Although much is known about this enzymatic process, implementation is significantly hampered due to the inherent recalcitrance of biomass to enzymatic degradation, requiring high enzyme loading and thus driving up costs within the biorefinery. This resistance to enzymatic degradation is caused by challenges related to the structural phenolic polymer lignin which limits enzyme accessibility and non-productively binds CAZymes, as well as low overall activity of the soluble enzymes on a highly insoluble, crystalline substrate. Limited accessibility to polysaccharides is typically alleviated by thermochemical pretreatment of biomass, but this does not assist in overcoming the challenges related to CAZyme functionality. Thus, CAZymes and their appended carbohydrate binding modules (CBMs) must be engineered for improved hydrolytic activity.
Lignin and cellulose both possess slight negative charges on their surfaces post pretreatment, and exploiting these electrostatics using enzyme supercharging is one promising route for improving enzyme activity. Supercharging is a rational design technique where several surface exposed amino acid residues are iteratively mutated to either positive (R, K) or negative (D, E) charged residues to produce high theoretical net charges on the enzyme’s surface. We hypothesize that supercharging can be applied to both CAZyme catalytic domains (CD) and their appended binding modules to tune enzymes surfaces to a critical net charge where binding affinity, and ultimately activity, are optimized. This hypothesis is based on the Sabatier principle which has often been applied to inorganic heterogenous catalysis and states that the binding affinity between catalyst and substrate should be neither too strong nor too weak. At the crux of this principle is the Sabatier Optimum where binding affinity is optimized to an intermediate strength, resulting in maximal catalytic turnover. Electrostatic interactions play a key role in modulating binding of CAZymes to cellulosic substrates. Thus, utilizing supercharging, some critical net charge can be reached to produce intermediate strength binding interactions between CAZyme and biomass optimizing catalysis.
This supercharging process was first applied to an industrially relevant endocellulase enzyme Cel5A and its family 2a CBM from the thermophilic microbe Thermobifida fusca. A combinatorial library of 33 mutant constructs containing different CBM and Cel5A designs spanning a net charge range of -52 to 37 was constructed using Rosetta macromolecular modeling software. Activity for all mutants was characterized from the soluble cell lysates after expression in E. coli and promising single mutant constructs (containing mutations either on the CBM, Cel5A catalytic domain, or both CBM and Cel5A domains) were then purified and systematically characterized. Remarkably, we found that often endocellulases with mutations on the CBM domain alone resulted in improved activity on cellulosic biomass, with three top-performing supercharged CBM mutants exhibiting between 2–5-fold increase in specific activity, compared to native enzyme, on pretreated lignocellulosic biomass enriched in lignin (i.e., corn stover) as well as isolated crystalline cellulose. Pull down binding assays were further conducted on top performing mutants to identify a strong correlation between CBM net charge and cellulose binding affinity (KA). Furthermore, we were able to clearly demonstrate that endocellulase net charge can be selectively fine-tuned using our proposed protein supercharging protocol for targeting distinct substrates and maximizing biocatalytic activity by several folds. Additionally, some supercharged CBM containing endocellulases exhibited a 5 °C increase in optimal hydrolysis temperature, compared to the native enzyme, which enabled further increase in hydrolytic yield at higher operational reaction temperatures.
Based on the success with the endocellulase Cel5A, this workflow was expanded to an exocellulase partner Cel6B and its native family 2a CBM (CBM2a) from the same microbial source. Once again, Rosetta macromolecular modelling software was used to computationally design a combinatorial library of 33 constructs, this time spanning a net charge range of -84 to 2. Cel6B is naturally much more negatively charged compared to Cel5A, resulting in the low theoretical positive net charge for this library. Once again, the activity of the entire library was screened from soluble E. coli cell lysates, and several constructs display increased activity with some near a 2-fold improvement compared to the native enzyme. These constructs that displayed higher activity were expressed and purified on a large scale for similar characterization as was done with Cel5A, as well as to understand the nuances of synergism between supercharged exo-endo partners.
Overall, this work demonstrates the first successful implementation of enzyme supercharging to improve the catalytic activity of cellulases. These improved constructs are directly relevant for industrial implementation. Future work is focused on expanding this supercharging technique to hemicellulolytic enzymes in order to construct a synergistic enzymatic cocktail for biomass deconstruction.