1

THE 9E3/CEF4 CYTOKINE: KINETICS OF

SECRETION, PROCESSING BY PLASMIN, AND
INTERACTION WITH EXTRACELLULAR MATRIX

Manuela Martins-Green 1 , Mark Stoeckle2 , Annie Hampe3* ,

Sandelle Wimberly1 and Hidesaburo Hanafusa3

1Dept. Biol., Univ. Calif. Riverside, CA 92521, 2Dept. Medicine, N.Y. Hosp.-
Cornell Med. Cntr, N.Y., 3Lab. Molec. Oncol., Rockefeller Univ., N.Y.

Running Title : 9E3 Cytokine: Secretion/Processing/ECM binding

Key Words: Angiogenesis/chemokine/chicken/collagen/laminin.

Corresponding Author: M. Martins-Green

Department of Biology
University of California
Riverside, CA 92521
Tel: (909) 787-2585
Fax: (909) 787-4509
email: MMGreen@ucrac1.ucr.edu

*Current Address: Labo. Recomb. Génétique, Hôpital St-Louis, Paris,

 2

ABSTRACT
The chicken gene, 9E3/CEF4, is a small inducible cytokine highly homologous to human IL-8
and gro . It is overexpressed during wound healing and in the tissues around tumors induced by

Rous sarcoma virus. More is known about the expression of 9E3 in vivo than any other of the small

cytokines, yet little is known about its biochemical characteristics and functions. Here we report on

some of the biochemical properties of the 9E3 gene product, the kinetics of protein secretion, the post-

secretory processing of the protein, and on its association with ECM molecules. The protein: (i) Is

synthesized and secreted in < 10 min; (ii) is not glycosylated and does not bind heparin with high

affinity; (iii) is secreted as a 9 kDa form and is processed to a 6-7 kDa form by plasmin, an enzyme

released at wound sites and produced in association with tumors; (iv) in its smaller form binds to

interstitial collagen, laminin and to a lesser extent to proteoglycan, and does not bind to collagen IV or

fibronectin. This is the most rapidly secreted protein yet described in eukaryotic cells and is the first of

the small inducible cytokines to be found to associate with ECM molecules other than

glycosaminoglycans. Our results suggest that, given the appropriate stimulus, the level of the 9E3

cytokine could be elevated very rapidly, resulting in similarly rapid biological responses. The different

modes of availability of the two forms of the molecule suggest that the two isoforms may play different

roles in vivo.
 3

INTRODUCTION
Over the past 10-12 years, many genes have been isolated that encode for small cytokines

whose expression is tightly regulated in normal cells but induced to high levels when stimulated by

serum factors or agents that cause inflammation, and by wounding (1-7). The genes encoding for

these proteins are early response genes. Therefore, they do not require new protein synthesis for

induction and are regulated at the level of transcription (1-3,8).

Several of these inducible cytokines are secreted proteins that are chemotactic for specific kinds

of leukocytes, suggesting that they play a role in recruiting inflammatory cells to sites of injury. These

factors are very small proteins, most of them 6-14 kDa in size (5-14); they are evolutionarily conserved

and highly homologous to each other (5, 6, 15-17). Multiple forms are frequently observed (18,19)

and appear to be related to activation of the molecule. Structurally, they have 4 cysteine residues in

well-conserved regions that are thought to be important for their function. They belong to a

superfamily of chemokines that is subdivided into two families based on the position of the first two

cysteines (15,20) -- the CC family in which the cysteines are adjacent (e.g. RANTES; MCAF; LD-78;

ACT-2) and the CXC family (which includes 9E3) in which the cysteines are separated by any single
amino acid (other members include PF-4; βTG; IL-8; IP-10; gro/MGSA).

9E3/CEF4 is the only currently known nonmammalian member of this superfamily and shows
high protein sequence homology to human IL-8 (51%) and gro (45%). Studies using the Boyden

Chamber assay suggest that the 9E3 protein is chemotactic for periferal blood monocytes and

heterophils (21). This gene is highly expressed by Rous sarcoma virus (RSV)-transformed chick

embryo fibroblasts (T-CEFs) but also is expressed in normal chick embryo fibroblasts (N-CEFs)
when they are stimulated to grow by the addition of serum (1,4). TGFα, TGFβ, aFGF and bFGF

(growth factors that are known to be released by tumors and are known to be angiogenic in vivo; 22-

24), also stimulate the expression of the 9E3 gene (4).

In newly-hatched chicks, 9E3 is expressed at very low levels in normal tissues that grow and

remodel by cell division (e.g. connective tissue, tendon and bone), whereas tissues that do not grow

by cell division (e.g. muscle) do not express this gene (14,25). It is overexpressed during wound
 4

healing, especially in areas of neovascularization, suggesting a role in wound repair. Furthermore,

this protein is also overexpressed in the tissues surrounding RSV-induced tumors but the tumor cells

themselves do not express it (14,25). The overexpression around tumors and the high stimulation of

expression by growth factors that are known to be released by tumors and to be angiogenic in vivo,

point to possible additional roles in stimulating or inhibiting tumor growth.

Although more is known about the distribution of 9E3 expression in vivo than for any other of

these small cytokines, little is known about its biochemical characteristics and behavior. Here we

report some of the biochemical properties of the 9E3 gene product, delineate the kinetics of protein

secretion, show that the protein is processed from a 9 kDa form to a 6-7 kDa form by plasmin, and that

the latter form binds to some extracellular matrix molecules.

RESULTS
Specificity of the Antibody: In the work presented here we used a polyclonal antibody developed

against the entire 9E3 protein (minus the signal peptide) produced in bacteria. The specificity of the

antibody was determined by immunoprecipitation from cell extracts and culture media of metabolically-

labeled N-CEFs and CEFs transformed with the engineered form of Schmidt-Ruppin subgroup A of

RSV (henceforth designated XD2; 26). We found that this antibody specifically immunoprecipitates a

large amount of the 9 kDa form of the protein from the supernatant of T-CEFs (Fig. 1, ln 8) and a

barely detectable amount of the same protein from the supernatant of N-CEFs (Fig. 1, ln 6). A small

amount of protein was also immunoprecipitated from the cell lysate of T-CEFs (Fig. 1, ln 4), but

nothing was detected from the lysate of N-CEFs despite the fact that the cells were metabolically

labeled for 3 hrs (Fig. 1, ln 2). The high molecular weight proteins precipitated by the immune serum

were also precipitated by the pre-immune serum (Fig. 1, odd-numbered lanes) and therefore are non-

specific. This antibody also recognized the smaller form of the protein when it was present (e.g. in the

culture medium of Prague C transformed cells, see below).

Stability of the 9E3 Protein and Kinetics of Secretion: We metabolically-labeled CEFs with 35S

methionine and 35S cysteine to determine the stability and the kinetics of secretion of the 9E3 protein.
 5

A 15 min pulse followed by a chase from 5 min to 24 hrs showed that the protein leaves the cells very

rapidly with an abrupt drop between 5 and 10 min after the pulse is removed (Fig. 2A) and that it is

stable in the supernatant over a period of 24 hr (Fig. 2B).

The presence of a significant amount of the 9E3 protein in the supernatant within five minutes

of the beginning of the chase showed that a 15 min pulse was too long to determine the time it takes for

the 9E3 protein to be synthesized and secreted. Therefore, we repeated the experiments with a pulse of

5 min and chased for 30 min at 5 min intervals (Fig. 3). A very small amount of the protein was

detected in supernatant collected 5 min after the end of the pulse (Fig. 3B,D), and again an abrupt drop

off of the protein inside the CEFs was observed between 5 and 10 min after pulse termination (Fig.

3A,C). That is, after 5 min of pulse and 5 more of chase, the protein was still essentially entirely

retained within the cells, but a small amount had been secreted. After 10 min of chase, however, the

protein had virtually all disappeared from the cell extracts. Thus, secretion had begun in less than 10

minutes after the start of the pulse, and essentially finished within 10 minutes of the end of the pulse;

secretion occurred approximately as a square wave. Identical results were obtained with T-CEF

cultures (not shown).

Kinetics of Synthesis of the 9E3 Protein : The very rapid secretion of the protein raises the

possibility of very rapid biological response. As a consequence, we investigated the kinetics of protein

production after stimulation of quiescent cultures in which the 9E3 protein was initially below our

detection limit (Fig. 1, ln 2) and in which 9E3 mRNA was present in much less than 0.1% of the cells

(4). Cultures were stimulated to produce 9E3 by treatment with the phorbol ester TPA (27) in the

presence of 35S methionine and 35S cysteine. Immunoprecipitation of the cell extracts of these TPA-

stimulated CEFs at various times after stimulation demonstrated the presence of the 9E3 protein within

10 min of treatment with TPA and rapid rise thereafter for the period of an hour (Fig. 4).

Biochemical Properties of the 9E3 Cytokine: Our previous observation that two forms of the 9E3

protein exist (14) led us to investigate the possibility that the larger form of the protein is glycosylated.

Computer analysis of the predicted sequence showed that this protein does not contain potential N-
 6

glycosylation sites, but the possibility remained that the protein could have O-linked oligosaccharides.

We looked for the presence of glycosylation by probing western blots with lectins that bind to specific

sugars (Fig.5) and by using a glycan detection kit (not shown). In O-linked oligosaccharides, N-

acetylgalactosamine invariably attaches to the OH of a serine or a threonine and galactose to the OH of

lysines (28,29). Except for the ABO blood groups, these oligosaccharides are generally very short, 1-

4 sugar residues in length (29). Based on these previous findings, we used four lectins with different

specificities to determine if the 9E3 protein is glycosylated (Fig.5A): Concanavalin A (ln 1,2)
recognizes α-linked mannose, soy bean agglutinin (ln 3,4) preferentially binds to oligosaccharide

structures with terminal α and β-linked N-acetylgalactosamine and to galactose residues, Ulex

europaeus agglutinin I (ln 5,6) binds to glycoproteins containing α-linked fucose residues and wheat

germ agglutinin (ln 7,8) binds to N-acetylglucosamine. We found that the glycosylated proteins used

as positive controls (ovalbumin, carbonic anhydrase, trypsin inhibitor and lysosyme) bind individual

lectins to varying degrees (Fig 5A, odd lanes) but 9E3 does not bind any of them (Fig. 5A, even

lanes). These results were confirmed by the glycan detection method which consists of oxidation of

sugars by sodium metaperiodate followed by binding of digoxigenin-3-0-succinyl-e-amino-caproic

acid hydrazide to the oxidized sugars and by immunodetection with an antibody to digoxigenin (not

shown). Fig. 5B is an immunoblot of the 9E3 protein showing the abundance of the protein in the

starting material.

Some members of the chemokine family have been reported to be active in dimeric or

multimeric forms (15). Therefore, we investigated the possibility that the 9E3 protein might form

dimers or multimers in solution. Western blot and glycerol gradient analysis of the supernatant of

RSV-transformed cells in the absence of detergents and under non reducing conditions, showed that

most of the protein is present as the monomeric form in the top bands of the gradient (not shown).

Some cytokines, especially those that are very basic, have been shown to have affinity for

heparin (5,30,31) which could have physiological relevance. For example, extracellular matrix

molecules, in particular those that are heparin-like, can bind and store bFGF when this growth factor is

found in high concentrations and release it at specific times for interaction with its receptor and

stimulation of plasminogen activator production and DNA synthesis (30). Furthermore, it has been
 7

shown that heparin and heparan sulfate protect bFGF from proteolytic degradation, enabling this

growth factor to stimulate cells at a distance (31). Therefore, using column chromatography we

investigated whether the 9E3 protein has a high affinity for heparin. We found that 9E3 binds to

heparin and elutes from it (Fig. 6A) much as it would bind and elute from a cation exchanger (300-500

mM salt) rather than with high affinity.

It is also known that platelet factor 4 (PF-4), the stem molecule for the small cytokine
superfamily (of which 9E3 is a member), is present in platelet α granules complexed with serglycin, a

chondroitin sulfate proteoglycan32. Therefore, we investigated the possibility that 9E3 could be

present in thrombocytes, the chicken equivalent to platelets. Immunostaining of blood smears with the

antibody to the 9E3 protein did not reveal the presence of this protein in thrombocytes (not shown).

Immunoblot analysis of plasma and serum produced by clotting in vitro were similarly negative

(Fig.6B).

Using the semipurified protein from the heparin-binding experiment, we determined its pI on

an isoelectric focusing gel with a pH gradient 3-10. By comparison with isoelectric focusing markers,

we found that the protein has an isoelectric point of about pH 8.7 (not shown).

Processing of the 9E3 Protein: As mentioned above, two forms of the 9E3 protein (9 kDa and 6-7

kDa) appear in the supernatant of CEFs transformed with wild-type Prague C-RSV (e.g. Fig. 6; see

also ref. 14). However, in this study we have found that only the larger form appears in the

supernatant of cells transformed with the XD2 modification (26) of the Schmidt-Ruppin strain of the

virus or in the supernatant of N-CEFs stimulated to produce 9E3 by serum (e.g. Fig. 2), TPA or LPS

(Fig. 7A). Because the protein has no modifications except for two disulfide bonds, the simplest

interpretation of these results is that the cells infected with the Prague C strain of the virus produce an

enzymatic activity (absent or inactive in the XD2 Schmidt-Ruppin strain and in normal cells) that

processes the protein to the smaller form. To test this possibility, we labeled cultures of XD2-

transformed CEFs and TPA-stimulated normal CEFs with 35S methionine and 35S cysteine to obtain

radio labeled 9E3 protein exclusively of the larger form. We incubated supernatants from these

cultures with supernatants from Prague C cultures. In both cases, the labeled 9 kDa protein was
 8

partially processed to the 6-7 kDa form (Fig. 7B), confirming the presence of the postulated enzymatic

activity in the supernatant of Prague C T-CEFs.

Because 9E3 is abundantly expressed in vivo during wound healing and in tissues around

tumors (14,25), we speculated that enzymes known to be associated with these situations might be

responsible for the processing of the protein. Tumors produce high levels of plasminogen activators

(33,34), and thrombin and plasmin are abundantly released during wounding. Therefore, we

incubated the 9 kDa form of the protein with plasmin, thrombin, tissue plasminogen activator (tPA),

and urokinase plasminogen activator (uPA). Thrombin, tPA and uPA had no observable effect on

protein size (Fig. 7C, lanes 3-5), whereas plasmin processed the larger form to the smaller form (Fig.
7C, lane 2). Using 0.1U of plasmin per µg of protein, approximately half of the protein was

processed to the smaller form in the first 5 min, with detectable but progressively lower levels of the

larger form remaining through 15 min, falling below our detectability limit at 20 min (Fig. 7D). We

have also found that the smaller form of the protein is associated with ECM molecules. Western blot

analysis of cells cultured on EHS and on rat tail collagen gels showed that both matrices bind the 6-7

kDa form of the 9E3 protein (Fig. 7E). Whereas the supernatants showed only the 9kDa form (ln

2&6), both forms of the molecule were present in the collagen gels, with the smaller form greatly

predominating (lane 8), and only the 6-7 kDa form of the molecule was found in association with EHS

(ln 5). Extracts of the matrices before cell culturing showed no detectable levels of the 9E3 protein (ln

5&9). Binding of the 9E3 protein to these two matrices was confirmed by immunostaining with the

antibody to the 9E3 protein. To determine if this protein binds to specific components of EHS we

analyzed for the binding to the major components, laminin, collagen IV and proteoglycan (Fig. 7F).

We found that the protein binds strongly to laminin (ln 3&4), to a lesser extent to proteoglycan (ln

1&2), and does not bind to collagen IV (ln 5&6). We also tested for binding to fibronectin, a

glycoprotein that is very abundant in extracellular matrices, and found that the 9E3 protein does not

bind to it (Fig. 7F, ln 7&8).

 9

DISCUSSION
Many of the members of the chemokine superfamily of cytokines show structural and

biochemical characteristics which are important for their functions. The results presented here show

that: (1) both synthesis and secretion of the 9E3 protein are very rapid; (2) the protein carries no

modifications other than two disulfide bonds and does not have high affinity for heparin; (3) the

protein is rapidly synthesized and secreted as a 9 kDa form which is processed to 6-7 kDa by plasmin,

an enzyme released at wound sites and produced in association with tumors; (4) the smaller form of the

protein binds to interstitial collagen, laminin and to a lesser extent to complex proteoglycan, but does

not bind to collagen IV and fibronectin.

It has been shown previously that the regulation of activation of the 9E3 gene occurs at the

level of transcription (1,2) and that this gene is also activated in the presence of cycloheximide,

indicating that new protein synthesis is not necessary for initiation of transcription (1,2). These

properties indicate that this gene is an early response gene. Here we have shown that TPA stimulation

of cultures with undetectable amounts of 9E3 mRNA (1) or protein leads to production of detectable

amounts of the protein in less than 10 min. The 9E3 gene is activated at the level of transcription (1,2)

and in quiescent cells the message is unstable (40) and not detectable. Moreover, we showed

previously that quiescent cultures show only very few cells expressing 9E3, that growth stimulation
leads to expression in more cells, and that expression begins in G1 and continues into S phase (4).

Therefore, the large increase in protein we observe after stimulation with TPA is unlikely to be the

result of a simple increase in translation of low levels of mRNA present in cells previously expressing

the gene. More likely, upon growth stimulation, the 9E3 gene was activated in the cells entering the

cell cycle and that the 9E3 mRNA was transcribed and the protein synthesized within them, all in less

than ten minutes. The activation of the gene in more and more cells led to the steep rise between 20

and 60 min. Such very rapid activation would suggest that this gene might be under the control of a
repressor protein, as suggest previously for the gro gene (3). The expression of 9E3 is tightly

controlled in normal cells but this regulation is quickly lifted upon activation by serum or a variety of

specific agents (e.g. phorbol esters, kinase and phosphatase inhibitors, ionophores, RSV; 27,36), and
 10

during wound healing and tumor development (14,25). Alternatively, this gene could behave like a

stress response gene; the appropriate transcription factors could be bound to the TATA box and the

polymerase II could be paused very early in transcription, just waiting for the activation of a

transcription factor which would bind to its activation site releasing the paused polymerase.

It is well-established that secreted proteins are synthesized into the ER where disulfide bonds

are formed and cores consisting of N-acetylglucosamine and Mannose sugars are N-linked to proteins

that are to be further glycosylated in the Golgi by addition of new sugars to these cores. Furthermore,

during their passage through the Golgi, some proteins may also get glycosylated by O-linkages to Ser

and Thr. If proteins do not contain the necessary sequences for retention in the ER, they go via

vesicular transport from the ER to the first compartment of the Golgi, progress from stack to stack for

processing as necessary, and then are shuttled via additional vesicles to specific organelles or to the cell

surface to be secreted (37-43). This is true not only for glycoproteins but is also thought to be the case

for proteins that do not require modifications in the Golgi and are carried through passively as part of

the "bulk" flow (40).

The results of our experiments on the kinetics of secretion show that the 9E3 protein is

synthesized and secreted in less than 10 min, a time frame slightly shorter than that reported for

secretion alone of synthetic tripeptides that consist only of the acceptor sequence for N-linked

glycosylation (Asn-X-Ser/Thr). These tripeptides are the minimum size and complexity requiring

transit through the Golgi (40). These investigators observed that secretion took about 10 min and

concluded that their results should provide a minimum time for secretion of any protein through this

pathway. One way that 9E3 might be secreted more rapidly than these tripeptides would be for the

protein to be shuttled directly to the cell surface by vesicles that bud from the ER. This alternative

secretory pathway would not be basically different from the classical ER-Golgi pathway; it would

represent simply a short-circuit of the Golgi for unmodified proteins. As yet, we have no direct

evidence for such a secretory pathway in our system, but support for this suggestion comes from a

recent study of the primitive eukaryote Giardia lamblia . This organism has a secretory machinery

similar to that of higher eukaryotes (43). The nonencysting form of this organism lacks a Golgi

apparatus, yet these investigators showed that it is capable of secreting simple, nonglycosylated
 11

proteins by a pathway that is inhibited by Brefeldin A and requires the function of ADP ribosylation
factor (ARF) and coat protein βCOP, hence vesicular transport must be involved in secretion of these

proteins.

The fact that many cytokines, including 9E3, are active during emergency situations such as

wound repair and inflammatory response, suggests that maximizing the kinetics of gene activation,

protein synthesis, and secretion would be an evolutionarily-favored process and could have led to

conservation of this simple pathway to the cell surface. Molecular and immunolabeling approaches are

in progress to investigate the possibility that 9E3 may be secreted directly from the ER.

Some of the cytokines in the chemokine superfamily are biologically active in multimeric

forms, such as PF-4 which forms tetrameres (41). Others, such as IL-8 (44), IP-10 (45) and JE (46),

are active in monomeric form. The 9E3 protein occurs primarily as monomers in solution suggesting

that this protein is biologically active in monomeric form.

The biological significance of the existence of two forms of the 9E3 protein is not yet known,

but because both forms are labeled by an antibody to the C-terminus peptide (14), it is clear that

processing to the smaller form by plasmin occurs at the N-terminus. Importantly, a number of other

proteins in the chemokine superfamily also are processed at the N-terminus and it has been shown that

the various forms of the proteins can have different functions (5,15,47). In the case of IL-8, the

chemokine most homologous to 9E3, two active forms have been identified (18), both of which

contain the C-terminus, with the smaller form being much more potent in neutrophil degranulation than

the larger form. Similar observations have been made for PF-4 (48,49). The functions of 9E3 may be

at least as varied: The C-terminus one-third of the protein is angiogenic on the chorioallantoic

membrane (50) and, as shown here, the smaller form binds to specific extracellular matrix molecules.

Many cytokines with higher molecular weights are known to interact with molecules of the

extracellular matrix (ECM) and these interactions are known to be important in their function (31,51-

54). For example, when monocytes and macrophages adhere to FN, they produce Granulocyte-

Macrophage-Colony Stimulating Factor and Colony Stimulating Factor (51,54) or when cells adhere to
collagen they produce IL-1 and TNFα (54,55). In other cases, cytokines complex with ECM

molecules which protects them from protease degradation (31). Here we have shown that 9E3 binds
 12

to interstitial collagen and to specific basement membrane proteins. In the latter case, we found only

the small form of the protein, which was not present in the supernatant, hence cleavage must either be

required to expose a specific binding site or binding must immediately induce cleavage. The situation

with collagen could be more complicated because both forms of the molecule were found in the gels.

On the one hand, the amount of the larger form was much less than that of the smaller form, hence the

presence of the larger form in the gel could have been the result of entrapment of supernatant within the

gel rather than specific binding (collagen gels are much looser than EHS, consistent with this

possibility). On the other hand, both forms could be bound directly to the collagen. One way this

might occur is if binding is followed by cleavage and the cleavage on collagen is not as efficient as on

laminin and proteoglycan.

The interaction of 9E3 with these ECM molecules could have important implications for its

function(s), as has been shown for other cytokines (31,52-54,56). The matrix could potentially

function as a presenter of the cytokine to its receptor, could control how cells respond to it, could

induce cells to produce it or could protect it from being degraded. Because 9E3 is expressed at low

levels in connective tissue, is overexpressed in the granulation tissue of wounds and in tissues

surrounding tumors (14,25), it is possible that interactions with the matrix are of fundamental

importance for the function of the protein in these various events. The fact that 9E3 binds to some

matrix molecules and not to others suggests that at least some of its functions could operate through

simultaneous, synergistic, binding of 9E3 and specific ECM motifs to cells57,58 .

The observation that it is plasmin that specifically cleaves the 9 kDa form of 9E3 to the 6-7 kDa

form suggests that the smaller form could be active in wound healing and tumor suppression. In

wound healing, plasmin is known to be abundant at wound sites where 9E3 is overexpressed (14,25).

In the case of tumors, because 9E3 expression is elevated around their periphery and tumors produce

high levels of plasminogen activator, it is possible that this enzyme produces plasmin from

plasminogen resident in the tissues (in vivo, as much as 40% of plasminogen is found extravascularly

in normal tissues; 59,60). Plasmin, in turn, then could process the 9E3 protein to the smaller form.

In conclusion, the rapid kinetics of both production of the 9E3 protein and of its secretion

suggest that, upon the appropriate stimulus, the level of this protein in tissues could be elevated very
 13

rapidly, causing similarly rapid biological responses. The secreted form is cleaved to 6-7 kDa by

plasmin and binds to extracellular matrix molecules, suggesting that the smaller form of the protein

could be importantly involved in wound repair and tumor suppression. The close homology between

9E3 and the mammalian members of the intecrine superfamily of genes suggests that these

characteristics may be broadly shared in higher vertebrates.

MATERIALS AND METHODS

Expression of Recombinant 9E3 in Bacteria and Preparation of 9E3 Antiserum: An inducible

bacterial expression vector which contained the 9E3 coding region minus the signal peptide coding

sequence was constructed. To remove the signal sequence, 9E3 cDNA was partially digested with PvuII

and the 1076-nucleotide PvuII-EcoRI fragment was isolated. ClaI linkers were added and the DNA

fragment was recut with ClaI and PstI. Following a partial Sau3A1 digest, the 393 nucleotide, ClaI-

Sau3A1 fragment was isolated and ligated into ClaI-BamHI site of pJL6 (61). The resulting plasmid,

designated pX4, contains the entire coding region for mature 9E3 peptide downstream of a lambda

bacteriophage pL promoter and the cII repressor. The pX4 vector was used to transform E.coli N99cI+

(Pharmacia) which contains a wild type repressor of the cII repressor and E.coli N4830-1 (Pharmacia)

which carries a temperature-sensitive repressor. The validity of the construct was confirmed by restriction

mapping and nucleotide sequencing of the insert junctions. Recombinant 9E3 was prepared by growing

E.coli N4830-1 harboring pX4 at 28-30o C to O.D.595 of 0.5. Cultures were then transferred to 42o C for

2-4 h. A lysate of the bacterial cultures was applied to a denaturing polyacrylamide gel. To prepare

antisera, gel sections containing the 9 kDa band were excised, washed to remove SDS, homogenized

together with complete (for initial immunization) or incomplete (for subsequent immunizations) Freund’s

adjuvant and administered subcutaneously to female New Zealand rabbits. Animals were sacrificed after a

total of 4 or 5 immunizations and the reactivity of pre- and post-immune sera was examined.

Cell Cultures on Plastic Petri Dishes: Cultures of primary chick embryo fibroblasts (CEFs) were

prepared from 10-day-old chick embryos as described previously (62). Briefly, cultures of secondary

CEFs were prepared by seeding primary CEFs onto uncoated 60mm tissue culture dishes. Normal cells

were plated at 1.2x106 cells/dish and grown to confluency in 199 medium containing 2% calf serum and
 14

1% chicken serum. Transformed cells were also plated at 1.2x106 cells/dish, allowed to adhere for 3 hrs,

then treated with DEAE dextran (0.02%) and infected with XD-2, an engineered form of Schmidt-Ruppin

Rous sarcoma virus (RSV) subgroup A (26). Cells were fully transformed 4-5 days post infection.

Samples were analyzed by immunoprecipitation or by western blot.

Immunoprecipitations Using 9E3-Specific Antiserum: For each gel, the samples to be

immunoprecipitated with the antibody against 9E3 contained equal amounts of total protein as determined

by the BioRad detergent compatible assay for proteins. The samples were incubated with the rabbit anti-
serum at 1:500 dilution, O/N at 4oC with gentle rocking. Each sample was then incubated with 50µl of a

10mg/ml solution of Protein A-Sepharose in PBS containing 0.02% Na azide for 1 hr at 4oC with gentle

rocking. The pelleted beads were washed 3 times in 1X RIPA buffer containing 300mM NaCl and 1 time
in 1X RIPA buffer containing 10mM NaCl. The samples were electrophoresed in 15% SDS-PAGE (63)

for Fig. 1, or in 20% SDS-glycerol-PAGE (64) for the remaining gels. Controls consisted of samples

incubated with the pre-immune serum in place of the antiserum.

Stimulation of synthesis of 9E3 by TPA: Confluent starved CEFs were stimulated with 500 ng/ml of

TPA (Sigma Co.) in serum free medium and incubated for 10, 20, 30, 45 and 60 min at 37oC in the

presence of 35S methionine and 35S cysteine. The supernatants and cell extracts were collected treated

with protease inhibitors and the 9E3 protein immunoprecipitated as described above.

Pulse-Chase Labeling of Cells with 35S Methionine and S Cysteine: Cells were incubated with

minimal essential medium methionine- and cysteine-free for 1 hr at 37oC. Fresh medium, containing
250µCi/ml of 35S-methionine and 35S-cysteine plus 4% dialyzed calf serum, was added and the cells

incubated for 5 min or 15 min as pulse. The cells were washed 2 times with Tris-glutamine and chased

with complete media containing 4% calf serum and 1% chicken serum. Samples were collected at zero

time after wash (hence zero-time supernatants always show no protein) and at various times thereafter,

immunoprecipitated with the antibody 9E3, and run on 20% glycerol-SDS polyacrylamide gels. All times

reported for pulses and collections are accurate to within 15 sec.

Lectin Binding Assays to Determine if 9E3 is Glycosylated: We used four different lectins to

determine if 9E3 is glycosylated: Concanavalin A (Con A); Soy Bean Agglutinin (SBA); Ulex europaeus

Agglutinin I (UAE I); and Wheat Germ Agglutinin (WAG). Western blotted proteins were incubated with
 15

streptavidin-HRP complex for 30 min at RT. The blots were washed as above and then ECL (Amersham

Co.) was performed to visualize the glycosylated proteins. We also used a glycan detection method

(Boehringer and Mannheim) and followed the procedure recomended by the manufacturer.

Western Blot Analysis: Samples were transferred onto nitrocellulose and blocking done with 5% nonfat

dry milk in TBS containing 150mM NaCl and 0.1% Tween 20 (TTBS). The membrane was incubated

O/N at 4 oC with the antiserum diluted 1:2000 in TTBS plus 1% milk, washed 3 times, 20 min each, in

TTBS, incubated with goat-anti-rabbit HRP (Amersham) diluted 1:20000 in TTBS + 1% milk for 1 hr at

RT with gentle shaking. The 9E3 protein was visualized using ECL (Amersham) following the procedure

recommended by the company. Transfer buffers consisted of: Anode 1 buffer (300mM Tris, pH 10.4 +

20% Methanol), Anode 2 buffer (25mM Tris, pH 10.4 + 20% Methanol) and Cathode buffer (25mM Tris

pH 9.4 + 40mM Amino Caproic acid + 0.037% SDS).

Glycerol Gradient: One ml of supernatant of T-CEF was concentrated on Centricon 3 (Amicon) and
loaded on a 5-30% glycerol gradient in a buffer without β-mercaptoethanol or detergents (Tris 10 mM pH

7.4, EDTA 5 mM, NaCl 10 mM). Centrifugation was performed for 16 hours at 44,000 rpm in a SW50.1

rotor. An aliquot of 1/9 from each gradient fraction was analyzed on a 16,25% low-bis SDS-PAGE and

silver-stained using the silver staining kit from BioRad.

HiTrap Heparin-binding Chromatography: 1 ml columns of HiTrap heparin (Pharmacia) and

equilibrated with 10mM phosphate buffer pH 7.0. Supernatant was applied to the column at 0.5ml/min

and after the column was washed with equilibration buffer the protein was eluted with a NaCl gradient

from 0.1M-1M. Samples of the fractions were electrophoresed and analyzed by western blot.

Test for Protease Activity in the Supernatant of Prague C Transformed CEFs: We obtained

radioactivity labeled 9 kDa form of 9E3 by incubating normal CEFs stimulated with TPA or transformed

with XD2 Schmidt Ruppin Subgroup A RSV with 35S methionine and 35S cysteine. We labeled the cells

with a pulse of 1hr, washed away the pulse and collected the supernatant containing the 9E3 protein 2 hrs

after removal of the pulse. The supernatants were incubated 1:1 with supernatant from Prague C RSV for

1 hr at 37oC. The 9E3 protein was then immunoprecipitated and run on SDS-PAGE, enhanced with

Amplify (Amersham), dried, and exposed to X-OMAT AR film (Kodak).

 16

Enzymatic Digestions of the 9E3 Molecule: Aliquots of FPLC-purified protein containing 1 µg of

protein were treated with various enzymes (Sigma) for 10 min at 37oC. Final concentrations were: 0.3U

of thrombin, 0.1U plasmin, 0.04U urokinase and 0.03U tissue plasminogen activator. After the

enzymatic treatment the samples were analyzed by western blot. For the time course of processing by

plasmin, the samples were incubated for 5, 10, 15 and 20 min, sample buffer was added, the samples

were boiled and analyzed by western blot.

Extracellular Matrix Derivatized Plastic Petri Dishes: CEFs were cultured on top of Interstitial

Collagen and EHS. Interstitial Collagen: Rat tail collagen was mixed with 10x 199 and 0.33N NaOH in a

8:1:1 proportion. Three ml were delivered to each 60mm petri dish, incubated at 37oC for 30 min to gel,

and then incubated with 4ml of serum-free 199 medium for 1 hr at 37oC to remove the NaOH. The

collagen gels were rinsed with fresh medium, and the cells plated onto them. EHS: This matrix was
prepared from Englebreth-Holm-Swarm tumors (65). 750µl were pipetted onto 60mm petri dishes and

incubated 30 min. at 37oC to gel and then cells were plated onto it. Individual ECM molecules: 750µl of

the purified molecules were applied the same way the EHS matrix was applied. CEFs were cultured on

these matrices for two days without change of medium. This conditioned medium was analysed by

western blot.

Separation of Cells from Collagen and EHS: Interstitial Collagen: The cells were washed briefly in

1X PBS then incubated in collagenase for 30 min. at 37oC. The collagen was digested by this enzyme,

releasing the cells which were centrifuged to a pellet. The supernatant containing the collagen was

removed, treated with protease inhibitors and frozen. The cell pellet was resuspended and the cells

washed 3 times with 1X PBS and then extracted with 1X RIPA in the presence of proteases inhibitors.

EHS and purified ECM molecules: We followed the procedure described by (66) with the some

modifications. Briefly, cells were washed and incubated for 10 min. at 37oC in hypotonic buffer (10mM
Tris-Glu, 0.1% BSA, 0.1mM CaCL2, pH 7.5). This caused the cells to swell and partially detach from

the EHS. The cells were removed by disrupting the cell membranes with 0.5% N-P40 in hypotonic buffer

containing protease inhibitors, twice, 2 min. each at 37oC. The matrices were solubilized in 2X SDS

buffer and boiled immediately. Supernatants, cells extracts and matrices were analyzed by western blot.
 17

ACKNOWLEDGMENTS
We thank J. Koon for technical support and I. Campbell for critical reading of the manuscript.

MMG also thanks M.J. Bissell and the Lawrence Berkeley Laboratory for laboratory space and

support in 1992 by DOE grant DE-AC03-76SF00098. This work was further supported by NIH grant

GM48436 to MMG, by NIH grant GM46890 and Amer. Canc. Soc. grant JFRA-386 to MS, and by

NIH grant CA44356 to HH.

 18

REFERENCES
1. Sugano S, Stoeckle M Y, Hanafusa H (1987) Transformation by Rous sarcoma virus induces a

novel gene with homology to a mitogenic platelet protein. Cell.49: 321-328.

2. Bedard P-A, Alcorta D, Simmons D L, Luk K-C, Erikson R L (1987) Constitutive expression of a

gene encoding a polypeptide homologous to biologically active human platelet protein in Rous

sarcoma virus-transformed fibroblasts. Proc. Natl. Acad. Sci. USA 84: 6715-6719.

3. Anisowicz A, Bardwell L, Sager R (1987) Constitutive over-expression of a growth-regulated gene

in transformed Chinese hamster and human cells. Proc. Natl. Acad. Sci., USA. 84: 7188-7192.

4. Martins-Green M, Tilley C, Schwarz R, Hatier C, Bissell M J (1991) Wound-factor-induced and

cell cycle phase-dependent expression of 9E3/CEF4, the avian gro gene. Cell Regulation 2: 739-

752.

5. Oppenheim J J, Zachariae C O C, Mukaida N, Matsushima K (1991) Properties of the novel

proinflammatory supergene "intercrine" cytokine family. Annu. Rev. Immunol. 9: 617-648.

6. Mukaida N, Matsushima K (1992) Regulation of IL-8 production and the characteristics of the

receptors for IL-8. Cytokines 4: 41-53.

7. Miller M D, Krangel M S (1992) Biology and biochemistry of the chemokines: a family of

chemotactic and inflammatory cytokines Critical Reviews in Immunology 12: 17-46.

8. Cochran B H, Reffel A C, Stiles C D (1983) Wound-factor-induced and cell cycle phase-

dependent expression of 9E3/CEF4, the avian gro gene. Cell 33: 939-947.

9. Richmond A, Lawson D H, Dixon D W, Stevens J S, Chawla R K (1983) Extraction of a

melanoma growth-stimulatory activity from culture medium conditioned by the Hs0294 human

melanoma cell line. Cancer Res 43: 2106-2112.

10. Richmond A, Lawson D H, Dixon D W, Chawla R K (1985) Characterization of autostimulatory

and transforming growth factors from human melanoma cells. Cancer Res 45: 6390-6394.

11. Lawson DH, Thomas H G, Roy RGB, Gordon DS, Chawla RK, Nixon D and Richmon A

(1987) Preparation of a monoclonal antibody to a melanoma growth-stimulatory activity released

 19

into serum-free culture medium by Hs0294 malignant melanoma cells. J. Cell. Biochemistry 34:

169-185.

12. Richmond A, Thomas H G (1988) Melanoma growth stimulatory activity: Isolation from human

melanoma tumors and characterization of tissue distribution. J. Cellular Biochem. 36: 185-198.

13. Derynck R, Balentien E, Han JH, Thomas HG, Wen D, Samantha AK, Zachariae CO, Griffin

PR, Brachmann R, Wong WL, Matsushima K, Richmond A (1990) Recombinant expression,

biochemical characterization, and biological activities of the human MGSA/gro protein.

Biochemistry 29: 10225-10233.

14. Martins-Green M, Aotaki-Keen A, Hjelmeland L, Bissell MJ (1992) The 9E3 protein:

immunolocalization in vivo and evidence for multiple forms in culture. J. Cell Science 101: 701-

707.

15. Stoeckle MY, Barker KA (1990) Two burgeoning families of platelet factor 4-related proteins:

Mediators of the inflammatory response. The New Biologist 2: 313-323.

16. Sager R, Haskill S, Anisowicz A, Trask D, Pike M C (1991) GRO: A novel chemotactic

cytokine. Plenum Press, New York pp 73.

17. Sherry B, Cerami A (1991) Small cytokine superfamily. Current Opin. in Immunol. 3: 56-60.

18. Hebert CA, Luscinskas FW, Kiely J-M, Luis E, Darbonne WC, Bennett GL, Liu CC, Obin MS,

Gimbrone MA Jr, Baker JB (1990) Endothelial and leukocyte forms of IL-8. Conversion by

thrombin and interactions with neutrophils. J. Immunol. 145: 3033-3043.

19. Schröder J-M, Sticherling M, Henneicke HH, Preissner WC, Christophers E (1990) Il-1α or

tumor necrosis factor-α stimulate release of three NAP-1/IL-8-related neutrophil chemotactic

proteins in human dermal fibroblasts. J. Immunol. 144: 2223−2232.

20. Sager R, Haskill S, Anisowicz A, Trask D, Pike MC (1991) GRO: a novel chemotactic cytokine.

Advances in Experimental Medicine and Biology 305: 73-7.

21. Barker K A, Hampe A, Stoeckle M Y, Hanafusa H (1993) Transformation-associated cytokine

9E3/CEF4 is chemotactic for chicken peripheral blood mononuclear cells. J. Virol. 67: 3528-

3533.
 20

22. Schreiber A B, Winkler M E, Derynck R (1987) Transforming growth factor α is a more potent

angiogenic mediator than epidermal growth factor. Science 232: 1250-1253.

23. Roberts A B, Sporn M B, Assoian R K, Smith J M, Roche N S, Wakefield L M, Heine U I,

Liotta L A, Falanga V, Kehrl JH, Fauci A S (1986) Transforming growth factor type beta: Rapid

induction of fibrosis and angiogenesis in vivo and stimulation of collagen formation in vitro.

Proc. Natl. Acad. Sci. USA 83: 4167-4171.

24. Yang E Y, Moses H L(1990) Transforming growth factor β1-induced changes in cell migration,

proliferation, and angiogenesis in the chicken chorioallantoic membrane. J. Cell Biol. 111: 731-

741.

25. Martins-Green M, Bissell M J (1990) Localization of 9E3/CEF-4 in avian tissues: Expression is

absent in Rous sarcoma virus- induced tumors but is stimulated by injury. J. Cell Biol. 110: 581-

595.

26. Cross F R, Hanafusa H (1983) Local mutagenesis of Rous sarcoma virus: The major sites of

tyrosine and serine phosphorylation of p60src are dispensible for transformation. Cell 34: 597-607.

27. Barker K, Hanafusa H (1990) Expression of 9E3 mRNA is associated with mitogenicity,

phosphorylation, and morphological alteration in chicken embryo fibroblasts. Mol. Cel. Biol. 10:

3813-3817.

28. Jentoft N (1990) Why are proteins O-glycosylated? TIBS 15: 291-294.

29. Esko J D (1991) Genetic analysis of proteoglycan structure, function and metabolism. Current

Opinion in Cell Biology 3: 805-16.

30. Flaumenhaft R, Moscatelli D, Saksela O, Rifkin DB (1989) Role of extracellular matrix in the

action of basic fibroblast growth factor: matrix as a source of growth factor for long-term

stimulation of plasminogen activator production and DNA synthesis. J. Cell. Physiol. 140: 75-81.

31. Flaumenhaft R, Moscatelli D, Rifkin DB (1990) Heparin and heparan sulfate increase the radius

of diffusion and action of basic fibroblast growth factor. J. Cell Biol. 111: 1651-1659.

32. Ruoslahti E, Yamaguchi Y (1991) Proteoglycans as modulators of growth factor activities.

Cell 64: 867-869.

 21

33. Mignatti P, Rifkin DB (1993) Biology and biochemistry of proteinases in tumor invasion.

Physiol. Rev. 73: 161-195.

34. de Vries TJ, Quax PHA, Denijn M, Verrijp KN, Verheijen JH, Verspaget HW, Weidle UH, Ruite

DJ, van Muijen GNP (1994) Plasminogen activators, their inhibitors, and urokinase receptor

emerge in late stages of melanocytic tumor progression. Am. J. Path. 144: 70-81.

35. Stoeckle M, Hanafusa H (1989) Processing of 9E3 mRNA and regulation of its stability in normal

and Rous sarcoma virus-transformed cells. Mol. Cel. Biol. 9: 4738-4745.

36. Vangainkar S, Bui T, Koon J, Martins-Green M (1995) Thrombin is a potent natural stimulator of
the small cytokine 9E3/CEF4. Wound Repair and Regeneration 3: 105.

37. Balch WE, Dunphy WG, Braeli WA, Rothman JE (1984) Reconstition of the transport of protein

between successive compartments of the Golgi measured by the coupled incorporation of N-

acetylglucosamine. Cell 39: 405-416.

38. Orci L, Glick BS, Rothman J E(1986) A new type of coated vesicular carrier that appears not to

contain clathrin: its possible role in protein transport within the Golgi stack. Cell 46: 171-184.

39. Becker CJM, Keller DS, Balch W E (1987) Semi-intact cells permeable to macromolecules: use

in reconstitution of protein transport from the ER to the Golgi. Cell 50: 523-534.

40. Wieland FT, Gleason ML, Serafini TA, Rothman JE (1987) The rate of bulk flow from the

endoplasmic reticulum to the cell surface. Cell. 50: 289-300.

41. St Charles R, Walz D A, Edwards BFP (1989) The three dimensional structure of bovine platelet

factor 4 at 3.0-A resolution. J. Biol. Chem. 264: 2092-2099.

42. Rothman J (1994) Mechanisms of intracellular protein transport. Nature 372: 55-63.

43. Lujan HD, Marotta A, Mowatt MR, Sciaky N, Lippincott-Schwartz J, Nash TE (1995)

Developmental induction of Golgi structure and function in the primitive eukaryote Giardia

lamblia. J B C 270: 4612-8.

44. Paolini J F, Willard D, Consler T, Luther M, Krangel MS (1994) The chemokines IL-8,

monocyte chemoattractant protein-1, and I-309 are monomers at physiologically relevant

concentrations. J. Immunology 153: 2704-2717.

 22

45. Luster AD, Ravetch JV (1987) Biochemical characterizatin of a γ interferon-inducible cytokine

(IP10). J. Exp. Med. 166: 1084-1097.

46. Graves DT, Jiang YL, Williamson MJ, Valente AJ (1989) Role of extracellular matrix in the

action of basic fibroblast growth factor: matrix as a source of growth factor for long-term

stimulation of plasminogen activator production and DNA synthesis. Science 245: 490-493.

47. Gimbrone et. al. (1989) Endothelial interleukin-8: a novel inhibitor of leukocyte-endothelial

interaction. Science 246: 1601-1603.

48. Maione TE, Gray SG, Petro J, Hunt A, Donner AL, Bauer SI, Carson HF, Sharpe RJ (1990)

Inhibition of angiogenesis by recombinant human platelet factor-4 and related peptides. Science

247: 77-79.

49. Sato Y, Abe M, Takaki R (1990) Platelet factor 4 blocks the binding of basic fibroblast growth

factor to the receptor and inhibits the spontaneous migration of vascular endothelial cells. Bioch.

Bioph. Res. Com. 172: 595-600.

50. Martins-Green M, Kelly T, Clemons B, Hanafusa H (1994) Angiogenic activity in vivo of the C-

terminus of the 9E3/CEF4 gene product, a small inducible cytokine overexpressed during wound

healing. Mol. Biol. of the Cell (Suppl) 5: 395a.

51. Nathan C, Sporn M (1991) Cytokines in context. J. Cell Biol. 113: 981-986.

52. Yamaguchi Y, Mann DM, Ruoslahti E (1990) Negative regulation of transforming growth
factor-β by the proteoglycan decorin. Nature (Lond.). 346: 281-184.

53. Lillien L E, Sendtner M, Raff MC (1990) Extracellular matrix-associated molecules collaborate

with ciliary neurotrophic factor to induce type-2 astrocyte development. J. Cell Biol. 111: 635-

644.

54. Eierman DF, Johnson CE, Haskill JS (1989) Human monocyte inflammatory mediator gene

expression is selectively regulated by adherence substrates. J. Immunol. 142: 1970-1976.

55. Dayer J-M, Ricard-Blum S, Kaufmann M-T, Herbage D (1986) Type IX collagen is a potent
inducer of PGE2 and interleukin 1 production by human monocyte macrophages. FEBS (Fed.

Eur. Biochem. Soc.) Lett. 198: 208-212.

 23

56. Ingber DE, Folkman J (1989) Mechanochemical switching between growth and differentiation

during fibroblast growth factor-stimulated angiogenesis in vitro: role of extracellular matrix. J.

Cell Biol. 109: 317-330.

57. Martins-Green M, Bissell MJ (1995) Cell-extracellular matrix interactions in development.

Sems. in Dev. Biol. 6: 149-159.

58. Martins-Green M (1996) Dynamics of Cell Interactions with Implications for Tissue

Engineering. In Press.

59. Isseroff RR, Rifkin DB (1983) Plasminogen is present in the basal layer of the epidermis. J.

Invest. Dermatol. 80: 297-299.

60. Clark RA, Henson PM (1995) The molecular and Cellular Biology of Wound Repair, Plenum

Press, New York.

61. Lautenberger JA, Court D, Papas TS (1983) High-level expression in Escherichia coli of the

carboxy-terminal sequences of the avian myelocytomatosis virus (MC29) v-myc protein.Gene 23:

75-84.

62. Bissell M J, Farson D, Tung A S-C (1977) Cell shape and hexose transport in normal and virus-

transformed cells in culture. J. Supramol. Struct. 6: 1-12.

63. Anderson BL, Berry RW, Telser A (1983) A sodium dodecyl sulfate-polyacrylamide gel

electrophoresis system that separates peptides and proteins in the molecular weight range of 2500-

90,000. Analytical Biochem. 132: 365-375.

64. Giulian GC, Moss RL, Greaser M (1983) Improved methodology for analysis and quantitation

of proteins on one-dimensional silver-stained slab gels. Anal. Biochem. 129: 277-287.

65. Kleinman HK, McGarvey ML, Hassell JR, Star VL, Cannon FB, Laurie GW, Marytin GR

(1986) Basement membrane complexes with biological activity. Biochemistry 25: 312-318.

66. Kramer RH, Fuh GM, Bensch KG, Karasek MA (1985) Synthesis of extracellular matrix

glycoproteins by cultured microvascular endothelial cells isolated from the dermis of neonatal and

adult skin. J. Cell. Physiol. 123: 1-9.