Hierarchically encapsulating enzymes with multi-shelled metal-organic frameworks for tandem biocatalytic reactions

Biocatalytic transformations in living organisms, such as multi-enzyme catalytic cascades, proceed in different cellular membrane-compartmentalized organelles with high efficiency. Nevertheless, it remains challenging to mimicking biocatalytic cascade processes in natural systems. Herein, we demonstrate that multi-shelled metal-organic frameworks (MOFs) can be used as a hierarchical scaffold to spatially organize enzymes on nanoscale to enhance cascade catalytic efficiency. Encapsulating multi-enzymes with multi-shelled MOFs by epitaxial shell-by-shell overgrowth leads to 5.8~13.5-fold enhancements in catalytic efficiencies compared with free enzymes in solution. Importantly, multi-shelled MOFs can act as a multi-spatial-compartmental nanoreactor that allows physically compartmentalize multiple enzymes in a single MOF nanoparticle for operating incompatible tandem biocatalytic reaction in one pot. Additionally, we use nanoscale Fourier transform infrared (nano-FTIR) spectroscopy to resolve nanoscale heterogeneity of vibrational activity associated to enzymes encapsulated in multi-shelled MOFs. Furthermore, multi-shelled MOFs enable facile control of multi-enzyme positions according to specific tandem reaction routes, in which close positioning of enzyme-1-loaded and enzyme-2-loaded shells along the inner-to-outer shells could effectively facilitate mass transportation to promote efficient tandem biocatalytic reaction. This work is anticipated to shed new light on designing efficient multi-enzyme catalytic cascades to encourage applications in many chemical and pharmaceutical industrial processes. Mimicking multi-enzyme catalytic cascades in natural systems with spatial organization in confined structures is gaining increasing attention in the emerging field of systems chemistry. Here, the authors demonstrate that multi-shelled metal-organic frameworks can be used as a hierarchical scaffold to spatially organize enzymes on nanoscale to enhance cascade catalytic efficiency.

NAD + -dependent ADH uses the interconversion of NAD + /NADH redox couple to catalyze the oxidation of alcohol to aldehyde. ADH exists as a dimer (that is, composed of two polypeptides), with each monomer containing a catalytic domain (that is catalytic zinc, which holds hydroxyl group on alcohol) and a coenzyme binding domain with a large cleft between the two. The active site is at the bottom of the cleft. According to previous studies [4][5][6] , the mechanism of NAD + -dependent ADH enzyme for the oxidation of alcohol to aldehyde is described as follows (summarized in Supplementary Fig. 34): First, the ADH structure is initially open to facilitate access to the active site for NAD + binding. When NAD + and alcohol bind, ADH undergoes a global conformational change, which involves a rotation of the catalytic zinc domain relative to the NAD + binding domain. This process closes up and isolates the active site from solvent, creating a hydrophobic environment for the productive holoenzyme complex (that is, E → E-NAD + → E*-NAD + → E*-NAD + -RCH 2 OH).
Second, the resulting E-NAD + -RCH 2 OH complex is poised for hydrogen transfer, involving alcohol to deprotonate, and transfer the proton via Ser-48 and His-51 (His-51 contacts solvent water on the protein surface) to solvent. During the proton relay, the reduced nicotinamide ring may become puckered. This process leads to the formation of E-NADH-RCHO.
Third, NADH and aldehyde dissociate from the abortive E-NADH-RCHO complex. Previous kinetics studies have shown that the dissociation of NADH is the rate-limiting step when ethanol is used as the substrate.
Overall, ADH relies on coenzyme dissociation and association that involves conformational changes and interconversion of NAD + /NADH redox couple to catalyze the oxidation of alcohol to aldehyde. Fig. 34 The mechanism of NAD + -dependent ADH enzyme for the oxidation of alcohol to aldehyde 5 .  Fig. 36 Catalytic efficiencies of ADH/NAD + in the absence (red) and in the presence (blue) of cobalt hydroxide. Note that the concentration of cobalt hydroxide was 1.1 mg mL -1 , which was calculated by assuming that ZIF-67 in Pro@ZIF-8@ADH/NAD + @ZIF-67@ZIF-8 was completely dissociated; the concentrations of ADH and NAD + were 19.3 μg mL -1 and 17.1 μg mL -1 , which was the same with respective enzyme concentrations in all control experiments of Pro-ADH/NAD + cascade.

Supplementary Fig. 37
Comparison of long-term stability. Cascade activities of the Pro@ZIF-8@ADH/NAD + @ysZIF-8 and supernatant solution of Pro@ZIF-8@ADH/NAD + @ysZIF-8 immediately after fresh preparation (red) and after storage at room temperature for a time-interval of 10 day (blue), respectively. Supplementary Fig. 41 The amine-reactive fluorophores are acylating reagents that form thioureas or carboxamides upon reaction with amino groups of enzymes. a) Reaction of a primary amine with an isothiocyanate in FITC. b) Reaction of a primary amine with an isothiocyanate in Rhodamine B isothiocyanate. c) Reaction of a primary amine with a succinimidyl ester in 7-hydroxycoumarin-3carboxylic acid N-succinimidyl ester.

Supplementary Tables
Supplementary Table 1   On the basis of enzyme loading characterization results and the final product weight, we determined the loadings of GOx and HRP to be 70.72 and 215.9 μg mg -1 ; and weight percentage of GOx, HRP, and ZIF-8 in the final product to be 7.1 wt%, 21.6 wt%, and 71.3 wt%. Through the following equation, the weight loss in GOx@ZIF-8@HRP@ZIF-8 can be estimated as 67.9 wt%, which is similar to the corresponding measured weight loss in TGA curve (72.4 wt%). Based on the above results, we can conclude that the weight loss in TGA matches with the enzymes loading.
weight loss GOx@ZIF-8@HRP@ZIF-8 % On the basis of enzyme loading characterization results and the final product weight, we determined the loadings of Pro, ADH, and NAD + to be 53.49, 193.0, and 170.9 μg mg -1 ; weight percentage of Pro, ADH, NAD + , and ZIF-8@ysZIF-8 in the final product to be 5.6 wt%, 19.3 wt%, 17.1 wt%, and 58.2 wt%. Through the following equation, the weight loss in Pro@ZIF-8@ADH/NAD + @ysZIF-8 can be estimated as 77.3 wt%, which is similar to the corresponding measured weight loss in TGA curve (72.4 wt%). Based on the above results, we can conclude that the weight loss in TGA matches with the enzymes loading.