Mechanisms of Fibrinogen Adsorption on Latex Particles Determined

ARTICLE pubs.acs.org/Langmuir

Mechanisms of Fibrinogen Adsorption on Latex Particles Determined by Zeta Potential and AFM Measurements Zbigniew Adamczyk,* Anna Bratek-Skicki, Paulina Da) browska, and Mazgorzata Nattich-Rak J. Haber Institute of Catalysis and Surface Chemistry, Polish Academy of Science, Niezapominajek 8, 30-239 Cracow, Poland

bS Supporting Information ABSTRACT: The adsorption of fibrinogen on polystyrene latex particles was studied using the concentration depletion method combined with the AFM detection of residual protein after adsorption. Measurements were carried out for a pH range of 3.5
11 and an ionic strength range of 10
3
0.15 M NaCl. First, the bulk physicochemical properties of fibrinogen and the latex particle suspension were characterized for this range of pH and ionic strength. The zeta potential and the number of uncompensated (electrokinetic) charges on the protein were determined from microelectrophoretic measurements. It was revealed that fibrinogen molecules exhibited amphoteric characteristics, being on average positively charged for pH 5.8 and a higher ionic strength of 0.15 M. It was also shown that in the latter case, variations in the zeta potential with the protein coverage could be adequately described in terms of the electrokinetic model, previously formulated for planar substrate adsorption. On the basis of these experimental data, an efficient procedure of preparing fibrinogen-covered latex particles of controlled monolayer structure and coverage was envisaged.

Malmsten,12 using ellipsometry, determined the kinetics of fibrinogen adsorption on methylated silica for physiological conditions (i.e., for pH 7.4 and I = 0.15 M). Systematic measurements of fibrinogen adsorption were performed by Ortega
Vinuesa et al.,5 who, using the same technique, determined the thickness of the fibrinogen layers on silicon plates for various pH values and ionic strengths. The highest coverage was observed for pH close to the isoelectric point (iep), 5.8, whereas for pH 9 and 4 the maximum coverage was much smaller. Precise kinetic measurements of fibrinogen adsorption on silicon and modified glass surfaces forming parallel-plate channels were performed using the in situ fluorescent TIRF technique by Toscano and Santore et al.7 and Wertz and Santore.13 These results were supplemented with an interesting AFM determination of the kinetics of fibrinogen adsorption for the low coverage range. Fibrinogen adsorption on various substrates such as stainless steel, nickel
titanium alloy, and pure titanium was also studied

1. INTRODUCTION The controlled adsorption of proteins at solid interfaces is a prerequisite for their efficient separation and purification by chromatography, filtration, for biosensing, bioreactors, immunological assays, and so forth. Numerous studies on adsorption mechanisms and kinetics were devoted to fibrinogen because of its fundamental role in blood clotting, platelet adhesion, thrombosis, angiogenesis, wound healing, and tumor growth. Fibrinogen’s geometrical dimensions and conformations on surfaces are known from the electron microscopy studies of Hall and Slayter1 and others.2
4 From the micrographs of fibrinogen adsorbed on mica, it was established that its molecule has a colinear, trinodular shape with a total length of 47.5 nm. The two equal end domains are spherical in shape and have a diameter of 6.5 nm; the middle domain has a diameter of 5 nm. These domains were connected by cylindrical rods, having a diameter of 1.5 nm. Similar conformations and fibrinogen dimensions were confirmed by numerous studies carried out using atomic force microscopy (AFM).5
11 Experimental works devoted to the kinetics of fibrinogen adsorption on macroscopic solid substrates are also abundant. r 2011 American Chemical Society

Received: September 29, 2011 Revised: October 25, 2011 Published: October 25, 2011 474

dx.doi.org/10.1021/la2038119 | Langmuir 2012, 28, 474–485

Langmuir

ARTICLE

concentration of fibrinogen needed in the mass balance is obtained from AFM measurements of its coverage of the mica surface. This procedure was previously standardized,11 and it was shown that the fibrinogen coverage after a fixed adsorption time is directly proportional to its bulk concentration in the solution. Using the mass balance procedure combined with AFM measurements, one can uniquely determine the surface concentration of fibrinogen on latex particles. This enables one to make a quantitative interpretation of the experimental data in terms of electrokinetic theories previously developed for flat interfaces. It should be mentioned that systematic measurements of this type were not carried out in the literature.

by Bai et al.14 These experiments were carried out for pH 7.4, I = 0.15 M at a temperature of 320 K. The amount of adsorbed fibrinogen was determined using ex situ wavelength dispersive spectroscopy (WDS). It should be mentioned that a significant spread in the maximum coverage of fibrinogen was reported in these publications, with the lowest and highest values equal to 1.4 and 11 mg m
2, respectively. As discussed in ref 15, this discrepancy can be attributed to different adsorption mechanisms of fibrinogen on various substrates and for various concentration regimes. For hydrophilic and smooth surfaces, fibrinogen adsorption occurs according to the side-on mechanism with the maximum amount of adsorbed protein equal to 1.4
1.7 mg m
2 (depending on the degree of hydration) as theoretically predicted in ref 16. This was also experimentally confirmed for the fibrinogen/mica system using the streaming potential method,11 which offers the unique possibility of in situ measurements of adsorption and desorption processes of proteins.17 However, for hydrophobic surfaces and a higher bulk fibrinogen concentration, the end-on adsorption of fibrinogen becomes feasible with the maximum coverage many times higher than for the side-on adsorption mechanism. Because the binding energy in the end-on configuration is much lower than for the side-on orientation, a partial reversibility of adsorption occurs, which can explain the source of discrepancies reported in the literature. This adsorption regime of fibrinogen was quantitatively analyzed in ref 18. It can be expected that the surface roughness, either of an intrinsic geometrical character or generated by the adsorption of polymeric species, can also promote the end-on adsorption mechanism, enhancing the maximum amount of adsorbed protein. Despite its vital significance, however, these interesting aspects of protein adsorption have not been studied in the literature. The only exception is the work of Kalasin and Santore,19 which determined by microelectrophoresis the zeta potential of negatively charged silica spheres (1000 nm diameter) covered by a controlled amount of fibrinogen (for pH 7.4 and an ionic strength varying between 5  10
3 and 0.176 M). The amount of adsorbed protein was determined by the TIRF method using fluorescently tagged fibrinogen. However, there was no attempt at quantitatively interpreting these experimental results in terms of the roughness of the silica particles, which was influenced by the ionic strength and pH. The lack of reliable experimental results concerning fibrinogen adsorption on colloid particles is astonishing in view of the vast body of literature data pertinent to analogous problems of albumins, immonoglobulins (antibodies), and other protein adsorption on polymeric particles.20
27 Therefore, the main goal of this article is to reveal mechanisms of fibrinogen adsorption on colloid latex particles, which are extensively used as carriers of proteins in many immunological assays. In particular, the reversibility problem is considered and the orientation of adsorbed fibrinogen molecules on polymeric surfaces is determined for various ionic strengths and pH values. Another important aspect of this work is determining the range of applicability of the electrokinetic theory, previously developed for particle-covered surfaces,28
30 in the present case of protein adsorption on polymeric surfaces. Our fibrinogen adsorption measurements exploit a simple but reliable concept of mass balance, where the amount of adsorbed protein per unit area (surface concentration) is quantitatively determined from the amount added to the suspension. The residue

2. MATERIALS AND METHODS Fibrinogen from bovine plasma, fraction I, type IV in the form of a crystalline powder containing 65% protein, 25% sodium chloride, and 15% sodium citrate was purchased from Sigma (F4753) and used without further purification. The protein was 93% clottable. The purity of the fibrinogen solution was checked by the dynamic surface tension measurements carried out using the pendant drop shape method. There was practically no change in the surface tension of the 10 ppm solution within 3 h, which is much longer than the typical time for experiments performed in this work. The effective bulk concentration of fibrinogen, after the dissolution of the powder in appropriate electrolyte solutions and filtration, was determined according to the procedure described in ref 11 using a high-precision densitometer (Anton Paar, type DMA5000M). The specific density of the fibrinogen solutions (nominal concentration range of 500
2000 ppm) was measured, as was the specific density of the supernatant solution acquired by the membrane filtration of these fibrinogen solutions. These stock fibrinogen solutions were then diluted prior to the adsorption experiments to the desired degree (usually 0.5
5 ppm). The densitometry technique combined with an appropriate dilution procedure was also used to prepare and characterize polystyrene latex suspensions of a desired weight concentration, ranging between 40 and 100 ppm. The negatively charged latex particle suspension was synthesized in a conventional emulsion polymerization carried out according to the procedure described in ref 31 involving a persulfate initiator. After the synthesis, the latex suspension was cleaned using steam distillation and by extensive membrane filtration. The specific conductivity of the stock latex suspension was 10 μS cm
1 at a weight fraction of 10%. Ruby muscovite mica obtained from Continental Trade was used as a substrate. The solid pieces of mica were freshly cleaved into thin sheets prior to every experiment. Water was purified using a Millipore Elix 5 apparatus. Other chemical reagents (sodium chloride, hydrochloric acid, and sodium hydroxide) were commercial products of Sigma-Aldrich and used without further purification. The temperature of the experiments was kept constant at 298 ( 0.1 K. Diffusion coefficients of fibrinogen and latex were determined by dynamic light scattering (DLS) using the Zetasizer Nano ZS Malvern instrument (measurement range of 0.6 nm to 6 μm) as previously described.10 The microelectrophoretic mobility of fibrinogen, bare latex, and fibrinogen-covered latex was measured using the laser doppler velocimetry (LDV) technique (measurement range of 3 nm to 10 μm) and the same Malvern device. Kinetic adsorption experiments aimed at determining the residual fibrinogen concentration were carried out analogously as described elsewhere.11 In the first stage, fibrinogen solutions were acquired by the membrane filtration of latex particle suspensions after the fibrinogen adsorption step. Afterward, freshly cleaved mica sheets were immersed 475

dx.doi.org/10.1021/la2038119 |Langmuir 2012, 28, 474–485

Langmuir

ARTICLE

in these fibrinogen solutions. The adsorption of fibrinogen proceeded for a desired time period (usually 5-30 min) under diffusion-controlled transport in a thermostatted cell at 298 K. The mica sheets with the adsorbed layer of fibrinogen were rinsed with a stream of pure buffer solution for more than 30 s. The micrographs of fibrinogen monolayers were acquired by AFM imaging in air using the NT-MDT Solver device with the SMENA - B scanning head. The measurements were performed in semicontact mode using a silicon probe (polysilicon cantilevers NSG-03) with resonance frequencies of 47
150 kHz (10%; the typical curvature radius of the tip was 10 nm, and the cone angle was less than 20°. Typically, the number of fibrinogen molecules was determined over 10
20 equally sized areas chosen randomly over the mica sheet. The overall number of adsorbed fibrinogen molecules was 500
1000, providing the relative precision of these to better than 5%. It was checked by precise variance tests that the fibrinogen surface concentration was statistically uniform over the entire surface area of the mica sheets exploited for this analysis. Figure 1. Dependence of the electrophoretic mobility and the zeta potential of fibrinogen on pH as regulated by the addition of HCl/ NaOH, T = 298 K, cf = 500 ppm. (1) 10
3 M NaCl. (2) 10
2 M NaCl. (3) 0.15 M NaCl.

3. RESULTS AND DISCUSSION a. Fibrinogen Characteristics in the Bulk. The bulk characteristics of fibrinogen used in our study were reported elsewhere.10 It was determined that for pH 7, fibrinogen formed stable aqueous solutions for a concentration range of up to 5000 ppm, enabling precise measurements of its diffusion coefficient and electrophoretic mobility. By exploiting this finding, we have determined that for the ionic strength range applied in this work (i.e., 10
3
0.15 M NaCl and pH 3.5
11), the average diffusion coefficient of fibrinogen molecules was 2.2  10
7 cm2 s
1 (2.2  10
11 m2 s
1), for T = 298 K. This value was practically independent of the bulk concentration of the protein, which suggests that there were no specific effects stemming from protein interactions in the bulk. From the diffusion coefficient measurements, one can determine the Stokes hydrodynamic radius of the protein using the Stokes
Einstein dependence

kT RH ¼ 6πηD

However, for pH >5.8 (which can be identified as the isolectric point of fibrinogen molecules) the electrophoretic mobility of fibrinogen became negative. Thus for pH 7.4, μe =
1.5 μm cm s
1 V
1 (
1.5  10
8 m2 s
1 V
1) for I = 10
3 M; μe =
1.1 μm cm s
1 V
1 (
1.1  10
8 m2 s
1 V
1) for I = 10
2 M; and μe =
0.63 μm cm s
1 V
1 (
0.63  10
8 m2 s
1 V
1) for I = 0.15 M. As previously discussed,10,11,15 by using such electrophoretic mobility data, one can calculate the average number of charges per molecule from the Lorenz
Stokes relationship Nc ¼

6πη  108 RH μe 1:602

ð2Þ

where Nc is expressed as the number of elementary charges (e) per molecule. It should be noted that e = 1.602  10
19 C. However, eq 2 becomes less accurate if ka > 1, that is, if the thickness of the electric double layer k
1 = (εkT/2e2I)1/2 (where I = 1/(2)(∑icizi2) is the ionic strength, ci is the ion concentrations, and zi represents their valencies) becomes smaller than the protein characteristic dimension a (e.g., the hydrodynamic diameter). For pH 3.5 and ionic strength equal to 1.3  10
3, 10
2, and 0.15 M, the numbers of uncompensated (electrokinetic) charges on the fibrinogen molecule Nc were 26, 22, and 13, respectively. Analogously, for pH 7.4 and ionic strength equal to 10
3, 10
2, and 0.15 M, Nc become negative, equal to
17,
12, and
7, respectively. (For the sake of convenience, these values are collected in Table 1.) To our knowledge, there are no such data for the pH and ionic strength range studied in our work, either theoretical or experimental, in the literature. MC simulations are too complex for the fibrinogen because of its high molecular weight and complicated structure. Besides, the simulations usually do not consider hydration effects and counterion condensation on protein molecules, which lead to a considerable compensation of the surface charge, usually by more than 90%. Analogously, the titration experiments usually give much higher values of the effective charge because of the ion replacement effects. Therefore, despite their limited precision, the above estimates of the electrokinetic charge of fibrinogen are useful for the interpretation of its adsorption behavior.

ð1Þ

where RH is the hydrodynamic radius, k is the Boltzmann constant, T is the absolute temperature, η is the dynamic viscosity of water, and D is the diffusion coefficient of fibrinogen. Using eq 1, it was calculated that RH = 10.7 nm and thus hydrodynamic diameter 2RH equals 21.4 nm, which is lower than the effective length of the molecule determined to be 47.5 nm as mentioned above. Another important parameter directly characterizing the electric charge on the protein is the electrophoretic mobility μe, defined as the average translation velocity of the protein U under a given electric field E (i.e., μe = U/E). This quantity is of primary significance because it is experimentally accessible via the microelectrophoretic method. The dependence of μe on pH for fibrinogen molecules is shown in Figure 1 for ionic strengths equal to 10
3, 10
2, and 0.15 M NaCl. As can be seen, for pH 10), F(ka) approaches unity, and for thick double layers (ka > 1), F(ka) approaches 3/2. Values of the zeta fibrinogen molecule's zeta potential calculated from eq 3 are also collected in Table 1. b. Latex Particle Characteristics in the Bulk. Analogously, bulk characteristics of the latex particles used as substrates for fibrinogen adsorption were carried out. In the first stage, the diffusion coefficient of particles was determined using the DLS method as a function of the ionic strength and pH. From these data, the hydrodynamic diameter of latex dH = 2RH was calculated using eq 1. In Figure 2, the dependence of this parameter on the ionic strength is shown for fixed pH values equal to 3.5 and 7.4, respectively. As can be seen, for high ionic strength I > 10
2 M, the hydrodynamic diameter of latex particles becomes practically constant, equal to 810 nm. However, for lower I = 10
3 M, it increases, attaining 860 nm for I = 10
3 M and 880 nm for I = 10
4 M. The relative standard deviation of these values was approximately (7%, which indicates that the latex suspension was fairly monodisperse. It should also be mentioned that because the latex particles were spherical in shape, as proven by SEM (inset in Figure 2), the hydrodynamic diameter is practically equal to the geometrical diameter of the latex particles. As discussed previously,32 the decrease in the hydrodynamic diameter with the ionic strength is a direct indication of the presence of the residue polymeric moieties on the latex particle surface forming loops or chains, often referred to as hairs.33,34 Because these polymeric moieties are charged, an increase in the

Figure 2. Dependence of the hydrodynamic diameter of the L800 latex on the NaCl concentration (ionic strength) at T = 298 K, cb = 100 ppm. (1) b, pH 7.4. (2) O, pH3.5. The inset shows an SEM micrograph of the L800 latex particles on mica.

By knowing the electrophoretic mobility, the zeta potential of fibrinogen ζ can be calculated from the Smoluchowski
Henry relationship

ζ¼

3η μ 2εFðkaÞ e

ð3Þ 477

dx.doi.org/10.1021/la2038119 |Langmuir 2012, 28, 474–485

Langmuir

ARTICLE

ionic strength decreases the electrostatic repulsion among them, leading to their collapse on the core latex surface and a decrease in the apparent size of the latex particles. This hypothesis can be confirmed using electrophoretic measurements of latex particles. Knowing the mobility, the zeta potential of latex can be determined using eq 2. The dependence of the zeta potential of latex on pH for ionic strengths of 10
3, 10
2, and 0.15 M is shown in the Supporting Information and in tabulated form in Table 1. It was determined that for a fixed ionic strength the zeta potential of latex was negative and independent of pH, which confirms the presence of strong acidic groups on the surface of the latex with a pKa value well below pH 3. Interestingly, however, the increase in ionic strength from 10
3 to 10
2 M causes a decrease in the zeta potential of latex, from
80 to
102 mV for pH 3.5. This indicates that the negative charge on the latex surface increased (in absolute terms) because of the collapse of polymeric chains toward the core of the latex. However, a further increase in the ionic strength reduced the negative zeta potential to
60 mV for pH 3.5 and I = 0.15 M. By knowing the zeta potential of latex particles, one can calculate its electrokinetic (uncompensated for) charge using the Gouy
Chapman relationship, which is valid for a symmetric 1:1 electrolyte15
 ð8εkTnb Þ1=2 eζ sinh σ0 ¼
ð4Þ 2kT 0:160

Figure 3. Dependence of the zeta potential of latex ζ on the initial bulk concentration of fibrinogen in the suspension cf. The points denote experimental results obtained by the microelectrophoretic measurements for pH 3.5, 10
2 M NaCl, T = 298 K. The dashed lines show the limiting zeta potential values for the bare latex (lower line) and fibrinogen (upper line). (1) b, bulk latex concentration of 40 ppm. (2) 2, bulk latex concentration of 60 ppm. (3) O, bulk latex concentration of 100 ppm.

of the bulk concentration of fibrinogen added to the latex suspension. To perform a quantitative analysis of this process, it is necessary to estimate the relaxation time of fibrinogen monolayer formation on latex particles. Because fibrinogen adsorption was governed by diffusion, this relaxation time can be estimated as30

where σ0 is the electrokinetic charge density of latex particles expressed in e nm
2 and nb is the number concentration of the salt (NaCl) expressed in m
3. Using the above zeta potential values determined for latex particles, one can calculate from eq 4 that for pH 3.5 σ0 =
0.068 e nm
2 (
0.0109 C m
2) for 10
3 NaCl, σ0 =
0.25 e nm
2 (
0.0400 C m
2) for 10
2 NaCl, and σ0 =
0.41 e nm
2 (
0.0656 C m
2) for I = 0.15 M. For pH 7.4, σ0 =
0.052 e nm
2 (
0.00833 C m
2) for 10
3 NaCl, σ0 =
0.28 e nm
2 (
0.0297 C m
2) for 10
2 NaCl, and σ0 =
0.40 e nm
2 (
0.0641 C m
2) for I = 0.15 M. (For the sake of convenience, these values are collected in Table 1.) As can be noticed, the electrokinetic charge density on the latex particles decreases significantly with ionic strength. The experimental results shown in Figure 1 and Table 1 suggest, therefore, that the electrostatically driven adsorption of fibrinogen on latex particles is expected for pH