Covalent functionalization of graphene oxide with D-mannose: evaluating the hemolytic effect and protein corona formation

Marcelo de Sousa *a, Carlos H. Z. Martins a, Lidiane S. Franqui bc, Leandro C. Fonseca a, Fabrício S. Delite c, Evandro M. Lanzoni b, Diego Stéfani T. Martinez *abc and Oswaldo L. Alves *a
aLaboratory of Solid State Chemistry, Institute of Chemistry, University of Campinas, Campinas, São Paulo, 13083-970, Brazil. E-mail: oalves@iqm.unicamp.br; marcelosousap2@yahoo.com.br
bBrazilian Nanotechnology National Laboratory (LNNano), Brazilian Center for Research in Energy and Materials (CNPEM), Campinas, São Paulo 13084-970, Brazil. E-mail: diego.martinez@lnnano.cnpem.br
cSchool of Technology, University of Campinas, Limeira, São Paulo 13484-332, Brazil

Received 16th November 2017 , Accepted 8th April 2018

First published on 10th April 2018


In this work, graphene oxide (GO) was covalently functionalized with D-mannose (man-GO) using mannosylated ethylenediamine. XPS (C1s and N1s) confirmed the functionalization of GO through the binding energies at 288.2 eV and 399.8 eV, respectively, which are attributed to the amide bond. ATR-FTIR spectroscopy showed an increase in the amine bond intensity, at 1625 cm−1 (stretching C[double bond, length as m-dash]O), after the functionalization step. Furthermore, the man-GO toxicity to human red blood cells (hemolysis) and its nanobiointeractions with human plasma proteins (hard corona formation) were evaluated. The mannosylation of GO drastically reduced its toxicity to red blood cells. SDS-PAGE analysis showed that the mannosylation process of GO also drastically reduced the amount of the proteins in the hard corona. Additionally, proteomics analysis by LC–MS/MS revealed 109 proteins in the composition of the man-GO hard corona. Finally, this work contributes to future biomedical applications of graphene-based materials functionalized with active biomolecules.


Introduction

In recent years, the interest in using graphene oxide (GO) in biological and biomedical applications has increased in the scientific community.1 GO is a two-dimensional material with a large surface area comprising single-layer sheets of sp2-hybridized carbon atoms and sp3 carbon sites where there are oxygen hydrophilic functional groups.2 Thus, GO has excellent dispersibility in water, amphiphilicity and many possibilities for surface functionalization.3 Covalent functionalization is possible because, during the exfoliation and oxidation processes of graphite, the formation of a large amount of carboxylic groups occurs. These groups can be modified with coupling reagents, such as 1-ethyl-3-(3-dimethylaminopropyl) carbodiimide (EDC) and N-hydroxysuccinimide (NHS),4 or converted to acyl chlorides using thionyl chloride (SOCl2).5 The GO functionalization methodology depends on the conditions of the reaction, solvent, temperature, functional groups of the binding molecules, and reaction time among other factors. The high non-covalent and covalent functionalization capacities make GO a unique nanomaterial to develop studies in biological systems. Recent studies have shown excellent results when functionalized GO was used as a targeted drug delivery system.6 This is possible because a large proportion of the amount of hydrophobic drugs can be adsorbed on surface of GO by non-covalent interactions and are slowly released to specific organism regions according to the specificity of the covalent functionalization. Among other biological applications related to the functionalization of the GO are bioimaging,7 photodynamic therapy,8 and antibacterial activity.9

Several targeting studies have shown interesting results with mannosylated nanomaterials.10 The interest in functionalizing nanomaterials with mannose occurs because many cells and tissues, such as macrophages and dendritic cells, over-express mannose receptors.11,12 These studies indicate that mannosylated nanomaterials present higher affinity to these cells than unmodified nanomaterials. Generally, the mannosylated nanosystems are captured by macrophages and are targeted to specific regions of the organism, such as the spleen, liver, lymph nodes, bone marrow and gut. Specifically, there are few reports of functionalized GO with mannose. In one study, GO was first functionalized with polyethyleneimine, to express free amino groups, followed by coupling with mannose by oxidative amidation.13 GO was also functionalized with 4-[(trimethylsilyl)ethynyl]aniline to form graphene-phenylacetylene, followed by the reaction with α-D-mannosyldendrons, a glycodendron with three mannose moieties.14 In this sense, new approaches are necessary to develop alternative GO functionalization methods with sugars and to verify their effects on biological systems.

Understanding and controlling the interaction of GO materials with complex biological systems are mandatory for the medical applications and safety evaluation of these materials.15 In this sense, different biological models have been used to evaluate the nano-biointeractions with GO-based materials.16 However, in most biological applications, the blood-circulatory system is the first bio-interface that interacts with nanomaterials, and human red blood cells (RBCs) are good models to probe the membranolytic effect of nanostructured materials.17 Accordingly, the hemolytic assay is considered an effective method to assess the toxicity and biocompatibility of GO materials. However, there are limited reports in the literature describing the interaction of GO with RBCs, and the results of these investigations are usually contradictory.18 For example, it was demonstrated that GO materials can disrupt the RBC membrane (hemolysis) in a dose-dependent manner.19 Other reports have indicated a size-dependent effect of GO on hemolytic activity.20 The superior hemolytic effect of pristine GO was observed compared with that of functionalized GO materials.19 Some studies did not observe the hemolytic effect of the GO materials investigated.21 These observations show that the hemolytic effect of GO is strongly modulated by the source of material, preparation method, physicochemical properties and functionalizations applied.16

Another key point regarding the biomedical application of graphene-based materials is a deep understanding of its interactions with biological media (e.g., human plasma).22 Once intravenously administered nanomaterials are exposed to a range of plasma proteins present in the bloodstream that will bind to the nanomaterial surface to form a biomolecular coating named the protein corona (PC). PC formation confers a new “biological identity” to the nanomaterial and dictates all the subsequent biological and toxicological responses in biological systems.23 Serum protein-coated GO materials greatly mitigate GO toxicity by PC formation and markedly reduce the cytotoxicity toward carcinoma cells by reducing the physical damage and interaction of GO with cell membranes.24 The biocompatibility of GO in RBCs showed that PC formation reduces hemolytic activity after its interaction with human plasma proteins.25

Recently, the development of advanced mass spectrometry technologies allowed the identification and quantification of the major proteins in human blood plasma (proteomics).26,27 It was possible to investigate the in vitro and in vivo biocompatibility of different surface-modified GO nanomaterials. It was shown that the biocompatibility of the surface-modified GO was due to the different compositions of PC, especially immunoglobulin G, which determined the adsorption onto the membrane and subsequent uptake by macrophages. These results indicate that deciphering the composition and structure of PC is crucial to understanding and improving the biodistribution and biocompatibility of GO materials towards safer biomedical applications.

Considering these observations, our goal was to functionalize GO with mannose using mannosylated ethylenediamine28 and perform an integrated physicochemical characterization. Additionally, we performed a comparative study of pristine GO and GO functionalized with mannose (man-GO) on the interactions of these materials with RBCs and human plasma proteins towards future biomedical applications (i.e., drug delivery) and safety issues (toxicity assessment). Finally, advanced mass spectrometry analysis (LC–MS/MS) was applied to identify the proteins comprising the PC associated with man-GO material.

Experimental section

Materials

The following materials were used for the experiments: natural graphite powder (98%) (Synth, Brazil), potassium permanganate (KMnO4, 99.0%) (Synth, Brazil), sulfuric acid (H2SO4, 95.0–98.0%) (Synth, Brazil), chloride acid (36.5–38.0%) (Synth, Brazil), hydrogen peroxide (30.0%) (Synth, Brazil), acetone (99.5%) (Synth, Brazil), sodium hydroxide (NaOH, >99.0%, pellets) (Synth, Brazil), ethanol (Neon, Brazil), phosphorous pentoxide (P2O5) (Sigma Aldrich, India), potassium persulfate (K2S2O8, 99.0%) (Sigma Aldrich, Japan), D-mannose (Sigma Aldrich, India), N-(3-dimethylaminopropyl)-N-ethylcarbodiimide (EDC) (Sigma Aldrich, Japan), N-hydroxysuccinimide (NHS) (Sigma Aldrich, Japan), iodine (Prods Labor Chemicals LTDA, Brazil), nitric acid (Synth, Brazil), dimethylformamide DMF (Synth, Brazil), ethylenediamine (Sigma-Aldrich, Belgium) and NaCl (Sigma Aldrich, Darmstadt, Germany).

Preparation and covalent functionalization of GO with mannosylated ethylenediamine

GO synthesis was performed using the modified Hummers’ method.29 Briefly, mannosylated ethylenediamine was synthesized according to Sousa et al.28 GO was completely dispersed in DMF (1 mg mL−1) using an ultrasound bath. EDC and NHS crosslinking reagents were added at a concentration of 4 mg mL−1 each, following the addition of the mannosylated ethylenediamine at a concentration of 6 mg mL−1, and were allowed to stand for 24 h. After the reaction, man-GO was washed several times with deionized water using a pump vacuum until neutral pH and was dialyzed for 3 days. One portion of this dispersion was dried and subjected to physico-chemistry characterizations, and another was stored in the refrigerator for the development of biological studies. Fig. 1 illustrates the functionalization steps performed in this work.
image file: c7tb02997g-f1.tif
Fig. 1 Functionalization of GO with mannosylated ethylenediamine using EDC and NHS crosslinking reagents.

GO and man-GO were characterized using the following techniques: (i) X-ray photoelectron spectroscopy (XPS) using the K-alpha X-ray model from Thermo Scientific. This instrument employs an Al Kα X-ray source with an Ar+ ion beam. The pressure in the analysis chamber was set to 1.6 × 10−9 mbar. Avantage software was used for data acquisition and analysis. The binding energy of the C1s of graphite, 284.5 eV, was used as the reference. (ii) Thermogravimetric analysis (TGA): 3.5–5 mg of each dried sample of GO and man-GO was weighed and analyzed in a flux of nitrogen at a heating rate of 10 °C min−1 between 30 °C and 1000 °C using the TA Instruments SDTQ600 model; (iii) FTIR spectra were obtained by attenuated total reflectance (ATR) in the range of 400 to 4000 cm−1, with 4 cm−1 resolution and a total of 32 scans using an Agilent Technologies Cary 630 spectrometer. (iv) Raman spectra were obtained using a HeNe laser with a slit width of 100 μm, a spectral resolution of 2 cm−1, a 100× objective, a potency of 0.5 mW, 3 scans and 60 second accumulation on a Raman T64000 spectrometer (Horiba). The calibration was performed using a silica standard; (v) GO and man-GO were observed by TEM using a Zeiss LIBRA 120 microscope coupled to an Omega Filter-spectrometer with an accelerating voltage of 80 kV. (vi) Atomic force microscope (AFM) images were acquired using a Park Systems NX10 instrument in tapping mode, for the samples deposited onto silicon wafers from deionized water dispersions (10 μg mL−1); (vii) the zeta potential was used to analyze the surface charge of GO and man-GO using 100 μg mL−1 of both samples in deionized water. The measurements were carried out using a Malvern Instrument model Zetasizer Nanoseries. The final pellets of man-GO were resuspended in deionized water (1 mL) by sonication for 3 min in an ultrasound bath prior to the measurements. The assays were performed in triplicate.

Hemolytic assays

Human RBCs concentrated (type O+) from the transfusion service (certified plastic bags) were provided by the Hemocenter from the School of Medicine, University of Campinas (UNICAMP), in Campinas, São Paulo, Brazil. First, a stock-dispersion (0.5 mg mL−1) of GO and man-GO was prepared in deionized water using an ultrasound bath (Cole-Parmer 8891) for 60 min. Before incubation with GO and man-GO samples, the RBCs were washed 3 times with NaCl 0.15 mol L−1 solution and were centrifuged at 10[thin space (1/6-em)]000 rpm for 10 min at 4 °C (Eppendorf, 5810R model). A stock-suspension of RBCs (10%, 20 mL) in NaCl 0.15 mol L−1 was prepared to be used in all hemolytic assays. The GO and man-GO samples were pre-incubated at concentrations ranging from 10 to 250 μg mL−1 with NaCl solution in sterile Eppendorf microtubes (1.5 mL) at 23 ± 1.0 °C. After 5.0 min, the previously washed RBCs (100 μL of the RBC stock-suspension) were added, and the microtubes were incubated for 60 min at room temperature after gentle homogenization. The final volume of the hemolytic assay in all experiments was 1 mL, and the assay was performed in independent triplicates. After the GO material interaction with the RBCs, the microtubes were centrifuged at 10[thin space (1/6-em)]000 rpm and 4 °C for 10 min, and aliquots of the supernatant (100 μL) were carefully removed from each microtube and transferred to a clean 96-well plate. The quantification of hemoglobin in the supernatant was performed by detecting the solution absorbance at 540 nm with a microplate reader (Thermo, Multiskan GO model). The positive control consisted of deionized water (900 μL) and RBC stock suspension (100 μL). The negative control consisted of a 0.15 mol L−1 NaCl solution (900 μL) and the RBC stock suspension (100 μL). The percentage of hemolysis was calculated using the linear equation y = mx + c, where the percentage of hemolysis (x) = [optical density (y) − negative control optical density (c)]/[(positive control optical density − negative control optical density)/100].

Protein corona characterization

Human blood plasma (type O+) preserved in a small plastic transfusion bag was provided by the Hemocenter of the School of Medicine, University of Campinas (UNICAMP), Campinas, São Paulo, Brazil. The assay was performed using a pool of plasma samples (n = 6). Stock-dispersion of the GO and man-GO samples (100 μg mL−1) were incubated with 55% plasma in 0.15 mol L−1 NaCl (1.5 mL Protein LoBind Eppendorf microtubes) for 60 min under a horizontal mixer (14 rpm) at 37 °C (Thermoblock Eppendorf). After the interaction of the GO samples with plasma, the microtubes were centrifuged for 60 min at 14[thin space (1/6-em)]000 rpm and 4 °C (Eppendorf centrifuge, 5810R model) and were washed three times with NaCl solution to remove the weakly bound plasma proteins from the nanomaterial surface (soft corona). After washing, the final pellets comprised GO and man-GO with hard coronas on the surface. The hard corona proteins were extracted from the nanomaterial surface using loading buffer (New England BioLabs, Ipswich, USA), and 20 μL of the protein solution was resolved by 15% SDS polyacrylamide gel electrophoresis (SDS-PAGE). The protein band contrast in the gels were enhanced by Coomassie blue staining. All experiments were performed in independent triplicates. For LC/MS-MS analysis, the protein spots were excised from the SDS-PAGE gel and were subjected to a procedure for protein digestion and peptide extraction according to Shevchenko.30 The extracts were purified on StageTips C18 and were analyzed using a mass spectrometer LTQ OrbitrapVelos (Thermo Fisher Scientific, Waltham, MA, USA) connected to a nanoflow LC (LC–MS/MS) instrument using an EASY-NLC system (Proxeon Biosystems, West Palm Beach, FL, USA).

Dispersion stability monitoring

The colloidal stability of GO and man-GO was evaluated by monitoring their sedimentation behavior throughout UV-Vis spectroscopy analysis. The GO materials were dispersed in deionized water, NaCl 0.15 mol L−1 solution and 55% human plasma in NaCl 0.15 mol L−1 solution, at a final concentration of 100 μg mL−1, and the supernatant absorbance was measured at 400 nm at 0, 1, 3, 6, 9 and 24 h.

Results and discussion

Structural and morphological characterization of pristine GO and man-GO

The coupling of the mannosylated ethylenediamine to GO was confirmed by XPS and ATR-FTIR spectroscopy. The XPS measurements identified the elemental composition and types of chemical bonds that constitute the sample providing the functionalization. The deconvolutions and attributions were based on results obtained from the literature.31–36 The survey and high-resolution XPS observations are shown in Fig. 2A–E.
image file: c7tb02997g-f2.tif
Fig. 2 XPS analysis: (A) survey spectra for GO and man-GO, (B) C1s spectrum of GO, (C) C1s spectrum of man-GO, (D) O1s spectrum of GO, (E) O1s spectrum of man-GO and (F) N1s spectrum of man-GO.

In Fig. 2A the survey observations show the presence of carbon, oxygen and nitrogen in man-GO but only carbon and oxygen in GO. The presence of nitrogen in man-GO is evidence of the covalent coupling of mannosylated ethylenediamine. Fig. 2B shows the C1s XPS spectrum for GO, which was deconvoluted into four components corresponding to carbon atoms in different chemical environments. The binding energies were 284.8(±0.2), 285.3(±0.1), 287.3(±0.1) and 289.0(±0.1) eV assigned to C[double bond, length as m-dash]C, C–C, C–O–C and C[double bond, length as m-dash]O bonds, respectively.32–36 Additionally, the spectrum shows very intense ether/epoxy (C–O–C) and sp3 carbon (C–C) binding energies, confirming that GO was highly oxidized. Furthermore, this material presented high dispersibility in water and a significant presence of carboxyl groups on its surface, which are essential to functionalization studies. On the other hand, the C1s XPS spectrum for man-GO was deconvoluted into five signals (Fig. 2C). In this case, the binding energies of the C[double bond, length as m-dash]C and C–C bonds were observed at 284.9(±0.1) eV, in agreement with the literature results.36,37 However, the intensity of the binding energy of the C[double bond, length as m-dash]C and C–C bonds increased while the intensity of the binding energy in 287.0 eV (oxygenated groups) reduced, indicating the occurrence of the reduction of GO after functionalization. This comportment was observed when GO was treated with diamines38 and hydrazine.39 Some literature references show that the C–N binding energy can be observed between 285.5 and 285.9 eV28,33,37,40 and that the aliphatic COH binding energy can be observed between 286.0 and 286.4.28,31,32,41 In the XPS spectrum, the binding energies of these bonds overlapped and were observed at 286.0(±0.3) eV. The binding energy at 288.2(±0.3) eV was attributed to the O[double bond, length as m-dash]C–N bonds,37,42 and those at 287.0(±0.1) and 289.2(±0.4) eV were attributed to C–O–C and O–C[double bond, length as m-dash]O bonds, respectively.37,42 The assignment of binding energies for the O1s spectrum confirmed the information obtained for the C1s XPS spectrum. Fig. 2D shows the deconvoluted O1s XPS spectrum of GO, where the binding energies at 532.1(±0.3), 533.0(±0.1) and 534.4(±0.4) eV were attributed to the C[double bond, length as m-dash]O (carbonyl), C–O (epoxy and ether) and O–C[double bond, length as m-dash]O (carboxyl) bonds, respectively. The O1s binding energies for man-GO in Fig. 2E were 530.9(±0.2), 531.9(±0.3), 532.8(±0.2) and 533.9(±0.2) eV and were assigned to quinone, C[double bond, length as m-dash]O (carboxyl), C–O (ether/epoxy/COH aliphatic) and O–C[double bond, length as m-dash]O bonds, respectively.31,33,35,36 The binding energy of quinone agrees with the literature results, and the aliphatic COH binding energy is very close in energy to the CO ether/epoxy binding energy, and it was not possible to separate it by data processing. However, through N1s XPS deconvolution, it was possible to identify amide and amine binding energies at 399.8(±0.1) eV and 401.9(±0.1) eV, respectively (Fig. 2D).28,43,44 The ATR-FTIR spectra for GO and man-GO reinforce the XPS results, showing characteristic signals of the mannosylated ethylenediamine (Fig. 3).


image file: c7tb02997g-f3.tif
Fig. 3 ATR-FTIR spectra for GO and man-GO.

This spectrum clearly shows the bands of the functional groups of GO. A broad and intense band with a maximum at 3310 cm−1 is attributed to νOH. The band at 1723 cm−1 is assigned to νC[double bond, length as m-dash]O from the carboxylic acid and ketone groups, and the band at 1625 cm−1 is attributed to νC[double bond, length as m-dash]C from network-conjugated bonds. The bands observed at 1160 cm−1 and 1047 cm−1 were assigned to νC–O–C from the epoxy groups and νC–O, respectively. The man-GO ATR-FTIR showed a band at 3310 cm−1 that was attributed to νOH. The band at 1723 cm−1 assigned to νC[double bond, length as m-dash]O from carboxylic acid showed a considerable reduction in intensity because the bond was converted to an amide bond after functionalization. The observed small band can probably be attributed to the ketone groups. The νC[double bond, length as m-dash]O of the amide bond overlaps with the νC[double bond, length as m-dash]C of the network GO, and both are observed at 1625 cm−1. The increase in the intensity of this band propose the occurrence reduction of GO after funtionalization and restoring of C[double bond, length as m-dash]C network, agreeing with the XPS results. This band becomes more intense and presents a shoulder at a lower frequency. Two clear bands were also observed at 1360 cm−1 and 1240 cm−1 for the mannosylated ethylenediamine and were attributed to δCH/δOH and amine νCN, respectively. The intense band observed at 1056 cm−1 is attributed to νC–O due to the contribution of the CO bonds of the mannosylated ethylenediamine. The clear signals observed for the sugar show that this molecule was efficiently coupled to GO.

The TGA and DTGA curves for GO and man-GO are shown in Fig. 4A and B, respectively.


image file: c7tb02997g-f4.tif
Fig. 4 TGA and DTGA curves of (A) GO and (B) man-GO.

The decomposition profile of man-GO was similar to that of mannosylated ethylenediamine (ESI). At approximately 200 °C, both compounds exhibited noteworthy weight loss related to their oxygen-containing functional groups. Up to 290 °C, the weight loss for GO was 38%, while that for man-GO was 30%. At 600 °C, both samples lost approximately 45% of their weight. For man-GO, a thermal event in the DTGA curve was observed between 230 °C and 500 °C, related to the functional groups of the sugar.

Raman spectroscopy is an essential technique for GO characterization because the C[double bond, length as m-dash]C bonds result in high Raman intensities.38 This material presented two main peaks: G and D bands. From the intensity ratio between the D and G peaks (ID/IG), the oxidation degree of the sample can be estimated. Fig. 5 shows the Raman spectra for GO and man-GO.


image file: c7tb02997g-f5.tif
Fig. 5 Raman spectra for GO and man-GO.

For GO, the G and D bands were observed at 1591.30 cm−1 and 1334.26 cm−1, respectively. In the literature, the G band was observed between 1580 cm−1 and 1601 cm−1, and the D band was observed between 1350 cm−1 and 1360 cm−1.45,46 For man-GO, the G band was observed at 1585.49 cm−1, and the D band was observed at 1331.46 cm−1, indicating a red shift of both bands after functionalization. The G band red shifted by 12.31 cm−1, and the D band red shifted by 2.8 cm−1. Studies have demonstrated that the red shift in the G band occurs as a function of the addition of epoxy and hydroxyl groups to GO.47 Actually, the red shift may have occurred due to the hydroxyl groups and C–O of the mannosylated ethylenediamine. The C–O groups of the sugar are more similar in chemical nature to the epoxy groups of the GO. Thus, it is reasonable to suppose that they are responsible for the red shift observed. It is important to verify that the intensity ratio of the D and G bands (ID/IG) had no considerable change after the functionalization of GO. The intensity ratio ID/IG was 1.329 for GO and 1.336 for man-GO. This indicates that GO showed low reduction after functionalization. Even the XPS showing the occurrence of the reduction, the man-GO has still many defects in its structure provided by oxygenated, amide and amine groups observed by intense binding energies between 286.0 eV and 289.2 eV, which they are responsible by defects in the material structure. Thus, these defects were also detected by Raman spectroscopy. In the literature can be found reports similar to this when GO was covalently functionalized with amines.48,49

TEM analysis was carried out to verify the morphology of GO and man-GO sheets. Fig. 6A and B show the GO and man-GO sheets, respectively.


image file: c7tb02997g-f6.tif
Fig. 6 TEM images of (A) GO and (B) man-GO.

Fig. 6A shows that the methodology used for GO sheet synthesis was very efficient at exfoliating and oxidizing graphite. Fig. 6B demonstrates that the morphology of GO was maintained after functionalization, indicating that this approach is a good alternative to the covalent mannosylation of nanomaterials. However, by AFM analysis (ESI, Fig. S2) it was possible to observe that the mannosylation of GO lead to its agglomeration. This higher degree of sheet agglomeration observed after functionalization can be attributed to the reduction of the oxygenated groups (e.g. epoxy and carboxylic groups) and consequent decreasing in the zeta potential value.

Hemolytic effect of pristine GO and man-GO

Previous results showed that GO was successfully functionalized with mannosylated ethylenediamine and that this functionalization did not affect the morphology and structure of the nanomaterial. Therefore, man-GO showed different behavior than pristine GO when applied to biological studies. Considering this aspect, the interaction of GO and man-GO with RBCs (hemolytic assay) was investigated to evaluate the potential blood compatibility and RBC toxicity of the samples prepared in this study (Fig. 7). The hemolytic assay was evaluated at six concentrations of each sample (GO and man-GO): 0, 10, 50, 100, 150, 200, and 250 μg mL−1. The toxic effect observed (hemolysis) was dose dependent for both samples (Fig. 7A). GO induced higher toxicity than man-GO, leading to membrane disruption of >70% of RBCs at the maximum concentration studied (250 μg mL−1). However, after the mannosylation of GO, a reduction of approximately 75% was observed in the hemolysis value (Fig. 7A). This effect could be associated with a decrease in electrostatic charges on the surface of man-GO (−30.3 mV) compared with that of GO (−42.7 mV), as observed through the zeta-potential measurements. Furthermore, because mannose is a non-toxic sugar used by living systems and a large part of the GO surface is covered by mannosylated ethylenediamine, this low toxicity is reasonable. This is because the RBCs do not directly interact with the functionalized nanomaterial surface, and the sugar is not toxic to the organism cells. Thus, mannosylation of GO reduces the interaction with RBC membranes, decreasing hemolysis. In this context, the mannosylation level of GO material is an interesting strategy to modulate its biological effects towards a better biocompatibility. A nanobiointerface with higher degree of GO mannosylation may result in low or no RBCs toxicity while a lower content of mannosylated ethylenediamine on GO surface may increase RBCs hemolysis. According to literature reports, hemolysis from pristine GO is related to distinct factors, including the surface charge, size distribution, morphology, surface area, and oxygen content, the latter being related to the formation of reactive oxygen species (ROS) that interfere in the lysis of the RBCs.50 However, there is no consensus regarding the hemolytic assay because of insufficient comparative studies on the incubation conditions (e.g., RBC type, storage duration and quality, medium type, osmolality, incubation pattern, and exposure time). In fact, some investigations have studied the hemolytic properties of pristine GO, but the results are largely controversial.51 Furthermore, covalent functionalization of GO with amines, proteins, and polymers has been described as an effective approach to reduce the toxicity of GO to RBCs.21,51–55
image file: c7tb02997g-f7.tif
Fig. 7 (A) Percentage of hemolysis of RBCs induced by GO and man-GO. (B) Hemolysis suppression after HC formation on GO and man-GO. The standard deviation is a result of three independent assays, where ***p < 0.01.

It is critical to understand the influence of surface chemical modification of man-GO on biomolecular interactions. GO and man-GO were incubated with 55% human blood plasma for 60 min to induce protein corona formation on the material surface. Thereafter, the hard corona (HC)-coated samples were incubated with RBCs in NaCl solution for 60 min to evaluate the influence of the HC coating on the hemolytic effect of GO and man-GO. A total hemolysis suppression effect (Fig. 7B) was observed, demonstrating that, regardless of the surface chemistry of GO, the biomolecular coating from human plasma (hard corona) on GO acts as a very effective surface shielding, thus mitigating the interaction of GO surface groups with the RBC membrane. Similar results have been recently reported for GO.25

Protein corona formation after mannosylation of GO

The PC composition modulates the biocompatibility and biodistribution of the nanoparticle in vivo;56,57 thus, the interaction of GO and man-GO with plasma proteins was evaluated. To compare the protein load on both samples, the proteins of the human plasma that form the HC coating on GO and man-GO were extracted and separated according to their molecular weights by 1D SDS-PAGE gel electrophoresis (Fig. 8A). The SDS-PAGE results reveal that surface modification of GO with D-mannose has considerably affected the amount and composition of adsorbed proteins. All protein bands visualized for man-GO were also observed for GO but with a higher intensity. Moreover, some protein bands (indicated by arrows in Fig. 8A) were only observed in the GO samples. These results are consistent with the lower content of the carboxylic groups available on man-GO in relation to GO due to the covalent attachment of mannosylated ethylenediamine. Another factor that may contribute to the different protein patterns presented by the samples is steric hindrance; i.e., the presence of mannosylated ethylenediamine on man-GO could generate steric effects capable of minimizing the access of proteins to the GO surface.
image file: c7tb02997g-f8.tif
Fig. 8 (A) SDS-PAGE of HC proteins extracted from GO and man-GO after interaction with 55% human blood plasma; (B) HC proteins associated with man-GO analyzed by LC–MS/MS mass spectrometry.

The structure and composition of nanomaterial–corona complex are influenced by several factors, such as size, shape, surface charge and surface functional groups of the nanoparticle; the characteristics of the biological environment, i.e., blood, temperature, pH, incubation time.58 The adsorption of proteins on the nanomaterial surface is a dynamic process governed by protein–nanoparticle binding affinities and affinity-based protein–protein interactions. The main forces that contribute to these interactions include van der Waals, H-bonds, salt bridge, Coulombic, electrostatic, hydrophobic, and π–π stacking.59 A competitive and selective binding of different human serum proteins on single-wall carbon nanotubes (SWCNTs) surface with different adsorption capacity and packing modes was demonstrated by experimental and theoretical approaches.60 The π–π stacking interactions between SWCNTs and protein aromatic residues were determinant in their adsorption capacity and greatly affected the subsequent cellular responses resulting in much reduced cytotoxicity for these protein-coated SWCNTs. Additionally, the plasma proteins present in the HC of man-GO were identified by mass spectrometry analysis. A list of the most abundant HC plasma proteins on man-GO is presented in Fig. 8B, and the whole array of the analyzed HC proteins (a total of 109 proteins were identified) is presented in the ESI (Table S1).

The proteins identified are mainly involved in immune response and transport.57,61 The most abundant group of proteins attached to GO materials were opsonins such as complement C3, inter-alpha-trypsin inhibitor heavy chain H4, fibrinogen alpha, fibrinogen beta and an isoform of the P02679 fibrinogen gamma chain. These proteins promote macrophage recognition and phagocytosis, leading to rapid removal of the nanomaterial from the circulation system via cells of the reticuloendothelial system (RES), resulting in less circulation time in the body.62,63 The PC of the GO materials was also significantly enriched by the transport proteins serum albumin and apolipoproteins, named dysopsonins. The binding of serum albumin prolongs the nanomaterial circulation time in blood, and the attachment of apolipoproteins promotes their translocation through the blood–brain barrier.27,64 Vitronectin and kininogen-1 are involved in tissue structuring and hemostasis, respectively.65 A recent study used the proteomics approach to characterize the bovine serum albumin corona and the human serum protein corona formed on different carbon-based nanomaterials, including GO. The proteins that were enriched in the PC of all the investigated nanomaterials were mainly complement factors, serum albumin and apolipoproteins, together with proteins involved in hemostasis and tissue structuring.66

The current study demonstrated that PC formation was drastically reduced on man-GO compared with that on pristine GO. This result indicates that mannosylated ethylenediamine linked to the GO surface prevented plasma protein binding, especially binding of proteins involved in opsonization. This phenomenon prolonged the nanomaterial circulation in the body, avoiding clearance from the blood circulation and limiting side effects caused by non-specific proteins. One additional advantage of this approach related with further biological GO material effects is the possibility to control the nanobiointerface by regulating the mannosylation degree on GO surface to prevent protein adsorption towards the design more biocompatible and safer nanomaterials. This behavior represents the first key factor for better in vivo performance and therapeutic response of the nanomaterial.56 Additionally, the enrichment of PC with dysopsonins represents another important method to increase the circulation time and improve man-GO bioavailability, as well as to enhance biodistribution and increase delivery to target sites.63,66 A commonly adopted method to improve the systemic circulation and decrease macrophage recognition of nanomaterials is their surface modification with poly(ethylene glycol) (PEG). It has been show that PEGylation reduce nonspecific protein adsorption, avoiding agglomeration and opsonisation.55 A comparative study regarding the interactions of serum proteins with GO and functionalized nano-GO with PEG (nGO-PEG) demonstrated that the latter showed a similar drastic reduction in protein binding but also certain selectivity and increased affinities toward a few proteins.67 Six nGO-PEG binding proteins are identified, and four of them are immune related factors, including C3a/C3a(des-Arg), an anaphylatoxin involved in local inflammatory responses. The nGO-PEG binds up to 2-fold amount of C3a/C3a(des-Arg) than unfunctionalized GO, and can efficiently decrease complement C3 activation. These results showed a new strategy to modulate the immune responses triggered by one nanomaterial through the addition of another type of nanomaterial. Therefore, the safer biomedical application of chemically surface-modified GO materials depends on a detailed knowledge of the interaction between proteins and nanosystems.

Another important feature required for biomedical applications of nanomaterials is their colloidal stability in biological medium. In this sense, we have evaluated the dispersion stability of GO and man-GO in deionized water, NaCl 0.15 mol L−1 solution and 55% human plasma over a period of 24 h (ESI, Fig. S3 and S4). Both materials presented a small portion unstable that precipitated in the first 3 h; the remaining material was very stable over a period of 24 h. GO and man-GO have completely lost their stability in NaCl solution, however in the presence of 55% plasma proteins their stabilities were restored due to protein corona formation. These results imply that for future biomedical applications of man-GO it will be necessary to perform a fractionation step in order to select the well stable fraction of man-GO; in addition to highlighted the role of the protein corona on GO and man-GO stabilization in biological environment.

Conclusion

Here, GO was covalently functionalized with D-mannose using mannosylated ethylenediamine as the intermediary precursor molecule. The spectroscopic techniques (XPS and ATR-FTIR) showed characteristic signals of the mannosylated ethylenediamine linked to GO. The TEM images demonstrated that the structure and morphology of the GO sheets did not change after the mannosylation procedure. Furthermore, covalent functionalization of GO drastically changed its biointeractions with RBCs and human plasma proteins. Considering the future biomedical applications of man-GO, our findings suggest that mannosylation can improve the blood compatibility of GO, resulting in a lower toxic effect to RBCs (hemolysis) than that of pristine GO. Moreover, it was demonstrated that the composition of hard corona proteins on the GO surface is critically affected by the mannosylation process. Finally, this work contributes to the understanding of the influence of GO surface chemistry bio-functionalization and its relationships with RBC toxicity and protein corona formation in the emerging field of biomedical nanotechnology.

Conflicts of interest

There are no conflicts to declare.

Acknowledgements

The authors gratefully acknowledge financial support from CNPq, INCT-Inomat, and NanoBioss-SisNANO/MCTI. The authors also extend gratitude to CNPEM open-facilities: LMN, LCS, and NBT at LNNano for X-ray photoelectron spectroscopy, Atomic force microscope, Hemolytic and protein corona studies, respectively; and MAS facility at Brazilian National Biosciences Laboratory (LNBio) for Mass spectrometry analysis.

References

  1. C. Xie, X. Lu, L. Han, J. Xu, Z. Wang, L. Jiang, K. Wang, H. Zhang, F. Ren and Y. Tang, ACS Appl. Mater. Interfaces, 2016, 8, 1707–1717 CAS.
  2. K. P. Loh, Q. Bao, G. Eda and M. Chowalla, Nat. Chem., 2010, 2, 1015–1024 CrossRef CAS PubMed.
  3. N. Mahmoudi and A. Simchi, Mater. Sci. Eng., C, 2017, 70, 121–131 CrossRef CAS PubMed.
  4. M. M. Song, H. L. Xu, J. X. Liang, H. H. Xiang, R. Liu and Y. X. Shen, Mater. Sci. Eng., C, 2017, 77, 904–911 CrossRef CAS PubMed.
  5. F. Panahi, R. Fareghi-Alamdari, S. K. Dangolani, A. K. Nezhad and M. Golestanzadeh, ChemistrySelect, 2017, 2, 474–479 CrossRef CAS.
  6. T. Yin, J. Liu, Z. Zhao, Y. Zhao, L. Dong, M. Yang, J. Zhou and M. Huo, Adv. Funct. Mater., 2017, 27, 1604620 CrossRef.
  7. A. Li, H. Ma and J. Liu, RSC Adv., 2016, 6, 63704–63710 RSC.
  8. Y. Wei, F. Zhou, D. Zhang, Q. Chen and D. Xing, Nanoscale, 2016, 8, 3530–3538 RSC.
  9. P. Li, S. Sunb, A. Dongc, Y. Haoa, S. Shi, Z. Suna, G. Gaoa and Y. Chen, Appl. Surf. Sci., 2015, 355, 446–452 CrossRef CAS.
  10. N. Soni, N. Soni, H. Pandey, R. Maheshwari, P. Kesharwani and R. K. Tekade, J. Colloid Interface Sci., 2016, 481, 107–116 CrossRef CAS PubMed.
  11. P. R. Taylor, L. Marinez-Pomares, M. Stacey, H. H. Lin, G. D. Bron and S. Gordon, Annu. Rev. Immunol., 2005, 23, 901–944 CrossRef CAS PubMed.
  12. A. Engel, S. K. Chatterjee, A. Al-arifi, D. Riemann, J. Langner and P. Nuhn, Pharm. Res., 2003, 20, 51–57 CrossRef CAS.
  13. B. Oh and C. H. Lee, Mol. Pharmaceutics, 2015, 12, 3226–3236 CrossRef CAS PubMed.
  14. M. E. Ragoussi, S. Casado, R. Ribeiro-Viana, G. Torre, J. Rojo and T. Torres, Chem. Sci., 2013, 4, 4035–4041 RSC.
  15. S. Gurunathan and J. H. Kim, Int. J. Nanomed., 2016, 11, 1927–1945 CrossRef CAS PubMed.
  16. N. Durán, D. S. T. Martinez, C. Silveira, M. Durán, A. C. M. Moraes, M. B. Simões, O. L. Alves and W. J. Fávaro, Curr. Top. Med. Chem., 2015, 15, 309–327 CrossRef.
  17. L. A. V. Luna, D. S. T. Martinez and O. L. Alves, in Nanotoxicoloy: Materials, Methodologies and Assessments, ed. N. Durán, S. S. Guterres and O. L. Alves, Nanomedicine and Nanotoxicology series, Springer, USA, 2014, ch. 16, pp. 347–361 Search PubMed.
  18. A. N. Ilinskaya and M. A. Dobrovolskaia, Nanomedicine, 2013, 8, 1013–1026 CrossRef PubMed.
  19. T. Wang, S. Zhu and X. Jiang, Toxicol. Res., 2015, 4, 885–894 RSC.
  20. S. K. Singh, M. K. Singh, P. P. Kulkarni, V. K. Sonkar, J. J. Grácio and D. Dash, ACS Nano, 2012, 6, 2731–2740 CrossRef CAS PubMed.
  21. N. Ma, J. Liu, W. He, Z. Li, Y. Luan, Y. Song and S. Garg, J. Colloid Interface Sci., 2017, 490, 598–607 CrossRef CAS PubMed.
  22. K. Bhattacharya, S. P. Mukherjee, A. Gallud, S. C. Burkert, S. Bistarelli, S. Belluci, M. Bottini, A. Star and B. Fadeel, J. Nanomed. Nanotechnol., 2016, 12, 333–351 CrossRef CAS PubMed.
  23. M. Hadjidemetriou and K. Kostarelos, Nat. Nanotechnol., 2017, 12, 288–290 CrossRef CAS PubMed.
  24. Q. Mu, G. Su and L. Li, ACS Appl. Mater. Interfaces, 2012, 4, 2259–2266 CAS.
  25. M. Papi, M. C. Lauriola, V. Palmieri, G. Ciasca, G. Maulucci and M. DeSpirito, RSC Adv., 2015, 5, 81638–81641 RSC.
  26. C. D. Walkey, J. B. Olsen, F. Song, R. Liu, H. Guo, D. W. H. Olsen, Y. Cohen, A. Emili and W. C. W. Chan, ACS Nano, 2014, 8, 2439–2455 CrossRef CAS PubMed.
  27. V. H. Nguyen and B. J. Lee, Int. J. Nanomed., 2017, 12, 3137–3151 CrossRef PubMed.
  28. M. Sousa, D. S. T. Martinez and O. L. Alves, J. Nanopart. Res., 2016, 18, 143–154 CrossRef.
  29. A. C. M. Moraes, P. F. Andrade, A. F. Faria, M. B. Simões, F. C. C. S. Salomão, E. B. Barros, M. C. Gonçalves and O. L. Alves, Carbohydr. Polym., 2015, 123, 217–227 CrossRef PubMed.
  30. A. Shevchenko, M. Wilm, O. Vorm and M. Mann, Anal. Chem., 1996, 68, 850–858 CrossRef CAS PubMed.
  31. S. Drewniak, R. Muzyka, A. Stolarczyk, T. Pustelny, M. K. Moranska and M. Setkiewicz, Sensors, 2016, 16, 1–16 CrossRef CAS PubMed.
  32. C. Tao, J. Wang, S. Qin, Y. Lv, Y. Long, H. Zhua and Z. Jiang, J. Mater. Chem., 2012, 22, 24856–24861 RSC.
  33. B. Yu, X. Wang, X. Qian, W. Xing, H. Yan, L. Ma, Y. Lin, S. Jiang, L. Song, Y. Hu and S. Lo, RSC Adv., 2014, 4, 31782–31794 RSC.
  34. K. Haubner, J. Murawski, P. Olk, L. M. Eng, C. Ziegler, B. Adolphi and E. Jaehne, Chem. Phys. Chem., 2010, 11, 2131–2139 CrossRef CAS PubMed.
  35. L. Zhang, Y. Li, L. Zhang, D. W. Li, D. Karpuzov and Y. T. Long, Int. J. Electrochem. Sci., 2011, 6, 819–829 CAS.
  36. S. Makharza, G. Cirillo, A. Bachmatiuk, I. Ibrahim, N. Ioannides, B. Trzebicka, S. Hampel and M. H. Rümmeli, J. Nanopart. Res., 2013, 15, 1–26 CrossRef.
  37. J. Briscoe, A. Marinovic, M. Sevilla, S. Dunn and M. Titirici, Angew. Chem., Int. Ed., 2015, 54, 4463–4468 CrossRef CAS PubMed.
  38. N. H. Kim, T. Kuila and J. H. Lee, J. Mater. Chem. A, 2013, 1, 1349–1358 CAS.
  39. S. Pei and H. M. Cheng, Carbon, 2012, 50, 3210–3228 CrossRef CAS.
  40. R. S. Vieira, M. L. M. Oliveira, E. Guibal, E. R. Castollón and M. M. Beppu, Colloids Surf., A, 2011, 374, 108–114 CrossRef CAS.
  41. C. M. Castilha, M. V. L. Ramón and F. C. Marín, Carbon, 2000, 38, 1995–2001 CrossRef.
  42. R. J. Waltman and J. Pacansky, Chem. Mater., 1993, 5, 1799–1804 CrossRef CAS.
  43. S. W. Lee, B. S. Kim, S. Chen, Y. Shao-Horn and P. T. Hammond, J. Am. Chem. Soc., 2009, 131, 671–679 CrossRef CAS PubMed.
  44. T. Ramanathan, F. T. Fisher, R. S. Ruoff and L. C. Brinson, Chem. Mater., 2005, 17, 1290–1295 CrossRef CAS.
  45. E. Nossol, A. B. S. Nossol, S. X. Guo, J. Zhang, X. Y. Fang, A. J. G. Zarbin and A. M. Bond, J. Mater. Chem. C, 2014, 2, 870–878 RSC.
  46. R. Ramachandran, M. Saranya, P. Kollu, B. P. C. Raghupathy, S. K. Jeong and A. N. Grace, Electrochim. Acta, 2015, 178, 647–657 CrossRef CAS.
  47. K. N. Kudin, B. Ozbas, H. C. Schiepp, R. K. Prud’homme, I. A. Aksay and R. Car, Nano Lett., 2008, 8, 36–41 CrossRef CAS PubMed.
  48. N. A. Carvajal, E. V. Basiuk, V. M. Laguna, I. P. Lee, M. H. Farías, N. Bogdanchikova and V. A. Basiuk, RSC Adv., 2016, 6, 113596–113610 RSC.
  49. S. Chakraborty, S. Saha, V. R. Dhanak, K. Biswas, M. Barbezat, G. P. Terrasid and A. Chakraborty, RSC Adv., 2016, 6, 67916–67924 RSC.
  50. A. Sasidharan, L. S. Panchakarla, A. R. Sadanandan, A. Ashokan, P. Chandran, C. M. Girish, D. Menon, S. V. Nair, C. Rao and M. Koyakutty, Small, 2012, 8, 1251–1263 CrossRef CAS PubMed.
  51. Y. Wang, B. Zhang and G. Zhai, RSC Adv., 2016, 6, 68322–68334 RSC.
  52. C. Cheng, S. Li, S. Nie, W. Zhao, H. Yang, S. Sun and C. Zhao, Biomacromolecules, 2012, 13, 4236–4246 CrossRef CAS PubMed.
  53. C. Chen, S. Nie, S. Li, H. Peng, H. Yang, L. Ma, S. Sun and C. Zhao, J. Mater. Chem. B, 2013, 1013(1), 265–275 RSC.
  54. N. L. Zhou, H. Gu, F. Tang, W. X. Li, Y. Chen and J. Yuan, J. Mater. Sci., 2013, 48, 7097–7103 CrossRef CAS.
  55. F. S. Kiew, L. V. Kiew, H. B. Lee, T. Imae and L. Y. Chung, J. Controlled Release, 2016, 226, 217–228 CrossRef PubMed.
  56. W. Hu, C. Peng, M. Lv, X. Li, Y. Zhang, N. Chen, C. Fan and Q. Huang, ACS Nano, 2011, 5, 3693–3700 CrossRef CAS PubMed.
  57. P. Aggarwal, J. B. Hall, C. B. McLeland, M. A. Dobrovolskaia and S. E. McNeil, Adv. Drug Delivery Rev., 2009, 61, 428–437 CrossRef CAS PubMed.
  58. M. Neagu, Z. Piperigkou, K. Karamanou, A. B. Engin, A. O. Docea, C. Constantin, C. Negrei, D. Nikitovic and A. Tsatsakis, Arch. Toxicol., 2017, 91, 1031–1048 CrossRef CAS PubMed.
  59. B. Kharazian, N. L. Hadipour and M. R. Ejtehadi, Int. J. Biochem. Cell Biol., 2016, 75, 162–174 CrossRef CAS PubMed.
  60. C. Ge, J. Du, L. Zhao, L. Wang, Y. Liu, D. Li, Y. Yang, R. Zhou, Y. Zhao, Z. Chai and C. Chen, Proc. Natl. Acad. Sci. U. S. A., 2011, 108, 16968–16973 CrossRef CAS PubMed.
  61. M. Xu, J. Zhu, F. Wang, Y. Xiong, Y. Wu, Q. Wang, J. Weng, Z. Zhang, W. Chen and S. Liu, ACS Nano, 2016, 10, 3267–3281 CrossRef CAS PubMed.
  62. C. C. Fleischer and C. K. Payne, Acc. Chem. Res., 2014, 47, 2651–2659 CrossRef CAS PubMed.
  63. R. Cukalevski, S. A. Ferreira, C. J. Dunning, T. Berggård and T. Cedervall, Nano Res., 2015, 8, 2733–2743 CrossRef CAS.
  64. C. Corbo, R. Molinaro, A. Parodi, F. N. E. Toledano, F. Salvatore and E. Tasciotti, Nanomedicine, 2016, 11, 81–100 CrossRef CAS PubMed.
  65. W. Yang, S. Liu, T. Bai, A. J. Keefe, L. Zhang, J. R. Ella-Menye, Y. Li and S. Jiang, Nano Today, 2014, 9, 10–16 CrossRef CAS.
  66. M. Sopotnik, A. Leonardi, I. Krizaj, P. Dusak, D. Makovec, T. Mesaric, N. P. Ulrih, I. Junkar, K. Sepcic and D. Drobne, Carbon, 2015, 95, 560–572 CrossRef CAS.
  67. X. Tan, L. Feng, J. Zhang, K. Yang, S. Zhang, Z. Liu and R. Peng, ACS Appl. Mater. Interfaces, 2013, 5, 1370–1377 CAS.

Footnote

Electronic supplementary information (ESI) available. See DOI: 10.1039/c7tb02997g

This journal is © The Royal Society of Chemistry 2018