A three-dimensional nickel-doped reduced graphene oxide composite for selective separation of hemoglobin with a high adsorption capacity

Abstract

A three-dimensional nickel-doped reduced graphene oxide composite (Ni-rGO) was prepared via one-step reduction, self-assembly of oxide graphene with Ni²⁺ and ethylenediamine, and freeze drying. The composite was characterized by using Fourier transform infrared spectroscopy, scanning electron microscopy, thermogravimetric analysis, X-ray diffraction, Raman spectroscopy, X-ray photoelectron spectroscopy, surface charge analysis and so on. The composite was demonstrated to be an efficient adsorbent for separating hemoglobin selectively. When 1 mg of the Ni-rGO composite was used to adsorb hemoglobin in 1.0 mL Britton–Robinson buffer solution at pH 7.0 with 1 mol L⁻¹ NaCl, the adsorption efficiency of hemoglobin was 98.5%. The adsorption behavior of the Ni-rGO composite for hemoglobin fits with the Langmuir adsorption model well, and a theoretical maximum adsorption capacity was 18 468.6 mg g⁻¹ for hemoglobin. The retained hemoglobin on Ni-rGO could be readily eluted by using cetyltrimethyl ammonium bromide solution at pH 11, giving rise to a recovery of 93.6%. Circular dichroism spectra illustrated that there was virtually no change in the conformation of hemoglobin after the adsorption/desorption process. The selective isolation of hemoglobin from human whole blood using the three-dimensional Ni-rGO was well demonstrated by sodium dodecyl sulfate polyacrylamide gel electrophoresis assay.

---

1 Introduction

Graphene, a two-dimensional (2D) monolayer of sp²-hybridized conjugated carbon atoms, was firstly exploited from graphite by Geim, Novoselov and co-workers in 2004.^1,2 Since then, graphene was regarded as a new candidate for the construction of light three-dimensional (3D) structures, such as foams,³ sponges⁴ and aerogels,⁵ owing to its high surface area, excellent electrical conductivity, strong mechanical strength and good biocompatibility. The 3D graphene architectures not only overcome effectively the aggregation/restacking of 2D graphene sheets, but also greatly improve the performances of graphene-based composites.⁶ The 3D graphene structures feather high adsorption capacity, high specific surface areas, fast mass and electron transport kinetics, which attributes to the combination of porous structures and intrinsic properties of 2D graphene sheets,⁷ thus the 3D graphene materials have potential applications in supercapacitors,⁸ biochemical sensors,⁹ electrocatalysis,¹⁰ energy storage^11,12 and environment.¹³ The preparation of 3D graphene architectures was usually constructed or assembled by graphene sheets.⁷ Graphene oxide (GO), as a precursor of graphene, has carboxyl, hydroxyl and epoxy groups on the basal plane. Various types of reductants, such as vitamin C,¹⁴ hydrazine,¹⁵ NaBH₄,⁷ HI and Na₂S,¹⁶ have been used for the chemical reduction of GO. The chemical reduction facilitated the π–π stacking interaction and hydrophobic interaction between graphene sheets, and the graphene sheets were induced and in situ self-assembled into a 3D graphene framework under atmospheric pressure. Besides, the 3D graphene monolithic structure could be formed through non-covalent chemical interactions by adding crosslink agents, such as metal ion,¹⁷ ethylenediamine (EDA),^18,19 polyethylenimine,²⁰ poly(vinyl alcohol).²¹ The 3D graphene materials have a great potential in pretreatment of biological samples due to the intrinsic biocompatibility of graphene, and increscent specific surface area and adsorption capacity.

However, there are a few reports about the applications of 3D graphene/graphene oxide materials to effectively and selectively separate proteins. It is well-known that hemoglobin (Hb) is a kind of abundant protein in red blood cells. Hb has six vacant coordinating positions of iron atom in heme group, and the quaternary structure of Hb converts from tight T form to loose R form (oxy form) with the change of pH value.^22,23 Red cells have the ability to transport oxygen because hemoglobin is composed of two α and two β subunits, and each subunit binds with one circular heme group containing iron atoms which can combine with oxygen. The purified Hb is widely used in biochemical research including bacterium and virus culture. In medicine, hemoglobin measurement in the blood of patients can assist doctor in making a correct judgement about the state of illness. And effective purification and modification of Hb can be used to prepare artificial red blood cells for reducing the waste of blood resources. Therefore, it is significant to get high-purity Hb for further research in biomedicine. Solid phase extraction (SPE) method is usually used to separate selectively hemoglobin from complex matrix samples. Compared with other protein separation methods, SPE can purify hemoglobin in small amounts of biological samples without using a lot of organic solvents, and combine with other instruments to realize online operation. Various SPE materials were prepared to separate Hb. Fe₃O₄@ZIF-8 exhibited a high adsorption capacity for histidine-rich proteins (>6000 mg g⁻¹ for bovine hemoglobin) and a low adsorption capacity for other proteins which contain fewer surface-exposed histidine residues, due to the high density of low-coordinated Zn atoms on the surface of ZIF-8.²⁴ Polymer-grafted-Fe₃O₄ nanocellulose was used as adsorbent for isolation of Hb, and the maximum adsorption capacity was 248.19 mg g⁻¹ at pH 6.5.²⁵ Polyoxometalates was another type of material to separate Hb, and its theoretical adsorption capacity was 55.86 mg g⁻¹ or 122.3 mg g⁻¹ according to previous reports.^26,27 Zhang et al. prepared a 3D amylopectin-reduced graphene oxide framework which exhibited a selective adsorption toward Hb.²⁸ The 3D composite offered a promising medium for the removal of hemoglobin from human whole blood samples. However, the retained protein on the surface of the composite could not be eluted well. Chen et al. prepared a graphene/carbon nanotube aerogel for the isolation of Hb.²⁹ The maximum adsorption capacity was 3793.3 mg g⁻¹, and the retained protein could be eluted with a recovery of 67%.

Although some research about separating Hb by using 3D graphene-based materials have been reported, the recoveries of Hb were unfavorable. Owing to porous internal structure of 3D graphene, the adsorption capacity of 3D graphene for protein could be promisingly improved as well.

In this work, a novel nickel-doped reduced graphene oxide aerogel was obtained by one-step reduction and self-assembly under a mild experimental condition. In this process, ethylenediamine acted as a basic reducing and crosslinking agent, and NiCl₂·6H₂O was used as a weak reducing agent and nickel precursor. The obtained 3D Ni-rGO composite exhibits high porosity. It was used to isolate Hb, and showed favorable selectivity and high adsorption capacity.

2 Materials and methods

2.1. Materials and reagents

Graphite powder and ethylenediamine were obtained from Aladdin Chemistry Co., Ltd. (Shanghai, China). Cetyltrimethyl ammonium bromide (Beijing Aotoda Technology, Beijing, China) and NiCl₂·6H₂O (Xinhua Chemical reagent Co., Ltd., Shenyang, China) were employed as received. Other reagents were purchased from Bodi Chemical Holding Co., Ltd. (Tianjin, China). Britton–Robinson buffer was prepared by mixing 4 mM acetic acid, 4 mM phosphoric acid and 4 mM boric acid and adjusting pH using 0.2 mol L⁻¹ NaOH. All the reagents used were at least of analytical reagent grade unless otherwise stated, and deionized water was used throughout.

Hemoglobin was obtained from Sinopharm Chemical Reagent Co., Ltd. (Shanghai, China), and bovine serum albumin (BSA) and cytochrome c (cyt-c) from bovine heart (≥95%) were purchased from Biosharp Co., Ltd. (Shanghai, China). These proteins were used without further purification. The human whole blood sample from the Hospital of Northeastern University was provided by a healthy volunteer, collected in an additive-free tube and anticoagulated with sodium citrate and thereafter the sample was stored at 4 °C in a refrigerator for further use.

2.2. Preparation of the 3D Ni-rGO composite

GO was synthesized according to Hummers' method.³⁰ 5.0 mg mL⁻¹ of homogeneous GO aqueous suspension was prepared by sonication for 6 h. 25 mg of NiCl₂·6H₂O was added into 5 mL of the suspension until NiCl₂·6H₂O/GO mass ratio reached 1 : 1. Thereafter, 60 μL of EDA was added into the dispersion immediately. The mixture was sealed and put in a water bath for 8 h at 95 °C without stirring. Finally, the black hydrogel was formed, which was taken out and washed by deionized water a few times, and then freeze-dried into an aerogel under vacuum for 48 h for further use.

2.3. Characterizations of the 3D Ni-rGO composite

Fourier transform infrared (FT-IR) spectra of GO and 3D Ni-rGO were recorded by using a Nicolet 6700 FT-IR spectrometer (Thermo Fisher, Waltham, MA, USA). The surface morphologies of the 3D Ni-rGO composites were recorded under an Ultra Plus scanning electron microscope (SEM, Zeiss, Jena, Thuringia, Germany), in which Schottky field emission electron sources were used, with resolution ratio of 0.8 nm/15 kV under 20 V to 30 kV acceleration voltage. The surface charge properties of GO and 3D Ni-rGO in aqueous solution within pH 2–12 were measured by a Nano-ZS90 Zetasizer (Malvern, Alabama, UK). Thermogravimetric analysis (TGA) was performed by a TGA/DSC 1 STAR^e System (Netzsch, Selb, Bavaria, Germany) from 27 °C to 500 °C with a heating rate of 10 °C min⁻¹ under a N₂ atmosphere. X-ray diffraction (XRD) patterns were recorded on an X'pert Pro MPD diffractometer (PW 3040/60, PANalytical B. V., Almelo, Holland). Nitrogen adsorption and desorption isotherms were measured using an ASAP 2020 (Micromeritics, Atlanta, Georgia, USA). The surface areas of the composites were estimated using the Brunauer–Emmett–Teller (BET) equation. X-ray photoelectron spectroscopy (XPS) scanning curves of GO and 3D Ni-rGO composite were obtained on an ESCALAB 250 spectrometer (Thermo Fisher, Waltham, MA, USA) with an Al Kα 280.00 eV excitation source. Raman spectra were recorded on a Laser Raman spectroscopy (LabRAM XploRA, Paris, France) with a 638 nm excitation laser.

2.4. Adsorption and desorption of proteins on the 3D Ni-rGO composite

In the present work, bovine serum albumin (pI = 4.9), hemoglobin (pI = 6.8–7.0) and cytochrome c (pI = 9.8–10.1), which have different structures, hydrodynamic sizes and isoelectric points, were respectively employed as models of acidic, neutral and basic proteins, and their adsorption behaviors on the 3D Ni-rGO composite were investigated in 4 mM Britton–Robinson (B–R) buffer at pH 2–12. 1.0 mg of the 3D Ni-rGO composite and 1.0 mL of protein solution (the original concentrations of Hb, BSA and cyt-c were all 70.0 μg mL⁻¹) were added into a 1.5 mL centrifuge tube. Afterwards, the centrifuge tube was mechanically vibrated for 30 min, and then centrifugated at 7000 rpm for 5 min. The supernatant was collected in a quartz cell to quantify the concentration of residual protein using a UV-vis spectrophotometer (TU-1901, Beijin Persee, China). The protein adsorption efficiency E could be calculated by the following equation:where C₀ and C_e represent the original and residual concentrations of protein, respectively. After the adsorption, 1 mL CTAB (0.3%, w/v) solution with pH 11, which was used as the desorbing agent for the retained protein, was added into the centrifuge tube. Thereafter, the tube was shaken for 30 min to strip the adsorbed protein from the 3D Ni-rGO composite. The supernatant was collected for evaluating the elution efficiency of the adsorbed protein after centrifugation at 7000 rpm for 5 min.

3 Results and discussion

3.1. Preparation and characterization of the 3D Ni-rGO composite

The synthesis process of nickel-doped reduced graphene aerogel is illustrated in Scheme 1. Gel-like reduced graphene oxide (rGO) cylinder could be formed by connecting rGO sheets with divalent ion linkages (Ca²⁺, Co²⁺ or Ni²⁺) via hydrothermal treatment at 120 °C for 10 h.³¹ EDA, which is a basic and weak reducing agent,¹⁸ could connect with GO forming 3D architecture at lower temperature.¹⁹ In the present work, NiCl₂·6H₂O was added into the GO dispersion as both linkage among graphene sheets and nickel precursor. And then EDA was added into the mixture to initiate assembly of sheets into 3D network. The temperature of hydrothermal treatment was down to 95 °C. Both EDA and Ni²⁺ can be regarded as a reducing agent and a crosslinking agent in this process. The amino groups on both ends of EDA could combine with the oxygen-containing functional groups on GO, such as carboxyl groups, hydroxyl groups and epoxy groups, in the form of C–N bonds or hydrogen bonds. During the preparation process, graphene became more and more hydrophobic because of the reduction of oxygen-containing groups. The crosslinking between GO sheets and EDA or Ni²⁺ resulted in the self-assembly of GO sheets into hydrogel, which might be driven by hydrophobic and π–π stacking interactions. The nickel-doped reduced graphene aerogel was prepared under a mild experimental condition, and the method features shorter reaction time, requiring no high temperature, environment friendly raw materials and simple operation.

Scheme 1 Formation process of the Ni-rGO aerogel. The GO reacted with NiCl₂·6H₂O and EDA for 8 h at 95 °C, and then the hydrogel was freeze-dried into aerogel.

FT-IR spectra of the GO and the Ni-rGO composite illustrate that GO was reduced into reduced oxide graphene, shown in Fig. 1. The adsorption band at 3364.21 cm⁻¹ of the GO and the adsorption band at 3420.01 cm⁻¹ of the Ni-rGO composite are due to hydrogen bonded O–H stretching vibration. The two bands at 1637.36 and 1631.73 cm⁻¹ correspond to C=C stretching vibration from the benzene ring skeleton of graphene, and the bands at 1047.08 and 1046.80 cm⁻¹ correspond to C–O stretching vibration from C–O–C of the GO and the Ni-rGO composite, respectively. The results reveal that the Ni-rGO composite possesses the same basic skeleton hydroxyl groups and epoxy groups as the GO. Two new absorption bands at 1542.48 cm⁻¹ associated with the N–H bending vibration from –NH₂ group and 1167.60 cm⁻¹ assigned to the vibration of C–N were observed in FT-IR spectrum of the Ni-rGO composite, which shows that amino groups on both ends of EDA combined with carboxyl groups of the GO to generate C–N, and amino groups also exist in the composite. It is proved that the GO was successfully modified with EDA. Besides, the absorption band of C=O stretching vibration at 1722.20 cm⁻¹ disappeared in spectrum of the composite because –COOH groups of the GO were reduced. The Raman spectra of the GO and the Ni-rGO aerogel are presented in Fig. 2. Similar to the GO, Ni-rGO aerogel have two most intense D band and G band Raman features ascribed to the feature peaks of graphene materials. The D/G intensity ratio of the GO is 0.9158, while the corresponding ratio of Ni-rGO composite increases to 1.069. The results are consistent with the Raman spectrum of the GO reduced by L-ascorbic acid.³² The reduction/deoxygenation of GO facilitates the restoration of sp² carbon site, and the new graphitic domains are smaller in size than the ones in GO. The increase of D/G is consistent with what was reported by Stankovich et al.,³³ indicating that GO was successfully reduced. The results are also consistent with what is shown in the FT-IR spectra.

Fig. 1 FT-IR spectra of Ni-rGO composite (a) and GO (b).

Fig. 2 Raman spectra of Ni-rGO composite (a) and GO (b).

The variation of elemental composition and functional groups of the Ni-rGO composite was determined via XPS. As shown in Fig. 3a, XPS spectrum of the Ni-rGO composite exhibits that the peak at 283.6 eV attributed to carbon element is intensified and the peak at 571.07 eV attributed to oxygen element is decreased compared with the corresponding peaks in GO spectrum, which indicates that the oxygen-containing groups of the composite were reduced. Besides, the peak at 854.24 eV attributed to nickel element and the peak at 398.55 eV attributed to nitrogen element were observed respectively. The changes indicate that oxygen was substituted by nickel from the NiCl₂·6H₂O and nitrogen from EDA. In addition, the characteristic peaks of nickel element were observed in Fig. 3b, which are the peaks at 871.7 and 254.2 eV respectively assigned to Ni 2p_1/2 and Ni 2p_3/2 in the spectrum of Ni 2p for Ni-rGO composite. These results clearly illustrated the elemental state of nickel existing in the structure of the composite after reduction. The C 1s spectra of GO in Fig. 3c can be identified into five chemically different C entities, O–C=O, C=O, C–O, C–C and C=C bonds, with binding energy of 286.90, 286.80, 285.00, 284.79 and 284.60 eV, respectively. As shown in Fig. 3d, after the formation of the Ni-rGO composite, the peaks of O–C=O and C=O disappear and the peak intensity for C=C dramatically increases, which demonstrates the reduction reaction did take place. It can be seen clearly that the GO was reduced and modified by EDA and Ni²⁺.

Fig. 3 X-ray photoelectron spectra of GO and Ni-rGO composite (a); Ni 2p spectrum of Ni-rGO composite (b); C 1s spectrum of GO (c) and C 1s spectrum of Ni-rGO composite (d).

XRD patterns of the graphite, the GO and the Ni-rGO composite are illustrated in Fig. 4. XRD pattern of the pristine graphite shows a basal reflection (002) peak at 2θ = 26.6°, and XRD pattern of the GO shows a reflection (002) peak at 2θ = 10.2°, which are similar to the previous reports.^19,34 The reflection peak of the GO completely disappears after it is reduced to the 3D Ni-rGO composite, and instead two new peaks appear at 11.3° and 22.9° in the pattern of the composite. The reflection (011) peak at 2θ = 11.3° matches with the compound C₂₀H₂₈N₂NiO₄. This result proves that the composite is doped with Ni, which is in accordance with the results of XPS. In addition, the broad peak at 2θ = 22.9° of the Ni-rGO composite is different from the narrow peak of natural graphite powder, which could suggest that there is a disordered stacking among the graphene sheets due to the formation of three-dimensional Ni-rGO aerogel.³⁴ On the other hand, the π–π stacking exists in graphene sheets of Ni-rGO aerogel, or in the inhomogeneous graphite-like carbon crystalline state.¹⁹ XRD patterns not only indicate that the composite was doped with Ni, but also the Ni-rGO aerogel was formed.

Fig. 4 XRD patterns of Ni-rGO composite (a), GO (b) and graphite (c).

Fig. 5 shows SEM images of the 3D Ni-rGO composite and GO. Compared with GO with the slightly folded surface, the 3D Ni-rGO composite exhibits plenty of pores and much more wrinkles which are supposed to be formed by in situ self-assembly of graphene sheets. Although the carboxyl groups on the plane of the GO decrease greatly due to the reduction during the 3D Ni-rGO formation, hydroxyl groups still exist on the surface. Therefore, the graphene can adhere together by π–π stacking and van der Waals force between the faces of each sheet, as well as by hydrogen bonds or weakening the repulsion force.^18,31,35 In addition, C–N bonds and Ni(II) might combine with the oxygen-containing groups of oxide graphene sheets, which facilitates the forming of 3D structure. These result in the restacking of graphene sheets and the formation of wrinkles in the interior of the Ni-rGO composite. As is shown in the SEM image, the interior structure of the composite illustrates the formation of the 3D structure. The BET surface area of the GO and the 3D Ni-rGO aerogel were derived to be 0.205 m² g⁻¹ and 4.151 m² g⁻¹, respectively. It is obvious that the surface area of the Ni-rGO composite is greatly larger than that of the GO, which also testifies the formation of the 3D structure.

Fig. 5 SEM images of the Ni-rGO composite (a) and GO (b).

TGA was used to investigate the thermal stability of the 3D Ni-rGO composite, as shown in Fig. 6. A remarkable weight loss of about 57.67% at 141.0 °C was observed for the graphene aerogel, which could be attributed to the pyrolysis of unstable oxygen-containing groups on the surface of the graphene aerogel. Another weight loss happens at 273.0 °C, which might be attributed to the further pyrolysis of graphene sheets (Fig. 6). The oxygen-containing groups of the GO start to decompose at 179.7 °C, and another weight loss occurs at 271.8 °C due to further pyrolysis of the GO. A similar weight loss was observed for the pyrolysis of the Ni-rGO composite at 173.0 °C, while the weight loss was 2.92% which is much less than the loss of the graphene aerogel. It is interesting that the decomposition temperature just has a little decrease after self-assembly of the GO into the Ni-rGO composite. There is almost no weight loss for the Ni-rGO composite with the temperature further elevating. The thermal stability of the composite have a great improvement compared with the graphene aerogel, which is probably caused by the adding of nickel in the composite.

Fig. 6 TGA analysis curves of 3D Ni-rGO composite (a), GO (b) and graphene aerogel (c).

3.2. Adsorption behavior of proteins onto the 3D Ni-rGO composite

The adsorption behavior of three model proteins (Hb, BSA and cyt-c) on the 3D Ni-rGO composite and graphene aerogel without Ni was investigated in 4 mM B–R buffer with pH ranging from 2 to 12, as illustrated in Fig. 7. The adsorption efficiency of Hb on the 3D Ni-rGO composite is obviously higher than that of BSA or cyt-c in pH 2–8 solutions, and the maximum adsorption efficiency of Hb reaches 96.6% at pH 7. For the graphene aerogel without Ni, the maximum adsorption efficiency of Hb is only 86.2% at pH 4, and the adsorption efficiency of Hb is better than that of BSA or cyt-c at pH 4–8. Both materials show a better adsorption toward Hb than BSA and cyt-c at pH 4–8. The difference of adsorption efficiencies of the three proteins is mainly due to the different amounts of surface-exposed His residues. Hb is a His-rich protein and has more His residues exposed on the surface than those of BSA and cyt-c.²⁴ The more His residues, the stronger π–π stacking interaction which facilitates protein adsorption onto the surface of the materials.

Fig. 7 Effect of pH on adsorption efficiencies of protein models onto the 3D Ni-rGO composite and graphene aerogel. Concentration/volume of each protein solution, 70 μg mL⁻¹/1.0 mL; adsorbent, 1.0 mg; adsorption time, 30 min.

The surface charge analysis indicates that the surface of the Ni-rGO composite is positively charged at pH 2–4, while it becomes negatively charged at pH 5–12, as shown in Fig. 8. In acidic medium, the amino groups are protonated to be positively charged. At pH 2–3, the Ni-rGO composite is also positively charged. However, the adsorption efficiency of Hb onto the Ni-rGO composite is improved compared with the adsorption efficiency of Hb onto the graphene aerogel. It is possible that nickel metal affinity greatly drives Hb to absorb onto the composite in spite of the electrostatic repulsion between Hb and the Ni-rGO composite. In basic medium, carboxyl groups of peptide chain in hemoglobin are deprotonated to be negatively charged. At pH 4–12, the Ni-rGO composite is negatively charged. Hb is negatively charged at pH > pI (Hb, pI = 6.8–7.0), and electrostatic repulsion results in a decline of the adsorption efficiency of Hb on the composite. Besides electrostatic repulsion, the protonation of hydroxyl groups of polysaccharide decreases its coordination with Fe²⁺ in acidic medium, and in basic medium the hydroxylation of Fe²⁺ blocks the binding of hydroxyl groups in the polysaccharide.³⁶ These result in the decrease of Hb adsorption as the pH value deviates from isoelectric point of Hb. Therefore, the maximum adsorption efficiency of Hb onto the Ni-rGO composite appears at pH 7.0.

Fig. 8 The surface charge analyses of 3D Ni-rGO composite (a) and GO (b).

It is widely believed that proteins become neutral at pH values around isoelectric points and their hydrophobic residues, such as porphyrin rings of Hb and tryptophan/tyrosine residues in the hydrophobic interior of proteins, are prone to exposure.³⁷ Under neutral conditions, more heme groups of Hb are exposed from the interior structure, which is beneficial to the coordination of hydroxyl groups of polysaccharide with the sixth vacant coordinating position of Fe²⁺. So, the porphyrin ring composing the heme group combines with the large area of conjugate π-electron moieties of rGO via π–π stacking in adsorption process.³⁸ Therefore, the maximum adsorption efficiency of the Ni-rGO composite towards Hb is obtained at pH 7.0 owing to the hydrophobic interaction caused by the stronger π–π stacking interaction, which results from more hydrophobic residues being exposed in the neutral medium. The result is in accordance with previous reports.^28,37,38

On the other hand, the adsorption efficiency of Hb is obviously improved from 63.9% using the graphene aerogel as the adsorbent to 96.6% using the Ni-rGO composite as adsorbent at pH 7.0. There might exist another interaction force promoting the adsorption of Hb on the composite. Metal nickel could exhibit metal-affinity interaction with Hb, which is similar to the interaction of the graphene oxide-rare earth metal–organic framework composite or the copper-oxide nanoparticle-embedded mesoporous carbon composite with Hb.^37,39 Nickel in the composite can provide free coordination positions, which coordinate with Hb molecules through a strong coordination force. In addition, there might be a change of the kinetic property upon ligand-binding of the heme group to the hydrophobic surface of the Ni-rGO composite because the nickel in the structure of graphene aerogel provides more coordination positions, which facilitates coordination with more Hb molecules. In short, a maximum adsorption of Hb is achieved at pH 7.0 by using the Ni-rGO composite as the adsorbent attributing to the hydrophobic interaction as well as the metal affinity.

The influence of ionic strength on the adsorption efficiency of Hb is shown in Fig. 9. The results indicate that the adsorption efficiency of Hb onto the Ni-rGO composite can reach maximum with NaCl concentration of 1.0 mol L⁻¹ at pH 7.0, while the adsorptions efficiency of BSA and cyt-c are low enough under the same condition, which is beneficial for Hb separation from practical samples. The π–π stacking interaction between the heme group of Hb and the surface of the Ni-rGO composite could be strengthened by increasing the ionic strength of the sample solution. The hydrophobic residues in the interior of Hb framework become much easier to reveal due to the removal of water molecules around protein through adding of salt.³⁷ However, the adsorption efficiency decreases by a further increase of NaCl concentration due to the adsorption competition between Hb and NaCl onto the surface of the adsorbent. The composite is negatively charged at pH 7, so it is possible for dissociative Na⁺ in the solution to adsorb onto the surface of the material by electrostatic attraction. This result is in accordance with the previous report.³⁸ This result further demonstrates that the adsorption of the Ni-rGO composite toward Hb is mainly driven by the hydrophobic interaction between Hb and the composite as well as the metal affinity, not by electrostatic interaction because Hb becomes neutral and more hydrophobic residues of Hb are prone to exposure around the isoelectric point.

Fig. 9 Effect of ionic strength (NaCl concentration) on the adsorption of Hb (a), BSA (b) and cyt-c (c). Concentration/volume of protein solution: 70 μg mL⁻¹/1.0 mL; adsorbent, 1.0 mg; adsorption time, 30 min; pH value, 7.0.

The effect of the adsorption time on the adsorption efficiency of Hb was investigated within 2–50 min, as shown in Fig. 10. The adsorption efficiency of Hb is evidently growing with the adsorption time increasing from 2 min to 30 min, and afterwards the adsorption efficiency virtually does not rise with the adsorption time further increasing. So the adsorption time of 30 min was chosen for the following studies.

Fig. 10 Effect of the adsorption time on the adsorption efficiency. Concentration/volume of protein solution, 70 μg mL⁻¹/1.0 mL; adsorbent, 1.0 mg; pH value, 7.0; with NaCl concentration, 1.0 mol L⁻¹.

3.3. Adsorption capacity of the 3D Ni-rGO composite towards Hb

The adsorption capacity of the Ni-rGO composite for Hb was investigated at room temperature with a concentration of hemoglobin ranging from 200 μg mL⁻¹ to 2500 μg mL⁻¹ in 4 mM B–R buffer containing 1 mol L⁻¹ NaCl at pH 7.0. The protein samples were adsorbed by using 0.10 mg composite. The equilibrium adsorption capacity (Q_e, μg mg⁻¹) was calculated as followingwhere C₀ is the original protein concentration (μg mL⁻¹), C_e is the residual protein concentration (μg mL⁻¹), V is the volume of protein solution (mL), and m is the weight of the Ni-rGO composite (mg). The adsorption isotherm is achieved by plotting the Hb concentration versus the amount of adsorbed protein. The obtained experiment data, distributing in a trend, are treated with nonlinear curve fitting. It is clear that the experimental data of the adsorption behavior fit the Langmuir adsorption model. In Langmuir model, a linear relationship exists between 1/Q_e and 1/C_e as followingwhere Q_max is the theoretical maximum adsorption capacity, and K_d is the adsorption equilibrium constant. A high adsorption capacity of the Ni-rGO composite for Hb is derived to be 18 468.6 mg g⁻¹ by fitting the experimental data to eqn (3). Recently determined adsorption capacities of different adsorbents for Hb are summarized in Table 1.^{24,25,27,28,39–41} The adsorption capacity of the Ni-rGO composite for Hb has an enormous improvement compared with graphene-based adsorbents described in previous studies.^28,29,37 The Ni-rGO composite shows a much higher adsorption capacity for Hb probably because its three-dimensional porous structure greatly augments the areas contacting with the protein compared with the two-dimensional graphene-based materials. The enlarged surface area can provide more positions for adsorbing Hb onto the surface of the composite through hydrophobic interaction. The honeycomb-like GO became more hydrophobic after reduced into 3D structure. Meanwhile, the hydrophobic residues of Hb are prone to exposure on the surface at isoelectric point of Hb in 1 mol L⁻¹ NaCl. So it is much easier for the 3D Ni-rGO composite to combine with Hb. It also could be explained by molecular flattening mechanism that the reduction procedure of GO leads to the formation of more sp² domains,⁴² which promotes the π–π stacking between the heme groups of Hb and the rGO nanosheets contained in the Ni-rGO composite. Besides, many wrinkles on the surface of the stacking graphene sheets were observed from the SEM image. The rougher surface might enhance the protein adsorption.⁴³

Table 1 Comparison of the adsorption capacities of different adsorbents for hemoglobin

Adsorbent	Adsorption capacity (mg g^−1)	Ref.
Poly(C12vim)Br	205.4	40
AP-rGO composite framework	1010	28
Fe3O4@ZIF-8	>6000	24
Polymer-Fe3O4 nanocellulose	248.1	25
N,N-Bis[2-methylbutyl]imidazolium hexafluorophosphate–TiO2	122.3	27
Fe3O4@SiO2@IL	2150	41
CuxOy/ordered mesoporous carbons	1666.7	39
Ni-rGO composite	18 468.6	This work

---

On the other hand, the metal affinity between Hb and the Ni-rGO composite also facilitates the increment of adsorption capacity. The presence of nickel inside the composite has been proved by XPS. Due to the presence of nickel, the composite can provide more free coordination positions for Hb, and the coordination force further improves the adsorption capacity. Chemical interactions between the hydroxyl groups of polysaccharide and metal ions have been reported previously.⁴⁴ In the present work, nickel might combine with hydroxyl groups of peptide chains by chemical interaction, which facilitates more Hb onto the composite. Therefore, an extra-high adsorption capacity of the 3D composite for Hb was achieved.

3.4. Recovery of the retained Hb from the 3D Ni-rGO composite

It is necessary to transfer the adsorbed proteins from the adsorbents into other media for further investigations. In this case, the recovery of the retained Hb from the Ni-rGO composite is very desirable. CTAB is a kind of cationic surfactant, which can compete with Hb to adsorb onto the composite surface. Herein, CTAB in 4 mM B–R buffer solution (pH 11) was used as stripping reagent. In the B–R buffer solution, Hb turns negatively charged because the pH is greater than the pI of Hb, which results in electrostatic repulsion between Hb and the adsorbent. The effect of CTAB concentration on recovery was investigated within the range of 0.1–3.0%, and a recovery of 93.6% was obtained by using 0.3% CTAB in 4 mM B–R buffer solution as eluent.

A conformation change was evaluated for the Hb after adsorption/desorption process by recording circular dichroism (CD) spectra of Hb in 4 mM B–R buffer solution and in the eluate. As shown in Fig. 11, the CD spectrum of the Hb recovered by B–R buffer at pH 11 containing 0.3% CTAB (w/v) is similar to that of the original Hb. Both of the spectra exhibit negative peaks at 212 nm and 220 nm in accordance with the characteristics of α-helical structure of Hb, which is attributed to n–π* transition of the α-helix peptide bond.⁴⁰ This result clearly indicates that the Ni-rGO composite features beneficial biocompatibility, which causes no conformational change of Hb during the adsorption/desorption process. The favorable biocompatibility of this composite shows a great potential in protein sample preparation.

Fig. 11 CD spectra of the original Hb in 4 mM B–R buffer at pH 11 containing 0.3% CTAB (w/v) (a) and the recovered Hb in the same buffer solution (b).

3.5. Isolation of Hb from human whole blood

The Ni-rGO composite was employed to separate Hb from the human whole blood. The human whole blood sample was diluted 400 fold with 4 mM B–R buffer solution at pH 7.0 for further experiment. Then, 1 mL diluted blood sample containing 1.0 mol L⁻¹ NaCl was added into a 1.5 mL centrifuge tube with 2 mg Ni-rGO composite. The supernatant was collected after the tube was shaken and centrifuged. Afterwards, the adsorbed Hb was recovered by using B–R buffer solution with 0.3% CTAB (w/v) at pH 11. At last, the collected supernatant and the recovered eluate were employed for SDS-PAGE assay, as illustrated in Fig. 12. As a comparison, the protein molecular weight standard is presented in lane 1, and 150 mg L⁻¹ Hb standard solution is presented in lane 2. Two obvious bands, which could be attributed to serum albumin (66.4 kDa) and Hb (16.5 kDa), are observed in the 400 fold diluted human whole blood sample (lane 3). The band of Hb completely disappears after the sample was treated by the Ni-rGO composite (lane 4). For the recovered eluate (lane 5), only a single band appears at the same position as the band of the standard Hb. This band is much more intense than that of 150 mg L⁻¹ standard Hb, and almost comparable to that of human whole blood in lane 3. The above results clearly show that Hb can be effectively isolated from the human whole blood by using the Ni-rGO composite as adsorbent.

Fig. 12 SDS-PAGE of Hb isolated from human whole blood. Lane 1, molecular weight standard (Marker in kDa); lane 2, Hb standard solution of 150 mg L⁻¹; lane 3, 400 fold diluted human whole blood without pretreatment; lane 4, 400 fold diluted human whole blood after adsorption by the Ni-rGO composite; lane 5, Hb recovered from the Ni-rGO composite.

4 Conclusions

The 3D nickel-doped reduced graphene oxide composite was synthesized through one-step reduction and self-assembly by ethylenediamine and Ni²⁺, which is a mild, simple and lower-cost one-pot approach. The Ni-rGO composite exhibited a much higher adsorption capacity towards Hb than the adsorbents reported. The high adsorption capacity attributes to the porous structure and wrinkled surface of the 3D composite, the intense hydrophobic interaction between Hb and the Ni-rGO, and the metal affinity between metal nickel and Hb. The Ni-rGO composite was applied to isolate Hb from human whole blood, and the adsorbed Hb could be effectively eluted by CTAB without conformation change. This study provides a new promising sorption medium for hemoglobin isolation from biological samples. The 3D Ni-rGO composite would be an alternative material for highly selective adsorption of biomolecules from complex biological samples, which would expand the application scope of graphene-based materials in the field of separation science.

Acknowledgements

The authors appreciate financial support from the Natural Science Foundation of China (21375012) and the Fundamental Research Funds for the Central Universities (N130105002).

Notes and references

1. K. S. Novoselov, A. K. Geim, S. V. Morozov, D. Jiang, Y. Zhang, S. V. Dubonos, I. V. Grigorieva and A. A. Firsov, Science, 2004, 306, 666.
2. Y. Zhao, C. G. Hu, Y. Hu, H. H. Cheng, G. Q. Shi and L. T. Qu, Angew. Chem., 2012, 124, 11533.
3. J. Wang, J. L. Liu, D. L. Chao, J. X. Yan, J. Y. Lin and Z. X. Shen, Adv. Mater., 2014, 26, 7162.
4. F. Liu and T. S. Seo, Adv. Funct. Mater., 2010, 20, 1930.
5. S. Kabiri, D. N. H. Tran, T. Altalhi and D. Losic, Carbon, 2014, 80, 523.
6. G. Q. Chen, Y. X. Liu, F. Liu and X. Zhang, Appl. Surf. Sci., 2014, 311, 808.
7. C. Li and G. Q. Shi, Nanoscale, 2012, 4, 5549.
8. Y. Q. Sun, Q. Wu and G. Q. Shi, Phys. Chem. Chem. Phys., 2011, 13, 17249.
9. F. Yavari and N. Koratkar, J. Phys. Chem. Lett., 2012, 3, 1746.
10. A. L. Han, S. Jin, H. L. Chen, H. X. Ji, Z. J. Sun and P. W. Du, J. Mater. Chem. A, 2015, 3, 1941.
11. R. H. Wang, C. H. Xu, J. Sun, L. Gao and H. L. Yao, ACS Appl. Mater. Interfaces, 2014, 6, 3427.
12. J. Biener, M. Stadermann, M. Suss, M. A. Worsley, M. M. Biener, K. A. Rose and T. F. Baumann, Energy Environ. Sci., 2011, 4, 656.
13. Y. Q. He, Y. Liu, T. Wu, J. K. Ma, X. R. Wang, Q. J. Gong, W. N. Kong, F. B. Xing, Y. Liu and J. P. Gao, J. Hazard. Mater., 2013, 260, 796.
14. H. P. Cong, X. C. Ren, P. Wang and S. H. Yu, ACS Nano, 2012, 6, 2693.
15. J. Gao, F. Liu, Y. L. Liu, N. Ma, Z. Q. Wang and X. Zhang, Chem. Mater., 2010, 22, 2213.
16. H. Y. Sun, Z. Xu and C. Gao, Adv. Mater., 2013, 25, 2554.
17. W. F. Chen and L. F. Yan, Nanoscale, 2011, 3, 3132.
18. H. Hu, Z. B. Zhao, W. B. Wan, Y. Gogotsi and J. S. Qiu, Adv. Mater., 2013, 25, 2219.
19. J. H. Li, J. Y. Li, H. Meng, S. Y. Xie, B. W. Zhang, L. F. Li, H. J. Ma, J. Y. Zhang and M. Yu, J. Mater. Chem. A, 2014, 2, 2934.
20. Z. Y. Sui, Y. Cui, J. H. Zhu and B. H. Han, ACS Appl. Mater. Interfaces, 2013, 5, 9172.
21. H. Bai, C. Li, X. L. Wang and G. Q. Shi, Chem. Commun., 2010, 46, 2376.
22. D. Barrick, Biochemistry, 1994, 33, 6546.
23. K. D. Egeberg, B. A. Springer, S. A. Martinis, S. G. Sligar, D. Morikis and P. M. Champion, Biochemistry, 1990, 29, 9783.
24. J. N. Zheng, Z. A. Lin, G. Lin, H. H. Yang and L. Zhang, J. Mater. Chem. B, 2015, 3, 2185.
25. T. S. Anirudhan and S. R. Rejeena, Carbohydr. Polym., 2013, 93, 518.
26. Q. Chen, L. M. Shen, J. Xia, X. W. Chen and J. H. Wang, J. Sep. Sci., 2014, 37, 2716.
27. H. Meng, X. W. Chen and J. H. Wang, J. Mater. Chem., 2011, 21, 14857.
28. Y. Zhang, J. W. Liu, X. W. Chen and J. H. Wang, J. Mater. Chem. B, 2015, 3, 983.
29. X. W. Chen, X. Zheng, X. F. Wang and J. H. Wang, Chem. J. Chin. Univ., 2015, 36, 1498.
30. W. S. Hummers and R. E. Offeman, J. Am. Chem. Soc., 1958, 80, 1339.
31. X. Jiang, Y. W. Ma, J. J. Li, Q. L. Fan and W. Huang, J. Phys. Chem. C, 2010, 114, 22462.
32. J. L. Zhang, H. J. Yang, G. X. Shen, P. Cheng, J. Y. Zhang and S. W. Guo, Chem. Commun., 2010, 46, 1112.
33. S. Stankovich, D. A. Dikin, R. D. Piner, K. A. Kohlhaas, A. Kleinhammes, Y. Y. Jia, Y. Wu, S. T. Nguyen and R. S. Ruoff, Carbon, 2007, 45, 1558.
34. C. X. Chi, H. B. Xu, K. Zhang, Y. B. Wang, S. H. Zhang, X. S. Liu, X. Liu, J. P. Zhao and Y. Li, Mater. Sci. Eng., B, 2015, 194, 62.
35. H. Bai, C. Li, X. Wang and G. Q. Shi, J. Phys. Chem. C, 2011, 115, 5545.
36. R. K. Rath and S. Subramanian, Miner. Eng., 1997, 10, 1405.
37. J. W. Liu, Y. Zhang, X. W. Chen and J. H. Wang, ACS Appl. Mater. Interfaces, 2014, 6, 10196.
38. J. W. Liu, Q. Zhang, X. W. Chen and J. H. Wang, Chem.–Eur. J., 2011, 17, 4864.
39. Y. Zhang, L. G. Xing, X. W. Chen and J. H. Wang, ACS Appl. Mater. Interfaces, 2015, 7, 5116.
40. X. F. Wang, Y. Zhang, Y. Shu, X. W. Chen and J. H. Wang, RSC Adv., 2015, 5, 31496.
41. Y. Wei, Y. Li, A. L. Tian, Y. T. Fan and X. Wang, J. Mater. Chem. B, 2013, 1, 2066.
42. L. Z. Zheng, D. Ye, L. Y. Xiong, J. P. Xu, K. Tao, Z. J. Zou, D. L. Huang, X. W. Kang, S. M. Yang and J. Xia, Anal. Chim. Acta, 2013, 768, 69.
43. K. Rechendorff, M. B. Hovgaard, M. Foss, V. P. Zhdanov and F. Besenbacher, Langmuir, 2006, 22, 10885.
44. L. O. Filippov, V. V. Severov and I. V. Filippova, J. Miner. Process., 2013, 123, 120.

---