00

,

400

,

,

450

500

550

Wavelength

~.

.

.

.

.

.

.

.

.

.

.

Figure 9 -- Time-course for the hydrolysis of erythrocyte lipid/Triton X-100 micelles by phospholipase D (PLD). Chol, cholesterol; PA, phosphatidic acid; PE, phosphatidylethanolamine; Cer, ceramide; PS, phosphatidylserine; PC, phosphatidylcholine; LysoPL, lysophospholipid; SM, sphingomyelin.

I E g g yolk SM (- / + PLB) • Egg yolk Chol (- / + PLD) • Ethenolamine (- / + PLD) I Choline ( - / + PLD) I S e r i n e ( - I + PLD) I E g g yolk PE ( - PLD)

•o 8o

'~. •~ . ~_ 60-

i

purified egg yolk PE is intrinsically carbamylated and that treatment with PLD releases carbamyl ethanolamine that subsequently reacts with diacetyl. The detectable chromogenic development in the absence of PLD treatment is likely due to reaction of diacetyl with either native carbamylated PE or the nonenzymatically hydrolyzed headgroup of carba-

• Buffer • PLD • Egg yolk PC (- / + PLD) I Egg yolk PS (- / + PLD)

12°1 I:;:;x'l

t

]Egg yolk PE ( + PLD)

I

20_/

.-

. 0.0

. 012

.

. 014

.

. 016

018

110

/ Egg yolk PE/PC/PB/SM/Ch~ (+ PLD)

______.

o.oo . . . .

1:2

114 '

116

Time (h)

Figure 10 - - Imaging densitometry analysis for the hydrolysis of erythrocyte membrane phospholipids by phospholipase D. PC, phosphatidylcholine; PS, phosphatidylserine; Chol, cholesterol; PA, phosphatidic acid. 342

Egg yolk PF.IPC/PS/SM/Chol ( - PLD)

I Egg yolk PC/PS/SM/Chol ( - / + PLD)

0

650

Figure 11 -- Absorption spectra for the reaction of diacetyl with native and phospholipase D (PLD)-treated erythrocyte lipid/Triton X-100 micelles. Buffer (50 mM Tris-HC1, 10 mM CaC12, 1% Triton X-100, pH 8.0).

PC

...... ~

600

(rim)

o.b5 . . . .

o.%o . . . .

o.~s . . . .

o.~,o . . . .

o.~5

kbsorbance (530 nm)

Figure 12 -- Reaction of phospholipid headgroups and phospholipid micelles, in the absence and presence of phospholipase D (PLD), with diacetyl. PC, phosphatidylcholine; PS, phosphatidylserine; SM, sphingomyelin; Chol, cholesterol; PE, phosphatidylethanolamine; Buffer (50 mM Tris-HC1, 10 mM CaC12, 1% Triton X-100, pH 8.0). CLINICALBIOCHEMISTRY,VOLUME29, AUGUST 1996

CAPBAMYLATIONOF ERYTHROCYTEMEMBRANE AMINOPHOSPHOLIPIDS

mylated PE (carbamyl ethanolamine) resulting from the high acidity and temperature of the assay. The presence of carb-PE in commercially purified egg yolk PE is a possibility. Commercial preparation of PE from egg yolk involves separation of PE from other egg yolk phospholipids by TLC in a solvent system of chloroform/methanol/acetic acid/water (60/50/1/4, v/v/v/v) (21). Because the present study demonstrates that this solvent system does not separate carb-PE from PE (Figure 2), any intrinsic carb-PE would not be separated from commercial preparations of purified egg yolk PE. The results of the above study, therefore, are consistent with the hypothesis that the PE (and probably PS) of the native erythrocyte membrane is intrinsically carbamylated and that incubation of native erythrocyte membrane lipid with PLD releases both carbamyl ethanolamine and carbamyl serine. However, because carbamyl serine is nonchromogenic with diacetyl, the observed absorbance (Figure 11) is likely specifically a measure of in vivo carb-PE. By measuring the concentration of phospholipid in membrane lipid/Triton X-100 micelles and quantitating the carbamyl ethanolamine released upon PLD treatment (Table 3), the level of carbamylated PE in a pool of erythrocytes from normal individuals (urea = 4.46 _+0.51 mM, n = 12) was calculated to be 2.85 + 0.65%. We believe that the above data provide preliminary evidence for the presence of carb-PE (and most likely carb-PS) as an intrinsic component of the native erythrocyte membrane. Discussion

Carbamylation of nonprotein amino groups has not been previously reported. The present investigation demonstrates that incubation of erythrocyte ghosts with cyanate, under physiological conditions of temperature and pH, results in carbamylation of the aminophospholipids phosphatidylethanolamine (PE) and phosphatidylserine (PS) on the same timescale as that observed :for membrane protein carbamylation. The degree of carbamylation of PE and PS following a 15-h incubation with 25 mM cyanate are essentially equivalent (15.76 +_0.09 and 13.34 +_0.81 mole% cyanate/phospholipid, respectively), indicating that both phospho]ipids are as equally reactive towards cyanate. Our investigation has also demonstrated that

treatment of erythrocyte lipid micelles with phospholipase D releases a species that produces a chromogen when reacted with diacetyl. The results are consistent with the identity of this species as carbamyl ethanolamine and, as such, provide evidence that carbamylated PE is an intrinsic component of the plasma membrane of erythrocytes from healthy individuals. Preliminary estimates indicate that approximately 3% of the total PE is present as carbamylated PE. Because the reaction of cyanate or any endogenous compound with aminophospholipids has not been previously reported, the physiological significance of this observation is unknown. We have previously demonstrated increased levels of carbamylated erythrocyte membrane protein in persons with uremia (5) and hypothesized that this may result in a destabilization of key cytoskeletal binding domains, and potentially contribute to the decreased erythrocyte lifetime observed in uremics (22). Although the present study has not directly shown that carb-PE and/or carb-PS are increased in uremics, it can be hypothesized, however, that, if this is the case, then the carbamylation of PE and/or PS may, likewise, contribute to membrane destabilization. This is because, in addition to the protein-protein interactions among the cytoskeletal matrix proteins (k d -~ 5 × 10 -8 M), it has been demonstrated that both spectrin and protein 4.1 bind with high affinity (k d -- 3.3 x 10 -7 M) and selectively to the headgroup of phosphatidylserine (PS) (23,24). The binding is electrostatic, with the negatively charged headgroup of PS interacting with positively charged regions within the binding site of the protein (24,25). It has been proposed that these proteinlipid interactions play a role in the maintenance of phospholipid asymmetry by their capacity to "fix" phosphatidylserine to the inner leaflet (26). Indeed, individuals with erythrocyte protein 4.1 deficiency display an altered pattern of membrane lipid distribution, with significant enrichment in the outer lipid leaflet of PS (27). Carbamylation can be expected to alter the protein-lipid interactions either through protein modification or lipid modification or, perhaps, both. First, the carbamylation of PE and/or PS removes a positive charge from the headgroup and, as such, changes the physiological charge of PE and PS from (0) to (-1) and (-1) to (-2), respectively. Second, protein lysine

TABLE 3

Quantitation of Total Phospholipid, Phosphatidylethanolamine, and Carbamylated-Phosphatidylethanolamine in Erythrocyte Lipid Micelles in the Absence and Presence of PLD Treatment

Micelle (-PLD) Micelle (+PLD)

Phospholipid* (nmol)

PE~ (nmol)

Diacetyl Assays (Abs 530 nm)

Carb-PE§ (nmol)

Carb PE/PE (%)

3653 _+711 same

1059 _+206 same

0.035 0.132 +_0.006

29.06 +_1.40

2.85 _+0.65

PE, Phosphatidylethanolamine; Carb-PE, carbamylated phosphatidylethanolamine. *Quantitated as phosphorus (9). tCalculated as 29% of total phospholipid (27). SMicelles reacted with diacetyl. §Calculated from a carbamyl ethanolamine standard curve. All values represent the mean of duplicate determinations. CLINICAL BIOCHEMISTRY,VOLUME 29, AUGUST 1996

343

TREPANIER AND THIBERT

carbamylation removes a positive charge (that may be directly or indirectly involved in PS binding) to generate a neutral species. Either case could induce altered protein-lipid binding affinities and potentially lead to membrane destabilization. The resulting destabilized membrane would not be expected to be as resilient to the mechanical stress incurred in the microcirculation as the native membrane and, as such, would lead to the premature erythrocyte destruction observed in both uremics (22) and upon in vitro carbamylation (28). Another intriguing mechanism of premature erythrocyte destruction in uremics may, possibly, involve PS directly. As a result of the lipid asymmetry of the erythrocyte membrane, PS normally resides entirely on the cytoplasmic side of the membrane (26). It has been proposed that the exposure of PS on the outer layer of the membrane may serve as a signal for recognition by phagocytes that, subsequently, eliminate the altered erythrocyte from circulation (26). In this regard, it is interesting to note that the erythrocytes from diabetes mellitus patients display: an altered lipid asymmetry, with PS appearing in the plasma membrane outer layer (29,30); elevated levels of adherence to mononuclear phagocytes and endothelial cells (31); and a decreased lifetime (32). In addition, recent studies (33) have shown that incubation of normal erythrocytes with high concentrations of glucose duplicates this effect. It was proposed that the PS reorientation may be caused by direct alteration of protein (glycosylation) or lipid (by fatty acyl chain oxidation). It is conceivable, therefore, that carbamylation of erythrocyte components m a y induce the exposure of carb-PS and PS on the outer layer of the membrane. C a r b a m y l a t i o n m a y also be involved in other pathophysiological conditions observed in uremics, such as the development of peripheral neuropathies (34). The condition involves both segmental demyelination and axonal degeneration in peripheral nerves and evidence of uremic neuropathy is present in approximately half of all patients on hemodialysis programs (35). It is believed that uremic neuropathy is secondary to retained, dialyzable toxins or metabolites normally excreted by the kidney; however, the cause remains elusive. In this regard, it is interesting to note that an unrelated study demonstrated that a number of patients with sickle-cell anemia being treated with cyanate, developed peripheral neuropathies (36). Microscopic examination of nerve biopsies clearly showed the presence of segmental demyelination. The neuropathy clinically improved with no specific treatment after cessation of cyanate administration (36). It is tempting, therefore, to hypothesize that carbamylation of either the abundant amounts of Schwann cell PE and PS making up the myelin sheath (7) or the myelin basic protein (7) m a y be playing a role here. Finally, because our results demonstrate the carbamylation of erythrocyte membrane PE and PS under physiological conditions, it can be hypothesized that the covalent modification of aminophospholip344

ids in vivo with other endogenous amine reactive species, such as glucose, may occur. In addition, it can be envisioned that covalent modifications of PE and PS may extend beyond the plasma membrane of the erythrocyte to subcellular locations, such as the endoplasmic reticulum or nuclear membrane within other cell types.

Acknowledgements This work was funded by research grants to R. J. Thibert from the Natural Sciences and Engineering Research Council of Canada.

References 1. Dirnhuber P, Schutz F. The isometric transformation of urea into ammonium cyanate in aqueous solutions. Biochem J 1948; 42: 628-32. 2. Stark GR, Stein WH, Moore S. Reactions of the cyanate present in aqueous urea with amino acids and proteins. J Biol Chem 1960; 235: 3177-81. 3. Kwan JTC, Carr EC, Barron JL, Bending MR. Carbamylated haemoglobin in normal, diabetic and uraemic patients. Ann Clin Biochem 1992; 29: 206-9. 4. Oimoni M, Ishikawa K, Kawasaki T, et al. Plasma carbamylated protein in renal failure. N Engl J Med 1981; 308: 655-6. 5. Trepanier DJ. Carbamylation of erythrocyte membrane components. PhD Thesis, University of Windsor, Windsor, Ontario, 1995. 6. Gennis RB. The structure and composition ofbiomembranes. In: Cantor CR, ed. Biomembranes molecular structure and function, Chapter 1. Pp. 1-35, New York: Springer-Verlag Inc., 1989. 7. Gennis RB. Lateral and transverse assymmetry in membranes. In: Cantor CR, ed. Biomembranes, molecular structure and function, Chapter 4. Pp. 138165, New York: Springer-Verlag Inc., 1989. 8. Dodge JT, Mitchell C, Hanahan DJ. The preparation and chemical characteristics of hemoglobin-free ghosts of human erythrocytes. Arch Biochem Biophys 1963; 100: 119-30. 9. Higgins JA. Separation and analysis of membrane lipid components. In: Findlay JBC, Evans WH, eds. Biological membrane: a practical approach, Chapter

4. Pp. 103-137, Oxford: IRL Press, 1987. 10. Ames BN. Assay of inorganic phosphate, total phosphate and phosphatases. Methods Enzymol 1966; 8: 115-8. 11. Marcel MA, Hass SM. Lowry protein as modified for membrane proteins. Anal Biochem 1978; 87: 206-10. 12. Folch J, Lees M, Stanley GHS. A simple method for the isolation and purification of total lipids from animal tissues. J Biol Chem 1957; 226: 497-509. 13. Vaskovsky VE, Kostevsky EY. Modified spray for the detection of phospholipid on thin-layer chromategrams. J Lipid Res 1968; 9: 396. 14. Conte MH, Bishop JKB. Nanogram quantification of nonpolar lipid classes in environmental samples by high performance thin-layer chromatography. Lipids 1988; 23: 493-500. 15. Holbrook PG, Pannell LK, Daly JW. Phospholipase D-catalyzed hydrolysis of phosphatidylcholine occurs with P-O bond cleavage. Biochem Biophys Acta 1991; 1084: 155-8. 16. Wybenga DR, Di Giorgio J, Pileggi VJ. Manual and CLINICAL BIOCHEMISTRY, VOLUME 29, AUGUST 1996

CARBAMYLATION OF ERYTHROCYTEMEMBRANE AMINOPHOSPHOLIPIDS

17.

18.

19. 20. 21. 22. 23. 24.

25.

26.

automated methods for urea nitrogen measurement in whole serum. Clin Chem 1971; 17: 891-5. Gennis RB. The structure and properties of membrane lipids. In: Cantor CR, ed. Biomembranes molecular structure and ]'unction, Chapter 2. Pp. 36-84, New York: Springer-Verlag Inc., 1989. Roberts JD, Caserio MJ. Organic nitrogen compounds. Amines, amides, nitriles, nitro, azo, diazo, and related compounds. In: Basic principles of organic chemistry, Chapter 19, Pp. 641-700. New York: WA Benjamin Inc., 1965. Dennis EA. Phospholipases. In: The enzymes, Volume 16. Pp. 307-53. New York: Academic Press, 1983. Genencor International Inc. Phospholipase D product description. Catalog of Enzymes, Rochester, NY, 1991. Technical Services Department, Sigma Chemical Company, St. Louis, MO, Personal communication, 1995. Eschbach JW, Korn D~ Finch CA. 14C cyanate as a tag for red cell survival in normal and uremic man. J Lab Clin Med 1977; 89: 823-8. Cohen AM, Liu SC, Lawler J, Derick L, Palek J. Identification of the protein 4.1 binding site to phosphatidylserine vesicles. Biochemistry 1988; 27: 614-9. Maksymiw R, Sen-fang S, Gaub H, Sackmann E. Electrostatic coupling of spectrin dimers to phosphatidylserine containing lipid lamellae. Biochemistry 1987; 26: 2983-90. Shifter KA, Goerke J:. Dfizgfines N, Fedor J, Shohet SB. Interaction of erythrocyte protein 4.1 with phospholipids. A monolayer and liposome study. Biochim Biophys Acta 1988; 937: 269-80. Schroit AJ, Zwaal RFA. Transbilayer movement of

CLINICAL BIOCHEMISTRY, VOLUME 29, AUGUST 1996

27. 28. 29.

30.

31. 32. 33.

34. 35. 36.

phospholipids in red cell and platelet membranes. Biochim Biophys Acta 1991; 1071: 313-29. Schwartz RS, Chiu DTY, Lubin B. Plasma membrane phospholipid organization in human erythrocytes. Curr Topics Hematol 1985; 5: 63-112. Lane TA, Burka ER. Decreased life span and membrane damage of carbamylated erythrocytes in vitro. Blood 1976; 47: 909-17. Wali RK, Jaffe S, Kumar D, Kalra VK. Alteration in organization of phospholipids in erythrocytes as factor in adherence to endothelial cells in diabetes mellitus. Diabetes 1988; 37: 104-11. Lupu F, Calb M, Fixman A. Alterations of phospholipid asymmetry in the membrane of spontaneously aggregated platelets in diabetes. Thrombosis Research 1988; 5: 605-16. Zachowski A, Craescu CT, Galacteros F, Devaux PF. Abnormality of phospholipid transverse diffusion in sickle erythrocytes. J Clin Invest 1985; 75: 1713-7. Peterson CM, Koenig RV, Jones RL, Melvin ET, Lehrman ML. Reversible hematologic sequelae of diabetes mellitus. Ann Intern Med 1977; 86: 425-9. Wilson MJ, Richter-Lowney K, Daleke DL. Hyperglycemia induces of loss of phospholipid asymmetry in h u m a n erythrocytes. Biochemistry 1993; 32: 11302-10. Asbury AK, Victor M, Adams RD. Uremic polyneuropathy. Arch Neurol 1963; 8: 413-28. Asbury AK: Uremic neuropathy. In: Dyck PJ, Thomas PK and Lamert, EH (eds): Peripheral Neuropathy, Vol II, P. 982. Philadelphia: WB Saunders, 1975. Peterson CM, Tsairis P, Ohnishi A, Lu YS, Grady R, Cerami A, Dyck PJ. Sodium cyanate induced polyneuropathy in patients with sickle-ceU disease. Ann Intern Med 1974; 81: 152-8.

345