Positional assembly of hemin and gold nanoparticles in graphene–mesoporous silica nanohybrids for tandem catalysis† †Electronic supplementary information (ESI) available. See DOI: 10.1039/c4sc02714k Click here for additional data file.

A hybrid catalyst in which two different types of enzyme mimics are positioned in spatially separate domains within a graphene–mesoporous silica support is presented.


Introduction
In a natural environment, enzymes are almost always spatially conned in crowded and tightly controlled cellular compartments, which can isolate the catalytic cycle, prevent interference and make biomolecular catalysts more efficient. 1 In order to mimic the natural compartmentalization process, researchers have long directed their attention to enzyme encapsulation or assembly. 2 Until now, much effort has been focused on using phospholipid liposomes or polymersomes as synthetic nano-or micro-capsules. 3 Furthermore, to be efficient, the biomolecular catalysts need not only to be presented in a conned reaction space but also positioned at specic sites within subcellular organelles. 3 To this end, van Hest and collaborators constructed a variety of biohybrid polymersome nanoreactors in which two or more different enzymes were spatially positioned and precisely ordered. 1a,4 Very recently, through loading different enzyme-containing organelle mimics inside larger polymersomes, they have even successfully created a structural and functional eukaryotic cell mimic. 5 Recently, Lu et al. demonstrated a promising approach by assembling and encapsulating enzymes within a thin polymer shell to form biomimetic enzyme nanocomplexes with precise compositional and spatial controls. 6 On the other hand, using synthetic systems to simulate the function of natural enzymes has attracted increasing attention for the last few decades. 7 Among the countless examples arising from these efforts, catalytically active nanomaterials as a new generation of articial enzymes are particularly impressive and lead to new opportunities in biomedical diagnosis, environmental monitoring, and therapeutics. 8 Until now, researchers have discovered a number of nano-sized materials that possess unique enzyme-mimicking activities, such as CeO 2 , 9 Fe 3 O 4 , 10 gold nanoparticles (AuNPs), 11 V 2 O 5 , 12 PtPd-Fe 3 O 4 , 13 graphene oxide 14 and graphene nanocomposites. 15 Nevertheless, creating such "static" articial enzymes is not sufficient to mimic smart enzymatic systems, just like simply combining individual biomolecules (e.g. protein, nucleic acid and lipid) together is not enough to construct a functional cell. 16 Recently, through the integration of articial enzymes with natural enzymes, catalytic ensembles with synergic and complementary functions have been achieved. 15b,17 Such studies take one important step towards mimicking complex natural systems. To mimic nature more completely, it would be desirable not only to explore "static" articial enzymes, create catalytic ensembles or design functional enzyme complexes with a high level of control over positional assembly, but also to position different types of articial enzymes (or prosthetic groups) in separate domains.
Herein, we describe the rational design of robust articial enzyme nanocomplexes to achieve this aim, as shown in Fig. 1. Specically, a graphene-mesoporous silica hybrid (GS) was used as a nanocontainer to anchor two articial enzymes (i.e. AuNPs as a glucose oxidase (GOx) mimic and hemin as a prosthetic group to mimic peroxidase) at different locations, namely, on the outer surface of coated silica and on the inner surface of exposed graphene, respectively. This environment allows the simple design of an articial enzymatic reaction system in which AuNPs and hemin can work in tandem catalysis. To the best of our knowledge, this is the rst example of the integration of multiple biomimetic catalysts through a controlled spatial positioning procedure. Meanwhile, our new ndings might pave the way to applying articial tandem catalytic systems for articially mimicking organelles or important chemical transformations.

Results and discussion
Synthesis of GSHA Fig. 1A illustrates the basic procedure using the GS as a scaffold for the precise positional assembly of hemin and AuNPs to form GS-hemin-AuNPs nanohybrids (GSHA). The structure of hemin and the corresponding TEM images are shown in Fig. 1B and C. Firstly, the 2D sandwich-like GS was prepared by Feng and Müllen's recently developed method with a slight modication. 18 Briey, the cationic surfactant, cetyltrimethylammonium bromide (CTAB), electrostatically adsorbed and self-assembled onto the surface of highly negatively charged graphene oxide (GO) in alkaline solution. Upon the hydrolysis of tetraethyl orthosilicate (TEOS), hydrazine reduction treatment, surface functionalization with amine groups (i.e. APTES treatment) and so-template removal, the GS products were successfully collected with mesoporous silica around the surface of singlelayer graphene. The resulting GS products were studied by transmission electron microscopy (TEM) imaging. As seen in Fig. 1C and S1, † as-prepared GS sheets with a morphology similar to that of graphene and a mesoporous structure were observed. 18 Additionally, atomic force microscopy (AFM) was conducted to further demonstrate the structural features of the GS sheets (Fig. S2 †). Then, as a at molecule, hemin (a wellknown natural metalloporphyrin) could be assembled onto the surface of the exposed graphene to form GS-hemin (GSH) through p-p stacking interactions. 15 The conjugation experiment between hemin and the GS was carried out in methanol solution, as hemin is monomeric under this condition (Fig. S3a †). Although it is difficult to distinguish between the GS and GSH using TEM (Fig. 1C), the attachment of hemin on the exposed graphene surface could be characterized by UV/vis absorption spectroscopy. As shown in Fig. S3b, † an absorption maximum at 418 nm was also observed aer reduction, which claried that the hemin molecules were attached to the exposed graphene. 15a,b The next step was to adsorb AuCl 4 À to the NH 2group-rich silica surface of GSH via electronic interactions. Aer that, highly dispersed AuNPs could be formed on the silica surface (GSHA) by in situ reduction of auric chloride ions with NaBH 4 , since the functional NH 2 groups present could serve as a stabilizing agent by providing an anchoring surface. 19 The nature of the GSHA structure was studied by TEM and highangle annular dark-eld scanning TEM (HAADF-STEM) and elemental mapping ( Fig. 1C and 2). Uniform distribution of N, O, Si, Au and Fe in the same graphene support was observed, which indicated that the AuNPs and hemin had been coimmobilized into the same GS support. Such a hybrid catalyst based on heterogeneous materials contained different catalytic species, which were expected to possess multiple enzyme-like  activities. Similarly, GS-AuNPs nanohybrids (GSA) and GS-AuNPs-hemin nanohybrids (GSAH) were also prepared (Scheme S1, S2 †). Characterizations of the resulting GSA and GSAH are described in detail in the ESI (Fig. S4 and S5 †). Compared to anchoring the AuNPs prior to the introduction of hemin (GSAH), the rst way, creating a catalytic GSHA ensemble, was thought to be more practical, as it was demonstrated that hemin could also adsorb onto the gold surface during the hemin assembly process (Fig. S6 †).
The peroxidase-like catalytic activity of GSH To demonstrate the proof of principle, the peroxidase-like catalytic activity of GSH was rst investigated under different conditions (Fig. 3), compared with that of hemin alone or GSA. Because of molecular aggregation and oxidative destruction, free hemin itself is generally inactive as a catalyst. Aer the assembly process, the adsorbed hemin species on the exposed graphene surface of the GS is monomeric and can function as a highly effective catalyst in various biomimetic oxidation reactions. For instance, like horseradish peroxidase (HRP) (Fig. S7 †), it can catalyze the reaction of the peroxidase substrate 3,3,5,5-tetramethylbenzidine (TMB) in the presence of H 2 O 2 (Fig. 3a). The TMB oxidation pathways by GSH/H 2 O 2 can be described as shown in Fig. 3b. As expected, GSH had high catalytic activity, whereas free hemin showed little activity at the same hemin concentration (Fig. S8 †). The oxidation of TMB produced a blue color with major absorbance peaks at 370 and 652 nm (Fig. 3c). Aer incubation of sulfuric acid, the reaction was stopped, and the blue color changed to yellow with maximum absorbance at 450 nm (Fig. 3c). Control studies indicated that neither H 2 O 2 nor GSH alone could efficiently oxidize TMB (Fig. 3d). Meanwhile, the ability of GSH/H 2 O 2 to oxidize TMB was dependent on catalyst concentration (Fig. 3e) and pH (Fig. S9 †).
The glucose oxidase-mimic activity of GSA Next, we systematically evaluated the glucose oxidase-mimic activity of GSA in solution (Fig. 4). As unsupported AuNPs have recently been found to exhibit intrinsic GOx-like activity, 11a-d we reasoned that the "naked" AuNPs supported on the GS could serve as a more effective GOx mimic. Like GOx (Fig. S10 †), GSA could catalyze the oxidation of glucose by means of molecular oxygen (in equilibrium with air), yielding gluconic acid and H 2 O 2 (Fig. 4a). The reaction solution was studied with a gluconic acid-specic colorimetric assay. 11a Upon the addition of hydroxamine and Fe 3+ , the color of the solution turned red with a characteristic absorbance peak at 505 nm ( Fig. 4b and c), which suggested that gluconic acid was indeed produced in this GSA-catalyzed reaction. The solutions containing glucose or GSA alone could not introduce any color change. However,  control experiments indicated that GSH without AuNPs and citrate-capped AuNPs (13 nm) had very little activity. This is because the GS support helps the formation of a high degree of ultrane AuNPs (Fig. S4 †). As a result, a larger fraction of active metal atoms are exposed to the surface, and thus these very small and stable AuNPs possess highly enhanced catalytic activity. 11a In addition, since gluconic acid is one of the organic acids, we reason that its production in the reaction can also decrease the ambient pH. To further conrm the reaction product, we used methyl red as a pH indicator (red in pH under 4.4 and yellow in pH over 6.2) and a pH meter to monitor the pH change of the solution (Fig. 4d and e). All the above results conrmed that GSA can act as a more effective GOx mimic than unsupported AuNPs. 11a-d

GSHA-catalyzed two-step reaction
So far, many studies have been reported in the literature with the objective of mimicking natural enzyme architectures. In terms of these studies, there are basically two major aspects concerning the construction of synthetic systems, namely (1) using enzyme encapsulation or assembly to mimic the natural compartmentalisation process (Fig. 5a, Route 1); 1-3,5 such a strategy has been developed to encapsulate and position different types of natural enzymes in separate domains, and (2) exploring articial enzymes that mimic the function of natural enzymes (Fig. 5a, Route 2); 8-15 researchers have recently directed their attention to the construction of catalytic ensembles for mimicking complex enzymatic systems. 15b,17 However, there has been no report of articial enzyme-loaded nanodevices with a high level of control over positional assembly for mimetic tandem catalysis. Based on the enzyme-mimicking activities of GSH and GSA, we expected that the integrated GSHA could function as a hybrid catalyst that could drive a two-step reaction to allow for in situ generation of H 2 O 2 for the oxidation of the peroxidase substrate TMB (Fig. 5b). Initially, the enzyme-like activities of GSHA were tested separately. GSHA could catalyze the oxidation reaction of both glucose and TMB (Fig. S12 †). In contrast, GSA or GSH could only possess one of the enzyme-like activities under our experimental conditions. These above experimental results demonstrated that GSHA could exhibit dual enzyme-mimicking activities. Inspired by these unique features, we further pieced them together to catalyze a two-step reaction, usually catalyzed by GOx and HRP. That is, GSHA rst catalyzed the glucose oxidation reaction to yield gluconic acid and H 2 O 2 , and then oxidized TMB resulting in the formation of a colored product of oxTMB [eqn (1)]: As shown in Fig. 5c and S13, † similar with the oxidaseperoxidase coupled enzyme system, GSHA could produce a blue color due to the oxidation of TMB by H 2 O 2 , indicating that the entire reaction was operational in solution. Control experiments conrmed that neither GSA nor GSH could catalyze the TMB oxidation reaction (Fig. 5c and S13 †), unless their catalytic reactions were coupled with enzymatic catalysis (Fig. S14 and S15 †). More importantly, positional assembly of hemin and AuNPs in spatially separate domains (GSHA) shows a clear advantage over other similar systems without such an assembly ( Fig. 5c and S13 †). For example, GSAH almost cannot catalyze this two-step cascade reaction ( Fig. 5c and S13 †), as the hemin adsorbed onto the gold surface can inhibit the GOx-like activity of GSAH. Taken together, articial enzyme complexes with signicantly improved compositional and spatial controls were developed for realizing more complex functions.

Conclusion
In conclusion, we have demonstrated that it is possible to nd a suitable support and to entrap hemin molecules and AuNPs within the support in a controlled way. Addition of glucose to a dispersion of such a multifunctional hybrid catalyst resulted in tandem catalysis, which was commonly completed by the oxidase-peroxidase coupled enzyme system. Overall, our studies show a general strategy to position articial enzymes with different functions into a unique support, which holds great promise for designing other hybrid catalysts with a high level of control. This will be important for the future development of catalytic ensembles that can function as "articial organelles" or enable important chemical transformations not otherwise readily possible.