2.1. Physicochemical Characteristics of Albumin and Polymer Particles
In order to facilitate a quantitative interpretation of the corona formation and particle deposition experiments, physicochemical characteristics of HSA and the polymer particle suspensions were acquired applying the above described techniques.
The diffusion coefficient of HSA molecules at pH range 3.5 to 5 and NaCl concentration of 10 to 30 mM directly measured by DLS was equal to 6.3
0.3×10
-7 cm
2 s
-1. This corresponds to the hydrodynamic diameter calculated from the Stokes–Einstein relationship equal to 7.9±0.4 nm. These values are similar to previously reported in References [
29] indicating that the HSA solutions were stable under these physicochemical conditions.
The dependencies of the albumin molecule zeta potential on pH for NaCl concentrations of 1, 10 and 30 mM are shown in
Figure 1a. As can be seen, at smaller pHs the zeta potential was positive for all NaCl concentrations attaining 36±2 mV at pH 4 in 10 mM NaCl. However, for larger pHs it rapidly decreased changing sign at pH 5. At pH 7.4 and 10 mM NaCl the zeta potential attained a negative value of - 36±2 mV.
Physicochemical characteristics were also carried out for the polymer particles. The dependence of their zeta potential on pH, calculated using the Ohshima formula for NaCl concentrations of 1, 10 and 30 mM are shown in
Figure 1b. The zeta potential was negative for the entire range of pHs assuming -40±2 mV and -50±3 mV at pH 4 and 7.4, respectively for 10 mM NaCl.
On the other hand, the diffusion coefficient of the particles determined by DLS was practically independent of pH and equal to 1.2×10
-8 cm
2 s
-1 for the NaCl concentration of 10 mM. This corresponds to the hydrodynamic diameter of 400 ±20 nm, calculated from the Stokes-Einstein formula. It should be mentioned that the hydrodynamic diameter alone, defined as the size of an equivalent sphere having the same diffusion coefficient, does not furnish information about the shape and true dimensions of the particles. These parameters were derived from scanning electron microscopy (SEM) micrographs, see
Figure 2, as an average taken from ca. 100 particles. Thus, the particle dimensions were 1100±50×220±20×220±20 nm, with the size distribution dispersity of ca 5%. It was also confirmed that their shape can be approximated by a prolate spheroid with the longer to shorter axis ratio λ=
a/b of 5 and the cross-section area in the side-on orientation equal to 0.19 mµ
2.
2.2. Formation of Albumin Corona at Polymer Particles
Functionalization of the spheroidal particles by adsorption of HSA was carried out according to the procedure described in Ref. [
29] where the changes in the electrophoretic mobility were in situ monitored by the LDV technique. The electrophoretic mobility of the particles acquired as a function of the initial albumin concentration was converted to the zeta potential and interpreted in terms of the electrokinetic model. It is worth mentioning that the time of corona formation at the polymer particles is very short compared to the adsorption at planar substrates and does not depend on the protein concentration. Using the formula derived in Reference [
48] and the parameters pertinent to our measurements: the bulk particle concentration after mixing of 50 mg L
-1,
dH = 400 nm, and the protein diffusion coefficient of 6.2×10
-7 cm
2 s
-1 one can calculate that the corona formation time was equal to ca 0.1 s, which is considerable shorter than the experimental incubation time.
In
Figure 3 the dependence of the zeta potential of particles on the initial concentration of HSA in the suspension acquired for pH 4 and 10 mM NaCl is shown. As can be seen, the initially negative zeta potential of the particles rapidly increased with the albumin concentration (denoted by
cp) and became positive for
cp larger than 0.5 mg L
-1. For still larger concentrations of albumin the zeta attained the plateau value of 15 mV, which was markedly lower than the bulk zeta potential of HSA molecules equal to 36 mV.
One can observe at for the initial albumin concentration above 1 mg L
-1 a saturated monolayer is formed at polymer particles. The mass coverage of the albumin layer can be calculated from the formula
where
is the protein mass coverage, conveniently expressed mg m
-2,
is the protein solution volume and
is the net surface area of the particles given by
where
is the particle volume,
is the particle suspension concentration and
,
are the surface area and mass of a single particle, respectively.
Considering that for spheroids
where
is the polymer particle density and
is the correction function of the axis ratio parameter accounting for the increase in the specific surface area compared to spherical particles given by
Considering Equations 2,3 one obtains the following formula for the mass average of albumin on the spheroidal particles
For spheres, where
= 1,
and for equal volumes of the protein and the particles Equation 5 simplifies to the usual form [
29,
49]
where
is the spherical particle diameter.
Considering the following parameters pertinent to the corona formation experiments: cpol = 100 mg L-1, , b = 110 nm, = 5, = 1.06 g cm-3 one can calculate from Equations 4,5 that the monolayer coverage of albumin on the particles was equal to ca 0.5 mg m-2 for cb = 1 mg L-1.
In order to obtain a more precise estimation of the monolayer coverage, the results shown in
Figure 3 were interpreted in terms of the general electrokinetic model developed in References [
30,
48] using the following formula for the zeta potential of polymer particles covered by a protein layer, denoted by
was derived
where
is the zeta potential of the particles with the protein corona,
is the dimensionless protein coverage,
ζi is the zeta potential of the polymer particles,
ζp is the zeta potential of the protein in the bulk, and
are the dimensionless functions. The
Fi function describes the damping of the flow near particle surface by the adsorbed molecule layer and the
Fp function characterizes the contribution to the zeta potential stemming from the electric double-layer surrounding the protein molecules. These functions were calculated in References [
30] applying the multipole expansion method. The dimensionless coverage occurring in Equation 7 is connected with the mass coverage via the constitutive dependence
where
Sg is the characteristic cross-section area of the albumin molecule,
is the Avogadro number and
is the molar mass of albumin, equal to 66,400 g mol
-1.
The theoretical results calculated from Equations 7,8 adequately reflected the experimental data for
smaller than 1.3 mg L
-1 (see
Figure 3) whereas at larger protein concentration the particle zeta potential attained the plateau value of 15 mV (this is depicted as the dashed horizontal line). Using this limiting concentration one can calculate from Equation 8 that the mass coverage of the protein corona was equal to 0.65 mg m
-2.
It is interesting to mention that an identical value was previously determined for silica sensor using the OWLS method [
50], whereas in Reference [
51] the value of 0.6 mg m
-2 was obtained by optical reflectometry. In Reference [
29] obtained 0.7 mg m
-2 for recombinant HSA layer on negatively charged polystyrene latex particle having the diameter of 800 nm.
In the next series of experiments the stability and of the spheroidal particles functionalized by the albumin corona, hereafter referred to as SHSA, was determined. The results, shown in
Figure 4 (for pH 4, 10 mM NaCl) were expressed as the dependence of the particle hydrodynamic diameter and the zeta potential as a function of the storage time. As can be seen, the changes in the hydrodynamic diameter and the zeta potential were negligible for the storage time up to 1600 min (27 h) that confirms an adequate stability of their suspensions.
It should be mentioned, however, that for the pH within the range 4.5 to 6.5 the hydrodynamic diameter of the particles markedly increased attaining a maximum value of ca. 1100 nm, see
Figure 5a. Most probably this increase is caused by a reversible association of the SHSA particles under end-on orientations in fibrous-like structures. This hypothesis is supported by the dependence of the zeta potential of the SHSA particles on pH (see
Figure 5b). As can be seen, zeta potential of the particles abruptly decreased with pH and vanished at pH 4.7, that can be treated as their isoelectric point. Such a low zeta potential value promoted a reversible particle association at pH around 5. At pH larger than 7 the zeta potential of the SHSA particles approached the zeta potential of the bare particles (marked as curve 2 in
Figure 5b).
2.3. Deposition of the SHSA Particles on Abiotic Surfaces
The deposition kinetics of the SHSA particles (bulk concentration 50 mg L
-1, pH 4, 10 mM NaCl) on bare mica under diffusion is shown in
Figure 6. As can be seen, the particle surface concentration (directly determined by the optical microscope enumeration technique) linearly increased with the square root of the deposition time. This behavior was adequately interpreted in terms of the theoretical results derived from the hybrid random sequential adsorption (RSA) approach, [
30,
48], depicted as the solid red line in
Figure 6.
The results shown in
Figure 6 confirm that the theoretical predictions calculated from the RSA model agree with the experimental data for
t1/2 up to 14 min
1/2 (ca. 200 min). This fact enables to conclude that the particle deposition kinetics attained the maximum value pertinent to a barrier-less transport conditions [
48]. However, one should mention, that such kinetic experiments are tedious to perform because they require an image analysis of the deposited particles layers acquired at discrete time intervals. Hence, they are not recommended for routine measurements but can rather serve as useful reference data for the interpretation of experiments performed by other techniques, such as the quartz microbalance (QCM), where the acquired signal cannot be directly related to the real particle coverage. On the other hand, an essential advantage of the QCM technique is that it yields real-time, and quasi-continuous signals that is the sensor oscillation frequency and dissipation changes, which can be related to real coverage upon a proper calibration [
52,
53,
54,
55]. Usually, the QCM kinetic measurements are carried out under flow conditions in order to accelerate the experimental run. However, this markedly increases the consumption of the particles, which is disadvantageous for expensive solutes such as proteins, virus or bacteria or the functionalized spheroidal particles, produced in a tedious synthesis. Therefore, in this work, the measurements were predominantly carried out under diffusion transport conditions that manifold reduced the consumption of SHSA particles. Primarily, in these experiments the dependence of the frequency shift on the deposition time was recorded for various overtones
(1 to 11 in our case). Then, the apparent QCM coverage,
was calculated from the commonly used formula often referred to as the Sauerbrey equation [
55]
where
Δf is the frequency shift, and
Cs is the Sauerbrey constant equal to 0.177 mg m
-2 Hz
-1 for the 5 MHz AT cut quartz sensor [
55].
A typical kinetic run recorded at pH 4, 10 mM NaCl, the bulk SHSA particle concentration of 50 mg L
-1 and the silica sensor is shown in
Figure 7a as the dependence of the QCM coverage calculated from Equation 9 on the time. For comparison, in
Figure 7b analogous results for the bare spheroidal particles without HSA corona deposited on PAH functionalized sensor are also presented.
As shown in
Figure 7a, the particle coverage abruptly increased with the time and attained after 1100 min plateau values of 20 and 5 mg m
-2 at for the first and the 11th overtone, respectively. Interestingly, the change in the particle coverage was negligible upon switching to the pure electrolyte flow (shown as the arrow and the dashed line), which can be interpreted as negligible desorption of particles. Analogous, results were obtained in the case of the bare spheroidal particle deposition on the silica sensor functionalized by the PAH macro-ion according to the procedure described in Reference [
56], see
Figure 7b. This similarity of the deposition kinetics indicates that the basic mechanism of the bare and functionalized spheroids was the same. One should mention that such a significant decrease in the QCM coverage calculated using Equation 9 with the overtone number was also observed in deposition kinetic experiments carried out for nano- and microparticles [
55,
57,
58] as well as for viruses [
33,
35]. In the case of particles forming a stiff contact with the sensor this effect can be attributed to hydrodynamic forces acting on the particle layer whose relative significance, compared to the inertia force, decreases with the oscillation frequency, i.e., the overtone number [
57].
For larger particles especially for spheroids, a decisive role plays the hydrodynamic slip effect, also inducing significant differences in the QCM coverage predicted from Equation 9 for various overtones [
56]. This indicates that the interpretation of virus and bacteria deposition kinetics investigate by QCM can be rather ambiguous without considering an adequate theoretical background developed in References [
56]. It was shown that the under the hydrodynamic slip regime, a more appropriate transformation of the frequency changes to obtain particle coverage has the following form
where
is the dimensionless correction factor of the order of unity mainly depending on the sensor roughness.
The exact value of
can be determined from the solution of the diffusion equation with the boundary condition at the sensor surface derived from the random sequential adsorption (RSA) approach [
56], which were carried out without introducing any empirical parameters.
Kinetics of the SHSA particle deposition at silica sensor expressed using the transformation defined by Equation 10 (with
equal unity) are shown in
Figure 8. As can be seen, the dependencies of the coverage on the deposition time calculated using Equation 10 for various overtones almost coincide with each other. This confirms the utility of this transformation because one does need to arbitrarily chose the overtone number. In order to increase the precision of the QCM measurements and decrease the noise ratio, one can also calculate the kinetics averaged over all overtones, which is shown as the solid green line in
Figure 8. As can be seen, the overtone averaged kinetics reasonably agrees with that derived from the RSA modeling (dashed blue line in
Figure 8). A quantitative agreement for the deposition time below 800 min can be achieved assuming that the correction factor
equal to 0.9. This indicates that the QCM results transformed according the proposed method can be applied for comparative studies of particle deposition kinetics under various physicochemical condition. Comparison of SHSA particle deposition kinetics on the silica and the gold/PAH sensors acquired at pH 4, 10 mM NaCl using the above transformation is shown in
Figure 9. It can be seen that kinetic runs calculated from Equation 10 (averaged over the overtones) for the bare spheroids on the gold/PAH sensor and the SHSA particles at the bare silica were practically identical for shorter deposition time. However, at longer times, the deposition of the SHSA particles was less effective, which was probably caused by a lower adhesion strength of the latter particles.
The utility of the approach based on the above transformation is also confirmed by the results shown in
Figure 10 where the SHSA deposition kinetics on the silica sensor at pH 4 and 7.4 are compared. One can observe that the deposition kinetics at pH 7.4 becomes practically negligible in comparison with pH 4 where it attained a maximum effectiveness governed by the bulk transport rate. It is worth mentioning that this behavior well correlates with the decrease in the particle zeta potential shown in
Figure 5b, from 20 to -50 mV at pH 4 and 7.4, respectively, whereas the zeta potential of the silica sensor was negative at both pHs, and equal to -40 and -50 mV, respectively. This behavior suggests that electrostatic interactions, which were predicted to be attractive at pH 4 and repulsive at pH 7.4, played a decisive role in the deposition of the SHSA particles.
Analogous results were reported in References [
33,
34] presenting results of QCM investigations of several bacteriophages deposition at silica and gold sensors modified by self-assembled amine- and carboxyl-terminated layers (SAMs). Electrophoretic mobility measurements showed that the capsids exhibited a positive zeta potential for pH below 5 and negative otherwise, analogously as in our case for the polystyrene particles with the HSA corona. It was confirmed in Reference [
33,
34] that the deposition kinetics of MS2 virus at the negatively charged carboxyl terminated self-assembled monolayer decreased with pH and vanished at pH 6.
Additional QCM experiments were performed with the aim to determine the influence of flow on the deposition kinetics of the SHSA particles acquired under different pHs. The results presented in
Figure 11 quite unexpectedly indicate that both at pH 4 and 7.4 the deposition rate of the SHSA particles was practically negligible compared to that observed for the bare spheroidal particles (shown as the dashed blue dashed line in
Figure 11). This effect can be interpreted as due to the flow induced desorption of the particles from the sensor because of the appearance of hydrodynamic shearing forces. Apparently these forces exceeded the particle/sensor adhesion force because of their low zeta potential equal to 20 and -50 mV for pH 4 and 7.4, respectively.
Therefore, the results presented in
Figure 11 indicate that in the case of the functionalized particles bearing protein coronas, the QCM measurements carried out under diffusion conditions furnish more reliable results compared to these carried out under flow conditions.