Self-assembled carbohydrate-based vesicles for lectin targeting

Colloids and Surfaces B: Biointerfaces 148 (2016) 12–18

Contents lists available at ScienceDirect

Colloids and Surfaces B: Biointerfaces journal homepage: www.elsevier.com/locate/colsurfb

Self-assembled carbohydrate-based vesicles for lectin targeting Marinalva Cardoso dos Santos a , Yasmine Miguel Serafini Micheletto b , Nadya Pesce da Silveira b , Luciano da Silva Pinto c , Fernando Carlos Giacomelli d , Vânia Rodrigues de Lima a , Tiago Elias Allievi Frizon e , Alexandre Gonc¸alves Dal-Bó e,∗ a

Universidade Federal do Rio Grande (FURG), 96203-900, Rio Grande, RS, Brazil Universidade Federal do Rio Grande do Sul (UFRGS), 91501-970, Porto Alegre, RS, Brazil Universidade Federal de Pelotas, Campus Universitário, 96010-900, Pelotas, RS, Brazil d Universidade Federal do ABC (UFABC), 09210-170, Santo André, SP, Brazil e Universidade do Extremo Sul Catarinense (UNESC), 88806-000, Criciúma, SC, Brazil b c

a r t i c l e

i n f o

Article history: Received 24 June 2016 Received in revised form 26 August 2016 Accepted 29 August 2016 Keywords: Glycosurfactant Glycosylated surface Vesicles Lectin

a b s t r a c t This study examined the physicochemical interactions between vesicles formed by phosphatidylcholine (PC) and glycosylated polymeric amphiphile N-acetyl-ˇ-d-glucosaminyl-PEG900 -docosanate (C22 PEG900 GlcNAc) conjugated with Bauhinia variegata lectin (BVL). Lectins are proteins or glycoproteins capable of binding glycosylated membrane components. Accordingly, the surface functionalization by such entities is considered a potential strategy for targeted drug delivery. We observed increased hydrodynamic radii (RH ) of PC + C22 PEG900 GlcNAc vesicles in the presence of lectins, suggesting that this aggregation was due to the interaction between lectins and the vesicular glycosylated surfaces. Furthermore, changes in the zeta potential of the vesicles with increasing lectin concentrations implied that the vesicular glycosylated surfaces were recognized by the investigated lectin. The presence of carbohydrate residues on vesicle surfaces and the ability of the vesicles to establish specific interactions with BVL were further explored using atomic force microscopy (AFM) and small-angle X-ray scattering (SAXS) analysis. The results indicated that the thickness of the hydrophilic layer was to some extent influenced by the presence of lectins. The presence of lectins required a higher degree of polydispersity as indicated by the width parameter of the log-normal distribution of size, which also suggested more irregular structures. Reflectance Fourier transform infrared (HATR-FTIR), differential scanning calorimetry (DSC), nuclear magnetic resonance (NMR) and ultraviolet-visible (UV–vis.) analyses revealed that the studied lectin preferentially interacted with the choline and carbonyl groups of the lipid, thereby changing the choline orientation and intermolecular interactions. The protein also discretely reduced the intermolecular communication of the hydrophobic acyl chains, resulting in a disordered state. © 2016 Elsevier B.V. All rights reserved.

1. Introduction The surface functionalization of nanoparticles and liposomes has become of increasing importance since Ehrlich publicized his vision of the “Magic Bullet” [1]. Vesicles and liposomes (phospholipid-based vesicles) have been extensively studied due to their widespread application as controlled drug delivery vehicles in the pharmaceutical industry and as biomembranes [2–4]. Because of their lipid composition, vesicles have lower toxicity than other vehicles, making them promising systems for the delivery

∗ Corresponding author. E-mail addresses: [email protected], [email protected] (A.G. Dal-Bó). http://dx.doi.org/10.1016/j.colsurfb.2016.08.053 0927-7765/© 2016 Elsevier B.V. All rights reserved.

of a wide range of drugs requiring specific treatments, controlled circulation times, reduced side effects and optimum drug action [5–11]. The possibility of directing a drug toward targeted tissues without changing its structure and hence, its biological activity, is fundamental for therapeutic applications [12]. Thus, efforts have been directed toward improving the vectorization of drug delivery systems to specific target tissues. In this context, the surface functionalization of vesicles with natural or synthetic glycolipids has been proposed to enhance the specificity of vesicles for lectins [12,13]. Lectins are non-immunological proteins or glycoproteins that specifically recognize sugar molecules and are capable of binding glycosylated membrane components. They are widely used to characterize carbohydrates on cell surfaces [14]. The Bauhinia variegata lectin (BVL) is particularly found in Caesalpinoideae plants. Its subunit has molecular weight of ∼33 kDa and diffraction pat-

M.C. dos Santos et al. / Colloids and Surfaces B: Biointerfaces 148 (2016) 12–18

ter similar to related lectins such as Bauhinia purpurea agglutinin (BPA) [15]. The carbohydrate-lectin binding typically involves two or three terminal sugar residues of mammalian glycans, including galactose, mannose, N-acetyl-neuraminic acid, fucose, and N-acetyl-glucosamine [16,17]. High levels of lectins, such as galectin-3, have been detected in various cancers [12]. The specificity of the carbohydrate-lectin interaction has been exploited to convey glycosylated liposomes to tumor cells [12,18–20]. Therefore, the development of nanoparticles with outer shells decorated with glycoconjugates for lectin targeting is considered a promising means to improve the delivery and internalization of antitumor drugs [16,21]. Previously, our research group described the physicochemical interactions between vesicles composed of phosphatidylcholinepurified soybean lecithin and the glycosylated polymeric amphiphile N-acetyl-ˇ-d-glucosaminyl-PEG900 -docosanate conjugate (C22 PEG900 GlcNAc). Structurally, the ∼100 nm composite vesicles self-assemble via attractive force between the lipidic region and the nitrogen groups of C22 PEG900 GlcNAc. The results also suggested discrete interaction among the hydroxyl and carbonyl regions of C22 PEG900 GlcNAc [22]. Amphiphiles containing PEG chains have become a topic of interest due to their biocompatibility and anomalous behavior in water [23]. Through the PEGylation of vesicles, nanoparticles, and proteins, the residence times of these carriers can be significantly extended and diminish their uptake by the organs (e.g., liver and spleen) of the reticulo endothelial system (RES) [24,25]. This present work reports on the interaction between selfassembled vesicles (composed of phosphatidylcholine-purified soybean lecithin and the glycosylated polymeric amphiphile, C22 PEG900 GlcNAc) and BVL. The physicochemical properties of the phosphatidylcholine-based vesicle system containing C22 PEG900 GlcNAc and lectin were investigated by several techniques, including zeta potential, dynamic light scattering (DLS), atomic force microscopy (AFM), small-angle X-ray scattering (SAXS), horizontal attenuated total reflectance Fourier transform infrared (HATR-FTIR), differential scanning calorimetry (DSC), nuclear magnetic resonance (NMR) and ultraviolet-visible (UV-vis.) spectroscopy measurements. The physicochemical properties of these self-assembled vesicles presented herein could contribute to improved vectorizations of drug delivery systems in cancer therapy. 2. Materials and methods

13

2.3. Vesicles preparation Vesicles were prepared by a reverse-phase evaporation method [26,29]. The preparation of the vesicles without the lectin was previously presented (reference [20]). Herein the preparation has been performed in phosphate buffered saline (PBS, 10 mM, pH 7.2) instead of water. The types of vesicle were prepared comprising self-assembled structures between PC and C22 PEG900 GlcNAc (PC + C22 PEG900 GlcNAc) in the absence (pure) and in the presence of Bauhinia variegata lectin at a PC:C22 PEG900 GlcNAc:lectin ratio of 1:1:1 (w/w). Lectin was previously solubilized in 10 mM PBS and added in the organogel hydration step. The solution was then filtered using 0.45-␮m pore-size nylon-membrane filters to remove dust and large aggregates. The final vesicle concentration was 15 mg/mL. 2.4. Electrophoretic light scattering The zeta potential measurements were performed on a ZetaPALS Zeta Potential Analyzer (Brookhaven Instruments) with a coherent He-Ne 632.8 nm laser. Vesicle suspensions (15 mg/mL) in phosphate buffered saline (PBS, 10 mM, pH 7.2) were diluted to 1 mg/mL at 20 ± 1 ◦ C for the zeta potential measurements. Ten measurements were performed in triplicate for each sample. 2.5. Dynamic light scattering (DLS) Dynamic light scattering (DLS) experiments were performed on a BI-200 goniometer with a BI-9000 AT digital correlator (Brookhaven Instruments) with a He-Ne laser (␭ = 632.8 nm) as a light source. For the DLS experiments, the scattering volume was minimized using a 0.4 mm aperture and an interference filter before detecting the signal on the photomultiplier at a 90◦ angle. The autocorrelation functions were obtained in a multi-␶ mode using 224 channels. The samples were filtered through membrane filters with a pore size of 0.45 ␮m directly into the sample cell and placed in the index matching liquid decahydronaphthalene (Aldrich). The autocorrelation functions were analyzed by using the nonlinear inverse Laplace transformation algorithm REPES [30] resulting the distributions of size. The temperature used for the analyses was 20 ± 1 ◦ C. The relaxation frequency,  = 1/, is a function of the scattering angle [31]. The apparent diffusion coefficient Dapp of the nanoparticles was calculated from Eq. (1): q → 0|

2.1. Materials

 = Dapp q2

(1)

where q is the scattering vector: Phosphatidylcholine (PC) from soybean lecithin (95% phosphatidylcholine, 5% lysophosphatidylcholine and phosphatidic acid) was a gift from Solae do Brasil S.A. The molecular composition of the soybean PC was approximately 75% distearoylphosphatidylcholine (DSPC, 18:0), 12% dioleoylphosphatidylcholine (DOPC, 18:2) and 8% dipalmitoylphosphatidylcholine (DPPC, 16:0) [26]. All reagents were of commercial grade and were used as received unless otherwise noted. The native Bauhinia variegata (BVL) lectin was extracted from seeds, according to the protocol described by Pinto et al. [15]. 2.2. Synthesis of C22 PEG900 GlcNAc and soybean lecithin purification

q=

4n sin 

   2

(2)

 is the wavelength of the incident laser beam (632.8 nm), n is the refractive index of the sample, and  is the scattering angle. The hydrodynamic radius RH (or diameter, 2RH ) was calculated using the Stokes-Einstein relation given in Eq. (3): RH =

Kb T 6Dapp

(3)

where kB is the Boltzmann constant, T is the temperature of the sample and  is the viscosity of the solvent (water). 2.6. Small angle X-ray scattering (SAXS)

C22 PEG900 GlcNAc was synthesized according to a previously reported method [27] and phosphatidylcholine (PC) was obtained through the purification of soybean lecithin according to previously reported methods [22,26,28].

Small angle X-ray scattering (SAXS) measurements were performed at the SAXS1 beamline of the Brazilian Synchrotron Light Laboratory (LNLS − Campinas, SP, Brazil) operating at wavelength

14

M.C. dos Santos et al. / Colloids and Surfaces B: Biointerfaces 148 (2016) 12–18

 = 1.55 Å. The system consisted of a temperature-controlled vacuum flow-through cell and a Pilatus 300 K 2D detector (Dectris). The 2D-images were found to be isotropic and were normalized by the sample transmission undertaken using the FIT2D software. The I(q) vs. q scattering curves were corrected by subtracting the pure solvent scatter.

2.7. Atomic force microscopy (AFM) AFM was used to investigate the morphology of the casting films of the PC + C22 PEG900 GlcNAc vesicles in the absence and the presence of lectin. The AFM measurements were carried out using a Shimadzu SPM-9700 instrument (Kyoto, JP) with dynamic mode scanning. Vesicle suspensions (15 mg/mL) were diluted to 0.1 mg/mL in phosphate buffered saline (PBS, 10 mM, pH 7.2), dropped onto a mica surface and dried at room temperature.

2.8. HATR-FTIR measurements FTIR spectra of the PC + C22 PEG900 GlcNAc vesicles were obtained on a Shimadzu-IR Prestige-21 spectrophotometer (Kyoto, JP). Interferograms were averaged over 45 scans at 2 cm−1 resolution and recorded from 4000 to 400 cm−1 [32]. The FTIR spectra were analyzed based on the frequency shifts and bandwidth (calculated at 75% of the peak height) variations of peaks attributed to the axial stretching vibrations of liposome PC groups possibly caused by interactions with lectin [32,33]. These lipid group vibrations were due to the following: antisymmetric stretching of choline groups (as N+ (CH3 )3 ); antisymmetric stretching of phosphate groups (as PO2 − ); stretching vibrations of esters ( C O C) and carbonyls ( C O); and symmetric and antisymmetric stretching vibrations of methylenes (s CH2 and as CH2 , respectively).

2.9. NMR assays NMR spin-lattice relaxation times (T1 ) of liposomal PC choline (at 3.2 ppm) were obtained at 60 MHz using an Anasazi Instrument equipment (Indianapolis, USA). Inversion recovery pulse sequences (180◦ –␶ 90◦ ) were executed at 20 ◦ C, with a time delay (t) ranging from 0.4 to 12.8 s. A water:deuterated water (80:20) solvent and TSP internal reference were used [34,35]. Exponential data were fitted to the NUTS code to obtain T1 values and relative intensities, according to Levy and Peat (1978). NMR T1 values were reported as the average of three independent measurements.

1H

2.10. DSC assays DSC measurements of PC + C22 PEG900 GlcNAc vesicles, pure or lectin-loaded, were obtained by a Shimadzu DSC-60 instrument (Tokyo, JP). The heating rate was set to 2.5◦ C/min from −60 to 60 ◦ C under nitrogen flow (10 mL min−1 ) [36]. An empty aluminum cell was used as reference [34]. The enthalpy variation ( H) was calculated by integrating the area under the DSC peak with TA 60WS software.

2.11. UV–vis spectroscopy measurements Turbidity assays were performed in a Shimadzu UV-2550 UV–vis spectrophotometer (Kyoto, JP) at 400 nm [37,38]. Turbidity values were expressed as averages from three independent measurements.

Fig. 1. Zeta potential profiles of PC + C22 PEG900 GlcNAc vesicles in the presence of Bauhinia variegata lectin BVL (1 mg/mL).

3. Results and discussion 3.1. Lectin effect on vesicle charge and size The zeta potential (ZP) was used to evaluate the surface charge of the PC + C22 PEG900 GlcNAc vesicles in the presence of lectin. The zeta potential of PC + C22 PEG900 GlcNAc vesicles in phosphate buffer saline was −47 ± 2 mV, indicating stable assemblies in suspension due to electrostatic repulsion between particles, resulting in relative colloidal stability, as previously reported by the authors [22]. Fig. 1 shows the zeta potential values with increasing amounts of lectin. As the lectin concentration increased, the zeta potential correspondingly increased toward zero, suggesting interactions between lectin molecules and the vesicles. PC membranes with negative surface zeta potential values indicated that choline molecules were positioned under the phosphate groups. As the zeta potential increased, choline was observed to locate closer to the phosphate plane [39]. The 80% increase in the zeta potential as induced by lectin in the PC + C22 PEG900 GlcNAc vesicles indicated a protein-induced change in choline orientation toward the phosphate plane of the lipids. The DLS technique was used to obtain the hydrodynamic dimension (RH ) of the PC + C22 PEG900 GlcNAc vesicles in the presence of lectin. Fig. 2A shows a typical autocorrelation function C(q,t) measured at 90◦ for PC + C22 PEG900 GlcNAc in the absence and the presence of lectin. The characteristic decay time related to the diffusion of the assemblies shifts towards the right-hand side in the presence of lectin. It accordingly means increase in the size of the supramolecular aggregates. The size distributions weighted by number are given in Fig. 2B. The dimension of the lectin-free assemblies is about ∼ 100 nm whereas in the presence of the protein the dimension shifts to about ∼ 130 nm. These experimental results suggest that the glycosylated surfaces of the vesicles were capable of recognizing lectins due primarily to the increased diameter. 3.2. Lectin effect on vesicle structure The morphology of the self-assembled objects was probed via SAXS analysis. The profiles of the PC + C22 PEG900 GlcNAc vesicles in the absence and presence of lectin are shown in Fig. 3. The vesicular morphology was confirmed by the SAXS data. The scattering patterns shown in Fig. 3 could be fitted using a bilayer vesicle form factor implemented in the SASfit software. The adjustable parameters included the radius of the inner compartment Rc , the

M.C. dos Santos et al. / Colloids and Surfaces B: Biointerfaces 148 (2016) 12–18

15

thickness of the hydrophilic outer section th , the thickness of hydrophobic inner segment tt of the bilayer containing the tail groups and respective electron densities. The electron density of water (0.33 e− /Å3 ) was kept fixed during the fitting procedures. The hydrophilic and hydrophobic regions were varied to allow for possible swelling effects. Based on the determined values of Rc , tt and th of 45.0, 3.6 and 0.5 nm, respectively, the vesicle wall thickness (2th + tt ) was 4.6 nm and the overall radius of the vesicle was 49.6 nm. In the presence of lectin, superficial changes were observed in the SAXS profile. Indeed, a shift of the characteristic shoulder toward the right was recorded. The fitting procedure using the bilayer form factor indicated a slight increase in th (from 0.5 to 0.8 nm), which suggested changes in the vesicular structure. The use of the bilayer form factor was not ideal because protein adsorption may have only occurred at the outer hydrophilic surface of the vesicles, i.e., in protein environment, th(inner) < th(outer) . On the one hand, these results suggested that the thickness of the hydrophilic layer was indeed influenced by lectin. On the other hand, the values of tt and Rc were essentially the same. Furthermore, the fitting procedure in the presence of lectin required a higher degree of polydispersity, as indicated by the width parameter of the log-normal distribution of size; this suggested more irregular structures. 3.3. Morphological study by atomic force microscopy (AFM)

I (q) a.u

Fig. 2. A) Typical autocorrelation function C(q,t) measured at 90 ◦ C for PC + C22 PEG900 GlcNAc in the absence (䊉) and the presence () of lectin. B) Distribution of 2RH by considering the contribution of the particles related to their total number for PC + C22 PEG900 GlcNAc in the absence (䊉) and the presence () of BVL lectin (4 ␮L, 1.0 mg/mL).

10

8

10

7

10

6

10

5

10

4

A number of techniques are available to characterize membrane domain topology, system molecular dynamics and morphology, such as fluorescence microscopy, dynamic light scattering, Xray diffraction, and electron microscopy. Despite this variety, only recently have the nanometer-level structures and properties of lipid films been determined. In this context, atomic force microscopy (AFM) is an essential tool in lipid film research. Fig. 4 presents a topographical image of the PC + C22 PEG900 GlcNAc vesicles in the absence and the presence of lectin bilayers supported on mica. Vesicles are shown to be homogenous and have an average size of 100 nm in the absence of lectin (Fig. 4A). However, as shown in (Fig. 4B), the specific interaction with lectin confirmed the bioavailability of carbohydrate residues on the surfaces of the vesicles because of an a significant increase the size of the vesicular aggregates. This result together with the techniques of ZP measurements, DLS and SAXS corroborate to the presence of GlcNAc residues on surfaces combination of the self-assemblies and its ability to interact specifically with lectin-specific carbohydrate. This experiment demonstrated that BVL specifically interacted with NAcetyl-glycosamine particles, resulting in large aggregates. These results support the assumption that the carbohydrates are exposed at the surface of the vesicle and can provide specific receptors in targeted nanocarrier systems [40–44]. 3.4. Lectin location and influence on PC + C22 PEG900 GlcNAc vesicles intermolecular dynamics

0.1

1 -1

q (nm ) Fig. 3. SAXS analysis of PC + C22 PEG900 GlcNAc vesicles in the absence (䊉) and the presence ( ) of BVL lectin.

In a previous work, [22] the influence of polymer type on PC vesicle dynamics was studied using FTIR. To investigate the lectin location and resultant effects on the system molecular dynamics, HATR-FTIR, NMR, DSC and UV–vis spectroscopy assays were used. Fig. 5 shows the HATR-FTIR spectra of (i) PC + C22 PEG900 GlcNAc and (ii) PC + C22 PEG900 GlcNAc vesicles in the presence of lectin (PC + C22 PEG900 GIcNAc + lectin). The FTIR frequency and bandwidths of the lipid peaks were observed to shift after the proteins were introduced (Table 1 and Fig. S1). In the PC + C22 PEG900 GlcNAc vesicle system, lectin affected the frequency values of the FTIR peaks of the choline, ester and acyl methylene groups. The variations in the frequency of the FTIR choline stretching vibrations indicated a change in the interactions between choline lipid groups and phosphate groups on neighbor-

16

M.C. dos Santos et al. / Colloids and Surfaces B: Biointerfaces 148 (2016) 12–18

Fig. 4. AFM images showing the surface morphology of the solid films of PC + C22 PEG900 GlcNAc vesicles in the absence (A) and in the presence (B) of BVL bilayers supported on mica. Table 1 Lectin influence on the PC + C22 PEG900 GlcNAc vesicle FTIR band frequency and bandwidth variations, in cm−1 . Functional group

+

as N (CH3 )3 C O C C O as CH2

PC + C22 PEG900 GlcNAc

PC + C22 PEG900 GlcNAc + lectin

Frequency (cm−1 )

Bandwidth (cm−1 )

Frequency (cm−1 )

Bandwidth (cm−1 )

977.99 1083.99 1734.07 2924.20

11.0 40.0 8.37 28.00

983.70 1078.21 1734.07 2926.77

7.40 23.70 4.00 28.00

Fig. 5. Influence of lectin presence on the recovery of PC + C22 PEG900 GlcNAc vesicles 1 H FID signal after several inversion pulses. The open circles represent the PC + C22 PEG900 GlcNAc vesicles; the open inverted triangles represent the protein interaction in PC + C22 PEG900 GlcNAc vesicles. 1 H T1 was calculated from these curves.

ing lipids or water molecules [33]. Lectin insertion in the vesicles increased the FTIR as N+ (CH3 )3 frequency value by 6.70 cm−1 . A narrowing effect of approximately 30% was observed in the FTIR lipid as N+ (CH3 )3 bandwidth after interacting with lectin. To better understand the lectin-choline interaction, NMR 1 H spin-lattice relaxation times (T1 ) were performed. The recovery of choline 1 H FID signals (␦ = 3.2 ppm) is shown in Fig. 5. From this recovery, the T1 values were calculated (Table S1). A discrete lectin-induced lipid choline ordering effect was observed in the PC + C22 PEG900 GlcNAc vesicles. The variation of PC choline 1 H T1 values after the lectin insertion corresponded

to approximately 12%. Thus, lectin influenced the intermolecular interactions and motional degree of lipid choline. In the PC/polymer system, the membrane lipid dynamic was likely affected by the lectin recognition to the polymer sugar residue (in this case, the N-acetyl-glucosamine content of C22 PEG900 GIcNAc). Therefore, a comparison between lectin and glycoprotein behavior and effects on these membranes would be of interest. For example, surfactantprotein A (SP-A) is a hydrophilic glycoprotein that contributes to pulmonary surfactant surface activity and possibly host defense. SP-A binds non-phosphatidylglycerol PC membranes [45]. Because the lipid structures of these two types of membrane differ by the presence of a choline group, this highlights the importance of the interaction between lipid choline and lectin carbohydrate recognition domains, as evidenced by the SAXS analyses and zeta potential measurements. The incorporation of charged amphiphiles into a bilayer may affect the torque applied to the electric dipole of the PC headgroup but not necessarily change the bilayer packing structure [46,47]. Morrow et al. [48] observed that the addition of a negative surface charge may tilt the PC headgroup toward the membrane surface due to the attraction of the choline group. This explanation justifies the FTIR and NMR results because the presence of negative partial charges, as detected in the lectin and polymer carbonyl and hydroxyl groups, can change the lipid choline dipole–dipole interactions by affecting the choline orientation. The lectin-induced change in the lipid choline orientation was also confirmed by the zeta potential results. No lectin-induced changes were observed in the lipid phosphate vibration (as PO2 − ). The frequency and bandwidth variations in the as PO2 − reflect changes in its hydration degree and mobility, respectively. These parameters were not influenced by lectin. In the liposomal interfacial region, the lectin influence on the lipid interface was significant, as shown by 5.78 and 18.85 cm−1 downward shifts in the FTIR  C O C frequency and bandwidth

M.C. dos Santos et al. / Colloids and Surfaces B: Biointerfaces 148 (2016) 12–18

17

Table 2 Turbidity values (400 nm) of PC + C22 PEG900 GlcNAc vesicles in the absence and the presence of lectin.

Fig. 6. DSC curves of PC + C22 PEG900 GlcNAc vesicles in the absence and the presence of lectin. In each DSC experiment, the heating rate was set to 10 ◦ C and the temperature ranged from −45 to 5 ◦ C under 10 mL min−1 nitrogen flow.

values, respectively (Table 1 and Fig. S1). The  C O bandwidth narrowed by approximately 52% after interaction with the protein in the PC-polymer liposomes. Thus, lectin seemed to promote an ordering effect on the PC-polymer liposome interface region more than on the lipid headgroup. The discrete ordering effect induced by BVL has also been reported for wheat germ agglutinin. These carbonyl chemical shift anisotropy studies, obtained by 13 C NMR, suggested that a complex between wheat germ agglutinin lectin and N-acetyl glicosamine slowed the dimyristoylphosphatidylcholine headgroup motion [49]. In the hydrophobic region of the lipid, an analysis of the lipid FTIR as CH2 peak showed that lectin induced a 2.63 cm−1 increase in the lipid peak frequency (Table 1 and Fig. S1). The variations of the methylene stretching band frequencies reflect the changes in lipid order and packing. This can be sensitive to an increase in the intermolecular space between lipid fatty acids, caused by the insertion of other substances [33,50]. Furthermore, an increase in the s CH2 peak frequencies suggests an increase of gauche bonds in the lipid hydrophobic region [51]. Thus, the shift observed in the s CH2 frequency value induced by lectin indicated that the protein fluidized the hydrophobic regions of the PC + C22 PEG900 GlcNAc vesicles. DSC measurements were also performed to further determine the role of lectin in the lipid hydrophobic region. DSC endothermic curves are shown in Fig. 6. The phase transition temperatures (Tm) and enthalpy changes ( H) calculated from DSC curves are listed in Table S2. Lectin provoked a reduction of 0.15 ◦ C in the Tm of the PC + C22 PEG900 GlcNAc vesicles, indicating a very discrete interaction between the protein and the hydrophobic regions of the membrane and a lectin-induced molecular rearrangement of lipids to a more disordered state. These results agreed with those observed for lipid acyl methylenes measured by FTIR. A lower temperature was required to reach the phase transition, as observed in the increased number of acyl chain gauche conformations [52]. A very discrete variation in lipid H values provoked by lectin was observed. Higher changes in H are attributed to locations within the membrane hydrophobic core and not in the surrounding polar and interfacial regions [52]. Thus, the discrete decreases in the lipid Tm and H parameters indicate that lectin may be located near the lipid choline and carbonyl regions and not in the hydrophobic core. However, lectin may diminish the intermolecular interactions of the lipid acyl chain methylene groups as indicated by variations in s CH2 .

Sample

Turbidity (Optical density at 400 nm)

PC + C22 PEG900 GlcNAc PC + C22 PEG900 GlcNAc + lectin

0.27 0.29

The discrete lectin membrane hydrophobic disordering effect, observed in FTIR and DSC assays, was also observed in the turbidity results as shown in Table 2. With the addition of lectin to the PC + C22 PEG900 GlcNAc vesicles, a 7% reduction in the turbidity values was observed. A similar discrete influence on membrane turbidity has also been reported for mellitin, a bee venom with lytic effects in unsaturated phosphatidylcholines [53]. Thus, lectin seemed to be located between the lipid choline and carbonyl groups. Lectin seemed to induce a structural order to these groups by affecting the choline orientation. Lectin also slightly decreased and disrupted the intermolecular interactions of hydrophobic acyl chains. 4. Conclusion We investigated the interaction between vesicles formed by phosphatidylcholine (PC) and the glycosylated polymeric amphiphile N-acetyl-ˇ-d-glucosaminyl-PEG900 -docosanate conjugate (C22 PEG900 GlcNAc) with the lectin BVL. Our results suggested that the lectin molecules recognized the glycosylated surfaces of the vesicles. Furthermore, the lectin molecules were located between the lipid choline and carbonyl groups, thereby changing the choline orientation and ordering the lipid polar head and interface regions. The presence of carbohydrate residues on the surfaces of the vesicles and their ability to establish specific interactions with BVL were further highlighted by light-scattering measurements. These results indicate the potential applications of such sugar-based vesicles in site-specific vectorization systems for drug delivery, diagnoses and biosensors. Conflicts of interest The authors declare no conflicts of interest. Acknowledgments The authors wish to thank the Conselho Nacional de Pesquisa e Desenvolvimento (CNPq– Brazil) (440619/2014-9) for financial support. The LNLS is acknowledged for supplying the SAXS beam time (proposals 20150020 and 20150021). A.G.D.B acknowledges the financial support from FAPESC 3805/2012, and F.C.G. thanks FAPESP (grant 2014/22983-9). Appendix A. Supplementary data Supplementary data associated with this article can be found, in the online version, at http://dx.doi.org/10.1016/j.colsurfb.2016.08. 053. References [1] T. Fritz, M. Hirsch, F.C. Richter, S.S. Müller, A.M. Hofmann, K.A.K. Rusitzka, J. Markl, U. Massing, H. Frey, M. Helm, Biomacromolecules 15 (2014) 2440. [2] R. Ghosh, J. Dey, Langmuir 30 (2014) 13516. [3] V.P. Torchilin, Nat. Rev. Drug Discov. 4 (2005) 145. [4] J.H. Collier, P.B. Messersmith, Annu. Rev. Mater. Res. 31 (2001) 237. [5] M. Umrethia, P.K. Ghosh, R. Majithya, R.S.R. Murthy, Cancer Invest. 25 (2007) 117. [6] J.M. Koziara, P.R. Lockman, D.D. Allen, R.J. Mumper, J. Nanosci. Nanotechnol. 6 (2006) 2712.

18

M.C. dos Santos et al. / Colloids and Surfaces B: Biointerfaces 148 (2016) 12–18

[7] E. Fattal, P. Couvreur, C. Dubernet, Adv. Drug Deliv. Rev. 56 (2004) 931. [8] M.H. Costa, O.A. Sant’Anna, P.S. de Araujo, R.A. Sato, W. Quintilio, L.V. Silva, C.R. Matos, I. Raw, Appl. Biochem. Biotechnol. 73 (1998) 19. [9] G. Gregoriadis, Trends Biotechnol. 13 (1995) 527. [10] D.D. Lasic, D. Papahadjopoulos, Science 267 (1995) 1275. [11] D.D. Lasic, Trends Biotechnol. 16 (1998) 307. [12] A. Mauceri, S. Borocci, L. Galantini, L. Giansanti, G. Mancini, A. Martino, L.S. Manni, C. Sperdut, Langmuir 30 (2014) 1130. [13] H. Lis, N. Sharon, Chem. Rev. 98 (1998) 637. [14] X. Gao, W. Tao, W. Lu, Q. Zhang, Y. Zhang, X. Jiang, S. Fu, Biomaterials 27 (2006) 3482. [15] L.S. Pinto, C.S. Nagano, T.M. Oliveira, T.R. Moura, A.H. Sampaio, H. Debray, V.P. Pinto, O.A. Dellagostin, B.S. Cavada, J. Biosci. 33 (2008) 355. [16] A.G. Dal Bó, V. Soldi, F.C. Giacomelli, C. Travelet, R. Borsali, S. Fort, Carbohydr. Res. 397 (2014) 31. [17] A.G. Dal Bó, V. Soldi, F.C. Giacomelli, C. Travelet, B. Jean, I. Pignot-Paintrand, R. Borsali, S. Fort, Langmuir 28 (2011) 1418. [18] I. Ofek, J. Goldhar, Y. Keisari, N. Sharon, Annu. Rev. Microbiol. 49 (18) (1995) 239. [19] M. Hill, D. Mazal, V.A. Biron, L. Pereira, L. Ubillos, E. Berriel, H. Ahmed, T. Freire, M. Rondán, G.R. Vasta, F. Liu, M.M. Iglesias, E. Osinaga, J. Histochem. Cytochem. 58 (2010) 553. [20] S. Nakahara, A. Raz, Anticancer Agents Med. Chem. 8 (2008) 22. [21] A. Varki, Glycobiology 3 (1993) 97. [22] Y.M.S. Micheletto, N.P. da Silveira, D.M. Barboza, M.C. dos Santos, V.R. de Lima, F.C. Giacomelli, J.C.V. Martinez, T.E.A. Frizon, A.G. Dal Bó, Colloids Surf. A Physicochem. Eng. Asp. 467 (2015) 166. [23] K. Tasaki, J. Am. Chem. Soc. 118 (1996) 8459. [24] S. Zalipsky, Bioconjug. Chem. 6 (1995) 150. [25] M.C. Woodle, D.D. Lasic, Biochim. Biophys. Acta 1113 (1992) 171. [26] O. Mertins, M. Sebben, P.H. Schneider, A.R. Pohlmann, N.P. da Silveira, Quím. Nova 31 (2008) 1856. [27] A.G. Dal Bó, V. Soldi, F.C. Giacomelli, B. Jean, I. Pignot-Paintrand, R. Borsali, S. Fort, Soft Matter 7 (2011) 3453. [28] P. da Silva Malheiros, Y.M.S. Micheletto, N.P. da Silveira, A. Brandelli, Food Res. Int. 43 (2010) 1198.

[29] F. Szoka, D. Papahadjopoulos, Proc. Natl. Acad. Sci. (1978) 4194. [30] J. Jakes, Collect. Czechoslov. Chem. Commun. 60 (1995) 1781. [31] M. Tammer, L. Horsburgh, A.P. Monkman, W. Brown, H.D. Burrows, Adv. Funct. Mater. 12 (2002) 447. [32] C. Chen, C.P. Tripp, Biochim. Biophys. Acta 1778 (2008) 2266. [33] M.M. Moreno, P. Garidel, M. Suwalsky, J. Howe, K. Brandenburg, Biochim. Biophys. Acta 1788 (2009) 1296. [34] V.R. De Lima, M.S. Caro, M.L. Munford, B. Desbat, E. Dufourc, A.A. Pasa, T.B. Creczynski-Pasa, J. Pineal Res. 49 (2010) 169. [35] G.W. Feigenson, S.I. Chan, J. Am. Chem. Soc. 96 (1974) 1312. [36] R. Koynova, M. Caffrey, Biochim. Biophys. Acta 1376 (1998) 91. [37] G.C. Kresheck, K. Kale, M.D. Vallone, J. Colloid Interface Sci. 73 (1980) 460. [38] R.S. de Sousa, A.O.M. Nogueira, V.G. Marques, R.M. Clementin, V.R. de Lima, Bioorg. Chem. 51 (2013) 8. [39] D.G. Fatouros, S.G. Antimisiaris, J. Colloid Interface Sci. 251 (2002) 271. [40] C. Zhang, H. Wang, G. Su, R. Li, X. Shen, S. Zhang, Q. Geng, F. Liu, I. Otsuka, T. Satoh, T. Kakuchi, Polym. Int. 61 (2012) 1158. [41] T. Isono, I. Otsuka, Y. Kondo, S. Halila, S. Fort, C. Rochas, T. Satoh, R. Borsali, T. Kakuchi, Macromolecules 46 (2013) 1461. [42] Y.C. Chiu, I. Otsuka, S. Halila, R. Borsali, W.C. Chen, Adv. Funct. Mater. 24 (2014) 4240. [43] A. Upadhyay, S. Karpagam, J. Photopolym. Sci. Technol. 28 (2015) 755. [44] I. Otsuka, T. Isono, C. Rochas, S. Halila, S. Fort, T. Satoh, T. Kakuchi, R. Borsali, ACS Macro Lett. 1 (2012) 1379. [45] F.X. McCormack, Biochim. Biophys. Acta 1408 (1998) 109. [46] J. Seelig, P.M. MacDonald, P.G. Scherer, Biochemistry 26 (1987) 7535. [47] P.G. Scherer, J. Seelig, Biochemistry 28 (1989) 7720. [48] M.R. Morrow, N. Abu-Libdeh, J. Stewart, K.M.W. Keough, Biophys. J. 85 (2003) 2397. [49] B.J. Hare, F. Rise, Y. Aubin, J.H. Prestegard, Biochemistry 33 (1994) 10137. [50] H.H. Mantsch, R.N. McElhaney, Chem. Phys. Lipids 57 (1991) 213. [51] D.C. Lee, D. Chapman, Biosci. Rep. 6 (1986) 235. [52] L. Zhao, S.S. Feng, N. Kocherginsky, I. Kostetski, Int. J. Pharm. 338 (53) (2007) 258. [53] P. Wessman, A.A. Strömstedt, M. Malmsten, K. Edwards, Biophys. J. 95 (2008) 4324.