Effect of cationic polyelectrolytes in contact-active antibacterial layer-by-layer functionalization

(1)

This is the published version of a paper published in Holzforschung.

Citation for the original published paper (version of record):

Chen, C., Illergård, J., Wågberg, L., Ek, M. (2017)

Effect of cationic polyelectrolytes in contact-active antibacterial layer-by-layer functionalization

Holzforschung, 71(7-8): 649-658 https://doi.org/10.1515/hf-2016-0184

Access to the published version may require subscription.

N.B. When citing this work, cite the original published paper.

Permanent link to this version:

http://urn.kb.se/resolve?urn=urn:nbn:se:kth:diva-211388

(2)

Holzforschung 2017; 71(7-8): 649–658

Chao Chen, Josefin Illergård, Lars Wågberg and Monica Ek*

Effect of cationic polyelectrolytes in contact-active antibacterial layer-by-layer functionalization

DOI 10.1515/hf-2016-0184

Received October 14, 2016; accepted April 4, 2017; previously published online May 12, 2017

Abstract: Contact-active surfaces have been created by

means of the layer-by-layer (LbL) modification tech- nique, which is based on previous observations that cellulose fibers treated with polyelectrolyte multilay- ers with polyvinylamine (PVAm) are perfectly protected against bacteria. Several different cationic polyelec- trolytes were applied, including PVAm, two different poly(diallyl dimethyl ammonium chloride) polymers and two different poly(allylamine hydrochloride) poly- mers. The polyelectrolytes were self-organized in one or three layers on cellulosic fibers in combination with polyacrylic acid by the LbL method, and their antibac- terial activities were evaluated. The modified cellulose fibers showed remarkable bacterial removal activities and inhibited bacterial growth. It was shown that the interaction between bacteria and modified fibers is not merely a charge interaction because a certain degree of bacterial cell deformation was observed on the modified fiber surfaces. Charge properties of the modified fibers were determined based on polyelectrolyte titration and zeta potential measurements, and a correlation between high charge density and antibacterial efficiency was observed for the PVAm and PDADMAC samples. It was demonstrated that it is possible to achieve antibacte- rial effects by the surface modification of cellulosic fibers via the LbL technique with different cationic polyelectrolytes.

Keywords: antibacterial, cellulosic fiber, fiber modifica-

tion, layer by layer, nonleaching, polyelectrolyte

Introduction

Antibacterial agents are frequently incorporated into a base material, which is subsequently leached out during use. The antibacterial agents, e.g. silver in various forms, have been found to induce antibiotic resistance, but they are also toxic to higher organisms (Sütterlin et al. 2014).

Triclosan, another conventional low-molecular-weight antimicrobial agent, is also toxic to humans and causes an environmental burden after release (Fuchs and Tiller 2006), which loses antibacterial effect over time. To overcome these problems, contact-active antibacterial materials have been developed, in which the antibacte- rial substance is permanently attached to the material.

Different immobilization techniques were applied, for example, those via quaternary ammonium compounds (Chen et al. 2000; Bieser and Tiller 2011; Bieser et al. 2011), antibacterial peptides (Bagheri et al. 2009) and cationic polymers (Kugler et al. 2005; Terada et al. 2006; Illergård et al. 2011). The immobilization is usually achieved by chemically grafting a positively charged polymer onto the material surface (Advincula 2006; Hou et al. 2009), but this approach usually requires organic solvents and high temperatures as demonstrated by Roy et al. (2007), who showed that fibers grafted with poly(2-(dimethylamino) ethylmethacrylate) (PDMAEM) are antibacterial. In con- trast to these rather complicated grafting procedures, it was demonstrated that physical adsorption processes of water-based polyelectrolytes (Lichter et al. 2009b;

Illergård et al. 2011) are effective for the formation of polyelectrolyte monolayers on solid surfaces, which are antibacterial and contact active. Moreover, the antibacte- rial effect is increasing when adsorbing the polymers in polyelectrolyte multilayers created by the layer-by-layer (LbL) approach (Illergård et al. 2011), in which the adsorp- tion steps are repeated via oppositely charged polyelectro- lytes until the desired number of layers is built up. The LbL treatment of wood surfaces is relevant in terms of the aggregation behavior and adsorption characteristics of wood and wood components (Dammak et al. 2013; Qian

*Corresponding author: Monica Ek, Department of Fiber and Polymer Technology, School of Chemical Science and Engineering, KTH Royal Institute of Technology, Teknikringen 56-58, SE-100 44, Stockholm, Sweden, e-mail: monicaek@kth.se

Chao Chen and Josefin Illergård: Department of Fiber and Polymer Technology, School of Chemical Science and Engineering, KTH Royal Institute of Technology, Stockholm, Sweden

Lars Wågberg: Department of Fiber and Polymer Technology, School of Chemical Science and Engineering, KTH Royal Institute of Technology, Stockholm, Sweden; and Department of Fiber and Polymer Technology, Wallenberg Wood Science Centre, School of Chemical Science and Engineering, KTH Royal Institute of Technology, Stockholm, Sweden

(3)

et al. 2014; Rao et al. 2015; Li et al. 2016; Xiong et al. 2016).

The process is entropically driven because of the release of counterions to the polyelectrolytes and surfaces, and this makes the initially charged materials, such as wood- based cellulose fibers and silica (Notley et al. 2004;

Andreasson and Wågberg 2009), suitable substrates.

Various parameters affect the end result. Of course, the choice of polymer system is the most influential para- meter. Three polyelectrolytes are very common, namely, polyvinylamine (PVAm), poly(diallyl dimethyl ammo- nium chloride) (PDADMAC) and poly(allylamine hydro- chloride) (PAH) (Lu and Xiao 2007; Lu et al. 2008a,b). In a series of articles, Illergård et al. (2011, 2012, 2013, 2015) described the antibacterial effect of a system of PVAm and polyacrylic acid (PAA) on cellulose fibers and found a perfect growth inhibition. By visualization of Escheri-

chia coli by SEM, it was also observed that the bacteria

suffered severe damage (Illergård et al. 2015). Microbial reductions were seen after blending of PDADMAC with different wood-based substrates (Lu et al. 2008a,b; Melo et al. 2010). Lichter et al. (2009a) studied the antibacte- rial activities of PAH adsorbed onto a glass substrate and found that the structure and organization of LbL bases on sodium polystyrene sulfonate (SPS)/PAH and PAA/PAH system can be pH controlled and utilized for engineer- ing antibacterial functionality. These three cationic poly- electrolytes have different charge densities, and only the PDADMAC is a strong polyelectrolyte, for which the charge density is pH independent. The other polyelectrolytes are weak, and the charge of the PVAm, PAH and PAA are pH dependent. Hence, by comparing the LbL adsorption and antibacterial activities of these different polyelectrolytes, it should be possible to identify the type of the most effi- cient polymer and the conditions, under which the poly- electrolytes are adsorbed to the fiber surfaces. It is known that factors such as molecular weight, charge density and polymer structure of the polyelectrolytes affect the satura- tion adsorption at the air/solid interphase (Illergård et al.

2010).

The LbL treatment of solid materials is a conveni- ent procedure for introducing contact-active antibac- terial properties into materials (Lichter et al. 2009a;

Illergård et al. 2013). However, the mechanisms behind these complex interactions are still unclear. There are controversial discussions concerning the importance of hydrophobic interaction and/or that of the surface charge. Pioneering work on contact-killing surfaces was done by Tiller et al. (2001, 2002) and Siedenbiedel et al.

(2013), who emphasized the importance of a hydropho- bic interaction with the bacterial cell membrane. These authors also assumed that the surface-grafted biocidal

polymer is able to penetrate the bacterial cell wall and disrupt the phospholipid membrane. Bieser and Tiller (2011) and Bieser et al. (2011) explored this mechanism and proposed a “phospholipid sponge effect”, in which the contact-active coating without long polymeric spacers can disturb the bacterial cell membrane by adsorb- ing negatively charged phospholipids. Electrostatic forces may also be related to the high density of cationic charges on the surfaces, which are in relation to bacte- rial death. Lichter et al. (2009a,b) found, for example, an increased activating effects of the positive charge on weak polyelectrolyte PAH, which can be achieved by a final acidic rinsing before drying the modified material.

This is in line with studies concerning grafted polymeric surfaces (Kugler et al. 2005; Murata et al. 2007), where a charge threshold was found between 10

¹²

and 10

¹⁶

posi- tive charges per square centimeter. The specific threshold is, of course, bacterium dependent, and the status of the bacterial culture is also influential. Lichter et al. (2009a,b) and Thompson et al. (2005) suggested that the stiffness of the layers regulates the bacterial adhesion, i.e. the direct contact between the bacteria and the material. Illergård et al. (2015) observed that the bacteria were affected in several layers, even without direct physical contact with the surface, which is indicative for the importance of the charge. It is virtually impossible to make a direct com- parison of previous works on the antibacterial LbLs with different polymers because different substrates and dif- ferent polyelectrolytes were studied.

This is the reason why in the present study, three of the above-mentioned cationic polymers (PVAm, PDADMAC and PAH) should be adsorbed onto cellulosic fibers in one and three layer. PAA will serve as the middle layer in the three-layers LbL assembly to study the underlying anti- bacterial mechanisms of electrostatically adsorbed poly- electrolytes. The goal is to compare the effect of different polyelectrolyte combinations on cellulose substrates for certain bacteria in order to differentiate the influences of different fundamental factors of this technique.

Materials and methods

Fiber modification: Disintegrated bleached chemical softwood fibers (88% cellulose content) were studied (SCA Hygiene Prod- ucts, Mölndal, Sweden), which has been used in hygiene products for years in Sweden. The fibers have a total net negative charge of 43 μEq g⁻¹, as determined by conductometric titration (Katz and Beat- son 1984). The fibers were first washed into their hydrogen form and then into their sodium form prior to the treatments to remove possible metal ions, surfactants, dissolved polymers and colloids (Wåg- berg and Björklund 1993a).

(4)

C. Chen et al.: Cationic polyelectronics by the LbL method 651

Polymers: An overview of the polymers applied is given in Table 1.

PVAm was supplied by BASF SE (Ludwigshafen, Germany), and PDAD- MAC, PAH and PAA by Sigma-Aldrich (Stockholm, Sweden). The molecular weight data are from the suppliers. The polymers were carefully dialysed and freeze-dried to remove possible contaminations. The charge densities of the cationic polymers were measured by polyelectrolyte titration (PET) at the desired pH and salt concentration.

Layer-by-layer (LbL) modification: The anionic polymer concentra- tion was kept constant (PAA), whereas that of the cationic polymer was changed. The polymers were adsorbed for 15 min at a concentration of 0.1 g l⁻¹ with 0.1 M NaCl added and at a fiber consistency of 5 g l⁻¹ under constant stirring. A pH of 3.5 was used for PAA, whereas the PVAm was adsorbed at pH 9.5, where its charge is lower; PAH was adsorbed at pH 7.5; no pH adjustment was applied for the strongly charged PDADMAC, when modifying the fibers with one layer; and pH 3.5 was used for the PAA treatment in case of multilayer modification. After each adsorption step, the fibers were thoroughly rinsed by deionized water. After treatments, the fibers were rinsed with acidic water (pH 3.5) prior to freeze-drying (Illergård et al. 2015). An overview of the modified samples is given in Table 2.

Fiber characterization: Nitrogen analysis on an ANTEK Multi- Tek instrument (PAC, Houston, TX, USA) was applied to assess the amount of cationic polymers (the only N-containing compounds) adsorbed onto the fibers. By means of calibration curves, the amount of polymers could be calculated. All samples were tested in triplicate and the mean value together with its 95% confidence interval is presented.

Polyelectrolyte titration (PET): The amount of adsorbed polymer and the surface charge of the fibers were assessed by PET according to Wågberg et al. (1989). The polymers in Table 1 were added at different concentrations to a 5-g l⁻¹ fiber suspension in the presence of 0.1 M NaCl and allowed to adsorb for 15 min. Thereafter, the fibers were filtered and the amount of polymer remaining in the filtrate was determined by titration with anionic potassium polyvinyl sulfate

(KPVS) from Wako Pure Chemicals, Osaka, Japan. The amount of adsorbed polymer was then calculated as c_adsorbed = c_initial–c_filtrate and normalized to the amount of solid fibers used. For fibers modified with one layer of the cationic polyelectrolyte, the net charge was calculated as q_net = qadsorbed polymer − q_fiber. The charge is defined as that found at saturation adsorption. The surface charge of the unmodified fiber can be determined by adsorbing five different concentrations of PDADMAC on unmodified fibers, and the amount of PDADMAC remaining in the filtrate after filtration is determined by titration with KPVS. The charge of the unmodified fiber was calculated by extrapo- lation based on the PDADMAC charge at saturation concentration according to the adsorption isotherm (Wågberg et al. 1989). For fibers modified with three layers, the charge was determined by adsorbing anionic KPVS to saturation adsorption for both one and three layers.

Zeta potential measurements: A Mütek SZP-10 zeta potential ana- lyzer (BTG Instruments GmbH, Säffle, Sweden) was available for this purpose. The conductivities were measured by a conductometer equipped with a 013005D cell with a conductivity constant of 0.475 (Thermo Orion, Stockholm, Sweden) at pH 7.0. A fiber consistency during measurement was 1% and the salt concentration for the zeta potential test was 0.01 M NaCl.

Antibacterial assay: Gram-negative E. coli K-12 (BioRad, Solna, Swe- den) and Gram-positive Bacillus subtilis (Merck, Solna, Sweden) were grown overnight in nutrient broth (BD Difco, Stockholm, Sweden) at 37°C with continuous shaking. The bacterial cells were harvested by centrifugation at 5000 rpm for 5 min in a bench-top centrifuge (VWR, Stockholm, Sweden) at room temperature. The cells were washed with ¼ diluted Ringer’s solution (Merck, Stockholm, Sweden) followed by an additional centrifugation step. The bacteria were thereafter redispersed in ¼ Ringer’s solution.

Bacterial removal activity (BRA) tests: The fibers were suspended at 1% consistency in 10 ml of ¼ Ringer’s solution with the addition of 0.1 M pH 7.5 TRIS buffer (Sigma-Aldrich, Stockholm, Sweden). As a control, a sample without added fibers was also tested in parallel Table 1: Overview of the polymers studied.

Parameters

Cationic Anionic

PVAm PDADMAC1 (PD1) PDADMAC2 (PD2) PAH1 PAH2 PAA

M_w (kDa) 340 250–300 400–500 17.5 56 250

Charge (mEq g⁻¹) 12.7 6.16 5.43 13.2 3.53 n.d.

The molecular weights are from the suppliers. Charges of the polymers were measured by polyelectrolyte titration (PET). n.d., not determined.

Table 2: A summary of the different fiber samples and the respective polymers in the various layers.

Fiber modified by

PVAm 1L PVAm 3L PD1 1L PD1 3L PD2 1L PD2 3L PAH1 1L PAH2 1L

First layer PVAm PVAm PD1 PD1 PD2 PD2 PAH1 PAH2

Second layer – PAA – PAA – PAA – –

Third layer – PVAm – PD1 – PD2 – –

1L, single layer; 3L, three layers.

(5)

(“ + ” control). A 10⁷ colony-forming unit (CFU) ml⁻¹ of E. coli was added to each sample, and thereafter, the samples were incubated for 18 h with continuous shaking at 90 rpm. Thereafter, 1.5 ml of each suspension was centrifuged at 100 rpm for 1 min to remove the fibers, and 1 ml of the supernatant was cultivated on Petrifilm™ (3 M, Sol- lentuna, Sweden) before or after dilution, depending on the cationic polymer, as their different bacterial removal capabilities had been assessed in trial tests. The Petrifilms were kept in an incubator at 37°C for 24 h, and the number of colonies was counted either manu- ally or by a particle analyzer by ImageJ (Murata et al. 2007).

Bacterial growth inhibition (GI) test: After the BRA tests, 10% nutri- ent broth was added to each fiber-bacteria suspension to monitor bacterial growth. The suspensions were incubated for 18 h with continuous shaking at 90 rpm at 37°C. The bacterial growth was moni- tored by reading the optical density (OD) at λ = 620 nm on a MultiSkan FC microplate spectrophotometer (Thermo Scientific, Stockholm, Sweden). When the bacterial concentration was below the sensitiv- ity of the instrument, the bacterial concentration was determined by cultivation on Petrifilm as described above.

Fiber-bacteria adsorption capacities: To find the capacity limit of the different fibers to adsorb bacteria, the procedure was repeated with different starting concentrations of E. coli ranging from 5 × 10⁷ to 3 × 10⁹ CFU ml⁻¹. To increase the bacterial load per gram of fiber, the consistency of the fibers was reduced from 1% to 0.5% in 10 ml of Ringer’s solution with 0.1 M and pH 7.5 TRIS buffer added. To under- stand better the interaction between different modified fibers and bacteria, the inhibition test data were compared with those having the same bacterial load as applied for the inhibition tests.

Scanning electron microscopy (SEM): Modified fibers that had been exposed to 3.5 × 10⁹ CFU ml⁻¹ of E. coli in the bacterial capacity tests were visualized by SEM. To fix the bacteria, the fibers were pelleted via centrifugation (2000 rpm, 3 min) and thereafter fixed by rinsing in 2.5% glutaraldehyde in 0.1 M phosphate buffer (pH 7.4) overnight at 4°C. The specimens were rinsed in deionized water and solvent exchanged in 30%, 50%, 70%, 80% and 96% ethanol for 15 min, in absolute ethanol for 15 min at 4°C and finally in acetone (Mustafa et al. 1998) followed by drying at 4°C overnight. Prior to SEM analysis, a 5-nm layer of platinum was coated on all specimens by sputter- ing, which were then analyzed in a TM1000 tabletop SEM (Hitachi, Tokyo, Japan).

Results and discussion

The fiber surface modifications focused on single layers and three layers with PAA in the middle layer. The PVAm/

PAA LbLs were assembled according to Illergård et al.

(2015) with the pH sequences of 9.5/3.5/9.5 for the dif- ferent layers. The formation of PAA and PDADMAC LbLs also depends on the pH control for the weakly anionic PAA; a recent finding for PAA/PDADMAC shows that LbL growth occurs at pH 3, but that the multilayer growth is inhibited at pH > 6 because of an instant formation of

Table 3: Amounts of cationic polyelectrolytes in the outer layer adsorbed onto bleached softwood fibers.

Polyelectrolytes

Adsorbed amount (mg g⁻¹ fiber)

n anal.^a Titr.^b

PVAm 1L 4.7 3.7

PVAm 3L 15 8.9

PDADMAC 1 1L 3.2 2.5

PDADMAC 1 3L – 0.9

PAH 1 1L 8.1 5.1

PAH 2 1L 3.7 3.4

1L, single layer; 3L: three layers. ^aDetermined by ANTEK nitrogen analysis (n anal.); ^bDetermined by polyelectrolyte titration (Titr.).

intermolecular complexes (Alonso et al. 2013). For this reason, the layer assembly of PAA and PDADMAC was per- formed at pH 3. Because the fiber charge at this pH is very low, the first layer assembly of PDADMAC on the fiber was applied at neutral pH. The adsorption of PAH in LbLs on fiber surfaces was studied by Eriksson et al. (2005), who found that LbL formation was successful at pH 7.5. For all the adsorption studies and layer buildup, the salt concen- tration was kept constant at 0.1 M NaCl.

Polymer amount adsorbed on fiber

Table 3 summarizes the amounts of cationic polyelectro-

lytes adsorbed in the external layers determined both by

nitrogen analysis and from adsorption isotherms. The

former gives the total amount of polymers in the three

layers, and to obtain the amount in the third layer, the

amount in the first layer must be subtracted from the

total adsorbed amount. The PET approach leads directly

to the adsorbed amount in the outermost layer. The

results in Table 3 show that the PET results are gener-

ally lower than those determined by nitrogen analysis

but the trends are the same. A relatively large differ-

ence is in the amount on PVAm 3L, possibly because of

the principally high error, when the difference between

two large values has to be calculated. PAH1, with the

lowest molecular weight of the polyelectrolytes, has a

much higher adsorption on the fiber, i.e. polymers with

low molecular weight are easy to be absorbed. This was

demonstrated by Wågberg and Hägglund (2001), who

showed that low molecular PAH penetrates the fiber wall

more easily than high molecular PAH. The conforma-

tion of the polyelectrolytes at the solid–liquid interface

may additionally affect the adsorption. The difference

(6)

between PVAm and PDADMAC in three layers is possibly due to the recharging of the PDADMAC being inferior to that of the PVAm.

Surface charge determination

Table 4 summarizes the charge densities and zeta potentials of the modified fibers. For the surface charge density calculations, two units were used, i.e. μEq g

⁻¹

to normalize the adsorbed anionic polymer charge with respect to the amount of fibers, and N

⁺

cm

⁻²

for com- paring the charge threshold suggested by Kugler et al.

(2005) for optimal biocidal efficiency on the surface. The calculation is based on a fiber surface area of 1 m

²

g

⁻¹

. As the substrate remains the same, the surface charge depends mainly on the amount of polymer adsorbed and the polymer charge. Both the PET and the zeta poten- tial measurements in Table 4 show that the charging of the fibers is much more extensive for the PVAm than for the other polyelectrolyte in both, in case of one and three layers. The data difference concerning the PVAm and the PDADMAC in the PET indicates that the recharg- ing ability of these polymers differs almost in an order of magnitude. This very significant observation can be interpreted in different ways. It is well known that highly charged, stiff polyelectrolytes, such as PDADMAC, are adsorbed in a flat conformation on cellulose surfaces (Wågberg et al. 1988), with few loops and tails to bind to polyelectrolytes and particles in the second layer.

However, it has been demonstrated that, at higher ionic strengths, the LbL layers formed will be thicker, but the ionic strength of 0.1 M NaCl is still low (Dubas and

Table 4: Surface charge densities of modified fibers determined by PET in both Eq. g⁻¹ and number of charges per m² (with the assump- tion that 1 g fiber has 1 m² external surface area) and the zeta potential of the modified fibers.

Samples

Charge density^a

Zeta pot. (mV)^b (μEq g⁻¹) (10¹⁴ N⁺ cm⁻¹)

PVAm 1L 41.8 25.2 135

PVAm 3L 97.9 58.8 129

PDADMAC1 1L 11.3 6.80 68

PDADMAC1 3L 6.41 3.86 −9.5

PDADMAC2 1L 6.79 4.09 59

PDADMAC2 3L 2.72 1.64 −12.3

PAH1 1L 27.5 16.6 109

PAH2 1L 16.8 10.1 96

aFiber modified under 0.1 M NaCl. ^bThe zeta potential was determined at 0.01 M NaCl and 7.0 pH.

Schlenoff 2001). Another feature of weak polyelectro- lytes is that they increase their charge because another polyelectrolyte/particle of opposite charge is exposed to the formed layer (Kharlampieva et al. 2003). It can be assumed that this charge increment is a result of the combination of the flat conformation of the PDADMAC and its degree of protonation at a given pH, when an oppositely charged polyelectrolyte/particle approaches the surface. The PAH shows a similar trend supporting this hypothesis, but these polyelectrolytes are of lower molecular mass and thus have a shorter extension from the solid-liquid interface. The difference between PVAm/

PAH and PDADMAC is striking, and further research is needed to interpret this observation.

Bacterial removal activity (BRA)

Materials with a removal efficiency of ≈100% are consid- ered as antibacterial. The initial bacterial loads in the tests were 10 times higher, 10

⁷

CFU ml

⁻¹

, than in previous studies to facilitate a better comparison. The fibers modified with different combinations of PVAm and PDADMAC show a greater removal of Gram-negative E. coli than of Gram- positive B. subtilis, whereas PAH-modified samples had the opposite effect (Figure 1a). Of all the tested samples, 3L PVAm shows the best results and has almost identical effects on both E. coli and B. subtilis. The number of layers does not play an essential role concerning the bacterial removal efficiency.

Bacterial growth inhibition (GI)

In this test, 1 ml of nutrient broth was added to each sample directly, after removal of the excess bacteria to provide the remaining bacteria with good conditions to grow. The growth inhibition (GI) is presented in Figure 1b, where GI is calculated as follows:

Bacteria in solution

GI(%) 1 100

Bacteria no. of “ ”control

 

= −   +   ×

The GI% can be compared with the BRA presented

in Figure 1a. PD1/2 3L samples with a lower adsorbed

amount of PDADMAC and a lower surface charge had a

surprisingly good effect on E. coli regarding both BRA and

GI. The 3L PD1 had a higher charge and also showed a

better attraction to B. subtilis with over 99.6% BRA com-

pared to less than 90% BRA for PD2. However, neither of

these samples inhibited B. subtilis growth. This indicates

(7)

that charge density influences on bacterial adhesion, but that the impact on bacterial growth depends on the bacteria species. This view is supported by the results for the PAH samples, where both PAH1 and PAH2 work better for B. subtilis than for E. coli. 1L samples including PVAm and two PD samples showed very similar antibac- terial activity for both BRA and GI. These treatments were able to remove ≈100% of E. coli and B. subtilis from an aqueous solution and inhibited the growth to 96% of E.

coli and to 60% of B. subtilis. However, PVAm 1L had a

surface charge level four times higher values than that of PD1 1L and nearly seven times that of PD2 1L, whereas the adsorbed amounts showed no great difference. This indicates that the role of surface charge is crucial but not the only important factor. The chemical structure of the polymer is another decisive parameter, including the length of the alkyl chain length, hydrophobic backbone and hydrophilic side groups. Thus, the highest amount of adsorbed PVAm in the PVAm 3L sample embedded with the highest surface net charge did not give the best results in the case of neither E. coli nor B. subtilis. Further model investigations are needed to separate the most important factors for both BRA and GI.

Figure 2: Supposed adsorption packing of E. coli (1 × 2 μm²) on a fluff pulp fiber surface for estimating the highest possible adsorp- tive capacity on the fiber surface.

0

Growth inhibition (%)

Removal percentage

1L PV Am

3L PV Am

1L PD1 3L PD1 1L PD2 3L PD2 1L PAH1 1L P

AH2 20

40 60 80 100 120

E. coli E. coli

B. subtilis B. subtilis

99.4%

99.6%

99.7%

99.9%

100.0%

a

b

Figure 1: Bacterial inhibition properties of fibers treated with differ- ent polyelectrolyte combinations in terms of E. coli and B. subtilis.

(a) Bacterial removal capacity (BRC 5). (b) Percentage growth inhibition (GI%). The error bars show 95% confidence intervals.

Bacteria-fiber adsorption (BFA) capacities

It was observed by Illergård et al. (2015) that the degree of BFA depends on the initial bacterial load. Therefore, the loading concentration of E. coli was increased to learn more about this effect. To determine the relevant bacterial concentrations, a theoretical maximum bacterial adsorp- tion limit was estimated by a simple surface saturation model assuming an external surface area of the pulp of 1 m

²

g

⁻¹

(Chang et al. 1981) and that the bacterial cells can be fully packed and evenly distributed on the fiber surface (Figure 2). These assumptions led to a theoretical saturation density of ca. 5 × 10

¹¹

bacterial cells per gram fiber, compared with the removal of only 10

⁹

CFU g

⁻¹

of 3L PVAm-modified fiber, as reported earlier. The bacte- rial removal tests were therefore extended with 20-fold, 40-fold and 160-fold of the original bacterial loads to test the saturation limits of the fibers.

With a concept borrowed from the adsorption iso-

therms of polymers, the amount of E. coli adsorbed on the

modified fibers is presented in Figure 3, where the concen-

tration of bacteria left in the bacterial solution after contact

with modified fibers is compared with the equilibrium

(8)

concentration of the bacteria. For the one-layer samples (Figure 3a), it can be observed that PDADMAC-modified fibers have a better BFA capacity than fibers treated with PVAm, although fibers with one layer of PVAm show a similar or even better BFA with a 10

⁷

CFU ml

⁻¹

bacterial solution (Figure 1). It is evident from Figure 3a that the PDADMAC-treated fibers have a higher BFA capacity than the other investigated polyelectrolytes and also that the two PDADMAC polymers (PD1 and PD2) give basically the same results, which is not unexpected considering the similarity of their molecular weights. For the fibers treated with three layers, the situation is quite different. Here, PVAm is superior to the PDADMAC. The 3L PVAm has the

Figure 3: Adsorption isotherms of E. coli on modified fibers.

(a) One-layer modified fiber. (b) Three-layer modified fiber with PAA as the intermediate second layer. The vertical axis is the adsorbed number of E. coli cells per gram fiber. The horizontal axis is the equilibrium concentration after 18 h of contact.

Table 5: Optical density (OD) values at 620 nm in test 1 (Ringer’s solution together with nutrient) and test 2 (only Ringer’s solution) after 18 h incubation.

+ control PVAm 1L PVAm 3L PD1 1L PD1 3L PD2 1L PD2 3L PAH1 1L PAH2 1L

Test 1 (OD) 0.120 0.038 0.037 0.038 0.039 0.037 0.040 0.044 0.040

Test 2 (OD) 0.120 0.075 0.036 0.036 0.084 0.038 0.085 0.062 0.071

The second intermediate layer is PAA in samples with 3L.

largest BFA capacity and the highest adsorption, 5.5 × 10

¹¹

CFU g

⁻¹

, and the isotherm indicates that the surface satu- ration has not been reached for this treatment.

In the standard protocol of the bacterial GI test, only the number of bacteria left in the solution is assessed, which does not permit in-depth interpretation. As visible in Figure 3, the modified fibers have a great potential for BFA even at a very high bacterial load. A test was there- fore designed for the situation, where the same amount of functional fiber was added to a batch of bacteria with a lower bacterial load together with nutrients. This test was based on the OD of the “ + ” control for this batch of bacte- ria, 0.12, which is approximately 4 × 10

⁸

CFU ml

⁻¹

bacteria after 18 h of growth. The initial load of the other bacteria batch was 4 × 10

⁸

CFU ml

⁻¹

and included Ringer’s solution instead of nutrients. The OD values at 620 nm are sum- marized in Table 5 after 18 h of incubation together with the modified fibers. For the samples of PVAm 3L, PD1/2 1L, the OD values showed no great difference between Test 1 and Test 2, probably because of their high adsorption capacity (Figure 3). For the samples of PVAm 1L, PD1/2 3L and PAH samples, however, batches with nutrients (Test 1) had much lower OD values than batches without (Test 2) As expected for a contact-active antibacterial surface, this indicates that the bacterial cell may be adsorbed on the surface at the beginning, but it stops for some reason multiplying on the functioned fiber surface even in the presence of nutrients. Otherwise, if we assume that the interaction between the bacteria and the fiber surface was only electrostatic in nature, these should not lead to different OD values between batches with or without nutrients after 18 h. This hypothesis was tested by SEM observations.

Scanning electron microscopy (SEM)

As seen in Figure 4, a larger number of bacteria were attached to the modified fibers (3, 4) than to the untreated reference (2). The PVAm 3L-modified fiber surface (4) had more attached E. coli than the PD2 3L-modified fiber (3).

Images (3) and (4) also illustrate that the modified fiber

(9)

surfaces can affect the morphology of the bacteria. By image analysis, it was possible to quantify that the healthy

E. coli cells (1) are 1.41 μm in length with a standard devia-

tion (StD) of 0.14 μm, whereas the E. coli cells on the modi- fied fiber surfaces show an elongation up to 4.17 μm with a StD of 0.63 μm. Moreover, the membranes of the E. coli bacterial cells appear damaged as if they had been plas- tically deformed under strain in the length direction, i.e.

indicating a physiological deformation. This explains the results of the bacterial GI on the treated fiber samples.

Accordingly, the modified fiber surfaces have an ability to disrupt the integrity of the bacterial cell wall. Illergård et al. (2015) and Hoque et al. (2015) speculated that the strong interaction on the fiber surface may prevent bacte- rial movements so that the bacteria are unable to respond to surrounding changes leading to their starvation.

According to Dickson and Koohmaraie (1989), the surface charge on both the bacteria and the substrate has a major impact on bacterial attachment. Kugler et al.

(2005) enforced this view and stated that there is a charge threshold of a surface to be contact-active antibacte- rial and coined the notion “outer-layer charge density”

(OLCD), which is 10

¹²

N

⁺

/cm

²

to kill E. coli in the dividing state and 10

¹⁴

N

⁺

/cm

²

in the quiescent state. The OLCD is even more important than the hydrophobic interactions.

As visible in Table 4, the OLCD requirement is satisfied

and explains the inhibitory properties of modified fibers towards E. coli. However, the growth inhibition of Gram- positive B. subtilis is relatively low compared with that of

E. coli, possibly because of its thicker cell wall.

Conclusions

The LbL approach for fiber modification proved to be effective, and the importance of the charge density for antibacterial properties could be confirmed. Both PET and zeta potential measurements revealed a strong positive correlation between the charge density and the antibacte- rial activities concerning the bacterial removal and bacte- rial growth inhibition. Among the samples, the three-layer PVAm-modified fiber containing PAA shows the best anti- bacterial activities and the recharging ability explains the relatively large amounts of adsorbed bacteria and the high positive charge surface. The SEM observations show bacterial cell deformation and elongation. Future investi- gations should focus on how different cationic surfaces influence the antibacterial properties of fibers modified by the LbL technique.

Acknowledgments: We thank the Chinese Scholarship

Council for financial support and RISE Bioeconomy for

Figure 4: SEM images of (1) a healthy E. coli colony, (2) a reference fiber sample after contact with E. coli, (3) E. coli adsorption on PD2 3L-modified fibers and (4) E. coli adsorption on PVAm 3L-modified fibers.

(10)

technical support with the nitrogen analysis. Lars Wåg- berg acknowledges the Wallenberg Wood Science Centre at KTH Royal Institute of Technology for financial support.

References

Advincula, R. (2006) Polymer brushes by anionic and cationic surface-initiated polymerization (SIP). In: Surface-Initiated Polymerization I. Ed. R. Jordan. Springer Berlin Heidelberg.

Alonso, T., Irigoyen, J., Iturri, J., Moya, S. (2013) Study of the multilayer assembly and complex formation of poly (diallyldimethyl- ammonium chloride)(PDADMAC) and poly (acrylic acid)(PAA) as a function of pH. Soft Matter 9:1920–1928.

Andreasson, B., Wågberg, L. (2009) On the mechanisms behind the action of wet strength and wet strength agents. In: Pulp and Paper Chemistry and Technology: Paper Products Physics and Technology. Eds. Ek, M., Gellerstedt, G., Henriksson, G.

de Gruyter, Berlin, pp. 185–208.

Bagheri, M., Michael, B., Margitta, D. (2009) Immobilization reduces the activity of surface-bound cationic antimicrobial peptides with no influence upon the activity spectrum. Antimicrob.

Agents Chemother. 53:1132–1141.

Bieser, A.M., Tiller, J.C. (2011) Mechanistic considerations on contact‐active antimicrobial surfaces with controlled functional group densities. Macromol. Biosci. 11:526–534.

Bieser, A.M., Yi, T., Tiller, J.C. (2011) Contact‐active antimicrobial and potentially self‐polishing coatings based on cellulose. Macro- mol. Biosci. 11:111–121.

Chang, M.M., Terry, Y.C., George, T.T. (1981) Structure, pretreat- ment and hydrolysis of cellulose. Bioenergy. Springer, Berlin, Heidelberg, pp. 15–42.

Chen, C.Z., Beck-Tan, N.C.D., Prasad, D., Tina, K., LaRossa, R.A.C., Stuart L. (2000) Quaternary ammonium functionalized poly(propylene imine) dendrimers as effective antimicrobials:

structure–activity studies. Biomacromolecules 1:473–480.

Dammak, A., Moreau, C., Beury, N., Schwikal, K., Winter, H.T., Bonnin, E., Saake, B., Cathala, B. (2013) Elaboration of multilayered thin films based on cellulose nanocrystals and cationic xylans: application to xylanase activity detection.

Holzforschung 67:579–586.

Dickson, J., Koohmaraie, M. (1989) Cell surface charge characteristics and their relationship to bacterial attachment to meat surfaces. Appl. Environ. Microbiology 55:832–836.

Dubas, S.T., Schlenoff, J.B. (2001) Swelling and smoothing of polyelectrolyte multilayers by salt. Langmuir 17:7725–7727.

Eriksson, M., Notley, S.M., Wågberg, L. (2005) The influence on paper strength properties when building multilayers of weak polyelectrolytes onto wood fibres. J. Colloid. Interface Sci. 292:38–45.

Fuchs, A.D., Tiller, J.C. (2006) Contact-active antimicrobial coatings derived from aqueous suspensions. Angew. Chem. Int. Ed.

Engl. 45:759–762.

Hoque, J., Akkapeddi, P., Yadav, V., Manjunath, G.B., Uppu, D.S., Konai, M.M., Yarlagadda, V., Sanyal, K., Haldar, J. (2015) Broad spectrum antibacterial and antifungal polymeric paint materials: synthesis, structure-activity relationship, and membrane-active mode of action. ACS Appl. Mater. Interfaces 7:1804–1815.

Hou, A., Zhou, M., Wang, X. 2009. Preparation and characterization of durable antibacterial cellulose biomaterials modified with triazine derivatives. Carbohydrate Polym. 75:328–332.

Illergård, J., Enarsson, L.E., Wågberg, L., Ek, M. (2010) Interactions of hydrophobically modified polyvinylamines: adsorption behavior at charged surfaces and the formation of polyelectrolyte multilayers with polyacrylic acid. ACS Appl. Mater.

Interfaces 2:425–433.

Illergård, J., Wågberg, L., Ek, M. (2011) Bacterial-growth inhibiting properties of multilayers formed with modified polyvinylamine.

Colloids Surf. B Biointerfaces 88:115–120.

Illergård, J., Römling, U., Wågberg, L., Ek, M. (2012) Biointeractive antibacterial fibres using polyelectrolyte multilayer modification. Cellulose 19:1731–1741.

Illergård, J., Römling, U., Wågberg, L., Ek, M. (2013) Tailoring the effect of antibacterial polyelectrolyte multilayers by choice of cellulosic fiber substrate. Holzforschung 67:573–578.

Illergård, J., Wågberg, L., Ek, M. (2015) Contact-active antibacterial multilayers on fibres: a step towards understanding the antibacterial mechanism by increasing the fibre charge. Cellulose 22:2023–2034.

Katz, S., Beatson, R.P. (1984) The determination of strong and weak acidic groups in sulfite pulps. Svensk Papperstid. 87:48–53.

Kharlampieva, E., Sukhishvili, S.A. (2003) Ionization and pH stabil- ity of multilayers formed by self-assembly of weak polyelectrolytes. Langmuir 19:1235–1243.

Kugler, R., Bouloussa, O., Rondelez, F. (2005) Evidence of a charge- density threshold for optimum efficiency of biocidal cationic surfaces. Microbiology 151:1341–1348.

Li, Y., Yang, D., Huang, Q., Li, R. (2016) Modified sodium lignosul- fonates (NaLS) with straight chain alcohols and their aggregation behavior and adsorption characteristics on solid surfaces.

Lichter, J.A., Rubner, M.F. (2009a) Polyelectrolyte multilayers with intrinsic antimicrobial functionality: the importance of mobile polycations. Langmuir 25:7686–7694.

Lichter, J.A., Van V., Krystyn, J., Rubner, M.F. (2009b) Design of antibacterial surfaces and interfaces: polyelectrolyte multilayers as a multifunctional platform. Macromolecules 42:8573–8586.

Lu, J., Xiao, C. (2007) Synthesis, characterization and antibacterial activity of poly (dimethyldiallyl ammonium) chloride. J. Wuhan Univ. Nat. Sci. Ed. 53:397.

Lu, Y., Sun, J., Shen, J. (2008a) Cell adhesion properties of patterned poly(acrylic acid)/poly(allylamine hydrochloride) multilayer films created by room-temperature imprinting technique. Lang- muir 24:8050–8055.

Lu, J., Wang, X., Xiao, C. (2008b) Preparation and characterization of konjac glucomannan/poly(diallydimethylammonium chloride) antibacterial blend films. Carbohydrate Polym. 73:427–437.

Melo, L.D., Mamizuka, E.M., Carmona-Ribeiro, A.M. (2010) Anti- microbial particles from cationic lipid and polyelectrolyte.

Langmuir 26:12300–12306.

Murata, H., Koepsel, R.R., Matyjaszewski, K., Russell, A.J. (2007).

Permanent, non-leaching antibacterial surface-2: how high density cationic surfaces kill bacterial cells. Biomaterials 28:4870–4879.

Mustafa, K., Blanca S.L., Kjeil H., Wennerberg, A., Arvidson, K.

(1998) Attachment and proliferation of human oral fibroblasts to titanium surfaces blasted with TiO2 particles. A scanning

(11)

electron microscopic and histomorphometric analysis. Clinical Oral Implants Res. 9:195–207.

Notley, S.M., Simon, C., Vincent, S.J., Wågberg, L. (2004) Adsorbed layer structure of a weak polyelectrolyte studied by colloidal probe microscopy and QCM-D as a function of pH and ionic strength. Physical Chem. Chemical Physics 6:2379–2286.

Qian, L., Dong, C., Liang, X., He, B., Xiao, H. (2014) Polyelectrolyte complex containing antimicrobial guanidine-based polymer and its adsorption on cellulose fibers. Holzforschung 68:103–111.

Rao, X., Liu, Y., Fu, Y., Liu, Y., Yu, H. (2015) Formation and properties of polyelectrolytes/TiO2 composite coating on wood surfaces through layer-by-layer assembly method. Holzforschung 70:361–367.

Roy, D., Knapp, J.S., Guthrie, J.T., Perrier, S. (2007) Antibacterial cellulose fiber via RAFT surface graft polymerization. Biomacro- molecules 9:91–99.

Siedenbiedel, F., Fuchs, A., Moll, T., Weide, M., Breves, R., Tiller, J.C.

(2013) Star‐shaped poly (styrene)‐block‐poly (4‐vinyl‐N‐meth- ylpyridiniumiodide) for semipermanent antimicrobial coatings.

Macromol. Biosci. 13:1447–1455.

Sütterlin, S.E., Petra, S., Linus, M., Thomas, M., Björn, Å.

(2014) Silver resistance genes are overrepresented among Escherichia coli isolates with CTX-M production. Appl. Environ.

Microbiology 80:6863–6869.

Terada, A., Yuasa, A., Kushimoto, T., Tsuneda, S., Katakai, A., Tamada, M. (2006) Bacterial adhesion to and viability on positively charged polymer surfaces. Microbiology 152:3575–3583.

Thompson, M.T.B., Michael, C.T., Irene, S.R., Michael, F., Van, V., Krystyn, J. (2005). Tuning compliance of nanoscale polyelectrolyte multilayers to modulate cell adhesion. Biomaterials 26:6836–6845.

Tiller, J.C., Liao, C., Lewis, K., Klibanov, A.M. (2001) Designing surfaces that kill bacteria on contact. Proc. Nat. Academy Sci.

98:5981–5985.

Tiller, J.C., Lee, S., Lewis, K., Klibanov, A.M. (2002) Polymer surfaces derivatized with poly (vinyl‐N‐hexylpyridinium) kill airborne and waterborne bacteria. Biotechn. Bioeng.

79:465–471.

Wågberg, L., Björklund, M. (1993a) Adsorption of cationic potato starch on cellulosic fibres. Nordic Pulp Paper Res. J 8:53–58.

Wågberg, L., Hägglund, R. (2001b) Kinetics of polyelectrolyte adsorption on cellulosic fibers. Langmuir 17:1096–1103.

Wågberg, L., Ödberg, L., Lindström, T., Aksberg R. (1988) Kinetics of adsorpton and ion-exchange reactions during adsorption of cationic polyelectrolytes onto cellulosic fibers. J. Colloid.

Interface Sci. 123:287–295.

Wågberg, L., Odberg, L., Glad, N.G. (1989) Charge determination of porous substrates by polyelectrolyte adsorption Part 1. Car- boxymethylated bleached cellulosic fibres. Nordic Pulp Paper Res. J. 4:71–76.

Xiong, W., Qiu, X., Zhong, R., Yang, D. (2016) Characterization of the adsorption properties of a phosphorylated kraft lignin-based polymer at the solid/liquid interface by the QCM-D approach.