Bp

JD

a

ARRAA

KPHPPA

1

tdbTosiim

mpcaipmefhb[b

0d

Colloids and Surfaces B: Biointerfaces 88 (2011) 115– 120

Contents lists available at ScienceDirect

Colloids and Surfaces B: Biointerfaces

jou rn al h om epage: www.elsev ier .com/ locate /co lsur fb

acterial-growth inhibiting properties of multilayers formed with modifiedolyvinylamine

osefin Illergård, Lars Wågberg, Monica Ekepartment of Fibre and Polymer Technology, School of Chemical Science and Engineering, Royal Institute of Technology, Stockholm, Sweden

r t i c l e i n f o

rticle history:eceived 13 January 2010eceived in revised form 17 June 2011ccepted 17 June 2011vailable online 8 July 2011

eywords:olyvinylamine

a b s t r a c t

New methods are needed to fight antibiotic-resistant bacteria. One alternative that has been proposed isnon-leaching, permanently antibacterial surfaces. In this study, we test multilayers formed with antibac-terial cationic polyvinylamine (PVAm) and polyacrylic acid (PAA) in a growth-inhibition assay. Bothhydrophobically modified and native PVAm were investigated. Multilayers did reduce the bacterialgrowth, as compared to single layers. However, the sampling time in the assay was critical, as the treatedsurface area is a capacity-limiting factor. After 2 h incubation, a maximal growth inhibition of more than99% was achieved with multilayers. In contrast, after 8 h we observed a maximal growth-inhibition of 40%.

ydrophobic modificationolyacrylic acidolyelectrolyte multilayersntibacterial surfaces

At longer incubation times, the surface becomes saturated, which explains the observed time-dependenteffectiveness. The polymers giving multilayers with the strongest growth-inhibiting properties werenative PVAm and PVAm modified with C8, which also were the polymers with highest charge density.We therefore conclude that this effect is mainly an electrostatically driven process. Viability staining usinga fluorescent stain showed a high viability rate of the adhered bacteria. The multilayers are therefore morebacteriostatic than antibacterial.

. Introduction

Bacteria are a natural element in our ecosystem. In some sensi-ive situations and environments, such as hospitals, it can though beesirable to control bacterial growth. This control has traditionallyeen achieved by direct application of disinfectants and biocides.he approach is not without problems. Given the wrong dosager wrongly chosen biocide, biocide resistant strains can evolve. Inome cases, as with triclosan or silver, antibiotic resistance can benduced [1]. This approach leads to the leaching of toxic compoundsnto the environment, where they are harmful to both beneficial

icroorganisms and higher organisms.An alternative method of bacterial control is non-leaching, per-

anently antibacterial surfaces [2,3]. By immobilising cationicolymers via covalent modification, it is possible to achieve a high,onstant biocide concentration [2]. The use of cationic substrateslso offers another advantage, as the positive charges electrostat-cally attract negatively charged bacteria [4]. Since the cationicolymers are immobilised, the bactericidal effect of these poly-ers must be based on physical disruption of the bacterial cell

nvelope [5]. Two major biocidal mechanisms have been proposedor antibacterial surfaces. The first claims that polycations withydrophobic modifications interact with the outer bacterial mem-

rane of Gram-negative bacteria, thereby causing bacterial lysis2,6]. To accomplish this, the polymers must have sufficiently longranches to be able to penetrate the bacteria [7] as the bacterial cell

927-7765/$ – see front matter © 2011 Elsevier B.V. All rights reserved.oi:10.1016/j.colsurfb.2011.06.023

© 2011 Elsevier B.V. All rights reserved.

envelope is around 50 nm [8]. The second proposed mechanism isbased on findings of a charge density threshold for antibacterialsurfaces [4,8–10]. According to this theory, the surfaces displacethe stabilising counterions of bacteria, mainly divalent Mg2+ andCa2+ in an ion-exchange process similar to polymer adsorption [4].The mechanism is known for antibacterial compounds in solutionsuch as EDTA and polylysine [11]. It is also interesting to note thatmany naturally occurring surfaces are neutral or bear a negativecharge.

A drawback of the contact-active surfaces has so far been thechemistry used for surface preparation. Using covalent attach-ment, the surface preparation is often elaborate and involvesorganic solvents [2,4,7,9,10,12]. An appealing alternative wouldinstead be physical adsorption, such as the polyelectrolyte mul-tilayer (PEM) technique. The multilayer technique, based on thestepwise adsorption of oppositely charged polymers, was firstintroduced by Decher in the 1990s [13]. The simple, yet precise,process can be carried out on any charged surface, regardless ofgeometry, and can be performed in aqueous solutions. Multilayershave been used to construct anti-adhesive surfaces [14,15] as wellas to incorporate antibacterial substances for controlled releaseapplications [16–18]. However, considering the advantages of thetechnique, surprisingly few articles on contact-active antibacte-

rial PEM systems have been published [19,20,21]. Wong et al.[21] was able to reduce bacteria by using multilayers of polyca-tionic N,N-dodecyl,methyl-polyethylenimine and polyacrylic acid

1 aces B: Biointerfaces 88 (2011) 115– 120

(sbloabuePnhtpf

Iwd

2

2

mGPmdptP

op

p

wmtMsUn

orFF(tAN

TTa

16 J. Illergård et al. / Colloids and Surf

PAA). However, the multilayers were constructed with organicolvents for the cationic polymer. In the study they found theactericidal effect to improve by higher number of layers. Simi-

arly, in the work of Westman et al. [20], the antibacterial activityf a hydrophobically modified polyvinylamine (PVAm) and PAAgainst Escherichia coli was found to increase with increasing num-er of layers. The highly charged weak polyelectrolyte PVAm wassed together with PAA to form PEM-covered surfaces on regen-rated cellulose in aqueous solutions in room temperature. TheVAm/PAA multilayer system has been studied in detail [22]. Theative PVAm and especially the hydrophobically modified PVAmsave been shown to have excellent antibacterial properties in solu-ion [23,24]. In the study by Westman et al. [24] the antibacterialroperties of PVAm in suspension were increased more than ten-old by hydrophobically modifying the polymer backbone.

In the present contribution, the work by Westman et al. andllergård et al. is continued by screening immobilized polymers

ith different modifications and different numbers of layers toevelop mechanism of the observed growth-inhibiting effect.

. Experimental

.1. Materials

Polyvinylamines with different substituents and degrees ofodification (Table 1) were supplied by BASF SE (Ludwigshafen,ermany). The molecular weight was 250 kDa for the unmodifiedVAm and 340 kDa for the modified PVAm. The details of the poly-er synthesis can be found elsewhere [25]. The polymers were

ialysed and freeze-dried prior to use. Polyvinylamines are weakolyectrolytes; i.e. their charges depend on pH and charge densi-ies have been previously determined for all polymers used exceptVAm-C4 [24]

Anionic polyacrylic acid (PAA) (Sigma) with a molecular weightf 240 kDa according to the supplier was used without furtherurification.

MilliQ water (MQ) (Millipore, Solna, Sweden) was used to pre-are all polymer solutions.

Commercially available cover slides (VWR, Stockholm, Sweden)ere used as substrates for multilayer formation. The slides wereade out of pure, white glass and had a diameter of 13 mm. Prior

o use, the slides were cleaned by rinsing with a sequence ofQ–EtOH–MQ and hydrolysed in 10% NaOH for 30 s. Finally, the

lides were plasma treated at 10 W for 30 s at reduced air pressure.ntreated glass slides were used without additional cleaning as aegative control in the growth experiments.

Two commercial antibacterial fabrics treated withctadecyldimethyl(3-trimethoxysilylpropyl) ammonium chlo-ide (AEGIS Microbe shield, Midland, USA), hereafter designatedabric1 and Fabric2, were used as positive reference samples.abric1 consisted of 100% polyester fabric for clean room usePrecision fabrics, Bamberg, Germany). Fabric2 consisted of cotton

erry intended for household usage (LaRedoute, Borås, Sweden).s the terry absorbed water, it was first saturated with 100 mMaCl before testing (Fig. 1).

able 1he different PVAm polymers used in the present study. The polymer properties areccording to the supplier’s specifications.

Polymer Substituent Hydrolysis Degree of substitution

PVAm – 100% –PVAm-C4 C4 92.5% 100%PVAm-C6 C6 90.7% 30%PVAm-C8 C8 90.7% 10%

Fig. 1. The principal polyvinylamine polymer structure. In this study, the hydropho-bically modified PVAm corresponds to a p of 1, 3, or 5.

2.2. Bacteria

The test organisms for the growth-inhibition assay were Gram-negative E. coli ATCC 11775 obtained from SIK (Göteborg, Sweden)and Gram-positive Bacillus subtilis (MERCK, Solna, Sweden). Thebacteria were grown in tryptone glucose extract broth (TGE) (BDDifco, Stockholm, Sweden) at 37 ◦C with continuous shaking.

2.3. Methods

2.3.1. Polyelectrolyte multilayersPolyelectrolyte multilayers were formed on negatively charged

cover slips using PVAm as the cationic polymer and PAA as theanionic polymer. The conditions for the multilayer build-up werechosen for maximal adsorption, based on results from a previousstudy [22]. The polymer concentration was 0.100 g/L, and a salt con-centration of 100 mM NaCl was used. The pH was adjusted to 7.5for PVAm and 3.5 for PAA. Salt solutions with the same salt con-centrations and pH as the polymer solutions were used to rinsethe surfaces after each polymer adsorption step. The samples wererinsed with salt solution after the final adsorption step and werethereafter air dried. All future references to the multilayer thin filmsin this report will be denoted as “Polymer name–x”, where x is thenumber of monolayers. Uneven numbers are capped with PVAm,whereas even implies that PAA is in the outer layer.

2.3.2. Contact angle analysisThe contact angles for water were measured with a CAM 200

(KSV, Helsinki, Finland).

2.3.3. Growth-inhibition experimentsGrowth medium consisting of 10% TGE and 90% 100 mM NaCl

was inoculated with E. coli or B. subtilis to a final bacterial concen-tration of 103 colony forming units (CFU)/mL. This growth-mediummixture was used to limit the maximal growth, and to facilitatespectroscopic readings. One milliliter of the suspension was addedto each test surface placed in a multiwell plate with a bottomdiameter of 15 mm (Nunc, Roskilde, Denmark). Each sample con-dition was tested by two samples. The plates were incubated at30 ◦C with 90 rpm shaking. The growth was analysed by transfer-ring two aliquots of 200 �L from each test sample to a transparent96-well microplate (Nunc) and reading the optical density (OD) ofthese aliquots at 540 nm (Labsystems Multiskan MCC/340, ThermoScientific, Sweden). The values were corrected for backgroundadsorption and were compared to the negative control by calcu-lating the relative growth according to Equation 1

Relative growth = ODsample

ODreference(1)

where ODsample is the OD averaged over the two test samples at540 nm and ODreference is the average OD for the reference samples(untreated glass slides) at the same wavelength.

J. Illergård et al. / Colloids and Surfaces B: Biointerfaces 88 (2011) 115– 120 117

0

10

20

30

40

50

60

70

80

90

control 1-PVAm 3-PVAm 5-PVAm 7-PVAm

Cont

act a

ngle

()

Layer number

PVAm

PVAm-C4

PVAm-C6

PVAm-C8

Fo

2

OLaOD

3

3

astlcHadrtr

3

8wu(tiwsaTl2ambu

Fig. 3. The relative growth of E. coli in the growth inhibition assay after (a) 2 h and(b) 8 h. The sampling time is an important factor, which indicates a capacity limitof the treated surfaces. A 100% relative growth corresponds to the reference. Errorbars in (b) represent the standard deviation.

Fig. 4. The relative growth of B. subtilis in the same growth inhibition assay after (a)

ig. 2. Contact angles of the water–air interface for the different layers and polymersn cleaned glass substrates.

.3.4. Viability testingThe multilayer-treated slides in the wells showing the lowest

D after 11 h incubation were stained using the fluorescent BaclightIVE/DEAD kit (Invitrogen, Täby, Sweden). The results were visu-lised using an epifluorescence microscope (Olympus BH-2-RFCA,lympus, Solna, Sweden) equipped with a digital camera (OlympusP10, Olympus).

. Results

.1. Surface analysis

As multilayers produced with modified and unmodified PVAmnd PAA have previously been studied [22], only supplementaryurface analysis by contact-angle analysis was performed. The driedreated surfaces had a contact angle between 55◦ and 80◦. Multi-ayers with unmodified PVAm generally gave a somewhat lowerontact angle, although still rendering the surface less hydrophilic.ere, the control sample was cleaned by the rinsing sequencend plasma treated prior to the contact angle measurements, asescribed in the experimental section for the multilayer substrate,esulting in a contact angle of 5◦. The untreated reference used inhe antibacterial assay had a contact angle of 79.7◦, i.e. in the sameange as the treated slides (Fig. 2).

.2. Microbial testing

The bacterial growth was analysed every other hour during h by measuring the optical density. The timing of the samplingas shown to be an important factor. After 2 h, all test surfacessed resulted in reduced growth for both E. coli and B. subtilisFigs. 3 and 4). For some samples, the growth was below the sensi-ivity limit for the spectrophotometer (around 105 CFU/mL). Whennstead comparing the data after 8 h, a higher relative growth

as observed although there were still differences between theamples. At this point, the bacterial growth of the reference waspproximately 1.5 × 108 CFU/mL according to a calibration curve.he sensitivity limiting factor of the spectrophotometer introducedarge standard deviations in the calculations for samples after

h. Therefore standard deviations are only presented for the data

fter 8 h. The unmodified polyvinylamine (PVAm) and the polymerodified with a C8 chain (PVAm-C8) showed the most growth inhi-

ition, although bacterial growth was present in both samples. These of multilayer increased the effect, but overall three layers max-

2 h and (b) 8 h. A 100% relative growth corresponds to the reference. Error bars in (b)represent the standard deviation. As with E. coli, the sampling time was importantfor the result.

118 J. Illergård et al. / Colloids and Surfaces B

0.00

0.02

0.04

0.06

0.08

0.10

0.12

0.14a

b

0 2 4 6 8 10

OD

540

nm

(a.

u.)

Time (hours)

Untreated

Fabric1

PVAm 3 layers

PVAm-C8 3 layers

0.00

0.02

0.04

0.06

0.08

0.10

0.12

0.14

0 2 4 6 8 10

OD

540

nm

(a.

u.)

Time (h ours)

Untreated

Fabric1

PVAm -B 3 layers

PVAm -C8 3 layers

Fig. 5. The growth curves monitored over 8 h for (a) E. coli and (b) B. subtilis in thegl

iet

titafi

wbitewbdaf

late a theoretical charge density in the range of 1014–1015 N+/cm2

rowth-inhibition assay. Using the treated surfaces delayed the entrance into theog phase.

mized the effect. After 8 h, a surface treated with three monolayersxhibited a bacterial growth of approximately 60% of the growth inhe control.

By analysing the growth curves over time for the negative con-rol, Fabric1, PVAm–3 and PVAm-C8–3, the effect of an early growthnhibition becomes even clearer (Fig. 5). The antibacterial Fabric2,he terry, showed an increased growth compared to the referencet all sampling times, and thus the results are not shown in thegures.

Epifluorescence microscopy with the Baclight LIVE/DEAD kit,hich stains the viable bacteria green and membrane-damaged

acteria red, was used to estimate the bacterial density and viabil-ty on surfaces treated with PVAm and PVAm-C8. As seen in Fig. 6,he bacterial density is sharply increased with the use of multilay-rs as opposed to a single layer. The maximum bacterial densityas achieved for three layers (Fig. 6b). The viability of the adhered

acteria is high, as indicated by the green stain. The multilayers

o thus have a bacteriostatic rather than biocidal effect. Bacterialdhesion to the reference was low. There were similar tendenciesor all surfaces modified by the multilayer technique and for both

: Biointerfaces 88 (2011) 115– 120

bacterial strains. The woven surface of Fabric1 made it difficult tocompare to the multilayered samples, but a majority of the bacteriawere viable (not shown). Fabric2 was impossible to compare due toboth high background fluorescence and the structure of the fabric.

Surfaces modified with an even number of layers, i.e. with PAAoutermost, were tested to validate if the anionic polymer PAA hadany influence on the bacteria-reducing effect. When tested thegrowth in the wells with PAA exposed at the surface was compa-rable to the reference (Fig. 7). The growth-inhibiting effect can befully attributed to the presence of an outer layer of PVAm.

4. Discussion

The results presented show that multilayers of PVAm and PAAdecrease bacterial growth. The most prominent observed differ-ence was between one single adsorbed PVAm layer and three layers(Figs. 3 and 4). After three layers, the effect generally was slightlyreduced. The difference is also prominent when observing thesurface-adhered bacteria. One single layer PVAm showed low bac-terial density, whereas for three monolayers the surface was fullybacteria covered (Fig. 6). After three layers, the density remainedhigh or was slightly reduced. The observed decrease in the mediasuspended bacteria appears therefore to be directly linked to thebacterial-adhering capacity of the surface.

It is also of interest to evaluate the performance of individ-ual polymers. Antibacterial evaluation in solution has shown thathydrophobic modification increases the antibacterial efficiency ofpolyelectrolytes [24]. It was therefore expected that the same effectin immobilised multilayers. Instead, multilayers with unmodi-fied PVAm provided effective growth-inhibiting surfaces, thoughmarginally. Previous adsorption studies of the polymers haveshown that multilayers formed with both unmodified and modi-fied PVAm/PAA incorporate the polymer in a flat conformation onthe surface [22]. The polymers are therefore most likely too rigidlyattached to be able to penetrate the bacterial cell envelope, a pro-cess that the hydrophobic modifications are thought to facilitate.The contact angle measurements gave no difference between themodified and unmodified PVAm, which also points towards thatthe hydrophobic chains are of minor importance for the resultingsurface properties (Fig. 6). The effect of the immobilized polymercan instead be considered as a result of surface charge and struc-ture at the solid–liquid interface, although a synergistic effect ofincreased total surface hydrophobicity and charge cannot be ruledout. Several studies have identified surface charge as the drivingmechanism for antibacterial surfaces [4,8,10,19]. Charge has alsobeen related to an inhibition in bacterial growth by viable bacteria,as also seems to be the case in the present work. Studies by Gotten-bos et al. [26] showed that for Pseudomonas aeruginosa, a commonpathogen, the adhesion on a positively charged surface was twiceas high as on a negatively charged surface. However, growth onthe positively charged surface was virtually absent. Murata et al.[8] found that a high bacterial adhesion reached a lower limit onlyslightly below the charge threshold, 5 × 1015 N+/cm2. As the chargedensity drops below 2 × 1015 N+/cm2, adhesion was found to be vir-tually absent. Since it is very difficult to experimentally measuresurface charge with the high precision needed, we have conductedno such measurements in the present work. Charge density mea-surements on polymers in solution have previously shown thatPVAm and PVAm-C8 have the highest charge densities in solutionboth at low and at physiological pH [24]. Combining the measuredcharge of the polymers with the reflectometry data [22], we calcu-

for the third polymer layer. However, a large fraction of this chargeis consumed in the formation of the multilayer. Experiments withan outermost PAA layer do indeed show that cationic charge is an

J. Illergård et al. / Colloids and Surfaces B: Biointerfaces 88 (2011) 115– 120 119

F , and (b f layer

ieffts

biawjasfrtitbirwpas

Foo

ig. 6. Fluorescence microscopy images of E. coli on surfaces treated with (a) 1, (b) 3y the green stain) whereas the bacterial densities are dependent on the number oeferences to color in text, the reader is referred to the web version of the article.)

mportant factor, as these samples exhibited no growth-inhibitingffect. As PAA is anionic, the bacteria will be repelled from the sur-ace by electrostatic interaction forces. Similar low adhesion resultsor PAA have been observed by Yang and Seo [15], who used mul-ilayers of polyacrylamide and PAA to make a cell anti-adhesiveurface.

A limiting factor of contact active surfaces is the distanceetween the bacteria and the treated surface. In this case the driv-

ng force for bacterial adhesion is the release of counterions upondhering to the charged surface. The bacteria close to the surfaceill logically be the first to adhere. Provided the bacteria are sub-

ected only to Brownian motion, i.e. with use of an immotile strainnd no external stirring, there will be a depletion of cells close to theurface. Other bacteria must diffuse over this distance to the sur-ace. Since bacteria in these circumstances are large (0.74–4 �m, aseported by Maillard [11]) the number of bacteria–surface encoun-ers will be low. With stirring, the number of possible encountersncreases by increasing the number of surface encounters, evenhough the probability of all bacteria adhering to the surface cane considered low. Therefore, bacterial growth in the suspension

s not unexpected. This limitation is indicated in the results. Theelative growth was comparably lower after 2 h than after 8 h. Still,

e observed a reduction even after this relatively long incubationeriod. To get a more unbiased view of the effect, it would be desir-ble to use a lower volume to surface ratio. Such an experimentaletup would, however, lead to other practical problems in the anal-

1-PVA m 2-PAA 3-PVA m 4-P AA 5-PVA m

02

04

06

08

01

00

12

0

Re

lative

gro

wth

(%

)

Layer nu mber

B. subtili sE. coli

ig. 7. The relative growth of bacteria on multilayers with PVAm or PAA as theutermost layer. No growth reduction is seen when the anionic PAA is used as theuter layer.

c) 9 layers of PVAm-C8–PAA. For all surfaces, the bacterial viability is high (indicatedrs used. The same tendency was observed for B. subtilis. (For interpretation of the

ysis, as the surfaces are slightly hydrophobic (Fig. 2). Another optionwould be to measure the effect in saline solutions, without any bac-terial growth. However, the growth state of the bacteria has beenshown to impact test results, so this possible confounding factormust be taken into account [4]. It is also interesting that the refer-ence fabrics had a higher effective surface area, especially the terry(Fabric2). In spite of this increased surface area, Fabric2 showedhigher bacterial growth than the negative control. This result isprobably due to the fact that the bacteria could use the fabric asan extra nutrient supply. The effectiveness of the silanol treatmenthas previously been questioned as well [27].

5. Conclusions

It has been shown that multilayers of polyvinylamine andpolyacrylic acid possess a bacterial growth-reducing effect. Weascribe this effect to the electrostatic interaction of PVAm, asmultilayers with an outermost PAA layer did not show any growth-inhibiting properties. The sampling time was an important factorin these results. As surfaces provide limited binding capacity,an underestimation of the growth-inhibiting effect will be seenwith longer sampling times. To get an unbiased view of theeffect, it is therefore desirable to keep the volume-to-surfaceratio low when testing these treated surfaces. Combining sur-face characterization data with growth inhibition data will allowus to understand the mechanisms behind the observed effects. Apreceding fundamental study on the multilayers has been pre-sented earlier [22], although further evaluation is needed. Insummary the multilayer technique with PVAm/PAA is a promis-ing alternative to covalently attaching polymers to render surfacesantibacterial.

Acknowledgements

BASF SE, SCA Hygiene Products AB and Vinnova are acknowl-edged for financing the project.

References

[1] A.D. Russell, Biocide use and antibiotic resistance: the relevance of labora-tory findings to clinical and environmental situations , The Lancet InfectiousDiseases 3 (12) (2003) 794–803.

[2] K. Lewis, A.M. Klibanov, Surpassing nature: rational design of sterile-surfacematerials , Trends in Biotechnology 23 (7) (2005) 343–348.

[3] A.E Madkour, G.N. Tew, Towards self-sterilizing medical devices: controllinginfection , Polymer International 57 (1) (2008) 6–10.

[4] R. Kügler, O. Bouloussa, F. Rondelez, Evidence of a charge-density threshold foroptimum efficiency of biocidal cationic surfaces , Microbiology 151 (5) (2005)1341–1348.

1 aces B

[

[

[

[

[

[

[

[

[

[

[

[

[

[

[

[

[

20 J. Illergård et al. / Colloids and Surf

[5] S.B. Lee, R.R. Koepsel, S.W. Morley, K. Matyjaszewski, Y. Sun, A.J. Russell, Perma-nent nonleaching antibacterial surfaces. 1. Synthesis by atom transfer radicalpolymerization , Biomacromolecules 5 (3) (2004) 877–882.

[6] J.C. Tiller, C.J. Liao, K. Lewis, A.M. Klibanov, Designing surfaces that kill bacteriaon contact , Proceedings of the National Academy of Sciences of the UnitedStates of America 98 (11) (2001) 5981–5985.

[7] J. Lin, S. Qiu, K. Lewis, A.M. Klibanov, Mechanism of bactericidal and fungici-dal activities of textiles covalently modified with alkylated polyethylenimine ,Biotechnology and Bioengineering 83 (2) (2003) 168–172.

[8] H. Murata, R.R. Koepsel, K. Matyjaszewski, A.J. Russell, Permanent, non-leachingantibacterial surfaces—2: how high density cationic surfaces kill bacterial cells, Biomaterials 28 (32) (2007) 4870–4879.

[9] A. Terada, A. Yuasa, T. Kushimoto, S. Tsuneda, A. Katakai, M. Tamada, Bacterialadhesion to and viability on positively charged polymer surfaces , Microbiology152 (12) (2006) 3575–3583.

10] J. Huang, R.R. Koepsel, H. Murata, W. Wu, S.B. Lee, T. Kowalewski, A.J. Russell,K. Matyjaszewski, Nonleaching antibacterial glass surfaces via “grafting onto”:the effect of the number of quaternary ammonium groups on biocidal activity, Langmuir 24 (13) (2008) 6785–6795.

11] J.Y. Maillard, Bacterial target sites for biocide action , Journal of Applied Micro-biology 92 (2002) 16S–27S.

12] A.M. Klibanov, Permanently microbicidal materials coatings , Journal of Mate-rials Chemistry 17 (2007) 2479–2482.

13] G. Decher, Fuzzy nanoassemblies: toward layered polymeric multicomposites, Science 277 (5330) (1997) 1232–1237.

14] F. Boulmedais, B. Frisch, O. Etienne, P. Lavalle, C. Picart, J. Ogier, J.C. Voegel, P.Schaaf, C. Egles, Polyelectrolyte multilayer films with pegylated polypeptidesas a new type of anti-microbial protection for biomaterials , Biomaterials 25(11) (2004) 2003–2011.

15] S.Y. Yang, J.-Y. Seo, Cellular interactions on nano-structured polyelectrolyte

multilayers , Colloids and Surfaces A: Physicochemical and Engineering Aspects313–314 (2008) 526–529.

16] P.M. Nguyen, N.S. Zacharia, E. Verploegen, P.T. Hammond, Extended releaseantibacterial layer-by-layer films incorporating linear-dendritic block copoly-mer micelles , Chemistry of Materials 19 (23) (2007) 5524–5530.

[

: Biointerfaces 88 (2011) 115– 120

17] A. Guyomard, E. Dé, T. Jouenne, J.-J. Malandain, G. Muller, K. Glinel, Incorpora-tion of a hydrophobic antibacterial peptide into amphiphilic polyelectrolytemultilayers: a bioinspired approach to prepare biocidal thin coatings ,Advanced Functional Materials 18 (5) (2008) 758–765.

18] M. Malcher, D. Volodkin, B. Heurtault, a. AndreÌ, P. Schaaf, P. Möhwald, H.Voegel, J.-C. Sokolowski, A. Ball, V. Boulmedais, F.B. Frisch, Embedded sil-ver ions-containing liposomes in polyelectrolyte multilayers: cargos films forantibacterial agents , Langmuir 24 (18) (2008) 10209–10215.

19] J.A. Lichter, M.F. Rubner, Polyelectrolyte multilayers with intrinsic antimicro-bial functionality: the importance of mobile polycations , Langmuir 25 (13)(2009) 7686–7694.

20] E.H. Westman, M. Ek, L. Wagberg, Antimicrobial activity of polyelectrolytemultilayer-treated cellulose films , Holzforschung 63 (1) (2009) 33–39.

21] S.Y. Wong, Q. Li, J. Veselinovic, B.-S. Kim, A.M. Klibanov, P.T. Hammond, Bacte-ricidal and virucidal ultrathin films assembled layer by layer from polycationicN-alkylated polyethylenimines and polyanions , Biomaterials 31 (14) (2010)4079–4087.

22] J. Illergård, L.E. Enarsson, L. Wågberg, M. Ek, Interactions of hydrophobicallymodified polyvinylamines: adsorption behavior at charged surfaces and theformation of polyelectrolyte multilayers with polyacrylic acid , ACS AppliedMaterials & Interfaces 2 (2) (2010) 425–433.

23] L. Bromberg, T.A. Hatton, Poly(N-vinylguanidine): characterization, and cat-alytic and bactericidal properties , Polymer 48 (26) (2007) 7490–7498.

24] E.H. Westman, M. Ek, L.E. Enarsson, L. Wagberg, Assessment of antibacte-rial properties of polyvinylamine (PVAm) with different charge densities andhydrophobic modifications , Biomacromolecules 10 (6) (2009) 1478–1483.

25] S. Champ, O. Koch, M. Ek, E. Westman, L. Wagberg, R. Feuerhake, H. Haehnle,Biocidal Coatings , BASF SE, 2008.

26] B. Gottenbos, H.C. Van der Mei, H.J. Busscher, D.W. Grijpma, J. Feijen, Initialadhesion and surface growth of Pseudomonas aeruginosa on negatively and pos-

itively charged poly(methacrylates) , Journal of Materials Science: Materials inMedicine 10 (12) (1999) 853–855.

27] P.R. Murray, A.C. Niles, R.L. Heeren, Microbial inhibition on hospital garmentstreated with Dow Corning 5700 antimicrobial agent , Journal of Clinical Micro-biology 26 (9) (1988) 1884–1886.