Fp

ARa

Vb

a

ARR2AA

K�ETPD

1

staeTro

N5f

h1

Journal of Molecular Catalysis B: Enzymatic 110 (2014) 77–86

Contents lists available at ScienceDirect

Journal of Molecular Catalysis B: Enzymatic

j ourna l ho me pa ge: www.elsev ier .com/ locate /molcatb

acile immobilization of enzyme by entrapment using alasma-deposited organosilicon thin film

dil Elagli a,b,1, Kalim Belhacenea,b,1, Céline Viviena, Pascal Dhulsterb,enato Froidevauxb, Philippe Supiota,∗

Institut d’Electronique, de Microélectronique et de Nanotechnologie (IEMN) – UMR CNRS 8520, Université Lille Nord de France – Sciences et Technologies,illeneuve d’Ascq, FranceInstitut Charles Viollette (ProBioGEM team) – EA 1026, Université Lille Nord de France – Sciences et Technologies, Villeneuve d’Ascq, France

r t i c l e i n f o

rticle history:eceived 11 June 2014eceived in revised form2 September 2014ccepted 23 September 2014vailable online 30 September 2014

eywords:-GalactosidasentrapmentMDSOlasma polymeriffusion

a b s t r a c t

Over the years, immobilization of biologically active species such as enzymes onto solid supportgave rise to a wide range of analytical and industrial applications. The development of fast, simpleand efficient immobilization strategies is becoming of great importance in specific Biological Micro-Electromechanical Systems (BioMEMS) manufacturing. Thus, the current work focuses on an originalmethodology and mild procedure for �-galactosidase immobilization. Using as support either siliconor a thin film obtained from polymerization of 1,1,3,3-tetramethyldisiloxane (ppTMDSO) deposited byPlasma Enhanced Chemical Vapor Deposition in afterglow mode, the strategy developed here consistedin adsorption of �-galactosidase followed by its overcoating by the same siloxane plasma polymer. Aftersample washing, the enzymes were characterized to be efficiently entrapped within the porous polymermatrix while allowing the penetration and hydrolysis of the synthetic substrate ortho-nitrophenyl-�-d-galactopyranoside (o-NPG) with stability over at least 8 assays. The entrapment procedure allowedobtaining bio-functionnal coatings where �-galactosidase was expected to be included in the plasma-

polymerized films while preserving its native structure and its activity. This latter was modulated bymass transfer limitations of the substrate according to the thickness of the ppTMDSO coatings. The dry-process-based-preparation of such a thin bio-functional film (from ∼200 nm to ∼650 nm) is fast andcompatible with biochip or microreactor fabrication processes while avoiding the use of lot of chemicalsand multi-step treatments commonly encountered in enzyme immobilization procedures.

© 2014 Elsevier B.V. All rights reserved.

. Introduction

Over the years, immobilization of enzyme molecules onto solidupports gave rise to a wide range of analytical or industrial applica-ions. Because of the huge potential of proteins in various fields suchs protein microarrays, biosensors, microreactors and drug discov-ry, approaches to immobilize enzymes are widely investigated [1].

his is due to their unique properties in terms of specificity, mildeaction conditions and stereoselectivity, which provide numer-us applications ranging from high volume catalysis down to

∗ Corresponding author at: Institut d’Electronique, de Microélectronique et deanotechnologie, Avenue Poincaré, BP 60069, Université Lille 1 – Cité Scientifique,9652 Villeneuve d’Ascq Cedex, France. Tel.: +33 320 43 6581;ax: +33 320 43 41 58.

E-mail address: philippe.supiot@univ-lille1.fr (P. Supiot).1 These authors contributed equally to this work.

ttp://dx.doi.org/10.1016/j.molcatb.2014.09.014381-1177/© 2014 Elsevier B.V. All rights reserved.

ultra-sensitive biosensors [2,3]. Enzyme immobilization canprovide greater thermal and operational stability than their solu-ble forms. It also allows enzymes to be held in place throughout thereaction, following which they are easily separated from the prod-ucts, making it easy to recycle the biocatalyst and consequentlyproviding cost-reduced protocols. Furthermore, for proteomicinvestigations, enzyme grafting reduces also the auto-digestion ofproteolytic enzymes. As a consequence and following the dynamicevolution of micro- and nanobiotechnologies, the development ofvarious types of enzyme interactions with material surfaces leadsto a growing interest in understanding and controlling the immo-bilization strategies [4–7].

The immobilization of proteins on surfaces can be accomplishedby both physical and chemical methods. Different approaches

are commonly encountered such as bio-affinity binding, covalentbinding, carrier-free cross-linking, physical adsorption, and entrap-ment [8]. Both bio-affinity interaction and physical adsorptioninvolve usually a reversible interaction under specific conditions

7 Cataly

aosidcifiausaoaBo[tuea

sboomatv

taWaciEtoaloiaaa

tOsaomAm

Dvuoamie

8 A. Elagli et al. / Journal of Molecular

nd may lead therefore to reversible immobilization. Attachmentf enzymes by a chemical procedure involves the formation oftrong covalent or coordination bonds between the protein and themmobilization support. The chemical attachment involves morerastic conditions for the immobilization reaction than a physi-al attachment or entrapment. This can lead to a significant lossn enzyme activity. In addition, covalent and coordinate bondsormed between the protein and the support can lead to a changen the structural configuration of the immobilized protein [5]. Such

change in the enzymatic structure may lead to reduced activity,navailability of the active site of an enzyme for the substrates or ahift in optimum pH, which can be resumed by a reduced bindingbility [5]. Functional groups, such as amines, carboxylates, thiols,r hydroxyls, in the side chain of exposed amino acids in the proteinre used in the formation of irreversible linkages with surfaces [9].ecause multiple copies of the same functional group usually existn protein surface, covalent immobilization is usually not selective10]. When the protein contains just one copy of the reactive func-ional group, oriented immobilization can be achieved, which issually realized through selective chemical reactions or by proteinngineering [11]. In other words, covalent immobilization requires

good control over the orientation.Unfortunately, protein engineering and enzyme immobilization

till remain costly, sometimes long and tricky. Oriented immo-ilization methods that maintain the native structure or properrientation of the enzyme are desirable in order to allow furtherptimal binding to their substrates. Moreover, new “green” and fastethodologies are welcome to overcome the huge use of reagents

nd solvents but also multi-step chemical treatments which arerickier to integrate in a microfabrication process, one of the moti-ations for this work.

Physical methods of immobilization are generally milder sincehey include the attachment of protein to surfaces by various inter-ctions such as electrostatic, hydrophobic/hydrophilic and van deraals forces [5]. The major advantage in this methodology is that it

voids multi-step chemical treatments commonly encountered inhemical procedure, and can be therefore an interesting approachn micro/nanobiofunctional layer conception for Biological Microlectro Mechanical Systems (BioMEMS) manufacturing. In this con-ext, entrapment in polymeric materials has also been the subjectf notable efforts [12]. Following the fast evolution of microfluidicsnd nanotechnology, the elaboration of efficient enzyme immobi-ization processes is becoming of great interest for the developmentf new and original analytical tools or microreactors [7,13–15]. Themmobilization of biologically active species constitutes therefore

crucial step in the fabrication of BioMEMS for which the potentialpplication fields may concern biological, biochemical and medicalnalysis, environmental investigations or clinical diagnosis.

The efficiency of immobilized enzyme depends strongly on bothhe attachment strategy and the material used as a carrier matrix.ver the last few years, plasma polymerized materials have been

uccessfully used as supports for immobilization of biologicallyctive species with covalent linkages [16–20], by adsorption [21],r by entrapment [22]. The increasing use of plasma polymers isainly due to their biocompatibility and easy chemical activation.dditional advantages rise from their high thermal stability, goodechanical properties, and resistance against organic solvents [23].Polymerization induced by Plasma Enhanced Chemical Vapor

eposition (PECVD) is a powerful way to deposit materials ofarious kinds on the basis of a common precursor [24]. Whensed in interaction with biological material, the major advantagef this process is that it can be realized at moderate temper-

ture, dry state, and mild conditions, especially in afterglowode conditions. Organosilicon thin films deposited by Chem-

cal Vapor Deposition have been the subject of considerablefforts in the last years in a wide set of applications including

sis B: Enzymatic 110 (2014) 77–86

enzyme immobilization [25]. Recently, cold plasma polymeriza-tion of 1,1,3,3-tetramethyldisiloxane (TMDSO) by Remote PlasmaEnhanced Chemical Vapor Deposition (RPECVD) has shown itspotentiality in BioMEMS conception and has been successfully usedfor the simple fabrication of microchannels [26,27]. In the contextof BioMEMS manufacturing, we present a fast, innovative, and bio-compatible method for the rapid conception of bioactive coatingsusing plasma polymerized 1,1,3,3-tetramethyldisiloxane (ppT-MDSO) as carrier matrix. Here, the fabrication of a bio-functionalmicro/nanolayer is realized either by addition of enzymes to thecoating precursor in an afterglow reactor (one step immobiliza-tion) or by combining enzyme adsorption with an overcoatingby a controlled thickness of ppTMDSO (two-step immobilization).Because of the highly branched and cross-linked polymer net-work structure of plasma polymerized films [28], the enzymesare expected to be entrapped in the growing polymer network(one-step procedure) or are overcoated by the polymer layer (two-step procedure). Herein, the study focuses mainly on the two-stepprocedure consisting in the coating of adsorbed enzymes by aporous polymer deposition using plasma technology. Since adsorp-tion could be a reversible mechanism, the polymer deposit couldprotect the enzyme from leaching and retain them while allow-ing access to the substrate molecules. Using �-galactosidase asenzyme, we aim to develop a facile immobilization procedure inorder to fabricate a bio-functionnal layer where the enzymes areexpected to be included into the polymer matrix while preser-ving their native structure and their activity. RPECVD technologywas used for plasma assisted polymer deposition. In addition, weinvestigate this methodology in continuous-flow dynamic modein a top-down approach for further potential integration of thisprocess in a BioMEMS microfabrication protocol. The main reac-tions investigated were the hydrolysis of two synthetic substrates,ortho-nitrophenyl-�-d-galactopyranoside (o-NPG) and resorufin-�-d-galactopyranoside (RBG), by �-galactosidase.

2. Materials and methods

2.1. Materials

�-Galactosidase (�-d-galoctoside galactohydrolase, E.C.3.2.1.23, 10.4 U/mg, MW: 105 kDa) from Aspergillus oryzae waspurchased from Sigma–Aldrich Chemical Co. Two synthetic �-galactosidase substrates purchased from Sigma–Aldrich ChemicalCo. were employed: ortho-nitrophenyl-�-d-galactopyranoside(o-NPG, C12H15NC, MW: 301.3 g/mol) was prepared in 0.02 Macetate buffer for UV/vis measurements and sensitive fluoro-genic resorufin-�-d-galactopyranoside was employed in 0.1 Mphosphate buffer (RBG, C18H17NO8, MW: 375.3 g/mol) for spec-trofluorimetric measurements. 1,1,3,3-Tetramethyldisiloxanesolution (TMDSO, MW: 124.33 g/mol, grade 97%) purchased fromABCR & Co. (Germany) was used as precursor (monomer solution)for plasma assisted polymerization. Enzymes solutions wereprepared (i) directly in the monomer with 10% absolute ethanol,used as co-solvent, (ii) in 0.02 M acetate buffer (pH 4.5), or (iii)in absolute ethanol only. In order to inactivate the enzymes inexperiments where the feasibility of the process was tested, 1 MNa2CO3 solution was prepared to inactivate �-galactosidase.All aqueous solutions were prepared in Milli-Q pure deionizedwater (Millipore) with a resistivity of 18.2 m� cm. Immobilizationexperiments were carried out on single side polished siliconwafers ((1 0 0) oriented, p-doped, resistivity <1 � cm), obtained

from Siltronix (France). Silicon wafers were firstly cut with dimen-sions of about 1 cm × 1 cm and degreased by ultrasonicating inacetone and ethanol for 5 min each, then dried under nitrogenstream.

A. Elagli et al. / Journal of Molecular Catalysis B: Enzymatic 110 (2014) 77–86 79

Fig. 1. Schemes of (a) the RPECVD setup used to generate ppTMDSO polymer and of the related the processes, (b) vaporization of TMDSO monomer solution containingthe enzyme, (c) enzyme adsorbed on the silicon substrate and coated by a ppTMDSO film, (d) enzyme absorbed on a pre-deposited ppTMDSO film subsequently coatedby a second ppTMDSO film. The enzymes were prepared directly in the precursor monomer (procedure (i)), in 0.02 M acetate buffer (procedure (ii)) or in absolute ethanol(procedure (iii)). The procedure (i) was applied in (b) whereas procedures (ii) and (iii) were both applied in (c) and in (d) leading to five different samples. (1) Precursorv , (5) Np

2

EamopalntT3g5oi2v

2s

om

apor feeding, (2) injector, (3) plasma assisted polymerization area, (4) substratere-deposited ppTMDSO thin film, (8) resulting sample

.2. Experimental setup and polymer deposition method

Organosilicon films were prepared by “cold” Remote Plasmanhanced Chemical Vapor Deposition (RPECVD) in afterglow of

nitrogen (N2) plasma (flow rate of 1.8 standard liter perinute) (Fig. 1a) whose typical operation conditions was previ-

usly described [29]. The liquid precursor (TMDSO monomer) wasumped from the reservoir by a peristaltic pump at 0.25 mL min−1

nd injected in the afterglow by a 1 mm diameter aluminum tubu-ar injector placed at 2.5 cm from a silicon substrate. Note that noitrogen or oxygen was used as carrier gas. The monomer containshe Si O Si bridge required to obtain a polysiloxane-like structure.he RF glow-discharge (helical coupling device [30], 13.56 MHz,00 W) in N2 generates N active atoms flowing to the late after-low where the deposition takes place under a total pressure of.2 mbar, measured by a capacitive gauge (Pfeiffer). Decompositionf the monomer in the reaction cone leads to its partial polymer-zation. Deposition times of the resulting films were 5, 10, 15, and0 min for SEM thickness analysis and varied from 1 to 30 min forarious enzyme catalytic assays.

.3. Experimental procedures for enzyme immobilization andample washing

�-Galactosidase was prepared using 3 different strategies inrder to compare its activity as function of different coatingethodologies (Fig. 1). Thus, the enzyme was either (i) directly

2-plasma remote afterglow (N atom active flux, mainly), (6) �-galactosidase, (7)

solubilized in TMDSO monomer solution with 10% ethanol asco-solvent before polymerization (�-Gal-ppTMDSO, Fig. 1b), (ii)solubilized in 0.02 M acetate buffer and adsorbed either on silicon(Si-�-GalB-ppTMDSO, Fig. 1c) or on a ∼200 nm pre-deposited ppT-MDSO thin film (Si-ppTMDSO-�-GalB-ppTMDSO, Fig. 1d) beforeits overcoating by another ppTMDSO film, or (iii) adsorbed eitheron silicon (Si-�-GalEtoH-ppTMDSO, Fig. 1c) or on a ∼200 nm pre-deposited ppTMDSO thin film (Si-ppTMDSO-�-GalEtoH-ppTMDSO,Fig. 1d) from an absolute ethanol-enzyme mixture before itsovercoating by the polymer. Adsorption was carried out by10 × 2 �L deposits with micropipette or syringe of the solubilizedenzyme (in acetate buffer) or the ethanol-enzyme mixture pre-pared at 3 mg mL−1 final concentration. To control the adsorbedenzyme quantity in (iii) (ethanol is rapidly evaporated), differ-ent weightings (Sartorius ISO 9001 microbalance) were realizedfor each experiment. An amount of about 60.0 ± 5.0 �g of �-galactosidase was deposited on the support for each experiment.Then, the samples were directly exposed to the afterglow fluxwithout any washing. After immobilization, all the samples con-taining �-galactosidase were finally washed with 3 × 3 mL of0.02 M acetate buffer, as well as between each activity assay.The amount of non-immobilized �-galactosidase was detecteddue to fluorescence emission of aromatic compounds of so-called

aromatic amino acids (mostly tryptophan and tyrosine) and mea-sured by Perkin-Elmer Luminescence spectrophotometer LS 50B.Excitation was fixed at 280 nm and emission was recorded at340 nm.

8 Cataly

2

slioomaU“c

ukmtwt5iqss3

bpeawfrA

2

wtdac

aod

pesb

pattaAwuard

(on the whole support area) of 204 ± 20 nm, 502 ± 27 nm, and658 ± 35 nm. The growth-rate varies from 0.34 to 0.42 nm s−1 fordeposition times larger than 10 min. These values correspond toO2-free RPECVD conditions to limit the possible oxidative effects

0 A. Elagli et al. / Journal of Molecular

.4. Enzymatic activity assays

�-Galactosidase activity was investigated with two differentynthetic substrates. Since ONP is a colored product easy to ana-yze, the hydrolysis of o-NPG in galactose and ONP was firstlynvestigated. The coloration is detectable at basic pH in presencef Na2CO3. Batch reactions were mainly carried out in presencef excess of substrate (10 mM) in 10 mL total volume. For eacheasure, 1 mL sample was mixed with 0.5 mL Na2CO3 1 M and

bsorbance was read at 415 nm (UV/visible spectrophotometer,ltraspec-1100 pro, Amersham Biosciences). For immobilized and

free state” enzyme kinetic parameters determination, o-NPG con-entrations varied from 0.5 to 10 mM.

In a second part, the RBG sensitive fluorogenic substrate wassed in order to study the enzyme behavior on shorter timeinetics with higher sensitivity. The activity was monitored byeasuring the amount of product formed using a continuous spec-

rofluorimetric assay. Measurements of the fluorescence emissionere performed by Perkin-Elmer Luminescence spectrophotome-

er LS 50B. Excitation and emission were fixed at 571 nm and90 nm respectively. The immobilized support (1 cm2) was directly

nstalled without hindering the excitation laser beam in a 1 cmuartz cuvette which was filled up with 3 mL of 26.65 �M RGBolution in 0.1 M phosphate buffer (pH 7.0). The emission inten-ity increase at 590 nm was continuously recorded for a period of600 s at 22 ◦C. All samples were assayed similarly.

Finally, a handmade continuous-flow reactor was implementedy setup of an immobilized �-galactosidase support (1.5 cm2),repared using the procedure (iii), within a ∼0.5 cm inner diam-ter tube. A continuous flow-rate containing 10 mM of o-NPG waspplied in the channel containing the support. 36 samples of 1 mLere collected in eppendorf tubes containing 0.5 mL Na2CO3 1 M

or flow-rates varying from 0.145 to 2 mL min−1. Absorbance wasead at 415 nm (UV/visible spectrophotometer, Ultraspec 1100 pro,mersham Biosciences).

.5. Material characterization techniques

The chemical bonding states of as-deposited ppTMDSO filmsere investigated by Fourier Transform Infrared (FTIR) Spec-

roscopy using a Perkin-Elmer Spectrum One spectrometer. Theeposits were analyzed under specular reflection mode with anngle of 45◦ in the spectral range 4000–400 cm−1. The spectra wereollected and processed with the manufacturer-supplied software.

Water contact angles were measured using a GBX Digidrop. Thenalyses were performed by deposition of deionized water dropletsn single side polished silicon wafers and on ∼500 nm thick as-eposited ppTMDSO thin films.

A Zeiss Gemini Ultra 55 FEG type instrument was used toerform scanning electron microscopy (SEM) measurements forvaluating the thicknesses of the deposits after samples transver-al cleaving. Structural aspects of the samples were also evaluatedefore and after enzyme catalytic assays.

A multimode/Nanoscope IIIA (Digital Instruments) was used torobe the surface topography, and evaluate both protein adsorptionnd polymer coatings. 3D topography allowed also appreciatinghe structural aspect of the samples. The images were acquired inapping mode, in air at room temperature, using a commerciallyvailable Si3N4 tip with a 10 nm radius curvature. For each sample,FM scans (50 �m × 50 �m, 2 �m × 2 �m and 500 nm × 500 nm)ere performed in various surface positions to check the surface

niformity. Data analysis was carried out using the freely avail-ble WSxM software [31]. It allowed calculating both the averageoot-mean-square (RMS) roughness, corresponding to the standardeviation of the heights (Z values in nm) in a three-dimensional

sis B: Enzymatic 110 (2014) 77–86

AFM map, and the peak-to-valley distance (P–V), corresponding tothe mean roughness depth or Z maximal value.

3. Results

3.1. FTIR analysis and SEM-assisted thicknesses determination ofppTMDSO coatings

FTIR spectroscopy was used to study the chemical compo-sition of as-deposited ppTMDSO films (Fig. 2). The spectrumfeatures indicates that the monomer (Fig. 2b) underwent par-tial decomposition during plasma-assisted polymerization [24].The resulting plasma-polymer is formed by the fragmentation,recombination, reorganization and cross-linking of the excited andbroken monomer units following a complex mechanism [30,32].Typical band assignments of ppTMDSO films FTIR spectra canbe found in literature [33]. Vibrational absorption bands in therange 1100–1000 cm−1 reveal the Si O Si stretching mode. Bandsat 1150–930 cm−1 confirm nitrogen fixation in the film. Si Cstretching modes are mainly located around 880–600 cm−1. Mul-tiple absorption peaks around 3000 cm−1 can be assigned to thestretching modes of methyl (C H groups). Deformation modes ofCH3 are revealed at 1400–1300 cm−1. Some small contributionsfrom N H around 3600–3400 cm−1, correlatively with N inclusion,and from Si H and C N stretching modes at 2250 cm−1 are alsodetected. These characteristics are consistent with the features offilm grown from polymerization in RPECVD in our setups underO2 poor reactive media. FTIR spectra of as-deposited ppTMDSOfilms and ppTMDSO films overcoating �-galactosidase, which weredeposited according to the procedure (iii), were similar and no spe-cific signal was detected from the entrapped enzyme. These twostatements suggest that the enzyme does not significantly affectthe chemical composition of the deposited polymer. Additionally, itis worth noting that after more of 72 h of immersion either in wateror in acetate buffer 0.02 M pH 4.5, the as-deposited ppTMDSO filmrevealed no degradation. This suggests a good chemical stability ofthe film.

Investigation on 500 nm thick deposits revealed 83 ± 2◦ watercontact angle (Fig. 2a), suggesting a moderate hydrophobicity ofthe polymer film, but considered as sufficiently hydrophobic forbiological applications such as biomolecular adsorption [34]. Deter-mination of thin films thicknesses were carried out by SEM aftertransversal cleaving of the samples. Deposition times of 10, 20and 30 min correspond respectively to the average thicknesses

Fig. 2. FTIR spectra of 500 nm thick as-deposited ppTMDSO film and, as insets,(a) water droplet on 500 nm thick ppTMDSO for contact angle measurement(� = 83 ± 2◦) and (b) semi-developed formula of the TMDSO monomer.

A. Elagli et al. / Journal of Molecular Cataly

Fig. 3. Immobilized �-galactosidase activity level after 1 h for (a) enzyme andmonomer solution after plasma-assisted polymerization (�-Gal-ppTMDSO), (b)polymer-coated enzyme after adsorption from acetate buffer on silicon (Si-�-GalB-ppTMDSO), (c) polymer-coated enzyme after adsorption from absolute ethanol onsilicon (Si-�-GalEtoH-ppTMDSO) (d) polymer-coated enzyme after adsorption on a200 nm thick ppTMDSO from acetate buffer (Si-ppTMDSO-�-GalB-ppTMDSO) and(f�

osTcm[si2fi

3a

3

adMewtmoeta(ibs

3

�pwtrcafiw

e) polymer-coated enzyme after adsorption on a 200 nm thick ppTMDSO layerrom absolute ethanol (Si-ppTMDSO-�-GalEtoH-ppTMDSO). For each preparation,-galactosidase is overcoated by a ∼200 nm thick ppTMDSO film.

f oxygen containing radicals and moderate N atom flux to pre-erve the monomer and the enzyme skeletons as much as possible.hese growth-rates are about 30–40 times smaller than those typi-ally obtained in presence of O2 as carrier gas and with a 2450 MHzicrowave plasma with the same power (about 14–16 nm s−1

24]). The absence of oxygen leads to the decrease of the depo-ition rate [29] while the 13.56 MHz plasma excitation frequencys known to lead to weaker N atom concentrations than under450 MHz [35]. These two factors are sufficient to explain the lowlm growth rate.

.2. Comparison of ˇ-galactosidase activity levels between one-nd two-step immobilization procedures

.2.1. One-step entrapmentIn the procedure (i) where enzyme and liquid monomer are

ssociated, the enzyme and polymer are simultaneously depositeduring 10 min to obtain a ppTMDSO film of approximately 200 nm.onomer fragmentation and recombination form a layer where

nzyme molecules are supposed to be physically incorporatedithin the growing polymer network. The major inconvenience of

his process is that it requires high enzyme concentration becauseost of the enzyme molecules are wasted, either in the reactor

r by the bad solubility in the strong a polar precursor solution,ven with the use of ethanol as a co-solvent. Moreover, the quan-ity of entrapped enzymes is not controlled and the activity levelfter 10 min deposition is lower than in the procedures (ii) andiii) (Fig. 3a). However, simultaneous addition of enzyme by atom-zation and coating precursor could be a good technical solutionut only if the polymer presents a chemical compatibility with theolvent of enzyme [21].

.2.2. Two-step entrapmentThe previously described processes, (ii) and (iii), are used. The

-galactosidase is adsorbed and then overcoated by a ∼200 nmpTMDSO film (procedures (ii) and (iii)), which results in a “sand-ich” where the enzyme is physically included. This simple

wo-step immobilization method does not involve any chemicaleaction between the activated monomer and the enzyme thatould affect its activity, except if strong bonds between the enzyme

nd the plasma-polymerized film such as covalent bindings areormed [36]. Enzyme quantities are more precisely controlled thann the first procedure where �-galactosidase and liquid monomer

ere associated. The differences in activity levels (Fig. 3b–e)

sis B: Enzymatic 110 (2014) 77–86 81

between the procedures can be explained by the crucial role of bothenzyme solvent and adsorption support.

3.2.2.1. Role of the enzyme solvent. Adsorption of �-galactosidase isfirstly achieved from acetate buffer (procedure (ii)) and secondly,in parallel, from absolute ethanol (procedure (iii)). Indeed, the maingoal remains to keep a clean carrier surface after �-galactosidaseadsorption, taking into account that the drying efficiency is specificfor a solvent. Acetate buffer evaporation at moderate temper-ature for long time (in order to avoid enzyme denaturation)promotes the appearance of large “coffee-ring” enzyme deposi-tions which creates strong leveling inducing a brittle adherenceof the polymer. Since covalent attachment of the growing ppT-MDSO on clean silicon is stronger than attachment on adsorbedenzymes, the spot of �-galactosidase must not be too large tomaintain the ppTMDSO adherence on the support. In addition,the enzyme deposition retains certainly more water traces thatsimple hydration shells around enzymes, which reinforce the brit-tle character of the polymer film overcoating the �-galactosidase.Indeed, adhesion of ppTMDSO on moisted surfaces is poor. Conse-quently, more enzyme molecules are lost during the first washingsequence and the activity level is lower than that using absoluteethanol as solvent for enzyme adsorption. The deposition of �-galactosidase from ethanol would hence be preferred even if itdoes not involve a real kinetic-dependent adsorption mechanismdue to its fast vaporization. This volatile solvent allows to avoidthe presence of moisture or water traces on the surface duringthe plasma polymerization. However, in this case, since enzymesolubility in ethanol is lower than in water, some intermolecu-lar aggregates appear in solution. Nevertheless, residual activityof free �-galactosidase does not decrease after pre-solubilizationor in presence of ethanol (result not shown) suggesting ethanoltolerance capacity of the �-galactosidase [37]. In addition, using amixture enzyme-ethanol (Si-�-GalEtoH-ppTMDSO, procedure (iii))seems to be the best methodology to maximize the immobilizationperformance and thus the activity level (Fig. 3c).

3.2.2.2. Role of the nature of the adsorption support. The adsorptioncapacity of �-galactosidase onto ppTMDSO and silicon as supportswas investigated by both Si-ppTMDSO-�-GalB-ppTMDSO (Fig. 3d)and Si-ppTMDSO-�-GalEtoH-ppTMDSO (Fig. 3e) preparations. Inthese cases, enzyme activities were lower than for both Si-�-GalB-ppTMDSO and Si-�-GalEtoH-ppTMDSO respectively, especiallywhen using ethanol as solvent. This can be due to the crucial roleof the adsorption surface which allows or not a good adsorptioncapacity and the enzyme to be adsorbed in a good structural con-formation. As revealed by Fig. 3, this effect has as much impacton the enzyme activity as the choice of the solvent. However, amoderate effect on enzyme activity when using acetate buffer assolvent is notable. In comparison with ethanol, this can be due tothe greater enzyme leaching when using the buffer as solvent (seeSection 3.2.2.1), which may act in favor of a low visible solvent-support combination effect. In fact, the solvent plays a crucial rolein the adsorption mechanism. Due to the brittle adherence of thepolymer favored by the water traces in the Si-ppTMDSO-�-GalB-ppTMDSO preparation, the enzymes that are not well adsorbed andprotected against desorption by the ppTMDSO thin film can logi-cally be desorbed. The buffer can also play a protective role againstenzyme aggregation and adsorption when the surface propertiesdisadvantage the �-galactosidase in term of thermodynamics. Itcan thus prevent the enzyme adsorption by favoring hydrophilicinteractions with water instead of hydrophobic interactions with

the support surface. In the case of enzyme deposition from ethanol,and considering its fast evaporation, enzyme-enzyme and enzyme-support interactions are predominant and lead to a more impactingrole of the nature of the support. Nevertheless, it should be

82 A. Elagli et al. / Journal of Molecular Catalysis B: Enzymatic 110 (2014) 77–86

Fs

ciotogiit

eee(opbtSttatdtfilta[baerdslaidoi

3ˇ

gs

Table 1RMS roughness and P–V maximal value of different samples calculated from2 �m × 2 �m images by WSxM software.

Preparation RMS roughness (nm) P–V maximal value (nm)

Si-�-galactosidase 1.24 15.20

ig. 4. AFM surface topography (30 �m × 50 �m) of �-galactosidase adsorbed onilicon after drying of ethanol.

onsidered that the �-galactosidase re-solubilization in the wash-ng buffer before the activity assays may involve a re-organizationf some enzyme molecules within the polymer. Thus, the effec-ive role of the chemical nature of the support is in fact the effectf the complex solvent-support combination. The ability of �-alactosidase to retain its activity on a surface results, apart fromts intrinsic properties, from a complex balance between the forcesnvolved in its repartition such as van der Waals forces and elec-rostatic, acido–basic and steric effects.

The effect of the nature of the support is clearer when usingthanol as solvent. Depending on its isoelectric point (pI), annzyme can be adsorbed easily or not, according to the overalllectrical charge of the surface [4,38]. Negatively charged proteinA. oryzae �-galactosidase pI ∼ 4.6 [39]) is more easily adsorbednto a positively charged surface by coulombic interactions. ThepTMDSO thin film is more hydrophobic than silicon, as revealedy contact angle determination and FTIR measurements charac-erized by the presence of methyl functions. Enzyme activity ofi-ppTMDSO-�-GalEtoH-ppTMDSO preparation was not higher thanhe Si-�-GalEtoH-ppTMDSO preparation (Fig. 4), probably due tohe enzyme repartition on the hydrophobic ppTMDSO film. Whenmino acids containing hydrophobic side groups are exposed athe surface of the enzyme, hydrophobic interactions can cause aehydration of the protein surface and a shift in its structure sohat the molecules rearrange each other or on the material sur-ace to minimize the free energy [40–42], which can cause a lossn enzymatic activity [43]. This hydrophobic association can alsoead to an unfolding of the hydrophobic core toward the surface sohat hydrophobic supports are considered to be not always suit-ble for immobilized enzymes applications except using lipases4]. However, since enzymes are amphiphilic in nature, havingoth hydrophilic and hydrophobic amino acids capable of inter-ctions at a solvent–material interface, retention of immobilizednzyme activity is not just a simple matter of rendering a mate-ial more hydrophilic [4,34]. Indeed, structural changes can occurue to the competitive hydrogen bonding between a hydrophilicurface, hydrophilic amino acid groups and water, resulting in aoss of enzyme conformation [4,44,45]. Thus, the �-galactosidasedsorption performances are as important as its ability to retaints activity. Nevertheless, using a single-step adsorption procedureirectly on silicon (Si-�-GalEtoH-ppTMDSO) is the best methodol-gy to retain enzyme activity (Fig. 3c) but also to avoid multi-stepn the process.

.3. AFM and SEM investigations of the immobilized-galactosidase for Si-ˇ-GalEtoH-ppTMDSO samples

In this section, the characterization of immobilized �-alactosidase samples were realized on the Si-�-GalEtoH-ppTMDSOtructure after their preparation based on the procedure (iii). AFM

Si-ppTMDSO 2.88 26.94Si-�-galactosidase-ppTMDSO 1.11 9.59

measurements were performed in order to determine the struc-tural aspects, the surface topography, and the RMS roughnessvalues related to the polymer as well as each step of the immo-bilization process. The images have the same Z-scale (26.5 nm) fordirect comparison purpose. SEM analysis was carried out to visual-ize the uniformity of the deposits and to evaluate the efficiency ofthe adsorption procedure.

Due to the drying effect of ethanol, �-galactosidase (1022 aminoacids) rearrangement on silicon reveals some aggregates logicallyderived from those formed in ethanol (Fig. 3, right side) but alsosmall clusters with height around 10–15 nm, suggesting a depo-sition of larger protein assemblies (Fig. 4, left side and Fig. 5a).3D images of coated enzymes suggest a uniformity of the peak-to-valley (P–V) distance throughout the polymer surface (Fig. 5c).The protein molecules seem to be arranged laterally on siliconwhen they are not deposited in aggregate form, which suggest acrucial role of protein–protein and protein–surface interactions.It exists probably a balance between coulombic interaction andaggregation forces among �-galactosidase molecules that creates arelative homogeneous film. The protein repartition on the surfacestrongly depends on the �-galactosidase solvent. Nevertheless, theRMS roughness measurements confirms that both silicon-enzymeand silicon-enzyme-polymer preparations do not change drasti-cally the roughness (Table 1). In addition, P–V separation of bothpreparations is almost the same, suggesting a uniformity of enzymedeposition and overcoating. In contrast, the silicon-polymer prepa-ration shows a RMS roughness higher value as well as the P–Vseparation value, which drastically increases (Fig. 5b). In fact, chainsmobility is strongly influenced by the surface adherence in terms ofexposed atoms or functions. Indeed, the polymer has greater affin-ity with silicon than with the enzymes surface. This favors localreorganizations of the polymer chains each other around enzymesat the early growth of the layer and results in a more homogeneousfilm. In contrast, in absence of enzyme, chains mobility is reducedbecause of the strong mutual affinity of the polymer chains. Thus,both RMS roughness and P–V separation values are larger becauseno reorganization of the film occurs.

SEM images of entrapped �-galactosidase are presented in Fig. 6.They suggest that the enzyme is well overcoated when depositedeither in aggregate (Fig. 6a) or in a thin film form (Fig. 6b).

3.4. Influence of the film thickness on ˇ-galactosidase activity

From here, and since the �-galactosidase activity was higher forSi-�-GalEtoH-ppTMDSO preparation (Fig. 3c), the activity assays ofimmobilized �-galactosidase were thus realized with samples pre-pared only from the procedure (iii) with enzymes adsorbed fromethanol and overcoated but exclusively on silicon. The influence ofthe deposit thicknesses on the enzyme retention was firstly investi-gated by fluorescence and was then correlated to enzyme activity.Enzyme activity was analyzed as function of different ppTMDSOdeposit thicknesses (Fig. 7). The homogeneity of the polymer layeris function of the geometry of the reactor. Presently, the precur-

sor feeding was perpendicular to the N2 flux (Fig. 1a). Thus, thethicknesses of the obtained deposits showed some gradients onthe samples of 1.5 cm2 area. This was due to the injector-to-sampledistance and injector dimensions. This may allow a certain amount

A. Elagli et al. / Journal of Molecular Catalysis B: Enzymatic 110 (2014) 77–86 83

Fig. 5. AFM surface topography (2 �m × 2 �m) of different steps and surfaces of the adsorption from ethanol-overcoating immobilization process: (a) adsorbed �-galactosidaseon silicon, (b) ∼200 nm thick ppTMDSO polymer deposited on silicon and (c) �-galactosidase overcoated by ∼200 nm thick ppTMDSO layer (Si-�-GalEtoH-ppTMDSO).

Fig. 6. MEB images of polymer-coated �-galactosidase (Si-�-GalEtoH-ppTMDSO structure): (a) large localized coated enzyme aggregates and (b) small enzyme clusters. Theparticles in front of the aggregates in (a) results from cleaving operation.

84 A. Elagli et al. / Journal of Molecular Catalysis B: Enzymatic 110 (2014) 77–86

Fig. 7. Immobilized �-galactosidase activity as function of coating thickness for 1,2 and 3 h of reaction.

Fig. 8. Dependence vs. ppTMDSO overlayer thickness of the released enzyme quan-tb

otbmtwtcEmedm

toeskwtkodimtn

Fig. 9. 1 h monitoring of fluorescent product (resorufin) auto-degradation (yellow)

which leads to the decrease in velocity. The kinetic parameter of

ity (�g) in 3 × 3 mL of washing sequence just after plasma polymerization and justefore activity assay.

f enzymes to be released due to the relative thickness decrease athe edges of the coated sample. This phenomenon can be reducedy increasing the ppTMDSO deposition time: the thicker the poly-er film is the fewer �-galactosidase is released (Fig. 8). Note that

his release shows a linear dependence on the film thickness andould be stopped for values larger than 950 nm under this assump-

ion. However, the enzyme molecules are mainly adsorbed in theenter of the sample because of the drying effect of ethanol (Fig. 4).nzyme activity gets larger for long time kinetics and larger poly-er layer thicknesses because the enzyme molecules are more

fficiently entrapped (Fig. 7). For optimal reactions, the polymereposition time has to be chosen as a compromise between entrap-ent efficiency and substrate diffusion time.Using spectrofluorimetry for evaluating the kinetics on shorter

ime, the influence of the thicknesses was investigated in termsf diffusional limitations, which are often associated with enzymentrapment. Indeed, the hydrolysis of o-NPG analyzed by UV/vispectrophotometry is not sensitive enough to analyze short timesinetics. Since o-NPG and RBG have a relative similar moleculareight (301.3 g/mol for o-NPG vs. 375.3 g/mol for RBG), investiga-

ions on 1 h kinetics were realized by the resorufin fluorescenceinetics monitoring at 590 nm (Fig. 9). Both considered valuesf polymer layer thicknesses (∼200 and ∼500 nm) induce smalliffusional limitations and increasing mass-transfer resistances

n terms of RBG diffusion seem to occur with increasing poly-

er thickness. This is shown by the curve slope at origin, i.e.

he initial enzyme velocity, which decreases when the film thick-ess increases, indicating a diffusion-limited transport of the

and production at 590 nm by spectrofluorimetry for ∼200 nm film thickness (green)vs. ∼500 nm film thickness (blue). (For interpretation of the references to colour inthis figure legend, the reader is referred to the web version of this article.)

reactive substrate trough the polymer matrix toward the immo-bilized enzyme. However, longer time kinetics reveals that moreenzymes seem to be entrapped in the thicker ppTMDSO layers(Fig. 7). Thus, the polymer film thickness seems to constitute a keyparameter that influences the retention of enzyme molecules in thepolymer carrier matrix as well as the enzyme reaction kinetics.

3.5. Kinetic parameters determination

The determination of kinetic parameters for the immobilizedenzyme is directly influenced by the mass transfer limitations dueto the polymer when the reaction starts. Contrary to batch reac-tion where activity is maximum at the beginning of the reactionand decreases with decreasing substrate concentration, immobi-lized enzyme activity slowly increases according to the diffusionof the substrate molecules inside the polymer matrix. Then, activ-ity increases with time and substrate diffusion before its decreaseaccording to substrate consumption. Thus, mass transfer limita-tions make difficult to compare the initial rate of both batch andimmobilized enzyme reaction by taking into account the same ini-tial time (tangent at (0,0) of the curve of product concentrationas function of reaction time). Moreover, in the same way as masstransfer limitations, numerous factors can influence the rate of areaction such as species mobility, enzyme orientation or also chargeinteractions between the material and the substrate (a boundary orNernst layer for example) [4]. Since partitioning and mass trans-fer effects strongly influence the velocity of a reaction but alsothe optimal operational pH and the apparent substrate affinity,the denomination of apparent Vmax (Vapp

m ) and apparent Km (Kappm )

seems to be more appropriate.According to the Lineweaver-Burk method, the kinetic param-

eters of o-NPG hydrolysis were calculated and are presented inTable 2. Lower Michaelis–Menten constant Kapp

m but also maxi-mal reaction velocity Vapp

m values were obtained for immobilized�-galactosidase in comparison with free �-galactosidase, whosekinetic parameters were consistent with literature [46]. Kapp

m andVapp

m of immobilized �-galactosidase are smaller than Km and Vmax

values of free enzyme, suggesting an improved apparent affinitycombined with a drop of the maximal rate. Less substrate is suf-ficient to reach the declining value of maximal velocity when the�-galactosidase is entrapped in the ppTMDSO. In fact, as often inentrapped enzyme system, this plasma-polymerized thin film actsas a barrier that reduces the exchange between the enzyme and itssubstrate. Thus, less substrate is available for the �-galactosidase,

a reaction depends on all the initial rates calculated for varioussubstrate concentrations. Taking into account that diffusional lim-itations and enzyme retention in the polymer matrix influence

A. Elagli et al. / Journal of Molecular Catalysis B: Enzymatic 110 (2014) 77–86 85

Table 2Kinetic parameters (Km, Vmax, Kcat, Kcat/Km) for free and immobilized �-galactosidase for ∼200 nm polymer coating thickness.

Kma (mM) Vmax

a (�mol min−1) Vmaxa (�mol min−1 mg enz−1) Kcat (s−1) Kcat/Km (mM−1 s−1)

Free �-galactosidase 2.23 0.59 9.83 1.72 0.77Immobilized �-galactosidase 0.74 0.07

a Apparent coefficients Kappm , Vapp

m are considered in the latter case.

Fig. 10. Immobilized �-galactosidase activity (left scale, histogram) as function ofnumber of assays. The relative stability of the enzyme amount can be appreciatedabb

ekfireeewteoteirgbVtl

3

tTs�tatepapoIi

ccording to the released enzyme quantity (right scale, points connected by thelack curve) determined by spectrofluorimetry before each assay in washing acetateuffer.

nzyme initial activity, it is important to note that these presentedinetic parameters in Table 2 are specific to the immobilized con-guration with ∼200 nm polymer coating thickness. Moreover, theeleased enzymes leached from the support must not be consid-red in the specific activity appreciation. Thus, the apparent specificnzyme activity can be calculated by taking into account thesenzymes leached from the plasma-polymerized film during theashing sequence just after the polymer deposition and just before

he first activity assay (Table 2). These results were confirmed by thexpression of the catalytic constant Kcat, which translate the turn-ver of the enzyme during the reaction (turn-over number) and byhe catalytic efficiency Kcat/Km, which reflects the specificity of thenzyme for a substrate. The Kcat value of the immobilized enzymes lower than for the free enzyme, surely due to the diffusional bar-ier which limits �-galactosidase accessibility for the substrate. Theap between the catalytic efficiency values of the free and immo-ilized forms, a factor of two, is not so large as that shown by bothmax and Kcat (a factor of ∼6.5). It is due to the smaller Km value forhe immobilized enzyme, showing that the immobilization processeads to a two-time decrease of the �-galactosidase efficiency.

.6. Stability of the enzyme activity

The reproducibility of the results obtained after immobiliza-ion was investigated to appreciate the enzyme stability (Fig. 10).he first assay exhibited a maximum of activity after which a verylow decrease took place during 8 assays. The amount of released-galactosidase leached from the ppTMDSO film is about 23% for

he first washing sequence but fall to 0% before each consecutivessay, suggesting an enzyme entrapment efficient enough to keephe same amount of �-galactosidase included in the polymer forach assay (Fig. 10, black curve, right scale). The substrate androduct transfers through the polymer “membrane” can influence,fter a certain time, the exchange properties by modifying the film

orosity itself. Substrate and product molecules can be adsorbedr retained within the polymer and therefore block their diffusion.ndeed, product (o-NP) appears progressively during the wash-ng sequences of the samples. Note that a five months old sample

1.52 0.26 0.35

was assayed and exhibited the same activity level as as-depositedsample, suggesting a good stability of the film and of the coating-to-enzyme interface over the time. It shows also that the decreasein activity is mainly related to a modification of the mass transferproperties since this phenomenon appears before any degradationof the surface of the assayed samples.

A continuous-flow reactor was implemented in a top-down per-spective (result not shown). As stability is a key factor for the useof Immobilized Enzyme Reactors (IMERs), the dynamic mode ofcontinuous flow reactor was used in order to appreciate the effi-ciency of the immobilization in terms of stability and becausemechanical constraints exist when a flow-rate is applied to thiskind of system. Using overcoated �-galactosidase support, 36 sam-ples of 1 mL were collected in eppendorf tubes for a flow-rateof 0.250 mL min−1. Activity stability around 0.02 �mol min−1 wasachieved after 5–10 collected samples. This configuration showsthat a microfluidic transposition is possible since the flow does notdisturb the exchange between immobilized enzymes and free sub-strate. The reaction was effective for a large interval of flow-rates(varying from 0.145 mL min−1 to 2 mL min−1), suggesting efficiencyand exchange good enough for top-down perspective. Moreover,diffusion enhanced effect specific to microfluidics can greatly influ-ence matter exchange through the polymer film for low weightsubstrate molecules.

The conservation modes of the prepared immobilized �-galactosidase samples were also evaluated in different ways.Keeping a sample dry seems to constitute the most simple andefficient conservation mode. However, alternation between dry(conservation) and wet (assay) state between each assay couldaffect the polymer structure and can also modify the propertiesof substrate and product transfer by causing a progressive blockingof the exchange.

4. Conclusion and perspectives

In summary, cold plasma technology allows fast immobiliza-tion of enzymes while retaining their activity after several assays.The immobilization procedure described here is simple, mild, fast,and allows obtaining thin films of various thicknesses with differ-ent deposit strategies. Since adsorption is a reversible mechanism,enzyme overcoating was performed in order to optimize its reten-tion and activity stability. The thickness of the deposit seems toconstitute a key parameter that influences the retention of enzymemolecules in the polymer matrix as well as the enzyme kinetics.Finally, the enzyme immobilization is not limited by the natureof the polymer and another polymer or various ppTMDSO pre- orpost-treatments can improve enzyme entrapment and mass trans-fer performances.

Since plasma-polymerized films are suitable for the design of aninterface between biological components and physical transducers,such a methodology of enzyme inclusion is a fast and facile proce-dure that can be incorporated in a microdevice fabrication in orderto obtain a bio-functionnal coating. Remote afterglow cold plasmatechnology could bring an interesting deposition technique in the

fabrication of functionalized microfluidic MEMS. As lab-on-a-chips,IMERs and BioMEMS are part of the new tools in chemistry and biol-ogy, the methodology presented here is easily compatible with suchfields while preventing the use of various and numerous hazardous

8 Cataly

scpcsmm

A

hn

R

[

[

[[

[

[

[

[

[

[

[[

[

[

[[[

[

[[

[

[

[

[

[

[

[[[[

[[[[43] S.S. Pai, T.M. Przybycien, R.D. Tilton, Langmuir 26 (2010) 18231–18238.

6 A. Elagli et al. / Journal of Molecular

olvent commonly encountered in enzyme immobilization proto-ols. Further investigations and optimizations of the technologicalrocess as well as improvement of enzyme activity stability willertainly enable the development of new biofunctional coatings forpecific microfluidic applications. Integration of this technology inicrosystems for patterning bio-functional coatings in ppTMDSOicrofluidic channels fits into this context [27].

cknowledgments

We are grateful to Christian Malas and David George for theirelp in RPECVD experiments, Dominique Deresmes for AFM tech-ical assistance and Christophe Boyaval for SEM measurements.

eferences

[1] L.S. Wong, F. Khan, J. Micklefield, Chem. Rev. 109 (2009) 4025–4053.[2] E.-L. Ticu, D. Vercaigne-Marko, R. Froidevaux, A. Huma, V. Artenie, D. Guillo-

chon, Process Biochem. 40 (2005) 2841–2848.[3] H. Muguruma, Trends Anal. Chem. 26 (2007) 433–443.[4] J.N. Talbert, J.M. Goddard, Colloids Surf. B: Biointerfaces 93 (2012) 8–19.[5] U. Hanefeld, L. Gardossi, E. Magner, Chem. Soc. Rev. 38 (2009) 453–468.[6] R.C. Rodrigues, C. Ortiz, A. Berenguer-Murcia, R. Torres, R. Fernández-Lafuente,

Chem. Soc. Rev. 42 (2013) 6290–6307.[7] A. Sassolas, L.J. Blum, B.D. Leca-Bouvier, Biotechnol. Adv. 30 (2012) 489–511.[8] R.A. Sheldon, S. van Pelt, Chem. Soc. Rev. 42 (2013) 6223–6235.[9] F. Rusmini, Z. Zhong, J. Feijen, Biomacromolecules 8 (2007) 1775–1789.10] J.N. Yewle, Y. Wei, D.A. Puelo, S. Daunert, L.G. Bachas, Biomacromolecules 13

(2012) 1742–1749.11] P.-C. Lin, D. Weinrich, H. Waldmann, Macromol. Chem. Phys. 211 (2010)

136–144.12] K.A. Heyries, C.A. Marquette, L.J. Blum, Langmuir 23 (2007) 4523–4527.13] S.D. Psoma, P.D. van der Wal, O. Frey, N.F. de Rooij, A.P.F. Turner, Biosens.

Bioelectron. 26 (2010) 1582–1587.14] T. Honda, M. Miyazaki, H. Nakamura, H. Maeda, Adv. Synth. Catal. 348 (2006)

2163–2171.15] T. Honda, M. Miyazaki, H. Nakamura, H. Maeda, Chem. Commun. (2005)

5062–5064.16] K.S. Siow, L. Britcher, S. Kumar, H.J. Griesser, Plasma Process. Polym. 3 (2006)

392–418.17] M. Ghasemi, M.J.G. Minier, M. Tatoulian, M.M. Chehimi, F. Arefi-Khonsari, J.

Phys. Chem. B 115 (2011) 10228–10238.

[[

[

sis B: Enzymatic 110 (2014) 77–86

18] A. Abbas, D. Vercaigne-Marko, P. Supiot, B. Bocquet, C. Vivien, D. Guillochon,Colloids Surf. B: Biointerfaces 73 (2009) 315–324.

19] I. Gancarz, J. Bryjak, M. Bryjak, G. Pozniak, W. Tylus, Eur. Polym. J. 39 (2003)1615–1622.

20] I. Gancarz, J. Bryjak, G. Pozniak, W. Tylus, Eur. Polym. J. 39 (2003) 2217–2224.21] C. Amorosi, C. Mustin, G. Frache, P. Bertani, A. Fahs, G. Francius, V. Toniazzo, D.

Ruch, V. Ball, L. Averous, M. Michel, J. Phys. Chem. C 116 (2012) 21356–21365.22] P. Heyse, A. van Hoeck, M.B.J. Roeffaers, J.-P. Raffin, A. Steinbüchel, T. Stöveken,

J. Lammertyn, P. Verboven, P.A. Jacobs, J. Hofkens, S. Paulussen, B.F. Sels, PlasmaProcess. Polym. 8 (2011) 965–974.

23] P.K. Chu, J.Y. Chen, L.P. Wang, N. Huang, Mater. Sci. Eng. R: Rep. 36 (2002)143–206.

24] P. Supiot, C. Vivien, K. Blary, V. Rouessac, Chem. Vap. Depos. 17 (2011) 321–326.25] T. Hayakawa, M. Yoshinari, K. Nemoto, Biomaterials 25 (2004) 119–127.26] V. Mille, N.-E. Bourzgui, C. Vivien, P. Supiot, B. Bocquet, J. Micromech. Microeng.

18 (2008) 125026.27] A. Abbas, P. Supiot, V. Mille, D. Guillochon, B. Bocquet, J. Michromech. Microeng.

19 (2009) 045022.28] H. Muguruma, Plasma Process. Polym. 7 (2010) 151–162.29] F. Callebert, P. Supiot, K. Asfardjani, O. Dessaux, P. Gourmand, P. Dhamelicourt,

J. Laureyns, J. Appl. Polym. Sci. 52 (1994) 1595–1606.30] C. Foissac, J. Kristof, A. Annusova, P. Veis, P. Supiot, Plasma Sources Sci. Technol.

21 (2012) 055021.31] I. Horcas, R. Fernández, J.M. Gómez-Rodríguez, J. Colchero, J. Gómez-Herrero,

A.M. Baro, Rev. Sci. Instrum. 78 (2007) 013705.32] V. Krishnamurthy, I.L. Kamel, J. Polym. Sci. A: Polym. Chem. 27 (1989)

1211–1224.33] S. Abou Rich, V. Mille, C. Vivien, S. Godey, P. Supiot, Plasma Process. Polym. 7

(2010) 775–784.34] C.P. Stallard, K.A. McDonnell, O.D. Onayemi, J.P. O’Gara, D.P. Dowling, Biointer-

faces 7 (2012) 31.35] P. Merel, M. Tabbal, M. Chakar, M. Moisan, M. Ricard, Plasma Sources Sci. Tech-

nol. 7 (1998) 550–556.36] M.M. Bilek, D.R. McKenzie, Biophys. Rev. 2 (2010) 55–65.37] Z. Wu, M. Dong, M. Lu, Z. Li, J. Mol. Catal. B: Enzym. 63 (2010) 75–80.38] M. Rabe, D. Verdes, S. Seeger, Adv. Colloid Interface Sci. 162 (2011) 87–106.39] P. Urrutia, B. Rodriguez-Colinas, L. Fernandez-Arrojo, A.O. Ballesteros, L. Wilson,

A. Illanes, F.J. Plou, J. Agric. Food Chem. 61 (2013) 1081–1087.40] C. Czeslik, R. Winter, Phys. Chem. Chem. Phys. 3 (2001) 235–239.41] Y. Moskovitz, S. Srebnik, Biophys. J. 89 (2005) 22–31.42] A. Kondo, F. Murakami, K. Higashitani, Biotechnol. Bioeng. 40 (1992) 889–894.

44] C.C. You, M. De, G. Han, V.M. Rotello, J. Am. Chem. Soc. 127 (2005) 12873–12881.45] C.H. Yu, A. Al-Saadi, S. Shih, L. Giu, K.Y. Tam, S.C. Tsang, J. Phys. Chem. C 113

(2009) 537–543.46] R. Gaur, H. Pant, R. Jain, S.K. Khare, Food Chem. 97 (2006) 426–430.