figure4.gif (11723 bytes)
Figure 4
ACE electrophoretograms for the interactions between 293- and CHO-derived glypican-AP fusion proteins and 3 extracellular ligands: Type I collagen, bFGF, and laminin-1.   The top gels are for 293-derived glypican-AP, and the bottom gels for CHO-derived glypican-AP.  The ligand concentrations for each binding assay are indicated at the bottom of the gels.  The gels for CHO-derived glypican-AP also show, in addition to the expected ACE patterns, a relatively dark, non-shifting horizontal band near the top (arrow).  The presence of this band is probably due to degraded products of the fusion protein and is accounted for as described in Materials and Methods.  Images of the gels were taken with a digital camera using 1D Image Analysis Software (Kodak Digital Science).

enough to make the interference due to endogenous PG insignificant. It should also be noted that the purified glypican-AP contains HS chains of varying length and degree of sulfation so that the K obtained represents an average value of a population of glypican molecules with the same core protein, but with different glycanation patterns.

Among the ligands examined, bFGF yields the highest binding affinity against glypican, which is consistent with previous findings that HSPGs are necessary for receptor binding and mitogenic activity of bFGF (Aviezer et al. 1994).  The affinities determined for the binding between 293-derived glypican and collagen are similar to those obtained for the binding of low-molecular weight heparin (K=315±83 nM) and syndecan (derived from normal murine mammary gland epithelial cells, K= 164±7 nM) against the same ligand (San Antonio et al. 1993).  Glypican, syndecan, and heparin share a common structural feature, namely the presence of sulfated polysaccharides; the similarity in the binding affinities thus suggests that glypican may have a similar collagen-binding determinant as heparin or syndecan within the sugar chains.  The K value obtained for laminin-1, on the other hand, is substantially lower than that obtained for the binding between laminin-1 and glypican (obtained in the rat brain; Herndon and Lander 1992). Brain-derived glypican tends to have longer, more heavily sulfated HS chains, suggesting that 1) glypican's binding affinity against laminin-1 depends primarily on its GAG chains rather than its core protein and 2) that the nature of the GAG chains (for example, their length and degree of sulfation) is more important for binding against laminin than against bFGF.


figure5a.gif (4876 bytes)
figure5b.gif (5828 bytes)
Figure 5
Determination of ACE retardation coefficients (R’s).  To determine the retardation coefficients (Figure 2) we need to measure the mobility of the PG front, taken to be the distance from the top of each lane to the “midpoint” of the PG band. To determine the midpoint of the PG band, relative AP-staining intensity along a 2 mm wide vertical band in the middle of the lane is measured as a function of the distance from the top of each lane.  A typical plot of AP-staining intensity versus migration distance is shown in (A).  The staining intensity ranges from 0 (lightest) to 255 (darkest).   The midpoint of the band is taken to be the point, or line, that divides the area under the intensity curve into equal halves (shown as the dashed line), after subtracting the background intensity level (about 50 in this case).  In some cases, significant amounts of non-binding AP-positive degraded products were present (Figure 4, arrow), which register as a secondary peak on the intensity curve (B, top curve).  The degraded products are accounted for by subtracting from the original intensity values the intensity curve due to the degraded product alone (bottom curve) which yields a more accurate intensity curve (middle curve).  The midpoint of the corrected curve is then determined as described.

Binding heterogeneity in the interaction of glypican and its ligands­as shown by highly extended bands on the ACE electrophoretograms­is lower for bFGF than for collagen (and possibly laminin). Further, the low Rvalues for collagen strongly suggest the presence of a glypican subpopulation in the PG sample that does not bind collagen.  Yet no such subpopulation is seen with bFGF. Apparently, there exists a glypican subpopulation that 1) does not bind, or binds very weakly to, collagen and 2) binds bFGF with moderate to high affinities.  Comparison of Kd values obtained for the binding of glypican-AP from 293 and CHO cells against the three ligands shows that glypican from the two cell lines binds bFGF or laminin-1 with similar affinities, yet binds collagen with significantly different affinities (Figure 7).

back.gif (287 bytes) back.gif (221 bytes)


Page 5


Huy Pham - Determination of Functionally Relevant Cell-Type... [1] [2] [3] [4] [5] [6] [7]