THE JOURNAL OF BIOLCGIC~L CHEMISTRY
Vol. 247, No. 8, Issue of April 25, pp. 2512-2521, 1972
Printed in U.S.A.
The Isolation and Characterization of Bovine Factor VIII
(Antihemophilic Factor) *
(Received for publication, September 20, 1971)
GOTTFRIED SCHMER,~: EDWARD P. I<IRBY,$ DAVID
C.
TELLER, AND EARL W. DAVIE
From the Department of Biochemistry, University of Washington, School of Medicine, Seattle, Washington
98195
SUMMARY
Bovine Factor VIII (antihemophilic factor) has been puri-
fied approximately lO,OOO-fold with an over-all yield of 10 to
15%. The purification procedure involves BaS04, kaolin,
and bentonite adsorption to remove contaminants, ethanol,
polyethylene glycol, and /Y-alanine fractionation, calcium
citrate-cellulose chromatography, concanavalin A precipita-
tion, and an Agarose gel filtration step. The final product is
homogeneous when examined by gel filtration, density
gradient centrifugation, and zone electrophoresis. It also
shows a single precipitin line when subjected to immuno-
electrophoresis employing a specific rabbit antibody against
Factor VIII. However, when examined by sedimentation
equilibrium, the preparation is physically heterogeneous,
apparently due to substantial aggregation of the protein.
In these experiments, the smallest species which could be
calculated has a molecular weight of approximately 1.1
million. After reduction in Z-mercaptoethanol, the protein
shows one band in polyacrylamide gel electrophoresis and
zone electrophoresis and a reduction in size. The smallest
species of the reduced protein has a molecular weight of
about 85,000. The protein contains approximately 11%
hexose, 2 % sialic acid, 7 % hexosamine, and no lipid. Other
general properties of this protein including its amino acid
composition are also reported.
A careful study of the role of Factor VIII (antihemophilic fac-
tor) in blood coagulation is dependent upon a highly purified
* This work was supported in part by Research Grants GM 10793
and HE 11857 from the National Institutes of Health and funds
from Initiative 171 from the State of Washington. The nomen-
clature employed in this manuscript for various coagulation fac-
tors is that recommended by an international nomenclature com-
mittee (Wright, I. (1959) J.-Amer. iYe& Ass.
170, 325).
1: Recinient of a Suecial Fellowship from the National Institutes
. _
of Health. Present address, Department of Laboratory Medicine,
University of Washington, School of Medicine, Seattle, Washing-
ton
98195.
$ Recipient of a Postdoctoral Fellowship from the National
Institutes of Health. Present address, Department of Biochem-
istry, Health Sciences Center, Temple University, Philadelphia,
Pennsylvania
19140.
preparation which has well defined physical-chemical character-
istics. Although many different methods of preparation for this
protein from various sources have been described during the
past 20 years, few have yielded preparations of high quality and
in sufficient quantity to permit detailed studies. Most methods
of purification have involved ethanol (l), ether (a), or polyeth-
ylene glycol (3) fractionation, precipitation with amino acids
(4), phosphate (5), citrate (6), and by decreased temperature
(7), and removal of fibrinogen contaminants with absorbents
such as bentonite (8). Gel filtration techniques (9) and con-
canavalin A precipitation (10) have also provided some success
in the purification of this protein. Recently, human Factor VIII
has been purified approximately lO,OOO-fold by Hershgold
et al.
(3) by means of a combination of several of these methods.
In the present communication, we wish to describe a relatively
simple method for the purification of milligram quantities of bo-
vine Factor VIII. This procedure leads to a product of high
purity and good yield. Various physical-chemical and biological
properties of this preparation are also described. Preliminary
studies dealing with this work have been published elsewhere
(11).
MATERIALS
Heparin sodium salt (Grade I) and e-amino-n-caproic acid
were purchased from Sigma. Benzamidine hydrochloride was
obtained from Aldrich. Barium sulfate (x-ray grade) was pur-
chased from Merck (U. S.), and kaolin N. F. was obtained from
J. T. Baker Chemical Co. Bentonite was a product from
Prolabo, Paris, France. Sodium barbital was purchased from
Mallinckrodt Chemical Works. Polyethylene glycol (Carbowax
6000) was obtained from Union Carbide Corp., and imidazole,
p-alanine, and calcium citrate (Ca3(C6H50& ‘4 HzO) were ob-
tained through Matheson Coleman and Bell, Cincinnati, Ohio.
Cellulose powder was Whatman CFll fibrous powder, Reeve
Angel, Clifton, N. J. The final step of gel filtration was carried
out with Agarose A-15m or A-50m supplied by Bio-Rad, Rich-
mond, Calif. Methyl-a-n-glucopyranoside (B grade) was ob-
tained from Calbiochem. Asolecithin, purchased from Associ-
ated Concentrates Lecithin, Woodside, N. Y., was used as a
platelet substitute. Dri Film silicone (SC 87) was a product
from General Electric, Waterford, N. Y., and all glassware was
coated with this material unless otherwise noted. Human
fibrinogen was purchased from Warner-Chilcott, Morris Plains,
N. J. Bovine thrombin employed for the fibrinogen assays was
2512
by guest, on July 18, 2011www.jbc.orgDownloaded from
Issue of April 25, 1972
G. Xchmer, E. P. Kirby, D. C. Teller, and E. W. Davie
2513
purchased from the Upjohn Co. Concanavalin A was prepared
according to the method of Agrawal and Goldstein (12). Human
plasma (Y~ glycoprotein was kindly provided by Dr. S. I. Hako-
mori, University of Washington, Seattle, Wash. All other chem-
icals were commercial preparations of the highest quality avail-
able.
METHODS
Protein concentration was measured by the ninhydrin reac-
tion following alkaline hydrolysis according to the method of
Moore (13) or absorption at 280 nm. The hexose content of
Factor VIII was determined by the phenol sulfuric acid method
of Dubois et al. (14), using glucose as a reference standard. Hex-
osamine was determined by the Elson and Morgan reaction (15),
employing galactosamine as a standard. For the quantitative
determination of sialic acid, the thiobarbituric acid assay of
Warren was used (16), employing human o(1 glycoprotein as a
reference standard. This protein contains approximately 11 y.
sialic acid (17).
Factor VIII was reduced and alkylated by dissolving 5 mg of
protein in a total volume of 1.5 ml containing 720 mg of recrys-
tallized urea, 0.08 ml of a 0.1
M
EDTA solution, pH 8.6, 0.02 ml
of 14
M
2-mercaptoethanol, and 0.6 ml of a 0.1
M
Tris-HCl solu-
tion, pH 8.6. The above reaction mixture was allowed to stand
for 4 hours at room temperature under nitrogen, followed by the
addition of 0.2 ml of 1
N
NaOH containing 54 mg of recrystallized
monoiodoacetic acid. After 15 min, the solution was dialyzed
overnight against 0.01
M
sodium phosphate buffer, 0.1
M
2-mer-
captoethanol, and 8
M
urea at room temperature before being
subjected to further analysis. In some experiments, it was
passed over a Sephadex G-25 column (2 X 40 cm) previously
equilibrated with 5% acetic acid and lyophilized using a Virtis
lyophilizer.
Antibody Preparation-Rabbits were immunized with either
the highly purified Factor VIII or material which had only been
purified through the early stages of the procedure (up through
the adsorption and elution from tricalcium citrate). Initial
immunization was achieved by multisite intramuscular or sub-
cutaneous injection of 1 to 2 mg of protein in the presence of
Freund’s complete adjuvant. After 4 to 6 weeks, the rabbits
were reinjected at intervals, and then bled. Reinjection was
intramuscularly in the presence of Freund’s incomplete adjuvant
for rabbits immunized against the high purity Factor VIII. For
immunization against the lower purity material, rabbits were in-
jected intravenously with alum-precipitated material.
Blood collected from the rabbits was allowed to clot overnight
at room temperature. The serum was treated with BaS04 (100
mg per ml) and centrifuged. Saturated ammonium sulfate was
then added to 330/, saturation, the precipitate centrifuged, and
redissolved in 0.15
M
NaCl to one-half the original serum volume.
Precipitation with 33% ammonium sulfate was repeated two
times, and the final precipitate dissolved in 0.15
M
NaCl to one-
half the original volume. This was then dialyzed extensively
against 1 mM phosphate buffer, pH 7.4, and a small precipitate
removed by centrifugation. Sodium chloride was added to make
the solution 0.15
M,
and sodium azide added (0.02’%) to retard
bacterial growth. The antibody was stored either in the cold
room or frozen.
EZectrophmesis--Polyacrylamide disc gel electrophoresis of the
reduced Factor VIII was carried out by the general method of
Davis (18). In the present experiments, electrophoresis was
performed in 8
M
urea and 0.025
M
Tris-HCl and 0.2
M
glycine
buffer, pH 8.7. A 3.5% gel was employed and electrophoresis
was carried out for 4 hours at 4”. The gels were stained for pro-
tein with Amido black and for carbohydrate by the method of
Zacharius et al. (19).
Zone electrophoresis was carried out in 0.05
M
sodium barbital,
pH 8.6, on microscope slides layered with 0.5% Agarose and 2%
unpolymerized acrylamide as described by Williams and Chase
(20). Samples (20 ~1 containing 10 to 20 pg of protein) were
placed in a small well and electrophoresis was carried out at room
temperature for 45 min with 150 volts and 5 ma per slide. The
slides were stained for protein with Amido black.
Immunoelectrophoresis on 0.5% Agarose containing 2% poly-
acrylamide on microscope slides (25 x 75 mm) was carried out
according to the method of Scheidegger (21). Essentially the
same conditions were employed as those for zone electrophoresis.
Samples were diluted in 0.05
M
sodium barbital buffer, pH 8.6,
and run for 45 min. Antibody was added to the center trough
and allowed to diffuse for at least 48 hours. The slides were
photographed employing indirect lighting.
Amino Acid Analysis-Samples for amino acid analysis were
prepared by the method of Moore and Stein (22). These sam-
ples (2 to 4 mg) were hydrolyzed in 5.8
N
constant boiling HCl
for 24, 48, 72, and 96 hours. A small black precipitate, presum-
ably due to humin formation, was removed by centrifugation
prior to analysis. The protein concentration for each sample
was determined by the ninhydrin reaction following alkaline hy-
drolysis as described earlier since the protein dried in vacua or by
lyophilization formed an insoluble, sticky gum which could not
be readily weighed. Amino acid analyses were carried out on
250~pg protein samples using a Spinco model 120 amino acid
analyzer according to the method of Spackman et al. (23). The
final protein concentration of each sample was calculated by
summation of the amino acid composition, and this was in good
agreement with that determined on the original sample by the
ninhydrin reaction employing leucine as a standard. Samples
from three different preparations of Factor VIII were employed
and gave essentially identical results. The reported values for
serine and threonine are extrapolations to zero time hydrolysis,
whereas isoleucine, leucine, and valine values are the average of
the 96-hour hydrolysis. Half-cystine and methionine were de-
termined after performic acid oxidation according to Hirs (24).
Tryptophan was determined by the method of Bencze and
Schmid (25).
Ultracentrifuge Analysis-Sedimentation equilibrium experi-
ments were performed as described by Harris et al. (26) and Seery
et al. (27). Calculations were made by the methods discussed
by Teller et al. (28). All experiments were conducted at 20”.
The partial specific volume of the protein moiety was calculated
using the ammo acid composition according to the method of
McMeekin et al. (29). To correct this value for carbohydrate
content (see Table IV), a value of 0.62 ml per g was employed
(30). Thus, the actual value of B was calculated as follows: B =
(0.74 ml per g x 0.80) + (0.62 ml per g
X
0.20) = 0.72 ml per g.
For the experiment in 6
M
guanidine HCl, the value was lowered
by 0.01 ml per g (31).
Density gradient ultracentrifugation studies of Factor VIII
were carried out in a Spinco model L with a SW 39 rotor employ-
ing 5-ml cellulose nitrate tubes. The gradient employed was
10 to 35% methyl-cy-n-glucopyranoside in 0.2
M
sodium perchlo-
rate and 0.05
M
borate buffer, pH 8.0. Centrifugation time was 5
by guest, on July 18, 2011www.jbc.orgDownloaded from
2514 Bovine Factor VIII Vol. 247, N-o. 8
hours at 39,000 rpm at room temperature. The Factor VIII
solution was collected in 5-drop samples and analyzed for protein
by the ninhydrin method and for Factor VIII activity.
Clotting
Assays-Platelet-deficient titrated plasma was pre-
pared from normal human blood and from the blood of patients
with various coagulative abnormalities by previously described
methods (32). In this procedure, plasma is not permitted to
come in contact with glass and is frozen at -90” until employed.
Pooled platelet-deficient oxalated bovine blood was employed as a
reference standard in the Factor VIII assay and was stored at
- 90”.
Factor VIII activity was measured by the kaolin-activated
partial thromboplastin time employing the method of Langdell
et al. (33) as modified by Proctor and Rapaport (34). In this
assay, 0.1 ml of Factor VIII-deficient plasma (Factor VIII level
less than 1%) is previously incubated with 0.1 ml of an appro-
priate serial dilution of Factor VIII (for the starting plasma
1: 10, 1:20, 1:40), 0.1 ml of a 0.1% asolecithin solution, and 0.1
ml of a 2% kaolin suspension for 6 min at 37”, followed by the
addition of 0.1 ml of 0.025
M
CaC12. The time which is required
to form a clot after the addition of Ca2+ is recorded with a stop
watch. The tilting method is used with siliconized glass tubes
@-mm internal diameter). A standard reference curve was
prepared from the most highly purified product. A straight
line was obtained when various concentrations of Factor VIII
and their clotting times were plotted on double log paper. The
slope of the line made from Factor VIII at different stages of
purification was parallel to that obtained with the most highly
purified product. For a given set of reagents (i.e. the same
deficient plasma, phospholipid, and length of prior incubation),
clotting times were readily reproducible. One unit of Factor
VIII was arbitrarily chosen as that amount of activity present
in 1 ml of fresh bovine plasma. For experiments other than
those shown in Table I, the units are only approximate since
fresh bovine plasma was not readily available as a reference
standard.
Fibrinogen was assayed employing the methods of Jacobsson
(35) and Clauss (36), and Factor V activity was measured on
artificially depleted Factor V substrate plasma (37). Pro-
thrombin concentrations were determined by the method of
Quick et al. (38). Various other coagulation factors (Factors
VII, IX, X, and XII) and plasminogen were determined essen-
tially as described by Biggs and Macfarlane (37).
Phospholipid extraction was carried out on 5-mg samples of
Factor VIII which were extracted with chloroform-methanol,
and the extract dried under nitrogen. The samples were then
heated to 210” for 1 hour after the addition of 0.5 ml of 70%
perchloric acid. They were then cooled and mixed with 2.5 ml
of a solution containing 0.4yo ammonium molybdate and 0.2 ml
of 8 x 10e3
M
aminonaphthalene sulfonic acid. After heating
for an additional hour at loo”, the samples were cooled, brought
to 5 ml with HzO, and the optical density determined at 820 nm.
By this method, levels as low as 1.4 pg of phosphate per sample
were readily determined. Free and bound fatty acids were de-
termined by gas chromatography according to the method of
Ways and Hanahan (39).
Purification of Bovine Factor VIII-Nine liters of bovine
blood were collected directly into plastic containers containing 1
liter of 0.10
M
oxalate, 0.10 RI benzamidine-HCl,l and 20,000
1 The beneficial effect of benzamidine-HCI on the yield of Fac-
tor VIII was first observed by Dr. A. Thompson in this laboratory.
units of heparin. Mixing of the blood with the anticoagulant
was facilitated by pouring the blood into a second container and
into a 20-liter polyethylene vessel. Care was taken so that
not more than 1 min elapsed between the death of the animal and
the collection of the blood. The blood was transported to the
laboratory (usually less than 1 hour) at ambient temperature
and centrifuged at 0” for 75 min at 2,500
x
g. The plasma was
decanted and stirred slowly with BaS04 (100 mg per ml) for 15
min at 4”. In this step, the BaS04 adsorbs prothrombin and
Factors VII, IX, and X. The suspension was centrifuged at 0”
for 15 min at 2,500 x g, and the supernatant was treated with
kaolin (100 mg per ml) for 15 min to remove the major portion
of the contact factors (Factors XI and XII). The plasma was
then centrifuged as above and stored at -80” in 2-liter plastic
containers. Under these conditions, the BaSOc-kaolin-adsorbed
plasma can be stored for months without losing Factor VIII
activity.
For further purification, 4 liters of deep-frozen BaSO1-kaolin-
adsorbed plasma were thawed overnight at room temperature.
After about 12 hours, a small amount of cryoprecipitate was ap-
parent and some ice still remained. The melting plasma was
stirred for an additional 30 min and adjusted to a pH of 6.3 with
a KHzPOI solution saturated at room temperature. This usually
amounts to about 150 ml of phosphate solution per preparation.
Ethanol (loo’% precooled to -90”) was slowly added with stir-
ring to give a final concentration of 5%, and the plasma which
still contained some ice was poured into plastic centrifugation
cups and placed in the centrifuge at -3” for 10 min prior to
centrifugation. The solution was centrifuged at 2,500 x g for
10 min at -3”, and the precipitate containing Factor VIII was
suspended in 800 ml of 0.02 in imidazole-HCl buffer, pH 6.5, at
0” for 5 min. The washed precipitate was centrifuged at 2,500
x g for 10 min at 0” and dissolved in 200 ml of 0.25
M
NaCl solu-
tion containing 0.02
M
imidazole-HCI, pH 6.5, and 0.02
M
e-amino
caproic acid by gently stirring at room temperature for about 30
min. Bentonite solution (40 ml) (100 mg per ml in 0.15
M
NaCl)
was then added slowly with stirring. After an additional 10 min,
the bentonite was removed by centrifugation at 10,000 x g at
22” for 10 min. Twenty per cent polyethylene glycol in 0.25
M
NaCl-0.02
M
imidazole, pH 6.5, was added slowly to the super-
natant to make the bentonite-adsorbed supernatant 4% with
respect to polyethylene glycol, and the solution was left for 10
min at room temperature after the addition of the last drop of
polyethylene glycol. A precipitate was removed by centrifuga-
tion at 4,000 x g for 10 min, and this contained most of the Fac-
tor VIII activity. Centrifugation at this lower speed provided a
pellet which was more easily dissolved than if spun at high speed.
The precipitate was redissolved in 50 ml (one-eightieth of
original volume) of 0.05
M
citrate, pH 6.8, containing 0.02
M E-
amino caproic acid by warming in a water bath at 37” under
slight agitation. The final volume of the solution was measured
and an equal volume of 3
M
fi-alanine solution in 0.05
M
citrate,
pH 6.8, was added dropwise with stirring. With the last few
milliliters of fl-alanine solution, a precipitate begins to form.
The solution again was left for 10 min at room temperature and
centrifuged at 10,000 X g for 10 min at 22”. The resulting
precipitate containing the Factor VIII activity was redissolved
in 40 ml (one-hundredth of original volume) of 0.05
M
citrate
and 0.02
M
e-amino caproic acid at 37” and further purified by
chromatography at room temperature on a calcium citrate-cel-
lulose column. In this procedure, 20 g of calcium citrate were
by guest, on July 18, 2011www.jbc.orgDownloaded from
Issue of April 25, 1972 G. Xchmer, E. P. Kirby, D. C. Teller, and E. W. Davie 2515
mixed with 20 g of cellulose in 0.05
M
citrate, pH 6.8, and poured
into a siliconized column (2.5
x
30 cm). After the column had
settled, the solution from the @alanine step was applied to the
column. The column was extensively washed with the same
citrate buffer until the extinction of the effluent was less than
0.03. This usually required about 1 liter of citrate buffer.
AfOer the citrate wash, the column was placed in the cold room
at 4” and the Factor VIII activity was eluted slowly at 4’ by a
solution containing 0.1
M
ethylenedinitrilo-tetraacetic acid tetra-
sodium salt, 0.2 RI Tris-HCl, and 0.02
M
e-amino caproic acid ad-
justed to a final pH of 8.6. The elution rate did not exceed 30 ml
per hour.
The eluate containing the Factor VIII activity was mixed with
an equal volume of 20% polyethylene glycol in 0.25
M
NaCl aud
0.02 31 imidazole buffer, pH 6.5, at 4”. After standing in an ice
water bath for 30 min, a small precipitate containing all the Fac-
tor VIII activity was centrifuged at 0” at 37,000
x
g for 10 min.
For final purification, the precipitate was dissolved in approxi-
mately 10 ml of 0.2
M
KCl-0.02
M
Tris-HCl, pH 7.4, at 37”. The
absorption at 280 nm was determined on a O.l-ml aliquot which
had been diluted with 0.8 ml of t.he same buffer. The resulting
optical density was multiplied by one-half the volume of the Fac-
tor VIII solution. This is equivalent to the volume of con-
canavalin A solution (9.0 O.D. units per ml; i.e. 7.9 mg per ml in
0.15
M
NaCI) to be added. The concanavalin A was added with
stirring at room temperature and the solution was left to stand
for 15 min. The flocculent precipitate consisting of a concana-
valin A-glycoprotein complex was centrifuged at 10,000
x
g for
5 min at 22”. The precipitate was then dissolved in 2 ml of
0.25
M
NaCl-0.02 Y imidazole, pH 6.5, containing 10% methyl-
a-n-glucopyranoside in a 37’ water bath. One milliliter of the
Factor VIII-concanavalin A solution was layered on an Agarose
A-15m column (1.5 X 100 cm column, Pharmacia) at room tem-
perature and elution was performed with the buffer containing
10% methyl-cr-n-glucopyranoside in addition to 0.02% sodium
azide to inhibit bacterial growth. The rate of elution was 15 ml
per hour and 5-ml samples were collected. The first protein
peak which appeared after 50 ml contained Factor VIII activity.
The latter third of this peak was discarded since it is contam-
inated with the second peak. This preparation can be stored
for months at 4” with little or no loss in activity. Factor VIII
activity also can be precipitated by the addition of polyethylene
glycol at a final concentration of 5% and leaving the mixture at
0” for 30 min. A small precipitate of Factor VIII is formed and
this is centrifuged at 20,000
x
g for 10 min at 0”. This prepara-
tion can be stored for months at -70” without appreciable loss
of activity.
During the early studies, the purification was carried out in 3
days by stopping after the kaolin step on the 1st day and the cal-
cium citrate step on the 2nd day. These preparations of Factor
VIII were prepared from 4-liter batches of plasma. More re-
cently, 16 liters of plasma have been employed routinely through-
out the whole procedure and the final yield has been 25 to 30 mg
of Factor VIII. This high yield (about 15 to 20%) is obtained
by continuing the purification procedure from the cryoprecipita-
tion up to the dissolving of the concanavalin A precipitate in 1
day. This requires about 18 hours of work, starting early in the
morning. The higher yield is presumably due to less proteolysis
at the early stages of purification and better recoveries from the
precipitates.
After each precipitation step, the Factor VIII precipitate was
slurried in a small volume of buffer and extracted with successive
washes of buffer until it completely dissolved. It was important
to dissolve all precipitates immediately, since on standing they
became more difficult to solubilize and recoveries were poor.
The Factor VIII purification procedure has been employed in
our laboratory at least 50 times during the past 2 years with only
a few failures. Once completely purified, the Factor VIII is
stable in solution in the cold room for several months, or may be
frozen without loss of activity.
RESULTS
Preparation
of
Bovine
Factor
VIII-The steps in the purifica-
tion procedure for bovine Factor VIII and the yield from a typi-
cal preparation are shown in Table I. The procedure involves
BaSO*, kaolin, and bentonite adsorption for the removal of pro-
thrombin, Factors VII, IX, and X, and kaolin and bentonite
adsorption for the removal of fibrinogen and the contact factors
(Factors XI and XII).
The removal of fibrinogen from Factor VIII preparations has
always been a major problem in the purification of this protein.
In the present procedure, fibrinogen is removed at three different
steps (see Table I). The calcium citrate-cellulose column re-
moves the last portion of the fibrinogen and is one of the most
effective purification steps in the entire procedure. The use of
calcium citrate-cellulose was originally described by Blomback
et
al. (40) as a batch procedure for the purification of human Fac-
tor VIII, but in a later publication, these authors reported less
success with this method (41). The poor yield which they ob-
tained may have been due in part to a substantial decrease in pH
Purification step
Plasma .........................
BaS04-kaolin. ..................
Ethanol. .......................
Bentonite
......................
Polyethylene glycol.. ...........
@-Alanine .......................
Calcium citrate-cellulose ........
Concanavalin A. ...............
Agarose
.......................
Purijication 0,
Protein
:oncentration Volume
w/ml ml w u?&s/mg
70 4,400 308,000 0.014
59 4,000 238,000 0.0178
55 200 11,000 0.28
15.7 240 3,770 0.70
15.7 100 1,570 1.4
15.7 50 785 2.8
15.6 30 47 28
13 2 26 42
0.2 15 3.1 140
Specific
activity
TABLE
I
f Factor VIII from bovine plasma
Total protein Total
units I
?ercentage
of yield Purification Percentage
If fibrinogen
4,400 100 1 100
4,000 90 1.2 40
3,080 70 20 40
2,640 60 50 10
2,200 50 100 10
2,200 50 200 10
1,320 30 2,000 0
1,100 25 3,000 0
440 10 10,000 0
by guest, on July 18, 2011www.jbc.orgDownloaded from
2516
Bovine Factor VIII
Vol. 247, No. 8
20 40 60 80 100 120 140
Column Effluent (m/l
-
160
FIG. 1. Gel filtration of partially purified Factor VIII on Aga-
rose A-50m (top panel) and Agarose A-15m (bottom panel). A l-ml
solution of the Factor VIII preparation (10 mg per ml in 0.25
M
NaCl-0.02
M
imidazole buffer, pH 6.5, and 10% methyl-cr-n-gluco-
pyranoside) from the concanavalin A step was applied to each
column (1.5 X 90 cm) as described under “Methods.” O--O,
Factor VIII activity; A--A, protein concentration as measured
by the ninhydrin method.
400
1
32 I
4
64 320
20 40 60 80 100 120
400
240
160
!40
I60
80
Column Effhent (ml)
FIG. 2. Gel filtration of the purified Factor VIII on Agarose
A-50m (top
panel)
and agarose A-15m (bottom
panel).
A l-ml
solution of the Factor VIII preparation (6 mg per ml in 0.25 M
NaCl-0.02 M imidazole buffer, pH 6.5, and 10% methyl-cu-n-gluco-
pyranoside) from each of the gel filtration steps shown in Fig. 1
was applied to a second corresponding Agarose A-50m or Agarose
A-15m column (1.5 X 90 cm). In these experiments, the carbo-
hydrate analysis was carried out as described under “Methods”
after extensive dialysis of each sample against 0.25 M NaCl and
0.025 M imidazole buffer, pH 6.5, to remove free methyl-cu-n-gluco-
pyranoside. 0-e) Factor VIII activity; A---A, protein
concentration; m--U, carbohydrate concentration.
TUBE NUMBER
60
FIG. 3. Density gradient centrifugation of Factor VIII. Fac-
tor VIII (0.7 mg in 0.2 ml of 0.2 M sodium perchlorate and 0.05 M
borate buffer, pH 8.0) was layered on a gradient of 10 to 35%
methyl-ol-n-glucopyranoside and centrifuged as described under
“Methods.” A--A, protein;
l
- - -0, Factor VIII activity.
which occurs during protein elution from calcium citrate when it
is carried out at low buffer concentrations. This problem was
avoided by employing 0.20 M Tris buffer, pH 8.6, in the present
experiments.
The use of concanavalin A in the purification of human Factor
VIII was first described by Kass et al. (10) but abandoned in a
purification procedure which was published later (42). In
preliminary experiments, concanavalin A was covalently linked
to Sepharose 2B by the cyanogen bromide method of Porath
et aZ. (43). With this procedure, optimal conditions for binding
and elution of bovine Factor VIII were established by column
chromatography experiments. Subsequently, free concanavalin
A was used in a batch process since this gave better recoveries.
The concanavalin A step also removed the remainder of the Fac-
tor V which was present in trace amounts after the calcium cit-
rate-cellulose step.
The contaminants present after the concanavalin A precip-
itation step are removed by gel filtration on Agarose. A typical
elution profile for gel filtration on Agarose is shown in Fig. 1.
The top portion of the figure shows the results with Agarose A-50
m, and the bottom portion of the figure shows the results with
Agarose A-15m. In these experiments, Factor VIII appears in
the first protein peak from both columns, followed by a second
contaminating peak which includes concanavalin A. Gel filtra-
tion of the first protein peak from Agarose A-15m or A-50m
shown in Fig. 1 gives a single symmetrical peak (Fig. 2). In
these experiments, a close correlation between protein, Factor
VIII activity, and carbohydrate was observed. These experi-
ments suggest that bovine Factor VIII is a high molecular weight
glycoprotein, and the final purification step yields a preparation
which is homogeneous by the criteria of gel filtration.
This preparation is also free of all other known clotting factors.
This was shown by testing 0.1 ml of an undiluted Factor VIII
solution (1 mg of protein per ml, approximately 140 units of
Factor VIII per ml) for fibrinogen, prothrombin, and Factors V,
VII, IX, X, XI, and XIII. It was also free of detectable plas-
minogen.
by guest, on July 18, 2011www.jbc.orgDownloaded from
Issue of April 25, 1972 G. Schmer, E. P. Kirby, D. C. Teller, and E. W. Davie 2517
One of the major reasons for the loss of Factor VIII activity in
plasma is its destruction by proteolytic enzymes (44). The
present procedure employs blood collected in oxalate, heparin,
and benzamidine to limit proteolysis. The addition of heparin
substantially improves the yield of Factor VIII apparently by
further inhibiting the coagulation process. Thrombin and other
serine proteases, such as plasmin, are also inhibited by benzami-
dine (45)) and the addition of this inhibitor increases the over-all
yield to 10 to 15y0 for Factor VIII. In the absence of benzami-
dine and heparin, the final yield is less than 5%.
Another major loss of Factor VIII occurs by adsorption of the
protein to various surfaces, particularly at the higher stages of
purification. This was especially evident during the calcium
citrate-cellulose column chromatography and the final gel filtra-
tion step on Agarose. In each of these steps, there was a 40 to
60% loss in Factor VIII activity. Losses by glass adsorption,
however, can be avoided to a major extent by employing silicone-
coated glassware during purification and storage.
Density
Gradient Centrifugation
of Bovine Factor VIII-Density
gradjent centrifugation was carried out on Factor VIII to provide
further evidence for the purity of this clotting factor preparation.
In these studies, the protein was centrifuged through a methyl-
cu-n-glucopyranoside gradient of 10 to 35% (w/v) in the presence
of 0.05
M
borate buffer, pH 8.0, and 0.2
M
sodium perchlorate to
minimize aggregation (46). Centrifugation was carried out at
room temperature since aggregation and precipitation of Factor
VIII occur at 4” even in the presence of a strong chaotropic agent
like perchlorate. As shown in Fig. 3, a close parallelism is ob-
served between activity and protein concentration throughout
the gradient.
Zone Electrophoresis-Factor VIII was then subjected to elec-
trophoresis at pH 8.6, employing microscope slides layered with
Agarose-nonpolymerized acrylamide (Fig. 4). After electro-
phoresis for 45 min, a single spot was observed for Factor VIII
(Sample 1). Upon reduction with 2-mercaptoethanol in 8
M
urea (Sample 2) or reduction and subsequent alkylation with
iodoacetate
(Sample
S), single spots were also obtained. The
distance of migration for the reduced and alkylated fractions is
increased relative to the native molecule. Also, the addition of
carbosymethyl groups by alkylation slightly increases the mobil-
ity of this fraction relative to the reduced form.
Immunoelectrophoresis-As another criterion for purity, Fac-
tor VIII was subjected to electrophoresis on Agarose-acrylamide
slides, followed by immunodiffusion against rabbit antibody pre-
pared from a highly purified Factor VIII preparation. In these
experiments, Factor VIII migrates toward the anode during
electrophoresis and forms a single sharp precipitin line after the
addition of antibody to the center trough (Fig. 5). Three dif-
ferent Factor VIII preparations are shown and each shows only a
single precipitin line.
Immunoelectrophoresis experiments were also carried out with
an antibody prepared from a partially purified Factor VIII prep-
aration. This sample contained about 10 to 20% Factor VIII
and was obtained from the calcium citrate-cellulose column step.
In these experiments shown in Fig. 6, the top well contained a
purified Factor VIII preparation, and the bottom well contained
Factor VIII purified only through the calcium citrate-cellulose
column step. It is clear from these studies that the cruder Fac-
tor VIII preparation shows many precipitin lines, and these are
reduced to only one after the final purification step.
The antibody employed in these experiments readily neutral-
1
anode 2 cathode
3
T
starting well
FIG.
4. Zone electrophoresis of Factor VIII, reduced Factor
VIII, and reduced and alkylated Factor VIII. Factor VIII (20
~1, containing 12~g of protein) was applied to each of the wells and
electrophoresis was carried out in 0.05
M
sodium barbital buffer,
pH 8.6, for 45 min at room temperature with a current of 5 ma per
slide and 150 volts. Samples were stained with Amido black.
Xample 1 contained Factor VIII,
Sample
2 contained reduced Fac-
tor VIII, and SamDle 3 contained reduced and alkvlated Factor
VIII.
ized Factor VIII activity as measured in the regular Factor VIII
assay (Table II). In these experiments, the antibody and Factor
VIII were previously incubated together and aliquots were as-
sayed in Factor VIII-deficient plasma. The neutralization of
Factor VIII activity at the higher antibody concentration was
greater than 90%. Similar results were obtained with the anti-
body prepared from the partially purified Factor VIII prepara-
tion.
Polyacrylamide Gel Electrophoresis
of
Bovine Factor VIII--In
preliminary studies, it was observed that bovine Factor VIII
does not enter a 3.25% polyacrylamide gel even in the presence
of 8
M
urea. After reduction of Factor VIII with 2-mercapto-
ethanol, however, a single protein and a single carbohydrate
These experiments support the conclusion that the single pro-
tein precipitin line observed in the immunoelectrophoresis ex-
periments (Figs. 5 and 6) is due to the presence of Factor VIII.
band are obtained following polyacrylamide gel electrophoresis
Furthermore, they provide strong evidence for the purity of the
bovine Factor VIII preparation.
(Fig. 7). In these experiments, the protein was detected with
Amido black and the carbohydrate with the periodic acid-Schiff
reagent. These results suggest that Factor VIII is made of sub-
units held together by disulfide bonds, and these subunits are
very similar in size or are perhaps identical.
Sedimentation Equilibrium Studies of Bovine Factor VIII-
Sedimentation equilibrium results demonstrated that Factor
VIII was physically heterogeneous both in the native and reduced
states (Table III). For the native protein, the two most concen-
trated samples (0.50 and 0.75 mg per ml) appeared to be in chem-
ical equilibrium, but the most dilute sample (0.25 mg per ml)
displayed much higher molecular weights. This result could
arise from either computation errors or a lack of chemical equi-
librium among all species (26). This data caused the rather
large standard deviations of :ICr, and J/I, shown in Table III.
The smallest species which could be calculated under these condi-
tions has a molecular weight in excess of lo6 g per mole. In the
presence of 8
M
urea, the molecular weight was also greater than
lo6 g per mole. In addition, a single sedimentation velocity ex-
periment using absorption optics displayed physical hetero-
geneity as seen by some spreading of the boundary.
In an attempt to determine subunit molecular weights, Factor
VIII was centrifuged to equilibrium in both 8
M
urea and 6
M
guanidine-HCl containing 2-mercaptoethanol. The experiment
in urea showed poor precision and the molecular weights ob-
by guest, on July 18, 2011www.jbc.orgDownloaded from
2518 Bovine Factor VIII Vol. 247, No. 8
FIG. 5 (top left). Immunoelectrophoresis of three different highly
purified Factor VIII preparations. For these experiments, Factor
VIII antibody was prepared against a highly purified Factor VIII
nrenaration. Factor VIII solution (20 ~1) (approximatelv 0.6 ma
I I
per ml) was placed in each of the wells and^electrophoresis was
carried out in 0.05
M
sodium barbital buffer, pH 8.6, for 45 min at
room temperature with a current of 5 ma per slide and 150 volts.
Following electrophoresis, 50 ~1 of rabbit Factor VIII antibody
solution (1.45 mg per ml) were added to the center trough. Photo-
graphs were taken after 24 hours at room temperature. The
anode is at the left of the photograph The wells in the slide on
the top contained samples from Factor VIII preparation 1, and
the slide on the bottom contained samples from Factor VIII prep-
arations 2 and 3.
FIG.
6
(bottom left).
Immunoelectrophoresis of a highly purified
and a partially purified Factor VIII preparation. In these ex-
served were functions of the initial concentration. This behavior
indicated that the protein was incompletely dissociated in this
solvent. The results for the preparation denatured in 6
M
guani-
dine-HCI and reduced in Z-mercaptoethanol are shown in Table
III. The protein was still heterogeneous, but the results show
good precision, indicating nearly complete dissociation of the
molecule under these conditions. In these experiments, there
was no discernible concentration dependence of the molecular
weight averages. From this data, it would appear that the sub-
periments, Factor VIII antibody was prepared against a partially
purified Factor VIII preparation obtained from the calcium cit-
rate-cellulose column. Electrophoresis was carried out as de-
scribed in the legend to Fig. 5. Following electrophoresis, 50 ~1
of
rabbit Factor VIII antibody (8 mg per ml) were added to the
center trough and photographs were taken after 72 hours at room
temperature. The anode is at the left of the photograph. The
well on the top of the slide contained the highly purified Factor
VIII, and the well on the bottom of the slide contained the partially
purified Factor VIII.
FIG. 7 (right). Polyacrylamide gel electrophoresis of reduced
bovine Factor VIII. Protein (25 pg) was applied to each gel as
described under “Methods.” The gel on the left was stained for
protein with Amido black and the gel on the right was stained for
carbohydrate with the periodic acid-Schiff reagent.
units of Factor VIII may not be identical and the smallest species
has a molecular weight of about 85,000.
It should be emphasized that these molecular weight deter-
minations are preliminary. Thus,
the data in Table III should
be taken as a general indication of the size of Factor VIII and
its subunits rather than the exact molecular weights for this pro-
tein and its subunits.
Lipid Content of Bovine Factor VIII-Hershgold et al. (3) re-
ported a lipid content of human Factor VIII of 11%. Also,
by guest, on July 18, 2011www.jbc.orgDownloaded from
Issue of April 25, 1972
G. Xchmer, E. P. Kirby, D.
TABLE
II
Neutralization of Factor VIII activity by Factor VZZZ antibody
Samples (10 ml) of Factor VIII (3 rg per ml) were previously
incubated for 30 min at 37” with 0.1 ml of rabbit antibody pre-
pared against the highly purified Factor VIII. Factor VIII
assays were then carried out as described under “Methods.”
Additions
None.
Antibody (0.37 pg).
Antibody (0.73 rg).
Antibody (1.45 pg).
Antibody (2.9 pg)
Antibody (14.5 pg). .
.
.
I
Clotting time Percentage of
inhibition
s
74.7, 75.6 0
77.8, 78.4 20
83.2, 83.9 42
93.0, 92.8 66
99.6, 98.6 74
122.8 93
C. Teller,
and
E. W. Davie 2519
nique, bovine Factor VIII was found to contain less than O.O04oj,
fatty acids.
Bovine Factor VIII was also tested for the presence of phos-
pholipid by phosphate analysis of a chloroform-methanol extract
of the protein. With 5-mg samples of Factor VIII, no phosphate
was found by this procedure, which readily detects levels as low
as 1.4 pg per sample. If one assumes that phosphate makes up
approximately 10% of a typical phospholipid molecule, then bo-
vine Factor VIII contains less than 1% phospholipid by this
analysis.
Carbohydrate
and Amino Acid Composition-The hexose, sialic
acid, and hexosamine composition of bovine Factor VIII is shown
in Table IV. Factor VIII contains approximately 11% hexose,
2% sialic acid, and 7% acetylhexosamine. Thus, the total car-
bohydrate level for Factor VIII is 20%. This table also
TABLE
III
Molecular weight of bovine Factor VIII by sedimentation equilibrium
Sedimentation equilibrium studies were carried out on native and denatured and reduced Factor VIII as described under “Methods.”
Molecular weights (X lo-3 g/mole)
Sample rPm 5
Mn MIX MS MP
ml/&-
Nativeb
4,027 0.728 2,180 f 170 3,100 f 660 4,700 f 2,400 1,150 f 160
Denatured and reduced in Z-mer-
captoethanolc.. 13,922 0.718
143 zt 3 169 f 5
198 f 5
85 f 5
a Ml refers to the smallest species which could be detected calculated by the methods described by Teller et al. (28). The uncer-
tainties are weighted root mean square values in this column, but standard deviations for all other values.
6 Buffer: 0.01 M Tris-HCl-0.2 M KCl, pH 7.5.
c Buffer 0.04 M Tris-HCl-6 M guanidine-HCl-0.2 M
2-mercaptoethanol, pH 7.5.
TABLE IV
Carbohydrate
content
of bovine Factor VIII
Carbohydrate and amino acid analyses were carried out on three
different preparations of bovine Factor VIII as described under
“Methods.”
Amino acidsa.. 100
80.4
Hexoseb. 13.8 11.1 61.6
Sialic acidc..
2.1 1.7 5.7
Hexosamined
8.5 6.8 32.2
0 Calculated from data from the amino acid analyzer.
b Calculated as glucose.
c Calculated as Xacetylneuraminic acid.
d Calculated as 9-acetylgalactosamine.
these authors state that their human Factor VIII preparation is
destroyed by phospholipase C, but activated 3-fold by phospholi-
pase D (47). Accordingly, it was of interest to examine bovine
Factor VIII for the presence of lipid. In these experiments, 5-
mg samples of bovine Factor VIII were analyzed for fatty acids
by the gas chromatographic method of Ways and Hanahan (39).
This method readily detects less than 0.1 /.~g of free and esterified
fatty acids which might be present in the protein. The procedure
involves a chloroform-methanol extraction followed by esterifica-
tion with methanol prior to gas chromatography. By this tech-
shows the approximate number of sugar residues per 100,000 g of
glycoprotein. These calculations indicate that there is about
twice as much hexose present in Factor VIII as compared to
hexosamine.
The amino acid composition for bovine Factor VIII is shown in
Table V. The levels of tryptophan and tyrosine in this protein
are lower than that found in many other plasma proteins. These
low levels account for the decreased absorption at 278 nm for
Factor VIII as compared to bovine fibrinogen. This was evident
from the 278:260 nm absorption ratios for the two proteins in
dilute phosphate buffer at neutral pH. This value is 1.7 for
fibrinogen and 1.3 for Factor VIII.
Stability of PurifLed Bovine Factor
VIII-A rather surprising
property of the highly purified Factor VIII preparation is its
stability. Preparations stored in 0.25
M
NaCl-0.02
M
imidazole
buffer, pH 6.5, and 10% methyl-ar-n-glucopyranoside or even
in 0.05
M
borate, pH 8.0, 0.2
M
perchlorate and 25% methyl-c-
n-glucopyranoside at 4” do not show any appreciable loss of
activity after 2 months. One preparation has been stored at 4” in
silicone tubes for 6 months without showing any detectable
loss
of activity. This stability is observed even in the absence
of a protease inhibitor and does not require the addition of meth-
yl-a-n-glucopyranoside.
The highly purified Factor VIII is also rather stable in the
general physiological pH range. For instance, incubation at
room temperature for 15 min in the pH range of 6 to 8.5 resulted
in no
loss of activity. However, at further pH extremes (pH
5.0 or 9.0), a rapid decrease of Factor VIII activity occurred.
by guest, on July 18, 2011www.jbc.orgDownloaded from
2520 Bovine Factor VIII Vol. 247, No. 8
TABLE V
Amino acid analysis of bovine Factor VZZZ
Amino acid analysis of Factor VIII was carried out as described
under “Methods.” Results are expressed in terms of protein
rather than glycoprotein.
Amino acid
Lysine ......................
Histidine
...................
Arginine
...................
Aspartic acid ...............
Threonine
.................
Serine ......................
Glutamic acid. .............
Proline. ....................
Glycine. ....................
Alanine .....................
Half-cystine. ...............
Valine ......................
Methionine .................
Isoleucine
..................
Leucine
....................
Tyrosine
Phenylalanine
..............
Tryptophan
...............
as/l00 P&T
protein
6.89
3.06
7.10
9.80
4.44
4.49
10.90
4.86
3.91
3.48
5.95
9.84
1.30
4.12
10.20
2.61
3.97
3.26
Residues/lOs g
protein
53.8
22.3
45.5
85.1
43.8
51.5
84.4
50.2
68.5
49.0
57.6
99.3
11.3
36.4
90.3
16.0
26.9
17.5
TABLE VI
Adsorption of Factor VZZZ on glass test tubes
A factor VIII solution (70 pg per ml) containing 10 units per ml
in 0.25 M NaCl-0.02 M imidazole buffer, pH 6.5, was diluted 40-
fold with 0.15 M NaCl-0.02 M imidazole buffer, pH 6.5, in silicon-
iced tubes. Duplicate samples (0.1 ml) were transferred into
Pyrex tubes and incubated 5 min at 0”. The Factor VIII solu-
tions then were transferred with a siliconized micropipette into
second, third, and fourth glass tubes after the same incubation
time. After incubation in the fourth tubes, the residual Factor
VIII activity was transferred into siliconized glass tubes and the
initial and final Factor VIII activity was measured by the partial
thromboplastin time (duplicate
samples
1 and 6). The nonsili-
conized Pyrex tubes (duplicate samples 2,3,4, and 5) were rinsed
repeatedly with 0.15 M NaCl and the adsorbed Factor VIII activ-
ity was measured after adding Factor VIII-deficient plasma, Ca2+
phospholipid, and kaolin. The total units of Factor VIII were
then estimated from a standard curve.
Conditions
Factor VIII stock solution
Factor VIII bound to the first
test tube
Factor VIII bound to the second
test tube
Factor VIII bound to the third
test tube
Factor VIII bound to the fourth
test tube
Factor VIII remaining in solu-
tion after glass adsorption
Buffer control (no Factor VIII)
Clotting time
s
70, 70
90, 92
94, 92
120, 125
147, 150
80, 82
300, 270
Activity
uaits
0.025
0.007
0.005
0.001
0
0.014
0
As mentioned earlier, Factor VIII is readily adsorbed to sur-
faces such as glass. When bound to glass, it also retains its
biological activity as shown in Table VI. In these experiments,
an aliquot from a Factor VIII stock solution was transferred
into a Pyrex glass tube, incubated for 5 min, and then trans-
ferred to a second, third, and fourth Pyrex tube after the same
incubation period. After thorough rinsing of each of the Pyrex
tubes, residual Factor VIII activity bound to the test tubes was
determined.
As seen in Table VI, about half of the original activity is ad-
sorbed to the first two glass tubes and little is adsorbed to the
third and fourth tubes. When solutions containing higher
levels of Factor VIII are incubated with Pyrex glass tubes under
similar conditions, even larger amounts of Factor VIII are ad-
sorbed and the partial thromboplastin time of added hemophilic
plasma in these tubes is readily corrected to normal levels. Thus
far, we have been unsuccessful in removing Factor VIII in a
native form from the glass surface to utilize this property as a
purification step.
DISCUSSION
The present data indicate that bovine Factor VIII is a glyco-
protein present in plasma at a concentration of about 7 pg per
ml. The purification procedu.re described results in a prepara-
tion purified approximately lO,OOO-fold. Establishing the purity
of a large molecule such as Factor VIII is difficult, especially
when the protein tends to absorb nonspecifically to surfaces such
as glass, cellulose, or Sephadex derivatives. Nevertheless, we
were able to demonstrate that the final product is homogeneous
when examined by gel filtration, density gradient centrifugation,
zone electrophoresis, and immunoelectrophoresis. When re-
duced with 2-mercaptoethanol, it shows a single protein and
carbohydrate band on polyacrylamide gel electrophoresis. The
preparation, however, is physically heterogeneous when sub-
jected to sedimentation equilibrium studies. In view of the
other evidence for purity, it appears probable that the hetero-
geneity observed in the sedimentation equilibrium experiments
is due to aggregation of a chemically pure species.
The presence of hexose, hexosamine, and sialic acid in bovine
Factor VIII is typical of plasma proteins. Preliminary results
employing gas chromatography have shown that the only two
hexoses present were mannose and galactose, and these were
found in a ratio of 1:3. Acetylgalactosamine was tentatively
identified as the galactosamine. Further studies will be re-
quired, however, to confirm these results.
In contrast to human Factor VIII (47), the bovine preparation
contains no lipid, as shown by the absence of fatty acids and
phospholipid. Factor VIII, however, requires phospholipid
during its interaction with factor IX, (48). Thus, it is not
surprising to find lipid in human preparations that are made by
procedures that do not include ethanol fractionation. The
destruction of human Factor VIII by phospholipase C (47),
however, suggests that the lipid may have an additional function
in the human preparation.
The present studies also show that bovine Factor VIII is a
large molecule with a molecular weight greater than one million.
The large size of human Factor VIII has been observed by others
by gel filtration studies in which the molecular weight was esti-
mated to be larger than two million (10, 49). In the present
experiments, there were no indications of Factor VIII activity
associated with molecules of smaller molecular weight. Indeed,
by guest, on July 18, 2011www.jbc.orgDownloaded from
Issue of April 25, 1972 G. Xchrner, E. P. Kirby, D. C. Tellw, and E. W. Davie
attempts to dissociate the molecule into subunits by urea, gua-
nidine-HCI, or chaotropic agents have been unsuccessful. In
the presence of a reducing agent, such as 2-mercaptoethanol,
the molecule is broken into subunits, the smallest species having
a molecular weight of about 85,000. This treatment which
breaks disulfide bonds completely destroys the biological activity
of Factor VIII.
Other workers (50-53) have reported that Factor VIII activity
can be markedly increased by incubation with trace amounts of
thrombin. This increase in activity induced by thrombin makes
it difficult to compare the present preparation of Factor VIII
with those of other investigators whose high specific activity
may be due in part to thrombin modification of Factor VIII.
The native bovine Factor VIII prepared by this procedure
contains about 140 units of Factor VIII per mg of protein. In-
cubation with traces of highly purified thrombir? causes a 50-fold
increase in apparent Factor VIII activity with a final specific
activity approaching 7,000 units per mg. This increase in ac-
tivity is as large or larger than that seen by other workers and
suggests that the activity of our native preparation has not
been made spuriously high by inadvertent exposure to traces
of thrombin during purification.
Bovine Factor VIII also reacts with Factor IX, in the presence
of phospholipid and calcium to form a complex capable of acti-
vating Factor X (54). A manuscript describing the details of
these experiments is in preparation.
Acknowledgments-We wish to express our sincere thanks to
Drs. Arthur Thompson, Robert Meyer, and Richard Counts for
valuable discussions and assistance, and to Mark Legaz who
performed the zone and immunoelectrophoresis experiments.
Thanks are also due to Richard Olsgaard, Charles Nicholas,
Barry Stewart, and Richard Cox for excellent technical assist-
ance at various stages of this project. We are also indebted to
the Cudahy Company and Auburn Packing Company for kindly
providing the bovine plasma employed in these studies.
REFERENCES
1.
B.YUMGAHTEN, W., SANDERS, B. E., BELKIN, B. D., PAGEN-
KPMPER, F. E., ALBERS, W. G., AND CIMINERA, J. L. (1963)
Thromb. Diath. Haemorrh. 9, 354
2. KEKWICK, It. A., AND WOLF, P. (1957) Lancet 1, 647
3. HERSHGOLD, E. J., DAVISON, A. M., AND JANSZEN, M. E. (1971)
J. Lab. Clin. Med. 77, 185
4. HURT, J. P., WBGNER, R. H., AND BRINKHOUS, K. M. (1966)
Thromb. Diath. Haemorrh. 16, 327
5. BIDWELL, E. (1955) Brit. J. Haematol. 1, 386-389
6. BIDWELL, E. (1955) Brit. J. Haematol. 1, 35-45
7. HERSHGOLD, E. J., POOL, J. G., AND PAPPENHAGEN, A. R.
(1966) J. Lab. Clin. Med. 67, 23
8. SOULIER, J.-P. (1959) Pathol. Biol. 7, 2451
9. JOHNSON, A. J., NEWMAN, J., HOWELL, M. B., AND PUSZICIN,
S.
(1967) Thromb. Diath. Haemorrh. 26, 377
10. KASS, L., RATNOFF, 0. D., AND LEON, M. A. (1969) J. Clin.
Invest. 48, 351-358
11. SCHMER,
G.,
AND DAVIE, E. W. (1969) Amer. Chem. Sot.
Monogr. Ser. 168, 188
12. AGRAWAL, B. B. L., AND GOLDSTEIN, I. J. (1965) Biochem. J.
96, 23C-25C
13. MOORE, S. (1968) J. Biol. Chcm. 243, 6281-6283
14. DUBOIS, M., GILLES, K. A., HAMILTON, J. K., REBERS, P. A.,
AND SMITH, F. (1956) Anal. Chem. 28, 35&356
15. GARDELL, S. (1957) in D. GLICK (Editor), Methods of biochem-
ica2 analysis, Vol. VI, p. 289, Interscience Publications,
Inc., New York
2 A. Thompson and E. Kirby, unpublished results.
16.
17.
18.
19.
20.
21.
22.
23.
24.
25.
26.
27.
28.
29.
30.
31.
32.
33.
34.
35.
36.
37.
38.
39.
40.
41.
42.
43.
44.
45.
46.
47.
48.
49.
50.
51.
W.IRREN, L. (1959) J. Biol. Chem. 234, 1971-1975
JEANLOZ, R. W. (1966) in A. GOTTSCH~LIC (Editor), Glyco-
proteins, their composition, structure, and function, Vol. 5,
p. 362, BBA Library, American Elsevier Publishing Co.,
New York
DAVIS, B. J. (1964) Ann. N. Y. Acad. Xci. 121, 404
ZXH~RIUS, R. M., ZELL, T. E., MORRISON, J. H., AND WOOD-
LOCI<, J. J. (1969) Anal. Biochem. 30, 148
WILLUMS, C. A., AND CHASE, M. W. (1971) Methods in im-
munology and immunochemistry, Vol. III, p. 103, Academic
Press, New York
SCHEIDEGGER, J. J. (1955) Znt. Arch. Allergy Appl. Immunol.
7, 103
MOORE,
S.,
AND STEIN, W. H. (1963) Methods Enzymol. 6, 819
SPACI~MAN, D. H., STEIN, W. H., AND MOORE, S. (1958) Anal.
Chem. 30, 1190
HIRS, C. H. W. (1967) Methods Enzymol. 11, 59
BENCZE,
W. L.,
AND SCHMID, K. (1957) Anal. Chem. 29, 1193-
1196
HARRIS, C. E., KOBES, R. D., TELLER, D. C., AND RUTTER,
W. J. (1969) Biochemistry, 8, 2442-2454
SEERY, V. L., FISCHER, E. H., AND TILLER, D. C. (1970)
Biochemistry 9, 3591-3598
TELLER, D.
C:,
HORBETT, T. A., RICHARDS, E. G., AND SCHACH-
M.~N. H. K. (1969) Ann. N. Y. Acad. Sci. 164. 66-101
MCME~KIN, T: L.,‘GRovEs, M. L., AND HIPP; N. J. (1949)
J. Amer. Chem. Sot. 71, 3298
LONGSWORTH, L. G. (1953) J. Amer. Chem. Sot. 76, 5705
REISLER, E., AND EISENBERG, H. (1969) Biochemistry 8, 4572-
4578
R.~TNOFF, 0. D., AND DAVIE, E. W. (1962) Biochemistry 1, 967-
974
LANGDELL, R. D., WAGNER, R. H., AND BRINKHOUS, K. M.
(1953) J. Lab. Clin. Med. 41. 637
PROCTOR, R. R., AND RAPAPORT, S. I. (1961) Amer. J. Clin.
Path. 36, 212
JACOBSSON, K. (1955) Stand. J. C&n. Lab. Invest. 7, Suppl.
14, 7
CLAUS, V. A. (1957) Acta Haematologica 17, 237
BIGGS. R.. AND MACFARLANE. R. G. (1962) Human blood coaau-
Zati&a and its disorders, Ed: 3, p. 3g9, F.‘A. Davis Co., Phiin-
delphia
QUICK, A. J., STANLEY-BROWN, M., AND BANCROFT, F. W.
(1935) Amer. J. Med. Sci. 190, 501
WAYS, P., .&ND HANAHAN, D. J. (1964) J. Lipid Res. 6,318-328
BLOMHBCI~. B.. BLOMB~~CK. M.. AND NILSSON. I. M. (1958)
Acta Chem. kand. 12, 18?8 ’ . ~I
BLOMB~CK, B., BLOMBSCIC, M., .~ND STRUWE, I. (1961) Ark.
Kemi 18, 223
RSTNOFF, 0. D., KASS, L., A~XD LANG, P. D. (1969) J. CZin.
Invest. 48, 957
PORATH, J., AXON, R., AND ERNBACK, S. (1967) Nature 216,1491-
1492
KEKWICK, R. A., AND WALTON, P. L. (1962) Nature 194,878-879
MARI~WARDT. F.. LANDMANN. H.. AND WALSMANN. P. (1968)
Eur. J. Bi&hek. 6, 502 ’ ’ , > ~,
H~TEFI, Y., AND HANSTEIN, W. G. (1969) Proc. Nat. Acad. Sci.
U. S. A. 62, 1129
HERSHGOLD, E. J., DAVISON, A. M., AND JANSZEN, M. E. (1971)
J. Lab. Clin. Med. 77, 206
LUNDBLAD, R. L., AND DAVIE, E. W. (1964) Biochemistry 3,
1720-1725
HERSHGOLD, E. J., SILVERMAN, L., DAVISON, A., AND JANSZEN,
M. (1967) Fed. Proc. 26, 488
RAPAPORT, S. I., SCHIFFMAN, S., PATCH, M. J., AND AMES,
S. B. (1963) Blood 21, 221
RSPAPORT, S. I., HJORT, P. F., AND PATCH, M. J. (1965) &and.
J. C&in. Lab. Invest. 17, Suppl. 84, 88
52. OZGE-ANWAR, A. H., CONNELL, G. E., AND MUSTARD, J. F.
(1965) Blood 26, 500
53. BIGGS, R., MACFARLANE, R. G., DENSON, K. W. E., AND ASH,
B. J. (1965) Brit. J. Haematol. 11,276
54. FUJIKAWA, K., THOMPSON, A. R., AND LEGAZ, M. E. (1971)
Fed. Proc. 30, 1075
by guest, on July 18, 2011www.jbc.orgDownloaded from
