INCREASED MEMBRANE THIOL OXIDATION IN SICKLE ERYTHROCYTES
by Benjamin Albert Hill Junior
SubmUted to.the Faculty of the School of Graduate Studies of the Medical College of Georgia in Partial Fulfillment of the Requirements for the Degree of Master of Science June 1988 INCREASED MEMBRANE THIOL OXIDATION IN SICKLE ERYrHROCYrES
This thesis submitted by Benjamin Albert Hill Junior has been examined ~nd approved by an appointed committee of the faculty of the School of Graduate Studies of the Medical College of Georgia. The signatures which appear below verify the fact that all required changes have· been irtcorporated and that the thesis has .re(:eived final approval with reference to content, form, and accuracy of presentation. This thesis is therefore accepted in partial fulfillment of the requirements for the degree of Master of Science.
___ _jjJ.:;:_JiJ}f__ _ Date (C) Benjamin Albert Hill Junior . All. Rights Reserved · ACKNOWLEDGEMENTS
I would like to express my sincere thanks to the following_ people for their individual and/or group c~:mtribution towards the. completion of this degree: Dr. Rashid Ahktar, .Or. David Lapp, Dr. Frederick Garver and Dr. Abdulla Kutlar, my advisory committee members, for their -constructive critisms throughout the process; Dr. Muthukrishnan Ramachandran who provided me initial guidance in the laboratory; Dr. Mruthinti ~atyanarayana ~wamy -w~ose help _wit~ HP.LC and Gel Electrophoresis techniques became paramount to r:nY work; Dr. Chang-Chun Qiu whose constant critiques helped me to understand the logic behind ~any of the procedures; Amminikutty "~ary" Abraham, whose expert advic~ regarding 'inicr
~ . . . . Ocasio, all who .in .some way assisted me in. the accum.uiation of- data; Dorothy Fulmer whose computer assistance provided m~ help during the writing phase; Kim Norris, Joy Chang and Dr. AI Hesser whose patience, understanding and morale support contributed immensely to my state of being; and lastly_ and certainly most importantly Dr. Edathara · C .. Abraham "The Boss", my major advisor, who went to bat for me countless times, guided me through all phases of the program giving me advice, encouragement and supJ)Ort when the going became rough. I truly appreciated the help.
iv DEDICATION
This thesis is dedicated to my wife, Donna, who has toiled, sweated and endured air the hardships throughout all phases of study; without her abundant love and curative encouragement, success would not have been possible. Special appreciation must be extended to Jennifer and Christopher who have had to spend most of the last two years without Dad. 1- hope they will understand someday that I . all this work was necessary so as to provide us a chance for a brighter, more promising future.
v TABLE OF CONTENTS·
Page UST OF ABBREVIATIONS vii UST OFTABLES AND FIGURES viii INTRODUCTION 1. Statement of the Problem 1
2. Revi~w of Related Literature . 1 MATERIALS AND METHODS 1. Overview 4 2. .Sample Collection 4 3. . Hematological Studies · 4 4. Stractan Density Gradient ·4 5. In Vitro 0);(idation 8 -6. Membrane Preparation· 9 7. ·Protein Determination 9 8. Thiol Disulfide Exchange Chromatography 9 9. Protein Characterization 12 RESULTS 1. Preliminary Baseline Studies 15 . 2. Membrane Thiol Oxidation Studies 16 3. · Stractan Density Gradient Centrifugation Studies 27 4. Chemical Oxidation Studies 37 DISCUSSION 45 SUMMARY 53 REFERENCES OF UTERATURE CITED 55.
vi LIST OF ABBREVIATIONS
M Normal individual (Hb A) PS Sickle cell trait AUFS Absorbance unit full scale E('A Bicinchoninic Acid
BSKG Buffered Saline with Pot~ssium and Glucose BME B-mercaptoethanol Cs Concentration of Stractan EDTA Disodium Ethylenediamine Tetraacetate FF First Filtrate Ge Gel-Bound Filtrate HMW High Molecular Weight HPLC High Pressure Liquid Chromatography ISC Irreversibly Sickled Cells ObsOsm Observed Osmolality PAGE Polyacrylamide Gel Electrophoresis PBS Phosphate Buffered Saline PMSF Phenylmethylsulfonylfluoride SDS Sodium dodecyl.sulfate $ Homozygous Sickle cell anemia TOsm True Osmolality
vii LIST OFTABLES Table Page
lA. Level of membrane thjol oxjdjzed protejns jn erythrocyte ghosts of normal (AA). 17
lB. Leyel of membrane thjol oxjdjzed protejns jn erythrocyte ghosts of normal and sjckle cell anemja (SS) patjents. 18
fl. Leyel of membrane thjol oxjdatjon jn other sjckle hemoglobjnopathjes and nOnsjckle hemoglobinopathies and hemolytic anemjas. 19
UST OF FIGURES
Figure Page-
1 . Scheme of research methodology. 5
- 2. Chemjcal structure of Thjol-actjyated 48 Sepharose. 10
3. Molecular wejght calibratjon of molecular sjeye columns. 14
4. Changes jn membrane thjol oxjdatjon wjth jncreasjng age - or decreasjng leyels of fetal hemoglobin jn a Sjckle cell _anemja child.- 21
5. Separation of normal and Sickle cell samples bySDS-PAGE. 23
viii Figure Page
6. Malw:ular- &ieve H~LC-- separation. of red . · protein& gt a nann I . ceU membrane a and a Sickle cell anem·Ja sample. 25
7. . Molecular sjeve HPLC---- characterizaf · . aKidi;;~;ed mernb . -- Jan gt the thial __ rane fract1on. 26
8. Djscontjnuous St__ ractan gradjent . s;ell& falm a carm I . -· s;entntuaaliac at red . . a acq a Sickle s;ell patiect sample&, 28
9. Percent djstrjbutjon of cells jn Stra . cs:!DIII.!I& ID 5) aDd S . ClaD ECIIC!i!:lD& iD M . _s pat1ents (n 4). 29
10. fBed ·cell jndjcjes (MCV -& MC__ I::IQl ()f the St caciKlDaled cell -.caclac acadiecl traer___ Ions from .normal ·-- (A cell anemja (SS) .A)and Sickle · ---· pat1ents. 30
11. Percent. ISC's ·mtractan 5 fractjons. of ·SS pat'ieDI&, 31
12. ~ea:ecl felal bemaalabic ll::lbE . . fcacliami of ss . · l iD lbe fllur Slraclac .; · -- -- patients, . 32
13. Percent retjculocytes jn normal AA . S!raclac fractigmUjgcs, ( l acd Sickle cell !SSl 33 .
14. Percent. membrane th'i!:l I QX!dabgn. . . · Sickle cell (SS) P r m carmal (M) and a Jent sampl f · 35
15. HPLC mo 1ecular · Sleye. separation . of . . . carmal acd Sickl . . Stcaclac tracliacated . e cell patient samples. . 36
ix Figure -Page
16. se aration (reducing conditions) ~f membrane SOS-PAGE: P . . d'!filr.ecttlml! · proteins after diamide OXIdation for I - intervals. 38
17. SDS-PAGE: liflParation (non-reduc!Og. cohditions)'th of : ~ro oxidized proteins after treatment WI JnVI . , 'd different concentrar !O ns -of d1am1 e. 39
18. . f el wells of an -so_S-PAGE (non-reducing ~ Clo!ifl UP YlflW o g . . . normal samplfl · .. g ac m Vlt/'.Q oxL.di;~;fld . cocdjtjocsl contaiD!~ . ntratiomi of djamidfl, treated with increas!Og conce - . - 40
19. . .HPLC separation. . . o_ f membrane- proteins- MQiflCUia[ Iii~~ > rna( r.ed ®llli wjtb 7,5 roM afl!lrin vitro gx!daMD Qf DQC . diamide for different tjme intervals. 42
20. . HPLC separation. o_ f membrane proteins Mg!fl®lac lii!M!. --: !lila! rfld Cllllli ll!lilh ft .1 ·n vitro oxidation of no . a fl[ d' 'de increast__'ng concentrations- _ _ oflaml0_ . 43
21. . . 'gc ll!litb Elflla-mflrcapJgfllbangl ~D.Ibfl (Df(UflDCfl gf r.tldU<:I! . . . f D in vitro QX!d!;i;fld molecular Sieve. HPLC -separation o_ a . normal sample. 44
22. Scheme depicting the probable. ro Ie of thiol oxjdatjon jn dense cell formation. 48
X INTRODUCTION
1. Statement of the Problem: Sickle Cells are known to be more susceptible to oxidant stress than normal cell.s (1 ). Indications of an increased oxidation of sickle cell
membranes~ came from a recent study of a few randomly chosen sickle cell anemia (SS) patients by Rank et al (2) who separated erythrocyte membrane proteins
I into thiol-oxidized and thiol-reduced fractions. In this study no attempt was made by ,these authors to relate membrane thiol changes to the age of the patient or to pertin.ent genetic and hematological information such as the levels of fetal hemoglobin (Hb F) and dense cells. Furthermore, no relationship has been establi.shed between· the biological age of the cell.s and the extent of thiol oxidation. In the present study I have focused on the effect(s) that dense cells, age of the cell and certain hematological parameters such as reticulocytes and fetal hemoglobin (HbF) have on membrane thiol oxidation. · 2. Review of Related Literature: The primary defect resulting in the pathophysiology of sickle cell disease undoubtedly lies in the low solubility of deoxy-hemoglobin S leading to intracellular hemoglobin polymerization. Corpuscular hemoglobin concentrations appears to be the principal factor that can determine the extent of intracellular polymerization of hemoglobin leading to morphological ·deformation of sickle red cells. The peculiar oblong shaped irreversibly sickled cells (ISC) are clearly evident in deoxygenated blood smears and may comprise up to 30-40o/o of the total red cell population in some SS patients. The diminished deformability of ISCs is reflected by the decreased ability of these cells to negotiate certain areas of the microvasculature. The obstruction of small blood vessels by these cells is generally believed to account for "crisis" as well as various chronic clinical manifestations. The sickle ·red cell, ISC's in particular, have abnormally low levels of potassium (K+), large amounts of calcium (Ca++), and elevated levels of sodium (Na+) ( 3-6). This phenomenon, known as the "Gardos. Effect", exhibits metabolically depleted cells
1 2.
losing K+ while gaining Na+. The increased loss in cellular K+, possibly in the form of · KCI, has an accompanying loss of cellular water thereby leading to dehydration and a resultant increase in the mean corpuscular hemoglobin concentration (MCHC).
It is indeed pos~ible that increased me~brane thiol oxidation contributes to the . . elevated levels of ca++ in the SS cell through the inhibition· of the thiol" containing.
enzyme ca++. ATPase resulting in a gradual loss in the function of t~e calcium pump. Furthermore, membrane thiol oxidation could directly modify ion channels thereby increasing cation permeabil~ty resulting in· an intracellular K+ loss leading to celiular
· dehydration and eventually de rise cell formation. Lastly, it could _be po~sible th~t increased thiol crosslinking of red cell membrane proteins (e.g. spectrin and band 3 protein) with denatured hemoglobin could lead to clustering and ultimate formation of
the cell senescent antigen which leads to the splenic sequestration and r~moval of the cell from circulation (7 -1 0). . . . . Extensive heterogeneity in the density distribution of red cells determined by Stractan density gradient method have been reported in SS patients (11 ,12). This
heterogeneity is the .r~~ult of the presence of a rather ~a_rg~ but variable number of young cells and a group of "dense" red cells in these patients .. The term "dense cell" has_ been widely used in the literature, to represent density gradient separated red cell fractions with a density > 1.120, i.e. with ~ MC~C? >37 g/dl (this criteria seems to vary slightly from laboratory to laboratory), whic·h would be the most dense fraction obtained ...... by any of the .centrifugation methods: . .Sin~e. ·IS9's constitute the major portion of the so~called dense cell fra_cti~n, synonymous usage of the terms ,;ISC'$" and "dense cells" have become a common, practice.
Oxidation of m·ambrane thiols occuring at an ~nhanced rate in sickle cells can lead to membrane changes that may contribute to der1se cell formation. Auto-oxidation of
hemoglobins produce·s oxygen fre~ radicals in the red cells (1 ). This process is .enhanced in red cells containing abnormal hemoglobins, particularly when the abnormal hemoglobins·, like Hemoglobin S, have high affinity for the red cell membrane. Hebbel et . al (13) have shown excessive oxidation in $ickle.erythrocyte membranes; and as
described before such events may contr~bute to the accelerated membrane se-nescence ?f these cells. Available evidence suggests that the soluble anti-oxidant system, i.e. superoxide dismutase, catalase, glutathione peroxidase, glutathione reductase and 3.
reduced glutat~ione is altered in sickle cells (14-19). However, the fe~ studies which have been reported have surfaced rather inconclusive findings. It appears that superoxide dismutase and glutathione peroxidase-levels are either normal or even higher in SS red ·cells (14, 15, 18, 19). Catalase, reduced glutathione and glutathione reductase levels are lower than in normal ABC's (11 ,15.-17,20). Increased levels of protein aggregates believed to be a consequence Qf thiol oxidation were shown in SS red cell membranes by Thiol-activated Sepharose 48 affinity column chromatography (2). SDS-PAGE characterization of the thiol oxidized proteins revealed a generalized oxidative perturbation of sickle .R8C membranes. An· proteins including cytoskeletal protein bands 1, 2, 2.1 and 4.1 were involved in oxidation and protein aggregation. In a recent microvesiculation study by Wagner et al (21) it was seen that bands 1 and 2 (spectrin) may not necessarily be thiol oxidized-proteins as they fail to appear in the first filtrate when fractionated by Thiol-activated 48 Sepharose chromatography. This is in sharp contrast to the findings of Rank et al (2) who showed the presence of spectrin in small amounts even. in normal controls. Nevertheless, it is currently not known whether these oxidative changes are occurring predominantly in ISC/dense cell membranes or contributary to their formation. Moreover, It has not been shown whether the age of the patient, the level of fetal hemoglobin (Hb F), or reticulocyte count can influe-nce these events. · In vjtro oxidation studies have been performed using diamide, hydrogen peroxide . ' and deoxygenation techniques (22,23). These experiments were designed in an attempt to mimic the oxidation processes that occur at an abnormal rat~ in the sickle cell. The resultant data so far has remained inconclusive as to the identity of the proteins involved in oxidation. The reason for this may lie in the fact that thiol oxidation could produce high molecular weight (HMW) aggregates which cannot be observed by present
techniques ~nd thus far there has been few techniques other than SDS-PAGE used for the characterization of the oxidized proteins. 4
METHODS 1. Studjes of Membrane ChangeS Resulting from Thjo! Oxjdatjon- An Overvjew: A systematic approach to the proposed study of thiol oxidation in sickle cell anemia is schematically presented in Figure 1. 2. Collection of Blood Samples: Whole blood was drawn in EDT A containing
"purple top" v~cutainer blood tubes from pediatric patients (1 month to 15 years of age) reporting to the Sickle Cell Clinic after obtaining consent from their parents. Patients of different ages were chosen and one patient was foHowed .from 4 weeks after birth to 40 weeks. Complete blood counts (CBC's) as well as the diagnosis of hemoglobinopathies was performed by the Core Laboratory of the Sickle Cell Center. 3. Hemato!ogjca! Studjes: Percent irreversibly sickle cells (ISC's) were determined by analyzing blood smears of whole blood and washed Stractan fractionated and chemically oxidized red cells. ISC counts were made in non-overlapping microscope fields until greater than 500 cells were tabulated. Cells were designated ISC when having a length two times greater than their width (24). The number of ISC's relative to the total number of eells was expressed as a percentage. The percent reticulocytes was determined by blood smear analysis also, however, these red cells were stained (1 :1) with 0.1 0°/o methylene blue (25). The -percent dense cells was evaluated by·the pthalate ester microcapillary centrifugation method (12,26-28). This was accomplished by first filling a capillary tube with 5 millimeters of a pthalate ester mixture (density 1.118) which was capable of segregating dense cells from other cells after centrifugation. The remainder of the the capillary tube was then filled with whole blood and centrifuged for15 minutes in a Damon-IEC microhematocrit (microcapillary) centrifuge. The percent dense cells could be calculated by determining the amount of cells sedimenting below and above th.is pthalate mixture by column measurement. The percent Hb F was determined by a cation exchange microcolumn method (29). 4. Separation of Red Cells by Discontinuous Stractan Gradjent Centrifugation:
Disconti~uous Stractan gradients were prepared by decreasing the number of Stractan fractions used by Clark eta! (30) who modified the work of Corash eta! (31 ). a. Preparation of Stractan: The Stractan density gradient centrifugation studies were dependent on the proper preparation of the Stractan (Arabinose-galactan). Four hundred and fifty grams of Stractan powder was dissolved (1 :1) in dH20, deionized FIGURE 1. Scheme of research methodology
Abbreviations:
Hb F - Fetal hemoglobin ABC's - Red blood cells SDS-PAGE - Sodium dodecyl sulfate polyacrylamide gel electrophoresis , 5
·Collection of Whole Blood
DeterJnation of
(t).% HbF (2) % dense
Separation of ------Wash ABC's ======~ Induce Oxidation Stractan Fractions
Prepare Membranes
. - ' . . ' . ' . .- .. Thioi-Activated 4B Sepharose Affinity Column Chromatography
Thiol-oxidized fractions Thiol-reduced fractions
. . ~--·. __;______;____ .,-----L----'---~--,--- t Separation by Gel· Separation by High · Electrophoresis Pressure Liquid SDS-PAGE Chromatography 6
using a mixture of Amberlite IR.~120(H) CP cation exchange and Amberlite IR-410 CP anion exchange resins. The.Stractan should have been deionized after passing through the · mixed resin·,- however after a deionization efficiency check was made, this usually was
' ' not the case. The deionization efficiency check involved the determination of true osmolality and was accomplished by completion of a step-by-step approach which provided sufficient information so that the true ~smolality could be calculated using the following formula:
[( Obs Osm BSKG/Stracian) + ( AW)] - [ 2 (Obs Qsm BSKG)]
(AW)
Where, Obs Osm BSKG is the observed osmolality of the BSKG (a nitrogenated buffered
saline solutio~ with potassiu~ and glucose), Obs Osm BSKG/Stractan is the observed
osmolality of the BSKG and Stractan mixture and AW is the available water in the
Stractan solution.
T~.e values_ in t~e preceding formula were derived as follows: The Observed . Osmolality values (Obs Osm) could only be determined by analyzing the solutions wi~h an osmom-eter. First, the concentration of the Stractan (Cs) had to be determined using a
graph which plotted Cs against the density of the Stractan solution which was determined by hydrometer measurements. Available water (AW) was calculated using the formula 7
AW = 0.64 (Cs)·_ Once all these data had been obtained, the above formula gave the True Osmolality (T Osm). If the calculated T Osm value of the Stractan was found to be greater than 1oo mOsm, then, th~ de ionization and calculation ofT Osm had to be repeated. If however, the T Osm·value was less than 100 mOsm,·the next step was to add the following chemicals so as to convert it into a "Stock" Stractan:
Bovine Serum Aibumin (BSA) 3.00g/1 00 ml Stractan solution Glucose 0.200_ g/1 00 ml Available Water 0.15 M K3P04 Buffer 11.1 0 mV1 00 ml Stractan solution MgCI2 · 6 H20 0.116 g/1 00 ml Available Water
------~---- . After adding the chemicals to make stock Stractan, final pH and T Osm checks
confirmed the quality of the solution. The T Osm ~hould now fall in the range betWeen 290 and 295 mOsm. If this failed to be the case, milligram amounts of NaCI (determined by the use of the following formula) were added to adjust the -final osmolality of the stock Stractan solution: ·
292.5 mOsm - (\----1>] (desired osmolality) (T Osm)
_____,;_ ____----"-....;.....__~- X · [volume of AW · 1=. A_djustment Factor [31.5 rriOsm] ·
Where, T Osm is the true osmolality, AW, is the vo,ume of available water in the · Stractan solution and the Adjustment Factor is the amount of NaCI which ~ad to be_ ad.ded to the stock ·stractan solution to bring the r· Osm to the de-sired osmolality range of 290-295 mOsm~
\ . b~ Preparation of the gradient and gradient fractions: Once the Stractan
was ready (i.e. pH~ 7.4), the gradient could be prepared by dilution o_f the stock Stractan solution with BSKG according to the percent Stractan desired fo~ each fraction~
The gradient was composedof.fo~:J_r fractions made up of different-percentage Stractan 8
solutions ranging from 19°/o, 22°/o, 2~0/o, . and 28°/o Stractan corresponding to densities 1.0785, 1.0919, 1.1 009 and 1.1195 respectively. The Stractan fractions were then carefully pipetted into the ultracentrifuge tubes beginning with the most dense fraction at the bottom. Care was taken not to disturb the previously pipetted layer to insure that a gradient was maintained. The sample Vias last to be applied into the ultracentrifuge tube. f!ed cells were washed three times and resuspended in nitrogenated BSKG. At this point, the ultracentrifuge tubes containing the samples were secured into swinging buckets for ultracentrifugation. The tubes were centrifuged i.n an SW41Ti rotor (LS-65 Beckman ultracentrifuge) at 20,000 rpm (68,000 x g) _for 50 minutes at soc. After centrifugation, Stractan density gradient fractionated samples ·were then collected by pi petting each fraction using modified 90° bent tipped. pasteur pipettes, pooled, thoroughly washed and resuspended in nitrogenated BSKG. The cell fractions prepared in this manner were characterized by determining hematological information and Hb F content. The recovery of the cells in each fraction was calculated on the basis of the total number of RBCs in each fraction. Cells sedimenting at 28°/o Stractan (density = 1.1195) were termed "dense cells". Membranes were prepared from each fraction of cells and later subjected to Thiol-activated Sepharose 4B chromatography for separation and quantitation of the oxidized and reduced thiol containing polypeptides. 5. lnducjng Membrane Thjo! Oxjdatjon by utilizing a Chemjcal Oxjdant: Diamide [ (CH3)2NCON=NCON(CH3)2 ] is a chemical oxidant which readily enters the red cells and oxidizes free thiols (32). Intact red cells were incubated in a Phosphate Buffered Saline (PBS) (pH 7.4) containing different concentrations of Diamide (i.e. 1.25mM to 40mM) for 5-120 minutes at 37°C in a metabolic shaker oscillating at 80 times per mi.nute~ After saline washes, membranes were prepared and then solubiiized in 2.00°/o SDS according to the method described by Rank et al (2) followed by Thiol-activated 48 Sepharc:>se chromatography. SDS-PAG.E and HPLC characterization of the thiol-oxidized membrane fractions were done under reducing and non reducing conditions. This
~ approach provided an opportunity to cOmpare the qualitative_ and quantitative differences in the response of the membrane. to an induced thiol oxidative challange between the In
~and jn vitro oxidation patterns of control and patients' samples. 9
6. Preparatjon of Membranes: This was done according to Dodge et al (33) as modified by Rank et al (2). Whole blood was centrifuged at 3000 rpm (4500 x g) in a Sorvall centrifuge to remove the plasma and the white cells followed by wahing three times with nitrogeriated PBS. Care was taken not to bubble air through the blood sample as this might have an adverse effect on the oxidation state of the membrane thiols .. After completing the last PBS wash, the red cells underwent hypotonic lysis and five washes in 20 volumes of a nitrogenated ice cold sodium phosphate membrane buffer solution containing 0.1 OOg/L phenylmethylsulfonylfluoride (PMSF) and 0.037g/L EDT A. The resultant cell lysate was centrifuged in a Sorvall centrifuge ~t 12,500 rpm (18,800 x g) for 15 minutes for the first spin with 10 minute subsequent spins. Centrifugation proceeded until the ·supernatant became clear and the membranes were creamy white. When satisfied with the color of the membrane·s, a final volume of membrane buffer was added (1 :1) adjusting the protein volume to approximately 2-3 mg/ml. The membranes prepared in this manner were immediately utilized for further studies or stored at -20°C. The influence of blood o.r membrane storage C?Qnditions on the thiol values was studied so that optimum storage conditions could be followed. 7. Qeterminatjon of Prcitejn Concentratjon: Protein concentration was determined using the bicinchoninic acid (BCA) method, an assay developed by Smith et al (34). This method provided an alternative to the Lowry method (35) which was equally sensitive yet was unable to tolerate higher concentrations of SDS. The sensitivity of the BCA method was determined by using bovine serum albumin (BSA) and the results were compared with_those obtained by. the Lowry method. Since the buffers ·utilized in the Thiol-activated 4B Sepharose column chromatography contained sufficient amounts of SDS to interfere with the Lowry method, it was necessary to use . the BCA method for all protein determinations . . 8. Thjol Disulfide Exchange Chromatography: Thiol-activated 4B Sepharose was used for this chromatographic method (36). This sepharose is a mixed disulfide formed between 2,2'-dipyridyl disulfate and glutathione coupled to CN-Br activated 4B Sepharose. The hydrophillic gluta_thioi'le residue acts as a spaeer group which decreases· steric interference at the terminal thiol group. Figure 2 shows a partial structure. FIGURE2. Chemical structure of Thjol-actiyated 48 Sepharose. 10
0 c- ""/- 0
NH 1 1
- Only the reduced thiol-containing proteins bind to this resin, whereas oxidized thiol-containing proteins and proteins not containing thiols run through the column. The bound proteins can then be eluted with a buffer containing B-mercaptoethanol. · a. Preparatjon of the Thjol-activated 4B Sepharose. Thiol-activated 4B Sepharose was dissolved, rinsed of the chemical preservatives. and allowed to swell in distilled water (dH20). The Sepharose contains approximately 1.0 micromole 2-pyridyl disulfide groups per ml swollen gel. The swelled Sepharose was poured directly onto a BOchner funnel w/Whatman filter (No.1) and suctioned-carefully until most of the moisture was·gone. The swelled Sepharose was set aside for subsequent preparation of column .. b. Preparatjon of Thjol Columns: (1) Macrocolumns for qualitative studies: A 1Omm. x 21 Omm. glass column was tared on scales for the direct transfer of 3.00 g swelled sepharose into the column. The column w~s then affixed to a ring stand for additional attachments which sufficiently provided a closed system, flushed with nitrogenated gel buffer to ensure that all of the preservatives have been removed and that air was excluded from the system, and nitrogenated by bubbling each column under N2 gas for 15 minutes. The column was then capped with a rubber column stopper and double sealed with parafilm squares held tightly in place with a large joint clamp. At this point, the column was ready for. the loading:of solubilized proteins. One milligram to 1.5 milligrams protein could be processed in such a column. This macrocolumn method was used to isolate enough quantities of the thiol oxidized and thiol reduced proteins for further characterization. (2) Microcolumns for quantitative studies. A method for microcolumn chromatography was developed so that small amounts of membrane proteins such as those recovered from Stractan density gradient fractionation, e.g. 125 to 150 ~g, could be processed tp datermine the percent level of thiol oxidized proteins. A 7mm. x 63mm. clean plastic microcolumn w/filter was tared on scales for the direct transfer of 0.200 g swelled Sepharose into the column. The microcolumn was nitrogenated and 9apped on both ends awaiting the loading of solubilized proteins. c. Loading of Sample on Thjol columns: ABC membranes were prepared as above by hypotonic lysis and membrane buffer wash~s. Membranes were heat 12
solubilized (1 :1) for 4 minutes in a 2.00% SDS solution; then applied to the column (1.00 mg protein for the macromethod and 0.125 mg protein for the micromethod) thoroughly saturated with a "gel buffer' containing 0.1 Og/L PMSf, 0.037g/L EDTA and 20g/L SDS ; and left for 15 hrs in a nitrogen saturated environment. The unbound protein fraction (containing the thiol-oxidized proteins) was collected in dialysis tubing and labelled as the first filtrate. The column was then flushed wittJ a "wash buffer''. which was similar to the gel buffer exc_ept that it contains less salt. The bound proteins were released from the gel by using 10°/o B-mercaptoethanol (0.200 ml in 2.00 ml wash buffer for the macromethod and 0.050 ml in 0.500 ml wash buffer for the micromethod), collected into another dialysis tubing labelled as the gel bound fraction. Dialysis continued for 8 to 24 hours against a "dialysis buffer'' which contained 3.72mg/L EDTA, 0.01g/L PMSF and 0.10g/L SDS. After lyophilization, proteins were resuspended in distilled water and the percent of protein which had eluted into each fraction was determined. 9. Characterjzatjon of thjol-oxjdjzed and thjol-reduced membrane fractions: a. Separatjon of Protejns by SDS-PAGE: Membrane proteins were fractionated by SDS-PAGE according to the procedure described by Laemmli (38).
Membrane proteins were solubilized in a buffer containing 15~6% glycerol, 0.86M B-mercaptoethanol (or without mercaptoethanol under non-reducing conditions), 5.00% SDS, 0.05°/o bromophenol blue. Twenty-five Jlg protein was applied on a .1 0°/o polyacrylamide gel and run at 20 milliamps for 30 minutes and 30 milliamps for the remaining 5-7 hours~ stai_ned for 30 minutes with 0.1 °/o Coomassie brilliant blue dye and lastly destained overnight (12-18 hours) in a 10:40:50 acetic acid/methyl alcohoVdistilled water solution. Thiol-oxidized (first filtrate) and thiol-reduced (membrane bound) fractions resulting from the Thiol-activated 48 Sepharose affinity chromatography were fractionated along with the original membrane. b. Separatjon of Protejns by Molecular Sjeye Hjgh pressure Ligujd ChromatOgraphy (HPLC): Membrane proteins were fractionated, by molecular sieve HPLC using a Beckman HPLC System including model11 OA pump, model 421 controller and-~ model160 fixed waveleng!h spectrophotometer wi_th a·1 0 Jll flow cell. Peaks were . recorded and integrated using a Hewlett Packard 3390A recording integrator. Molecular sieve HPLC size exclusion separation of proteins have only recently become available 13
with the development of rigid silica based packing material. Th~ columns used in the separation of the .erythrocyte membranes were TSK (Toya Soda-Aitex) size exclusion columns. These columns have a microparticulate rigid silica support with a polyvinyl matrix. This provides a neutral, hydrophobic, matrix of varying pore sizes for the separation of molecules based on size and molecular shape (38-47). Membrane preparations were heat solubilized (1 :1) for 4 minutes in a 2.00% SDS solution then applied to the HPLC column (1 00-150 J.Lg protein); equilibrated with the mobile phase containing 0.100M sodium phosphate, 0.100°/o'SDS, pH 7.0.
Isocratic flowrate was fixed by the ~ntroller at 1.0 ~Vmin and absorbance was measured at 280 nm; the detector AU FS was 0.01 . The recording integrator settings were as follows: area reject: 10,000, threshold: 5, peak width 0.265, chart speed 0.2, attenuation: 5. The establishment of reducing conditions was accomplished by solublizing membrane proteins in 0.80M B-mercaptoethanol then running them through the column equilibrated in an HPLC buffer containing 0.1 OM sodium phosphate , w/ 0.100% SDS & 0.70M B-mercaptoethanol. The erythrocyte membrane proteins were then separated using a TSK 3000 SW, 7 .5mm x 600mm column tandemly coupled with a TSK 4000 SW, 7.5mm x 300mm column. This combination resulted in a molecular weight range of approximately 1.5 X 1o6 to 2.0 X 1o3 daltons. All chromatograms were run at ambient temperature . · The molecular sieve HPLC system was calibrated using a mixture of proteins of known molecular weights. A mixture of thyroglobulin (mol.wt. 670,000), gam·ma globulin (mol.wt. 157,000), ovalbumin (mol.wt. 44,000),' and myoglobin (mol wt. 17,000)was injected into the molecular sieve HPLC system. The result$ were plotted as · the log molecular weights as Kav (Kav = (Ve-Vo/(Vi-Vo) where Ve =elution volume of the peak, Vi =column volume, and Vo =void volume) (Figure 3). FIGURE3. Molecular weight caljbratjon of molecular sjeye columns. The molecular sieve HPLC system was calibrated using thyroglobulin (mol.wt. 670,000), gamma globulin (mol.wt. 157,000), ovalbumin (mo!.wt. 44,000) and myoglobin (mol.wt. 17,000). The results were pl9tted as the log. molecular weights (Kav) where Kav = (Ve-Vo)/(Vi-Vo).
Abbreviations:
HPLC - High pressure liquid chromatography Kav - log molecular weights .mol. wt. - molecular weight Ve - elution volu.me of the peak Vi - column volume Vo - void volume 14
--..,
6 Thyroglobulin 670 kd
Gamma globulin 157 kd 5 .s:::. -C) Ovalbumin 44 kd ; .... Myoglobin 17 kd m '3 4 (,) Cl) 0 E C) .9 3
1 0 1 5 20 25 30 35 40
Retention time (minutes) RESULTS
1. Influence of Sample preparatjon and Storage Condjtjons on Oxjdjzed Thiol
Values: One of the concerns of any measurement of~ generated thiol oxidation is that the oxidation of the membr~ne could be altered during the handling of the sample. Different exp_erimental parameters such as (a) the effect of using rtitrogenated buffers, (b) the length of tirne required for sample processing and membrane preparation and (c) the length of time membrane were stored had to be considered as possible vehicles for the unintentional inducing of membrane thiol oxidation and therefore had to be critically analyzed. a. Buffer njtrogenatjon data: An experiment was designed to show a comparison of using nitroge~ated versus non-nitrogenated BSKG and 0.200 M sodium phosphate membrane buffer solutions during cell washing and membrane preparation. The results of the experiment (data not shown) showed that there was no statistical difference in the percent membrane thiol oxidation between the use of nitrogenated and non-nitrogenated buffers in cell washing and the preparation of membranes. Nevertheless, as a precaution, only nitrogenated buffers were used for further experimentation. b. Membrane p[eparatjon dqta: Blood samples from SS patient were drawn at the Sickle Cell Clinic then normally transfered to the Core Laboratory before they were received in our laboratory. The elapsed time (three hours or less) the sample spent en route from the clinic to our laboratory, termed sample processing time, could possibly have an effect upon the oxidative state of the membrane thiols. Furthermore, after arrival in the laboratory, red cell washings and membrane preparation time take an additional five hour period (minimum). To simulate these conditions, normal and sickle celi whole blood samples were drawn, processed and prepared into membranes at different times after the collection of the samples. The control sample used in this experiment was washed and prepared into membranes immediately whereas other samples were allowed to sit at room temperature for up to three hours simulating 1 6
various sample processing times·at the Clinic and Core Laboratory (slow or busy day). Sample aliquots were taken at 1 .and 3 hours for washing and membrane preparation. After three hours had passed, the sample was placed and kept on ice for up to twelve hours simulating the handling procedures within our laboratory. Sample aliquots were taken and processed at the 6 and 12 hour mark. After twelve hours, about 5 mg glucose was added .to the sample before being placed overnight in the cold -room at 4°C. The next morning, twenty-four hours after drawing the sample the last aliquot was washed and
m~mbranes prepared, Thiol-activated 4B Sepharose chromatography was performed on each membrane sample .. The conclusion based on the findings of these data was that true oxi.dation value of the membrane thiols was not significantly altered during sample processing and membrane preparation. c. Membrane storage data: The membranes which were prepared and were immediately processed in the previous experiment were used as the experimental control for this experiment. When the membranes were initially prepared, they were · stored at -20°C. To conduct this experiment, these same membranes were removed from storage for analysis byThiol-activated 4B Sepharose chromatography at 1, 3, 6, 9 and 12 days. The result of the experiment showed no statistical difference in the thiol oxidation values due to membrane storage at -20°C over a twelve day period. Based on results of these three experiments, it was decided that all normal and SS patient samples would be washed in nitrogenated buffers and membranes prepared within a 24 hours period. Once samples had been made into membranes, they could be stored at -20°C for a period of time not. to exceed 12 days. 2. Membrane Thjol Oxjdatjon (whole blood samples) jn Patjents wjth Sjckle cell anemja and related condjtjonS as compared to normal controls: Blood samples from normal controls (M), sickle cell patient (SS) and other sickle and non·sickle hemoglobinopathies were collected and chromatographed to determine the percent · membrane thiol oxidation (Tables !A; 18 and II respectively). Other hemoglobinopathies . . - . include AS, CC, SC, SS- alpha thalassemia and SB-thalassemia. A heriditary spherocytosis control was chosen as an example of hemolytic anemia without any hemog!obinopathy. TABLE lA. Level of·membrane thjo! oxjdjzed protejns jn erythrocyte . ·ghosts Of normal (AA).. , Membrane thiol-oxidized proteins were separated and quantitated by· chromatography on columns of · · Thiol-actiVated 48 Sepharose utilizing·the microcolumn method as outlined in the text. All the controls ·(AA) are· ·: adult. volunteers.
Abbreviations:
Std. Dev. - Standard Deviation 1 7
TABLE lA. Level of membrane thiol oxidized proteins in erythrocyte ghosts of normal . . I (M) controls. .
Sample 0/o Thiol Oxidation B.R.(AA) 11.4 B.H.(AA) 11.7 K.N.(AA) 12.1 B.H.(AA) 12.2 B.P.(AA) 12.4 J.B.(AA) 13.3 M.S.(AA) 13.7 J.H.(AA) 14.3 R.W.(AA) 14.4 T.V.(AA) 15.0 C.M.(AA) 15.6 H.O.(AA) 15.8 M.R.(AA) 16..Z
Mean 13.8 Std. Dev. 1.7 Range 11.4-16.7 TABLE lB. .. Leye! of m·embrane thjo! oxjdjzed proteins jn erythrocyte ghosts Of Sjckle cell anemja (SS) patjents. ·Level of membrane thiol oxidation was estimated-as in Table lA. All the SS patients listed here have been regularly attending. the P~diatric Sickle Cell Clinic.
Abbreviations:
Retic - Reticulocytes Hb F - Fetal hemoglobin ISC's ~ Irreversibly sickled cells Std. Dev.- Standard Deviation 1 8
TABLE lB. Level of membrane thiol oxidized proteins in erythrocyte ghosts , age and pertinent hematology of sickle cell anemia (SS) patients.
Sample Age 0k Retic o/o HbF % Thiol Oxidation %1SC's C.R.(SS) 4wks 2.8 87.1 12.1 0.0 K.S.(SS) 1 yr 5.0 32.1 21.6 0.0 G.S.(SS) 2yrs 11.6 -16~ 1 31.6 C.T~(SS) 2 yrs 26.5 11.8 21.6 17.8 C.W.(SS) 3yrs 19.3 5.0 31.9 20.0 S.W.(SS) 6 yrs 10.0 11.5 24.1 4.4 M.M.(SS) 8 yrs 1.0 12.0 30.0 7.8 F.B.(SS) 9yrs 9.4 14.6 25.0 27.0 T.Z.(SS) 9yrs 4.4 11.6 26.1 8.0 T.R.(SS) 10 yrs 10.0 8.0 22.5 25.4 J.W.(SS) -10 yrs 6.5 11.6 28.7 7.7 T.S.(SS) 12 yrs 7.0 10.3 26.5 32.5 T.S.(SS) 14 yrs 10.4 6.8 24.4 27.9 R.D.(SS) 14 yrs 11.6 3.0 31.7 21.0 T.K.(SS) 14 yrs 6.4 6.8 24.9 8.9
K.S.(SS) ·~ u. ~ 2U 1.a3.
Mean 9.4 15.8 25.5 13.1 Std. Dev 6.2 20.2 5.0 11.3
Range 1.0-26.5 3.0-87.1 12.1-31.9 0.0-32.5 TABLE II. Leyel of membrane thjol oxjdatjon jn other sjckle and non-sjckle hemoglobjriopathjes and hemolytic anemjas. Level of membrane thlol oxidation was estimated as in Table lA.
·Abbreviations:
Retic - Reticulocytes Hb F- Fetal hemoglobin 1 9
TABLE II. Level of membrane thiol oxidation in other sickle and nonsickle hemoglobinopathies and hemolytic anemias.
Sample % Retic %HbF % Thiol Oxidation
K.S. (CC) 8.2 1.5 19.6· B.A. (CC) 3.7 3.2 20.3 G.~.(SS alpha thai) 22.8 J.T. (5. B thai) 11.0 V.T. (SC) 22.2 Y.Y. (SC) 24.0 A.M.(AS) 15.5 B.S.(AS) 18.6 J.O .(HS) 15.6 11.7 R.M.(HS) 11.0 -. 16.6 20
a. ABC Membrane Thjol Oxjdatjon Data: Erythrocyte ~embrane thiols in normal M individuals (Table lA) were found to be susceptible to some degree of oxidation as seen by a mean percent membrane thiol oxidation value of13.8o/o ± 1. 7 (range 11.4°/o to 16.7°/o). This mean value for control samples (n=13) is not significantly different from mean percent membrane thiol oxidation value 13.6°/o in controls used in other laboratories (2). Sickle cell (SS) individuals (Table IB) had almost a twofold increase in the mean percent membrane thiol oxidation value of 25.5°/o ± 5.0 ranging from 12.1 °/o to 31.9°/o. This mean value for SS patient samples (n=16) was slightly higher than the mean percent membrane thiol oxidation value of 21.5o/o, (range 15.9°/o to 28.4°/o) in SS patient samples used in the other laboratory (2) which is not surprising since the sampling popu.lation of patients from the other laboratory was unknown . . b. Correlatjon of Thjol Oxjdatjon Data wjth Leyels of Betjculocytes. ISC's and HbF: Statistical analyses were done to determine if relationships may exist between the percent level of membrane thiol oxidation and percent level of reticulocytes, irreversibly sickled cells (ISC's), and HbF. Correlational studies with SS patient samples (n=16) showed poor correlation (r= 0.161) between the level of reticulocytes and membrane thiol oxidation. This may indicate that the extent of hemolytic anemia or reticulocytosis does not influence the level of thiol oxidation. Normal levei of membrane thiol oxidation observed in a hereditary spherocytosis case also confirms that the mere presence of red cell hemolysis and overproduction of reticulocytes does not lead to increased membrane thiol changes. There was a poor correlation between the percent membrane thiol oxidation (r= 0.163) relative to the percent ISC's. ~ Hb F levels showed· the strongest correl~tion (r= -0.768) with the level of membrane thiol oxidation. c. Case Study Data: The SS patient C. B. was followed from 4 to 40 weeks so as to observe the effect of decreasing levels of fetal hemoglobin (HbF) upon the level of membrane thiol oxidation (Figure 4). As the HbE level decreased and sickle hemoglobin increased over time, increased level of thiol oxidation was expected because of the increased presence of the sickle hemoglobin. In the presence of high HbE, during FIGURE4 Changesjn membranethiol.qxidatjon with·jncreasjncj age or decreasjng levels of fetal hemoglobjn jn a Sickle cell anemja (SS) child. · _ The percent Hb F was determined by a cation exchange microcolumn method and percent thiol o~idation was· determined by Thiol-activated 4B $epharose chromatography during a postnatal period of 40 weeks in . an SS infant. ·
Abbreviations:
Hb F - Fetal hemoglobin 21
100
t: 80 0 iii "C )( 60 0 JJ 0/o Retics 0 0 :E + /o HbF 1- 40 + 0/o Thiol oxid. t: -G) eG) D. 20
0 4 26 29 31 32 39 40 Age (Weeks) 22
the first 29 weeks, the thiol oxidation values of this SS patient remained In the normal (AA) range. A transition then occured between the 29th and the 32nd week as thiol oxidation values increased. By 32 weeks, membrane thiol oxidation values rose to 30o/o and remained at ·this high level until the 40th week and probably beyond. d. SDS-PAGE Characterjzatjon: Characterization of the membrane polypeptides involved utilization of SDS-PAGE technique. Solubilized membranes
prepared from untreated or unf~actionated whole blood samples from normal controls and SS patients were loaded onto a 10.0°/o SDS polyacrylamide gel (Figure 5). Membrane proteins proceeded to separate into different" molecular weight oands which were identified according to the system_ used by Beck as follows: 1 (spectrin alpha · subunit), 2 (spectrin beta subunit), 2.1 (ankyrin), 3 (band 3 protein), 4.1a, 4.1b, 4.2, 5 (actin), 6 (glyceraldehyde-3-phosphate dehydrogenase), 7 and globin were visible on normal control samples. On the-same SDS polyacrylamide gel, it was seen that in SS patient samples, there appears to be a noticable decrease in the intensity of bands 4.1 a (80 kd) .and 7 (32 _kd) coupled with an increase in intensity of the globin band and lastly the appearance of a new band below band 7 (22 kd). The decrease in band 4.1 a (48,49) and increased assoCiation of globin to sickle cell membrane have been reported before (49). Whole membranes were then fractionated by Thiol- activated 4B Sepharose chromatography to separate the thiol oxidized proteins for further SDS-PAGE analysis which showed only the 32 kd Band 7, a periodic acid schiff base (PAS) protein, in the oxidized fraction of normal controls and SSsamples. In SS patient samples, it was also seen that there is a noticable decrease in the intensity of band 7 coupled with a new 22 kd band appearing below band 7. . FIGURES. Separation of normal and Sickle cell samples by SPS-PAGE. Membranes from normal controls and sickle cell patients were solubilized in a buffer containing 15.6°k glycerol, 0.86M BME, 5.0°/o SDS and 0.05°/o bromophenol blue. Twenty five Jlg protein was applied on the 10°/o polyacrylamide gel and run at 20 milliamps for 30 minutes and 30 ·milliamps for the remaining 5-7 hours, stained for 30 minutes with 0.1 °k Coomassie brilliant blue dye and lastly destained overnight (12 to18 hours) in a 10:40:50 acetic acid/ methyl alcohoV distilled water solution.' Lanes 1, 3, 4, 5, 7 and 8 contain ·control (AA) samples. Lanes 2 and 6 contain patient (SS) samples and lane 9 contains molecular marker SDS-7. Molecular weights of marker proteins are 66K, 45K, 36K, 29K, 24K and 20K daltons ..
Abbreviations:
SDS-PAGE- Sodium dodecyl sulfate polyacrylamide gel electrophoresis BME .. B-mercaptoethanol 23
LANES 1 2 3 4 5 6 7 8 9
AA ss AA AA AA ss AA AA SDS-7 mol wt band 1 band2 band 2.1 band3 band 4.1 (a&b) band 4.2 66K
.--··; band5 45K band 6 36K band7 29K 24K
·------globin 20K 24
Conspicuously absent was the protein spectrin (bands 1 and 2) which closely PB:rallels the findings of Chiu and Lubin (14). In contradiqtion to these data, Rank et al (2) found that the oxidized fractions do contain bands 1 and 2 (spectrins). The reason for this apparant contradiction is not clear. One major limitation of SDS-PAGE methodology was its inability to allow HMW aggregates, such as crosslinked thiol oxidized proteins, to enter the gel, thereby failing to undergo separation by molecular weight. e. HPLC Characterjzatjon: High Pressure Liquid Chromatography · (HPLC) studies were considered after modification of SDS-PAGE couJd not show the potential presence of HMW proteins. HPLC gel permeation chromatography allowed proteins as large as 7 x 1o6 daltons to be separated. Control and SS patient membrane samples were solubilized and loaded on the HPLC molecular sieve column (Figure 6). The resultant separation showed five distinct peaks eluting at retention times (R.T.) 15.30, 17.60, 21.00, 25.90 and 33.50 minutes which correlates with the molecular weights 1400 kd, 750 kd, 307 kd, 90 kd and 12 kd respectively based on the elution times of known molecular weight standards (See M~thods). It is important to note that a HMW protein appeared as a shoulder of the first peak (See arrow Figure 6). At first glance, the largest protein appeared to be peak 1 with a retention time of 15.30 minutes (1400 kd). However, closer scrutinization showed the presence of a shoulder, a potential1800 kd protein with an approximate R.T. of 14.00 minutes which appears to be eluting ahead of this foremost peak. Since the shoulder was not sensed by the integrator, a retention time and a value was not given for this component. The control and SS patient samples were then fractionated by Thiol-activated 48 Sepharose chromatography into thiol oxidized and thiol reduced fractions.. These fractions were then loaded onto the HPLC column for the separation of the proteins. Two distinctive peaks were observed (Figure 7); the first of these separating at a retention· time of about 14.00 minutes similar to the R.T. of the shoulder component: and the second peak having the same retention time as Peak 4. · Peaks 1, 2 and 3 are not clearly visible. FIGURES. Molecular sjeye HPLC separatjon of red cell membrane protejns of a normal and a sjck!e cell anemja sample. Untreated and unfractionated normal control (M) and patient (SS) sample membranes were heat solubilized (1 :1) for 4 minutes in a 2.0o/o SDS solution, then 100-150 J.Lg protein was applied to the Toya-Soda manufactured TSK SW-3000 and TSK SW-4000 HPLC molecular exclusion (sieve) columns in tandem. HPLC columns were equilibrated with the mobile phase containing 0.1 M Sodium Phosphate, 0.1°/o SDS, pH 7.0. Flowrate was fixed at 1mVmin; absorbance was measured at 280 nm; detector AUFS was 0.01.
Abbreviations:
HPLC - High pressure liquid chromatography SDS- Sodium dodecyl sulfate AUFS - Absorbance unit full scale 25
.. ·----~~---~ ·-··--·--- AA
E r:::: 0 co N ....ca \ I CD (,) .... •.;·;. r:::: ss ca .c... 0 f/1 .c '· Time (minutes) FIGURE7. Molecular Sjeye HPLC characterizatjoh of the thjol oxjdjzed membrane fractjOn. · Membranes from normal and Sickle cell patients were .separated using Thiol-activated 4B.Sepharose chromatography into thiol oxidized and thiol reduced proteins and then underwent the same processing as described in Figure 6. Abbreviations: · HPLC- High pressure liquid 9hrom.(itography 26 AA .. FF E C· , .... . ·r·. 0 \ cc ('I '\1' ...... }(, J 1\.. ctJ \ I CD u c ss ctJ .a FF . ~ 0 UJ 5.:· .a •"!;.:...... - < ..··t., . ·· ·.. · \, ·~ I '\ :.:·; j,J._/;-, \~ Time (minutes) 27 - - - These data confirmed that a large molecular weight protein (1800 kd) wa~- present in the thiol oxidized first filtrate fraction. - - - 3. Stydjes. pf Stractan G@djent fratjonatedqells: With~ four fraction gradient syste~ using 19°k, 22%, 24% and 28°k ~tractan, thre_e and four fractions were separated from normals and SS patients respectively (Figure 8). a. Hematology: Red cells from five normal (AA) ·controls and four (SS) . patient were fractionated by di~continuous Stractan gradient centrifugation met~od. Consistently it was shown that dense cens (fourth fraction) were present in sickle cen· patient samples and absent in normal controls (Figure 8 & 9). As expected MCV decreased and MCHC increased with increasing. cen density in both normal and SS cases (Figure 10). The change in MCV and MCHC in the 28°/o Stractan fraction (i.e. dense cells) were dramatic. These changes in MCV and MCHC further show the excenent quality of the stractan fractionation and cen separation because of the ·maintainence of the data linearity throughout the increasing density of the Stractan fractions. · Investigation of other hematological parameters (i.e percent irreversibly sickled cells, HbF and reticulocytes relative to the distribution of cens in the different . stractan fractions) provided additional data. Each stractan fractionated SS patient sample was analyzed for the percent presence of irreversibly sickled cens (ISC's). ISC's ar~ usually found to be dense cells but the converse is not necessarily true, that is, an dense cell are not always found to be ISC's (20). These ISC's were found mostly in the 28°k Stractan fractions (Figure 11 ). Stractan fra~tionated samples from SS . . . patients were also analyzed for the presence of fetal hemoglobin (HbF). Cens containing · the highest leve-l of HbF appeared to sepa.rate primarily into the 22°k Stractan fraction (Figure 12)-. The dense cell fraction contained the lowest level of HbF. This was expected from previous reports where Stractan or other types of gradients were utilized for cell fractionation, and likewise the enrichment of HbF in medium ~ensity cells was ·also shown in other. studies (12,50). Increased levels of reticulocytes were evident in the SS patient samples primarily in the 19°/o and 22°/o Stractan fractions~ (Figure 13). FIGURES. pjscontjnuous Stractan gradjent centrjfugatjon of red cells from a normal and a Sjckle cell patient sample. Normal control and a Sickle cell patient red cells were washed and resuspended in a nitrogenated BSKG buffer; and 1.0 ml of the _washed blood cells was layered onto the - gradient comprised of Stractan ranging from 19°k, 22°/o, 24% and 28°/o corresponding to densities 1.0785, 1.0919, 1.1 009 and 1.1195 respectively. The tubes. were centrifuged in an SW41Ti rotor (LS-65 Beckman ultracentrifuge) at 20,000 rpm (68,000 x g) for 50 minutes at 5°C. . Abbreviations: BSKG- Buffered Saline with Potassium and Glucose ) . 28 % Stractan AA ss 19 --> 19 --> 22--> 22--> 24--> 24--> 28--> '. J FIGURES.. percent djstrjbutjon of cells jn Stractan Eractjons jn M controls (n-5) and SS patients (n-5). The recovery of the cells in each Stractan fraction was calculated on the basis of the number of ABC's per fraction as determined by CBC analysis. The number of red cells in · each fraction relative to the number of red cells in all four fractions was expressed as a percentage. ·Cells · sedimenting at 28°/o Stractan (density = 1.1195) were · termed "dense cells". 24°/o Stractan n) FIGURE 10. Red cell jndjcjes (MCV & MCHC) of the Stractan gradient fractionated cell fractions from normal (AA) and sjckle cell anemja (SS) patients. The MCV and MCHC in each fraction was deterr:nined with a Sysmex CC-720 Hematology Analyzer. Error bars indicate standard deviation. Abbreviations: MCHC - Mean corpuscular hemoglobin concentration MCV - Mean corpuscular volume 3'0 100 ~ 80 tn .E 0 :I: 60 -m- MCVAA 0 ::iE -a- MCV SS(4) -o- MCHCAA olS 40 :;:: i .i i ... MCHC SS(4) .E > 0 20 ::iE 0 18 20 22 24 26 28 30 Fractions (Percent Stractan) ~·· FIGURE 11. Percent ISC's jn Stractan fractjons of SS patients. Percent irreversibly sickle cells (ISC's) was determined by 'analyzing blood smears of Stractan fractionated red cells. ISC oounts were made in non-overlapping fields until greater than 500 cells were tabulated. Cells were designated ISC _when their length was two times greater than their width. The number of ISC's relatiVe to the total number of cells was expressed as a percentage. Error bars indicate standard deviation. · ...... 31 . f' 50 40 fl) 0 30 ~ t: -CD Irm 1sc ss{4) I e 20 c.CD 10 0 19°/o 22o/o 24°/o 28o/o Fractions (Percent Stractan) FIGURE 12 .. percent fetal hemoglobjo fHbFl·jn the four Stractan fractjons of SS patjents. The percent Hb F was determined by a c~tion exchange microcolumn method. Error bars indicate- standard deviation. 32 50 40 LL .Jl 30 ::t: 1: -Q) lm Hb F ss<4> I e 20 a.Q) 10 0 19°/o Fractions (Percent Stractan) FIGURE.13. percent reticulocytes jn normal lAAl and sjckle cell lSSl Stractan fractjonatjons. The percent reticulocytes was determined by blood smear analysis as red cells were stained (1 :1) with 0.1 °/o methylene blue and counted in non-overlapping microscope fields until greater than 500 cells were tabulated. The number of reticulocytes relative to the total number of cells was expressed as a percentage. Error bars indicate standard deviation. Abbreviations: Retic - Reticulocytes 33 50 40 C1) G) ~ () 0 30 '5 () i • ReticAA a: 1J Retic SS(4) 20 1: -G) e G) D. 10 0 19°/o Fractions (Percent Stractan) The high presence of reticulocytes is not surprising because of the fact that these patients have various degree of hemolytic anemia. Since younger cells tend to separate into the less dense fractions, it is logical that the youngest cell, a reticulocyte, would be in the lighter fractions. These observations further confirm that cells are indeed being separated by their density and to some extent according to their age. b. Thjo! Oxjdatjon Oata: After undergoing Stractan density gradient fractionation, the red cells from the five normal control and four .SS patient samples were washed (to remove the presence of any residual Stractan); lysed, processed into membrane and chromatographed by Thiol-activated 48 Sepharose to determine the percent membrane thiol oxidation of red cells with respect to the age of the cell. Normal controls showed a linear increase in thiol oxidation from the lightest (19°k Stractan) to the densest fractions (24% Stractan) (Figure 14). (Remember, no cells partitioned into the 28o/o fraction in normal samples) Unpaired t-test analysis at a 95°k confidence level showed that the increase of thiol oxidation betWeen the 19°k and 24% fractions was significantly different, (P> .005)_. These results demonstrate an incr~ase in membrane thiol oxidation during ~ormaJ senescence of red cells since density and-biological age of the· red cell are believed· to be directly related except for ISC's and dense cells. Four SS samples also showed a linear increase in thiol oxidation from the lightest.(19o/o Stractan) to the.densest fractions (28°/o Stractan). Unpaired t-test analysis showed that the increase of thiol oxidation between the 19°k and 28°k fractions was also significantly different, ( p> .05), and that there was not a significant difference between adjacent Stractan fractions. Interestingly; dense cells had the highest level of thiol changes. Considering the corresponding fractions of normal and SS samples as pairs, SS samples always had significantly higher levels of thiol oxidation. c. HPLC Characterjzatjon: HPLC separation was performed to see if there was an increased pr~sence of HMW aggregate with increasing cell density (Figure-15). There appeared to be an increase in the size of the HMW shoulder relative to the density of the Stractan fraction especially in the SS pati_ent samples (see arrows). FIGURE 14. Percent membrane thjol oxjdatjon jn normal (AA) and sickle cell (SS) patjent samples after Stractan fractjonatjon. Normal and sickle cell membranes were prepared from each BSKG washed fraction of cells and later subjected. to Thiol-activated 48 Sepharose ·chromatography for separation and quantitation of thiol oxidized proteins. Abbreviations: BSKG -Buffered Saline with Potassium and Glucose 35 50 AA c: 40 0 ';:: m "tJ ";( 30 -a- thiol ox AA1 0 ..... thiol ox AA2 0 :E ... thiol ox AA3 ..... 20 .... thiol ox AA4 c: thiol ox AA5 -G) ...... (,) I 1 G) r- D. 10 0 18 20 22 24 26 28 30 Fraction (Percent Stractan) 50 c: 40 :;0 "tJ ";( 3o. 0 -a- thiol ox SS1 :2 -+- thiol·ox SS2 ..r:::: - thiol ox SS3 ..... 20 c: ..... thiol ox SS4 -G) e G) D. 10 04---~--~--~~~~--~~------~ 18 20 22 24 2'6 28 30 Fraction (Percent Stractan) FIGURE 15. HPLC molecular sjeye separatjon of Stractan fractjonated normal and Sickle cell patjent samples. , Normal (AA) and patient (SS) samples were separated by Stractan density gradient and further separated by molecular sieve HPLC as described in Figure 6; 36 ...... U) ...,Q) ::::1 c ~---- ...... E . CD ·-E .- (: :" l 1- ! c::::::: ~-··-···-· (: ~·· _____ ... ~..-----· <.___ ... --- ... -·-·-·~ en y i en . : i f ~: e:::::::- . . --~____--7 . . ~:-.·====------._ ··------::::::- -=::::.:::::.::.. - . --=- ----·---...... __ ---·----·· .;:;, ,-- Absorbance at 280 ·nm 37 4. Chemjcal Oxjdatjon Studjes: In yjtro chemical oxidation studies using diamide permitted the controlled oxidation of reduced thiols in the red cell membranes of normals and SS patients~ The purpose was to characterize the membrane thiol oxidation in normal and SS cells in response to- an _exogenous oxidant. , a. The tjme and djamjde concentratjon dependence of membrane thjol oxjdatjon: The time dependent changes in membrane thiol oxidation by a fixed . ' ' concentration (5.0 mM) of diarnide was perfomed .for both a normal control and an SS .~· ~ . . . patient: These data showed that membrane thiol oxidation increased with the time exposed to diamide in red cells. The concentration dependent changes were studie~ using a constant incubation time of 30 minutes; the millimolar concentration of the diamide were as follows: 1.25 mM, 2.5 mM, 5.0 mM, 10 mM, 20 mM and 40 mM. The results -o-i this experiment showed a diamide Concentration dependent increase in membrane thiol - - ' oxidation in normal and SS patient sa!flples. _Both sets of data showed a more dramatic -increase in the level of membrane thiol oxidation in the normal red cells versus red cells from the SS patient. This is most likely due to the fact that normal samples have increased presence of reduced ~hiols which are readily accessable to diamide oxidation whereas SS patient samples already contain a twofold increase in membrane thiol oxidation. However, the maximum lev-el of oxidation that could be achieved was _essentially the same for both types of membranes. b. SDS-PAGE characterjzatjon: Membranes prepared from the chemically-oxidized cells were then subjected to SDS-PAGE in the presence of B-mercaptoethanol (BME), a reducing gel, and also _in its absence, a .non reducing gel, so that the HMW thiol~oxidized disulfide-linked proteins could be identified~ In the presence of BME all of the usual bands migrated through the gel with very little of the protein aggregating in the· wells. The only distinctive diffe~ence between the individual samples of the time course was the increased intensity of the globin band (Figure 16). In the· absence of BME, the non-reducing gel, a decreased intensity in the spectrin bands - was_ seen because of the impossibility of oxidized spectrins (bands 1 & 2) as well as most of the remaining bands to enter the gel (Figure 17), prefering instead to - accumulate in the wells in -the form of disulfide linked aggregates (Figure 18). FIGURE 16. SDS-PAGE separatjon (reducjng condjtjons) of membrane protejns after djarOjde oxjdatjon for djfferent tjme intervals. A normal sample was washed and resuspended 1:1 0 in PBS and then treated with .5.0 mM diamide at 37°C . Time intervals were varied at 5, 10, 15, 30, 60 and 120 minutes. Red cells were washed x 3 in PBS to remove the oxidant and then separated under reducing conditions on a 10°/o · SDS-PAGE as described in Figure 5. Lanes 1 and 10 contain the molecular marker SDS-6H . Molecular weights of marker proteins are 205K, 116K, 97K, 66K, 45K and 29K. La·ne 2 contains a control sample at ambient temperature. · Lane 3 contains a control sample which went through the shaking (80 oscillations/min) at 37° c but'.received no chemical oxidant. Lanes 4 through 9 contain samples which were exposed to 5.0 mM diamide as follows: . · Lane 4- 5 min., Lane 5 - 10 min., Lane 6- 15 min., Lane 7- 30 min., Lane 8- 60 min., and Lane 9- 120 min. Abbreviations: Mol. Wt. - Molecular weight SDS-PAGE ~Sodium dodecyl sulfate polyacrylamide ·gel electrophoresis 38' LANES 1 3 . 4 5 6 7 8 9 10 ' molwt 205K ...... t· •t l I • ., f' , < , • 0~ ·' 116K 97K ,;. __. :~:< ..~.,.. ,.• ·_,"~~·~ "1 ·• .-~~• .. A ~iii 11 ',•1\; •• 'f?::l~:~ 66K -..... ' ;~.~'.:·._ ,·:~- ,_.... • /i. 45K .~~·~£~-~~· \ A"J ;.,"·'~~:,.;..- 1. . . : .: '' ~ •· 'f« .. 29K globin FIGURE 17. SQS-PAGE separatjon (oon-reducjog coodjtjoOSY of ;iz vitro oxjdjzed protejos after treatment with different concentratjoos of djamide. · .· . . A normal sample was washed and resuspended 1:1 o· in PBS and then treated with 1.25 mM~ 2.5. ·mM, 5.0 mM, 10 mM, · 20 mM and 40 mM diamide at 37~ C for 30 minutes. Red cells were ·washed x 3 in PBS to remove the oxidant and then separated on a 10°/o SDS~PAGE_,_ under non·-reducjng· conditions, where both the solubilizing an~ running buffers · · do not contain beta-mercaptoethanol. Except_for the pre.sen~e _of a non-reducing buffer, the procedure is identical to that described in Figure 5. Lanes 1 and 10 contain the mo_lecular marker SDS-6H. Lanes 2 through 7 contain sampl~s which were exposed to decreasing concentrations of diamide as follows: Lane 2 - 40 mM, Lane 3- 20 mM, .Lane 4- 19 mM, Lane 5- 5.0 mM, Lane 6 - 2.5 mM, ·and Lane. 7 - 1.25 mM diamide. Lane 8 contains a control sample which ~ent through the shaking · (80 oscillations /min) at 37° C but ·received no chemical oxidant. Lane 9 contains a control sample at ambient temperature. Abbreviations: _Mol. Wt. - Molecular weight SDS-PAGE- Sodium dodecyl sulfate polyacrylamide gel electrophoresis 39 LAI'{ES 1 2 3 4 5 6 7 8 9 .10 -~ ..o-- ---.~ ~ ---- rr==-""'l --...... ,- I moiWt 0 205K 116K 97K 66K -- 45K I: - I - .. -- ...... _J c-' --~ I 29K I L. <~-·-- __-_ - ____ ,_· _- __ ,._·~- ___ L-- ---~- ___-_",- ,__ !- globin FIGURE 18. ·Close up vjew of gel wells of an SQS-PAGE (non-reducjng condjtjons) containing an in vitro oxjdjzed normal sample treated wjth jncreasjng co'ncentratjons of diamjde. · Normal cells were chemically oxidiz.ed and SDS-PAGE was performed as described in Figure 17. Lanes 1 and 1o contain the molecular marker SDS-6H. Lanes 2 through 7 contain samples which were exposed to -decreasing concentrations of diamide as follows: Lane 2 - 40 mM, Lane 3-20 mM, Lane 4-10 mM, Lane 5-5.0 mM, Lane 6 ~ 2.5 mM, and Lane 7 - 1.25 mM diamide. Lane 8 contains a· control sample which went through the shaking (80 oseillations /min) at 37o· C but received no chemical oxidant. Lane· 9 contains a control sample at ambient temperature. Abbreviations: Mol. Wt. - Molecular weight SDS~PAGE -Sodium dodecyl sulfate polyacrylamide gel electrophoresis ' 40 LANES 9 10 1 2 3 4 5 6 7 8 205K 116K 97K 66K 45K molwt 41 Even after modification of the running gels by decreasing the amount of polyacrylamide to 7.5% to allow the proteins to enter the gel, aggregates of protein were still observed il1 the wells. Again it was seen that the limitations of the SDS-PAGE system for separating and characterizing proteins were restrictive to the gathering of meaningful data. It was concluded that further characterization could only proeeed by using molecular exclusion HPLc.· c. HPLC Characterjzatjon: Since HMW proteins were unable to become separated by SDS-PAGE, gel permeation HPLC techniques were employed. Five mM and 7.5 mM diamide oxidized AA samples were solubilized and loaded onto the HPLC column under non-reducing conditions,- that is, in the absence of thiol reducing chemicals such as BME (Figures19 & 20). Separation of these ·membranes showed a progr~ssive . increase in the size of.the.HMW shoulder (RT of about 14.00 as indicated by the arrow) in conjunction with a progressive decrea~e in the size and shape.of peak 1 (spectrin). If the HMW. prote.in was formed by cross-linked thiols, reducing conditions should eliminate it (Figure ·21). ·Indeed, these data showed the disappearance of the HMW shoulder in the presence of BME which confirmed the involvement of membrane thiols. FIGURE 19. Molecular sjeye HPLC separatjon of membrane protejns afterjn vjtro oxjdation of normal red cells wjth 7.5 roM djamjde for different tjme jntervals. A normal sarnple was washed and resuspended 1:10 in PBS ·then treated with 7.5 mM diamide at 37° C for 5, 10, 15, · 30 and 120 minutes. Red cells were washed, membranes prepared and HPLC separation proceeded as described in Figure 6. · Abbreviations: HPLC - High Pressure Liquid Chromatography 42 U) £: E U) •:•r··.-.. ·.~. c '-'."l:Jf,~ . ·- E ., 0 0,... N,.. .. , U) U) £: , 'j~. E·· .. -- .. -:: c ·- ,. ·- G) E I E E It) 0 ·-f- CD 0 U) ... c ...c ·- 0 E 0 0 M -.---....:.·-··--· .A' '• .__ ;~ Absorbance at 280 nm FIGURE 20. Molecular sjeye HPLC separation of membrane ptotejns after in vitro oxidation of normal red cells wjth · increasing concentrations of djamide. . .. A normal saniple underwent diamide oxidation (Figure 17). then~ HPLC separation proceeded as described in Figure 6. Abbreviations: HPLC - High Pressure Liquid Chromatography ··:, ... , . c 43 ______"'::) :E r •. E .. ·~·;'':.'~! E .'hi\- ...... i. ~ I ·:·;~ 1 ··yiJ·.··t r •• ...... tn cr. CD .: . ======-- ...::s _£-~ c: :;;.~ ·-··---·." . ·-~-· E. :E ...... E CD ln E N ,..... ·-t- 0 :s ... I .....a.. " ...... •' .... E :·~·~ :: / c: "· .... 0 0,.... 0 ...: Absorbance at 280 -nm · FIGURE 21. Influence of reductjon wjth Beta-mercaptoethanol on the molecular sieve HPLC separatjon of an in vitro oxidized . normal sample·. A normal sample underwent diamide oxidation as in Figure 17; then, HPLC separation proceeded as described in Figure 6. except that the membrane sample was solubilized in 2.0°k SDS and O.?M BME. The molecular sieve columns were equilibrated ·with a reducing mobile phase also containing O.?M BME. r 44 E ____ :E --- H/. r: ~: ·:::::===-::-;·.... ···--~- ·--·. . 0 ,.... \. '····- - :e ==E E Q) E i= --- :e ...0 ----···--·-··"'' ... f· ... ' ; E c L!. : 0 0 :.\ --=:::::::::=------~' ~----- Absorbance at 280 nm DISCUSSION Rank et al (2), using Thiol-activated 48 Seph~rose affinity chromatography, found_that sickle cell anemia patients .have higher percentage thiol oxidation in their red cell membranes than normal controls. To learn more about this phenomenon, a systematic study of membrane thiol oxidation in sickle ~II anemia and other related conditions was instituted .(Figure 1). In th~ present study we have correlated the level of membrane thiol oxidation occuririg ~to the level of reticulocytes, irreversibly sickled cells (ISC's) or dense cells, fetal hemoglobin (Hb F) and to the age of the patient. In addition, an attempt was made to identify. the. membrane polypeptides that are subjected to membrane thiol changes under ~:conditions and jn vjtro under the . influence of exogenous oxidants. ;,J " After the results of the preliminary baseline studies have shown ·th~t the increases in membrane thiol oxidation was indeeQ not the unintentional result or an artifact of the experimental procedure itself, work began by building a data base of normal (AA), sickle (SS) and other sickle and non sickle hemoglobinopathies (Tables lA, IB & II). The percent ofthiol oxidized proteins present in· red cell membranes prepared fro.m SS patients were twofold higher-than in the membranes of normal controls. A statisical analysis was performed to see how closely the oxidation values obtained in our laboratory using the same methodology and tec~niques paralleled .the values obtained in another laboratory (Rank et al)(2). A statistically significant difference was not found to exist between control values (p > 0.4) but was found to exist in the sickle ceU patient samples (p s 0.025) (present values are higher than those reported before). The closeness in the value of ·normal controls showed that the techniques employed were sound. The statistical level of difference in the sickle cell patient sa111ples is not of extreme concern because the variations in values could be due to non-experimental reasons such as hematological f~ctors (level of fetal hemoglobin. HbF, ISC's etc.) or age of the individual patients. 46 Since other hematological parameters were measured in addition to_ th~ level of thiol oxidation, relationships between various factors could be analyzed. The strong negative co_rrelation (r= -0. 768) of the HbF levels with the membrane tt:Jiol oxidation in the SS cases which were studied further underscores the beneficial effect of the presence of HbF in these cells. A similar relation~hip between Hb F levels and membrane oxidation was evident in an SS patient who was studied shortly after birth for a period of40 weeks (Figure 4). As expected, shortly after birth, this SS infant's HbF was very high but membrane thiol oxidation values were in the range of the normal controls. In the presence of a low level of sickle hemoglobin, percent membrane thiol oxidation values rernained·low. However, as the patient grew older, hemoglobin switching occured (exchange of sickle for fetal hemoglobin) and the level of membrane thiol oxidation rose. About the 29th week, sickle hemoglobin dominated and the patients ·percent membrane thiol oxidation leveled off at about 30°/o. The benefits of elevated fetal hemoglo~in prevents participation of HbF in polymerization (51), inhibits gelation (52) and helps against cellular sickling (53). The poor correlation (r= 0.161) between reticulocyte count and membrane thiol oxidation values and the fact that in a hemolytic disease like hereditary spherocytosis no increased !"embrane thiol changes were seen confirm that the increased number of thiol changes in sickle cell disease is not related to hemolytic anemia or reticulocytosis. Increased membrane thiol oxidation is not correlated with an increased presence of 1 • irreversibly sickled cells (ISC's) or dense cells. Thus the elevated levels of membr~ne thiol oxidation in SS cases does not appear to be due to the elevated levels of ISC's, although ISC fractions were found to have the highest level of thiol oxidation. The correlation between ISC's and thiol oxidation would have· been more significant if the presence of large population young cells (reticulocytes) having the lowest level of thiol . oxidation does not compensate for the effect of the other. Stractan density gradient centrifugation techniques can be employed to· separate cells according to the density and thereby separating the cells according to their age. In St.ractan fractions, normal cells and sickle cells containing a twofold difference in thiol oxidation can separate into the same density fraction which means percent thiol oxidation does not directly determine 47 . cell density. However, once increased membrane thiol oxidation exists, a cascade of events can be triggered ultimately leading to the formation of dense cells or ISC's (Figure 22). A recent study on jn vjtro membrane thiol oxidation by Quina et al (54) has shown that more den~e _cells were formed in a· buffer containing CaCI2 versus EGTA. · Furthermore, the production of dense cells were reversed by the reduction of membrane disulfides using trace amounts. of mercaptoethanol. It is highly suggestive that oxidation of thiols of membrane-bound Ca-ATPase may contribute to dense cell generation by causing an accumulation of Ca2+, loss of K+ and the concurrent loss of cellular water leading to cellular dehydration and dense cell formation (Figure 22). The age of the cell · again seems to be a determining factor. . Retic~locytes have the lowest while senescent and dense cells have the highest leyel of'thiol oxidation. SDS-PAGE separation of normal control· and sickle cell patient whole membranes ·shows distinctive changes in sickle cell samples over normal controls (48,55) (Figure 5): a decrease in the intensity of Band 4.1 a (88kd) and band 7 (32 kd), an increa~e in the intensity ~f globin band suggesting that more hemoglobin has bound to the membrane, and the appearance of a new 22 kd band. Band 4.1 acts as part _of the Spectrin-actin- 4.1 interaction. Such a defect in band 4.1 has be_en reported before in SS patients (49) indicating that a reduced presence can bring about membrane disorders (56). The separation of the membrane _proteins with Thiol-activated 4B Sepharose followed by SDS-PAGE showed the presence of Band 7 and a 22 kd band in the thiol oxidized fraction. The absence of spectrin bands 1. & 2 was in direct contradiction to the findings of Rank et al· (2) who found these bands in the oxidized fraction but supports the findings of Chiu et al (14). Morever, in the oxidized fraction of stored blood undergoing microvesiculation studies, Wagner et al (21) found an absence of spectrin which supported our findings. High Pressure Liquid Chromatog~aphy (HPLC) using amolecular sieve column was then considered as the the best possible choice for the separation of potential large molecular weight proteins. To this point the only HPLC methodology employed was that . . - . using reversed phase columns for the separation of erythrocyte membrane proteins (57). FIGURE 22. Scheme depjcijng the probable role of thjol oxjdatjon jn dense cell formatjon .. 48 ·Sickle Red Cells Ca - ATPase Inhibition I ba Accumulation lon Channel Modification , Membrane Tliiol Oxidation I . ,: Altered 9ation P~rmeabiliiy I K+ Leak · j Cell Dehydration De.nse Cell Formation 49 This method was not appealin·g because it did not provide separation of proteins-by molecuiar weight. There was an abundance of current l_iterature surfaced showing HPLC methodology using molecular e_xclusion (sieve) chromatography (38-4 7) but nothing showing the separation of erythrocyte membrane proteins. -The molecular sieve_ HPLC methodology previously established by Perry _and Abraham (58) for separation of disulfide linked lens crystallin aggregates was adopted for membrane proteins. Gel permeation HPLC added the dimension of allowing proteins as large as 7 X 1o6 daltons to be separated and when erythrocyte membranes were loaded (Figure 6); five distinct - . peaks eluting at retention times (R.T.) 15.30,-17.60, 21.00, 25.90 and 33.50 minutes which correlates with the molecular weights 1400 kd, 750 kd, 307 ·kd, 90 kd _ and 12 kd respectively. SDS-PAGE characterization showed peak 1 to contain spectrin ' ' _and ankyrin ; Peak 2 was seen to possess band 3 and 4.1 and 4.2_; and Peak 3 was seen to contain band 5, 6 and 7. The 1800 kd protein with R.T. of 14.00 minutes was not obtained in sufficient quantites for SDS-PAGE characte~ization. It is known that spectrin subunits have a molecular weight of approximately 240 kd and 220 kd totaling approximately 460 kd for a spectrin molecule (monomer). Spectrins have been seen to exist in dimers (920 kd) as well. It is possible that peak 1 (1400 kd) is the spectrin ~imer associated with other proteins. This is logical since th_i~ peak is not seen to contain-pure spectrin proteins. Peak 2 (750 kd) probably contains spec~rin monomers (460 kd) associated with 4.1 (88 kd) and 4.2 (80 kd). Band 3 protein (118 kd) is associated in some manner with the monomer. After fractionation by Thiol-activated 48 Sepharose chromatography, HPLC separation of the proteins was performed (Fig!Jre _7) which shoWed two distinctive peaks; the first separating at a retention time of about14.00 minutes (1800 kd): and the second peak_ having the same retention time as Peak 4 (307 kd). · T~ese data confirm a that a large molecular weight protein (1800 kd) is present in t~e thiol oxidized was to find out whether normal and sickle cells responded similarly to an exogenous oxidant and whether the same proteins were involved in ~ and jn vitro thiol oxidation. HPLC gel permeation chromatography was the preferred method of protein characterization for reasons already explai~ed. Diamide oxidation of human eythrocytes have been pei1omed with success (22,23,32). It was seen that in both a time dependent and concentration dependent study that thiol oxidation values increased during jn vitro ,oxidation. Percent membrane thiol oxidized did not exceed 50°k even at higher concentrations of the oxidant (e.g. 40mm diamide). When loaded on SDS-PAGE or HPLC, if the protein migration patterns of the jn vitro oxidized cells resembled the diseased (SS) cell then it might be concluded that identical proteins are involved. However, the SDS-PAGE was not able to even show the oxidized differences due to its molecular weight limitation. Nevertheless, the HPLC was able to show that a HMW protein not only exists in normal controls as evidenced by a HMW component but that this was more evident in SS patients. This compc)nent also appears larger especially in the Stractan fractionated SS patient samples··(Figure 15). This HMW component which was seen migr~ting before peak 1 is most likely spectrin since it increased during jn vjtro diamide oxidation, a known oxidi;zer of spectrin, at the expense of peak 1 and peak 2 both ofwhich have been seen to contain spectrin. Furthermore, the HMW .component was seen to disappear in reducing buffer attesting to the fact that it is indeed a sulfhydyl containing protein (Figure 21 ). · An interestin,g ph~nomenon occuring in this oxidation study was the oxidative denaturation of the red cell membranes which suggested that hemoglobins were being attached to the membranes. The SDS-PAGE was able_ to show increase in the binding of globins both in sickle cells and in normal cells treated with increasing concentrations of a chemical oxidant (Figures 16 & 17). The Increase in the intensity of the globin band showed proportionality to length of incubation time or concentration of diamide.· Palek et al (59) showed that the anti-oxidant system is AlP-dependent and the absence of ATP leads to oxidative denaturation~ Indeed a high molecular weight aggregate (HMW) appeared as a product of denaturation. These denaturation studies on modified normal hemoglobins were informative but didn't provide insights into disease cell 51 pathophysiology, in particular, to those cells containing unstable hemoglobin. Jacob and . Winterhalter (60,61) investigated Heinz body formation in people with unstable hemoglobins. Some of the hemoglobinopathies they looked at were Hemoglobin ·Hammersmit~ (B42 Phe--~Ser), Hemoglobin KOin (Bas.vai--~Met) ~nd Hemoglobin Zurich (B63 His---Arg) because of th~irpotential.influence upon the heme pocket. They found that unstable hemoglobi~s display an increased loss of the heme structure due to imperfections in attachment of t~e h~me to the pocket thereby increasing· the capability of the molecule to undergo denaturative processes terminating iri Heinz body formation. ·Is it possible that increased membrane thiol oxidation is not only related to . ir;ltramembrane protein crosslinkings but also to thiollinkages with the oxidatively denatured hemoglobins referred to as hemichromes? If this is indeed the case, then, the . . .disulfide linked ~emichromes would bring into juxtaposition the iron cation which would promote oxygen radical formation (Haber-Weissreaction) leading to more thiol oxidations and lipid peroxidations (1 ). Also, the membrane binding of hemich.romes would promote the clustering of band 3 protein which has been shown in· recent literature to be closely associated to the cell senescent antigen (9). It is known that the homeostasis of the ~irculatory system is dependent upon the removal o_f aged and damaged cells from the blood (7). Autologous a~tibodies are known to bind to, the senescent cells which initiates the removal process by macrophages and phagocytes (8). Since old and new cells alike contain· band 3, what is the relationship between hemichromes, band 3 and the aging process? The following hypotheses aro_se: Some researchers felt that band 3 protein underwent some kind of cleavage or covalent modification upon aging making the senescent cell different (9). Others adhered toa_cryptic site theory which proposed . . that a cryptic antigenic site is exposed upon cell aging. Lastly _there were those who felt · that somehow a cha.nge in the ·lateral distribution of band 3. brought about the indication of cellular senesce-nce_ (1 0). This last idea was the thinking of Low et al (62) when they decided to investigate the relationship between band 3 bound phenylhydrazine denatured · hemichromes and autologous lgG's. They found that there was increased lgG binding to phenylhydrazine or acradine orange denatured red cells. SchlUter and Drenckhahn (63) 52 found similar results while working with patients having unstable hemoglobins. Using immunofluorescent techniques; they found that lgG binding always occured in association with the presence of Heinz bodies). Their conclusion was that the denaturation of hemoglobin leads to hemichrome formation; hemichrome copolymerization; then ultimately, Heinz body formation. The clustering of the band 3 protein occurs as a resuh of aggregation during Heinz body formation. Once the band 3 protein had been clustered a new antigen, which can be recognized by the autologous antibodies labels the cell as a senescent cell. Phagocytes then remove the cell from circulation. This concept still needs further experimentation but is a definite front runner in explaining hemichrome involvement in cellular senescence. Many studies, including this one, appear to support the idea that increased membrane thiol oxidation is critically involved in the eventual demize of the cell . (8,9, 10,60,61 ,62). This involvement may be due to increased cell density, increased hemoglobin binding leading to auto-oxidation and/or band 3 clusterings advancing cellular senescence. SUMMARY Sickle Cells are known to be more susceptible to oxidant stress (1 ). Using a ' ' systematic approach to study. thiol oxidation in sickle cell anemia, the present study - - ·focused on the effect(s) that dense cells, age of the cell and certain hematological parameters such as reticulocytes and fetal hemoglobi·n (HbF) have on membrane thiol oxidation. Thiol-activated 4B Sepharose macr~columils and microcolumns were used for this chromatographic method (36). Baseline studies such as the influence of sample preparation and storage conditions on oxidized thiol-values had to be performed. Blood samples from normal controls (AA), sickle cell patient (SS) were collected and chromatographed to determine the percent membrane thiol oxidation. Erythrocyte membrane t~iols in normal AA individuals were found to be susceptible to some degree of oxidation 13.8% ± 1. 7. SS samples had almost a twofold increase in the mean percent membrane thiol oxidation value of 25.5% ± 5.0. _Correlational studies with SS patient samples (n=16) showed two poor correlations (r= 0.161 ), (r= 0.163) and one strong negative correlation (r= -0. 768) between the level of membrane thiol oxidation and the levels of reticulocytes, ISC's and HbF respectively. Characterization of thiol oxidized and thiol reduced membrane fractions was accomplished by separation of Proteins by SDS-PAGE according to the procedure described by Laemmli (38) and by Molecul~r Sieve High Pressure Liquid Chromatography (HPLC) using a Beckman HPLC System. -Discontinuous Stractan gradients were prepared by decreasing the number of Stractan fractions used by Clark et al (30) who modified the work of Corash et al (31). Using a four fraction gradient system using 19°k, 22°/o, 24°/o and 28o/o Stractan three and four fractions were separated from normals and SS patients respectively. After _undergoing Stractan density gradient fractionation samples were processed into membrane and chromatographed by Thiol-activatedAB Sepharose to determine the percent membrane thiol oxidation of red cells with respect to the age of the cell. Normal controls showed a linear increase in thiol oxidation from the lightest to the densest 54 fractions. These results demonstrate an increase in membrane thiol oxidation during normal senescence of red cells. Four SS samples also showed a linear increase. HPLC separation was performed to see if there was an increased presence of HMW aggregate with increasing cell density and there appeared to be an ·increase in the size of the HMW shoulder relative to the density of the Stractan fraction especially in the SS patient samples. In vjtro chemical oxidation using diamide [(CH3)2NCON=NCON(CH3)2] permitted the controlled oxidation of reduced thiols in the red cell membranes of normal and SS patients. The purpose was to characterize. the membrane thiol oxidation in normal and SS cells in response to an exogenous oxidant. Membranes prepared from the chemically-oxidized cells were then subjected to SDS-PAGE in the presence of. B-mercaptoethanol (BME), a reducing gel, and al~o in its absence, a non reducing gel, so that the HMW thiol-oxidized disulfide-linked proteins could be identified. Since HMW proteins were unable to become separated by SDS-PAG.E, gel permeation HPLC techniques were employed. REFERENCES OF UTERATURE CITED 1. Hebbel R.P. (1986) Erythrocytes antioxidants and membrane vulnerability .. J. Lab. Clin. Med. ~ 401. 2. - Rank, B.H., Catlsson, J. ~nd Hebbel, R.P. (1985) Abnormal redox status of membrane-protein tt1iols in sickle erythrocytes. J. Clin. Invest. ~ 1531. 3. Eaton, J.W., Skelton, T.D., Swofford, H.S., Koplin, C.E. and Jacobs, H.S. (1973) Elevated erythroc~e calcium in sickle cell disease. Nature ~ 105. 4. Erickson, B.N~, Williams, H.H. and Hummel, F. C. (1937) The lipid and mineral distribution of the serum and erythrocytes in tlie hemolytic and hypochromic anemias of childhood. J. Bioi. Chem.!JJt 567. 5. Hellerstein, S. and Bunthrarungraj, T. (1974) Erythrocyte composition in sickle-cell anemia. J. Lab. Clin. Med . .aa: 61.1. 6. Palek, J~ (1977) Red cell. calcium content and transmembrane calcium movements in sickle_cell anemia. J. Lab. Clin. Med. Ba: 1365. 7. Mollison, P.L.·. (1983) _Blood Transfusion in Clinical Medicine. Blackwell,· Oxford, 7th edition, p. ·198. 8. K~y, M.M.a. (1975) Mechanism of removal of senescent cells by human macrophages .irlsiW. Proc. Natl. Acad. Set ~ 3521. 56 9. Kay, M.M.B., Sorenson, K., Wong, P. and Bolton, P. (1982) Antigenicity, storage and aging: Physiologic autoantibodies to cell membrane and serurri proteins and the senescent cell antigen. Mol. Cell Bio,chemistry. ~65. 10. Walder, J.A.,_ Chatterjee, R., Steck, T. L., Low, P.S., Musco, G.F., Kaiser, E.T., Rogers, P.H. and Arnone, A. (19~4) Journal Bioi. Chern. ~1 0238.. 11. Embury, S.H., Clark,.M.R., Monroy, G. and Mohandas, N. (1984) Concurrent sickle cell anemia and alpha-thalassemia: Effect on pathological properties of sickle erythrocy·tes. J. Clin. Invest.. ~ '116. 12. Noguchi, C.T., Dover, G.J., Rodgers, G.P., Serjeant, G.R., Antonarakis, S.E., Angnou, N.P., Higgs, D.R., Weatherall, D.J. and Schechter, A.N. (1985) Alpha Thalassemia changes e'rythrocyte heterogeneity in sickle cell disease. J. Clin. Invest. ~ 1632. 13. Hebbel, R.P., Eaton, J.W., Balasingam, M. and Steinberg, M.H. (1982) Spontaneous oxygen radical generation by sickle erythrocytes. J. Clin.lnvest. ~ 1253. 14. Chiu, D. and Lubin, B. (1979.) Abnc:>rmal vitamin E and glutathione· peroxidase levels in sickle cell anemia. J ~ lab. Clin. Med. ~ 542. 15. Das, S.K. and Nair, R.C. (1980) Superoxide dismutase, glutathione peroxidase,· catalase ~nd lipid p~roxidati<:m of normal and sickled erythrocytes. Br. J. Haematol.-~: 87. · . 16. Schacter, L.P., DeiVillano, B.C., Gordon, E.M. and Klein, B.L. (1985) Red cell superoxide dismutase and sickle cell anemia symptom severity. Am~ J. Hematol. ~ 137.. 57 . . 17. Varma, R-.N., Mankad, V.N., Ph~elps, D.O., Jenkins, LD. and Suskind, R.M. (1983) Depressed erythrocyte glutathione reductas~ activity in sickle cell disease. Am. J. Clin. Nutrit. 3a.;, 884. 1a. Wett~rstroem, N., Brew~r, G.J_., /Warth, J.A., Mitchinson, A. and Near, K. (1984) Relationship of glutathio-ne levels and heinz body formation to irreversibly sickled cells in sickle cell anemia. J. _Lab. Clin. Med . .1.Q3.;. 589. ·19. Zimmerman, C.P. ~nd· Natta, C. (1981) Glutathiorye,peroxidase activity in whole blood of patients with sickle cell anaemia~ Scand. J... Haematol. ~ 177. 20. Beretta, L., Gerli, G.C., Ferraresi, R., Agostoni, A., Gualandri, V. and Orsini, G.B. (1983) Anti~xidant system in sickle red cells. Acta Haematol. ZQ;,-194 .. 21. Wagner, G.M., Chiu, D.T.Y., Qju, J.H., Heath, R.H. and Lubin, B. H. (1987) Spectrin oxidation correlates with membrane vesiculation in stored · RBC's. Blood~ 1777. 22. Becker, P.S., Cohen, C.M., and Lux, s~E. (1985) The effect of mild ~ diamide oxidation on the structure and function of human'erythrocyte · spectrin. J. of Bioi Chern ~ 4620. 23. Deuticke, B.,. Poser, a., LOtkemeier P. and Haest C.W.M .. (1983) Formation of aqueous pores in the human erythr~qyte membrane after oxidative cross linking of spectrin by diamide~ Biochemica and B!ophysica_Acta ~ 196. _ 24. Bartles, J. F. and Milner P. F. A. (1968) ·Irreversibly sickled erythrocytes: A consequence of the heterogeneous distribution of hemoglobin types in sickle cell anemia. J Clin Invest. £ 1731. 58 . 25. Daci~, J. V. and Lewis, S.M. (1975) The Reticulocyte Count. In: Practical Hematology, p. 30 (Grune and Straton) New York. 26. Danon, D. and Marikovsky, Y. (1964) Determination of density distribution of red cell population. J Lab Clin Med. M.; 668. 27. Rodgers, G.P., Schechter, A.N., and ,Noguchi, C.T. (1985) Cell heterogeneity in sickle cell disease: Quantitation of the erythrocyte density profile. J. Lab. Clin. Med. ~ 30 .. 28. Weems, H.B. and Lessin, L.S. (1984) Erythrocyte density distribution in sickle cell anemia. Acta Haematol. 11.: 361. 29. Abraham, E.C., Carver, J.,· DObler, J., Milner~ P.F.·, and Huisman, T.H.J. (1979) Microchromatographic quantitation of fetal hemoglobin in · patients with s.ickle cell disease. Hemoglobin ~ 341. 30. Clark, M.A., Greenquist, A.C. and Shohet, S.B. (1976) Stabilization of the shape of. sickled cells by calcium and A23187. Blood ~ 899. 31. Corash, L.M., Piomelli, S., Chen, H.C., Seaman, C.· and Gross, E. (1979) Separation of erythrocytes according to age on a simplified density gradient. J. Lab. Clin. Mad. ~ 147. - 32. Kosower, N.S., Kosower, · E.M., Wertheim B. (1969) Diamide, a new reagent for the intracellular oxidation of glutathione to the disulfide. Biochemic~l and Biophysical Research Communications ~ 5.93. 33. Dpdge, J.Tw Mitchell, C. and H~nahan, D.J~ (1963) The preparation and chemical characteristics of hemoglobin.;free ghosts of human erythrocytes. Arch. Biochem. ,Biophys. ~ 119. 59 34. - Smith, P.K., Krohn, R.I., Hermanson, G.T., Mallia, A.K., Gartner, F.H., Provenzano, M.D., Fujimoto, E.K., Goeke,_ N.M.,. Olson, B.J. and Klenk, D.C. (1985) Measurement of protein using Bicinchoninic acid. Analyt. Biochem. 15Q: 76. 35. Lowry, a·.H.,. Rosebrough, N.J., .Farr, A.L. and Randall, R.J. (1951) Protein measurement with the folin phenol reagent. J. Bioi Chern. ~ 265-279 . . 36. Egorov, T.A., Svenson, A., Ryden, L. and Carlsson, J.. (1975) .A rapid and specific method for isolation of thiol-containing peptides from large prot~ins by thiol~disulfide exchange on a solid suppOrt. Proc. Natl. Acad. Sci. USA ~ 3029. 37. Laemmli, U.K. (1970), Cleavage of_structural proteins during the assembly of the head of bacteriophage T4. Nature . .22Z: 680. 38. · Fukano, K., Komiya, K., Sasaki, H. and Hashimoto, T. (1978) Evaluation of new supports _for high pressure aqueous gel permeation chromatography TSK-GEL SW type columns. J. of Chromatogr.·~ 47. 39. Hashimoto, T., Sasaki, H., Makoto, A. and Kato, Y. (1978) High speed aqueous gel-permeation chromatography. J. of Polymer Science: Polymer Physics edition .1.6..;. 1789. 40. Hashimoto, T., Sasaki, H., Makoto, A. and Kato, Y. (1978) High speed aqueous gel-permeation chromatography of proteins. J. of Chromatogr . .wl;. 3_01. 41. Kato, Y., ·and Hashimoto, T. (1983) Solute support interactions in high performance gel filtration of proteins in TSK-GEL SW Type column. J. of High Resolutiqn Chromatography Communications ~ 4. 60 42. Kato, Y., Komiya, K., Sasaki, H. and Hashimoto, T. (1980) Comparison of TSK-GEL PW type and SW type in high speed aqueous gel-permeation chromatography. J. of Chromatogr. ~ 311. 43. Kato, Y., Komiya, K., Sasaki, H. and Hashimoto, T. (1980) High speed gel filtration of proteins in 6M guanadine hydrochloride TSK-GEL SW columns. J. of Chromatogr.laa: 458. 44. Kato, V., Komiya, K., Sasaki, H. and Hashimot~, T. (1980) Separation range and separation efficiency in high speed gel filtration on TSK-GE~ SW columns. J. of Chromatogr. ~ 297. 45. Kato, V., Kido, S., Yamamoto, M., Hashimoto,T. (1974) High speed gel-permeation chromatography. J. of Polymer Science: Polym$r Physics edition ~ 1339. 46. Kato, Y., Sasaki, H., Makoto, A. and Hashimoto, T. (1978) High performance aqueous gel permeation chromatography of oligomers. J. of Chromatogr. .!5a: 546. 47. Welling, G.W., Slopsema, K., Welling-Wester, S. (1986) Size-exclusion high-performance liquid _chromatography of sendai virus membrane proteins in different deterge'nts: A comparison of different ~olumns. J. of Chromatogr. ~ 307. 48~ Ans~lstetter, ·V. (1978) Gel electrophoresis of the human erythrocyte membrane proteins: Aberrant Patterns in Hematological and Non-hematological diseases. Blut ~ 144. 49. Riggs, M~G. and Ingram, V.M~ (1977) Differ~nces in erythrocyte membrane proteins and glycoproteins in sickle cell diseas~~ Biochem . . Biophys. Res. Comrriun. ~ 191. 61 50. ~braham, E.C., Walker, D., Gravely, M. and Huisman, T.H.J. (1975) Minor hemoglobins in sickle cell anemia beta-thalassemias and related conditions: A study of red cell fractions isolated by density gradient . centrifugation. Biochemical Medicine ~ 56. 51. Nagel R.L., Bookchin A.M., Johnson J., Labia, D., Wajcman, H., lssac-Sodeye, W.A., Honig, G. A., Schiliro, G., Crooksto_n, J.H. and Matsumoto, K. (1979) ·structural bases of the inhibitory effects of Hemoglobi~ F and hemoglobin A2 on the polymerization of hemoglobin S. Proc. Natl. Acad. Sci. USA 16.; 670. 52. Bookchin A.M. and Nagel R.L. (1974) Interactions between human hemoglobins: sickling and related phenomena. Semin Hematol. ~ 577 . . 53. . Sewcha~d, L.S., Johnson C.S. and Meiselman, H.J. (1983) The effect of fetal hemogl9bin on the sickling dynamics of SS e,Ythrocytes. Blood Cells~ 147. 54. Quina, W.R., Mankad, V.N. and Moore, A.a. (1987) The patential role of sulfhydryl oxidation in the formation of irreversibly sickled cells. Federal Proceedings ~ 2197a. 55. Fairbanks, G., Steck T. L., Wallach, O.F.H. (1971) Electrophoretic Analysis of the Major Polypeptides of the Human Erythrocyte Membrane. Biochemistry .1.0.: 2606. 56. Palek, J and Lux, S. E. (1983) Red cell membrane skeltetal defects in heriditary and squired hemolytic anemia.· Seminars in Hematology .2.Q;, 189. 57. Goheen, S~_C. and Chow, T.M. (1986) Reversed phase high performance liquid chromatography of red blood cell membranes. J. of Chromatogr. ~297. 62 58. Perry R.E. and Abraham, E.C. (1986) High-performance liquid chromatographic separation of lens crystallins and their subunits. J. Chromatogr . .351.:1 03. 59. Palek, J., Liu, S.C. and Snyder M .. (1978) Metabolic dependence of _ protein arrangement in human erythrocyte membranes I. Analysis of spectrin-rich c6mplexes in ATP-depleted red cells. Blood 5..1!3l: 385. 60. Jacob, H. and Winterhalter, K. ·(1970) The role of hemoglobin heme loss in Heinz body formation: study with a partially heme deficient . . . hemoglobin and with genetically unstable hemoglobins. Journal Clinical Investigations -~ 2008. 61. Jacob, H. and Winterhalter, K. (1970) The role of heme loss in Heinz body formation. Proc. Natl. Acad.-Sci. USA ~ 697. 62. Low, P.S., Waugh, S.M., Zinke K. and Drenckhahn, D. (1985) The role of hemoglobin dena~uration and band 3 clustering in red blood cell aging. Science ~ 533. 63. Schluter, K and Drenckhahn, D. (1986) Co-clustering of denatured hemoglobin with band 3: Its role in binding of aut~antibodies a~ainst band 3 to abnormal and aged erythroctes. Proc. Natl. Acad. Sci. USA 83(16): 6137.