Proteoglycans (PGs), proteins bearing glycosaminoglycan (GAG) side chains and present in the extracellular matrix and plasma membranes of almost all cell types, are thought to play important roles in cell growth, morphogenesis, and cancer (Bonneh-Barkay et al. 1997). The biological activities of PGs are mediated in part by their associated GAG chains, polysaccharides having varying patterns of sulfation and consisting of two types: chondroitin sulfate and heparan sulfate (HS). The most abundant cell-surface PGs are those that bear HS side chains which interact with a variety of heparin-binding proteins such as extracellular matrix components and growth factors (Bonneh-Barkay et al. 1997). The biological importance of heparan sulfate proteoglycans (HSPGs) and the potential contribution of their GAG chains has prompted many attempts to correlate HS glycanation patterns to PG functions, in particular their binding against potential ligands. However, these attempts have often been frustrated by difficulties in isolating or labeling any specific PG; chemical analyses have often been done on relatively heterogeneous PG populations or mixtures of GAG chains obtained from such populations. The binding affinity of a heterogeneous PG population does not accurately reflect the contribution of the GAG chains due to potential variability in the core proteins. Measuring the binding of GAG mixtures, on the other hand, fails to consider the native spatial configuration of the GAG chains, and ignores the core protein altogether. To accurately assess the effect of glycanation patterns on the binding activity of a PG, one needs to keep the core protein constant and vary only its glycanation pattern.
One way of achieving this is to express the same PG in different cell types that glycanate it differently. In this study, the gene coding sequence for the PG of interest is first fused with the gene sequence for heat-stable alkaline phosphate (AP); (Flanagan and Leder 1990) the resultant fusion plasmid is transfected into various cell lines. AP-tagging allows convenient tracking of the fusion protein in subsequent biochemical assays. The binding affinity of the fusion protein is determined with affinity co-electrophoresis, or ACE (Herndon and Lander 1997), a method that allows rapid determination of the binding affinity of moderately to highly charged proteins such as PGs. The combination of AP-tagging and ACE allows one to conveniently determine the binding affinity of a specific PG.
We look at the effects of cell-type dependent glycanation on the binding affinities of the PG glypican against three common extracellular molecules: Type I collagen, basic fibroblast growth factor (bFGF), and laminin-1.
Glypican is a cell-surface HSPG that is highly expressed in brain and skeletal muscle (Litwack et al. 1994; Karthikeyan et al. 1994), and that has been implicated in regulating cellular responses to fibroblast growth factors (Bonneh-Barkay et al. 1997) and the transforming growth factor decapentaplegic (Jackson et al. 1997). We find that the glypican-AP fusion proteins expressed in two mammalian cell lines, 293 and Chinese hamster ovary (CHO) cells, exhibit similar binding affinities against bFGF and laminin-1; however, they have significantly different affinities against Type I collagen.
Materials and Methods
Production of glypican-AP fusion plasmid:
The glypican-AP fusion plasmid was constructed as described by Flanagan and Leder (1990). The extracellular domain of glypican was inserted into the expression vector APtag-2, immediately upstream of the AP coding sequence (Figure 1). APtag-2 confers ampicillin and tetracycline resistance to transformed bacterial cells and contains a SV40 origin of replication that selectively enhances the plasmid's production in certain cell lines. E. coli transformed with the fusion construct were used to inoculate 500 ml of Luria Broth medium containing 40 µg/ml ampicillin and 10 µg/ml tetracycline, and were grown for 12 to 24 h at 37ºC. The expressed plasmid was collected according to the Quiagen Maxi Protocol (Quiagen, Santa Clarita, CA) and its concentration determined with ultraviolet (UV) spectrophotometry.
Expression of the glypican-AP fusion plasmid in cell lines:
Cos-7 and 293 cells were grown in an environment of 100% humidity, 5 to 8% CO2, at 37ºC, and to 50 to 60% confluence in Dubelco's Modified Eagle Medium containing 8 to 10% fetal bovine serum (Hyclone Inc., Logan, UT), penicillin, and streptomycin. Five µg of DNA in 250 µl optimem (Life Technologies Inc., Gaithersburg, MD) were mixed with 30 µl lipofectamine (Life Technologies Inc., Gaithersburg, MD) in 250 µl optimem and incubated for 20 to 40 min. The mixture was then added to each 100 mm plate of cells along with 10 ml of optimem. The next day, the cell media were replaced with normal growth medium; the media harvested 3 to 4 d after the start of transfection. The media were sterile-filtered and stored at 4ºC. Transfection of CHO cells, using protocols obtained from Life Technologies (http://www.lifetech.com), was similar except that 1) the medium used was Ham's F-12 (Life Technologies) and 2) only 5 ml of optimem were added at the time of transfection, followed by an additional 5 ml, 5 h later.