Fibronectin Exchange at Variable Surface Concentrations

Another set of experiments that enables determination of the characteristics of FN layers containing various amounts of protein will now be described. As it will be seen later on, this kind of analysis can be used to derive additional features about the interaction strength of adsorbed FN molecules and to conclusions about conformational changes of the adsorbed proteins.

FN layers were adsorbed from solutions of 2.5, 5, 10 and 50 ^g/ml FN in PBS onto hydrolysed POMA and PPMA surfaces; the adsorption process wasmonitoredby QCM-D. Thechangesinmassandviscoelasticproperties caused by the formation of adlayers were determined from the measured changes in resonance frequency and damping characteristics (energy dissipation) of the quartz crystal. As discussed in detail elsewhere (Renner et al. 2005), the difference in the relationship between mass increase and change in damping allowed us to consider FN layers as being "stiffer" on POMA, with a more compact protein conformation and stronger binding to the surface. Strong hydrophobic interactions apparently cause the tight bindingofFNtobemostpronouncedatlow coverage. Atveryhighcoverage the FN molecules may bind more loosely even on POMA due to proteinprotein in plane interactions and competition of proteins for surface sites, leading to a more weakly bound layer structure with higher dissipation. This interpretation is supported by comparing the FN surface coverage determined by QCM-D and HPLC. As in the QCM-D measurement the water mass content incorporated or coupled to the protein layer or bound as hydration shells to FN molecules is included in the mass determination, an effective density, £>eff, of the adsorbed layer can be calculated as introduced by Hook etal. (2001):

Vlayer mEto1 +

where the values of mlayer and mProt(_ mHPLC) are taken from measurement data. The solvent and protein densities are set to £>Solv _ 1,000 kg/m3 and

Fig. 8. Effective density peg- of the FN layer versus the surface coverage determined by HPLC (according to Eq. 4)

pProt = 1,370 kg/m3 according to Quillin and Matthews (2000). By using this equation, the density can be calculated for the dependence on the FN solution concentration during adsorption, as presented in Fig. 8. A strong change in the layer composition is observed for POMA surfaces, in contrast to a smaller change in the density for PPMA surfaces. The decrease in surface density corresponds to the change of coupling of water into the FN layer.

One can conclude that FN adsorbs on POMA at lower coverage in a very compact form with strong conformational changes due to the strong hy-drophobic interaction. At higher coverage it still adsorbs in a more globular form with a strong affinity to the surface. The globular shape, similar to the shape in solution, probably leads to the high water content in the FN layer and the higher surface coverage in comparison to the PPMA surface.

On the hydrophilic PPMA, the adsorption of FN was assumed to occur in a more expanded state due to the influence of electrostatic interactions. The lower affinity to the surface does apparently not strongly change the conformation. However, the electrostatic interactions may open the globular structure of the dissolved FN. Thus, less water can be coupled into FN molecules. Furthermore, the electrostatic proteinsubstrate interaction probably decreases the number of free charges, which may cause a thinner hydration shell and smaller changes in the effective layer density.

The hypothesis of varying conformational changes with varying surface coverage is further supported by the estimation of the coverage of a densely packed monolayer of FN dimers with a monomer of 60 nm length and 2.5 nm in diameter. A surface coverage of approximately 250 ng/cm2 is determined. As seen in Fig. 8, the decrease in the effective layer density levels off at about this surface coverage, because above this coverage the adsorbing proteins cannot spread as much and the overlapping of FN molecules during adsorption should increase. By that, the adsorbing FN molecules exhibit only a less intense interaction, with the polymer substrate inducing only minor changes of the conformational state of the whole FN molecule.

The conformational changes were verified by immunofluorescence analysis with monoclonal antibodies to the heparin-binding domain near the C-terminus (Clone FNH3-8) and a domain near the primary cell binding domain (Clone FN12-8). Figure 9 shows the differences in antibody binding on the FN-coated polymer surfaces. It clearly demonstrates stronger con-formational changes near the cell binding domain (FN12-8) for POMA by the lower antibody binding, especially at low FN coverage. For the heparin-binding domain (FH3-8), only a slightly lower antibody binding is observed for POMA. These findings support the QCM-D data, pointing to stronger conformational changes of FN on the hydrophobic POMA at low coverage.

Based upon this characterisation of the state of adsorbed FN at variable surface coverage, FN heteroexchange experiments were conducted to reveal the correlation of the above findings to the FN displacement characteristics. As shown in Fig. 10, at low initial surface coverage FN was barely displaced by HSA on the hydrophilic PPMA surface. Increasing the initial surface coverage led to an increase in the relative amount of displaced FN by as high as 50% for the highest absolute surface coverage of 409 ng/cm2 (for details on HPLC results see Renner et al. 2005). In contrast (Fig. 10B), FN adsorbed onto POMA was displaced by to up to 30%, with no significant

Fig.9. Relative fluorescence intensities of antibody binding (Clone FN12-8) to FN layers on POMA (filled triangles) and PPMA (open circles)
Fig. 10. FN surface coverage versus time during protein exchange by HSA (50 |g/ml) on PPMA (A) and on POMA (B)

difference caused by the initial surface coverage. A slightly higher relative displacement of the pre-adsorbed FN species was determined only for the lowest surface coverage. Plotting the absolute FN surface coverage versus time (Fig. 11) for different initial coverages on each substrate indicates that different absolute amounts were displaced on PPMA, while approximately similar amounts were exchanged on POMA surfaces. The comparison of the absolute surface coverage at late stages of displacement on PPMA reveals a limiting non-displaceable fraction of surface-bound FN, which was concluded to be independent of the initial surface coverage.

0 10 20 30 40 SO time [h]

Fig. 11. FN surface coverage versus time during protein exchange by HSA rescaled from Fig. 10 with respect to HPLC data at initial coverage on PPMA (a) and POMA (B)

Fig. 11. FN surface coverage versus time during protein exchange by HSA rescaled from Fig. 10 with respect to HPLC data at initial coverage on PPMA (a) and POMA (B)

The displacement rate was then further evaluated by fitting the exponential decay of the observed kinetics as described earlier. As postulated earlier, in this study a non-displaceable FN fraction might occur for a low FN coverage on PPMA. By using the model equation described above (Eq. 2), this should manifest itself as a drastic increase of the second time constant. Figure 12 shows the parameters of the kinetic fits as they depend on the surface coverage. The regression coefficient of the fits was always R2 > 0.97. For the POMA surface the time constants of the fast and slowly desorbed FN species show no clear trend. However, on the PPMA surface the slow time constant was found to be one order of magnitude higher for low coverage

Fig. 12. Time constants of the exponential fit of the FN displacement in dependence on FN surface coverage. Values were determined by fits to the averaged values shown in Fig. 10 with R2 > 0.97

than for higher ones, indicating almost no displacement, comparable to the results on POMA.

Insummary, the followingconclusions canbedrawnfromtheFNadsorp-tion and heteroexchange experiments with FN layers at different degrees of surface coverage:

1. During adsorption FN builds a more rigid layer on POMA in a more compact form with stronger conformational changes as a result of the strong hydrophobic interactions with the long alkyl chains of the comonomer. The more globular structure of the protein molecules allows a high amount of water to be coupled into the FN layer.

2. Similar amounts are thus displaced independent of the surface coverage, because every molecule adsorbs in a compact form and there is no potential for further spreading due to "fixation" of the protein in the initially adsorbed shape by strong hydrophobic interactions with the substrate.

3. On PPMA the characteristics of the FN layer is apparently more determined by electrostatic interactions involving the carboxylic acid groups, leading to looser binding of molecules in a more open form, however with less overall conformational changes. We suppose that FN molecules adsorb in totally lower amounts. Less water is coupled into this FN layer, where the open form of the molecules and the electrostatic origin of their binding are thought to be the main reasons for the smaller coupled hydration shell in comparison to POMA.

Fig. 13. Schematic representation of the suggested interfacial structure of FN on hydrophilic PPMA (left) and hydrophobic POMA (right) surfaces

4. The stronger influence of electrostatic interactions on PPMA substrates leads to a weaker binding and a higher spreading potential of FN. By varying the surface coverage, FN molecules adsorb with a different degree of spreading due to the different space available. This behaviour results in a dependence of the exchange characteristics on the surface coverage with less exchange at low surface coverage.

This interpretation is based on the earlier suggested random sequential adsorption model (Evans 1993; Schaaf and Talbot 1989; Van Tassel et al. 1996) with relaxation and spreading of adsorbed proteins being dependant on the available space, and hypothetically sketched in Fig. 13. For POMA the strong interaction prevents further spreading and the HSA displacement follows a similar pattern at each surface coverage. However, on PPMA the FN molecules spread very well at low surface coverage and remain in a non-displaceable state. Above a surface coverage of approximately 100ng/cm2 the FN molecules cannot fully spread and may partially overlap, resulting in an increase of the exchange probability by the HSA molecules.

The reported findings on the different states of FN on both substrates agree well with established models and reported data for protein adsorption. Norde et al. (Arai and Norde 1990a b; Norde and Favier 1992) discussed the influence of electrostatic charge and hydrophobic properties on the adsorption of various model proteins. Regarding the surface properties, FN attaches with a higher affinity to POMA by exposing the hydrophobic core of the protein towards the hydrophobic surface, which results in an increased binding strength and stronger conformational changes at the level of the secondary structure. On PPMA the electrostatic interactions may lead to a change in the tertiary structure by opening the globular form; however, less changes occur in the secondary structure due to the weaker interaction with the substrate (as compared to POMA). Importantly, the hypothesised change in the tertiary structure is supported by the obser vations of Bergkvist et al. (2003) revealing elongated FN molecules on hydrophilic surfaces and globular compact molecules on hydrophobic substrates. Studies on FN and FN-fragment adsorption onto CH3- and COOH-terminating self-assembled monolayers (Keselowsky et al. 2003; Wilson et al. 2004) found similar conformational changes in dependence on the substrate surface chemistry by binding different monoclonal antibodies and using molecular dynamic simulations.

Taken together, it is obvious that a wealth of details on the anchorage, the kind of interaction and the conformational changes of proteins at interfaces can be derived from protein heteroexchange experiments enabled by the combination of complementary analytical techniques.

Was this article helpful?

0 0

Post a comment