Buruli toxin genes decoded
The ancient scourges of leprosy and tuberculosis are diseases caused by mycobacteria. These afflictions remain widespread in much of the world. It is estimated that one-third of Earth's population is infected with tuberculosis, and >2 million die of the active form of the disease annually. Tuberculosis claims the greatest number of victims of any infectious disease and has reawakened public health concerns everywhere with the emergence of multiply drug-resistant strains. A lesser-known but devastating skin disease known as Buruli ulcer is caused by Mycobacterium ulcerans. This human pathogen is carried by aquatic insects (1), and the occurrence of the disease is spreading in central and west Africa, where its incidence now exceeds that of leprosy and is similar to tuberculosis in highly affected areas. Infection with M. ulcerans causes progressive necrotic lesions that, if untreated, can extend to 15% of a victim's body and lead to lifelong disability and occasionally death. Curiously, even advanced disease is usually marked by little inflammatory response and no physical pain. Pathogenesis by M. ulcerans is caused by the secretion of a small family of lipophilic toxins exemplified by mycolactones A and B. In this issue of PNAS, Stinear et al. (2) report the isolation and characterization of a biosynthetic cluster comprising six genes, three of which encode giant polyketide synthases (PKSs) that are central to mycolactone production. Their discovery reveals several highly unusual features of this virulence system and these proteins.,
Structures of mycolactone A (z isomer at C-4′) and mycolactone B (E isomer).
Structures of erythromycin A and rapamycine.
Genetic subtraction experiments led to identification of the M. ulcerans-specific PKS genes (3). Unexpectedly, Stinear et al. found that the PKS genes are harbored in a large, 174-kb plasmid, whose primary function appears to be mycolactone production. This is the first reported example of plasmid-mediated virulence in a Mycobacterium. In addition, complete decoding and translation of the biosynthetic gene cluster have revealed two PKSs thought responsible for assembling the macrolactone core of the toxin and a third PKS for generation of the side chain. These proteins are members of the modular class of PKSs discovered in 1990 (4, 5). This family of enzymes produces several medicinally important compounds having antibiotic, immunosuppressive, anticancer, antifungal, and antiparasitic activities in animals, e.g., erythromycin A and rapamycin. Intense effort during the past 12 years has led to the characterization of many examples of these multienzymes, whose hallmarks are their comparatively large size and their linear arrangement of catalytic domains. These domains function as an “assembly line” to polymerize and modify, by reduction and dehydration steps, simple monomer CoA esters to much higher molecular weight products. These natural products are richly adorned with ketone, hydroxyl, carbon–carbon double bond, and alkyl substituents at specific locations and in specific stereochemical configurations, which are “programmed” by the organization of catalytic domains grouped into “modules” for each monomer addition (6).
Plasmid-mediated virulence in Mycobacterium has been previously unreported.
Against this body of information, two features stand out about the mycolactone PKSs found by Stinear et al. (2). First, they have predicted sizes that place them among the largest of known proteins. The largest of these proteins is 1.8 MDa, which together with its smaller partner, 0.26 MDa, constructs the ketolide core of mycolactone. The side chain is synthesized by the third PKS of 1.2 MDa. Second, and most provocatively, the extent of similarity is exceptionally high at the gene and protein levels among domains of comparable function in all 16 modules that carry out chain extension chemistry. Although primary sequence identities typically fall between 40% and 70%, the key ketosynthase domains in the mycolactone PKSs, for example, are >97% identical. Other domains of analogous chemical and stereochemical function (7) similarly show extremely high levels of sequence identity (causing, no doubt, great difficulties in their sequencing). The fact that the genetic machinery of mycolactone biosynthesis is housed in a plasmid and that high extents of mutual sequence identity exist suggests that the acquisition of this virulence marker took place quite recently and evolved rapidly by gene recombination and duplication events. That said, one would also expect that the high sequence identity among multiple regions of the PKS genes would lead to great instability and mutability. One wonders how the organism is able to maintain these enormous genes to sustain biosynthesis of the toxin.
On another level, the engineering of PKSs to carry out the synthesis of “unnatural products” and combinatorial biosynthesis (8–10) must contend with the practical limitations imposed by downstream domains and their tolerance for structural variation(s) engineered at earlier steps in a growing chain. Clearly, for the mycolactone system, virtually no variation exists among the domains of analogous function, and yet a chain of variant structure at each biosynthetic step is accepted, and the expected downstream reactions are executed with fidelity. Although no experiments have yet tested this idea, the authors speculate that the mycolactone PKS modules could constitute a toolbox of “universal” extender units for the creation of PKSs that are both plastic in their synthetic potential and fully functional. This intriguing idea notwithstanding, at the very least, the nexus of chemistry, biology, and medicine embodied in this article opens approaches for the treatment and control of Buruli ulcers.
