Abstract
The arabinogalactan (AG) component of the mycobacterial cell wall is an essential branched polysaccharide which tethers mycolic acids (m) to peptidoglycan (P), forming the mAGP complex. Much interest has been focused on the biosynthetic machinery involved in the production of this highly impermeable shield, which is the target for numerous anti-tuberculosis agents. The galactan domain of AG is synthesised via a bifunctional galactofuranosyltransferase (GlfT), which utilises UDP-Galf as its high-energy substrate. However, it has proven difficult to study the protein in its recombinant form due to difficulties in recovering pure soluble protein using standard expression systems. Herein, we describe the effects of glfT co-induction with a range of chaperone proteins, which resulted in an appreciable yield of soluble protein at 5 mg/L after a one-step purification procedure. We have shown that this purified enzyme transfers [14C]Galf to a range of both beta(1-->5) and beta(1-->6) linked digalactofuranosyl neoglycolipid acceptors with a distinct preference for the latter. Ligand binding studies using intrinsic tryptophan fluorescence have provided supporting evidence for the apparent preference of this enzyme to bind the beta(1-->6) disaccharide acceptor. However, we could not detect binding or galactofuranosyltransferase activity with an n-octyl beta-d-Gal-(1-->4)-alpha-l-Rha acceptor, which mimics the reducing terminus of galactan in the mycobacterial cell wall. Conversely, after an extensive bioinformatics analysis of the H37Rv genome, further cloning, expression and functional analysis of the Rv3792 open reading frame indicates that this protein affords galactofuranosyltransferase activity against such an acceptor and paves the way for a better understanding of galactan biosynthesis in Mycobacterium tuberculosis.