IDENTIFICATION OF A LOW MOLECULAR WEIGHT FORM OF EPITHELIAL TRANSFORMING GROWTH FACTOR
D. Dunnington*, W. Prichett, M. Moyer and R. Greig
SmithKline Beecham Pharmaceuticals, Department of Cell Sciences, 709 Swedeland Road, King of Prussia, PA 19406-2799
(Received in final form September 25, 1990)
Summary
We have purified a novel form of epithelial transforming growth factor (TGFe) from bovine kidney. Acid ethanol extracts of kidney were fractionated by size exclusion, reverse phase and cation exchange chromatography and activity was monitored by measuring growth of SW13 adrenocortical carcinoma cells in soft agar. The purified material was highly cationic, bound weakly to heparin and gave a band at 13-15000 Mr by SDS-PAGE following Bolton- Hunter iodination. This band corresponded to the migration of biological activity extractable from gel slices. The results suggest that we have isolated a truncated form of TGFe which nonetheless retains biological activity.
The human adrenocortical carcinoma cell line SW13 has been used to isolate an apparently novel growth factor, termed TGFe, from bovine kidney (1). Formation of SW13 colonies in soft agar is stimulated either by TGFe or basic fibroblast growth factor (bFGF:1,2). Purified bovine TGFe is a single chain protein of Mr 25,000 that requires disulfide bonds for full biological activity (1), and does not bind to receptors for epidermal growth factor or bFGF (1). In addition to SW13 cells, soft agar growth of A431 squamous carcinoma cells, T24 bladder carcinoma cells and AKR-2B mouse embryo fibroblasts is stimulated by bovine TGFe (1). Extraction of 26 of 32 freshly excised human carcinomas under highly acidic conditions that would destroy bFGF activity yielded material that stimulated growth of SW13 cells in soft agar (2). Thus TGFe-like activity appears to be present in numerous epithelial cells and tissues (2).
Although methods have been devised for purification of TGFe to near homogeneity (3), determination of its amino acid sequence has yet to be achieved, owing to the extremely low abundance of TGFe in tissue or cellular sources. During scale-up of the purification procedure, we observed partitioning of biological activity into two separate components. The first component was the 25,000 Mr form seen previously (1); characterization of the second is described in the present report.
*To whom correspondence should be addressed.
Materials and Methods
Soft agar assay. 5000 SW13 cells were plated in 1 ml of 0.3% agarose (FMC, Rockland, ME) over a base layer of 0.5% agarose in Dulbecco’s modified Eagles medium with 10% fetal bovine serum. Samples were added with the cells and cultures were incubated for 10 days, stained with iodonitrotetrazolium and counted by image analysis as described (4). The SW13 cell line was obtained from ATCC, Rockville, MD and was free from mycoplasma contamination as determined by Gen Probe assay (San Diego, CA).
Purification of TGFe. Bovine kidney was extracted with acidified ethanol as described (4). The extract was diluted 4-fold with 0.05M ammonium acetate pH 4.0 and loaded onto a 15 x 25 cm column of S-sepharose (Pharmacia, Piscataway, NJ). The column was washed with 0.05M ammonium acetate and eluted batchwise with 2M sodium chloride in 0.05 M ammonium acetate pH 4.0. Eluted material was concentrated 40-fold by ultrafiltration (Amicon YM5 membranes, Danvers, MA) and further fractionated by size exclusion chromatography and preparative reverse phase HPLC essentially as described (4) except that Sephacryl S-200 resin (Pharmacia) was used instead of Bio Gel P60 (4). Active material from the reverse phase step was lyophilized, reconstituted in 0.05M ammonium acetate, pH 4.0, 8M urea and loaded onto a 1.6 x 10 cm Mono S column (Pharmacia), equilibrated with 0.05M ammonium acetate, 8M urea and eluted with a linear gradient of 0-60% 1M ammonium acetate, 8M urea pH 4.0. Active fractions were pooled and subjected to reverse phase HPLC using a 5 x 0.46 cm phenyl column (Vydac 219TP,
The Separations Group, Hesperia, CA), equilibrated with 0.05% trifluoroacetic acid and eluted with a gradient of 5-30% acetonitrile, 0.05% trifluoroacetic acid. Material from this step was lyophilized, reconstituted in 0.05M ammonium acetate pH 6.0 and loaded onto a 0.5 x 1 cm column of red sepharose (Pharmacia). The column was washed with 1M sodium chloride in 0.05M ammonium acetate pH 6.0 and with 0.05% trifluoroacetic acid and eluted with 1M ammonium acetate in 50% n-propanol. The eluate was diluted 10-fold with distilled water, lyophilized, desalted on a 5 x 0.46 cm C18 reverse phase column (The Separations Group) and examined by iodination and SDS-PAGE as described below. The purification scheme is summarized in Fig. 1.
Electrophoretic analysis of TGFe. Samples of purified TGFe were iodinated with the Bolton- Hunter reagent (ICN, Costa Mesa, CA) as recommended by the supplier. Labeled protein was separated from free iodine by gel filtration on a Sephadex PD10 column (Pharmacia). Aliquots of iodinated material were run on 15% polyacrylamide gels in the absence of mercaptoethanol and bands were located by autoradiography on XAR film (Kodak, Rochester, NY). Further samples of unlabeled material were subjected to electrophoresis as above and the gel was sliced into 0.5 cm sections. Each section was placed in an Eppendorf tube containing 0.5 ml 0.05% trifluoroacetic acid and homogenized with a small pestle. After 18 hr at 4ºC, the tubes were centrifuged for 5 min in a microfuge and the supernatants were bioassayed on SW13 cells to locate growth factor activity.
Heparin affinity of TGFe. Purified TGFe was dissolved in 0.15M sodium chloride in 0.05M NaH2PO4 PH 7.0 and loaded onto a column of heparin-sepharose (Pharmacia). The column was eluted stepwise with increasing concentrations of sodium chloride in 0.05M NaH2PO4 PH 7.0. Aliquots of the fractions were acidified to pH2 with trifluoroacetic acid and assayed for biological activity.
Results and Discussion
Identification of a second form of TGFe. After capture on S-sepharose and purification by size exclusion and reverse phase chromatography, bovine TGFe was further fractionated by cation exchange chromatography in the presence of 8M urea. This step yielded 2 peaks containing
biologically active material (Fig. 2). The least cationic peak (“Peak A” in Fig. 2) was identified by SDS-PAGE (see below) as containing a 25k form of bovine TGFe similar to that described by Halper and Moses (1). The second peak (“Peak B”) was further purified by reverse phase HPLC and dye affinity chromatography on red sepharose. The protein content of this material was too low for accurate measurement; thus we were unable to determine the specific activity of the Peak B TGFe. Of the total activity recorded from the Mono S column, approximately 60% was Peak A and 40% Peak B.
KIDNEY EXTRACT
S-SEPHAROSE (CAPTURE STEP)
FIG. 1 Purification scheme for TGFe Peaks A and B. Activity was followed by measuring growth of SW13 cells in soft agar. Purified material was characterized by iodination, SDS-PAGE and gel slicing (see text).
SEPHACRYL S-200
C4 REVERSE PHASE HPLC
MONO-S FPLC
PEAK A
PEAK B
PHENYL REVERSE PHASE
PHENYL REVERSE PHASE
125 1, SDS-PAGE
RED SEPHAROSE
DESALT, 125 I,SDS-PAGE
200
0.12
4
Colonies > 60um
Peak
Peak B
Absorbance (280nm)
150
0.61
0.08
FIG. 2
M AmAc
100
Chromatography of bovine TGFe on Mono S cation exchange resin in the presence of 8M urea. Ma- terial was eluted from the column with a linear gradient of ammonium acetate (AmAc), pH 4.0 in 8M urea. Aliquots of each fraction were assayed for SW13 stimulatory activity ( 0 - 0 ).
0.0
0.04
50
0
0.00
0
20
40
60
Fraction No.
Molecular weight of Peak B TGFe. Because TGFe does not readily stain with silver reagents, we used radioiodination and autoradiography to visualize TGFe in SDS gels. Samples of both Peak A and B material were examined. The Peak A sample was tested at the phenyl HPLC stage (Materials and Methods) and contained many low molecular weight impurities that were strongly labeled with iodine (Fig. 3). Longer exposures of the autoradiograph showed a weak band at Mr 25000. Elution of biological activity from gel slices showed 2 peaks of SW13-stimulating ma- terial, one at 25k and a second at 13-15k Mr (Fig. 3). The Peak B sample, tested after the C18 desalting step, contained a major peak of activity that corresponded with a radioactive band at 13-15k Mr (Fig. 3). One additional radiolabeled band was seen in this sample at Mr 5000-6000. A minor peak of activity was seen at 6500 Mr and did not appear to correspond with the 5-6000
Mr labeled band. Currently we do not have convincing evidence for additional low molecular weight forms of TGFe but this possibility cannot be ruled out.
MW Markers
4 6
30
21.5
14.3
6.5
3.4
0
100
Mono S Peak A
Colonies > 60um
200
140
Mono S Peak B
70
0
0
4
8
12
16
20
24
Slice No.
Electrophoretic analysis of bovine TGFe. Samples of Peak A material purified to the phenyl HPLC stage and Peak B TGFe purified to the C18 step (Materials and Methods) were iodinated with the Bolton- Hunter reagent and run on 15% polyacrylamide gels in the absence of reducing agents. Gels containing corresponding samples of unlabeled material were cut into 0.5 cm slices. Activity eluted from the slices was measured by bioassay on SW13 cells.
These results suggest that Peak A TGFe contains material similar to the 25,000 Mr TGFe described by Halper and Moses (1) and in addition, a truncated but nonetheless biologically active component of Mr 13-15k which is present in purified form in Peak B. Given the similarity in chromatographic behavior of the 25k and 13-15k components and the presence of both forms in the Peak A sample, it seems likely that Peak B was derived from the 25,000 Mr material, possibly by proteolytic cleavage. This would indicate that the majority of biological activity of TGFe resides in a 13-15k subfragment of the 25,000 Mr protein although further studies at the structural level will be necessary to confirm this possibility.
Relationship of Peak B TGFe to bFGF. Although the acidic conditions employed in our purification scheme would be expected to destroy bFGF activity (5), we examined the affinity of Peak B TGFe for heparin to determine whether TGFe belongs to the family of cationic heparin- binding growth factors (6). These proteins have very high affinity for heparin and typically require 1-1.5M salt for dissociation from a heparin matrix (6). Peak B TGFe bound to heparin as expected from its cationic nature; however, it was readily eluted from the matrix with 0.5M salt (Fig. 4). This salt concentration is roughly similar to that required to elute TGFe from weak cation exchangers (not shown). Previous studies have shown that the 25k form of TGFe is also eluted from a heparin matrix with 0.5M salt (3). Thus the two forms of TGFe have lower affinities for heparin than bFGF and related molecules.
200
Colonies > 60um
150
100
50
0
0.15
0.25
0.5
0.75
1.25
2
M NaCl
FIG. 4
Heparin affinity of Peak B TGFe. Purified Peak B material was loaded onto a column of heparin-sepharose and eluted stepwise with increasing concentrations of sodium chloride. Biological activity in the fractions was measured by bioassay on SW13 cells.
Conclusion
We have isolated a novel form of TGFe from bovine kidney. Our material has a molecular weight of 13-15,000 by SDS-PAGE under non-reducing conditions, is highly cationic and has only moderate affinity for heparin. It is possible that this protein is a truncated form of the 25,000 Mr TGFe described previously, which nonetheless retains biological activity in the SW13 soft agar assay.
Acknowledgement
We thank Jo Anne Mackey for excellent secretarial assistance.
References
1. J. HALPER and H.L. MOSES, Cancer Res., 47 4552-4559 (1987).
2. J. HALPER and H.L. MOSES, Cancer Res., 43 1972-1979 (1983).
3. P.G. PARNELL, J. WUNDERLICH, B. CARTER and J. HALPER, J. Cell. Biochem., 42 111-116 (1990).
4. D.J. DUNNINGTON, S. PINSKY, D. MATTES, W. PRICHETT, C.Q. EARL, R. GREIG and M.A. ANZANO, Analytical Biochem. 174 257-264 (1988).
5. M. SENO, R. SASADA, M. IWANE, K. SUDO, T. KUROKAWA, K. ITO and K. IGARASHI, Biochem. Biophys. Res. Comm. 151 701-708 (1988).
6. R. LOBB, J. SASSE, R. SULLIVAN, Y. SHING, P. D’AMORE, J. JACOBS and M. KLAGSBRUN, J. Biol. Chem. 261 1924-1928 (1986).