Rifamycin monooxygenase

Introduction

Antibiotics are powerful drugs used to treat bacterial infections and they typically work by killing or inhibiting the growth of bacterial cells. Between 1950 and 1970 many novel antibiotics were discovered and subsequently made their way into the clinic, revolutionizing modern medicine. Their success, however, comes with a price: many bacteria have acquired resistance to antibiotics and are therefore difficult or impossible to treat, resulting in a global health crisis. In Europe, approximately 25000 people annually die due to infections of multi-drug resistant bacteria, while in the United States about 2 million people are infected with antibiotic resistant bacteria of which 23.000 die each year (1). If left unchecked, it is expected that 10 million people die each year globally by 2050 due to increased antibiotic resistance (figure 1 adapted from 2). Additionally, the lack of newly developed antibiotics contributes to exacerbation of the current antibiotic crisis. However, the development of resistance is a natural phenomenon and the result of billions of years of evolution. In fact, metagenomic studies on uncontaminated samples from e.g. permafrost have established the presence of antibiotic resistance genes (3). This indicates that antibiotic resistance is ancient and occurs in the absence of human activity, although the current widespread (mis)use of antibiotics in human medicine, livestock and agriculture has dramatically increased the selective pressure for the appearance of antibiotic resistance (2). Bacteria can acquire resistance to antibiotics through different clinically relevant mechanisms (4). Firstly, some bacteria are naturally resistant because they lack the respective target. Antibiotics commonly target cell wall synthesis, protein production or DNA replication. If a bacterium contains, for example,  another version of an enzyme involved in one of these processes that is normally targeted by an antibiotic it is naturally resistant to its mode of action. Secondly, resistance can be obtained via minimizing the antibiotic concentration in the cell. This can be accomplished through efflux or reduction of uptake/penetration. Genetic or posttranslational modification of the antibiotic target represents an additional mechanism of developing resistance. Finally, some antibiotics are enzymatically inactivated. Of grave concern is the development of drug-resistant Mycobacterium tuberculosis (MTB), the causative agent of tuberculosis in humans. Globally 1.7 million people are estimated to be infected with MTB and without treatment the mortality rate is about 70%. The global burden of infections with drug-resistant MTB was in 2016 approximately 490.000 cases with about 4% of new infections attributed to drug-resistant MTB (5). Tuberculosis is treated by a combination of two antibiotics, namely: rifampicin and isoniazid (6). Rifampicin is a member of the rifamycin antibiotic family. These were first isolated in 1959 from actinomycetes and it is well established that rifamycins inhibit bacterial RNA polymerase (7). Rifamycins are mainly active against Gram positive bacteria, while rifampicin, a semi-synthetic rifamycin derivative, displays potent activity against mycobacteria. Not surprisingly, rifampicin is the drug of choice to treat mycobacterial infections, including tuberculosis. Resistance against rifampicin can be obtained via distinct mechanisms. For example, mycobacteria typically develop resistance through point mutations in RNA polymerase, while other soil bacteria and pathogens rely on inactivation of this antibiotic by rifamycin monooxygenase (Rox). This enzyme represents an NADPH-dependent flavoprotein monooxygenase that was recently structurally characterized (8,9). This revealed a unique catalytic mechanism to detoxify rifampicin. Here I will discuss the biochemical and structural features of rifamycin monooxygenase.

Rifamycin monooxygenase of Nocardia farcinica

Recently, high resolution structures of Rox bound to FAD and rifampicin from Nocardia farcinica and Streptomyces venezuelae were determined at 1.8 and 3.3 Å, respectively (8,9). N. farcinica Rox is homodimer in solution, whereas the structure of the monomeric enzyme is shown in figure 3 (8), revealing that the enzyme comprises three domains (bottom right panel), namely: a FAD binding domain (in orange), a middle domain (in blue) and a C-terminal domain (in red). FAD is in yellow ball and sticks and rifampicin is in green ball and sticks. The FAD binding domain is made up a three helix bundle and a central five stranded β sheet, resulting in a β β α substructure. Moreover, this domain is also involved in dimerization. The middle domain is inserted in between the FAD binding domain and is folded into two β α β substructures. The middle domain is unique to Rox and caps the FAD binding domain, and thereby creating two openings that provide access to FAD. One of these is used for the accommodation of the substrate, while the other exposes the isoalloxazine ring as well as the ribityl and ADP moieties to the solvent (8). The C-terminal domain adopts an α/β conformation and is folded against the FAD binding domain. Although the function of this domain is not clear, it could be involved in proper folding and stability of the enzyme. A close up of the FAD-binding site is presented in the left top panel and reveals that the naphthalene ring of rifampicin is perpendicular to the isoalloxazine ring of FAD. Moreover, FAD is oriented towards the substrate binding site in accordance with the “in” conformation. In contrast, in the “out” configuration FAD is pointed away from the substrate binding pocket with the nucleotide portion exposed. This configuration enables interaction with NADPH followed by reduction. Subsequently, FAD rotates into the “in” conformation, thereby pointing towards the substrate. 

Of particular importance is Arg41 (shown in magenta ball and sticks). This residue arches over the active site opening in the FAD-binding domain to interact with FAD and prevents movement of FAD. Conceivably, rotation of FAD is enabled by movements in the enzyme that also change the position of Arg41. A detailed view of the substrate binding pocket is presented in the bottom left panel. The rifampicin binding site is located in the middle domain and is relatively large and flexible. The substrate primarily interacts with non-polar residues (shown in grey ball and sticks) of which the Pro283-Gly286 loop (shown in red ball and sticks) interacts with the naphthalene ring. N. farcinica Rox was also crystalized with only FAD and a comparison of the structures with FAD or FAD and rifampicin is shown in figure 4. The structure in complex with both ligands is in grey with FAD in red and substrate in green ball and sticks. The FAD-containing structure is in blue with FAD in yellow ball and sticks. This reveals that in the presence of substrate the isoalloxazine ring is rotated differently relative to the structure devoid of rifampicin. Moreover, structural changes are also observed in the 280s loop that interacts with the naphthoquinone moiety (indicated by a red arrow). It is therefore conceivable that the structural changes in the 280s loop control the rearrangements of the isoalloxazine ring (8). 

Rifamycin monooxygenase of Streptomyces venezuelae

Although the structure of N. farcinica Rox was the first to be reported, the 3D conformation of another bacterial Rox was determined at 3.3Å recently (9). This structure is shown in figure 5 and includes FAD (in yellow ball and sticks) and rifampicin (in orange ball and sticks) and reveals that, similar to N. farcinica Rox, the enzyme is composed of three domains: a FAD-binding domain (in blue), a middle domain (in red) and a C-terminal domain (in green). The substrate is located in the middle domain with its naphthalene ring oriented vertically towards the isoalloxazine of FAD. This redox cofactor is directed towards the substrate indicating that it is in the “in” conformation as observed in the crystal structure of N. farcinica Rox (8). The overall structures of the monooxygenases from N. farcinica Rox and S. venezuelae are highly similar and is therefore likely that both enzymes employ the same catalytic mechanism. 

Catalytic mechanism of rifamycin monooxygenase

Biochemical characterization of Rox with rifampicin showed that NADPH is only able to efficiently reduce FAD when substrate is present, suggesting that rifampicin binds before NADPH to the enzyme in the catalytic cycle (8). Profound insight into the catalytic mechanism of Rox enzymes was provided recently by structures of N. farcinica Rox in complex with FAD and the hydroxylated rifampicin product at 2.1 Å (10). Figure 6 shows the FAD-substrate (left panel) and FAD-product complex (right panel) in N. farcinica Rox with FAD in yellow ball and sticks and substrate or product in ball and stick representation with coloring according to element (green: carbon; blue: nitrogen; red: oxygen). Both structures reveal that the naphthalene moiety is located next to the N5/C4a edge of FAD’s isoalloxazine ring. Whereas the substrate-containing enzyme (left panel) clearly contains rifampicin with its characteristic macrocyclic structure and ansa bridge, the structure with the rifampicin product (right panel) contains a linear molecule, suggesting that Rox converts macrocyclic rifampicin into a linear compound (figure 3 right top panel). Specifically, the structural data are consistent with a model (bottom panel) in which the C4a-hydroxyperoxyflavin hydroxylates C2 of the substrate (indicated by a red arrow) through a nucleophilic attack followed by cleavage of the ansa bridge by deamination of the ansa amide (indicated by a red arrow), resulting in linearization of the molecule (9,10). The linearized product is unable to bind to RNA polymerase, the target of rifampicin, and thereby explaining the inactivation of this antibiotic. Thus, Rox is a C-hydroxylating enzyme that cleaves the ansa bridge of rifampicin, which blocks of the drug to RNA polymerase.