Nanostructured collagen layers obtained by adsorption and drying

Nanostructured collagen layers obtained by adsorption and drying

(Parte 1 de 2)

Journal of Colloid and Interface Science 278 (2004) 63–70

Nanostructured collagen layers obtained by adsorption and drying

I. Jacquemart, E. Pamuła1, V.M. De Cupere, P.G. Rouxhet, Ch.C. Dupont-Gillain∗

Unité de chimie des interfaces, Université catholique de Louvain, Croix du Sud 2/18, 1348 Louvain-la-Neuve, Belgium

Received 7 November 2003; accepted 24 May 2004 Available online 2 July 2004


The supramolecular organization of collagen adsorbed from a 7 µg/ml solution on polystyrene was investigated as a function of the adsorption duration (from 1 min to 24 h) and of the drying conditions (fast drying under a nitrogen flow, slow drying in a water-saturated atmosphere). The morphology of the created surfaces was examined by atomic force microscopy (AFM), while complementary information regarding the adsorbed amount and the organization of the adsorbed layers was obtained using radioassays, X-ray photoelectron spectroscopy (XPS), and wetting measurements. The collagen adsorbed amount increased up to an adsorption duration of 5 h and then leveled off at a value of 0.9 µg/cm2. For samples obtained by fast drying, modeling of the N/C ratios obtained by XPS in terms of thickness and surface coverage, in combination with the adsorbed amount, water contact angle measurements and AFM images, indicated that the adsorbed layer formed a felt starting from 30 min of adsorption, the density and/or the thickness of which increased with the adsorption time. Upon slow drying, the collagen layers formed after adsorption times up to about 2 h underwent a strong reorganization. The obtained nanopatterns were attributed to dewetting, the liquid film being ruptured and adsorbed collagen being displaced by the water meniscus. At higher adsorption times, the organization of the collagen layer was similar to that obtained after fast drying, because the onset of dewetting and/or collagen displacement were prevented by the high density of the collagen felt. 2004 Elsevier Inc. All rights reserved.

Keywords: Collagen; Protein adsorption; Adsorption kinetics; Dewetting; XPS; AFM; Radiolabeling; Water contact angle

1. Introduction

Understanding and controlling the spatial organization of adsorbed protein layers on the nanometer scale is expected to bring significant progress in several fields (biosensors, immobilized enzymes, biomaterials, etc.). Organized layers of biological molecules may also serve as templates or as inspiring models to create other nanostructured materials [1–3].

Lithographic methods, which were first used in the micrometer range, have been adapted to the creation of protein patterns down to the nanometer scale [4–6].H owever, these methods are not suitable for the modification of large areas. Methods based on the deposition of colloidal particles, which create excluded areas for the proteins and then are

* Corresponding author. Fax: 32-10-472005. E-mail address: (Ch.C. Dupont-Gillain).

1 Present address: Department of Biomaterials, Faculty of Materials

Science and Ceramics, AGH—University of Science and Technology, 30 Mickiewicza Ave., 30-059 Kraków, Poland.

removed from the surface, have also been proposed [7,8]. These offer the advantage of a higher throughput. Another possible approach relies on the self-organizing properties of molecules, a feature which is commonly encountered among biological molecules [9]. For example, proteins originating from the so-called S-layer of bacteria form ordered monolayers, which can be used as nanotemplates or nanoresists [10]. Dewetting of thin liquid films on solid substrata may also produce patterns in the nanometer range. Dewetting has been investigated mainly for pure water [1] or polymer films [12,13], but it has also been applied to collagen solutions [14,15]. In comparison to other methods allowing surfaces to be nanostructured (nanolithography, selfassembly), dewetting offers different possibilities in terms of characteristic sizes and order and presents the advantage of a high throughput, which is important in view of industrial applications.

Collagen is an extracellular matrix protein (molecular mass ∼300 kg/mol) found in a wide range of vertebrate and invertebrate species. The type I collagen molecule (length ∼300 nm, diameter = 1.4 nm) is made of three helical

0021-9797/$ – see front matter 2004 Elsevier Inc. All rights reserved. doi:10.1016/j.jcis.2004.05.040

64 I. Jacquemart et al. / Journal of Colloid and Interface Science 278 (2004) 63–70 polypeptides wrapped together in a helix. Nonhelical portions, named telopeptides, are found at each end of the molecule. In vivo, collagen assembles into fibrils with a periodic structure. It can also be precipitated in vitro into several fibrillar forms [16]. Collagen offers promising perspectives for creating nanostructured protein layers, due to its shape and dimensions and to its self-assembling properties. The basement membrane of corneal epithelium was shown to exhibit a complex network of pores and fibers, with dimensions in the range of 30–400 nm. This was thought to modulate cell behavior [17]. Recently, it was also shown that the morphology of smooth muscle cells was affected by the density of collagen fibrils present on the substratum [18]. Controlling the nanometer-scale organization of adsorbed collagen layers could thus lead to a better control of cell–material interactions.

The patterns formed from collagen films prepared by spin-coating were found to depend on drying conditions and were interpreted by the occurrence of two distinct dewetting mechanisms: heterogeneous pore nucleation and spinodal dewetting [19,20].Collagen layers adsorbed on hydrophobic substrates may rearrange to form a net-like structure upon slow drying [15,21]. Such nets were used as templates to design nanostructured polymer surfaces [2].

The aim of the present study was to explore further the influence of experimental conditions on the formation of nanostructures from adsorbed collagen layers, in order to get deeper insight into the nanostructure formation process and to create a variety of nanostructures. Polystyrene, a lowcost polymer frequently used as a support for cell culture, was chosen as a substratum for adsorption. Two parameters able to affect the organization of the adsorbed layers were examined: the adsorbed amount (through the variation of the adsorption duration) and the rate of drying (with the aim of forming dewetting patterns). While the morphology of the surfaces was examined using atomic force microscopy (AFM), complementary information regarding the adsorbed amount and the organizationof the adsorbed layers was obtained using radioassays, X-ray photoelectron spectroscopy (XPS), and wetting measurements.

2. Materials and methods 2.1. Sample preparation

Polystyrene (PS) (PS05232 from BP, Zwijndrecht, Belgium) was spincoated (speed 5000 rpm, acceleration 20,0 rpm/s, time 60 s, volume 30 µl) on 12-m-diameter glass coverslips (Menzel–Gläzer, Germany) from a 10.4% (w/w) solution in toluene (analytical grade; VEL, Leuven, Belgium). The PS substrates obtained in these conditions were shown to be smooth(root-mean-squareroughnessmeasured on 5 × 5µ m2 AFM images equal to 0.2 nm) and featureless [2].

Collagen (type I from calf skin; Roche Diagnostics,

Mannheim, Germany) was received as an aqueous solution( 3m g/ml at pH 3.0). It was diluted to a concentration of 7 µg/ml in phosphate-buffered saline (PBS) (pH7.2;137mM NaCl(Merck, Leuven,Belgium), 6.44mM

KH2PO4 (VEL), 2.7 mM KCl (Merck), 8 mM Na2HPO4 (VEL)). At this pH, collagen aggregation could occur, as in- dicated in the literature [16]; we found that the aggregation in 150 µg/ml solutions, monitored using the absorbance at 400 nm, reached a maximum at pH 5.0 and was not detected at pH 7.2 (Dupont-Gillain et al., to be published).

Adsorption of collagen on PS substrates was performed as follows. Substrates were deposited in the wells of a tissue culture plate (Falcon 3043 from Becton Dickinson, Franklin Lakes, NJ, USA). Two milliliters of the collagen solution was added to each well and the samples were incubated at 37◦C for the desired period of time. When the adsorption was completed, the samples were rinsed without being exposed to air by adding 2 ml of deionized water (produced by a Milli-Q plus system from Millipore, Molsheim, France), stirring gently, pumping 3 ml of the solution, adding 3 ml of water, and repeating these two last steps four times.

Two different drying modes were then applied to the samples. A fast drying procedure was performed by flushing the samples with a nitrogen flow and storing them in a desicca- tor containing P2O5. A slow drying procedure consisted in placing the samples on a metal grid in a vessel containing a saturated solution of Na2CO3 (VEL), which ensured a relative humidity of about 95%. Before the vessel was closed a 200-µl drop of deionized water was placed on top of the samples and gentle agitation produced a uniform film of water on the sample surface. The samples were left to dry in the vessel for at least 24 h at a constant temperature of 30 ◦C.

2.2. Surface characterization

2.2.1. Atomic force microscopy

AFM images were obtained in air in the contact mode, using a commercial microscope (Nanoscope I, Digital Instru- ments, Santa Barbara, CA, USA) equipped with Si3N4 triangular levers (ThermoMicroscopes, Sunnyvale, CA, USA; nominalradius of curvature = 20 nm). Images were flattened with a third-order polynomial, using the image analysis software provided with the instrument.

2.2.2. X-ray photoelectron spectroscopy (XPS)

XPS spectra were recorded using an SSX-100 spectrometer (Model 206 from Surface Science Instruments, Mountain View, CA, USA) equipped with a monochromatizedand microfocused X-ray source (aluminum anode; 10 kV, 12 mA). The angle between the normal to the sample and the direction of photoelectron collection was 5◦. The pressure in the analysis chamber was about 10−7 Pa. Charge stabilization was achieved using an electron flood gun set at 6–8 eV and placing a grounded nickel grid 3 m above the sample surface. The analyzed area was approximately 1.4 mm2 and the

I. Jacquemart et al. / Journal of Colloid and Interface Science 278 (2004) 63–70 65 pass energy was set at 50 eV to record individual peaks. Under these conditions, the full width at half maximum of the

Au4f7/2 peak of a standard gold sample was around 1.1 eV. The order of peak analysis was C1s,O 1s,N 1s,C 1s,f ol- lowed by a survey spectrum. No modification of the C1s peak shape under X-ray irradiation was noted, indicating constant charge stabilization with time and the absence of sample degradation during analysis [23]. Data treatment was performed with the CasaXPS software (Casa Software Ltd, UK). The binding energy scale was fixed by setting the component due to carbon only bound to carbon and/or hydrogen at 284.8 eV. A linear background subtraction was used. Intensity ratios were converted into molar concentration ratios by using the sensitivity factors proposed by the manufacturer (Scofield photoemission cross sections [24]; variation of the electron mean free path according to the 0.7th power of the kinetic energy; constant transmission function).

2.2.3. Water contact angle measurement (θ)

Water contact angles were measured at ambient temperature and humidity using the sessile drop method and an image analysis of the drop profile (custom-built equipment provided by Electronish Ontwerpbureau de Boer, The Netherlands). The contact angle was determined both by measuring the tangent at the three-phases contact point and from the drop profile (height to width ratio). The water (Milli-Q) droplet volume was 0.3 µl. The surface tension of water was 71.9 mN/m. The reported values were measuredi mmediatelya fter thed ropw as depositedo nt he sample; there was no significant change of contact angle in the following seconds. For each sample, the determination was performed by averaging the results obtained on at least five droplets.

2.3. Radioassays

Collagen (type I from calf skin, Sigma–Aldrich, Bornem,

Belgium, received lyophilized and dissolved in 0.2 M acetic acid to a concentration of 1.5 mg/ml) was labeled by reductive methylation using 14C formaldehyde. Note that comparison of AFM images of adsorbed layers obtained with this collagen (before labeling) and the collagen from Roche (as used above) did not show significant differences. The labeling procedure was adapted from Means and Feeney [25] and was described in detail by Dewez et al. [26]. Comparison of adsorption isotherms obtained using a labeled collagen solution and a 20:80 mixture of labeled and unlabeled solutions showed that radiolabeling did not alter the adsorption properties of collagen.

Adsorption of labeled collagen was adapted from the method described above. PS was spin-coated on both sides of the glass coverslips and the samples were maintained in vertical position in the wells of the tissue culture plate. A volume of 5 ml of collagen solution was used. Rinsing was performed 10 times by pumping 4 ml of the liquid and adding 4 ml of water (Milli-Q). The samples were not dried before counting.

The radioactivity of PS after adsorption of labeled collagen was measured using a liquid scintillation counter (Tri- Carb 1600, Packard Instruments, Downes Grove, IL, USA). The samples were introduced into glass vials and the PS was allowed to dissolve for one night in 7 ml of scintillating liquid (PicoFluor 40, Packard Instruments). A calibration curve was obtained by counting standards containing increasing amounts of collagen and the same matrix, including a blank substratum.

3. Results

AFM images (5 × 5µ m2 areas; z = 10 nm) of collagen layers obtained by adsorption on PS for different periods of time and by either fast or slow drying are presented in Fig. 1. In the case of fast drying, after adsorption for 1 min, the adsorbed layer appears to be discontinuous and unstable under the action of the AFM tip. The same result was obtained after 10 min (not shown). When the adsorption duration is increased, a smooth layer is first observed, followed by a progressive coverage of the surface with small filamentous structures. After 24 h, a rather dense layer of these filaments (height 3–6 nm; length 250–500 nm) is formed at the surface.

In the case of slow drying, a discontinuous layer presenting a dendritic structure is observed after 1 min of adsorption. After up to 2 h of adsorption, the layer remains clearly discontinuous, with a progressive closing of the structures, leading to the formation of a netlike pattern. The height of the structures observed up to 2 h of adsorption is in the range of 6–8 nm. A transition occurs at 2 h of adsorption, at which time the reproducibility of the results is poor and the obtained layer may resemble a net or be rather continuous,with the occasional presence of holes. This is illustrated by the two AFM half-images, obtained on two independent samples. When the adsorption duration is increased beyond 2 h, the collagen layer becomes continuous and presents filamentous structures similar to those observed after fast drying. Small holes are still visible at the sample surface after 5 h of adsorption; they disappear after 24 h.

Fig. 2 presents the results of XPS analysis (Fig. 2a)a nd water contact angle measurement (Fig. 2b) onc ollagenl ayers obtained by adsorptionon PS for different periods of time and either fast or slow drying. The abscissa scale was based on the square root of the adsorption duration, for the sake of clarity. For each adsorption duration, two independent samples were characterized. The XPS results are expressed in terms of N/C molar concentration ratio, the nitrogen signal coming exclusively from the collagen. In the case of fast drying, a progressive increase of N/C is recorded up to 5 h of adsorption. Beyond an adsorption duration of 5 h, the N/C ratio is leveling off at a value of about 0.2. In the case of slow drying, the N/C ratio remains very low (close to 0.02)

6 I. Jacquemart et al. / Journal of Colloid and Interface Science 278 (2004) 63–70

Fig. 1. AFM images (5 × 5µ m2; z = 10 nm) of PS surfaces after collagen adsorption from a 7 µg/ml solution for different periods of time (from top to bottom: 1 min, 30 min, 2 h, 5 h, 24 h) followed by rinsing and fast (left) or slow (right) drying. The two half-images presented for 2 h of adsorption and slow drying illustrate the low reproducibility obtained under these conditions.

up to an adsorption duration of 30 min. Low reproducibility is observed at 2 h of adsorption, with either a low (0.05) or a much higher (0.17) N/C value. Beyond 2 h of adsorption, the N/C value levels off at a value similar to that observed after fast drying.

On pure PS, the water contact angle θ is equal to 90◦

(Fig. 2b). θ measured on adsorbed layers obtained after fast drying decreases progressively from 90◦ to about 40◦ when the adsorption duration is increased up to 2 h. Beyond 2 h, θ remains close to 40◦. On the collagen layers obtained after slow drying, θ remains close to 83◦ up to an adsorption duration of 30 min. After 2 h of adsorption, low reproducibility is found again, with either a high (81◦)o ram uch lower( 57◦)

(Parte 1 de 2)