Exploring the structure-function landscape of a multimeric family of glycoside hydrolases

Mycobacteria such as M. tuberculosis have a complex cell wall, including glycans and lipoglycans which contain an unusual sugar; D-arabinose. This sugar is rarely found outside mycobacteria and their relatives. We have identified a new family of glycoside hydrolase enzymes (GH172s) which are able to degrade this cell wall. Several of these enzymes are from gut bacteria which are proficient in degradation of complex glycans in the human distal gut, but we have also identified homologs of these enzymes in human and animal mycobacterial pathogens. Some family members are also able to target fructo-oligosaccharides, a common component of the human diet. D-arabinose containing glycans are essential to mycobacterial viability, indeed, their synthetic machinery is the target of a front-line anti-tuberculosis drug, Ethambutol. Understanding the specificity of D-arabinan targeting enzymes from pathogenic mycobacteria could provide future drug targets, and transform our understanding of how mycobacteria may remodel or recycle their cell wall. A toolbox of characterised gut bacterial D-arabinan degrading enzymes will be of use to mycobacterial cell wall research, diagnostics, and development of new therapeutics.

Unusually, this family of enzymes is multimeric, with the active site formed at the interface between monomers. We have structural data showing enzyme assemblies of 2-,3-.6- and 12-mers, but we do not yet understand the contribution of multimerisation to enzyme function or specificity. To fully understand the biological role of these enzymes, we must interrogate their precise specificity and activity.

In this proposal we aim to characterise diverse members of the GH172 family, from both commensal gut microbes and pathogens. We will dissect the relationship between sequence, specificity and multimerisation, to provide predictive motifs for uncharacterised GH172 members. Our main aims are:
1) characterise diverse GH172 enzymes from commensals and pathogens
2) understand how these enzymes multimerise and how it affects activity
3) use cryo electron microscopy to discover how these multimeric enzymes access the heterogeneous, branched glycans of the mycobacterial cell wall

Outcomes
In addition to broadening our understanding of an unusual family of enzymes, we will have a suite of enzymes with defined activities against D-arabinose containing glycans that will be of broad utility to glycobiologists and mycobacteriologists. These enzymes are found in members of the Mycobacterium avium complex (MAC), a group of species that are responsible for many pulmonary infections, particularly in immunocompromised people. Treatment of MAC infections is difficult, and reoccurrence is high. The incidence of these non-tuberculous mycobacterial infections is increasing worldwide. A GH172 enzyme is also found in M. avium paratuberculosis, the causative agent of Johne's disease in livestock. This is a major worldwide problem for cattle herds. leading to wasting disease, reduced milk yields, weight loss and even death. Understanding the role of GH172 cell-wall targeting enzymes could lead to the development of new therapeutics or diagnostics for these infections.

The mycobacterial cell wall contains arabinogalactan (AG) and lipoarabinomannan (LAM). The D-arabinan component of these two molecules is similar, and it's biosynthesis is essential to mycobacterial viability. Few enzymes are known which can target D-arabinan, and a degradative pathway within mycobacteria themselves is unknown. We have identified, through a screen of glycolytic gut bacteria, a few family of multimeric glycoside hydrolases (GH172s) which are able to degrade mycobacterial D-arabinan. In addition to the gut microbes in which we first discovered these activities, GH172 enzymes are found in several mycobacterial pathogens such as non-tuberculous mycobacteria from the M. avium complex. The enzymes themselves have a structure related to viral capsid proteins, and we have identified GH172 enzymes with 2,3,6 and 12 protomers.

In this proposal we will investigate the activity of multiple GH172 family members from both commensal and pathogenic bacteria, to unpick how multimerisation occurs, and how it contributes to enzyme activity and specificity. We will use cryo-EM to obtain high-resolution structures of multimeric GH172 enzymes in complex with branched and linear mycobacterial cell wall fragments. This will enable us to visualise how variable multimerisation affects substrate access and specificity. Together these data will enable us to predict multimerisation and activity for uncharacterised GH172 enzymes, provide us with a toolkit of defined enzymes active on biologically important glycans, and provide insight into the cell wall biology of mycobacterial pathogens.