Surface Induced Conformational Changes of a Soft Protein BSA

In general, proteins with low structural stability tend to adsorb on both hydrophobic and hydrophilic supports via a gain in conformational entropy. In the case of hydrophilic support, the soft proteins undergo more severe unfolding of the ordered structure because of a stronger increase in con-formational entropy to compensate for the repulsive coulombic interaction on an electrostatically repelling surface.

Changes in Solvation due to Adsorption

The NH/ND exchange that BSA undergoes in response to exposure to the deuterated buffer is monitored by the amide II:amide I' ratio measurements and is shown in Fig. 1. At time zero, the exchange always involves the fully CONH protein. Within 10 min of incubation in deuterated buffer, 70% of the peptide CONH in BSA becomes COND, corresponding to the fast-exchangeable peptide protons of the readily water-accessible external residues of the protein. It takes more than 2h to further exchange 12% of peptide NH of the BSA backbone. These slower-exchanging protons are located in the more internal domains or in the more hydrophobic secondary structures of the protein (Gregory and Lumry 1985). The right part of Fig. 1 shows the residual NH content and the protein adsorption density (r) recorded over time concomitantly during BSA adsorption on negatively charged hydrophilic (Fig. 1A), neutral hydrophilic (Fig. 1B) and hydrophobic support (Fig. 1C).

On hydrophobic support (Fig. 1C), BSA is adsorbed quite fast, within 10 min, while the peptide NH content decreases drastically to 5%. The larger solvation of the peptide carbonyls induced by adsorption is due to the diffusion of D2O molecules into the protein core and is directly correlated with the changes in secondary structure of BSA (Fig. 2C). The BSA molecules lose 12% of their helical domains when they become adsorbed onto the CH3-terminated SAMs, while the bent domains remain unchanged. The large unfolding of both the hydrophobic and polar helices involves the increase in the extent of self-associated and hydrated random domains, respectively (Servagent-Noinville et al. 2000). The internal hydrophobic domains of the protein are drafted towards the hydrophobic surface, increasing the solvation of 15% of the overall polypeptide backbone (Fig. 1C). The penetration of D2O molecules into the adsorbed proteins is correlated specifically with the increase in the content of the hydrated random domains (Fig. 2C). A similar structural transition has been observed for HSA adsorbed onto a reversed-phase chromatographic support (Boulkanz et al. 1997).

40 T

30-

20-

10-

40-

-t—'

30-

C

(1)

c

20-

8

X

10-

40-

30-

20-

10-

n-

0

solution

adsorption -.■—■—■—■—■—■—■

I (C)

rv^ t -

(B)

1

(A)

—5—$

60 120 180 240

60 120 180 240

Time (min)

Fig. 1. Solvation changes (-0-) and adsorbed amount (-■-) of bovine serum albumin (BSA) adsorbed at a bulk concentration of 500 |g ml-1 in deuterated phosphate buffer and pD 7.5 onto different solid supports: apure silica surface (A), ND2-terminated SAMs (B)and exterminated SAMs (C). The left part of the graphs (i.e. left of time 0) shows the isotope exchange of the BSA in solution prior to adsorption (right part). The residual NH content (expressed as a percentage of the overall peptide carbonyls) is monitored by the amide II:amide I' ratio measurements. The adsorbed amount (r) of BSA is calculated from the adsorption density equation (Sperline et al. 1987)

In contrast with the adsorption onto a hydrophobic support, the adsorption onto hydrophilic supports leads to weaker adsorbed amounts of BSA. At a pH above the isoelectric point, the negatively charged BSA molecules do adsorb onto the negatively charged silica surface, but change their sol-vation states. Indeed the residual NH content of BSA decreases from 20% to around 5% as the surface coverage increases on the negatively charged silica. This change in tertiary structure due to the contact of the protein with the repelling silica surface is also linked to the large unfolding of a-helices involving 13% of the overall peptide carbonyls (Fig. 2A).

The adsorption of BSA on the ND2-terminated SAMs is as slow as those on the repelling hydrophilic surface, but entails a weaker NH/ND exchange

Fig. 2. Comparison of secondary structures of BSA resulting from our spectral analysis in a deuterated solution and adsorbed at a bulk concentration of 500 |g ml-1 and pD 7.5 on different solid supports: a pure silica surface (A), ND2-terminated SAMs (B) and CH3-terminated SAMs (C). The given percentage of amide I' values were recorded at 4 h in solution (clear bars) and for adsorption (filled bars)

(Fig. 1B), which is corroborated by a weaker conformational change of the protein. The adsorption of BSA onto the electrostatically neutral support unfolds the most hydrophobic helical domains at 1,660 cm-1, involving around 4% of the overall backbone. The loss of these hydrophobic helical domains implies an increase in self-association. The structural change of the homologous HSA adsorbed to a positively charged hydrophilic surface such as an ion-exchanger chromatographic support also entails a small fraction of the overall peptide carbonyls (Pantazaki et al. 1998).

Adsorption Kinetics and Conformational Changes

The rates at which BSA was adsorbed onto the three different types of support were measured at a bulk concentration of 500 |g ml-1 (Fig. 1). The adsorbed amounts of BSA (r) increase with time and reach a plateau after 10 h for the negatively charged silica (not shown on Fig. 1A), after 4 h for the electrostatically neutral support (Fig. 1B) and after 60min for the hydrophobic support (Fig. 1C).

The adsorbed amount of BSA on the repelling silica surface reaches a value of 0.45 mg m-2 and 0.7 mg m-2 after 2 and 10 h, respectively. Under identical conditions of bulk concentration at a D2O-silica surface, the adsorbed amount of BSA determined by specular neutron reflection reached a limiting value of 2.5 mg m-2 at a pH close to the isoelectric point of the protein, but was found to be only 0.5 mg m-2 at pH 7 (Su et al. 1998a). The maximum amount of adsorbed protein molecules is generally observed at the adsorption pH close to the isoelectric point thanks to the minimisation of electrostatic repulsions between the protein and the support (Giacomelli et al. 1999b; Norde 1998). The structural stability of BSA is known to be strongly pH-dependent. The higher a-helical content for BSA molecules at the isoelectric point suggests that the molecular shape of BSA is more compact than at higher levels of pH (Servagent-Noinville et al. 2000). However, the negatively charged silica induces a large unfolding of the a-helices by attracting the external positively charged Lys (aminium) and Arg (iminium) residues of the BSA molecules, causing the disruption of internal salt bridges between carboxylate side chains and the positively charged residues. The adsorption kinetics of BSA at a bulk concentration of 186 ^g ml-1 on hydrophilic silica-titania surfaces reported by Kurrat et al. (1997) gave a r value of 1.3 mg m-2 in HEPES buffer and a value lower than 0.1 mg m-2 in phosphate buffer. Another research team found adsorption densities for BSA adsorbed on pure silica reaching 1.4 mg m-2 at a higher bulk concentration of 1,800 ^g ml-1 (Kridhasima et al. 1993). It should be noted that the limit value for adsorption density depends strongly upon the protein concentration (Malmsten 1998). Furthermore, the above discrepancy between the reported r values for BSA could be due not only to the variable charge densities between the various oxide surfaces used in the different works, but also to the influence of counterions on the external charge distribution of the protein and on the protein structure itself.

By assuming that the adsorption plateau corresponds to the formation of a close-packed monolayer, the average cross-sectional area of the protein adsorbed is calculated and compared to the theoretical values calculated from the molecular dimension of the protein. The calculated cross-sectional areas of a BSA molecule adsorbed onto bare silica and CH3-terminated SAMs are 155 and 97 nm2 molecule-1, respectively. One work also found cross-sectional areas of 99 and 100 nm2 molecule-1 for BSA adsorbed onto hydrophobic polyurethane films at pH 7.0 by the ATR technique (Jeon et al. 1994). According to the three-dimensional heart-shape of the BSA molecule (Table 1), the cross-sectional area of the native BSA is either 24 or 28 nm2 molecule-1, depending on an end-on or side-on orientation, respectively. The five-fold or four-fold increase in the average cross-sectional area of an adsorbed BSA molecule corresponds to a spreading of the protein either on the repelling hydrophilic or on the hydrophobic support, and is correlated in both cases to the large unfolding of the protein stereoregular structure upon adsorption.

In contrast, the average conformation of the adsorbed BSA molecules on the neutral hydrophilic support is unchanged compared to the solvated state (Fig. 2B). If the molecular shape of the protein corresponds to that of the native state, the adsorption density of 0.45 mg m-2 obtained on the ND2-terminated SAMs corresponds to the formation of an incomplete or loosely-packed monolayer. This is in agreement with the general findings of the poor affinity of proteins towards neutral surfaces (Herrwerth et al. 2003; Ostuni et al. 2001; Silin et al. 1997; Tengvall et al. 1998).

Was this article helpful?

0 0

Post a comment