The bacterium Mycobacterium tuberculosis (figure 1, adapted from Oxford University) is the causative agent of tuberculosis (TB) in humans. This is an infectious disease that typically affects the lungs but it can also spread to other sites of the body. Approximately, 2 billion people (one-third of the world’s population) are infected with M. tuberculosis and in 2016, 10.4 million people developed TB with an estimated 1.3 million TB-related deaths (1). It is therefore one of the most deadliest diseases. In fact, it is globally the ninth leading cause of death (1). Owing to the ability of M. tuberculosis cells to evade the human immune response as well as its insensitivity to many drugs (2), the treatment of TB requires the prolonged combination of chemotherapy, lasting up to nine months (3). Non-compliance of these treatments contributes to the emergence of drug-resistant TB strains. These have developed into a growing threat with 490 million cases of multi-drug resistant TB in 2016 (1). The discovery of new drug targets as well as novel antimycobacterial therapeutics is therefore of pressing urgency. Ideally, these should harness biochemical processes that are unique to M. tuberculosis cells such as the metabolism of coenzyme F420.
This compound is a flavin derivative and is used as redox cofactor similar to FMN and FAD, which are ubiquitously present in all organisms and play a central metabolic role. In contrast, F420 is exclusively produced and used by some bacteria as well as archaea and is therefore considered to be one of Nature’s rarest cofactors (4). Although structurally related to FAD and FMN, P420 displays distinct redox properties such as a low redox potential and it serves as a carrier of two electrons. Owing to its unique chemical properties, F420 is able to catalyze some of the most challenging reduction reactions in biology. Physiologically, it has been established that F420 plays a key role in the central metabolism of methanogens, mycobacteria and streptomycetes as well as the production of secondary metabolites, cell wall intermediates and biodegradation pathways (4). Actinobacteria contain numerous potential F420H2-dependent oxidoreductases and these can be divided into two superfamilies, the luciferase-like hydride transferases (LLHT) superfamily and the superfamily of flavin/deazaflavin oxidoreductases (FDOR). Phylogenetically, FDOR enzymes can be further subclassified into two groups, FDOR-A and FDOR-B. These enzymes typically hydrogenate a variety organic substrates, emphasizing their potential as novel biocatalyst. Here, I will discuss biochemical and structural features of recently characterized mycobacterial F420H2-dependent reductases.
Structure of F420 and overview of its biosynthesis
FAD and FMN serve as redox cofactor for a plethora of key enzymatic reactions in metabolism throughout all domains of life. However, some bacteria and archaea produce and utilize F420, another flavin derivative as redox cofactor (4). This compound was initially discovered in mycobacteria as a yellow and fluorescent pigment in the early 1960s, while it was chemically defined in 1972 as a redox active flavin-derivative that is prominent in methanogens (5,6). F420 is named after its characteristic absorbance band at 420 nm and is involved in several key metabolic redox reactions in methanogens. Although it is widely accepted that F420 is restricted to some bacteria and archaea, accumulating genetic evidence suggests that it is also produced by different aerobic soil bacteria (7). F420 is not essential for cell viability in mycobacteria under ideal conditions, unlike methanogens. However, F420 seems to contribute to the ability of mycobacteria to survive under harsh conditions, thereby increasing the virulence and antimicrobial resistance of pathogenic mycobacteria (8).
The structures of F420, FMN and FAD are shown in the upper panel of figure 1 (adopted from 9), revealing an overall core structure made up of a tricyclic (isoxazoline) ring. However, the tricyclic ring of F420 contains a carbon atom at position 5, while the isoxazoline ring of FMN and FAD contains a characteristic nitrogen atom at this position. F420 contains, therefore, a deazaisoxazoline ring, which is connected to a side chain comprising a ribitylphospholactyl moiety and polyglutamate chain of variable length (4). Primarily due to the presence of a carbon atom at position 5, F420 exhibits distinct redox properties, such as a lower redox potential (-360 mV) than FAD and FMN (-210 mV) and ability to serve as two (hydride) electron carrier. These features enable F420 to catalyze some of the most challenging reduction reactions in biology, including those of C=C and C≡C bonds as well as reduction of enone, imine, enamine and nitro groups (4). F420 is produced from different building blocks, namely: F0, lactate, glutamate and GTP. F0 (8-hydroxy-5-deazaflavin) is a chromophore used by DNA photolyases and is a precursor of F420 (4). In bacteria and archaea, the biosynthesis of F420 comprises 3 major steps of which the first comprises the formation of F0 through a condensation of a riboflavin precursor, 5-amino-6-(D-ribitylamino)uracil, and tyrosine. This step is catalyzed by the enzyme F0 synthase, which comprises two subunits, CofG and CofH, respectively. These are sometimes fused into a single enzyme (CofGH). In the second step, LPPG (L-lactyl-2-diphospho-5’-guanosine) is most likely produced from 2-phospho-L-lactate via GTP-dependent reaction catalyzed by CofC. Subsequently, LPPG is transferred to F0 by CofD, resulting in the intermediate F420-0. Finally, the LPPG side chain is elongated by addition of glutamate residues by the ligase CofE. The number of glutamate residues is species dependent and ranges from 3 to 9 (4). Although the precise reason for the difference in glutamate residues is not clear, it has been shown that the length of the polyglutamate chain modulates the catalytic properties of F420H2-dependent reductases (10). Specifically, short glutamate chains promoted an accelerated substrate turnover, while longer chains bound with greater affinity to the enzyme and favored an increased substrate affinity but reduced turnover rate.
Biochemical features of F420H2-dependent reductases
The proteomes of actinobacteria contain numerous putative F420H2-dependent oxidoreductases. In fact, these enzymes carry out the biological role of F420 in these organisms. actinobacterial F420H2-dependent reductases are commonly divided into two superfamiles, the superfamily of luciferase-like hydride transferases (LLHT) and the flavin/deazaflavin oxidoreductase (FDOR) superfamily (4). The latter superfamily comprises the largest known group of F420H2-dependent enzymes and members of this family are prominent in Mycobacteria: 30 in M. smegamatis, 15 in M. tuberculosis and 3 in M. leprae. FDOR proteins are small single-domain enzymes of about 150 residues. These enzymes are folded into a characteristic split β-barrel conformation with a cofactor-binding channel and substrate-binding pocket. Phylogenetically, FDOR family members can be further subclassified into two groups, FDOR-A and FDOR-B (9). Enzymes belonging to the FDOR-A group are ubiquitously F420-dependent and are particularly abundant in actinobacteria and chloroflexi bacteria. Proteins of this group can be subdivided into FDOR A1-A4 and AA1-AAG (9). The available biochemical data suggests that FDOR-A reductases exhibit a rather broad substrate spectrum and it is therefore conceivable that in vivo these enzymes detoxify a range of organic substrates, including antimicrobial drugs. Reductases of the FDOR-B group are found in a variety of bacteria, including ones that do not produce F420. It is therefore not surprising that the FDOR-B group comprises FMN and FAD-dependent enzymes. FDOR-B reductases are subclassified into FDOR B1-B12 (9). Hence, mycobacteria contain 22 different FDOR subgroups. A variety of flavin-dependent enzymes rely on reduced nicotinamide cofactors (NADH/NADPH) for subsequent reduction of FAD or FMN, while NAD/NADP are in turn recycled by dehydrogenases. Likewise, the reduction of F420 is in mycobacteria accomplished by glucose-6-phosphate dehydrogenase, which directly links mycobacterial carbohydrate catabolism to F420 reduction (4).
In contrast to the FDOR superfamily, biochemical and structural information of LLHT family members is scant and these will not be considered further.
Structural features of a FDOR-A reductase
Several high resolution structures of FDOR family members from different mycobacteria have been elucidated. These include Ddn, the deazaflavin-dependent nitroreductase from M. tuberculosis. This enzyme is a prototypical member of the FDOR-A subgroup and is responsible for the activation of several experimental anti-tuberculosis prodrugs such as pretomanid (PA-824). This is a bicyclic 4-nitroimidazole-like compound (figure 3) that is active against replicating and non-replicating M. tuberculosis cells. Its mode of action is not fully understood but it appears to kill M. tuberculosis cells through inhibition of cell wall biosynthesis (11). Moreover, this compound is currently evaluated in phase III clinical trials. The structure of F420-bound Ddn was solved at 2.1 Å (12) and is shown in figure 3 (upper right panel). This reveals an overall conformation made up of six antiparallel β-sheets (in yellow) organized into a barrel with a Greek key topology and four α helices (in red). F420 (in magenta) is located in a groove on the surface of the enzyme. FDOR family members contain functionally important motifs that are conserved amongst different subclasses (9). For example, Ddn is composed of four motifs (1 to 4) that are typically found in enzymes belonging to the FDOR-A group and these are shown in the bottom left panel of figure 3. Motif 1 is indicated in blue and interacts with the ribityl and polyglutamate chain of F420. Likewise, motif 2 (in orange) also interacts with the ribityl and polyglutamate chain. Motif 3 (in grey) is involved in cofactor and substrate binding and is, moreover, essential for catalysis. Motif 4 (yellow) is primarily involved in substrate binding (9). Although attempts to obtain crystals containing F420 and PA-824 were unsuccessful, residues important for catalysis were identified through modelling and subsequently confirmed by mutagenesis. This revealed that motif 3 contains a serine residue at position 78 (shown in green ball and sticks, bottom right panel) that is crucial for catalytic activity. Importantly, this residue is conserved amongst FDOR-A enzymes. Additionally, motif 4 contains a tyrosine at position 136 (shown in green ball and sticks) that ensures proper aligning of the substrate in the active site as well as tyrosine at position 130 which is probably involved in substrate binding and orientation. Residues involved in F420-binding are shown as red ball and sticks in the bottom right panel and include pro63 of motif 1 that stabilizes the the ribityl and polyglutamate chain and is highly conserved amongst FDOR-A enzymes. Motif 2 contains trp88 and asn91 that are part of the conserved so-called WXXN motif, involved in ribityl and polyglutamate chain stabilization. In addition to catalytically important ser78, motif 3 also contains lys79, involved in stabilizing the F420 complex and ensuring a catalytically competent position of ser78 (12). Likewise, ala76 and tyr65 also stabilize the F420 complex. Motif 4 contains tyr133, which probably anchors F420 into the active site (12).
Interestingly, modelling of PA-824 into the active site suggests that it is bound in such a way that its nitroimidazole group is facing towards the Re face of the deazaflavin cofactor. This is shown in the left panel of figure 4 (adopted from 12) with PA-824 in purple and F420 in yellow. A comparison of the apo (PDB 3R5P) and holo enzyme (PDB 3R5W) is presented in the right panel of figure 5 (the holo enzyme is colored according to secondary structure in red, yellow and green and with F420 in orange. The apo structure is in blue, magenta and pink). This shows that the overall structures are strikingly similar. Significant conformational changes are, however, observed in the F420-binding loop of motif 4, suggesting local flexibility in this region. Conceivably, these conformational movements are functionally relevant as they could allow replacement of oxidized F420 with reduced F420 during the catalytic cycle.
Structural features of a FDOR-B reductase
In addition to high resolution structures of FDOR-A enzymes, detailed structural information of FDOR-B members is also available. For example, Rv1155, a putative F420H2-dependent reductase, from M. tuberculosis and typical FDOR-B enzyme was structurally characterized together with F420 at 2.3 Å (13). The structure of the holo enzyme is shown in figure 5, revealing that the Rv1155 is a dimer (upper left panel with one protomer in blue and the other in green) with F420 (in orange) bound at the interface of the two protomers. The overall conformation comprises a central barrel-like structure made up of antiparallel β-sheets surrounded by small α helices. Despite structural similarity, F420-dependent reductases of the FDOR-B group are not related in sequence to FDOR-A enzymes. In contrast, FDOR-B members posses a F420-binding site that is similar to the cofactor-binding site other flavin-dependent enzymes as well as hemoproteins. FDOR-B reductases can’t be identified through the characteristic FDOR-A motifs, although these enzymes harbor two other distinct motifs, termed motif 5 and 6 (9). These are shown in the upper right panel of figure 5 with motif 5 in yellow and 6 in red. Motif 5 is made up of the β-sheets of the central barrel and is involved in the binding of F420, FAD and FMN, while motif 6 comprises a small α-helical structure and is exclusively found in F420-binding FDOR-B enzyme. A detailed view of the F420-binding site is presented in the left bottom panel of figure 6. Residues that are involved in the binding of F420 are shown in ball and stick orientation, revealing that the cofactor-binding site comprises residues of both protomers. The negatively charged phosphate group and the initial residues of the polyglutamate chain of F420 are fixed in the active site by a tunnel of positively charged and aromatic residues of motif 5 and 6 as well as adjacent β-sheets. Motif 6 contains a lysine at position 57 that is highly conserved amongst FDOR-B reductases, which is part of a hydrogen bond network that anchors the phosphate into the active site. Although different potential substrates were tested, Rv1155 did not show any activity towards any of the tested compounds and therefore no structure of the substrate-bound enzyme could be determined. However, a putative active site lined with potential substrate-binding polar and hydrophobic residues is present beneath F420’s heterocycle (right bottom panel of figure 6, adopted from 13).
Catalytic mechanism of F420H2-dependent reductases
The available structures of FDOR reductases reveal that these enzymes, and in particular members of the FDOR-A group, typically contain active sites with a substrate binding region of variable architecture, enabling the accommodation of a variety of compounds. This is consistent with their broad substrate spectrum, suggesting that they have evolved to hydrogenate various substrates. To probe the biotechnological potential of these enzymes in more detail, the substrate scope of 11 mycobacterial reductases of the FDOR-A and B groups was assessed in a recent study (14). It was found that these enzymes are active against structurally and chemically diverse substrates, including monocyclic compounds as well as more complex and bulky compounds such as coumarin and quinone (14). Based on this a common mechanism for the reduction of substrates was proposed, namely: (i) binding of deprotonated F420 (F420H-) to the enzyme, (ii) this is followed binding of the substrate into the active site, (iii) subsequent alignment of the reactive C5 center of the deazaisoxazoline ring and the activated alkene group enables direct hydride transfer and (iv) ensuing steps will complete the catalytic cycle, resulting in protonation of the substrate and release of the product from the active site (figure 6 adopted from 14). Conceivably, the reduced cofactor binds directly to the enzyme from the aqueous environment and subsequently facilitates hydride transfer to the substrate. Following dissociation from the enzyme, F420 is regenerated by glucose-6-phosphatase dehydrogenase.
Redox reactions play a central role in cellular metabolism and to enable the transfer of electrons during these conversions the enzymes that catalyze these reactions are typically equipped with a cofactor such as heme, FAD or FMN. The flavin derivatives FAD and FMN serve as key redox cofactor and are ubiquitously present throughout all domains of life. In contrast, F420, another flavin-based redox cofactor, is produced and utilized by some bacteria and archaea and is therefore thought to be one of rarest cofactors in biology (4). Although F420 is structurally related to FMN and FAD, it exhibits distinct redox properties, such as a lower redox potential and ability to serve as two (hydride) electron carrier. These features enable F420 to catalyze some of the most challenging reduction reactions in Nature, including those of C=C and C≡C bonds as well as reduction of enone, imine, enamine and nitro groups (4). It has been well established that F420 plays a crucial role in the metabolism of methanogens and actinobacteria. For example, the proteomes of actinobacteria contain numerous putative F420H2-dependent oxidoreductases. In fact, these enzymes carry out the biological role of F420 in these organisms and can be divided into two superfamilies: LLHT and FDOR superfamily. The latter superfamily comprises the largest known group of F420H2-dependent reductases and members of this family are prominent in mycobacteria. Phylogenetically, FDOR enzymes can be further subclassified into two groups, FDOR-A and FDOR-B (9). The structural data available shows that members of the FDOR-A typically contain active sites with a substrate binding region of variable architecture, enabling the accommodation of a variety of compounds. To probe the biotechnological potential of these enzymes in more detail, the substrate scope of FDOR-A and B reductases was assessed recently (14). It was found that these enzymes are active against structurally and chemically diverse substrates, including monocyclic compounds as well as more complex and bulky compounds such as coumarin and quinone. This demonstrates their potential as novel biocatalyst for the reduction of a variety of α,β-unsaturated compounds.
1. WHO. Global tuberculosis report 2017.
2. Barry CE, Boshoff HI, Dartois V, Dick T, Ehrt S, Flynn J, Schnappinger D, Wilkinson RJ, Young D. 2009. The spectrum of latent tuberculosis: rethinking the biology and intervention strategies.Nat Rev Microbiol. 12: 845-855.
3. Tang P, Johnston J. 2017. Treatment of Latent Tuberculosis Infection. Curr Treat Options Infect Dis. 9: 371-379.
4. Greening C, Ahmed FH, Mohamed AE, Lee BM, Pandey G, Warden AC, Scott C, Oakeshott JG, Taylor MC, Jackson CJ. 2016. Physiology, Biochemistry, and Applications of F420- and Fo-Dependent Redox Reactions. Microbiol Mol Biol Rev. 80: 451-493.
5. Cousins FB. 1960. The prosthetic group of a chromoprotin from mycobacteria. Biochim Biophys Acta. 40: 532-534.
6. Cheeseman P, Toms-Wood A, Wolfe RS. 1972. Isolation and properties of a fluorescent compound, factor 420 , from Methanobacterium strain M.o.H. J Bacteriol. 112: 527-531.
7. Ney B, Ahmed FH, Carere CR, Biswas A, Warden AC, Morales SE, Pandey G, Watt SJ, Oakeshott JG, Taylor MC, Stott MB, Jackson CJ, Greening C. 2017. The methanogenic redox cofactor F420 is widely synthesized by aerobic soil bacteria. ISME J. 11: 125-137.
8. Jirapanjawat T, Ney B, Taylor MC, Warden AC, Afroze S, Russell RJ, Lee BM, Jackson CJ, Oakeshott JG, Pandey G, Greening C. 2016. The redox cofactor F420 protects mycobacteria from diverse antimicrobial compounds and mediates a reductive detoxification system. Appl. Environ. Microbiol. 82: 6810-6818.
9. Ahmed FH, Carr PD, Lee BM, Afriat-Jurnou L, Mohamed AE, Hong NS, Flanagan J, Taylor MC, Greening C, Jackson CJ. 2015. Sequence-Structure-Function Classification of a Catalytically Diverse Oxidoreductase Superfamily in Mycobacteria. J Mol Biol. 427: 3554-3571.
10. Ney B, Carere CR, Sparling R, Jirapanjawat T, Stott MB, Jackson CJ, Oakeshott JG, Warden AC, Greening C. 2017. Cofactor Tail Length Modulates Catalysis of Bacterial F420-Dependent Oxidoreductases. Front Microbiol. 8: 1902.
11. Denny WA, Palmer BD. 2010. The nitroimidazooxazines (PA-824 and analogs): structure-activity relationship and mechanistic studies. Future Med Chem. 8: 1295-1304.
12. Cellitti SE, Shaffer J, Jones DH, Mukherjee T, Gurumurthy M, Bursulaya B, Boshoff HI, Choi I, Nayyar A, Lee YS, Cherian J, Niyomrattanakit P, Dick T, Manjunatha UH, Barry CE, Spraggon G, Geierstanger BH. 2012. Structure of Ddn, the deazaflavin-dependent nitroreductase from Mycobacterium tuberculosis involved in bioreductive activation of PA-824. Structure. 20: 101-112.
13. Mashalidis EH, Gittis AG, Tomczak A, Abell C, Barry CE 3rd, Garboczi DN. 2015. Molecular insights into the binding of coenzyme F420 to the conserved protein Rv1155 from Mycobacterium tuberculosis. Protein Sci. 24: 729-740.
14. Greening C, Jirapanjawat T, Afroze S, Ney B, Scott C, Pandey G, Lee BM, Russell RJ, Jackson CJ, Oakeshott JG, Taylor MC, Warden AC. 2017. Mycobacterial F420H2-Dependent Reductases Promiscuously Reduce Diverse Compounds through a Common Mechanism. Front Microbiol. 8: 1000.