Download Isolation, Characterization, and Immunoprecipitation

Isolation, Characterization, and Immunoprecipitation Studies of Immune
Complexes From Membranes of f3-Thalassemic Erythrocytes
By Jie Yuan, Rama Kannan, Eilat Shinar, Eliezer A. Rachmilewitz, and Philip S.Low
p-Thalassemia, a hemoglobinopathythat results in the precipitation of denatured *globin chains on the membrane, is
characterized by erythrocytes with significantly reduced
lifespans. We have demonstrated previously that hemoglobin denaturation on the membrane can promote clustering of
integral membrane proteins, and that this clustering in turn
leads t o autologous antibody binding, complement fixation,
and rapid removal of the cell by macrophages. To evaluate
whether this pathway also occurs in pthalassemic cells, we
have isolated and characterizedthe immune complexes from
the membranes of these cells. We observe that autologous
IgG-containingcomplexes obtained by either immunoprecipitation or simple centrifugation of nondenaturing detergent
extracts of pthalassemic cell membranes contain globin,
band 3,IgG, and complement as major components. Absorp-
tion spectra of these complexes demonstrate that the globin
is, indeed, mainly in the form of hemichromes. Immunoblotting studies further show that much of the band 3 protein in
the aggregates is covalently cross-linked t o a dimeric or
tetrameric form, consistentwith the preferenceof the autologous IgG for clustered band 3. Although the insoluble aggregates constitute only -1.6% of the total membrane protein,
they still contain 27% of the total IgG and 35% of the total
complement C3 on the thalassemic cell surface. Because cell
surface IgG and complement component C3 are thought t o
trigger removal of erythrocytes from circulation, the hemichrome-induced clustering of band 3 may contribute t o the
p-thalassemiccell’s shortened lifespan.
o 1992by The American Society of Hematology.
A
MATERIALS AND METHODS
LTHOUGH P-thalassemic erythrocytes are known to
have shorter lifespans than normal
the molecThalassemic erythrocyteswere obtained from untransfused spleular basis of their accelerated clearance has never been
nectomized and nonsplenectomized patients with @-thalassemia
intermedia from Kurdish Jewish and Arabic extraction. The
determined. However, in the process of investigating noridentification of their genetic mutations has been reported elsemal and sickle cell clearance, we have formulated a hypothwhere.16These samples as well as control erythrocytesfrom healthy
esis that explains how the lifespan of a cell can be
donors were transported on ice for 1 to 2 days before use.
determined by the integrity of its hemoglobinT6 Because
IODO-BEADS and protein A beads were obtained from Pierce
thalassemic cells are characterized by the early precipitaChemical Co (Rockford, IL) and Na’=I/NaOH was purchased
tion of denatured globin chains on the membrane as well as
from Amersham Corp (Arlington Heights, IL). Octaethylene glycol
in the c y t ~ p l a s m ,we
~ - ~felt the premature demise of the
mono-n-dodecylether (C12E8) was purchased from Nikko Chemithalassemic cell might arise in part from the mechanism we
cal, Ltd (Tokyo, Japan) and nitrocellulose membranes were from
have proposed.
Schieicher and Schuell (Keene, NH). Antihuman C3c goat IgG
Briefly, the clearance mechanism hypothesizes that shortly
(“cyto grade,” affinity purified, polyclonal) was obtained from The
Binding Site (San Diego, CA) and affinity-purifiedantihuman IgG
before an erythrocyte’s removal by macrophages, hemoglowas from Miles Biochemical (Elkhart, IN). Polyclonalantibodies to
bin begins to denature within the cell’s cytoplasm.6Because
human hemoglobin and erythrocyte membrane proteins were
the resultant hemichromes exhibit a high affinity for the
cytoplasmic domain of the membrane protein band 3,loJ1 raised in rabbits and affinity purified against their respective
antigens. All other reagents were purchased from Sigma (St Louis,
pulling the associated copies of the anion transporter
MO), Bio-Rad (Richmond, CA), or Boehringer Mannheim (Indiatogether into localized clusters in the membrane,3-6J1J2 napolis, IN) and were the highest purity available.
hemoglobin denaturation, in effect, induces microscopic
Isolation of hemichrome-rich membrane protein aggregates from
changes in the external topography of the ce11.12 These
Pthalassemic erythrocytes. Erythrocyte membrane proteins were
microscopic clusters, which commonly contain less than 1%
prepared from @-thalassemicand normal cells according to the
procedure of Dodge et all7 by lysis at 4°C in 5 mmol/L sodium
of the band 3 population of the ce11,5S6 are then rapidly
1mmol/L EDTA, pH 8.0 in the presence of phenylmeopsonized with autologous IgG and ~ o m p l e m e n t . ~ . ~ , ~ Jphosphate,
~J~
thylsulfonyl fluoride (PMSF) (20 p,g/mL final). The membranes
Finally, the deposition of IgG and complement at one or
more of these clustered sites triggers the recognition and
removal of the cell by macrophages.13-15
Although substantial evidence exists for the participation
From the Department of Chemishy, Purdue Universiiy, West
of the proposed clearance mechanism in sickle cells and
Lafayette, IN; and the Department of Hematology, Hadassah Medical
senescent normal ~ e l l s , 3 - ~ ,the
” ~ ~involvement of hemiCenter, Jerusalem, Israel.
Submitted October 31, 1991; accepted January 31, 1992.
chrome-induced integral protein aggregation in thalassemic
Supported by National Institutes of Health Grant GM24417.
cell removal has never been evaluated. In this study, we
Address reprint requests to Philip S. Low, PhD, Department of
address this question by investigating the occurrence and
Chemishy,
Purdue Universiv, West Lafayette, IN 47907.
composition of hemichrome-rich membrane protein aggreThe publication costs of this article were defrayed in part by page
gates in P-thalassemic cells. We report that such aggregates
charge payment. This article must therefore be hereby marked
do exist, and that they contain elevated amounts of band 3,
“advertisement” in accordance with 18 U.S.C. section I734 solely to
autologous IgG, and complement. We suggest that these
indicate this fact.
membrane protein clusters may contribute to the premaQ 1992 by The American Socieiy of Hematology.
ture clearance of P-thalassemic cells from circulation.
0006-4971/92/7911-0013$3.00/0
Blood, Vol79, No 11 (June 1). 1992: pp 3007-3013
3007
YUAN ET AL
3008
were subsequentlydepleted of spectrin and actin and converted to
inside-out vesicles (IOVs) by incubation at 37°C for 30 minutes in
40vol of 0.3 mmol/L EDTA, 0.2 mmol/L dithiothreitol, 20 kg/mL
PMSF, pH 8.0.l8 The resulting vesicles were collected by centrifugation at 17,ooOg. The protein content of the various IOV fractions
was then determined by the BCA protein assay bicinchoninic acid
(Pierce), and the hemoglobin content of the same vesicles was
assayed by measuring the absorbance at 541 nm (Si: = 8.63).19
Equal amounts of IOVs, corrected for hemoglobin content, were
then stirred for 20 minutes on ice in 5 vol of 5 mmol/L sodium
phosphate, pH 8.0, containing 1% CIZE~.
The insoluble pellet, if
present, was collected by centrifugation at 35,ooOg for 45 minutes
in an SS-34 Sorvall rotor and washed three times in 5 mL of 5
mmol/L sodium phosphate, pH 8.0, to remove weakly bound
protein.
Quantitative analysis of autologous IgG on iWuh.wemic erythrocytes and their derived aggregates. p-Thalassemic cells were washed
four times in phosphate-buffered saline (PBS), pH 7.4, to remove
plasma and butQ coat, after which they were incubated at 50%
hematocrit for 3 hours with 30 kg/mL of '4-labeled GAH-IgG
(1.96 x 102 cpm/mg). Radioiodination of the antihuman IgG was
performed with IODO-BEADS (Pierce) and NalZI.6 The unl
bound antibody was removed by washing three times in PBS
containing 1% bovine serum albumin, and spectrin-depleted IOVs
were prepared as described above. The IOVs were then assayed for
protein content and the quantity of tightly bound lBI-GAH-IgG
was determined by counting gamma emission. Detergent insoluble
aggregates were then prepared as described earlier and the
lZI-GAH-IgG content in the derived aggregates was also determined by gamma counting. The quantity of lZI-GAH-IgG bound
per ghost was calculated from the cpm/mg protein in the IOVs,
assuming that 0.65 mg of vesicle protein was equivalent to 1 mg of
ghost protein, as determined experimentally. The reason for not
measuring the counts on the intact cells directly was that some
loosely associated IgG (presumably nonspecifically bound) were
found to dissociateduring ghost and IOV preparation. Because the
nature of these easily eluted antibodies was unknown, we felt it
would be wisest to assign them neither to the aggregate nor to
nonaggregated regions of the membrane.
Immunoprecioiation of immune complms. Erythrocyte ghosts
(7.2 mg/mL) 0.4 mL, from p-thalassemic or normal cells were
solubilized in 10 vol of PBS containing c&8 (1% final). The
resulting detergent solution was incubated for 30 minutes at 0°C
with 100 FL of agarose beads (Sigma) to pre-absorb any polypeptides that might bind nonspecifically to the beads. After removing
the beads and other pelletable material by centrifugation, the
supernatant was allowed to react overnight at 4°C with 100 FL of
protein A-linked agarose beads. These beads were then collected
by centrifugation ( 1 , w 3 minutes) and washed five times in PBS
containing 0.1% C12E8. Immune complexes were eluted and
prepared for sodium dodecyl sulfate-polyacrylamide gel electrophoresis (SDS-PAGE) by boiling for 5 minutes in 100 ~ L osample
f
buffer containing 4% SDS with or without 5% f3-mercaptoethanol.
Analytical procedures. Aggregates were prepared for SDSPAGE by solubilizing in SDS electrophoresis buffer, sonicating to
disrupt tightly associated proteins, and heating for 5 minutes in a
boiling water bath. The polypeptide composition of the aggregates
was then analyzed on a 6% to 12% polyacrylamide gradient gel
according to the method of Laemmli?O For immunoblotting studies, proteins were transferred to nitrocellulose membranes" and
blocked with 4% bovine serum albumin in blotting buffer, consisting of 20 mmol/L Tris, 500 mmol/L NaCI, pH 7.4. The resulting
blots were washed with blotting buffer containing 0.05% Tween 20
and reacted with the desired antibody diluted in blotting buffer.
After further washing and labeling with second antibody conju-
gated to horseradish peroxidase, the immunoblotswere developed
using 4-chloronaphthol as the substrate.
RESULTS
SDS-PAGE analysis of hemichrome-stabilized aggregates
isolated from @thalassemic cells. To evaluate the possible
existence of hemichrome-stabilized membrane protein aggregates in p-thalassemic cells, membranes were prepared
and extracted according to the protocol described in Materials and Methods. As shown in Fig 1lanes a through d, the
Coomassie blue staining pattern of total membrane proteins from thalassemic cells was not significantly different
from the staining pattern of total proteins from normal
ghosts. Except for a new band appearing at -68,000
daltons, which was previously noted by JSahane and Rachmilewitz,22 and an unusually large amount of retained
globin, no major distinction was observed. Whether the new
68,000-dalton band is a breakdown product of a higher
molecular weight component, an over-expressed minor
isoform of band 4.1, or a protein absorbed from the serum
or cytoplasm cannot be evaluated from the data. However,
band 3
4.19
4.2/
?
5
6
7.
Hblr
Fig 1. Gel electrophoresis of detergent insoluble macromolecular
aggregates isolated from equal amounts of control and p-thalassemic
erythrocytes. The total detergent insoluble material (aggregates)
isolated from 12 mg of @thalassemic lOVs (0.29 mg) and normal
(control) IOVs (not detectable) was suspended in 2 vol of SDS
electrophoresis buffer containing 5 % 2-mercaptoethanol, sonicated
for 5 seconds to disrupt tightly associated proteins, heated for 5
minutes in a boiling water bath, and loaded onto a 6% to 12%
Laemmli gradient SDS-polyacrylamide gel. The lanes contain (a)
p-thalassemic ghosts from a splenectomized patient (50 pg), (b)
pthalassemic ghosts from a nonsplenectomized patient (50 pa), (c)
normal erythrocyte ghosts (travel control) (50 pg), (d) normal erythrocytes ghosts from a local donor (50 pg), (e) lOVs from a p-thalassemic
splenectomized patient (50 pg), (f) lOVs from a pthalassemic nonsplenectomized patient (50 pg), (9) lOVs from the normal travel control
(50 pg), (h) lOVs from the normal local donor (50 pg), (i) total
detergent-insoluble aggregate from the p-thalassemic splenectomized patient (50 pg), (j) total aggregate from the p-thalassemic
nonsplenectomized patient (4 pg), (k) total aggregate from the travel
control (not detectable), (I)total aggregate from the local control (not
detectable). The unidentified -68,000-dalton protein is designated
with a question mark.
IMMUNE COMPLEXES FROM P-THALASSEMIC CELLS
it should be pointed out that this band is not enriched in the
protein aggregate derived from thalassemic cells.
When membrane protein aggregates were isolated from
equal concentrations (6.8 mg/mL) of thalassemic and
normal (control) IOVs, vastly different yields were obtained. Whereas virtually no detergent insoluble protein
could be collected from control IOVs, (Fig 1,lanes k and I),
roughly 1.6% of the total membrane protein could be
harvested from thalassemic 10% Further, under identical
conditions, there was an approximately 10- to 30-fold
greater yield of aggregates from erythrocytes of splenectomized thalassemia patients than from erythrocytes of nonsplenectomized patients (Fig 1, compare lanes i and j).
Densitometric analysis of the Pthalassemic membrane
aggregates. The contribution of individual polypeptides to
the total protein content of the aggregates (Fig 1, lane i)
derived from erythrocytes of splenectomized thalassemic
individuals was determined by scanning densitometric analysis. For this purpose the lane containing intact ghost was
first scanned to determine the position of each major
membrane polypeptide in the gel. Then, because the
aggregate polypeptides migrated rather diffusely, the corresponding region of the scan from the aggregates lane was
integrated for each polypeptide and listed in Table 1. As
predicted, globin monomer represents the major polypeptide in the aggregates, comprising 43% to 52% of the total
protein. The other major polypeptides were in the band 3
region (5% to lo%), bands 1,2, and 2.1 region (2% to 5%),
band 4.1 (1% to 3%), band 5 (3% to 5%), and band 7 (5%).
Thus, as with sickle cells,11globin and band 3 were found to
be prominent proteins of the detergent insoluble thalassemic cell aggregates.
Immunoblotting analysis of the Pthalassemic membrane
aggregates. Because of the diffuse nature of the gel staining, we decided to rely on immunologic methods for
identificationof the proteins in the aggregates. Immunoblotting analysis with affinity-purified polyclonal antibodies to
Table 1. Relative Content of the Major Erythrocyte Membrane
Polypeptides Isolated in the Detergent-Insoluble Aggregates
Obtained From p-Thalassemic Erythrocytes
Protein
1,212.1
Band 3, monomer
Band 3, oligomer
4.1
4.2
5
6
7
Globin
Unidentified polypeptides
Protein in
Aggregate (X)
2.1
5.7
3.2
1.6
1.6
3.4
1.9
4.7
44
32
Data were obtained by scanning densitometry using a Shimadzu
CS9000 dual wavelength scanner of the Coomassie blue-stained gel in
Fig 1, lane i. Because the aggregate polypeptides migrated diffusely in
the polyacrylamide gel, the band area for each major membrane
polypeptide was integrated over the banding region occupied by the
same polypeptide in the gel of control ghosts. The location of the band
3 oligomer was determined by Western blotting, as in Fig 6c. lane 1.
3009
band 3 clearly demonstrated the presence of the anion
transporter (Fig 2, lane C). Nonreducible cross-linked
forms of band 3 were also seen in the same pellet at
190,000 daltons (band 3 dimer) and at the top of the gel
(lane C). Normal cell ghosts stained for band 3 only at
100,000 daltons (monomer) and -62,OOO daltons (common proteolytic fragment; lane A). The control cell pellet
(lane B) contained no detectable band 3.
To evaluate whether autologous IgG and complement
were also present at these sites of membrane protein
reorganization, 40 pg of thalassemic cell aggregates were
examined as above except they were immunoblotted with
horseradish peroxidase conjugated to either antihuman
IgG (Fig 3) or antihuman C3c (Fig 4) under nonreducing
conditions. After staining with 4-chloronaphthol, the IgG
(Fig 3, lane C) and complement (Fig 4, lane C) were easily
detected in the aggregates from thalassemic cells but not in
pelletable material from normal erythrocyte IOVs (lanes B
of Figs 3 and 4). Furthermore, IgG and complement were
undetectable in 50 pg of whole thalassemic cell ghosts
immunoblotted under identical conditions (not shown).
This indicates that autologous IgG and complement are
significantly concentrated in hemichrome-enriched aggregates of @-thalassemiccells.
Quantitative analysis of cell surface autologous IgG and
complement component C3 associated with &thalassemic cell
aggregates. To obtain a more quantitative estimate of the
degree of autologous IgG enrichment at sites of integral
membrane protein clustering, @-thalassemiccells were
labeled with *=I-antihuman IgG and the antibody content
of the whole membranes and derived aggregates was
determined. As shown in Table 2, an average of 27% of the
total cell surface autologous IgG was found to co-isolate
with the aggregates from splenectomized individuals. Because the aggregates constitute only 1.6% of the membrane
protein, this corresponds to a relative enrichment of autologous IgG at the aggregated sites over the remainder of the
thalassemic cell membrane of 23-fold (Table 2). This
suggests that the sites of integral protein clustering in
thalassemic cells are indeed a major locus of IgG binding.
Similar analysis of the enrichment of complement component C3 at the same sites shows a 40-fold enhancement of
complement in the detergent-insolublemembrane material
(Table 3).
Absorption spectrum of membrane-associated globin in
Pthalassemic cells. To evaluate whether the membraneassociated globin in the thalassemic cells was native or
denatured, an absorption spectrum of extensively washed
membranes was obtained. For this purpose thalassemic and
normal cell membranes were washed in cold lysis buffer
until no hemoglobin could be detected in the supernatant,
after which the membranes were immediately solubilized in
10 vol of PBS containing C12E8(2% final) and scanned in a
UV-visible spectrophotometer. The spectrum of p-thalassemic cell membranes shown in Fig 5 was obtained using
control membranes with no bound hemoglobin or hemichromes as the reference. When compared with standard
spectra of hemoglobin and hemichromes under similar
conditions,1°the spectrum indicates that the globin bound
-
-
YUAN ET AL
3010
A
FF-
B
~iw@?J!p
C
L i F F i V X
ePolymer
eDimer
A agarose beads and determined whether the other three
proteins were also present. After extensive washing, the
proteins bound to the beads were eluted in SDS,separated
by SDS-PAGE, transferred to nitrocellulose, and visualized
with antibodies against IgG, C ~ Cband
,
3, and Hb,as shown
in Fig 6. Importantly, the immunoprecipitated proteins
stained positively for IgG (lane al), complement compo-
A
B
C
a Monomer
._
Fig 2. lmmunoblotting analysis of the presence of band 3 in the
insoluble aggregates isolated from 2.0 mg of normal and 6-thalassemic IOVs. Nitrocellulose blots of samples separated by SDS-PAGE
were incubated in affinity-purified polyclonal rabbit anti-band 3 for 3
hours followed by goat anti-rabbit IgG conjugated t o horseradish
peroxidase (GAR-HRP)for 2 hours and then developed with 4-chloronaphtol. The lanes contain (A) normal red blood cell ghosts (40 pg), (6)
detergent-insoluble material from normal cells (not detectable), (C)
detergent-insoluble material from 6-thalassemic cell (40 pg).
to thalassemic cell membranes is mainly in the form of
hemichromes. This observation is consistent with our earlier finding that hemichromes, not hemoglobin, have a
strong affinity for the cytoplasmic domain of band
I"unopwc@itation
Of
complms from pthalassemic Cell membranes. The fact that IgG, complement,
band 3, and globin are present in the same detergentinsoluble membrane pellet suggests, but does not prove,
that the four components are physical1y associated in a
To examine this possibili@, we
associated material from the detergent extracts on protein
3.1071*
Fig 3. lmmunoblotting analysis of the presence of IgG in the
insoluble aggregates isolated from 2.0 m g of normal and 6-thalassemic cell IOVs. Detergent-insoluble pellets were dissolved in 2 vol of
Laemmli electrophoresis buffer without 6-mercaptoethanol, heated
for 5 minutes, separated on a 6% t o 12% polyacrylamide gel, and
transferred t o nitrocellulose. The blots were then incubated in GAHIgG for 3 hours followed by rabbit antigoat IgG conjugated t o
horseradish peroxidase for 2 hours. The lanes contain (A) pure IgG (3
pg), (6) the entire normal cell pellet (not detectable), (C) the 6-thalassemic cell aggregates (40 pg).
3011
IMMUNE COMPLEXES FROM p-THALASSEMIC CELLS
A
.=w-?&p
B
?mvw
C
Table 2. Estimation of the Fraction of Autologous Cell Surface IgG
Associated With @-ThalassemicCell Aggregates Obtained From
Splenectomized Individuals
m=
.-
Total
Aggregate
IgG
Membrane
GAH-IgGI Associated Concentratedin
Protein in Enrichment
Patient
Cell*
GAH-IgGICeIl Aggregate (%I Aggregate (%) Factort
.i'
is; .,
1
2
3
Average
: $ -185
kd
561
832
678
690
157
191
203
184
28
23
30
27
1.8
1.3
1.6
1.6
21
23
26
23
*Represents the number of GAH-IgGIcell. If more than one GAH-IgG
can opsonize a surface-bound IgG, then the total number of IgG will be
fewer by this factor.
tThe enrichment factor represents the fraction of IgG in the aggregate ),;(f
normalized to the fraction of protein in the aggregate (ftroIein)
divided by the fraction of IgG in the remainder of the membrane
normalized to the fraction of protein in the remainder of the membrane
(fzoIein), as described in the following equation:
(ftG)
Enrichment Factor =
;GIf
$rotein
:GIf
ZoIein
also observed, suggesting abnormal reactions may be occurring in these immune complexes. Taken together, it would
appear that autologous IgG, band 3, complement, and
globin are all present in the same complexes, and that
similar complexes can be isolated by both simple centrifugation and immunoprecipitation.
DISCUSSION
Y
Fig 4. lmmunoblotting analysis of the presence of complement in
the insoluble aggregates isolated from 2.0 mg of normal and @-thalassemic cell IOVs. Total pelletable material collected from detergenttreated lOVs was dissolved in an equal volume of electrophoresis
buffer without @-mercaptoethanol,heated for 5 minutes, separated
on a 6% t o 12% polyacrylamide gel, and transferred t o a nitrocellulose
membrane. The blots were then incubated in GAH-C3c for 3 hours
followed by rabbit antigoat horseradish peroxidase for 2 hours. The
lanes contain (A) serum (5 pL), (B) the entire normal cell pellet (not
detectable), (C) the @-thalassemiccell aggregates (40 Fg).
nent C3 (lane bl), band 3 (lane cl), and globin (lane dl). In
contrast, collection of similar immune complexes from
normal cell membranes yielded no stainable material (Fig
6, lanes 2). One significant finding in the band 3 blot (lane
cl) was that the dimer and polymer were present in
enriched amounts. This could infer that autologous IgG on
thalassemic cells display a preference for clustered over
dispersed band 3, as has been recently demonstrated for
other types of erythrocytes.z In the hemoglobin blot (lane
dl), many nonreducible cross-linked species of globin were
Numerous alterations in the structure and function of
thalassemic eqthrocytes have been described,' among them
Shinar et alZ4have shown that 3% to 8% of the protein in
isolated P-thalassemia cell membranes is tightly bound
globin. It would now appear that at least some of this globin
is associated with immune complexes containing autologous IgG, complement component C3, and band 3. If the
mechanism documented in model studies3J0J1J3 and observed at various stages in sickle cells5J2 and the densest
fraction of normal cells6is correct, we suggest the instability
Table 3. Estimation of the Fraction of Cell Surface C3 Associated
With @-ThalassemicCell Aggregates
Aggregate
Total
Associated
Membrane
GAH-C3cl GAH-C3cl C3 Concentratedin Protein in Enrichment
Sample Cell" (xlO3)Cell (xlO3) Aggregate (%) Aggregate (%I
Factort
1
2
3
Average
11.6
10.9
9.52
10.7
4.42
3.67
3.06
3.72
38
34
32
35
1.1
1.2
2.3
1.5
55
42
23
40
*Represents the number of GAH-C3clcell. If more than one GAH-C3c
can opsonize a surface-bound C3, then the total number of C3 will be
fewer by this factor.
tThe enrichment factor represents the fraction of C3 in the aggregate
(fe3) normalized to the fraction of protein in the aggregate (f$roIein)
divided by the fraction of C3 in the remainder of the membrane (f&)
normalized to the fraction of protein in the remainder of the membrane
(fZotein),as described in the following equation:
Enrichment Factor =
-
fkJf$rotein
f
%
:oleinIf
YUAN ET AL
3012
I
I
510
I
I
I
.... ........
590
Wavelength (nm)
550
630
I ]
Fig 5. The absorption spectrum of solubilized 6-thalassemic cell
membranes. Ghosts from normal and pthalassemic cells (7.2 mg/
mL), 0.4 mL, were washed in cold lysis buffer until no hemoglobin
could be detected in the supernatant. At this point the membranes
from control erythrocyteswere white, while the membranesfrom the
6-thalassemic cells were distinctly reddish-brown. The ghosts were
solubilized in 10 vol of PBS containing C,*E8 (2% final) and their
UV-visible absorption spectrum was obtained in an IBM UV-visible
The instrument was
9420 spectrophotometer (Danbury, CT) (-).
balanced for this spectrum using control white ghosts (free of
hemoglobin or hemichromes) in the reference cell. For comparison,
and 0.12
the standard spectra of 0.2 mg/mL hemichrome (---)
mg/mL oxyhemoglobin (--)shown.
of a-globin chains along with their ability to cluster band 3
strongly influence the cell's lifespan. Once the denatured
globin chains collect a few copies of band 3 into an
a g g r e g a t e ? ~ ~ J autologous
@~~
antibodies and complement
likely associate spontaneously and promote the cell's removal.l3 Lessin et alZ have directly observed the clustering
of integral membrane proteins at sites of brilliant cresyl
blue-induced hemichrome binding to the a-thalassemic cell
1 2
-
membrane (hemoglobin H disease) by electron microscopy.
We have isolated related clusters from P-thalassemic erythrocytes and found them to also contain complement and
IgG. While membrane
altered ion permeability,27,28 and exposure of cryptic p-galactosyl residues29 etc
may all contribute to the cell's early demise, an additional
participant in the removal process must be the opsonization
of clustered integral membrane proteins.
Although the hemichrome-rich membrane protein aggregates contain 23 times as much autologous IgG per mg
membrane protein as other regions of the membrane (see
Table 2), this enrichment factor is still well below the value
measured for similar aggregates isolated from sickle cells5
and the densest (most rapidly phagocy~tosed,3~.~*)
fraction
of normal cells? Thus, corresponding autologous IgG
enrichment factors for these cells were calculated to be
275-fold and 680-fold, respectively. Whereas integral membrane protein clusters likely represent the predominant
opsonization site on sickle and normal cells, other regions
of P-thalassemic cell membranes could be equally or more
important in accumulating IgG. It is possible that other
mechanisms yet to be explored are involved in causing
damage to the p-thalassemic cell surface. It would be
interesting to know what fraction of the immune complexes
in p-thalassemic cell membranes are in fact devoid of
clustered band 3.
Finally, as with sickle and dense normal cells,5S6 the
globin in the p-thalassemic cell aggregate was heavily
cross-linked. In fact, in the absence of reducing agent, little
protein in any of these three types of aggregates was found
to enter an SDS polyacrylamide gel. Because electrophoretic analysis of the unfractionated p-thalassemic cell
1 2
c3
-
1 2
1 2
band 3-
globinD
a
b
C
d
Fig 6. lmmunoblottlng analysis of proteins collected by adsorption of detergent-insolublemembrane aggregates to protein A agarose beads.
IEI (1%final). After
Erythrocyte ghosts (7.2 mg/mL), 0.4 mL, from thalassemic and normal cells were solubilized in 10 vol of PBS containing C2
pre-absorptionwith unmodified agarose beads, IgG-containingcomplexes were collected on protein A-agarose beads and extensively washed
(see Materials and Methods). The absorbed proteins were eluted from the protein A beads with SDS and then analyzed by SDS-PAGE and
immunoblotting with the desired antibody. Antibodies used were specific for human IgG (a), complement component C3c (b), band 3 (c), and
human hemoglobin (d). Lanes 1 contain proteins collected from solubilized pthalassemic cell membranes, while lanes 2 contain proteins
collected by the same procedurefrom solubilized normal cell membranes.
IMMUNE COMPLEXES FROM P-THALASSEMIC CELLS
3013
membranes displays a relatively normal polypeptide banding pattern even in the absence of reducing agent, it can be
concluded that intense oxidative reactions are concentrated
in the compact membrane clusters. Perhaps once a dense
mat of hemichromes forms at a site on the membrane,
enzymes responsible for protecting the cell against reactive
oxygen species are excluded and oxidative reactions proceed unabated.
REFERENCES
1. Vigi V, Volpato S, Galiurro D: The correlation between red
cell survial and excess of a-globin synthesis in P-thalassemia. Br J
Haematol56:25,1969
2. Rigas DA, Koler RD: Decreased erythrocyte survival in
hemoglobin H disease as a result of abnormal properties of
hemoglobin H: The benefit of splenectomy.Blood 18:1,1961
3. Low PS, Waugh SM, Zinke K, Drenckhahn D: The role of
hemoglobin denaturation and band 3 clustering in red cell aging.
Science 227531,1985
4. Low PS, Kannan R: Effect of hemoglobin denaturation on
membrane structure and IgG binding: Role in red cell aging. Red
Cell 319:525,1989
5. Kannan R, Labotka RJ, Low PS: Isolation and characterization of the hemichrome-stabilized membrane protein aggregates
from sickle erythrocytes. J Biol Chem 263:13766,1988
6. Kannan R, Yuan J, Low PS: Isolation and partial characterization of antibody- and globin-enriched complexes from membranes
of dense human erythrocytes. BiochemicalJ 278:57,1991
7. Rachmilewitz EA, Shalev 0, Galili U, Schrier S L Erythrocyte membrane alterations in P-thalassemia. Clin Haematol 14:
163,1985
8. Polliack A, Rachmilewitz E A Ultrastructure studies in
P-thalassemia major. Br J Haematol24:319,1973
9. Wickramasinghe SN, Hughes M, Hollan SR: Electronmicroscope and high-resolution autoradiographic studies of the
erythroblasts in hemoglobin-H diseases. Br J Haematol 45:401,
1980
10. Waugh SM, Low PS: Hemichrome binding to band 3:
Nucleation of Heinz bodies on the erythrocytemembrane. Biochemistry 24:34, 1985
11. Waugh SM, Walder JA, Low PS: Partial characterization of
the copolymerization reaction of erythrocyte membrane band 3
with hemichromes. Biochemistry 26:1777,1987
12. Waugh SM, Willardson BM, Kannan R, Labotka RJ, Low
PS: Heinz bodies induced clustering of band 3, glycophorin, and
ankyrin in sickle cell erythrocytes. J Clin Invest 78:1155,1986
13. Turrini F, Arese P, Yuan J, Low PS: Clustering of integra
membrane prateins of the human erythrocyte membrane stimulates autologousIgG binding, complement deposition and phagocytosis. J Biol Chem 266:23611,1991
14. Singer JA, Jennings LK, Jackson CW, Docker ME, Morrison
M, Walker WS: Erythrocyte homeostasis: Antibody-mediated recognition of the senescent state by macrophages. Proc Natl Acad Sci
USA 83:5498,1986
15. Lutz HU, Fasler S, Stammer P, Bussalino F, Arese P:
Naturally-occurring anti-band 3 antibodies and complement in
phagocytosis of oxidatively-stressedand in clearance of senescent
red cells. Blood Cells 14:1755,1980
16. Rund D, &hen T, Filon D, Dowling CE, Warren T, Barak I,
RachmilewitzEA, Kazazian HH Jr, Oppenheim A Evolution of a
genetic disease in an ethnic isolate: Beta-thalassemiain the Jews of
Kurdistan. Proc Natl Acad Sci USA 88:310,1991
17. Dodge TG, Mitchell RJ, Hanahan DJ: The preparation and
chemical characteristic of hemoglobin-free ghosts of human erythrocytes. Arch Biochem Biophys 100:119,1963
18. Marchesi SL, Steers E, Marchesi VT: Physical and chemical
properties of a protein isolated from red cell membranes. Biochemistry 950,1970
19. Williams RC, Tsay K A convenient chromatographicmethod
for the preparation of human hemoglobin. Anal Biochem 54:137,
1973
20. Laemmli U K Cleavage of structural proteins during assembly of head of bacteriophage-T4.Nature 227680,1970
21. Towbin H, Staehlin R, Fordon J Electrophoretic transfer of
proteins from polyacrylamide gels to nitrocellulose sheetsprocedure and some application. Proc Natl Acad Sci USA 76:4350,
1979
22. Kahane I, Rachmilewitz E A Alterations in red blood cell
membrane and effect of vitamin-E on osmotic fragility in p-thalassemia major. Israel J Med Sci 12:11,1976
23. Lutz H U Red blood cell aging. Blood Cells Subcell Biochem
17:81,1990
24. Shinar E, Shaley 0, Rachmilewitz EA, Schrier S L Erythrocyte membrane skeleton abnormalities in severe beta thalassemia.
Blood 70:158,1987
25. Lessin LS,Jensen W, Klug P: Ultrastructure of the normal
and hemoglobinopathic RBC membrane. Freeze-etching and stereoscan electron microscopic studies. Arch Intern Med 129:306,
1972
26. Tillman W, Schroter W: Deformability of erythrocytes in
iron deficiency anemia. Blut 40179,1980
27. Gunn RB, SilversDN, Rosse WF: Potassium permeability in
P-thalassemia minor red blood cells. J Clin Invest 51:1044,1972
28. Vettore L, Falezza GC, Cetto GL, De Matteis M L Cation
content and membrane deformability of heterozygous thalassemic
red blood cells. Br J Haematol27:429,1974
29. Galili U, Rachmilewitz EA, Peleg A, Flechner I: A unique
natural human IgG antibody with anti-a-galactosyl specificity. J
Exp Med 160:1519,1984
30. Tenbrinke M, Regt JD: Cr-51-half life time of heavy and
light human erythrocytes. Scand J Haematol723,1970
31. Mechanism of red cell aging, relationship evaluated by
ultracentrifugation in a discontinuous density gradient. J Lab Clin
Med 69:659,1967