What enables the activity at all? The gold matrix is porous with a hierarchical structure. The smallest building blocks are gold nanocrystals of around 16 nm (determined from XRD analysis13), which are tightly aggregated into sub-micron structures, which are further aggregated into micron-size particles, as can be seen in the HR-SEM imaging in Fig. 5. The surface area is ~ 30 m2/gr, which is typical for porous materials, the porosity of which is interstitial. This porous structure allows diffusion of substrate molecules to the buried enzyme molecules, and diffusion of product molecules out, but the tight aggregation around the entrapped enzyme molecules prevent their leaching out (there is zero activity of the supernatant solutions). That property of efficient entrapment on one hand, and free molecular diffusion on the other hand is possible because the cage walls are perforated with interstitial pores which are too small for the enzyme to leach out, as illustrated in Fig. 6 (roughly, with an Au nanocrystallite average size of 16 nm, one gets interstitial pores to be around 4 nm, while GOx diameter is around 8 nm29). Adding to this physical entrapment property are the strong interactions of the enzyme with gold and CTAB, detailed below. The known cost of entrapment within a porous matrix is the diffusional limitation which slows the reaction rate, but then the gains are the enhanced stability (Fig. 3), the recyclability (Fig. 4), the ability to construct a device, and the prospect of modifying the activity.
Widening the scope of enzymatic activity: In general, affecting and widening of enzyme activity has been achieved by two types of approach: Tampering with the primary structure of the protein by various enzyme-engineering and mutation methods, which, from that point of view can be considered as forming new enzymes30–32; or, affecting the tertiary and quaternary conformational structure of the original native enzyme, leaving the sequence of amino acids untouched. Such conformational changes have been induced mainly by adsorptive interactions33,34. The main target of both approaches has been, as expected, the active site32,34,35, but far less attention has been devoted to affecting the channel leading to the active site. The rationale in approaching the (dynamic) channel is that it acts like a highly stereoselective separation column, allowing entrance to the active site only for a substrate that fits the stereoselective screening. Thus, changing the conformation of the channel – particularly widening that entrance - may lead to alteration of the selectivity; this, we propose below, is the main mechanism that explains the observations of this report. Affecting the channel was reported by enzyme engineering methods36, but affecting it by conformational changes that do not involve changes in the primary, to the best of our knowledge, has not been reported.
The relevant interactions in the GOx-CTAB-gold system
The conformational changes that GOx undergoes in its entrapped state are due to three types of powerful interactions, each of which has been separately well documented, operating all together in the GOx/[email protected] system. The first is the interaction of proteins with gold. These interactions are with functional groups of most amino acids on the outer surface of the protein37,38, and particularly with the SH groups of the cysteine residues39, with the S of methionine side chain40, and with the exposed S-S moieties of cystine41,42. Furthermore, since the buffered pH of the entrapped GOx – pH 5.1 - is higher than its pI (4.243) rendering the enzymes interface negatively charged, one should also consider interactions with the anionic aspartate and glutamate37.
The second interaction to consider is that of CTAB with gold. It is a well-known strong interaction44,45, which has been used, for instance, to control the morphology and shape of gold nano-structures45–49. A special feature of the CTAB-gold interaction, first proposed by El-Sayed et al45 and confirmed by numerous subsequent studies, is the formation of an adsorbed CTAB double layer on the gold surface45,50,51. The proposed structure of that bilayer is the following (Fig. 6, right): In the first layer the cationic (Me3)R-N+ faces the gold surface through a bromide bridge50: Au --- Br−--- N+. Supporting that proposition is that fact that the bromide anion is known to adsorb strongly on gold (particularly on 111 planes but also on lower index plains49, and it has also strong attraction to CTA+ 49. The cetyl chain of CTAB, perpendicular to the gold surface, accepts the second layer through cetyl --- cetyl hydrophobic interactions and a double-layer forms, much like in liposomes51: Au—Br−---N+-R—R---N+. That second layer terminal N+ of CTA+ is then free to interact with the negatively charged protein52,53, releasing the bromide. We assume that the bilayer structure is not perfect inside the gold cage, and that there are also isolated CTAB molecules adsorbed on the gold surface which do not take part in a bilayer – these may add hydrophobic interactions between the alkyl tail of the CTAB and the residues of the hydrophobic amino acids on the surface of the protein54,55, such as valine and leucine 55. That holding of the enzyme inside the cage more rigidly by these combined types of interactions, is expressed by the major increase in the thermal stability (Fig. 3), and by the ability to recycle the entrapped enzyme (Fig. 4). But how do these interactions also affect the conversion of GOx into a general oxidase? This is discussed next:
The proposed induced conformational change
GOx is a homodimeric oligomer of which the two monomeric units are held by non-covalent bonds56. As common in many homodimeric enzymes, the interface between these two units forms the channel leading to the two active sites, a channel that evolved also to act as a stereoselective filter for substrates which can enter and reach the active site57,58. Unlike conformational changes which occur by 2D adsorption, which is a non-isotropic process, the 3D gold cage in our case, pulls apart the protein in an isotropic manner, that is, in all directions (Fig. 6). Since the interface between the two monomeric units is not held together by strong covalent bonds, it constitutes an initial “crack” that already exists in the enzyme, and which can be further affected. Thus, we propose that the combined pull of the direct protein-gold interaction and the gold-CTAB double layer interactions, opens that crack to a degree that the channel loses its evolutionary built strict stereoselective gate-keeper property, tailored for D-glucose exclusively. This, in turn, allows all substrates described in this report, to diffuse to the two active sites and reach them.
Further confirmation to this proposed mechanism comes from the comparison of the activity of GOx when entrapped with or without entrapment CTAB (a comparison possible for the monosaccharides which can be entrapped without CTAB). It is seen – Fig.’s 3,4 and Table 2 – that GOx/[email protected] greatly outperforms [email protected] in any parameter, including even the performance on the native D-glucose. This comparison allows shows the two types of pulling effects – with pure gold, and through CTAB. In fact, that additional pull, overshadows the pull of the gold itself, as seen in Table 2. The observation that the partial widening of activity with gold only is generalized by co-entrapping CTAB, strengthens the proposed conformational opening the access to the active site. The focus on the channel is also highlighted by discussing what do these observations mean from the point of view of the active site – next.
The active site: Our observations indicate that once the access to the active site is opened to saccharides other than D-glucose, the active site is capable of oxidizing non-specifically other saccharides. That is, our observations indicate that selectivity of GOx is not a sole feature of the active site. Furthermore, these observations also suggest an interpretation that the active site of GOx (the peptide fragment containing Glu412, His516 and His55959) resembles an early-evolutionary preserved structure of this biocatalyst, and that the strict stereospecificity evolved over the ages with the build-up of the encompassing protein and its dimerization, to form the exclusive activity towards D-glucose of the modern enzyme we know. That non-specificity of the active site towards the stereochemistry of the saccharide CH-OH is particularly evident by the ability of the entrapped GOx to oxidize methyl glucoside and sucrose, two molecules in which the D-glucose analogous cyclic β-C(1)H-OH is blocked. That free GOx is capable of oxidizing methyl glucoside at all, utilizing air oxygen and releasing H2O2 is evident from the low but detectable activity in solution – Fig. 7. The efficient activity of the exposed active site is also compatible with the observation - Table 1 – that the higher KM values observed of methyl glucoside, raffinose, sucrose and glucose-6-phosphate, are accompanied by higher Vmax values: The lower affinity (higher KM) means shorter residence time at the active site, and that increases Vmax if the oxidation step is fast.
In conclusion, we have been able to convert GOx into a general sugar oxidase, including saccharides which lack natural specific oxidases. High-specificity and wide scope activity of enzymes are two desirable properties, which are complementary, and answer different needs in biotechnology, medicine, pharmaceutics, and enzymatic devices. Enzymes in their native form and environments usually answer the high-specificity requirement. And since this is the starting point, widening the scope of activity requires manipulation of the native enzyme. Here we have shown how adsorptive-induced conformational changes can be utilized for that purpose. As GOx is common and robust, and since most of the saccharides employed in this study are of high-volume industrial use but lack specific oxidases, our study opens new directions to be considered. For instance, one could envisage that GOx/air/D-glucose fuel cells60 can now be generalized to other common sugars or their mixtures, using GOx only61. An added advantage of the entrapment in gold developed in this study is that the protective heterogenization, renders the use of the general oxidative GOx easier to implement in the variety of potential uses.