Bacteria 1325 Structure, function and biosynthesis of the Mycobacterium tuberculosis cell wall: arabinogalactan and lipoarabinomannan assembly with a view to discovering new drug targets L.J. Alderwick*, H.L. Birch*, A.K. Mishra*, L. Eggeling† and G.S. Besra* 1 *School of Biosciences, University of Birmingham, Edgbaston, Birmingham B15 2TT, U.K., and †Institute for Biotechnology 1, Research Centre J ¨ ulich, D-52425 J¨ ulich, Germany Abstract In spite of effective antibiotics to treat TB (tuberculosis) since the early 1960s, we enter the new millennium with TB, currently the leading cause of death from a single infectious agent, killing more than three million people worldwide each year. Thus an understanding of drug-resistance mechanisms, the immunobiology of cell wall components to elucidate host–pathogen interactions and the discovery of new drug targets are now required for the treatment of TB. Above the plasma membrane is a classical chemotype IV PG (peptidoglycan) to which is attached the macromolecular structure, mycolyl-arabinogalactan, via a unique diglycosylphosphoryl bridge. This review will discuss the assembly of the mAGP (mycolyl-arabinogalactan- peptidoglycan), its associated glycolipids and the site of action of EMB (ethambutol), bringing forward a new era in TB research and focus on new drugs to combat multidrug resistant TB. Introduction The Mycobacterium tuberculosis cell envelope differs sub- stantially from the cell wall structures of both Gram-negative and Gram-positive bacteria. This unique cell wall structure accounts for its unusual low permeability and resistance towards common antibiotics. The main structural element consists of a cross-linked network of PG (peptidoglycan) in which some of the muramic acid residues are replaced with a complex polysaccharide, AG (arabinogalactan). The AG is attached to PG through a unique linker unit, and in turn is acylated at its distal end to PG with mycolic acids. The entire complex is abbreviated as the mAGP (mycolyl- arabinogalactan–peptidoglycan) and is essential for viability in M. tuberculosis and other mycobacteria [1]. Our previous studies have elucidated the gross structural features of the mAGP complex, including aspects of its biosynthesis [1,2]. However, there are still substantial gaps in our knowledge of AG biosynthesis and the identification of potential new drug targets and in the biosynthesis of the related M. tuberculosis essential immunogen, LAM (lipoarabino- mannan) [3]. Key words: arabinogalactan, cell wall, drug target, lipoarabinomannan, Mycobacterium tuberculosis. Abbreviations used: AG, arabinogalactan; DP, decaprenol phosphate; (t-)Araf , (terminal) arabinofuranose; DPA, decaprenylmonophosphoryl-d-arabinose; EMB, ethambutol; Galf , galactofuranose; LAM, lipoarabinomannan; LM, lipomannan; mAGP, mycolyl-arabinogalactan– peptidoglycan; Manp, mannopyranose; PG, peptidoglycan; PI, phosphatidyl-myo-inositol; PIM, phosphatidyl-myo-inositol mannoside; PPM, polyprenyl monophosphomannose. 1 To whom correspondence should be addressed (email g.besra@bham.ac.uk). Biosynthesis of mycobacterial AG The biosynthesis of AG (Figure 1) begins with the formation of the linker unit, through the transfer of GlcNAc-1-P and Rha from their respective sugar nucleotides, to form polyprenol-P-P-GlcNAc and polyprenol-P-P-GlcNAc-Rha [4]. Several studies have defined the genetics and enzymology surrounding the biosynthesis of dTDP-Rha and the wbbL rhamnosyltransferase [5,6]. Surprisingly, the GlcNAc trans- ferase (rfe, Rv1302) has been largely ignored. Polyprenol-P- P-GlcNAc-Rha then serves as an acceptor for the sequential addition of Galf (galactofuranose) residues from UDP-Galf [7] to form polyprenol-P-P-GlcNAc-Rha-Gal 30 . The bulk of galactan polymerization is performed by a novel enzyme designated GlfT (Rv3808c). This latter enzyme expresses two galactofuranosyltransferase activities, a UDP-Galf :β - D-(1 → 5)-Galf and a UDP-Galf :β -D-(1 → 6)-Galf , both activities being required for alternating β (1 → 5) and β (1 → 6) linkages during galactan polymerization [8,9]. Chemical ana- lysis of the mature lipid-linked galactan, synthesized in vitro, suggests that this intermediate then serves as an acceptor for the subsequent addition of Araf (arabinofuranose) residues from β -D-arabinofuranosyl-1-monophosphoryldecaprenol [DPA (decaprenylmonophosphoryl-D-arabinose)] in the formation of the Araf portion (α1 → 5, α1 → 3 and β 1 → 2 linkages) of AG [10,11]. The key steps leading to the biosynthesis of DPA, from pRpp (5-phospho-ribofuranose- pyrophosphate) [12] and DP (decaprenol phosphate), have recently been investigated by us [13] and others [14,15]. The AG-lipid intermediate is then at some point transglycosylated to PG and mycolylated [16]. C  The Authors Journal compilation C  2007 Biochemical Society