Novel glycoprotease discovery and characterisation for enhanced detergent performance

Many secreted animal proteins are heavily glycosylated and are known as glycoproteins (GPs) - they are decorated with carbohydrate chains covalently linked to the protein backbone, either via N- or O-linkages. The type and variety of glycosylation varies significantly, but the key point is that the glycosylation can limit access of classical proteases to allow breakdown of the GP.
Automatic dishwashing (ADW) failures from animal-derived foods such as egg, milk, cheese, and meat are amongst the most frustrating for consumers, as they form stains that are difficult to remove, even with state-of-the-art protease-containing detergents. These foodstuffs are rich in glycoproteins such as albumin, ovomucin, lactoferrin and glycosaminoglycans, which might contribute to stain recalcitrance. Recently discovered glycoproteases (by the academic supervisory team and others; that specifically recognise glycans and cleave peptide/protein components of GPs, provide an exciting potential solution as these enzymes could remove GPs in animal-based food stains more easily and effectively (e.g., less time and/or lower enzyme load) than current P&G ADW proteases
(1) In the first instance we will use our human gut microbiota library, currently comprising tens of different species from the main bacterial phyla of the gut, to assess the potential of the microbiota to use defined glycoproteins or
stains as a sole carbon source and use proteomics to analyse the enzymatic apparatus upregulated and identify potential targets. The power of this approach is that the human gut microbiota has adapted to degrade the same food molecules that cause problem stains in ADW failures so should provide a rich resource for novel glycoproteases targeting GPs in foods.
(2) As an alternative and complementary approach, we will use advanced bioinformatics techniques to identify predicted proteases with associated carbohydrate binding modules (CBMs). CBMs are discrete non-catalytic modules that bind complex glycans and are normally found as components of carbohydrate-active enzymes (CAZymes) such as glycoside hydrolases. The presence of CBMs within non-carbohydrate active enzymes suggests the attached enzyme targets glyco-conjugates, and initial analysis by R. Hirt of predicted peptidase/proteases in the MEROPS protease database which have CBMs, suggests large numbers of uncharacterised glycoproteases exist. The feasibility of this approach is evidenced by the fact that this was how the first characterised glycoprotease was discovered by Bolam & Hirt. To narrow down the number of novel glycoproteases to test we will select predicted proteases from MEROPS with CBMs from families that are known to bind to animal-derived glycans (e.g., CBM32 or CBM51) as these enzymes will more likely target animal-derived glycoproteins such as those from eggs, cheese, and meat.
Novel glycoproteases discovered by both approaches will be tested in recombinant form for activity against a range of relevant animal glycoproteins (e.g., egg albumin, bovine lactoferrin, proteoglycans) in the academic lab, as well as problem food stains at P&G. Assays will involve initial screening by assessing GP breakdown on SDS-PAGE gels, followed by mass spectrometry analysis to identify the protease target sites. Experimental and AI based structural techniques (crystallography and Alpha-Fold) will be used to investigate substrate recognition. CBM specificity will be analysed using isothermal titration calorimetry. In addition, we will test the glycoproteases in conjunction with existing P&G ADW proteases to quantify their impact on cleaning i.e., not replacement but improved cleaning of problem stains in less time and/or with lower enzyme load. Further testing will also look at cooked vs. uncooked stains, with most cooked soils showing differences as to how proteins are presented to the enzymes (i.e., likely denatured in cooked).