THE JOURNAL OF BIOLOGICAL CHEMISTRY

Val. 258, No. 20, Issue of October 25, pp. 12548-12552,1983

Printed in U.S. A.

Isolation of Human Serum Spreading Factor*

(Received for publication, June 7, 1983)

David W. Barnes and Janet Silnutzer

From the Department of Biological Sciences, University of Pittsburgh, Pittsburgh, Pennsylvania15260

Serum spreading factor (SF) was isolated from hu-
man serum by a four-step procedure employing affinity
chromatography on glass beads, concanavalin A-Se-
pharose, DEAE-agarose, and heparin-agarose. The fi-
nal product was purified approximately 260-fold from
thestarting material and was maximally active in
assays of cell spreading-promoting activity at 300 ng/
ml. The isolated human SF preparation consisted of
two proteins of apparent molecular weights approxi-
mately 65,000 (SF65) and 75,000 (SF75). Both SF65
and SF75 have been shown previously to exhibit cell
identified the factor in soluble extracts of washed human
platelets (Barnes et al., 1983).
Holmes (1967) originally described a spreading-promoting
activity present in a serum fraction isolated by glass bead
affinity chromatography. This fraction contains both spread-
ing-promoting and growth-promoting activity when assayed
on cells plated in serum-free media without additional sup-
plementation. The spreading-promoting activity in this par-
tially purified preparation, which was found to be separable
chromatographicallyfromthegrowth-promotingactivity
(Barnes and Sato,1980a; Barnes et al., 1981), is the material

Downloaded from http://www.jbc.org/ by guest on November 12, 2017

spreading-promoting activity and to bind monoclonal
antibody to human serum SF. that we have termed “serum spreading factor” (Barnes and
Sato, 1980a, 1980b, 1980c; Barnes et al., 1980). In this paper
we describe the isolation of serum SF (SF65 and SF75)using
as a first step glass bead affinity chromatography, followed
SF’ is aglycoprotein component of humanserumthat by chromatographyonconcanavalin A-Sepharose, DEAE-
promotes the attachment and spreading of a wide variety of agarose, and heparin-agarose. Because serum SF exhibits a
both fibroblastic and epithelialcells in vitro (Barnes and Sato, strong tendency to adsorb to the surfaces of glass or plastic
1979; Barnes et al., 1980,1981,1982a). In studies using serum-vessels or dialysis tubing, theprocedure wasdeveloped in such
free hormone-supplemented cell culture medium, Barnes and a way that dialysis or other manipulations between chromat-
Sat0 (1979, 1980a, 1980b, 1 9 8 0 ~ )observed that this factor ographic steps were unnecessary. Serum SF was eluted a t each
mediates cell spreading-promoting effects that in some cases step in this procedure in a solution thatallowed direct loading
are not mimicked by fibronectin or cold-insoluble globulin of the material onto the next column in the series, and the
(HynesandYamada, 1982), anotherspreading-promoting isolated material was obtained at the endof the procedure in
proteininserum.SubsequentlyBarnes et al. (1980, 1981, a solution of approximately physiological pH and salt concen-
1982a, 1982b, 1983) reported that serum SF is distinct im- tration. This approach resulted in a purification procedure
munologically, biochemically and inbiological activities from that canbe carried out fairly rapidly with reasonable yields.
fibronectin,laminin,andchondronectin, allglycoproteins
capable of stimulating attachment and spreading of some EXPERIMENTALPROCEDURES
types of cells in culture (Kleinmanet al. 1981), and that serum Materials-Glass beads (Number 1014, Class IV-A) were obtained
SF contributes a significant portion of the total spreading- from Ferro, Cataphote Division, Jackson, MS. Chemicals and bio-
promoting activity of human serum in serum-supplemented chemicals were obtained from Fisher and Sigma. Materials for gel
culture medium. In addition to effects on cell attachment, chromatography and SDS-polyacrylamide gel electrophoresis were
obtained from Bio-Rad and Pharmacia Fine Chemicals. Immuno-
spreading, and gross morphology, serum SF also stimulates chemicals were obtained from N. L. Cappel Laboratories Inc. Human
growth and affects differentiationof a number of cell types in plasma fibronectin (cold-insoluble globulin) was obtained from Meloy
culture in hormone-supplemented serum-free media (Barnes Laboratories Inc. and was judged to be greater than 98% pure upon
andSato, 1980b, 1980c; Barnes et al., 1980, 1981, 1982a). analysis by SDS-polyacrylamide gel electrophoresis.
Experiments using monoclonal antibody that inhibits serum Monoclonal Antibody-binding Assay for Measurement of Serum
SF-promoted cell spreading established that serumSF exists SF-We have previously described procedures for measurement of
serum SF by this assay, as well as the derivation, characterization,
in two forms in human serum, oneform migrating in polyac- and isolation of the monoclonal antibody to serum SF used in the
rylamide gel electrophoresis in a manner consistent with M, assay (Barnes et al., 1983). The standard assay procedures are based
= 65,000-70,000 (SF65) and the other in a manner consistent in general on the methods of Rennard et al. (1980). Briefly, a fixed
with M , = 75,000-78,000 (SF75) (Barnes et al., 1982b, 1983). amount of monoclonal antibody (0.2 pg) in 100 pl of PBS containing
Experiments withmonoclonal antibodytoserum SF also 1 mg/ml of bovine serum albumin was incubated 16 h with samples
” (100 pl) to be assayed for serum SF, diluted to various concentrations
* This work was supported by American Cancer Society Grant BC- in PBS with 1 mg/ml of bovine serum albumin. The amount of
368 and National Institutes of Health Grant CA-35214-01. The costs antibody remaining free at the end of this incubation was determined
of publication of this article were defrayed in part by the payment of by transfer of the incubation mixtures to serum SF-coated microtiter
page charges. This article must therefore be hereby marked “adver- wells and subsequent measurements by standard ELISA, using per-
tisement” in accordance with 18 U.S.C. Section 1734 solely to indicate oxidase-conjugated goat anti-mouse IgG as the second antibody and
this fact. 2,2’-azino-di(3-ethylbenzthiazelinesulfoni~ acid) as the peroxidase-
The abbreviations used are: SF, spreading factor; SF65, form of dependent chromogen (Barnes et al., 1983). Extent of the enzymatic
serum SF of molecular weight approximately 65,000; SF75, form of reaction was determined by measuring the absorbance of the peroxi-
serum SF of molecular weight approximately 75,000; PBS, phosphate- dase reaction mixture at 415 nm. In this assay, samples containing
buffered saline; ELISA, enzyme-linked immunosorbent assay; SDS, relatively high amounts of serum SF bound a relatively larger amount
sodium dodecyl sulfate. of anti-SF and thus contained less free anti-SF at the end of the
 Isolation ofSpreading
Serum
Human
incubation period. ELISA of these samples gave rise to lower absorb-
ance at 415 nm at the endof subsequent incubations than ELISA of
samples containing relatively less serum S F in the initial incubation
mixtures. One unit of anti-SF antibody-binding activity was defined
as the amountof activity in the initialincubation mixture necessary
to bind half of the total anti-SF antibody in the standard assay as
described above. Protein concentration was measured by the method
of Bradford (1976) using bovine y-globulin as a standard.
Cell Spreading Assay-Procedures for assay of spreading-promot-
ing activity were as described previously (Barnes et al., 1980, 1983).
Briefly, cells were plated in serum-free medium containing 1 mg/ml
of bovine serum albumin onto tissue culture dishes, either pretreated
with the samples to be assayed (1-h incubations at room temperature)
or ontocontrol plates preincubated with PBS. The number of spread
cells was determined 90 min after plating. For the assay presented,
Hela cells were used; we also have used other epithelioid and fibro-
blastic cell types in this assay. One unit of cell spreading-promoting
activity was defined as the amount of activity necessary to elicit an
effect that was 50% of the maximum effect under the conditions of
the assay.
Purification of Human Serum SF-Fresh frozen human plasma
was obtainedfrom thePittsburghCentral Blood Bank, dialyzed
overnight against 0.15 M NaCl, and clotted by the addition of 1 mg/
ml of CaC12. The resulting serum was stored frozen at -20 “C in 25-
ml aliquots. Serum S F levels in serum do not differ markedly from
those in plasma (Barnes et aL, 1983), and we have used serum as the
starting material in our isolation procedures. Both SF65 and SF75
are present in serum and plasma. Human serum S F was purified in a
four-step procedure described below.
For step 1, serum (25 ml) was adjusted to pH 8.0 with 1 N NaOH
and chromatographed on a glass bead column (2.5 X 40 cm; 50 ml/h)
previously equilibrated with 0.6 M NaHC03 (pH 8.0). This and all
subsequent procedures were carried out at 4 “C. Fraction volumes of
1-3 ml, depending onthe expected sharpness of the peak, were
collected in polypropylene tubes. Serum on the column was followed
by 100 ml of 0.6 M NaHC03, 50 ml of 0.6 M NaHC03, 0.2 M Na&O,
(pH 9.3), 50 mlof water, 50 mlof 0.6 M KHCO,, 0.2 (pH 9.5),
and 100 ml of 0.2 M K2C03 (pH11.0). Elution with 0.2 M K2C03 (pH
11.0) allowed regeneration of the column for reuse. Elution of protein
was followed by measurement of absorbance at 280 nm, and peak
fractions eluting with 0.6 M NaHCO,, 0.2 M Na2C03were pooled. For
step 2, the pooled peaks from four to six glass bead column runs were
combined and chromatographed directly on a concanavalin A-Se-
pharose column (2.5 X 9 cm; 15 ml/h). Following sample application,
the column was eluted sequentially with 90 ml of 25 mM NaH2P0,/
Na2HP04 (Na/P,) at pH 6.0, 75 ml of 25 mM Na/Pi (pH 6.0) contain-
ing 50 mM mannose, and 25 ml of PBS containing1 M a-methylman-
noside. A total of approximately 75mgof protein could be loaded
onto a concanavalin A-Sepharose column (2.5 X 9 cm) under the
conditions described without exceeding the glycoprotein-binding ca-
pacity of the column.
Serum S F was also present in the peak eluting from the glass bead
column with 0.6 M KHC03, 0.2 M K,CO,. This material could be
chromatographed in a manner identical with that described for the
peak eluted with 0.6 M NaHC03, 0.2 M Na2C03;however, we found
that betteryields were obtained if the peak eluted fromthe glass bead
column with KHC03/K2C03was dialyzed against PBSbefore loading
onto the concanavalin A column. This is because the glycoprotein-
binding capacity of the lectin column was reduced significantly if the
sample was loaded in the presence of KHC03/K,C03, compared to
the capacity of the column if the sample was loaded in NaHCO,/
Na2C0,. Approximately 70% of the serum spreading factor retained
by the glass beads was recovered in the peak eluting with 0.6 M
NaHC03, 0.2 M Na2C03. Theremaining portion of the serum spread-
ing factor eluted from the column in the later portion of the 0.6 M
NaHC03, 0.2 M Na2C0, wash (about 10% of the total retained) and
in the 0.6 M KHCOs, 0.2 K&03 wash (about 20% of the total). No
serum spreading factor was recovered in the water wash, but inclusion
of this stepallowed a sharp demarcation between the sodium carbon-
ate and potassium carbonate washes that was useful in subsequent
processing of the fractions for concanavalin A-Sepharose chromatog-
raphy. Although all of the bound serum spreading factor could be
eluted eventually with extensive washing with 0.6 M NaHC03, 0.2 M
Na2C03,changing to potassium carbonate after an initial elution with
sodium carbonate allowed the recovery of the remaining portion of
the serum spreading factorin reasonably concentrated peak fractions.
Elution of serum spreading factor with potassium carbonate after
elution with sodium carbonate was first reported by Holmes (1967),
Factor 12549
and the chemical basis for this phenomenon is unclear, although the
relative chaotropic natures of sodium carbonate and potassium car-
bonate in solution may be involved.
In step 3, fractions constituting a broad peak eluting from the
concanavalin A-Sepharose column with 50 mM mannose in 25 mM
Na/P,(pH 6.0) were pooled and chromatographeddirectlyona
DEAE-agarose column (1.5 X 5 cm; 15 ml/h) previously equilibrated
with 25 mM Na/Pi (pH 6.0). After loading the sample, the column
was eluted with 20mlof25 mM Na/Pi (pH 6.0), 20 ml of 100 mM
NaCl in 25 mM Na/Pi (pH 6.0), and 20 ml of 1.5 M NaCl in 25 mM
Na/Pi (pH 6.0). In step 4,fractions containing material eluting from
the DEAE-agarose column in a sharp peak with 100 mM NaCl in 25
mM Na/P, (pH 6.0) were pooled and loaded directly onto a heparin-
agarose column (0.7 X 14 cm; 7.5 ml/h) previously equilibrated with
100 mM NaCl in 25 mM Na/Pi (pH6.0). The sample was followed on
the column by 15 ml of 100 mM NaCl in 25 mM Na/Pi (pH 6.0), 15
ml of 100 mM NaCl in 50 mM Na/Pi (pH 7.2), and 10 ml of 1.5 M
NaCl in 50 mM Na/P, (pH 7.2). Most of the serum SF initially bound
to the column eluted in a sharp peak with 100 mM NaCl in 50 mM
Na/P, at pH7.2 (“step 4” material). A relatively small amount of SF
remained bound to thecolumn under these conditions and was eluted
with 1.5 M NaCl in 50 mM Na/Pi (pH 7.2).
Analysis by SDS-polyacrylamide gel electrophoresis indicated that
the step 4 materialconsisted of proteins migrating in two bands
corresponding to molecular weights of approximately 65,000 (SF65)
and 75,000 (SF75). Protein staining of overloaded gels indicated that
Downloaded from http://www.jbc.org/ by guest on November 12, 2017
some preparations appeared to contain only SF65 and SF75. Other
preparations,althoughgreater than 98%SF65 + SF75, contained
trace contaminants,primarily a proteinof apparent molecular weight
approximately 50,000. These preparations could be freed of trace
contaminants by direct gel filtration chromatography of the step 4
material.
RESULTS
Isolation of Serum SF-Table I summarizes the relative
specific activities calculated from the antibody-binding and
cell spreading-promoting assays, total amount of protein re-
covered, and the percentageof total activity recovered at each
step in the isolationprocedure. Analysisby SDS-polyacrylam-
ide gel electrophoresis of the product obtained at each step is
shown in Fig. 1. The greatest purification in a single step was
obtained in theglass bead affinity column (step 1)procedure.
Although in previous work we have used partially purified
serum SF isolated by glass affinity chromatography asa first
step (Barnes et al., 1980, 1981), the ratio of serum volume to
glass bead column volume was about four times higher in
these earlier procedures than that of the step 1 procedure
described here. In attempts to maximize total yields, we found
that best results were obtained if the amount of serum chro-
matographed, aswell as thevolume of the elutingbuffers, was
decreased from those of our earliermethods. The adhesion of
serum proteins to glass represents rather complicated inter-
actions (Haas and Culp,1979), and the amount of serum
protein loaded on the glass beadcolumn affects to some extent
which proteinspredominate in thefractions subsequently
eluted.’
When step 1 material was passed through a concanavalin
A-Sepharose column (step 2), over half of the total protein
passed throughthe columnessentially unretarded.These
proteins were primarily of molecular weight less than 30,000
with some additional material of molecular weight approxi-
mately 88,000. Serum SF was eluted from the convanavalin
A-Sepharose column with 50 mM mannose in pH 6.0 buffer.
Elution with mannose at 50 mM gave more consistent results
with fewer contaminating proteins present than did elution
with a-methylmannoside at50 mM or less. Elution of serum
SF from the lectin columnat pH 6.0 in 25 mM Na/Pi allowed
direct loading of the eluted material onto a DEAE-agarose
column (step 3). Serum SF was eluted from the column by
D. W. Barnes, L. Mousetis, B. Amos, and J. Silnutzer, submitted
for publication.
12550 Isolation of Serum
Human Spreading Factor
raising the salt concentration(100 mM NaCl in 25 mM Na/Pi
(pH 6.0)).
Elution of serum SF from DEAE-agarose in the indicated
buffer solution allowed direct application of the eluted mate-
rial onto a heparin-agarose column (step 4). Serum SF was
TABLEI
Isolation of human serum spreading factor
Procedures for assay of specific monoclonal antibody-binding ac-
tivity (immunoassay) and cell spreading-promoting activity are de-
scribed under “ExDerimental Procedures.” FRACTION

Specific activity
Cell Total Activity
Purification stage Immuno- spread-
protein recove@
assaj ing
assayb
unitsfpg mg %
Human serum 87500.0620.012 Starting
material
Glass bead column (step 4571.8
3.12
0.661
1)
Concanavalin A-Sephar- 1.66 NDd 30 19.0
ose (step 2) FRACTION
DEAE-agarose (steu 3) 2.21 ND 4.59 10 FIG.2. Heparin-agarose column chromatography of iso-

Downloaded from http://www.jbc.org/ by guest on November 12, 2017

Heparin-agarose ( s b p 4) 19.63.23 2.48 8 lated human fibronectin ( A )and human serum spreading fac-
Anti-SF monoclonal antibody-binding assayas illustrated in Fig. tor ( B ) .Methods are described under “Experimental Procedures”
3. and “Results.” Arrows indicate change of elution conditions: Fraction
* Cell spreading-promoting assay as illustrated in Fig. 4. 17, change from pH 6.0 to 7.2 (100 mM NaCl, 50 mM Na/Pi); Fraction
Calculated from specific activity determinedby immunoassay. 42, change from 100 mM NaCl to 1.5 M NaCl (50 mM Na/Pi, pH 7.2).
ND. not determined.
subsequently elutedfrom the heparin column by adjusting the
eluting solution to 100 mM NaCl and 50 mM Na/Pi (pH 7.2).
The heparin column procedure removed several contaminat-
ing proteins from the preparation, particularly a protein of
molecular weight approximately 50,000 that co-purified with
serum SF through the first three steps, butremained bound
to the heparin column under conditions that eluted most of
the bound serum SF.
Like serum SF, fibronectin also bindsto heparin (Stathakis
and Mossesson, 1977; Yamada et al., 1980), although this
I binding is maintained under conditions that elute serum SF
from the column. This difference between the two spreading-
promoting serum proteins is illustrated in the experiment of
Fig. 2. Human serum SF (step 4 material) and human fibro-
nectin were loaded onto identical heparin-agarose columns

W il (0.7 X 2.5 cm) in 50 mM Na/Pi (pH 6.0) containing 100 mM

ii
a b c d e f g
FIG.1. SDS-polyacrylamide gel electrophoresis of human
serum SF preparations. The four-step method for isolation of
serum SF is described under “Experimental Procedures.” Samples of
material isolated a t each step were reduced with mercaptoethanol,
denatured a t 100 “C,and subjected to electrophoresis using a 7.5%
polyacrylamide slab gel with a 2.5% stacking gel. a, molecular weight
standards: 5 pg each of (top to bottom) myosin (200,000), @-galacto-
sidase (116,000), phosphorylase b (92,500), bovine serum albumin
(66,700), and ovalbumin (43,000); b, 50 pg of human serum; c, 40 pg
of preparation from glass bead column affinity chromatography (step
1); d, 20 pg of preparation from concanavalin A-Sepharose chroma-
tography (step 2); e, 20 pg of preparation from DEAE-agarose chro-
matography (step 3);f , 15 pg of isolated human S F (SF65 and SF75)
from heparin-agarose chromatography (step 4); g, molecular weight
standards asin a above.
NaC1. The columns were eluted with 15 ml of 100 mM NaCl
in 50 mM Na/Pi (pH 6.0), 20 ml of 100 mM NaCl in 50 mM
Na/Pi (pH7.2), and 15 ml of 1.5 M NaCl in 50 mM Na/Pi (pH
7.2). Although serum SF eluted from the column under con-
ditions of approximate physiological salt concentration and
pH, humanfibronectin remained tightly bound to the column
under these conditions and was eluted with 1.5 M NaCl, as
previously reported by others (Stathakis and Mossesson, 1977;
Yamada et al., 1980).
Total yield in thefour-step isolation procedure from 1unit
of human plasma was approximately 2.5 mgof serum SF,
representing an overall recovery of about 8% of the total
activity inthe startingmaterial. The step4 material consisted
of a mixture of SF65 and SF75 in about equal proportions.
The total yield and relative amounts of the serum SF forms
varied somewhat among different preparations, the totalyield
varying roughly in relation to the concentrationof serum SF
in plasma. Our measurements3 indicate that serum SF con-
centration in normal adults ranges from about 100 pg/ml to
about 400 pg/ml, with average concentration approximately
M. C. Shaffer, E. D. Avner, T. P. Foley, and D. W. Barnes,
submitted for publication.
 Isolation of Human Sejw m Spreading Factor 12551

200 pg/ml. Immunoblots of the final preparations indicated

that monoclonal antibody to serum SF recognized both the
SF65 and SF75 bands; no fibronectin was detected in the step
4 material assayed by ELISA using rabbit antiserumto human
fibronectin (not shown). SDS electrophoresis of the isolated
serum SF preparations in 12 and 15% polyacrylamide gels did
not detect additional lower molecular weight protein bands
that might not have been detected in the 7.5% acrylamide
gels of Fig. 1.
Amino acid compositions of two independently isolated step
4 serum SF preparations are given in Table 11. Also included
in Table I1 for comparisons are the amino acid compositions
of a spreading-promoting factorof M , = 62,000 isolated from PROTEIN ( p q l m l 1
fetal calf serum (Whateley and Knox, 1980) and a partially

.,
FIG.3. Monoclonal antibody-binding assay for measure-
purified spreading-promoting preparation isolated from fetal ment of serum SF. Methods are described under “Experimental
calf serum that contains some components in the molecular Procedures.” 0, isolated human SF (SF65 and SF75) from heparin-
weight range from 65,000 to 80,000 (Grinnell et al., 1977). agarose column chromatography (step 4); X, human SF preparation
The amino acid compositions of the two preparations of serum from glass bead column chromatography (step 1); Cohn fraction
spreadlng factor were quite similarand considerably different IV from human plasma; 0, human serum.
from those of the other two spreading factor preparations, I I I I 1 I
1
particularly in relative percentage of tyrosine, proline, and
arginine. NHz-terminal analysis of isolated serum spreading

Downloaded from http://www.jbc.org/ by guest on November 12, 2017

factor identified aspartic acid (asparagine) as thesingle NHz-
terminal amino acid in the preparation. The NHP-terminal
amino acid of the spreading factor of Whateley and Knox
(1980) is reported to be alanine.
Activity of Isolated Serum SF-Measurement of the relative
anti-serum SF monoclonal antibody binding activity (Barnes I I
I
I

et al., 1983) for human serum, Cohn fraction IV from human

plasma, a step 1 preparation eluted from the glass bead
column, and step 4 material eluted from the heparin-agarose
j 40
W I I
I
I

I
I
TABLEI1
Amino acid composition
Amino acid composition of human serum spreading factor was
determined on a Durrum D-500analyzer after hydrolysis of samples
in 6 N HCl for 24 h at 110 “C. Values shown are from single deter-
0
I% .01 0.I
c“-.-
1.0
/
/
/I
e
10
I I
I00
P R O T E I N (pqlrnl)
minations of two different serum spreading factor preparations (A
and B). Hydrolytic losses of serine and threonine were not corrected. FIG.4. Assay of cell spreading-promoting activity of iso-
Values for the fetal calf serum spreading factors of Whateley and lated human SF (0)and human serum (0).Methods are described
Knox (1980) andthe mixed factor of Grinnell et al. (1977) are under “Experimental Procedures.” Hela human carcinoma cells were
calculated from data in the publications cited. used in the assay.
Serum Spreading
Serum
factor of Spreading column is shown in Fig. 3. Cohn fraction IVwas enriched
spreading spreading factor of
Amino acid factor Whateley approximately %fold in serum SF over human serum; other
factor prep- Grinnell et
aration A prepara- and Knox
al. (1977) Cohn fractions tested containedless serum SF/mg of protein
tion B (1980) than did human serum. The protein concentration of step 4
mo1f103 mol amino acid serum SF isolated from the heparin-agarose column necessary
Aspartic acid 107 110 85.8 105 to reduce the amount of free antibody in the assay mixture
Threonine 38.6 50.1 37.8 76.0 by 50% was approximately 260-fold lower than the protein
Serine 62.9 73.7 61.0 71.6
Glutamic acid 149 127 144 113 concentration of human serumnecessary to produce the same
Proline 78.7 49.1 83.9 58.6 effect in the assay. Similarly, the step4 serum SF preparations
Glycine 82.6 94.8 85.0 73.8 were about 300-fold more active in promoting cell spreading
Alanine 63.4 75.6 67.1 56.5 in culture than was the starting material (Fig. 4). Maximal
Half-cystine 11.7 31.9 19.5 28.6 effect of the isolated serum SF preparations on cell spreading
Valine 52.0 64.0 51.1 78.3 was seen at about 300 ng/35 mm-diameter plate (8 cm’), and
Methionine 8.76
6.69 11.2 0
Isoleucine 29.6 32.0 28.4 40.9 a detectable effect was observed at 10 ng/plate.
Leucine 65.3 81.3 61.2 76.6
Tyrosine 46.2 25.2 46.6 33.5 DISCUSSION
Phenylalanine 49.8 39.2 49.8 36.8 Using anti-serum SF monoclonal antibody-binding activity
Histidine 23.3 48.3 20.3 24.0 and cell spreading-promoting activity in serum-free culture
Lysine 53.8 64.2 47.4 46.9
Arginine 77.1 75.7 40.8 47.2 medium asan assay, we have purified human serum SF
Tryptophan ND“ ND ND 32.2 approximately 260-fold from human serum by afour-step
Ornithine 0 0 10.5 0 procedure. The greatest purification in asingle step occurred
in the initial glass bead affinity column chromatography.
Total 1000 1000 1000 1000 Analysis by SDS-polyacrylamide gel electrophoresis indicated
ND,not determined. that thefinal productof the procedure was a mixture of SF65
12552 Isolation of Serum
Human Spreading Factor
and SF75. We have reported previously that both SF65 and protein of M,= 62,000, somewhat smaller thanour SF65, and
SF75 are active independently in assays of monoclonal anti- contains no material comparable to SF75 in our preparations.
body-binding and cell spreading-promoting activity (Barnes Furthermore, comparisons of amino acidcomposition and
et al., 1982a, 1983; Hayman et al., 1982). Yields of serum SF NH,-terminal point out significant biochemical differences
isolated by the proceduredescribedfrom 1 unit of human between human serum spreading and thetwo cell spreading-
plasma were sufficient to provide about 2.5 mg of material promoting preparationsfrom fetal calf serum.
maximally active in thebiological assay at 0.3 pg/ml. Epibolin is a M, = 65,000 protein isolated from human
Although serum S F exhibits some similaritiestoserum serum (Stenn, 1981) that promotes spreading of epidermal
fibronectin or cold-insoluble globulin (Hynes and Yamada, cells in culture. Spreading-promoting activity of epibolin in
1982) in biological properties and heparin-binding activity, the concentration rangeeffective for serum spreading factor
the two glycoproteins are biochemically and immunologically ( i e . 1 pg/ml or less) requires the presenceof a second factor
distinct by several criteria (Barnes et al., 1980, 1981, 1982a, synergistic with epibolin andtermed co-epibolin (Stenn,
1983). Serum SF does not bind to gelatin-Sepharose (Barnes 1981). In collaboration with Dr. Stenn, we have observed
et al., 1980), and we have made serum SF preparations by some immunological cross-reactivity between epibolin prepa-
procedures described in this paper using serumderived from rations and serum spreading factor preparations (Barnes et
human plasma that had been passed over gelatin-Sepharose al., 1983). Our serum spreading factor preparations appear to
to remove fibronectin, and found no reductionin yield of be unique, however, inthepresence of two forms of the
serum SF. Including the gelatin-Sepharose step in thebegin- spreading-promoting factor (SF65 and SF75) and absence the
ning allowed the isolation of both of these spreading-promot- of requirement for a synergistic cofactor. More detailed bio-
ing proteins independently from the same plasma sample. chemical comparisons of epibolin and serum spreading factor
Furthermore, because plasma fibronectin and serum SF eluted are inprogress.
from heparin-agarose under different conditions, the combi-

Downloaded from http://www.jbc.org/ by guest on November 12, 2017

nation of gelatin-Sepharose and heparin-agarose chromatog- Acknowledgments-We thank Dr. M. Sussman for suggestions and
raphy makes it unlikely that SF preparations contained con- M. Shaffer for help and advice. We also thank Dr. C. Coffee for NHZ-
taminatingfibronectin or fibronectinfragments.We have terminal amino acid analysis of isolated serum spreading factor.
confirmed this by ELISA for fibronectin under conditions
that would detect less than 0.05% fibronectin contamination REFERENCES
in our preparations.Our data indicate that, unlike fibronectinBarnes, D., and Sato, G. (1979) Nature (Lond.)281, 388-389
binding to heparin (Stathakis andMossesson, 1977; Yamada Barnes, D., and Sato, G. (1980a) in Cell Biology of Breast Cancer
et al., 1980), strong interaction
between heparin and the major (McGrath, C. M., Brennan, M. J., and Rich, M. A., eds) pp. 277-
portion of the serum SF initially bound to the column at pH 287, Academic Press, New York
Barnes, D., and Sato, G. (1980b) Anal. Biochem. 102, 255-270
6.0 did not occur at physiological pH and salt concentration. Barnes, D., and Sato, G. (1980~)Cell 2 2 , 649-655
This observation suggests that the heparin-binding phenom- Barnes, D., Wolfe, R., Serrero, G., McClure, D., and Sato, G. (1980)
enon that we found useful in devising an isolation procedure J. Suprarnol. Struct. 1 4 , 47-63
for human serum SFmay not reflect an interactionof impor- Barnes, D. W., van der Bosch, J., Masui, H., Miyazaki, K., and Sato,
tance between plasma or tissue SF and heparin or heparin G. (1981) Methods Enzyrnol. 79, 368-391
proteoglycans in vivo. Barnes, D. W., Darmon, M., and Orly, J. (1982a) Cold Spring Harbor
Conf. Cell Proliferation 9, 155-167
Recently Hayman et al. (1983) also have observed binding Barnes, D., Silnutzer, J., See, C., and Shaffer,M. (1982b) J. Cell Biol.
of serum spreading factor to heparin-agarose and reported 95,119a
that monoclonal antibody identifies serum spreading factor Barnes, D. W., Silnutzer, J., See, C., and Shaffer, M. (1983) Proc.
at the cell surface of cultured fibroblasts and in tissues. In Natl. Acad. Sci. U. S. A. 80, 1362-1366
ourexperimentswiththe monoclonal antibodytoserum Bradford, M. (1976) Anal. Biochem. 72, 248-254
spreading factor used in this andprevious studies (Barnes et Grinnell,
175-190
F., Hays, D. G., and Minter, D. (1977) Exp. Cell Res. 110,
al. 198213, 1983), we havebeen unableto identifycell- or Haas, R., and Culp, L. A. (1979) J. Cell. Physiol. 101, 279-292
tissue-associated serum spreading factor in cultured human Hayman, E. G., Engvall, E., A’Hearn, E., Barnes, D., Pierschbacher,
cells or frozen sections of human kidney and liver.3 In the M., and Rouslahti, E. (1982) J. Cell Biol. 95, 20-23
latter studies, however, immunostaining with our antibody Hayman, E. G., Pierschbacher, M. D., Ohgren, Y., and Rouslahti, E.
clearly identified serum spreading factor in Bowman’s space (1983) Proc. Natl. Acad. Sci. U. S. A. 8 0 , 4003-4007
in kidney and sinusoidsof liver, both plasma-containing areas Holmes, R. (1967) J. Cell Bid. 32, 297-308
Hynes, R. O., and Yamada, K. M. (1982) J. Cell Biol. 95, 369-377
of these tissues. Kleinman, H. K., Klehe, R. J., and Martin, G. R. (1981) J. Cell Bid.
Cell spreading-promoting activity hasbeen reported previ- 88,473-485
ously in preparations containing proteins in the molecular Rennard, S. I., Berg, R., Martin, S. R., Foidart, J. M., and Robey, P.
weight range from 60,000 to 80,000 isolated from fetal calf G. (1980) Anal. Biochem. 1 0 4 , 205-214
serum(Grinnell et al., 1977; WhateleyandKnox, 1980). Stathakis, N. E., and Mosesson, M. W . (1977) J. Clin. Inuest. 60,
855-865
Although the “mixed factor” of Grinnell et al. contains several Stenn, K. S. (1981) Proc. Natl. Acad. Sci. U. S. A. 78,6907-6911
spreading-promoting molecules,including fibronectin,the Whateley, J. G., and Knox, P. (1980) Biochern. J . 1 8 5 , 349-354
spreading factor of Whateley and Knox is immunologically Yamada, K. M., Kennedy, D. W., Kimata, K., and Pratt,R. M. (1980)
distinct from fibronectin. This preparation contains a single J , Biol. Chern. 255,6055-6063
 Isolation of human serum spreading factor.
D W Barnes and J Silnutzer
J. Biol. Chem. 1983, 258:12548-12552.

Access the most updated version of this article at http://www.jbc.org/content/258/20/12548

Alerts:
• When this article is cited
• When a correction for this article is posted

Click here to choose from all of JBC's e-mail alerts

This article cites 0 references, 0 of which can be accessed free at

http://www.jbc.org/content/258/20/12548.full.html#ref-list-1

Downloaded from http://www.jbc.org/ by guest on November 12, 2017