Pf

XMH

a

ARRAA

KPCSAL

1

olaoaasedPta

0h

Colloids and Surfaces B: Biointerfaces 115 (2014) 340– 348

Contents lists available at ScienceDirect

Colloids and Surfaces B: Biointerfaces

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

olypropylene non-woven meshes with conformal glycosylated layeror lectin affinity adsorption: The effect of side chain length

iang-Yu Ye, Xiao-Jun Huang, Zhi-Kang Xu ∗

OE Key Laboratory of Macromolecular Synthesis and Functionalization, Department of Polymer Science and Engineering, Zhejiang University,angzhou 310027, China

r t i c l e i n f o

rticle history:eceived 13 September 2013eceived in revised form 6 December 2013ccepted 10 December 2013vailable online 21 December 2013

eywords:olypropylene non-woven meshonformal graftingurface glycosylationffinity membraneectin adsorption

a b s t r a c t

The unique characteristics of polypropylene non-woven meshes (PPNWMs), like random network ofoverlapped fibers, multiple connected pores and overall high porosity, make them high potentials for useas separation or adsorption media. Meanwhile, carbohydrates can specifically recognize certain lectinthrough multivalent interactions. Therefore glycosylated PPNWMs, combing the merits of both, can beregarded as superior affinity membranes for lectin adsorption and purification. Here, we describe a versa-tile strategy for the glycosylation of PPNWMs. Two hydrophilic polymers with different side chain length,poly(2-hydroxyethyl methacrylate) (PHEMA) and poly(oligo(ethylene glycol) methacrylate) (POEGMA),were first conformally tethered on the polypropylene fiber surface by a modified plasma pretreatmentand benzophenone (BP) entrapment UV irradiation process. Then glucose ligands were bound throughthe reaction between the hydroxyl group and acetyl glucose. Chemical changes of the PPNWMs surfacewere monitored by FT-IR/ATR. SEM pictures show that conformal glucose ligands can be achieved through

the modified process. After deprotection, the glycosylated PPNWMs became superhydrophilic and hadhigh specific recognition capability toward Concanavalin A (Con A). Static Con A adsorption experimentswere further performed and the results indicate that fast adsorption kinetics and high binding capacitycan be accomplished at the same time. We also found that increasing the side chain length of polymerbrushes had positive effect on protein binding capacity due to improved chain mobility. Model studiessuggest a multilayer adsorption behavior of Con A.

© 2013 Elsevier B.V. All rights reserved.

. Introduction

Polypropylene non-woven meshes (PPNWMs) are one kindf the most popular membrane materials due to their excel-ent mechanical properties, high thermal and chemical stabilities,nd comparatively low-cost. In addition, the random network ofverlapped fibers offers many unique characteristics, including rel-tively large specific surface area, engineered interconnected poresnd overall high porosity. Computational fluid dynamics (CFD)imulation results reveal that such disordered fibrous structurendows non-woven meshes with high permeability, low pressurerop and reduced mass transfer/diffusion resistance [1]. Therefore,PNWMs exhibit promising prospects for comprehensive applica-

ions in air/liquid filtration or adsorption media, biomedical textilend protective cloth [2–6].

∗ Corresponding author. Tel.: +86 571 8795 2605; fax: +86 571 8795 1773.E-mail address: xuzk@zju.edu.cn (Z.-K. Xu).

927-7765/$ – see front matter © 2013 Elsevier B.V. All rights reserved.ttp://dx.doi.org/10.1016/j.colsurfb.2013.12.025

Among these promising applications of PPNWMs, membranechromatography for protein bioseparation has captured growingattention in recent years [3,7–10]. Industrial protein purificationprocess is conventionally carried out using packed bed or columnchromatography, which has several major limitations such as slowintraparticle diffusion, large pressure drop, high cost and hence lowprotein binding capacity/productivity and difficulty in scale-up.The unique structural characteristics of PPNWMs make them themost promising alternative materials to replace traditional packed-bed resins [11,12]. For example, PPNWMs grafted with conformalanion exchange ligands and hydrophilic spacers were preparedthrough a UV pretreatment-UV grafting process, and it was foundthat such activated meshes had high bovine serum albumin bind-ing capacity and permeability coefficient [9]. Even when stacking150 layers in a column, the pressure drop was still acceptable andit helped to create sharp chromatographic elution profiles [13]. Itwas also reported that the presence of a molecular spacer between

the ligand and the membrane matrix could be beneficial forprotein adsorption [14]. Further attempts to increase the bindingcapacity led to the creation of PPNWMs supported/macroporousgel based 3D binding domain composite membrane [10]. However,

s B: B

tiitt

bs[imkicpfallrn

caFmuhfipoTqma

2

2

tm58ad(MtiefiptwTA(te1aiL

X.-Y. Ye et al. / Colloids and Surface

he separation mechanism was mainly based on non-specificnteractions like electrostatic attraction/repulsion, hydrophobicnteraction and Vander Waals forces. It is a logical expectation thathe introduction of affinity ligands with specific interaction witharget proteins can result in high separation efficiency and purity.

Carbohydrates are widely regarded as energy sources anduilding elements. With the development of glycobiology, theirtructurally and functionally diverse roles are gradually discovered15]. Carbohydrate–protein recognition mediates various biolog-cal processes and is the first step in cell–cell interactions via

ultivalent interactions, or glycocluster effect [16]. Lectin, as aind of carbohydrate-binding proteins or glycoproteins, can specif-cally recognize certain carbohydrate moieties [17]. Meanwhile,arbohydrates are highly hydrophilic molecules that non-specificrotein adsorptions can be greatly reduced or eliminated. There-ore, glycosylated PPNWMs are reasonably expected as superiorffinity membranes for lectin adsorption and purification. In theiterature, a few studies have reported the fabrication of glycosy-ated poly(ethylene terephthalate) fibers [18–20]. However, it isarely addressed on the bioseparation performance of glycosylatedon-woven meshes.

In this work, our goal is to develop a facile method for theonstruction of glycosylated PPNWMs and further evaluate theirpplication prospects as affinity membranes for lectin purification.irst, a modified plasma pretreatment and benzophenone entrap-ent UV irradiation process was employed to graft conformal and

niform poly(oligo(ethylene glycol) methacrylate) and poly(2-ydroxyethyl methacrylate) (POEGMA and PHEMA) brushes on theber surfaces of PPNWMs. Then surface glycosylation was accom-lished by using different acetyl glycosyl donors in the presencef boron trifluoride diethyl etherate and subsequent deprotection.he surface wettability, qualitative fluorescein-labeled lectin anduantitative static Con A adsorption were characterized. Further-ore, the effect of side chain length on adsorption capacity was

lso evaluated.

. Experimental

.1. Materials

Commercial PPNWMs (Jiangyin Golden Phoenix Special Tex-ile Co., Ltd, China) used in this work were produced with a

elt-blown process. The fiber diameter was in the range of–10 �m. The density and porosity were about 35 g/m2 and0%, respectively. The samples were cut into rotundity with

diameter of 2.5 cm (S = 4.91 cm2), washed with acetone andried in a vacuum oven at 40 ◦C. 2-Hydroxyethyl methacrylateHEMA, 97%) and oligo(ethylene glycol) methacrylate (OEGMA,

n = 360, 99%) were supplied by Sigma–Aldrich and passedhrough neutral Al2O3 flash column chromatography to removenhibitors. Benzophenone (BP) was recrystallized from coldthanol. Boron trifluoride diethyl etherate (BF3·Et2O) was puri-ed by vacuum distillation. Dichloromethane was distilled fromhosphorus pentoxide immediately before use. Trichloroacetoni-rile, 2,3,4,6-tetraacetyl-�-d-glucose and �-d-glucose pentaacetateere bought from Beijing Chemsynlab Pharmaceutical Science &

echnology Co. Ltd. and used as received. Concanavalin A (Con), fluorescein-labeled Con A (FL-Con A) and peanut agglutinin

FL-PNA) (Vector, USA) were purchased and used directly. Allhe other chemicals like sodium methoxide, heptane, petroleumther (60–70 ◦C), potassium carbonate, and 2-[4-(2-hydroxyethyl)-

-piperazinyl]ethanesulfonic acid (HEPES) were of analytical gradend used without further purification. Water used in all exper-ments was deionized and ultrafiltrated to 18 M using an ELGAabWater system (ELGA Classic UF, France).

iointerfaces 115 (2014) 340– 348 341

2.2. Graft polymerization of OEGMA and HEMA onto PPNWMs

The experimental procedure is schematically illustrated in Fig. 1.Dielectric barrier discharge plasma was used to pretreat the nascentPPNWM at atmospheric pressure by a plasma apparatus (NanjingSuman Electronics Co., Ltd., China). Two quartz glass plates witha gap of 2 mm served as the dielectric layer and argon (99%)/air(1%) was used as the discharge gas. Firstly, the sample was irra-diated at 35 V and 10 kHz for a given time and exposed in the airfor 10 min. After that, the pretreated PPNWMs were immersed in20 mM BP heptane solution for 45 min to immobilize the photoini-tiator in the surface layer of the polypropylene fiber and then driedin the air. Thereafter, the PPNWMs were presoaked with acetoneand immediately dipped into OEGMA or HEMA solutions in petridishes and fixed between two sheets of filter paper. Finally, UVgrafting polymerization on the surfaces of PPNWMs was carriedout under a homemade high pressure mercury lamp (232–400 nm,intensity 3 mW/cm2) for a predetermined time. The grafted PPN-WMs were washed thoroughly with ethanol overnight to removeunreacted monomer and homopolymer, and then dried in a vac-uum oven at 40 ◦C. They were weighed with an analytical balanceto a precision of 0.01 mg (XP105DR, Mettler Toledo, Switzerland).The grafting density (GD, �g/cm2) was calculated by the followingequation:

GD = W1 − W0

S(1)

where W0 and W1 are the mass of the nascent and thePOEGMA/PHEMA grafted PPNWMs, respectively. S represents thesurface area of each sample.

2.3. Coupling of glucose ligands to hydroxyl groups of thePOEGMA/PHEMA grafted PPNWMs

Surface glycosylation reaction was conducted as reported in ourprevious work [21]. Three pieces of POEGMA/PHEMA grafted PPN-WMs were fully immersed in 6 mL freshly dried dichloromethanesolution containing 20 equiv glucose pentaacetate and 100 equivBF3·Et2O. The reaction was sealed in a Schlenk tube and carried outat 0 ◦C for 2 h followed by 20 h at room temperature. After that,the samples were washed extensively with ethanol and dried in avacuum oven at 40 ◦C.

Glycosyl trichloroacetimidate exhibiting excellent glycosyldonor properties was also tested, and the general synthesis proce-dure was as follows [22]. 2,3,4,6-Tetraacetyl-�-d-glucose (2 g) wasdissolved in freshly dried dichloromethane (20 mL), and treatedwith trichloroacetonitrile (4 mL) and finely powdered K2CO3 (2 g).The solution was stirred at room temperature and monitoredby thin layer chromatography (TLC) till the complete consump-tion of the raw material. The reaction mixture was filtered andconcentrated. Crude product was further purified by column chro-matography [ethy1 acetate–petroleum ether (1:2, v/v)] to afford2,3,4,6-tetraacetyl-�-d-glucopyranosy trichloroacetimidate as ayellow syrup in yield of 80.3% ([�]20

D = +92.5 (CHCl3), 1H NMR(500 MHz, CDCl3) ı (ppm):8.69 (s, 1H, C(NH)CCl3), 6.56 (d, 1H,J = 3.6 Hz, H-1), 5.56 (t, 1H, H-3), 5.18(t, 1H, H-4), 5.13 (dd, 1H, H-2), 4.25–4.12 (m, 3H, H-5, H-6), 2.07(s, 3H, oAc), 2.04 (s, 3H, oAc),2.03(s, 3H, oAc), 2.01(s, 3H, oAc)). The glycosylation procedure wasthe same as described before. The binding density (BD, �g/cm2)of glucose ligands and the reaction ratio (R, %) of hydroxyl groupswere calculated according to the following equations:

W2 − W1

BD =S

(2)

R = Mn(W2 − W1)330(W1 − W0)

× 100% (3)

342 X.-Y. Ye et al. / Colloids and Surfaces B: Biointerfaces 115 (2014) 340– 348

on of P

wcr

5tTPgt

2

es4aSWf(soww

2

aflqgt

Fig. 1. Schematic representation for the glycosylati

here W1 and W2 are the mass of POEGMA/PHEMA grafted andorresponding acetyl glucose modified PPNWMs, respectively. Mn

epresents the molecular weight of OEGMA (360) or HEMA (130).The acetyl groups were deprotected by dipping the samples in

0 mg/mL sodium methoxide/methanol solution for 90 min at roomemperature and washed with de-ionized water several times.he as-prepared glycosylated PPNWMs were assigned as PPNWM-OEGMA-Glu and PPNWM-PHEMA-Glu, respectively. Each of therafting and glycosylation data reported was an average of at leasthree parallel experiments.

.4. Characterization

FT-IR/ATR spectra were carried out on Nicolet Nexus 6700quipped with an ATR accessory (ZnSe crystal, 45◦). Thirty twocans were taken for each spectrum at a normal resolution of

cm−1. Surface morphology of PPNWMs was observed using field-emission scanning electron microscope (FESEM, Hitachi4800, Japan) after samples were sputtering coated with gold.ater contact angle (WCA) was measured to characterize the sur-

ace hydrophilicity of PPNWMs on a CTS-200 contact angle systemMighty Technology Pvt. Ltd., China) at room temperature by ses-ile drop method. Briefly, a water drop (2 �L) was carefully droppednto the sample surface with a microsyringe. Then images of theater droplet were recorded and contact angles were calculatedith system software in equal time intervals.

.5. Static protein adsorption

The bound glucose amount on the PPNWM surface was fixedt 0.3 mg for comparison in the following experiments. Firstly,

uorescein-labeled lectin adsorption assays were performed toualitatively evaluate the specific recognition capability of thelycosylated PPNWMs. Samples with diameter of 3 mm were wet-ed by ethanol for 30 min and replaced by HEPES buffer solution

PNWMs and the lectin affinity adsorption process.

(10 mM, pH 7.5, containing 0.1 mM Ca2+, 0.15 M Na+, 0.01 mM Mn2+

(not for PNA)). Subsequently, they were immersed in 200 �L ofFL-Con A and FL-PNA HEPES buffer solutions with a concentrationof 50 �g/mL for 2 h at 25 ◦C. The samples were then rinsed withfresh HEPES buffer solution 6 times by gentle shaking, each timeusing 200 �L solution for 10 min. After being dried under vacuumat room temperature, fluorescent images of the sample surfaceswere captured by fluorescence microscopy (Nikon ECLIPES Ti-U,Japan). Blue light was adopted to excite the fluorophore FITC andexposure time was 100 ms.

Bradford method was further used to quantitatively determinethe lectin binding capacity and adsorption kinetics of nascent PPN-WMs, PPNWM-POEGMA-Glu and PPNWM-PHEMA-Glu using Con Aas the protein standard, on a UV spectrophotometer (UV-2450, Shi-madzu, Japan). The amount of adsorbed Con A was estimated fromthe differences of Con A concentration before and after incubationat 25 ◦C. Prior to adsorption experiments, the samples were soakedin ethanol for 120 min and washed completely by HEPES buffersolution just as before. The dynamic Con A adsorption process wasmeasured by immersing a piece of the glycosylated PPNWMs in15 mL Con A solution with a concentration 0.1 mg/mL, and then0.3 mL solution was taken out every 10 min until the concentra-tion reached a constant value. The Con A adsorption isothermswere determined by incubating the glycosylated PPNWMs in 3 mLCon A solutions with different concentrations (0.1, 0.2, 0.4, 0.6,0.8, 1.0 mg/mL) and the adsorption time was set for 3 h to ensureequilibrium adsorption.

3. Results and discussion

3.1. Conformal grafting of POEGMA and PHEMA by theplasma-pretreatment and BP-entrapment UV irradiation process

As shown in Fig. 1, the glycosylation procedure of PPNWMsincludes two steps: the surface grafting of functional polymerchains, and the reaction between the functional groups and

s B: Biointerfaces 115 (2014) 340– 348 343

ssatPari(tc

rmshasptiimotd[aaua

gtiartitattroia

ppsnaemsdcTifmsrnf

5 10 15 20 25

0

100

200

300

400

500

600

GD

(g/

cm2 )

UV Time (min)

01020 4060

Plasma Pretra tment Ti me (s )

(a)

4 6 8 10 12 14 16 18 20 220

200

400

600

800

100 0

120 0

GD

(g/

cm2 )

OEGMA Concentration (% w/v)

Plasma 0 sPlasma 40 s

(b)

X.-Y. Ye et al. / Colloids and Surface

accharide derivatives. Advantages of this strategy are complexynthesis and purification of vinyl sugar monomers can be avoidednd the glycosylation reaction does not affect the bulk proper-ies even at high glycosylation density [23,24]. Here, we choseOEGMA and PHEMA as functional polymer chains because theyre widely regarded as hydrophilic polymers that can greatlyesist non-specific protein adsorption, which is beneficial to affin-ty membrane application [25]. Furthermore, the hydroxyl group

OH) is readily coupled with different acetylated saccharides inhe presence of BF3·Et2O and one can compare the effect of sidehain length ( CH2CH2O ) on protein binding capacity.

Surface UV grafting has been commonly employed due to fasteaction rate, simple equipment and versatile for various vinylonomers [26]. BP is the most widely used photoinitiator for the

urface modification of PPNWMs, and different initiating typesave been reported including soaking, adsorption, entrapment,nd sequential living grafting [27]. BP entrapping method is veryimple, based on pre-swelling and subsequent exchanging, thehotoinitiators were tightly immobilized in the surface layer ofhe polypropylene fibers [28]. This process has distinct advantagesncluding the more control of grafted layer structure leading to anmproved membrane adsorber performance and less homopoly-

erization reactions in solution caused by dissolved BP. However,ne important issue has to take into consideration is the spa-ial grafting uniformity of the polymer chains, because UV lightecays continuously as passing through the non-woven meshes9,29]. What’s more, polypropylene fibers of commercial PPNWMsre intact, non-porous and contain small amount of inhibitors orntioxidants. Therefore, a modified UV process is necessary for theniform surface modification of PPNWMs and BP distribution plays

key role here.A plasma pretreatment was introduced and its effect on UV

rafting of OEGMA was carefully studied (Fig. 2). It can be seenhat all the grafting kinetic curves have a ∼5 min polymerizationnduction period which is ascribed to the residual oxygen in thequeous solutions. After this period, the grafting density increasesapidly with increasing UV irradiation time and then levels off dueo the steric inhibition from nearby grafted chains. It is interest-ng to see that the plasma pretreatment has a significant impact onhe UV grafting process, and even when the pretreatment time iss short as 10 s the grafting density increases dramatically. 40 s ishe optimum pretreatment time, however, further increasing timeo 60 s shows adverse effect as long plasma exposure time wouldesult in decomposition (etching effect) and cross-linking reactionn the polypropylene fiber surface. The plasma pretreatment alsoncreases the grafting density at the same monomer concentrationnd UV irradiation time.

The increased UV grafting density by the introduction of plasmaretreatment may be mainly attributed to three aspects. First,lasma pretreatment can introduce polar groups onto the fiberurface and usually results in increased surface waviness or rough-ess [30], even sometimes obvious wrinkled or grooved structureppears (Fig. 3b). Both of them tend to promote the adsorption andntrapment of BP to the fiber surface. Second, plasma pretreat-ent will generate peroxide/hydroperoxide groups on the fiber

urface and thus free radicals can be produced under UV irra-iation [31–33]. We found that a very small amount of OEGMAould be grafted on PPNWMs even without BP-entrapment step.hird, the plasma pretreatment greatly increases the hydrophilic-ty of PPNWMs and facilitates the diffusion of hydrophilic monomerrom solution to the fiber surface. Furthermore, plasma pretreat-

ent improves the conformity of grafted POEGMA on the fiber

urface (Fig. 3c and d). PPNWMs without pretreatment have manyibbon-like fibers while the plasma pretreated one preserve theascent cylindrical shape after grafting. The improvement in con-

ormity is ascribed to the enhanced BP and monomer distribution

Fig. 2. Effect of (a) plasma pretreatment time and UV irradiation time (OEGMA con-centration was fixed at 15% (w/v)) and (b) OEGMA concentration (UV irradiationtime was fixed at 18 min) on the grafting density.

uniformity [9], or else autoacceleration effect and chain transferreaction would even cause the formation of POEGMA gel [34].Under the optimum plasma pretreatment time 40 s, the effects ofUV irradiation time and monomer concentration on the graftingbehavior of HEMA were further conducted and similar results werealso found (Fig. S1 in Supplementary material). In short, by using amodified plasma-pretreatment and BP-entrapment UV irradiationprocess, we achieved conformal grafting of POEGMA and PHEMAon the fiber surfaces of PPNWMs with high grafting density.

3.2. Surface glycosylation and structure characterization

The glycosylated PPNWMs were prepared through chemicalreaction between the grafted chains and glucose pentaacetateunder BF3·Et2O catalysis followed by a deacetylation process. Theinfluences of grafting density of POEGMA or PHEMA on the bind-ing density of glucose ligands and the reaction ratio of hydroxylgroups with excess glucose pentaacetate (20 equiv) and BF3·Et2O(100 equiv) were studied. Typical results are shown in Fig. 4. It canbe seen that the binding density increases remarkably with increas-ing grafting density and ∼140 �g/cm2 of acetyl glucose moiety can

be reached. The reaction ratios of hydroxyl groups lie in the rangeof 10–30% and the case of POEGMA only has slightly higher valuesthan that of PHEMA. The glycosylation reaction here is heteroge-neous and the steric hindrance effect is predominant. Therefore, it

344 X.-Y. Ye et al. / Colloids and Surfaces B: Biointerfaces 115 (2014) 340– 348

F pretrw olypr

dtitttacpm[gtbnthc[

cPiCbhoattptdp

ig. 3. SEM pictures of the pristine and modified PPNWMs: (a) nascent; (b) plasmaithout plasma pretreatment. The insets show the surface morphology of a single p

eclines at first for the reaction ratio of hydroxyl groups. With fur-her increasement of the grafting density, the reaction ratio slowlyncreases due to a significant increase in the concentration of reac-ion substrate and catalyst. 2,3,4,6-Tetraacetyl-�-d-glucopyranosyrichloroacetimidate was also synthesized and used, as glycosylrichloroacetimidates are very powerful glycosyl donors undercidic conditions. It was reported that a high degree of lactoseonjugation onto the hydroxyl-terminated PEO star and dendriticolymers could be achieved by trichloroacetimidate glycosidationethodology, using BF3·Et2O in homogeneous solution reaction

35]. However, we found that the conjugation ratio of hydroxylroups by using glycosyl trichloroacetimidate was approximateo glucose pentaacetate (data not shown). This result is mainlyecause both of them experience a SN2 type reaction mecha-ism and form oxocarbenium ion intermediates [36], and thushe reaction ratio of hydroxyl groups is limited by the stericindrance effect. Finally, the deprotection of acetyl groups of sac-haride residues was accomplished to yield glycosylated PPNWMs21].

FT-IR/ATR spectroscopy was used to monitor the chemicalhanges of PPNWMs surface (Fig. 5). Compared with the nascentPNWMs, POEGMA and PHEMA grafted ones exhibit additionalntense adsorption peak at 1725 cm−1, which is ascribed to the

O stretching vibration aroused by the major characteristic car-onyl groups. It should be noted that POEGMA grafted PPNWM hasigher peak intensity of C O C (1165 cm−1) than PHEMA graftedne as it contains more CH2CH2O chain segments. When thecetyl glucose moieties were bound, the O H stretching vibra-ion (3100–3500 cm−1) decreases yet not vanishes, demonstratinghe existence of partial unreacted hydroxyl groups. Furthermore,

eak at 1750 cm−1 can be observed, associated with the acetyl pro-ected glucose groups. After deprotection, the band at 1750 cm−1

isappears in accordance with the mass reduction caused by com-lete removal of the acetyl groups, and the band ranging from

eatment; (c) POEGMA grafted with plasma pretreatment and (d) POEGMA graftedopylene fiber.

3100–3500 cm−1 becomes much more intense, which is ascribedto the exposed free hydroxyl groups on the glucose ligands.

The surface morphology of the glycosylated PPNWMs wasobserved by FESEM (Fig. S2 in Supplementary material). Veryconformal glucose ligands can be attained on both POEGMAand PHEMA grafted PPNWM surfaces. Besides, the glycosy-lated PPNWMs can well preserve the nascent cylindrical fibershape and the pore structure with high glucose binding den-sity, which are important physical factors for protein affinityadsorption. Surface wettability was also evaluated by WCAexperiments, and it is found that glycosylation can endowPPNMWs with superhydrophilicity. (Fig. S3 in Supplementarymaterial).

3.3. Lectin affinity adsorption by the glycosylated PPNWMs

FL-Con A and FL-PNA were used to qualitatively investigatethe specific recognition capability of the glycosylated PPNWMs.The bound glucose amount was fixed at 0.3 mg (ensure glyco-cluster effect) for comparing the influences of the side chainlength. Con A shows specificity for mannose, glucose and acetylglucosamine with 3-,4-,6-hydroxyl groups while PNA binds pref-erentially to galactose and acetyl galactosamine [37]. As shownin Fig. 6, slight green fluorescence was detected on the nascentPPNWMs incubating in FL-Con A and FL-PNA solutions, whichresults from non-specific protein adsorption on the hydrophobicsurfaces. In contrast, no fluorescence was observable on the glyco-sylated PPNWMs surface after exposure to FL-PNA solution whilenotably green fluorescence appeared after exposure to FL-Con Asolution. It indicates that the glycosylated PPNWMs can bind a

large amount of Con A and resist non-specific PNA adsorption atthe same time. These glycosylated PPNWMs with high specificrecognition capability can be used as affinity membranes for ConA purification, and the quantitative Con A binding capacity was

X.-Y. Ye et al. / Colloids and Surfaces B: Biointerfaces 115 (2014) 340– 348 345

20

40

60

80

100

120

140

160

100 200 300 400 500 60 0 70 0

40

30

10

Glu

cose

BD

(g/

cm2 )

POEGMA GD ( g/cm2)

Ratio (%

)

20

(a)

(b)5

10

15

20

25

30

100 20 0 300 40 020

40

60

80

100

120

140

160

180

Glu

cose

BD

(g/

cm2 )

Ratio (%

)

PHEMA GD ( g/cm2)

Fig. 4. Effect of the GD of (a) POEGMA and (b) PHEMA on the binding density ofglucose ligands and reaction ratio of hydroxyl groups. Glucose pentaacetate was2P

ftnrm

PirtPti6ciamneicpWp

3500 3000 2500 2000 1500 1000

Wavenumber (c m-1)

bc

a

d

(A)

3500 3000 2500 2000 1500 1000

Wavenumber (cm-1)

d

cba

(B)

Fig. 5. FT-IR/ATR spectra of (A) POEGMA and (B) PHEMA tethered PPNWMs: (a)nascent; (b) OH grafted; (c) acetyl glucose grafted; (d) glycosylated. The insets show

0 equiv and BF3•Et2O was 100 equiv relative to the monomer unit grafted on the

PNWMs.

urther determined by Bradford method. It should be mentionedhat both POEGMA and PHEMA grafted PPNWMs greatly resistedon-specific protein adsorption (Fig. S4 in Supplementary mate-ial) and no protein adsorption could be detected using Bradfordethod.The adsorption kinetics of Con A on the nascent PPNWMs,

PNWM-POEGMA-Glu and PPNWM-PHEMA-Glu were first stud-ed (Fig. 7a). It can be clearly seen that Con A had fast adsorptionate on the studied PPNWMs and the amount of adsorbed Con Aended to reach an equilibrium value after 60 min. The nascentPNWMs showed about 10 �g/cm2 Con A binding capacity due tohe non-specific hydrophobic interaction, while the binding capac-ty of PPNWM-POEGMA-Glu and PPNWM-PHEMA-Glu was around0 �g/cm2 and 40 �g/cm2, respectively. The introduction of glu-ose ligands greatly increased the Con A binding capacity and thencreased side chain length had a positive effect on the Con Adsorption. Mammen and Stenzel suggested it was important toatch the distance between the saccharide ligand and the recog-

ition sites of lectin for good binding configuration [38,39]. It isvident here that POEGMA has a flexible (CH2CH2O)6 spacerncreasing the chain mobility and then enhancing the lectin bindingapacity. Compared with our previous glycosylated microporous

olypropylene membrane [40], we found that glycosylated PPN-Ms had faster adsorption kinetics (60 min vs. 5 h) and higher

rotein binding capacity (0.3 mg glucose/0.3 mg Con A vs. 0.64 mg

the spectra in the range of 1650–1850 cm−1.

glucose/0.12 mg Con A) under identical conditions, which can bemainly attributed to the unique inter-connected pore structure thatsignificantly reduces the mass transport and diffusional resistance.

The static adsorption isotherm was carried out in different ConA concentrations and the adsorption time was fixed at 3 h to reachequilibrium adsorption. As shown in Fig. 7b, the amount of ConA adsorbed on the glycosylated PPNWMs increased sharply atthe very beginning and then leveled off approaching saturationvalues. The Con A binding capacity decreased in the follow-ing sequence: PPNWM-POEGMA-Glu > PPNWM-PHEMA-Glu > thenascent PPNWM, which further demonstrates the roles of glucoseligands and flexible CH2CH2O spacer.

Langmuir (monolayer) and Freundlich (multilayer) adsorptionmodels were both used to quantitatively analyze the adsorptionisotherms, detailed instructions and calculated data are listed inTable 1. We found that both of the two models verified thatPPNWM-POEGMA-Glu had the highest adsorption capacity, and theFreundlich model fitted the adsorption isotherm in a better waycompared with Langmuir model. This result suggests a multilayeradsorption behavior of Con A on the glycosylated PPNWMs. It canbe further theoretically evaluated by concerning the 3D size of ConA (molecular weight 104 000 Da, 6.7 nm × 11.3 nm × 12.2 nm [41]),the total surface area of polypropylene fibers and adsorbed amount

of Con A in the sample. In our case, the theoretical saturationcapacity is in the range of 0.13–0.23 �g/cm2 when Con A is in mono-layer adsorption using geometric consideration. For glycosylated

346 X.-Y. Ye et al. / Colloids and Surfaces B: Biointerfaces 115 (2014) 340– 348

F -Con

P ale ba

Pc

q

TA

C

r

ig. 6. Fluorescent pictures of the PPNWMs surfaces after incubating with (a–c) FLHEMA-Glu. The bound glucose amount was fixed at 0.3 mg for comparison. The sc

PNWMs, the theoretical Con A binding capacity on single fiberan be calculated by the following equation:

= W4�D

4W0(4)

able 1dsorption behaviors of Con A on the studied PPNWMs.

Sample Langmuir adsorptiona

Ka (× 105) Qe (�g/cm2)

Nascent PPNWMs 3.77 57.24

PPNWM-PHEMA-Glu 3.25 158.98

PPNWM-POEGMA-Glu 5.03 188.32

a Data were calculated by equation [C]/Q = [C]/Qe + 1/Qe × 1/Ka , where Q is the measuredon A, [C] is the equilibrium concentration of Con A in solution, and Ka is the value of adsb Data were calculated by equation log Q = log Kf + n log[C], where Kf and n are the Freundl

espectively.

A and (a’–c’) FL-PNA: (a, a’) nascent; (b, b’) PPNWM-POEGMA-Glu; (c, c’) PPNWM-r is 50 �m, 400× magnification.

where q is the Con A binding capacity �g/cm2, W0 and W4 arethe mass of the nascent PPNWMs (16 mg) and adsorbed Con A,respectively. � is the density of polypropylene (0.91 g/cm3) and D

is the average diameter of polypropylene fibers (8 �m). Calculatedresults show that there are approximately at least 15 layers of ConA molecules stacked around each fiber. The multilayer adsorption

Freundlich adsorptionb

R2 Log Kf n R2

0.975 1.21 0.49 0.9560.903 1.63 0.49 0.9690.905 1.81 0.43 0.958

amounts of adsorbed Con A, Qe is the theoretical equilibrium amounts of adsorbedorption equilibrium constant.ich characteristic constants indicating adsorption capacity and adsorption intensity,

X.-Y. Ye et al. / Colloids and Surfaces B: B

0 20 40 60 80 10 0 12 0

0

10

20

30

40

50

60

70

PPNWM-PHEMA-Glu

PPNWM-POEGMA-Glu

Q (

g/cm

2 )

Time (min)

Nascent(a)

0.0 0.2 0.4 0.6 0.8 1.0-20

0

20

40

60

80

100

120

140

160

PPNWM-POEGMA-Glu

Q (

g/cm

2 )

Equilibrium protein concenration (mg/mL)

Nascent

PPNWM-PHEMA-Glu

(b)

Faa

bwom

4

bprtccBPcsiaPs

A

N

[

[

[

[

[

[

[

[

[

[

[

[

[

[

[

ig. 7. (a) Adsorption kinetics of Con A (concentration 0.1 mg/mL) and (b) Con Adsorption isotherms (adsorption time 3 h) on PPNWMs at 25 ◦C. The bound glucosemount was fixed at 0.3 mg for comparison.

ehavior of Con A is reasonable, because Con A is a glycoproteinith intermolecular interactions and can diffuse into the inner layer

f the flexible polymer brushes [14,42]. Furthermore, such affinityembranes can be easily regenerated by 1 M HAc [42].

. Conclusions

In summary, glycosylated PPNWMs were successfully preparedy a modified UV grafting and chemical reaction process. The rapidlasma pretreatment and BP entrapment UV irradiation methodesulted in conformal and uniform POEGMA/PHEMA brushes onhe polypropylene fiber surface with high grafting density. Theonjugation ratio between the hydroxyl groups and acetyl glu-ose ligands was limited by the steric hindrance effect underF3·Et2O catalysis. After deprotection, the as-prepared glycosylatedPNWMs had superhydrophilicity and high specific recognitionapability toward Con A. The introduction of flexible CH2CH2Opacer further increased the protein binding capacity due tomproved chain mobility. Con A had fast adsorption rate, multilayerdsorption behavior and high binding capacity on the glycosylatedPNWMs, which make them promising affinity membranes for theeparation and purification of lectins.

cknowledgments

The authors are grateful to the financial support from theational Natural Science Foundation of China (Grant no. 50933006)

[

[

iointerfaces 115 (2014) 340– 348 347

and Zhejiang Provincial Innovative Research Team (Grant no.2009R50004).

Appendix A. Supplementary data

Supplementary data associated with this article can befound, in the online version, at http://dx.doi.org/10.1016/j.colsurfb.2013.12.025.

References

[1] S. Jaganathana, H.V. Tafreshib, B. Pourdeyhimi, A realistic approach for model-ing permeability of fibrous media: 3-D imaging coupled with CFD simulation,Chem. Eng. Sci. 63 (2008) 244–252.

[2] R.S. Barhate, S. Ramakrishna, Nanofibrous filtering media: filtration problemsand solutions from tiny materials, J. Membr. Sci. 296 (2007) 1–8.

[3] V. Orr, L.Y. Zhong, M.M. Young, C.P. Chou, Recent advances in bioprocess-ing application of membrane chromatography, Biotechnol. Adv. 31 (2013)450–465.

[4] A. Prudente, C.L.Z. Riccetto, M.M.S.G. Simoes, B.M. Pires, M.G. Oliveira, Impreg-nation of implantable polypropylene mesh with S-nitrosoglutathione-loadedpoly(vinyl alcohol), Colloid. Surf. B: Biointerfaces 108 (2013) 178–184.

[5] S. Lee, S.K. Obendorf, Developing protective textile materials as barriers toliquid penetration using melt-electrospinning, J. Appl. Polym. Sci. 102 (2006)3430–3437.

[6] R. Uppal, G. Bhat, C. Eash, K. Akato, Meltblown nanofiber media for enhancedquality factor, Fibers Polym. 14 (2013) 660–668.

[7] E. Klein, Affinity membranes: a 10-year review, J. Membr. Sci. 179 (2000) 1–27.[8] H.Y. Liu, Y. Zheng, P.V. Gurgel, R.G. Carbonell, Affinity membrane development

from PBT nonwoven by photo-induced graft polymerization, hydrophilizationand ligand attachment, J. Membr. Sci. 428 (2013) 562–575.

[9] Y. Zheng, H.Y. Liu, P.V. Gurgel, R.G. Carbonell, Polypropylene nonwoven fabricswith conformal grafting of poly(glycidyl methacrylate) for bioseparations, J.Membr. Sci. 364 (2010) 362–371.

10] D.M. Kanani, E. Komkova, T. Wong, A. Mika, R.F. Childs, R. Ghosh, Separationof human plasma proteins HSA and HIgG using high-capacity macroporousgel-filled membranes, Biochem. Eng. J. 35 (2007) 295–300.

11] Z. Ma, S. Ramakrishna, Electrospun regenerated cellulose nanofiber affinitymembrane functionalized with protein A/G for IgG purification, J. Membr. Sci.319 (2008) 23–28.

12] L.J. Zhu, L.P. Zhu, Z. Yi, J.H. Jiang, B.K. Zhu, Y.Y. Xu, Hemocompatible and antibac-terial porous membranes with heparinized copper hydroxide nanofibers asseparation layer, Colloids Surf. B: Biointerfaces 110 (2013) 36–44.

13] T.J. Menkhaus, H. Varadaraju, L.F. Zhang, S. Schneiderman, S. Bjustrom, L. Liu,H. Fong, Electrospun nanofiber membranes surface functionalized with 3-dimensional nanolayers as an innovative adsorption medium with ultra-highcapacity and throughput, Chem. Commun. 46 (2010) 3720–3722.

14] W. Guo, E. Ruckenstein, A new matrix for membrane affinity chromatographyand its application to the purification of concanavalin A, J. Membr. Sci. 182(2001) 227–234.

15] T.W. Rademacher, R.B. Parekh, R.A. Dwek, Glycobiology, Annu. Rev. Biochem.57 (1988) 785–838.

16] J.J. Lundquist, E.J. Toone, The cluster glycoside effect, Chem. Rev. 102 (2002)555–578.

17] H. Lis, N. Sharon, Lectins: carbohydrate-specific proteins that mediate cellularrecognition, Chem. Rev. 98 (1998) 637–674.

18] L. Bech, B. Lepoittevin, A.E. Achhab, E. Lepleux, L.T. Gay, C.B. Laporte, P.Roger, Double plasma treatment-induced graft polymerization of carbohy-drated monomers on poly(ethylene terephthalate) fibers, Langmuir 23 (2007)10348–10352.

19] L. Renaudie, C.L. Narvor, E. Lepleux, P. Roger, Functionalization of poly(ethyleneterephthalate) fibers by photografting of a carbohydrate derivatized with aphenyl azide group, Biomacromolecules 8 (2007) 679–685.

20] L. Bech, T. Meylheuc, B. Lepoittevin, P. Roger, Chemical surface modificationof poly(ethylene terephthalate) fibers by aminolysis and grafting of carbohy-drates, J. Appl. Polym. Sci. 45 (2007) 2172–2183.

21] M.X. Hu, Z.K. Xu, Carbohydrate decoration of microporous polypropylene mem-branes for lectin affinity adsorption: comparison of mono- and disaccharides,Colloids Surf. B: Biointerfaces 85 (2011) 19–25.

22] R.R. Schmidt, J. Michel, Facile synthesis of �-and �-O-glycosyl imidates: prepa-ration of glycosides and disaccharides, Angew. Chem. Int. Ed. Engl. 19 (1980)731–732.

23] Z.W. Dai, L.S. Wan, Z.K. Xu, Surface glycosylation of polymeric membranes, Sci.China. Ser. B: Chem. 51 (2008) 901–910.

24] A.F. Che, X.J. Huang, Z.K. Xu, Polyacrylonitrile-based nanofibrous membranewith glycosylated surface for lectin affinity adsorption, J. Membr. Sci. 366(2011) 272–277.

25] T.B. Stachowiak, F. Svec, J.M.J. Frechet, Patternable protein resistant surfacesfor multifunctional microfluidic devices via surface hydrophilization of porouspolymer monoliths using photografting, Chem. Mater. 18 (2006) 5950–5957.

26] J.P. Deng, L.F. Wang, L.Y. Liu, W.T. Yang, Developments and new applications ofUV-induced surface graft polymerizations, Prog. Polym. Sci. 34 (2009) 156–193.

3 es B: B

[

[

[

[

[

[

[

[

[

[

[

[

[

[

48 X.-Y. Ye et al. / Colloids and Surfac

27] D.M. He, H. Susanto, M. Ulbricht, Photo-irradiation for preparation, modifi-cation and stimulation of polymeric membranes, Prog. Polym. Sci. 34 (2009)62–98.

28] M. Ulbricht, H. Yang, Porous polypropylene membranes with different carboxylpolymer brush layers for reversible protein binding via surface-initiated graftcopolymerization, Chem. Mater. 17 (2005) 2622–2631.

29] Y. Zheng, P.V. Gurgel, R.G. Carbonell, Effects of UV exposure and initiator con-centration on the spatial variation of poly(glycidyl methacrylate) grafts onnonwoven fabrics, Ind. Eng. Chem. Res. 50 (2011) 6115–6123.

30] Y.J. Hwang, M.G. Mccord, J.S. An, B.C. Kang, S.W. Park, Effects ofhelium atmospheric pressure plasma treatment on low-stress mechani-cal properties of polypropylene nonwoven fabrics, Text. Res. J. 75 (2005)771–778.

31] C. Wang, Y. Fan, M.X. Hu, W. Xu, J. Wu, P.F. Ren, Z.K. Xu, Gly-cosylation of the polypropylene membrane surface via thiol–yne clickchemistry for lectin adsorption, Colloids Surf. B: Biointerfaces 110 (2013)105–112.

32] S.F. Luan, J. Zhao, H.W. Yang, H.C. Shi, J. Jin, X.M. Li, J.C. Liu, J.W. Wang, J.H. Yin,P. Stagnaro, Surface modification of poly(styrene-b-(ethylene-co-butylene)-b-styrene) elastomer via UV-induced graft polymerization of N-vinyl pyrrolidone,Colloids Surf. B: Biointerfaces 93 (2012) 127–134.

33] C. Zhang, J. Jin, J. Zhao, W. Jiang, J.H. Yin, Functionalized polypropylene non-woven fabric membrane with bovine serum albumin and its hemocompatibilityenhancement, Colloids Surf. B: Biointerfaces 102 (2013) 45–52.

34] Y. Chang, C.Y. Ko, Y.J. Shih, D. Quemener, A. Deratani, T.C. Wei, D.M. Wang, J.Y.Lai, Surface grafting control of PEGylated poly(vinylidene fluoride) antifouling

[

[

iointerfaces 115 (2014) 340– 348

membrane via surface-initiated radical graft copolymerization, J. Membr. Sci.345 (2009) 160–169.

35] S.M. Rele, W.X. Cui, L.C. Wang, S.J. Hou, G.B. Zarae, D. Tatton, Y. Gnanou, J.D. Esko,E.L. Chaikof, Dendrimer-like PEO glycopolymers exhibit anti-inflammatoryproperties, J. Am. Chem. Soc. 127 (2005) 10132–10133.

36] R.R. Schmidt, X.M. Zhu, Glycosyl trichloroacetimidates, in: R.B. Fraser, K.Tatsuta, J. Thiem (Eds.), Glycoscience, Springer-Verlag, Berlin, Heidelberg,Germany, 2008, pp. 452–517.

37] I.J. Goldstein, C.E. Hollerman, E.E. Smith, Protein-carbohydrate interaction 2.Inhibition studies on interaction of concanavalin a with polysaccharides, Bio-chemistry 4 (1965) 876–883.

38] M. Mammen, S.K. Choi, G.M. Whitesides, Polyvalent interactions in biologicalsystems: implications for design and use of multivalent ligands and inhibitors,Angew. Chem. Int. Ed. 37 (1998) 2755–2794.

39] J. Kumar, L. McDowall, G.J. Chen, M.H. Stenzel, Synthesis of thermo-responsiveglycopolymers via copper catalysed azide–alkyne ‘click’ chemistry for inhibi-tion of ricin: the effect of spacer between polymer backbone and galactose,Polym. Chem. 2 (2011) 1879–1886.

40] C. Wang, P.F. Ren, X.J. Huang, J. Wu, Z.K. Xu, Surface glycosylation of polymermembrane by thiol-yne click chemistry for affinity adsorption of lectin, Chem.Commun. 47 (2011) 3930–3932.

41] J. Bouckaert, F. Poortmans, L. Wyns, R. Loris, Sequential structural changes uponzinc and calcium binding to metal-free concanavalin A, J. Biol. Chem. 271 (1996)16144–16150.

42] M.X. Hu, L.S. Wan, Z.K. Xu, Multilayer adsorption of lectins on glycosylatedmicroporous polypropylene membranes, J. Membr. Sci. 335 (2009) 111–117.