Get 20M+ Full-Text Papers For Less Than $1.50/day. Start a 14-Day Trial for You or Your Team.

Learn More →

Structural insights into simocyclinone as an antibiotic, effector ligand and substrate

Structural insights into simocyclinone as an antibiotic, effector ligand and substrate Downloaded from https://academic.oup.com/femsre/article/42/1/fux055/4604775 by DeepDyve user on 13 July 2022 FEMS Microbiology Reviews, fux055, 42, 2018, 100–112 doi: 10.1093/femsre/fux055 Advance Access Publication Date: 8 November 2017 Review article REVIEW ARTICLE Structural insights into simocyclinone as an antibiotic, effector ligand and substrate 1,† 2,‡ 3,# Mark J. Buttner ,MartinSchafer ¨ , David M. Lawson 3,§ and Anthony Maxwell Department of Molecular Microbiology, John Innes Centre, Norwich Research Park, Norwich NR4 7UH, UK, 2 3 Department of Biochemistry, Duke University School of Medicine, Durham, NC 27710, USA and Department of Biological Chemistry, John Innes Centre, Norwich Research Park, Norwich NR4 7UH, UK Corresponding author: Department of Biological Chemistry, John Innes Centre, Norwich Research Park, Norwich NR4 7UH, UK. Tel: +44 1603 450771; E-mail: tony.maxwell@jic.ac.uk One sentence summary: Simocyclinones are actinomycete natural products that target bacterial DNA gyrase; structural work has revealed their molecular mode of action, aspects of their biosynthesis and the mechanism underlying their inducible export. Editor: Kenn Gerdes Mark J. Buttner, http://orcid.org/0000-0002-3505-2981 Martin Schafer ¨ , http://orcid.org/0000-0003-0891-0626 David M. Lawson, http://orcid.org/0000-0002-7637-4303 Anthony Maxwell, http://orcid.org/0000-0002-5756-6430 ABSTRACT Simocyclinones are antibiotics produced by Streptomyces and Kitasatospora species that inhibit the validated drug target DNA gyrase in a unique way, and they are thus of therapeutic interest. Structural approaches have revealed their mode of action, the inducible-efflux mechanism in the producing organism, and given insight into one step in their biosynthesis. The crystal structures of simocyclinones bound to their target (gyrase), the transcriptional repressor SimR and the biosynthetic enzyme SimC7 reveal fascinating insight into how molecular recognition is achieved with these three unrelated proteins. Keywords: antibiotics; Streptomyces; DNA gyrase; aminocoumarins; DNA topoisomerases; transcription factor INTRODUCTION ral products that were first isolated nearly 20 years ago from Streptomyces antibioticus Tu¨ 6040, which produces simocycli- The increase in resistance to antimicrobials has become a seri- nones A1, B1, B2, C2, C4, D4, D6, D7 and D8 (Schimana et al. 2000, ous challenge in the 21st Century, with rising antibiotic-resistant 2001). More recently, Kitasatospora sp. and Streptomyces sp. NRRL pathogens, particularly in hospital settings, and a paucity of B-24484 have been identified as producers of the novel simocy- new agents becoming available (Boucher et al. 2009; Bush et al. clinones D9, D10 and D11 (Bilyk et al. 2016); representative simo- 2011). It is therefore essential that we continue our search cyclinones are shown in Fig. 1. As most work has been carried for new antibacterial compounds, particularly novel natural out on simocyclinone D8 (SD8), it will be the main topic of this products, which have the possibility of exploiting new chemi- review. cal space. Actinomycetes, most notably the genus Streptomyces, The target of simocyclinones is the type II DNA topoiso- have proved to be a rich source of bio-active molecules, with merase, DNA gyrase (Fig. 2). DNA topoisomerases (topos) are most antibiotics in current clinical use being actinomycete nat- enzymes found in all organisms that catalyze the intercon- ural products or their derivatives (Clardy, Fischbach and Walsh versions of different topological forms of DNA, e.g. relaxed– 2006). In this review, we discuss the simocyclinones, natu- Received: 12 September 2017; Accepted: 7 November 2017 FEMS 2017. This is an Open Access article distributed under the terms of the Creative Commons Attribution License (http://creativecommons.org/ licenses/by/4.0/), which permits unrestricted reuse, distribution, and reproduction in any medium, provided the original work is properly cited. 100 Downloaded from https://academic.oup.com/femsre/article/42/1/fux055/4604775 by DeepDyve user on 13 July 2022 Buttner et al. 101 Figure 1. Chemical structures of simocyclinones D8, D4 and C4. around the A B complex and the passage of one segment of 2 2 DNA, termed the ‘T’ or ‘transported’ segment, through a double- stranded break in another, the ‘G’ or ‘gate’ segment. Catalytic supercoiling requires the hydrolysis of ATP. As this reaction pro- ceeds via transient double-strand breaks in DNA, agents that can stabilize the broken DNA intermediate, such as the fluo- roquinolones, are very effective antibacterial agents. A number of other compounds inhibit gyrase (and other topoisomerases) via this ‘cleavage-complex stabilization’ mechanism (Collin, Karkare and Maxwell 2011; Bush, Evans-Roberts and Maxwell 2015). In addition, gyrase and other type II topoisomerases can be inhibited by compounds that act at the ATP-binding site Figure 2. Schematic representation of the DNA gyrase A B complex with bound (Maxwell and Lawson 2003); this includes aminocoumarin an- 2 2 G-segment DNA. Each subunit consists of an N-terminal domain (NTD) and a tibiotics, such as novobiocin. As will be shown below, the inhibi- C-terminal domain (CTD). For reference, the DNA gyrase structure shown in Fig. 3 tion of gyrase by simocyclinones occurs by a different, previously corresponds to a homodimer of a 55 kDa fragment of the GyrA NTD. unknown, mechanism: they prevent the enzyme from binding DNA. It is possible that this mode of action can be exploited to- wards the development of novel, clinically relevant antibiotics. supercoiled, knotted–unknotted, catenated–decatenated (Vos It is interesting to note that there are a number of peptide and et al. 2011; Bush, Evans-Roberts and Maxwell 2015)and are protein inhibitors that exhibit the three mechanisms of gyrase essential for DNA replication and transcription (Wang 2002). inhibition (Collin, Karkare and Maxwell 2011). For example, mi- They are classified as type I or II depending upon whether their crocin B17, CcdB and ParE can stabilize the cleavage complex, reactions proceed via single- or double-stranded breaks in DNA, MfpA and Qnr proteins seem to prevent DNA binding, and FicT and further divided into sub-types: IA, B, C and IIA and B, de- proteins modify GyrB and prevent ATPase activity (Harms et al. pending of mechanistic and evolutionary considerations (Wang 2015). 1996; Forterre et al. 2007). DNA gyrase (the target of simocycli- Simocyclinones (D4 and D8) were discovered during the nones) is a type IIA topoisomerase, and the only enzyme that search for novel secondary metabolites from Streptomyces strains can catalyze the introduction of negative supercoils into DNA. derived from soil samples (Schimana et al. 2000). These com- It is essential in all bacteria but lacking from animals, includ- pounds showed antibiotic activity against certain Gram-positive ing humans, making it an ideal target for antibiotics. The type bacteria and cytotoxic effects on tumor cell lines. By varying mi- II topoisomerase in humans, topo II, has been developed as an crobial growth and fermentation conditions the yield of these anti-cancer target (Pommier et al. 2010) and can relax and de- compounds was analyzed and optimized (Theobald, Schimana catenate DNA but cannot supercoil. Most bacteria, in addition and Fiedler 2000; Schimana et al. 2001). Using 2D NMR, the struc- to gyrase, have a second type II enzyme, topo IV, which is also tures of SD4 and SD8 were determined (Holzenkampfer et al. a relaxing/decatenating enzyme, and is also a target for antibi- 2002) and shown to consist of an aminocoumarin moiety linked otics. via a tetraene linker and olivose sugar to an angucyclinone DNA gyrase consists of two subunits, GyrA and GyrB, which polyketide moiety (Fig. 1). The presence of an aminocoumarin form an A B complex in the active enzyme (Fig. 2; Collin, 2 2 group and the discovery that some of the biosynthetic Karkare and Maxwell 2011; Bush, Evans-Roberts and Maxwell genes were related to those of the ‘classical’ aminocoumarin 2015). The supercoiling reaction involves the wrapping of DNA Downloaded from https://academic.oup.com/femsre/article/42/1/fux055/4604775 by DeepDyve user on 13 July 2022 102 FEMS Microbiology Reviews, 2018, Vol. 42, No. 1 antibiotic novobiocin (Galm et al. 2002; Trefzer et al. 2002;see ATP. This was directly tested using surface-plasmon resonance below), suggested that these compounds were likely to target (SPR) in which DNA was tethered to the chip surface and the bacterial DNA gyrase and that this was the likely cause of their binding of gyrase monitored in the absence and presence of SD8 antibacterial activity. (Flatman et al. 2005). The presence of SD8 at relatively low con- Simocyclinones have been studied most intensively as centrations (50 nM) blocked DNA binding. When different do- gyrase-inhibiting antibiotics, but the second section of this re- mains of gyrase were examined for their ability to bind SD8 us- view covers the role of SD8 as an effector molecule controlling ing SPR and isothermal titration calorimetry (ITC), it was found the activity of a transcription factor called SimR, responsible for that interaction occurred only with the N-terminal domain of linking the biosynthesis and export of SD8 in the producing or- GyrA (Flatman et al. 2005), which was already known to con- ganism, S. antibioticus. In addition, the SD8 precursor, 7-oxo-SD8, tain the binding site for the G-segment DNA (Morais Cabral et al. has been thoroughly characterized as a substrate for the enzyme 1997). These biochemical and biophysical experiments therefore SimC7, which reduces a carbonyl to a hydroxyl group at the C-7 supported the idea that simocyclinones act by binding to the position in the angucyclinone moiety of the molecule. This enzy- GyrA subunit at a DNA-binding site to prevent the binding of matic step, which is critical because it converts an almost inac- the enzyme to DNA; a completely novel mode of action. This tive precursor into the mature antibiotic, is covered in the third idea was later corroborated by X-ray crystallography (Edwards and final section of this review. et al. 2009b, Hearnshaw et al. 2014), see below. Binding of SD8 to the N-terminal domain of GyrA was also seen using circu- lar dichroism experiments (Sissi et al. 2010); this method also SIMOCYCLINONES AS ANTIBIOTICS showed evidence for a second binding site in the C-terminal do- main of GyrB, albeit of lower affinity than the GyrA-binding site. Activity of simocyclinones against bacteria Subsequent ITC experiments (Hearnshaw et al. 2014) also found In general, the antibiotic activity of simocyclinones was found to evidence for a binding site in GyrB, but estimated that it was be relatively weak, except against some Gram-positive bacterial ∼1000-fold weaker than the GyrA site; it is unlikely that the GyrB species, for example Bacillus brevis (MIC 10 μg/ml) and Strepto- site contributes to the activity of simocyclinones. myces viridochromogenes (MIC 1 μg/ml) (Schimana et al. 2000). Lit- SD8 has also been found to inhibit E. coli topo IV and human tle activity was detected against Gram-negative bacteria. This is topo II, albeit with a lower potency than against gyrase (Flatman almost certainly due to the inability of simocyclinones to pene- et al. 2005;Sadiq et al. 2009). In other work, SD8 was found to also trate the outer membrane, since imp mutants of E. coli,which are inhibit S. aureus gyrase, but was much less effective against topo specifically compromised in outer membrane integrity, become IV from E. coli and S. aureus (Oppegard et al. 2009); S. aureus gyrase sensitive to SD8 (Edwards et al. 2009b), although multidrug efflux wasfoundtobe3–4-foldlesssensitivetoSD8than E. coli gyrase. pumps like AcrB may also contribute to resistance (Oppegard Elsewhere, it was found that the difference in SD8 potencies be- et al. 2009). However, it has been pointed out that most of these tween these enzymes was ∼20-fold (Alt et al. 2011); however, it susceptibility tests have been carried out using stock lab strains, should be stressed that the absolute IC values are likely to be and SD8 has shown more promising activity against some clini- affected by assay conditions, which differ between the two en- cal isolates of E. coli and Klebsiella pneumoniae (Richter et al. 2010). zymes. Taken together, it seems that gyrase is the preferred tar- More recently, the discovery of new simocyclinones (Bilyk et al. get for simocyclinones, and that they act by binding to the GyrA 2016) and the capacity for engineering novel compounds, as has subunit of gyrase preventing the binding of DNA. been carried out with the classical aminocoumarins (Heide et al. 2008; Heide 2009, 2014) and to a limited extent with simocycli- nones (Anderle et al. 2007a,b), has raised the possibility of com- How simocylinone D8 binds to gyrase pounds with increased antibacterial potency. However, given that SD8 has been shown to inhibit human topo II (Flatman et al. Biochemical and biophysical data (described above) strongly 2005;Sadiq et al. 2009), the potential for mammalian toxicity suggested that the simocyclinones bind to GyrA in a region of the must be borne in mind. protein involved in DNA binding. This proposal was confirmed by X-ray crystallography. Crystallization trials using simocycli- none D8 (SD8) and the N-terminal domain of the DNA gyrase How simocyclinones inhibit DNA gyrase A protein (GyrA59), whose structure was already known (Morais The similarity between the structures of simocyclinones (Fig. 1) Cabral et al. 1997), gave diffracting crystals (Edwards et al. 2009a). and those of the classical aminocoumarins led to the expec- This first structure (initially solved at 2.6- A resolution) revealed tation that simocyclinones would inhibit bacterial DNA gyrase a tetramer of GyrA59 that consisted of two GyrA59 dimers cross- by competitively binding to the ATPase active site in the GyrB linked by four molecules of SD8 (Edwards et al. 2009b). Two bind- subunit. It was shown that simocyclinone D8 (and D4) did in- ing pockets were observed for SD8 in each subunit, both lying deed inhibit DNA supercoiling catalyzed by E. coli gyrase but, within the DNA-binding ‘saddle’ (Morais Cabral et al. 1997)ofthe surprisingly, also inhibited DNA relaxation (Flatman et al. 2005), GyrA59 dimer, one accommodating the aminocoumarin moiety an ATP-independent reaction. Moreover, ATPase assays showed and the other accommodating the angucyclinone moiety. Selec- that SD8 and SD4 did not inhibit this reaction under conditions tion of spontaneous SD8-resistant E. coli mutants showed that where novobiocin was effective. The most common mode of ac- the mutations occurred in both pockets, corroborating the crys- tion of topoisomerase-targeted drugs (e.g. fluoroquinolones) is tal structure (Edwards et al. 2009b). Further site-directed mutants the stabilization of the enzyme-DNA cleavage complex. It was also supported the structure, while others could not be fully shown that SD8 did not act in this way but was found to antago- rationalized (see below), suggesting that this structure might nize the ability of fluoroquinolones, and other agents, to induce not reflect the situation in vivo. Although the crystal structure cleavage-complex formation (Flatman et al. 2005). showed a protein tetramer, it was suspected that this dimer– Taken together, these data suggested that simocyclinones dimer interaction was stabilized in the crystal and may not rep- might interfere with the binding of gyrase to DNA rather than to resent the physiologically relevant form of the complex. Downloaded from https://academic.oup.com/femsre/article/42/1/fux055/4604775 by DeepDyve user on 13 July 2022 Buttner et al. 103 Figure 3. (a) Structure of E. coli DNA GyrA homodimer (55 kDa N-terminal fragment) with two molecules of SD8 bound; one subunit is colored blue and the other in yellow; the SD8 molecules are shown in two shades of green (PDB accession number 4CKL). A DNA duplex taken from a superposed structure of a Staphylococcus aureus gyrase-DNA-drug complex (PDB accession number: 2XCS) is also shown in pink to illustrate that SD8 would block the interaction of G-segment DNA with the DNA-binding ‘saddle’. (b) Top view of panel a, looking down the dimer 2-fold axis. (c) Enlarged view of the boxed region shown in panel b, with the SD8 ligands in stick representation with atom coloration (carbon, green; oxygen, red; nitrogen, blue, chlorine, gray). This clearly shows that the antibiotic spans the dimer interface with distinct binding pockets for the terminal angucyclinone (ANG) and aminocoumarin (AC) groups. (This figure and the other structural figures were p repared using CC4MG; McNicholas et al. 2011.) Figure 4. Schematic representation of the SD8-binding pocket of GyrA showing all residues within 4 A of the ligand, as revealed in the crystal structure of the GyrA-SD8 complex (PDB accession number 4CKL). One subunit is shown in blue and the other in yellow. Hydrogen bonds are indicated by dotted lines; van der Waal contacts are indicated by orange arcs, and water molecules are shown as filled blue circles labeled ‘W’. For clarity, all hydrogens have been omitted. Analysis of the SD8-GyrA59 complex using nanoelectro- discrete protein dimer with two SD8 molecules bound (Fig. 3; spray ionization mass spectrometry showed that the tetrameric Hearnshaw et al. 2014). This structure, solved at 2.05-Areso- species observed in the crystal could be reproduced in solution, lution, proved to be entirely consistent with all the mutations but only at high SD8 concentrations, while at lower concentra- to SD8 resistance that had been previously made or selected tions, a dimeric species was present with two SD8 molecules (Edwards et al. 2009b); additional mutants made in response to bound per dimer; this result was potentially at odds with the the revised structure were also shown to be consistent (Hearn- previous structural data (Edwards et al. 2009b). Further mass shaw et al. 2014). In addition to the new structure being dimeric, spectrometry suggested that the binding of SD8 to the pro- rather than tetrameric, the conformation of SD8 is significantly tein dimer showed strong allosteric cooperativity (Edwards et al. different, compared with the earlier structure: the orientation of 2011). A subsequent crystal structure of a shorter version of the the aminocoumarin within the aminocoumarin pocket is some- N-terminal domain of GyrA (GyrA55), which lacks residues that what different, while the angucyclinone ‘pocket’ has shifted stabilize dimer–dimer interactions in the tetramer, revealed a such that it now spans the interface between the two monomers Downloaded from https://academic.oup.com/femsre/article/42/1/fux055/4604775 by DeepDyve user on 13 July 2022 104 FEMS Microbiology Reviews, 2018, Vol. 42, No. 1 (Fig. 4) and thus could provide a structural explanation for the (SC4; Fig. 1), could dissociate SimR from its operators (Le et al. cooperative binding observed by mass spectrometry. This new 2009). Subsequently, crystal structures of SimR alone (apo; 1.95- position for the angucyclinone group suggests that the binding A resolution) (Le et al. 2011b), in complex with its operator DNA of SD8 effectively ‘staples’ the GyrA dimer closed so inhibiting (2.99-A resolution) (Le et al. 2011a), and in complex with either the conformational changes that need to occur upon DNA bind- SD8 or SC4 (both 2.3-A resolution) (Le et al. 2011b), showed how ing and cleavage (Hearnshaw et al. 2014). SimR binds its effector ligand and how ligand binding prevents The SD8-binding site on gyrase is close to the binding site of SimR from binding to its operator DNA. Unsurprisingly, there is the fluoroquinolone antibiotics (Laponogov et al. 2009;Bax et al. no similarity between the ligand-binding pockets in gyrase and 2010; Laponogov et al. 2010), raising the possibility of generat- SimR. ing hybrid compounds. To this end, a series of ciprofloxacin- aminocoumarin hybrids has been synthesized, designed to bind How SimR binds SD8 to the aminocoumarin pocket of SD8 and to the fluoroquinolone TFRs function as homodimers, with each subunit having two do- pocket (Austin et al. 2016); some of these compounds retain good inhibitory activity against gyrase. It remains to be seen whether mains, an N-terminal DNA-binding domain (DBD) containing a helix-turn-helix (HTH) motif, and a C-terminal ligand-binding such compounds can be developed as viable antibiotics. Also flavone-base analogs of simocyclinones have been made in or- domain (LBD) (Ramos et al. 2005;Yu et al. 2010; Cuthbertson and Nodwell 2013). The ligand-binding pocket of SimR is unusual; in der to bind to a hydrophobic cleft in the protein and further stabilize binding (Verghese et al. 2013). Although some of these other characterized TFRs, one ligand-binding pocket is typically contained within each subunit and so, for example, in the closely compounds are effective gyrase inhibitors, they also stabilize the gyrase-DNA cleavage complex and probably act via a mecha- related protein, ActR, there is only one ligand contact with the second subunit (Willems et al. 2008), while in TetR itself there nism involving DNA intercalation, i.e. they do not bind at the intended site. is none (Orth et al. 2000). In contrast, the ligand-binding pocket in SimR spans the two protein subunits, with the angucyclinone Taken together, we conclude that simocyclinones bind to the A subunit of DNA gyrase, in a region that is normally oc- of SD8 bound in one subunit, while the olivose sugar, tetraene cupied by the G-segment DNA (Figs 2–4) and prevent the ini- and aminocoumarin parts of the molecule are bound in the other (Le et al. 2011b)(Figs 5 and 6). This split binding pocket is tial interaction of DNA with the enzyme and thus all the en- suing catalytic events. This is quite distinct from the mode of ∼30 A in length, with SD8 bound in an extended conformation. Although SD8 has 19 atoms that could potentially participate in action of fluoroquinolones (cleavage-complex stabilization) and aminocoumarins (competitive inhibitors of ATP binding) and hydrogen bonding, there are only five direct hydrogen bonds be- tween SimR and SD8, three with the aminocoumarin and two raises the possibility of developing other agents that use this mode of action, which would be less likely to be cross-resistant with the angucyclinone (Fig. 6). However, the dearth of hydrogen bonding is compensated for by extensive van der Waals contacts to known antibiotics. with the protein along the length of the ligand (Fig. 6). The way cognate ligands are bound by TFRs is highly variable. For exam- ple, when the SimR-SD8 structure is compared with that of the SIMOCYCLINONE D8 AS A TRANSCRIPTION complex between the closely related TFR protein ActR and its FACTOR EFFECTOR MOLECULE cognate ligand, the antibiotic actinorhodin, the long axis of the actinorhodin molecule lies almost perpendicular to that of SD8 SD8 has been studied most intensively as a gyrase-inhibiting in the SimR–SD8 structure (Willems et al. 2008;Le et al. 2011b). antibiotic. However, it has also been characterized as an effec- tor molecule controlling the activity of a transcription factor called SimR, responsible for linking the biosynthesis and ex- How simocyclinone D8 prevents SimR from binding port of SD8 in the producing organism, S. antibioticus (Le et al. DNA 2009, 2011a,b). It is perhaps under-appreciated that antibiotics are often potentially toxic to the organisms that produce them Available evidence suggests that apo-TFRs sample a range of (Cundliffe 1989; Hopwood 2007). Therefore, producing organ- conformations in solution and that ligand binding simply cap- isms must have mechanisms to ensure that the antibiotic ex- tures one of these conformations, rather than inducing the con- port machinery is in place when antibiotic biosynthesis be- formational change (Reichheld, Yu and Davidson 2009;Yu et al. gins. The relevant mechanism in the simocyclinone producer 2010; Cuthbertson and Nodwell 2013). SimR-apo did not crys- is specified by two adjacent genes, simR and simX, which sit tallize in its DNA-binding form (apparent from the distance be- within the simocyclinone (sim) biosynthetic gene cluster (Galm tween its recognition helices), and indeed this is generally true et al. 2002; Trefzer et al. 2002;Le et al. 2009). The SimR and of TFR apo-proteins (Yu et al. 2010). However, comparison of the SimX proteins resemble the TetR/TetA repressor/efflux pump SimR-apo, SimR-SD8 and SimR-DNA structures provided clear pair found in a number of human pathogens, which confer resis- insight into the likely mechanism of ligand-mediated derepres- tance to clinically important tetracyclines (Chopra and Roberts sion. 2001). SimX is an efflux pump, a member of the major facilitator The ligand-binding sites of TFRs are remote from their superfamily, which exports simocyclinone from the producing DBDs and derepression generally involves allosteric mecha- organism. simX transcription is repressed by SimR, a TetR-family nisms (Ramos et al. 2005;Yu et al. 2010; Cuthbertson and Nod- transcriptional regulator (TFR) that binds to two separate opera- well 2013). Ligand-bound and DNA-bound structures have been tors in the intergenic region between the divergently transcribed determined for several TFRs, including QacR, DesT, CgmR and simR and simX genes (Le et al. 2009). Simocyclinone abolishes TetR itself, and in these cases conformational changes appear DNA-binding by SimR, thereby derepressing transcription of the to be transmitted largely within the same subunit (Orth, Saenger simX efflux pump gene, and this provides the mechanism that and Hinrichs 1999; Orth et al. 2000; Schumacher et al. 2001; 2002; couples the biosynthesis of simocyclinone to its export. It was Itou et al. 2010; Miller et al. 2010). Specifically, they suggest that also shown that the biosynthetic intermediate simocyclinone C4 ligand binding traps a conformational state in which the DBD (in Downloaded from https://academic.oup.com/femsre/article/42/1/fux055/4604775 by DeepDyve user on 13 July 2022 Buttner et al. 105 Figure 5. Comparison of the SimR-SD8 (a,b,c) and SimR-DNA (d,e,f) structures with one SimR subunit shown in blue and the other shown in yellow. The two recognition helices are highlighted in magenta and bound SD8 molecules are shown in green. Note that the ligand-binding pocket in SimR spans the two protein subunits, with the angucyclinone (ANG) end of SD8 bound in one subunit while the aminocoumarin (AC) end is bound in the other such that SD8 skewers the two subunits. Note also that in the apo form of SimR (structure not shown), Arg122 is buried in its cognate subunit; however, in the SimR-SD8 complex, each copy of this residue (shown as red sticks) projects across the dimer interface into a pocket in the surface of the opposing subunit. Arg122 is not ordered in the SimR-DNA structure. (PDB accession numbers: SimR-SD8: 2Y30; SimR-DNA: 3ZQL; SimR-apo: 2Y2Z). particular the HTH motif) is repositioned relative to the LBD such DNA-bound conformation and the derepressed ligand-bound that the two recognition helices in the homodimer are too far structure, and all five of these link the wrapping arm with the apart to bind appropriately in consecutive major grooves of the LBD of the other subunit. As a consequence, when the subunits DNA. In contrast, comparison of the repressive SimR-DNA struc- rotate in the ligand-bound form, the wrapping arm moves with ture with the derepressed SimR-SD8 structure shows that the them. Because the ligand-binding pocket passes through both relative dispositions of the LBDs and DBDs within each indi- subunits, SD8 effectively skewers the dimer, rigidifying the com- vidual SimR subunits remain essentially unchanged on ligand plex, and because it is a relatively hydrophobic molecule, SD8 binding. Instead, SD8 binding captures a conformation in which contributes to the hydrophobic core of the SimR dimer, stabi- there is a rigid-body rotation of one SimR subunit relative to the lizing the overall structure. In addition, in the apo and DNA- other, and this rigid-body rotation moves the recognition he- bound structures, the two SimR subunits present essentially lices ∼5 A further apart in the derepressed (SD8-bound) state, flat surfaces to one another, allowing them to rotate relative preventing DNA binding (Fig. 7). It may well be significant that to each other. In contrast, in the SD8-bound form, the side- the ligand-binding sites in the previously characterized TFRs are chain of Arg122 from each subunit projects across the dimer contained almost entirely within individual subunits, whereas interface into a pocket in the surface of the opposing sub- the ligand-binding pocket in SimR spans the two subunits. unit, potentially acting as locating pins to lock the subunits Two helices of the SimR LBD (α9-α10) form a wrapping arm together (Fig. 5). that folds around the LBD of the opposing subunit (Figs 5 and The biosynthetic intermediate simocylinone C4 (SC4) lacks 7). These two helices form the end of the ligand-binding pocket the aminocoumarin ring present in the mature antibiotic (Fig. 1) responsible for binding the angucyclinone of SD8 (Figs 5, 6 and is essentially inactive as a DNA gyrase inhibitor; the SD8 IC and 7), and the wrapping arm changes conformation in the is 0.1 μM, whereas the SC4 IC is >100 μM(Edwards et al. 2009b). ligand-bound state. Only five reciprocal inter-subunit hydrogen However, despite the absence of the aminocoumarin ring, SC4 bonds (i.e. 10 in total) are maintained between the repressive binds SimR and prevents it from binding DNA (Le et al. 2009). The Downloaded from https://academic.oup.com/femsre/article/42/1/fux055/4604775 by DeepDyve user on 13 July 2022 106 FEMS Microbiology Reviews, 2018, Vol. 42, No. 1 Figure 6. Schematic representation of the SD8-binding pocket of SimR showing all residues within 4 A of the ligand. One subunit is shown in blue and the other shown in yellow. Hydrogen bonds are indicated by dotted lines and van der Waal contacts are indicated by orange arcs. The two water molecules that link Gln136 to the olivose sugar are shown as filled blue circles labeled ‘W’. For clarity, all hydrogens have been omitted. structure of the SimR-SC4 complex has also been determined (Le Bilyk et al. (2016) sequenced the Kitasatospora sp. and Streptomyces et al. 2011b). Comparison of the SD8-SimR and SC4-SimR struc- sp. NRRL B-24484 biosynthetic clusters, there were no type I PKS tures shows that the two molecules bind SimR in the same way, genes present, and the tetraene was instead shown to be synthe- meaning the parts common to both molecules (the angucycli- sized by an iterative type II PKS. This type II PKS is also present none, tetraene and olivose sugar) occupy equivalent positions in in S. antibioticus,leaving theroleofthe typeIPKS unknown. To the binding pocket. SC4 is slightly less effective than SD8 at dere- date, only two biosynthetic enzymes have been characterized pressing SimR in vitro (Le et al. 2009) and this is probably a conse- biochemically: SimL and SimC7. SimL catalyses the presumed quence of the fewer favorable interactions that SC4 makes with last step in the pathway, acting as an amide bond-forming lig- the protein, due to the absence of the aminocoumarin. These ase that attaches the aminocoumarin to the tetraene linker (Luft results show that a pathway intermediate that is not an active et al. 2005;Pacholec et al. 2005;Anderle et al. 2007b). antibiotic can induce expression of the efflux pump, and simi- As noted above, the second enzyme, SimC7, was originally lar observations have been made in other antibiotic pathways, proposed to be involved in the biosynthesis of the tetraene particularly for actinorhodin (Otten, Ferguson and Hutchinson linker. It was subsequently shown to be an NAD(P)H-dependent 1995; Jiang and Hutchinson 2006;Ahn et al. 2007;Tahlan et al. ketoreductase that catalyzes the reduction of a carbonyl to a 2007;Ostash et al. 2008;Tahlan et al. 2008; Willems et al. 2008). hydroxyl group at the C-7 position of the angucyclinone, high- These data raise the possibility of a ‘feed-forward’ mechanism, lighting the dangers of relying on speculative gene annotations in which inactive intermediates ensure expression of the efflux (Fig. 8;Schafer ¨ et al. 2015). This enzymatic step is essential pump prior to the build-up of a toxic concentration of the poten- for antibiotic activity, converting the almost inactive 7-oxo-SD8 tially lethal mature antibiotic (Hopwood 2007;Tahlan et al. 2007; (IC ∼ 50–100 μM) into the potent gyrase inhibitor SD8 (IC ∼ 50 50 Le et al. 2009). 0.1–0.6 μM) (Schafer ¨ et al. 2015). Based on the intermediates pro- duced by S. antibioticus, it seems the biosynthesis of SD8 starts with assembly of the angucyclinone, followed by the attach- 7-OXO-SIMOCYCLINONE D8 AS A SUBSTRATE ment of the olivose sugar, and then the tetraene linker, and lastly the aminocoumarin (i.e. as drawn in Figs 1 and 8,SD8 is While the functions of most of the biosynthetic enzymes en- coded within the S. antibioticus sim cluster have been predicted assembled from right to left) (Schimana et al. 2001). Therefore, the natural substrate of SimC7 is probably a 7-oxo angucycli- (Galm et al. 2002; Trefzer et al. 2002), the biosynthetic pathway remains largely uncharacterized experimentally. This lack of none intermediate lacking the attached olivose sugar, tetraene linker and aminocoumarin, an intermediate that is detectable knowledge about the biosynthesis of simocyclinones is well il- lustrated by the tetraene moiety. Trefzer et al. (2002) proposed only in simC7 mutants (Schafer ¨ et al. 2015). Nevertheless, the enzyme readily accepts as a substrate the full-length interme- that the tetraene linker would be the product of the large modu- lar type I polyketide synthase (PKS), SimC1ABC, working in trans diate 7-oxo-SD8, the product made by simC7 mutants (Schafer et al. 2015). with two monofunctional enzymes, SimC6 and SimC7. Yet when Downloaded from https://academic.oup.com/femsre/article/42/1/fux055/4604775 by DeepDyve user on 13 July 2022 Buttner et al. 107 Figure 7. Structures of (a) SimR-SD8 and (b) SimR-DNA together with schematic representations illustrating the rigid-body rotation of the subunits relative to one another. To emphasize the subunit rotation, the position of the blue subunit is fixed in each panel so that the rotation of the yellow subunit accounts fo r all the movement in the dimer. The asterisk indicates the pivot point around which rotation occurs. Note that the net effect of subunit rotation is that the distance separating the two recognition helices increases to 41.7 A in the SD8-bound form, a distance incompatible with DNA binding. Note also that helices α9–α10 form a wrapping arm that engages the LBD of the opposing subunit and that these helices additionally form the angucyclinone end of the ligand-binding pocket. (PDB accession numbers: SimR-SD8: 2Y30; SimR-DNA: 3ZQL). SimC7 is a member of the short-chain dehydroge- that is characteristic of the so-called extended SDR subfamily nase/reductase (SDR) superfamily. These proteins have diverse (Kavanagh et al. 2008). This latter domain contains a ‘lid’ motif biochemical activities, including functioning as dehydratases, consisting of two anti-parallel α-helices that sits over the active reductases, epimerases, dehydrogenases and decarboxylases site. The apo, binary and ternary SimC7 structures are very sim- (Kallberg, Oppermann and Persson 2010; Persson and Kallberg ilar except for the orientation of this lid, which closes somewhat 2013). Classical SDR enzymes have a characteristic Ser-Tyr-Lys over the bound substrate (maximum Cα-Cα shift 5.35 A). The catalytic triad in their active site, in which the latter two underside of the lid forms part of the tight, highly hydrophobic residues form a YxxxK motif. The conserved tyrosine acts as a substrate binding pocket (Fig. 9) that provides the environment central acid-base catalyst that donates a proton to the substrate. needed for catalysis (Schafer ¨ et al. 2016). The adjacent lysine serves to lower the pKa of the tyrosine hydroxyl group and often contributes directly to a proton relay How SimC7 binds 7-oxo-SD8 mechanism. Lastly, the hydroxyl group of the serine polarizes the carbonyl group of the substrate (Kavanagh et al. 2008). In the SimC7 ternary complex with substrate and NADP bound, The catalytic mechanism of SimC7 was investigated because the angucyclinone ring system of 7-oxo-SD8 binds adjacent and it shares little sequence similarity with other characterized parallel to the nicotinamide ring of the cofactor (Fig. 9c), where ketoreductases, even with functionally analogous polyketide it adopts an essentially planar conformation. This differs from ketoreductases involved in the biosynthesis of related angucy- the conformations seen in the SimR-SD8 and gyrase-SD8 com- clinone antibiotics. Most of all, alignments of SimC7 with other plexes, where the A-ring of the angucyclinone in SD8 is tilted SDR proteins suggested that SimC7 lacked the classical catalytic upwards towards the epoxide (Le et al. 2011b, Hearnshaw et al. triad, including the tyrosine that acts as the central acid-base 2014;Schafer ¨ et al. 2016). The substrate pocket has several dis- catalyst in classical SDR proteins. This possibility was investi- tinctive characteristics (Fig. 9). The pocket is very hydrophobic gated by determining the structures of SimC7 alone (apo; 1.6-A and highly constricted in shape, features that are likely to en- resolution), the binary complex with NADP (1.95-A resolution) force the planar conformation on the angucyclinone ring sys- and the ternary complex with both NADP and 7-oxo-SD8 (1.2-A tem. Strikingly, within the hydrophobic pocket, 7-oxo-SD8 is resolution) (Schafer ¨ et al. 2016). As might be expected, there is bound by just one direct hydrogen bond, connecting the side- no similarity between the simocyclinone-binding pockets in chain of Ser95 and the C-7 carbonyl oxygen of the angucycli- gyrase, SimR and SimC7. ¨ none (Fig. 10;Schafer et al. 2016). However, even this single SimC7 has two domains (Fig. 9), a larger Rossmann-fold do- hydrogen bond is not required for enzymatic activity, since main that binds NADP and a smaller substrate-binding domain a constructed S95A variant shows almost wild-type levels of Downloaded from https://academic.oup.com/femsre/article/42/1/fux055/4604775 by DeepDyve user on 13 July 2022 108 FEMS Microbiology Reviews, 2018, Vol. 42, No. 1 Figure 8. SimC7 catalyzes the reduction of 7-oxo SD8 to simocyclinone D8. Labels A-D denote the four rings of the angucyclinone; the C-7 carbonyl/hydroxyl is highlighted in red. substrate conversion (Schafer ¨ et al. 2016). Thus, although this quinone-like C-ring and the phenyl-like D-ring of the angucy- hydrogen bond may help to position the C-7 carbonyl above the clinone promote the formation of an intramolecular hydrogen C-4 position of the nicotinamide ring ready for direct hydride bond between the proton on the C-8 hydroxyl and the oxygen of transfer, and provide additional polarization of the C-7 carbonyl the neighboring C-7 carbonyl (Fig. 11b). This intramolecular hy- group, as proposed for the structurally equivalent Ser or Thr drogen bond polarizes the carbonyl, enhancing the electrophilic- residues in classical SDR proteins (Kavanagh et al. 2008; Kallberg, ity of C-7 and making it a good acceptor for hydride attack from Oppermann and Persson 2010; Persson and Kallberg 2013), nei- the 4-pro-S position of the nicotinamide ring, which is only 3.0 ther proposed effect is crucial for catalysis. As discussed above, A away. Then, internal proton transfer from the neighboring the natural substrate for SimC7 is probably a 7-oxo angucycli- C-8 hydroxyl group forms the C-7 hydroxyl group, generating a none intermediate lacking the olivose sugar, tetraene linker and phenolate intermediate where the aromatic D-ring stabilizes the aminocoumarin. Consistent with this suggestion, only the an- negative charge on the C-8 oxygen. In the second step of the re- gucyclinone is buried in the active site of SimC7, with the rest of action, the phenolate intermediate leaves the substrate-binding the molecule projecting out of the enzyme (Fig. 9). Indeed, the pocket and the C-8 hydroxyl group re-forms by abstracting a pro- aminocoumarin and roughly half of the tetraene linker are not ton from bulk water (Fig. 11b), something that cannot happen resolved in electron density. within the confines of the active site. The hydrophobic active site cavity would accelerate expulsion of the charged phenolate intermediate created during catalysis. Lastly, the direct hydride How SimC7 converts 7-oxo-SD8 into SD8 attack from below the angucyclinone explains why simocycli- nones have 7S-stereochemistry. In summary, the SimC7 mecha- The structures confirmed the prediction made from sequence nism involves the intramolecular transfer of a substrate-derived alignments that SimC7 lacks a canonical SDR Ser-Tyr-Lys cat- proton to generate a phenolate intermediate, and this obviates alytic triad (Schafer ¨ et al. 2016). While the serine is conserved the need for proton transfer from a canonical SDR active-site (Ser95), the other two residues (the YxxxK motif), including the tyrosine. key tyrosine residue that acts as the acid/base catalyst in clas- sical SDR proteins, are replaced by Ile108 and His112, respec- tively (Fig. 11). The structures also demonstrate that there is Why is 7-oxo-SD8 almost inactive as a DNA gyrase no alternative residue that could act as an acid/base catalyst, inhibitor? and instead suggest that SimC7 has a novel reaction mecha- nism (Schafer ¨ et al. 2016). This unusual mechanism does not It is striking that SD8 is very potent as a gyrase inhibitor depend on catalytic residues in the protein, but instead ex- (IC ∼ 0.1–0.6 μM) and yet 7-oxo-SD8 is almost inactive (IC ∼ 50 50 ploits the chemical characteristics of 7-oxo-SD8 itself, and is 50–100 μM) (Schafer ¨ et al. 2015). Why does such a small struc- thus a new example of substrate-assisted catalysis (Dall’Acqua tural difference, the presence of a carbonyl group at the C-7 po- and Carter 2000). In the first step, the hydrophobic environment sition in 7-oxo-SD8 (Fig. 8), have such a drastic effect on the of the substrate-binding pocket and the juxtaposition of the antibiotic activity of the molecule? The likely answer becomes Downloaded from https://academic.oup.com/femsre/article/42/1/fux055/4604775 by DeepDyve user on 13 July 2022 Buttner et al. 109 Figure 9. (a) and (b) Crystal structure of the SimC7 ternary complex with NADP and 7-oxo-SD8. The nucleotide-binding domain, the substrate-binding domain and the lid motif are shown in yellow, blue and magenta, respectively. NADP is shown in pink and 7-oxo-SD8 is shown in green. For the latter, only the crystallographically resolved atoms are shown, i.e. the angucyclinone, the olivose and roughly half of the tetraene linker. (c) Close-up showing the active site of the ternary complex including the Ser95-Ile108-His112 ‘catalytic triad’ residues, and Asn137, which is important in maintaining the syn-conformation of the cofactor. C-4 of the cofactor nicotinamide ring and C-7 of the substrate are highlighted by black spheres, which are 3 A apart, indicating that the substrate is exactly positioned for direct hydride transfer. (d) Cross-section through the active site pocket, showing how tightly the cofactor (pink) and substrate (green) are bound. For clarity, only the nicotinamide ribosyl moiety of the cofactor is shown in panel d, and only the angucyclinone moiety of the substrate is shown in panels c and d (PDB accession number: 5L4L). clear from analysis of the structure of the GyrA-SD8 complex: ring system, which may well affect other bonding interactions both the C-7 and C-8 hydroxyls are involved in a hydrogen bond- with GyrA. ing network that helps secure the angucyclinone in its binding pocket (Fig. 4). However, in 7-oxo-SD8, an intramolecular hy- CONCLUDING REMARKS drogen bond between the C-7 carbonyl and the C-8 hydroxyl is preferred over these intermolecular interactions and this si- In the three different systems we have described in this review, multaneously breaks the direct contact between the angucycli- the interaction of the ligand with the protein has entirely differ- none and His80 and the indirect contacts with Pro79 and Arg121 ent downstream consequences. For gyrase, it results in inhibi- (Fig. 4). His80, in particular, is known to play a crucial role in tion, leading to cell death, for SimR, it results in derepression, binding simocyclinone, since mutating this residue to alanine leading to antibiotic export, and for SimC7, it results in cataly- causes a 230-fold increase in the IC of SD8 for gyrase (Edwards 50 sis, leading to potentiation of an antibiotic. Given that SimC7 is et al. 2009b). In addition, the presence of a carbonyl group at an enzyme, the interaction with the ligand is transient, whereas C-7 would alter the overall conformation of the angucyclinone the interaction with gyrase and SimR will be much longer-lived. Downloaded from https://academic.oup.com/femsre/article/42/1/fux055/4604775 by DeepDyve user on 13 July 2022 110 FEMS Microbiology Reviews, 2018, Vol. 42, No. 1 The extensive nature of these double-headed interactions leads to very tight binding, commensurate with the physiological con- sequences. Indeed, molecules lacking either the angucyclinone or the aminocoumarin bind much more weakly to DNA gyrase and, as a consequence, the potency of SD8 as an antibiotic is severely compromised through loss of either ‘warhead’ (Edwards et al. 2009b). The proportion of hydrogen bonds is highest for the complex with gyrase because the binding site is largely solvent- exposed and would otherwise interact with the G-segment DNA, which is polar. In SimR, the ligand-binding site threads through the hydrophobic core of the homodimer, and so the interac- tions are dominated by van der Waals contacts. In contrast, in the 7-oxo-SD8 complex with SimC7, only the angucyclinone in- teracts with the enzyme, this being consistent with the site of ketoreduction and the expectation that the natural substrate in vivo is the angucyclinone alone. Given the necessity to pre- cisely position the SimC7 substrate for catalysis, the dearth of hydrogen bonds seems counterintuitive. Indeed, a Ser95 to Ala substitution that removes the only hydrogen bond shows that even this is dispensable. However, the necessity to provide a hy- drophobic environment for efficient catalysis would be consis- Figure 10. Schematic representation of the hydrophobic substrate-binding tent with a paucity of hydrogen bonding partners and bound pocket of SimC7 showing all residues within 4 A of the ligand, as revealed in the water molecules. Instead, the highly constrained nature of the crystal structure of the SimC7 ternary complex with NADP and 7-oxo-SD8 (PDB accession number 5L4L). Hydrogen bonds are indicated by dotted lines and van SimC7 active site is a key factor in sterically guiding the sub- der Waal contacts are indicated by orange arcs. Note that the substrate is bound strate to its catalytically competent position adjacent to the by only one direct hydrogen bond, connecting the C-7 carbonyl of the angucy- cofactor with hydride donor and hydride acceptor atoms jux- clinone and the side-chain hydroxyl of Ser95. This hydrogen bond may assist in taposed. The transient nature of this interaction would be pro- positioning the substrate and facilitate the reaction. However, this interaction is moted by the negative charge that develops on the phenolate not required for enzymatic activity, since a constructed S95A variant of SimC7 intermediate, which would be unfavorable in the hydrophobic shows near wild-type enzymatic activity (Schafer ¨ et al. 2016). Note that one face of the pocket is formed by the NADP cofactor itself. In the natural SimC7 sub- active site, and possibly also by the increased puckering of the strate, R = H; in the substrate used here, R includes the deoxysugar, tetraene angucyclinone ring system that would occur when the C7 keto linker and the aminocoumarin. For clarity, all hydrogens have been omitted. group is reduced to a hydroxyl. Finally, although SD8 itself is not viable as a clinical antibi- In both gyrase and SimR, there are a substantial number of otic, due at least in part to its poor penetration into bacteria, the way in which it inhibits DNA gyrase is unique. It therefore has interactions with the terminal aminocoumarin and angucycli- none groups, which are bound by separate subunits; addition- the potential to guide the development of new, clinically rele- vant compounds acting against this enzyme, and the detailed ally, there are a handful of contacts involving the linker region. Figure 11. Comparison of the canonical SDR ketoreduction mechanism with the novel SimC7 reaction mechanism. (a) In classical SDR proteins, the conserved active site tyrosine serves as a central acid-base catalyst that donates a proton to the substrate. The adjacent lysine residue lowers the pKa of the tyrosine hydroxyl group and often contributes directly to the proton relay mechanism; the hydroxyl group of the serine stabilizes and polarizes the carbonyl group of the substrate. (b) SimC7 has an atypical catalytic triad consisting of Ser95, Ile108 and His112. In the first step of the SimC7 mechanism, the C-7 carbonyl group of the angucyclinon e is reduced by transfer of the 4-pro-S hydride of the cofactor onto the C-7 carbon of the substrate. This transfer from below the C-ring results in the characteristic 7-S-stereochemistry of SD8. Ketoreduction at position C-7 is completed by intramolecular proton transfer from the neighboring C-8 hydroxyl group of the angucyclinone; the resultant negative charge on C-8 is stabilized by the adjacent aromatic ring system. In the second step, the C-8 phenolate intermediate regains a proton from bulk water after leaving the substrate binding pocket. In the natural SimC7 substrate, R = H; in the substrate used here, R includes the deoxysugar, tetraene linker and the aminocoumarin. Note that there are no water molecules in the active site pocket that could contribute to the reaction mechanism. In the ternary complex, the nearest ˚ ˚ water to O-7 of the angucyclinone is ∼5.5 A away, and the nearest water to O-8 is ∼4.9 A away. Because of steric constraints within the pocket, neither could approach the substrate oxygen atoms without either a repositioning of the substrate or a conformational change in the protein. Downloaded from https://academic.oup.com/femsre/article/42/1/fux055/4604775 by DeepDyve user on 13 July 2022 Buttner et al. 111 structural information available should potentiate such devel- Dall’Acqua W, Carter P. Substrate-assisted catalysis: molecular opment. basis and biological significance. Protein Sci 2000;9:1–9. Edwards MJ, Flatman RH, Mitchenall LA et al. Crystallization and preliminary X-ray analysis of a complex formed between the ACKNOWLEDGEMENTS antibiotic simocyclinone D8 and the DNA breakage-reunion domain of Escherichia coli DNA gyrase. Acta Crystallogr Sect F We thank Hans-Peter Fiedler for providing simocyclinone D8. Struct Biol Cryst Commun 2009a;F65:846–8. Edwards MJ, Flatman RH, Mitchenall LA et al. A crystal struc- FUNDING ture of the bifunctional antibiotic simocyclinone D8, bound to DNA gyrase. Science 2009b;326:1415–8. This work was funded by a BBSRC studentship to M.S., Edwards MJ, Williams MA, Maxwell A et al. Mass spectrom- BBSRC grant BB/I002197/1 to M.J.B. and D.M.L., BBSRC grant etry reveals that the antibiotic simocyclinone D8 binds BB/I002049/1 to A.M. and D.M.L. and by BBSRC Institute Strategic to DNA gyrase in a “bent-over” conformation: evidence Programme Grant BB/J004561/1 to the John Innes Centre. of positive cooperativity in binding. Biochemistry 2011;50: Conflict of interest. None declared. 3432–40. Flatman RH, Howells AJ, Heide L et al. Simocyclinone D8, an in- hibitor of DNA gyrase with a novel mode of action. Antimicrob REFERENCES Agents Chemother 2005;49:1093–100. Ahn SK, Tahlan K, Yu Z et al. Investigation of transcription re- Forterre P, Gribaldo S, Gadelle D et al. Origin and evolution of pression and small-molecule responsiveness by TetR-like DNA topoisomerases. Biochimie 2007;89:427–46. transcription factors using a heterologous Escherichia coli- Galm U, Schimana J, Fiedler HP et al. Cloning and analysis of the based assay. J Bacteriol 2007;189:6655–64. simocyclinone biosynthetic gene cluster of Streptomyces an- Alt S, Mitchenall LA, Maxwell A et al. Inhibition of DNA gy- tibioticus Tu 6040. Arch Microbiol 2002;178:102–14. rase and DNA topoisomerase IV of Staphylococcus aureus and Harms A, Stanger FV, Scheu PD et al. Adenylylation of gyrase and Escherichia coli by aminocoumarin antibiotics. J Antimicrob Topo IV by FicT toxins disrupts bacterial DNA topology. Cell Chemother 2011;66:2061–9. Rep 2015;12:1497–507. Anderle C, Hennig S, Kammerer B et al. Improved mutasynthetic Hearnshaw SJ, Edwards MJ, Stevenson CE et al. Anew crys- approaches for the production of modified aminocoumarin tal structure of the bifunctional antibiotic simocyclinone D8 antibiotics. Chem Biol 2007b;14:955–67. bound to DNA gyrase gives fresh insight into the mechanism Anderle C, Li SM, Kammerer B et al. New aminocoumarin antibi- of inhibition. J Mol Biol 2014;426:2023–33. otics derived from 4-hydroxycinnamic acid are formed after Heide L. Genetic engineering of antibiotic biosynthesis for heterologous expression of a modified clorobiocin biosyn- the generation of new aminocoumarins. Biotechnol Adv thetic gene cluster. J Antibiot (Tokyo) 2007a;60:504–10. 2009;27:1006–14. Austin MJ, Hearnshaw SJ, Mitchenall LA et al. A natural product Heide L. New aminocoumarin antibiotics as gyrase inhibitors. Int inspired fragment-based approach towards the development J Med Microbiol 2014;304:31–36. of novel anti-bacterial agents. MedChemComm 2016;7:1387– Heide L, Gust B, Anderle C et al. Combinatorial biosynthesis, metabolic engineering and mutasynthesis for the genera- Bax BD, Chan PF, Eggleston DS et al. Type IIA topoisomerase tion of new aminocoumarin antibiotics. Curr Top Med Chem inhibition by a new class of antibacterial agents. Nature 2008;8:667–79. 2010;466:935–40. Holzenkampfer M, Walker M, Zeeck A et al. Simocyclinones, Bilyk O, Brotz E, Tokovenko B et al. New Simocyclinones: sur- novel cytostatic angucyclinone antibiotics produced by Strep- prising evolutionary and biosynthetic insights. ACS Chem Biol tomyces antibioticus Tu 6040 II. Structure elucidation and 2016;11:241–50. biosynthesis. J Antibiot (Tokyo) 2002;55:301–7. Boucher HW, Talbot GH, Bradley JS et al. Bad bugs, no drugs: no Hopwood DA. How do antibiotic-producing bacteria ensure their ESKAPE! An update from the Infectious Diseases Society of self-resistance before antibiotic biosynthesis incapacitates America. Clin Infect Dis 2009;48:1–12. them? Mol Microbiol 2007;63:937–40. Bush K, Courvalin P, Dantas G et al. Tackling antibiotic resistance. Itou H, Watanabe N, Yao M et al. Crystal structures of the mul- Nat Rev Microbiol 2011;9:894–6. tidrug binding repressor Corynebacterium glutamicum CgmR Bush NG, Evans-Roberts K, Maxwell A. DNA topoiso- in complex with inducers and with an operator. J Mol Biol merases. EcoSal Plus 2015;6. http://www.ncbi.nlm.nih.gov/ 2010;403:174–84. pubmed/26435256. Jiang H, Hutchinson CR. Feedback regulation of doxoru- Chopra I, Roberts M. Tetracycline antibiotics: mode of action, ap- bicin biosynthesis in Streptomyces peucetius. Res Microbiol plications, molecular biology, and epidemiology of bacterial 2006;157:666–74. resistance. Microbiol Mol Biol Rev 2001;65:232–60; second page, Kallberg Y, Oppermann U, Persson B. Classification of the short- tableofcontents. chain dehydrogenase/reductase superfamily using hidden Clardy J, Fischbach MA, Walsh CT. New antibiotics from bacterial Markov models. FEBS J 2010;277:2375–86. natural products. Nat Biotechnol 2006;24:1541–50. Kavanagh KL, Jornvall H, Persson B et al. Medium- and short- Collin F, Karkare S, Maxwell A. Exploiting bacterial DNA gyrase chain dehydrogenase/reductase gene and protein families: as a drug target: current state and perspectives. Appl Microbiol the SDR superfamily: functional and structural diversity Biotechnol 2011;92:479–97. within a family of metabolic and regulatory enzymes. Cell Mol Cundliffe E. How antibiotic-producing organisms avoid suicide. Life Sci 2008;65:3895–906. Annu Rev Microbiol 1989;43:207–33. Laponogov I, Pan XS, Veselkov DA et al. Structural basis of gate- Cuthbertson L, Nodwell JR. The TetR family of regulators. Micro- DNA breakage and resealing by type II topoisomerases. PLoS biol Mol Biol Rev 2013;77:440–75. ONE 2010;5:e11338. Downloaded from https://academic.oup.com/femsre/article/42/1/fux055/4604775 by DeepDyve user on 13 July 2022 112 FEMS Microbiology Reviews, 2018, Vol. 42, No. 1 Laponogov I, Sohi MK, Veselkov DA et al. Structural insight into Reichheld SE, Yu Z, Davidson AR. The induction of folding coop- the quinolone-DNA cleavage complex of type IIA topoiso- erativity by ligand binding drives the allosteric response of merases. Nat Struct Mol Biol 2009;16:667–9. tetracycline repressor. P Natl Acad Sci USA 2009;106:22263–8. Le TB, Fiedler HP, den Hengst CD et al. Coupling of the biosyn- Richter SN, Frasson I, Palumbo M et al. Simocyclinone D8 turns thesis and export of the DNA gyrase inhibitor simocyclinone on against Gram-negative bacteria in a clinical setting. Bioorg in Streptomyces antibioticus. Mol Microbiol 2009;72:1462–74. Med Chem Lett 2010;20:1202–4. Le TB, Schumacher MA, Lawson DM et al. The crystal structure Sadiq AA, Patel MR, Jacobson BA et al. Anti-proliferative effects of of the TetR family transcriptional repressor SimR bound to simocyclinone D8 (SD8), a novel catalytic inhibitor of topoi- DNA and the role of a flexible N-terminal extension in minor somerase II. Invest New Drugs. 2009. groove binding. Nucleic Acids Res 2011a;39:9433–47. Schafer M, Le TB, Hearnshaw SJ et al. SimC7 Is a Novel NAD(P)H- Le TB, Stevenson CE, Fiedler HP et al. Structures of the TetR- dependent ketoreductase essential for the antibiotic activ- like simocyclinone efflux pump repressor, SimR, and the ity of the DNA gyrase inhibitor simocyclinone. J Mol Biol mechanism of ligand-mediated derepression. J Mol Biol 2015;427:2192–204. 2011b;408:40–56. Schafer ¨ M, Stevenson CE, Wilkinson B et al. Substrate-assisted Luft T, Li SM, Scheible H et al. Overexpression, purification and catalysis in polyketide reduction proceeds via a phenolate characterization of SimL, an amide synthetase involved in intermediate. Cell Chem Biol 2016;23:1091–7. simocyclinone biosynthesis. Arch Microbiol 2005;183:277–85. Schimana J, Fiedler HP, Groth I et al. Simocyclinones, novel cy- Maxwell A, Lawson DM. The ATP-binding site of type II topoiso- tostatic angucyclinone antibiotics produced by Streptomyces merases as a target for antibacterial drugs. Curr Top Med Chem antibioticus Tu 6040. I. Taxonomy, fermentation, isolation and 2003;3:283–303. biological activities. J Antibiot (Tokyo) 2000;53:779–87. McNicholas S, Potterton E, Wilson KS et al. Presenting your struc- Schimana J, Walker M, Zeeck A et al. Simocyclinones: diversity of tures: the CCP4mg molecular-graphics software. Acta Crystal- metabolites is dependent on fermentation conditions. JInd logr D Biol Crystallogr 2011;67:386–94. Microbiol Biotechnol 2001;27:144–8. Miller DJ, Zhang YM, Subramanian C et al. Structural basis for the Schumacher MA, Miller MC, Grkovic S et al. Structural mecha- transcriptional regulation of membrane lipid homeostasis. nisms of QacR induction and multidrug recognition. Science Nat Struct Mol Biol 2010;17:971–5. 2001;294:2158–63. Morais Cabral JH, Jackson AP, Smith CV et al. Crystal struc- Schumacher MA, Miller MC, Grkovic S et al. Structural basis for ture of the breakage-reunion domain of DNA gyrase. Nature cooperative DNA binding by two dimers of the multidrug- 1997;388:903–6. binding protein QacR. EMBO J 2002;21:1210–8. Oppegard LM, Hamann BL, Streck KR et al. In vivo and in vitro Sissi C, Vazquez E, Chemello A et al. Mapping simocyclinone D8 patterns of the activity of simocyclinone D8, an angucy- interaction with DNA gyrase: evidence for a new binding site clinone antibiotic from Streptomyces antibioticus. Antimicrob on GyrB. Antimicrob Agents Chemother 2010;54:213–20. Agents Chemother 2009;53:2110–9. Tahlan K, Ahn SK, Sing A et al. Initiation of actinorhodin export Orth P, Saenger W, Hinrichs W. Tetracycline-chelated Mg2+ ion in Streptomyces coelicolor. Mol Microbiol 2007;63:951–61. initiates helix unwinding in Tet repressor induction. Biochem- Tahlan K, Yu Z, Xu Y et al. Ligand recognition by ActR, a TetR-like istry 1999;38:191–8. regulator of actinorhodin export. J Mol Biol 2008;383:753–61. Orth P, Schnappinger D, Hillen W et al. Structural basis of Theobald U, Schimana J, Fiedler HP. Microbial growth and pro- gene regulation by the tetracycline inducible Tet repressor- duction kinetics of Streptomyces antibioticus Tu 6040. Antonie operator system. Nat Struct Biol 2000;7:215–9. Van Leeuwenhoek 2000;78:307–13. Ostash I, Ostash B, Luzhetskyy A et al. Coordination of export Trefzer A, Pelzer S, Schimana J et al. Biosynthetic gene cluster and glycosylation of landomycins in Streptomyces cyanogenus of simocyclinone, a natural multihybrid antibiotic. Antimicrob S136. FEMS Microbiol Lett. 2008;285:195–202. Agents Chemother 2002;46:1174–82. Otten SL, Ferguson J, Hutchinson CR. Regulation of daunorubicin Verghese J, Nguyen T, Oppegard LM et al. Flavone-based ana- production in Streptomyces peucetius by the dnrR2 locus. JBac- logues inspired by the natural product simocyclinone D8 as teriol 1995;177:1216–24. DNA gyrase inhibitors. Bioorg Med Chem Lett 2013;23:5874–7. Pacholec M, Freel Meyers CL, Oberthur M et al. Characterization Vos SM, Tretter EM, Schmidt BH et al. All tangled up: how cells of the aminocoumarin ligase SimL from the simocyclinone direct, manage and exploit topoisomerase function. Nat Rev pathway and tandem incubation with NovM,P,N from the Mol Cell Biol 2011;12:827–41. novobiocin pathway. Biochemistry 2005;44:4949–56. Wang JC. DNA topoisomerases. Annu Rev Biochem 1996;65:635–92. Persson B, Kallberg Y. Classification and nomenclature of the su- Wang JC. Cellular roles of DNA topoisomerases: a molecular per- perfamily of short-chain dehydrogenases/reductases (SDRs). spective. Nat Rev Mol Cell Biol 2002;3:430–40. Chem Biol Interact 2013;202:111–5. Willems AR, Tahlan K, Taguchi T et al. Crystal structures of the Pommier Y, Leo E, Zhang H et al. DNA topoisomerases and their Streptomyces coelicolor TetR-like protein ActR alone and in poisoning by anticancer and antibacterial drugs. Chem Biol complex with actinorhodin or the actinorhodin biosynthetic 2010;17:421–33. precursor (S)-DNPA. J Mol Biol 2008;376:1377–87. Ramos JL, Martinez-Bueno M, Molina-Henares AJ et al. The TetR Yu Z, Reichheld SE, Savchenko A et al. A comprehensive analysis family of transcriptional repressors. Microbiol Mol Biol Rev of structural and sequence conservation in the TetR family 2005;69:326–56. transcriptional regulators. J Mol Biol 2010;400:847–64. http://www.deepdyve.com/assets/images/DeepDyve-Logo-lg.png FEMS Microbiology Reviews Oxford University Press

Structural insights into simocyclinone as an antibiotic, effector ligand and substrate

Download PDF
13 pages

Loading next page...
 
/lp/oxford-university-press/structural-insights-into-simocyclinone-as-an-antibiotic-effector-c2XH65uPyq

References (86)

Publisher
Oxford University Press
Copyright
Copyright © 2022 Federation of European Microbiological Societies
ISSN
0168-6445
eISSN
1574-6976
DOI
10.1093/femsre/fux055
pmid
29126195
Publisher site
See Article on Publisher Site

Abstract

Downloaded from https://academic.oup.com/femsre/article/42/1/fux055/4604775 by DeepDyve user on 13 July 2022 FEMS Microbiology Reviews, fux055, 42, 2018, 100–112 doi: 10.1093/femsre/fux055 Advance Access Publication Date: 8 November 2017 Review article REVIEW ARTICLE Structural insights into simocyclinone as an antibiotic, effector ligand and substrate 1,† 2,‡ 3,# Mark J. Buttner ,MartinSchafer ¨ , David M. Lawson 3,§ and Anthony Maxwell Department of Molecular Microbiology, John Innes Centre, Norwich Research Park, Norwich NR4 7UH, UK, 2 3 Department of Biochemistry, Duke University School of Medicine, Durham, NC 27710, USA and Department of Biological Chemistry, John Innes Centre, Norwich Research Park, Norwich NR4 7UH, UK Corresponding author: Department of Biological Chemistry, John Innes Centre, Norwich Research Park, Norwich NR4 7UH, UK. Tel: +44 1603 450771; E-mail: tony.maxwell@jic.ac.uk One sentence summary: Simocyclinones are actinomycete natural products that target bacterial DNA gyrase; structural work has revealed their molecular mode of action, aspects of their biosynthesis and the mechanism underlying their inducible export. Editor: Kenn Gerdes Mark J. Buttner, http://orcid.org/0000-0002-3505-2981 Martin Schafer ¨ , http://orcid.org/0000-0003-0891-0626 David M. Lawson, http://orcid.org/0000-0002-7637-4303 Anthony Maxwell, http://orcid.org/0000-0002-5756-6430 ABSTRACT Simocyclinones are antibiotics produced by Streptomyces and Kitasatospora species that inhibit the validated drug target DNA gyrase in a unique way, and they are thus of therapeutic interest. Structural approaches have revealed their mode of action, the inducible-efflux mechanism in the producing organism, and given insight into one step in their biosynthesis. The crystal structures of simocyclinones bound to their target (gyrase), the transcriptional repressor SimR and the biosynthetic enzyme SimC7 reveal fascinating insight into how molecular recognition is achieved with these three unrelated proteins. Keywords: antibiotics; Streptomyces; DNA gyrase; aminocoumarins; DNA topoisomerases; transcription factor INTRODUCTION ral products that were first isolated nearly 20 years ago from Streptomyces antibioticus Tu¨ 6040, which produces simocycli- The increase in resistance to antimicrobials has become a seri- nones A1, B1, B2, C2, C4, D4, D6, D7 and D8 (Schimana et al. 2000, ous challenge in the 21st Century, with rising antibiotic-resistant 2001). More recently, Kitasatospora sp. and Streptomyces sp. NRRL pathogens, particularly in hospital settings, and a paucity of B-24484 have been identified as producers of the novel simocy- new agents becoming available (Boucher et al. 2009; Bush et al. clinones D9, D10 and D11 (Bilyk et al. 2016); representative simo- 2011). It is therefore essential that we continue our search cyclinones are shown in Fig. 1. As most work has been carried for new antibacterial compounds, particularly novel natural out on simocyclinone D8 (SD8), it will be the main topic of this products, which have the possibility of exploiting new chemi- review. cal space. Actinomycetes, most notably the genus Streptomyces, The target of simocyclinones is the type II DNA topoiso- have proved to be a rich source of bio-active molecules, with merase, DNA gyrase (Fig. 2). DNA topoisomerases (topos) are most antibiotics in current clinical use being actinomycete nat- enzymes found in all organisms that catalyze the intercon- ural products or their derivatives (Clardy, Fischbach and Walsh versions of different topological forms of DNA, e.g. relaxed– 2006). In this review, we discuss the simocyclinones, natu- Received: 12 September 2017; Accepted: 7 November 2017 FEMS 2017. This is an Open Access article distributed under the terms of the Creative Commons Attribution License (http://creativecommons.org/ licenses/by/4.0/), which permits unrestricted reuse, distribution, and reproduction in any medium, provided the original work is properly cited. 100 Downloaded from https://academic.oup.com/femsre/article/42/1/fux055/4604775 by DeepDyve user on 13 July 2022 Buttner et al. 101 Figure 1. Chemical structures of simocyclinones D8, D4 and C4. around the A B complex and the passage of one segment of 2 2 DNA, termed the ‘T’ or ‘transported’ segment, through a double- stranded break in another, the ‘G’ or ‘gate’ segment. Catalytic supercoiling requires the hydrolysis of ATP. As this reaction pro- ceeds via transient double-strand breaks in DNA, agents that can stabilize the broken DNA intermediate, such as the fluo- roquinolones, are very effective antibacterial agents. A number of other compounds inhibit gyrase (and other topoisomerases) via this ‘cleavage-complex stabilization’ mechanism (Collin, Karkare and Maxwell 2011; Bush, Evans-Roberts and Maxwell 2015). In addition, gyrase and other type II topoisomerases can be inhibited by compounds that act at the ATP-binding site Figure 2. Schematic representation of the DNA gyrase A B complex with bound (Maxwell and Lawson 2003); this includes aminocoumarin an- 2 2 G-segment DNA. Each subunit consists of an N-terminal domain (NTD) and a tibiotics, such as novobiocin. As will be shown below, the inhibi- C-terminal domain (CTD). For reference, the DNA gyrase structure shown in Fig. 3 tion of gyrase by simocyclinones occurs by a different, previously corresponds to a homodimer of a 55 kDa fragment of the GyrA NTD. unknown, mechanism: they prevent the enzyme from binding DNA. It is possible that this mode of action can be exploited to- wards the development of novel, clinically relevant antibiotics. supercoiled, knotted–unknotted, catenated–decatenated (Vos It is interesting to note that there are a number of peptide and et al. 2011; Bush, Evans-Roberts and Maxwell 2015)and are protein inhibitors that exhibit the three mechanisms of gyrase essential for DNA replication and transcription (Wang 2002). inhibition (Collin, Karkare and Maxwell 2011). For example, mi- They are classified as type I or II depending upon whether their crocin B17, CcdB and ParE can stabilize the cleavage complex, reactions proceed via single- or double-stranded breaks in DNA, MfpA and Qnr proteins seem to prevent DNA binding, and FicT and further divided into sub-types: IA, B, C and IIA and B, de- proteins modify GyrB and prevent ATPase activity (Harms et al. pending of mechanistic and evolutionary considerations (Wang 2015). 1996; Forterre et al. 2007). DNA gyrase (the target of simocycli- Simocyclinones (D4 and D8) were discovered during the nones) is a type IIA topoisomerase, and the only enzyme that search for novel secondary metabolites from Streptomyces strains can catalyze the introduction of negative supercoils into DNA. derived from soil samples (Schimana et al. 2000). These com- It is essential in all bacteria but lacking from animals, includ- pounds showed antibiotic activity against certain Gram-positive ing humans, making it an ideal target for antibiotics. The type bacteria and cytotoxic effects on tumor cell lines. By varying mi- II topoisomerase in humans, topo II, has been developed as an crobial growth and fermentation conditions the yield of these anti-cancer target (Pommier et al. 2010) and can relax and de- compounds was analyzed and optimized (Theobald, Schimana catenate DNA but cannot supercoil. Most bacteria, in addition and Fiedler 2000; Schimana et al. 2001). Using 2D NMR, the struc- to gyrase, have a second type II enzyme, topo IV, which is also tures of SD4 and SD8 were determined (Holzenkampfer et al. a relaxing/decatenating enzyme, and is also a target for antibi- 2002) and shown to consist of an aminocoumarin moiety linked otics. via a tetraene linker and olivose sugar to an angucyclinone DNA gyrase consists of two subunits, GyrA and GyrB, which polyketide moiety (Fig. 1). The presence of an aminocoumarin form an A B complex in the active enzyme (Fig. 2; Collin, 2 2 group and the discovery that some of the biosynthetic Karkare and Maxwell 2011; Bush, Evans-Roberts and Maxwell genes were related to those of the ‘classical’ aminocoumarin 2015). The supercoiling reaction involves the wrapping of DNA Downloaded from https://academic.oup.com/femsre/article/42/1/fux055/4604775 by DeepDyve user on 13 July 2022 102 FEMS Microbiology Reviews, 2018, Vol. 42, No. 1 antibiotic novobiocin (Galm et al. 2002; Trefzer et al. 2002;see ATP. This was directly tested using surface-plasmon resonance below), suggested that these compounds were likely to target (SPR) in which DNA was tethered to the chip surface and the bacterial DNA gyrase and that this was the likely cause of their binding of gyrase monitored in the absence and presence of SD8 antibacterial activity. (Flatman et al. 2005). The presence of SD8 at relatively low con- Simocyclinones have been studied most intensively as centrations (50 nM) blocked DNA binding. When different do- gyrase-inhibiting antibiotics, but the second section of this re- mains of gyrase were examined for their ability to bind SD8 us- view covers the role of SD8 as an effector molecule controlling ing SPR and isothermal titration calorimetry (ITC), it was found the activity of a transcription factor called SimR, responsible for that interaction occurred only with the N-terminal domain of linking the biosynthesis and export of SD8 in the producing or- GyrA (Flatman et al. 2005), which was already known to con- ganism, S. antibioticus. In addition, the SD8 precursor, 7-oxo-SD8, tain the binding site for the G-segment DNA (Morais Cabral et al. has been thoroughly characterized as a substrate for the enzyme 1997). These biochemical and biophysical experiments therefore SimC7, which reduces a carbonyl to a hydroxyl group at the C-7 supported the idea that simocyclinones act by binding to the position in the angucyclinone moiety of the molecule. This enzy- GyrA subunit at a DNA-binding site to prevent the binding of matic step, which is critical because it converts an almost inac- the enzyme to DNA; a completely novel mode of action. This tive precursor into the mature antibiotic, is covered in the third idea was later corroborated by X-ray crystallography (Edwards and final section of this review. et al. 2009b, Hearnshaw et al. 2014), see below. Binding of SD8 to the N-terminal domain of GyrA was also seen using circu- lar dichroism experiments (Sissi et al. 2010); this method also SIMOCYCLINONES AS ANTIBIOTICS showed evidence for a second binding site in the C-terminal do- main of GyrB, albeit of lower affinity than the GyrA-binding site. Activity of simocyclinones against bacteria Subsequent ITC experiments (Hearnshaw et al. 2014) also found In general, the antibiotic activity of simocyclinones was found to evidence for a binding site in GyrB, but estimated that it was be relatively weak, except against some Gram-positive bacterial ∼1000-fold weaker than the GyrA site; it is unlikely that the GyrB species, for example Bacillus brevis (MIC 10 μg/ml) and Strepto- site contributes to the activity of simocyclinones. myces viridochromogenes (MIC 1 μg/ml) (Schimana et al. 2000). Lit- SD8 has also been found to inhibit E. coli topo IV and human tle activity was detected against Gram-negative bacteria. This is topo II, albeit with a lower potency than against gyrase (Flatman almost certainly due to the inability of simocyclinones to pene- et al. 2005;Sadiq et al. 2009). In other work, SD8 was found to also trate the outer membrane, since imp mutants of E. coli,which are inhibit S. aureus gyrase, but was much less effective against topo specifically compromised in outer membrane integrity, become IV from E. coli and S. aureus (Oppegard et al. 2009); S. aureus gyrase sensitive to SD8 (Edwards et al. 2009b), although multidrug efflux wasfoundtobe3–4-foldlesssensitivetoSD8than E. coli gyrase. pumps like AcrB may also contribute to resistance (Oppegard Elsewhere, it was found that the difference in SD8 potencies be- et al. 2009). However, it has been pointed out that most of these tween these enzymes was ∼20-fold (Alt et al. 2011); however, it susceptibility tests have been carried out using stock lab strains, should be stressed that the absolute IC values are likely to be and SD8 has shown more promising activity against some clini- affected by assay conditions, which differ between the two en- cal isolates of E. coli and Klebsiella pneumoniae (Richter et al. 2010). zymes. Taken together, it seems that gyrase is the preferred tar- More recently, the discovery of new simocyclinones (Bilyk et al. get for simocyclinones, and that they act by binding to the GyrA 2016) and the capacity for engineering novel compounds, as has subunit of gyrase preventing the binding of DNA. been carried out with the classical aminocoumarins (Heide et al. 2008; Heide 2009, 2014) and to a limited extent with simocycli- nones (Anderle et al. 2007a,b), has raised the possibility of com- How simocylinone D8 binds to gyrase pounds with increased antibacterial potency. However, given that SD8 has been shown to inhibit human topo II (Flatman et al. Biochemical and biophysical data (described above) strongly 2005;Sadiq et al. 2009), the potential for mammalian toxicity suggested that the simocyclinones bind to GyrA in a region of the must be borne in mind. protein involved in DNA binding. This proposal was confirmed by X-ray crystallography. Crystallization trials using simocycli- none D8 (SD8) and the N-terminal domain of the DNA gyrase How simocyclinones inhibit DNA gyrase A protein (GyrA59), whose structure was already known (Morais The similarity between the structures of simocyclinones (Fig. 1) Cabral et al. 1997), gave diffracting crystals (Edwards et al. 2009a). and those of the classical aminocoumarins led to the expec- This first structure (initially solved at 2.6- A resolution) revealed tation that simocyclinones would inhibit bacterial DNA gyrase a tetramer of GyrA59 that consisted of two GyrA59 dimers cross- by competitively binding to the ATPase active site in the GyrB linked by four molecules of SD8 (Edwards et al. 2009b). Two bind- subunit. It was shown that simocyclinone D8 (and D4) did in- ing pockets were observed for SD8 in each subunit, both lying deed inhibit DNA supercoiling catalyzed by E. coli gyrase but, within the DNA-binding ‘saddle’ (Morais Cabral et al. 1997)ofthe surprisingly, also inhibited DNA relaxation (Flatman et al. 2005), GyrA59 dimer, one accommodating the aminocoumarin moiety an ATP-independent reaction. Moreover, ATPase assays showed and the other accommodating the angucyclinone moiety. Selec- that SD8 and SD4 did not inhibit this reaction under conditions tion of spontaneous SD8-resistant E. coli mutants showed that where novobiocin was effective. The most common mode of ac- the mutations occurred in both pockets, corroborating the crys- tion of topoisomerase-targeted drugs (e.g. fluoroquinolones) is tal structure (Edwards et al. 2009b). Further site-directed mutants the stabilization of the enzyme-DNA cleavage complex. It was also supported the structure, while others could not be fully shown that SD8 did not act in this way but was found to antago- rationalized (see below), suggesting that this structure might nize the ability of fluoroquinolones, and other agents, to induce not reflect the situation in vivo. Although the crystal structure cleavage-complex formation (Flatman et al. 2005). showed a protein tetramer, it was suspected that this dimer– Taken together, these data suggested that simocyclinones dimer interaction was stabilized in the crystal and may not rep- might interfere with the binding of gyrase to DNA rather than to resent the physiologically relevant form of the complex. Downloaded from https://academic.oup.com/femsre/article/42/1/fux055/4604775 by DeepDyve user on 13 July 2022 Buttner et al. 103 Figure 3. (a) Structure of E. coli DNA GyrA homodimer (55 kDa N-terminal fragment) with two molecules of SD8 bound; one subunit is colored blue and the other in yellow; the SD8 molecules are shown in two shades of green (PDB accession number 4CKL). A DNA duplex taken from a superposed structure of a Staphylococcus aureus gyrase-DNA-drug complex (PDB accession number: 2XCS) is also shown in pink to illustrate that SD8 would block the interaction of G-segment DNA with the DNA-binding ‘saddle’. (b) Top view of panel a, looking down the dimer 2-fold axis. (c) Enlarged view of the boxed region shown in panel b, with the SD8 ligands in stick representation with atom coloration (carbon, green; oxygen, red; nitrogen, blue, chlorine, gray). This clearly shows that the antibiotic spans the dimer interface with distinct binding pockets for the terminal angucyclinone (ANG) and aminocoumarin (AC) groups. (This figure and the other structural figures were p repared using CC4MG; McNicholas et al. 2011.) Figure 4. Schematic representation of the SD8-binding pocket of GyrA showing all residues within 4 A of the ligand, as revealed in the crystal structure of the GyrA-SD8 complex (PDB accession number 4CKL). One subunit is shown in blue and the other in yellow. Hydrogen bonds are indicated by dotted lines; van der Waal contacts are indicated by orange arcs, and water molecules are shown as filled blue circles labeled ‘W’. For clarity, all hydrogens have been omitted. Analysis of the SD8-GyrA59 complex using nanoelectro- discrete protein dimer with two SD8 molecules bound (Fig. 3; spray ionization mass spectrometry showed that the tetrameric Hearnshaw et al. 2014). This structure, solved at 2.05-Areso- species observed in the crystal could be reproduced in solution, lution, proved to be entirely consistent with all the mutations but only at high SD8 concentrations, while at lower concentra- to SD8 resistance that had been previously made or selected tions, a dimeric species was present with two SD8 molecules (Edwards et al. 2009b); additional mutants made in response to bound per dimer; this result was potentially at odds with the the revised structure were also shown to be consistent (Hearn- previous structural data (Edwards et al. 2009b). Further mass shaw et al. 2014). In addition to the new structure being dimeric, spectrometry suggested that the binding of SD8 to the pro- rather than tetrameric, the conformation of SD8 is significantly tein dimer showed strong allosteric cooperativity (Edwards et al. different, compared with the earlier structure: the orientation of 2011). A subsequent crystal structure of a shorter version of the the aminocoumarin within the aminocoumarin pocket is some- N-terminal domain of GyrA (GyrA55), which lacks residues that what different, while the angucyclinone ‘pocket’ has shifted stabilize dimer–dimer interactions in the tetramer, revealed a such that it now spans the interface between the two monomers Downloaded from https://academic.oup.com/femsre/article/42/1/fux055/4604775 by DeepDyve user on 13 July 2022 104 FEMS Microbiology Reviews, 2018, Vol. 42, No. 1 (Fig. 4) and thus could provide a structural explanation for the (SC4; Fig. 1), could dissociate SimR from its operators (Le et al. cooperative binding observed by mass spectrometry. This new 2009). Subsequently, crystal structures of SimR alone (apo; 1.95- position for the angucyclinone group suggests that the binding A resolution) (Le et al. 2011b), in complex with its operator DNA of SD8 effectively ‘staples’ the GyrA dimer closed so inhibiting (2.99-A resolution) (Le et al. 2011a), and in complex with either the conformational changes that need to occur upon DNA bind- SD8 or SC4 (both 2.3-A resolution) (Le et al. 2011b), showed how ing and cleavage (Hearnshaw et al. 2014). SimR binds its effector ligand and how ligand binding prevents The SD8-binding site on gyrase is close to the binding site of SimR from binding to its operator DNA. Unsurprisingly, there is the fluoroquinolone antibiotics (Laponogov et al. 2009;Bax et al. no similarity between the ligand-binding pockets in gyrase and 2010; Laponogov et al. 2010), raising the possibility of generat- SimR. ing hybrid compounds. To this end, a series of ciprofloxacin- aminocoumarin hybrids has been synthesized, designed to bind How SimR binds SD8 to the aminocoumarin pocket of SD8 and to the fluoroquinolone TFRs function as homodimers, with each subunit having two do- pocket (Austin et al. 2016); some of these compounds retain good inhibitory activity against gyrase. It remains to be seen whether mains, an N-terminal DNA-binding domain (DBD) containing a helix-turn-helix (HTH) motif, and a C-terminal ligand-binding such compounds can be developed as viable antibiotics. Also flavone-base analogs of simocyclinones have been made in or- domain (LBD) (Ramos et al. 2005;Yu et al. 2010; Cuthbertson and Nodwell 2013). The ligand-binding pocket of SimR is unusual; in der to bind to a hydrophobic cleft in the protein and further stabilize binding (Verghese et al. 2013). Although some of these other characterized TFRs, one ligand-binding pocket is typically contained within each subunit and so, for example, in the closely compounds are effective gyrase inhibitors, they also stabilize the gyrase-DNA cleavage complex and probably act via a mecha- related protein, ActR, there is only one ligand contact with the second subunit (Willems et al. 2008), while in TetR itself there nism involving DNA intercalation, i.e. they do not bind at the intended site. is none (Orth et al. 2000). In contrast, the ligand-binding pocket in SimR spans the two protein subunits, with the angucyclinone Taken together, we conclude that simocyclinones bind to the A subunit of DNA gyrase, in a region that is normally oc- of SD8 bound in one subunit, while the olivose sugar, tetraene cupied by the G-segment DNA (Figs 2–4) and prevent the ini- and aminocoumarin parts of the molecule are bound in the other (Le et al. 2011b)(Figs 5 and 6). This split binding pocket is tial interaction of DNA with the enzyme and thus all the en- suing catalytic events. This is quite distinct from the mode of ∼30 A in length, with SD8 bound in an extended conformation. Although SD8 has 19 atoms that could potentially participate in action of fluoroquinolones (cleavage-complex stabilization) and aminocoumarins (competitive inhibitors of ATP binding) and hydrogen bonding, there are only five direct hydrogen bonds be- tween SimR and SD8, three with the aminocoumarin and two raises the possibility of developing other agents that use this mode of action, which would be less likely to be cross-resistant with the angucyclinone (Fig. 6). However, the dearth of hydrogen bonding is compensated for by extensive van der Waals contacts to known antibiotics. with the protein along the length of the ligand (Fig. 6). The way cognate ligands are bound by TFRs is highly variable. For exam- ple, when the SimR-SD8 structure is compared with that of the SIMOCYCLINONE D8 AS A TRANSCRIPTION complex between the closely related TFR protein ActR and its FACTOR EFFECTOR MOLECULE cognate ligand, the antibiotic actinorhodin, the long axis of the actinorhodin molecule lies almost perpendicular to that of SD8 SD8 has been studied most intensively as a gyrase-inhibiting in the SimR–SD8 structure (Willems et al. 2008;Le et al. 2011b). antibiotic. However, it has also been characterized as an effec- tor molecule controlling the activity of a transcription factor called SimR, responsible for linking the biosynthesis and ex- How simocyclinone D8 prevents SimR from binding port of SD8 in the producing organism, S. antibioticus (Le et al. DNA 2009, 2011a,b). It is perhaps under-appreciated that antibiotics are often potentially toxic to the organisms that produce them Available evidence suggests that apo-TFRs sample a range of (Cundliffe 1989; Hopwood 2007). Therefore, producing organ- conformations in solution and that ligand binding simply cap- isms must have mechanisms to ensure that the antibiotic ex- tures one of these conformations, rather than inducing the con- port machinery is in place when antibiotic biosynthesis be- formational change (Reichheld, Yu and Davidson 2009;Yu et al. gins. The relevant mechanism in the simocyclinone producer 2010; Cuthbertson and Nodwell 2013). SimR-apo did not crys- is specified by two adjacent genes, simR and simX, which sit tallize in its DNA-binding form (apparent from the distance be- within the simocyclinone (sim) biosynthetic gene cluster (Galm tween its recognition helices), and indeed this is generally true et al. 2002; Trefzer et al. 2002;Le et al. 2009). The SimR and of TFR apo-proteins (Yu et al. 2010). However, comparison of the SimX proteins resemble the TetR/TetA repressor/efflux pump SimR-apo, SimR-SD8 and SimR-DNA structures provided clear pair found in a number of human pathogens, which confer resis- insight into the likely mechanism of ligand-mediated derepres- tance to clinically important tetracyclines (Chopra and Roberts sion. 2001). SimX is an efflux pump, a member of the major facilitator The ligand-binding sites of TFRs are remote from their superfamily, which exports simocyclinone from the producing DBDs and derepression generally involves allosteric mecha- organism. simX transcription is repressed by SimR, a TetR-family nisms (Ramos et al. 2005;Yu et al. 2010; Cuthbertson and Nod- transcriptional regulator (TFR) that binds to two separate opera- well 2013). Ligand-bound and DNA-bound structures have been tors in the intergenic region between the divergently transcribed determined for several TFRs, including QacR, DesT, CgmR and simR and simX genes (Le et al. 2009). Simocyclinone abolishes TetR itself, and in these cases conformational changes appear DNA-binding by SimR, thereby derepressing transcription of the to be transmitted largely within the same subunit (Orth, Saenger simX efflux pump gene, and this provides the mechanism that and Hinrichs 1999; Orth et al. 2000; Schumacher et al. 2001; 2002; couples the biosynthesis of simocyclinone to its export. It was Itou et al. 2010; Miller et al. 2010). Specifically, they suggest that also shown that the biosynthetic intermediate simocyclinone C4 ligand binding traps a conformational state in which the DBD (in Downloaded from https://academic.oup.com/femsre/article/42/1/fux055/4604775 by DeepDyve user on 13 July 2022 Buttner et al. 105 Figure 5. Comparison of the SimR-SD8 (a,b,c) and SimR-DNA (d,e,f) structures with one SimR subunit shown in blue and the other shown in yellow. The two recognition helices are highlighted in magenta and bound SD8 molecules are shown in green. Note that the ligand-binding pocket in SimR spans the two protein subunits, with the angucyclinone (ANG) end of SD8 bound in one subunit while the aminocoumarin (AC) end is bound in the other such that SD8 skewers the two subunits. Note also that in the apo form of SimR (structure not shown), Arg122 is buried in its cognate subunit; however, in the SimR-SD8 complex, each copy of this residue (shown as red sticks) projects across the dimer interface into a pocket in the surface of the opposing subunit. Arg122 is not ordered in the SimR-DNA structure. (PDB accession numbers: SimR-SD8: 2Y30; SimR-DNA: 3ZQL; SimR-apo: 2Y2Z). particular the HTH motif) is repositioned relative to the LBD such DNA-bound conformation and the derepressed ligand-bound that the two recognition helices in the homodimer are too far structure, and all five of these link the wrapping arm with the apart to bind appropriately in consecutive major grooves of the LBD of the other subunit. As a consequence, when the subunits DNA. In contrast, comparison of the repressive SimR-DNA struc- rotate in the ligand-bound form, the wrapping arm moves with ture with the derepressed SimR-SD8 structure shows that the them. Because the ligand-binding pocket passes through both relative dispositions of the LBDs and DBDs within each indi- subunits, SD8 effectively skewers the dimer, rigidifying the com- vidual SimR subunits remain essentially unchanged on ligand plex, and because it is a relatively hydrophobic molecule, SD8 binding. Instead, SD8 binding captures a conformation in which contributes to the hydrophobic core of the SimR dimer, stabi- there is a rigid-body rotation of one SimR subunit relative to the lizing the overall structure. In addition, in the apo and DNA- other, and this rigid-body rotation moves the recognition he- bound structures, the two SimR subunits present essentially lices ∼5 A further apart in the derepressed (SD8-bound) state, flat surfaces to one another, allowing them to rotate relative preventing DNA binding (Fig. 7). It may well be significant that to each other. In contrast, in the SD8-bound form, the side- the ligand-binding sites in the previously characterized TFRs are chain of Arg122 from each subunit projects across the dimer contained almost entirely within individual subunits, whereas interface into a pocket in the surface of the opposing sub- the ligand-binding pocket in SimR spans the two subunits. unit, potentially acting as locating pins to lock the subunits Two helices of the SimR LBD (α9-α10) form a wrapping arm together (Fig. 5). that folds around the LBD of the opposing subunit (Figs 5 and The biosynthetic intermediate simocylinone C4 (SC4) lacks 7). These two helices form the end of the ligand-binding pocket the aminocoumarin ring present in the mature antibiotic (Fig. 1) responsible for binding the angucyclinone of SD8 (Figs 5, 6 and is essentially inactive as a DNA gyrase inhibitor; the SD8 IC and 7), and the wrapping arm changes conformation in the is 0.1 μM, whereas the SC4 IC is >100 μM(Edwards et al. 2009b). ligand-bound state. Only five reciprocal inter-subunit hydrogen However, despite the absence of the aminocoumarin ring, SC4 bonds (i.e. 10 in total) are maintained between the repressive binds SimR and prevents it from binding DNA (Le et al. 2009). The Downloaded from https://academic.oup.com/femsre/article/42/1/fux055/4604775 by DeepDyve user on 13 July 2022 106 FEMS Microbiology Reviews, 2018, Vol. 42, No. 1 Figure 6. Schematic representation of the SD8-binding pocket of SimR showing all residues within 4 A of the ligand. One subunit is shown in blue and the other shown in yellow. Hydrogen bonds are indicated by dotted lines and van der Waal contacts are indicated by orange arcs. The two water molecules that link Gln136 to the olivose sugar are shown as filled blue circles labeled ‘W’. For clarity, all hydrogens have been omitted. structure of the SimR-SC4 complex has also been determined (Le Bilyk et al. (2016) sequenced the Kitasatospora sp. and Streptomyces et al. 2011b). Comparison of the SD8-SimR and SC4-SimR struc- sp. NRRL B-24484 biosynthetic clusters, there were no type I PKS tures shows that the two molecules bind SimR in the same way, genes present, and the tetraene was instead shown to be synthe- meaning the parts common to both molecules (the angucycli- sized by an iterative type II PKS. This type II PKS is also present none, tetraene and olivose sugar) occupy equivalent positions in in S. antibioticus,leaving theroleofthe typeIPKS unknown. To the binding pocket. SC4 is slightly less effective than SD8 at dere- date, only two biosynthetic enzymes have been characterized pressing SimR in vitro (Le et al. 2009) and this is probably a conse- biochemically: SimL and SimC7. SimL catalyses the presumed quence of the fewer favorable interactions that SC4 makes with last step in the pathway, acting as an amide bond-forming lig- the protein, due to the absence of the aminocoumarin. These ase that attaches the aminocoumarin to the tetraene linker (Luft results show that a pathway intermediate that is not an active et al. 2005;Pacholec et al. 2005;Anderle et al. 2007b). antibiotic can induce expression of the efflux pump, and simi- As noted above, the second enzyme, SimC7, was originally lar observations have been made in other antibiotic pathways, proposed to be involved in the biosynthesis of the tetraene particularly for actinorhodin (Otten, Ferguson and Hutchinson linker. It was subsequently shown to be an NAD(P)H-dependent 1995; Jiang and Hutchinson 2006;Ahn et al. 2007;Tahlan et al. ketoreductase that catalyzes the reduction of a carbonyl to a 2007;Ostash et al. 2008;Tahlan et al. 2008; Willems et al. 2008). hydroxyl group at the C-7 position of the angucyclinone, high- These data raise the possibility of a ‘feed-forward’ mechanism, lighting the dangers of relying on speculative gene annotations in which inactive intermediates ensure expression of the efflux (Fig. 8;Schafer ¨ et al. 2015). This enzymatic step is essential pump prior to the build-up of a toxic concentration of the poten- for antibiotic activity, converting the almost inactive 7-oxo-SD8 tially lethal mature antibiotic (Hopwood 2007;Tahlan et al. 2007; (IC ∼ 50–100 μM) into the potent gyrase inhibitor SD8 (IC ∼ 50 50 Le et al. 2009). 0.1–0.6 μM) (Schafer ¨ et al. 2015). Based on the intermediates pro- duced by S. antibioticus, it seems the biosynthesis of SD8 starts with assembly of the angucyclinone, followed by the attach- 7-OXO-SIMOCYCLINONE D8 AS A SUBSTRATE ment of the olivose sugar, and then the tetraene linker, and lastly the aminocoumarin (i.e. as drawn in Figs 1 and 8,SD8 is While the functions of most of the biosynthetic enzymes en- coded within the S. antibioticus sim cluster have been predicted assembled from right to left) (Schimana et al. 2001). Therefore, the natural substrate of SimC7 is probably a 7-oxo angucycli- (Galm et al. 2002; Trefzer et al. 2002), the biosynthetic pathway remains largely uncharacterized experimentally. This lack of none intermediate lacking the attached olivose sugar, tetraene linker and aminocoumarin, an intermediate that is detectable knowledge about the biosynthesis of simocyclinones is well il- lustrated by the tetraene moiety. Trefzer et al. (2002) proposed only in simC7 mutants (Schafer ¨ et al. 2015). Nevertheless, the enzyme readily accepts as a substrate the full-length interme- that the tetraene linker would be the product of the large modu- lar type I polyketide synthase (PKS), SimC1ABC, working in trans diate 7-oxo-SD8, the product made by simC7 mutants (Schafer et al. 2015). with two monofunctional enzymes, SimC6 and SimC7. Yet when Downloaded from https://academic.oup.com/femsre/article/42/1/fux055/4604775 by DeepDyve user on 13 July 2022 Buttner et al. 107 Figure 7. Structures of (a) SimR-SD8 and (b) SimR-DNA together with schematic representations illustrating the rigid-body rotation of the subunits relative to one another. To emphasize the subunit rotation, the position of the blue subunit is fixed in each panel so that the rotation of the yellow subunit accounts fo r all the movement in the dimer. The asterisk indicates the pivot point around which rotation occurs. Note that the net effect of subunit rotation is that the distance separating the two recognition helices increases to 41.7 A in the SD8-bound form, a distance incompatible with DNA binding. Note also that helices α9–α10 form a wrapping arm that engages the LBD of the opposing subunit and that these helices additionally form the angucyclinone end of the ligand-binding pocket. (PDB accession numbers: SimR-SD8: 2Y30; SimR-DNA: 3ZQL). SimC7 is a member of the short-chain dehydroge- that is characteristic of the so-called extended SDR subfamily nase/reductase (SDR) superfamily. These proteins have diverse (Kavanagh et al. 2008). This latter domain contains a ‘lid’ motif biochemical activities, including functioning as dehydratases, consisting of two anti-parallel α-helices that sits over the active reductases, epimerases, dehydrogenases and decarboxylases site. The apo, binary and ternary SimC7 structures are very sim- (Kallberg, Oppermann and Persson 2010; Persson and Kallberg ilar except for the orientation of this lid, which closes somewhat 2013). Classical SDR enzymes have a characteristic Ser-Tyr-Lys over the bound substrate (maximum Cα-Cα shift 5.35 A). The catalytic triad in their active site, in which the latter two underside of the lid forms part of the tight, highly hydrophobic residues form a YxxxK motif. The conserved tyrosine acts as a substrate binding pocket (Fig. 9) that provides the environment central acid-base catalyst that donates a proton to the substrate. needed for catalysis (Schafer ¨ et al. 2016). The adjacent lysine serves to lower the pKa of the tyrosine hydroxyl group and often contributes directly to a proton relay How SimC7 binds 7-oxo-SD8 mechanism. Lastly, the hydroxyl group of the serine polarizes the carbonyl group of the substrate (Kavanagh et al. 2008). In the SimC7 ternary complex with substrate and NADP bound, The catalytic mechanism of SimC7 was investigated because the angucyclinone ring system of 7-oxo-SD8 binds adjacent and it shares little sequence similarity with other characterized parallel to the nicotinamide ring of the cofactor (Fig. 9c), where ketoreductases, even with functionally analogous polyketide it adopts an essentially planar conformation. This differs from ketoreductases involved in the biosynthesis of related angucy- the conformations seen in the SimR-SD8 and gyrase-SD8 com- clinone antibiotics. Most of all, alignments of SimC7 with other plexes, where the A-ring of the angucyclinone in SD8 is tilted SDR proteins suggested that SimC7 lacked the classical catalytic upwards towards the epoxide (Le et al. 2011b, Hearnshaw et al. triad, including the tyrosine that acts as the central acid-base 2014;Schafer ¨ et al. 2016). The substrate pocket has several dis- catalyst in classical SDR proteins. This possibility was investi- tinctive characteristics (Fig. 9). The pocket is very hydrophobic gated by determining the structures of SimC7 alone (apo; 1.6-A and highly constricted in shape, features that are likely to en- resolution), the binary complex with NADP (1.95-A resolution) force the planar conformation on the angucyclinone ring sys- and the ternary complex with both NADP and 7-oxo-SD8 (1.2-A tem. Strikingly, within the hydrophobic pocket, 7-oxo-SD8 is resolution) (Schafer ¨ et al. 2016). As might be expected, there is bound by just one direct hydrogen bond, connecting the side- no similarity between the simocyclinone-binding pockets in chain of Ser95 and the C-7 carbonyl oxygen of the angucycli- gyrase, SimR and SimC7. ¨ none (Fig. 10;Schafer et al. 2016). However, even this single SimC7 has two domains (Fig. 9), a larger Rossmann-fold do- hydrogen bond is not required for enzymatic activity, since main that binds NADP and a smaller substrate-binding domain a constructed S95A variant shows almost wild-type levels of Downloaded from https://academic.oup.com/femsre/article/42/1/fux055/4604775 by DeepDyve user on 13 July 2022 108 FEMS Microbiology Reviews, 2018, Vol. 42, No. 1 Figure 8. SimC7 catalyzes the reduction of 7-oxo SD8 to simocyclinone D8. Labels A-D denote the four rings of the angucyclinone; the C-7 carbonyl/hydroxyl is highlighted in red. substrate conversion (Schafer ¨ et al. 2016). Thus, although this quinone-like C-ring and the phenyl-like D-ring of the angucy- hydrogen bond may help to position the C-7 carbonyl above the clinone promote the formation of an intramolecular hydrogen C-4 position of the nicotinamide ring ready for direct hydride bond between the proton on the C-8 hydroxyl and the oxygen of transfer, and provide additional polarization of the C-7 carbonyl the neighboring C-7 carbonyl (Fig. 11b). This intramolecular hy- group, as proposed for the structurally equivalent Ser or Thr drogen bond polarizes the carbonyl, enhancing the electrophilic- residues in classical SDR proteins (Kavanagh et al. 2008; Kallberg, ity of C-7 and making it a good acceptor for hydride attack from Oppermann and Persson 2010; Persson and Kallberg 2013), nei- the 4-pro-S position of the nicotinamide ring, which is only 3.0 ther proposed effect is crucial for catalysis. As discussed above, A away. Then, internal proton transfer from the neighboring the natural substrate for SimC7 is probably a 7-oxo angucycli- C-8 hydroxyl group forms the C-7 hydroxyl group, generating a none intermediate lacking the olivose sugar, tetraene linker and phenolate intermediate where the aromatic D-ring stabilizes the aminocoumarin. Consistent with this suggestion, only the an- negative charge on the C-8 oxygen. In the second step of the re- gucyclinone is buried in the active site of SimC7, with the rest of action, the phenolate intermediate leaves the substrate-binding the molecule projecting out of the enzyme (Fig. 9). Indeed, the pocket and the C-8 hydroxyl group re-forms by abstracting a pro- aminocoumarin and roughly half of the tetraene linker are not ton from bulk water (Fig. 11b), something that cannot happen resolved in electron density. within the confines of the active site. The hydrophobic active site cavity would accelerate expulsion of the charged phenolate intermediate created during catalysis. Lastly, the direct hydride How SimC7 converts 7-oxo-SD8 into SD8 attack from below the angucyclinone explains why simocycli- nones have 7S-stereochemistry. In summary, the SimC7 mecha- The structures confirmed the prediction made from sequence nism involves the intramolecular transfer of a substrate-derived alignments that SimC7 lacks a canonical SDR Ser-Tyr-Lys cat- proton to generate a phenolate intermediate, and this obviates alytic triad (Schafer ¨ et al. 2016). While the serine is conserved the need for proton transfer from a canonical SDR active-site (Ser95), the other two residues (the YxxxK motif), including the tyrosine. key tyrosine residue that acts as the acid/base catalyst in clas- sical SDR proteins, are replaced by Ile108 and His112, respec- tively (Fig. 11). The structures also demonstrate that there is Why is 7-oxo-SD8 almost inactive as a DNA gyrase no alternative residue that could act as an acid/base catalyst, inhibitor? and instead suggest that SimC7 has a novel reaction mecha- nism (Schafer ¨ et al. 2016). This unusual mechanism does not It is striking that SD8 is very potent as a gyrase inhibitor depend on catalytic residues in the protein, but instead ex- (IC ∼ 0.1–0.6 μM) and yet 7-oxo-SD8 is almost inactive (IC ∼ 50 50 ploits the chemical characteristics of 7-oxo-SD8 itself, and is 50–100 μM) (Schafer ¨ et al. 2015). Why does such a small struc- thus a new example of substrate-assisted catalysis (Dall’Acqua tural difference, the presence of a carbonyl group at the C-7 po- and Carter 2000). In the first step, the hydrophobic environment sition in 7-oxo-SD8 (Fig. 8), have such a drastic effect on the of the substrate-binding pocket and the juxtaposition of the antibiotic activity of the molecule? The likely answer becomes Downloaded from https://academic.oup.com/femsre/article/42/1/fux055/4604775 by DeepDyve user on 13 July 2022 Buttner et al. 109 Figure 9. (a) and (b) Crystal structure of the SimC7 ternary complex with NADP and 7-oxo-SD8. The nucleotide-binding domain, the substrate-binding domain and the lid motif are shown in yellow, blue and magenta, respectively. NADP is shown in pink and 7-oxo-SD8 is shown in green. For the latter, only the crystallographically resolved atoms are shown, i.e. the angucyclinone, the olivose and roughly half of the tetraene linker. (c) Close-up showing the active site of the ternary complex including the Ser95-Ile108-His112 ‘catalytic triad’ residues, and Asn137, which is important in maintaining the syn-conformation of the cofactor. C-4 of the cofactor nicotinamide ring and C-7 of the substrate are highlighted by black spheres, which are 3 A apart, indicating that the substrate is exactly positioned for direct hydride transfer. (d) Cross-section through the active site pocket, showing how tightly the cofactor (pink) and substrate (green) are bound. For clarity, only the nicotinamide ribosyl moiety of the cofactor is shown in panel d, and only the angucyclinone moiety of the substrate is shown in panels c and d (PDB accession number: 5L4L). clear from analysis of the structure of the GyrA-SD8 complex: ring system, which may well affect other bonding interactions both the C-7 and C-8 hydroxyls are involved in a hydrogen bond- with GyrA. ing network that helps secure the angucyclinone in its binding pocket (Fig. 4). However, in 7-oxo-SD8, an intramolecular hy- CONCLUDING REMARKS drogen bond between the C-7 carbonyl and the C-8 hydroxyl is preferred over these intermolecular interactions and this si- In the three different systems we have described in this review, multaneously breaks the direct contact between the angucycli- the interaction of the ligand with the protein has entirely differ- none and His80 and the indirect contacts with Pro79 and Arg121 ent downstream consequences. For gyrase, it results in inhibi- (Fig. 4). His80, in particular, is known to play a crucial role in tion, leading to cell death, for SimR, it results in derepression, binding simocyclinone, since mutating this residue to alanine leading to antibiotic export, and for SimC7, it results in cataly- causes a 230-fold increase in the IC of SD8 for gyrase (Edwards 50 sis, leading to potentiation of an antibiotic. Given that SimC7 is et al. 2009b). In addition, the presence of a carbonyl group at an enzyme, the interaction with the ligand is transient, whereas C-7 would alter the overall conformation of the angucyclinone the interaction with gyrase and SimR will be much longer-lived. Downloaded from https://academic.oup.com/femsre/article/42/1/fux055/4604775 by DeepDyve user on 13 July 2022 110 FEMS Microbiology Reviews, 2018, Vol. 42, No. 1 The extensive nature of these double-headed interactions leads to very tight binding, commensurate with the physiological con- sequences. Indeed, molecules lacking either the angucyclinone or the aminocoumarin bind much more weakly to DNA gyrase and, as a consequence, the potency of SD8 as an antibiotic is severely compromised through loss of either ‘warhead’ (Edwards et al. 2009b). The proportion of hydrogen bonds is highest for the complex with gyrase because the binding site is largely solvent- exposed and would otherwise interact with the G-segment DNA, which is polar. In SimR, the ligand-binding site threads through the hydrophobic core of the homodimer, and so the interac- tions are dominated by van der Waals contacts. In contrast, in the 7-oxo-SD8 complex with SimC7, only the angucyclinone in- teracts with the enzyme, this being consistent with the site of ketoreduction and the expectation that the natural substrate in vivo is the angucyclinone alone. Given the necessity to pre- cisely position the SimC7 substrate for catalysis, the dearth of hydrogen bonds seems counterintuitive. Indeed, a Ser95 to Ala substitution that removes the only hydrogen bond shows that even this is dispensable. However, the necessity to provide a hy- drophobic environment for efficient catalysis would be consis- Figure 10. Schematic representation of the hydrophobic substrate-binding tent with a paucity of hydrogen bonding partners and bound pocket of SimC7 showing all residues within 4 A of the ligand, as revealed in the water molecules. Instead, the highly constrained nature of the crystal structure of the SimC7 ternary complex with NADP and 7-oxo-SD8 (PDB accession number 5L4L). Hydrogen bonds are indicated by dotted lines and van SimC7 active site is a key factor in sterically guiding the sub- der Waal contacts are indicated by orange arcs. Note that the substrate is bound strate to its catalytically competent position adjacent to the by only one direct hydrogen bond, connecting the C-7 carbonyl of the angucy- cofactor with hydride donor and hydride acceptor atoms jux- clinone and the side-chain hydroxyl of Ser95. This hydrogen bond may assist in taposed. The transient nature of this interaction would be pro- positioning the substrate and facilitate the reaction. However, this interaction is moted by the negative charge that develops on the phenolate not required for enzymatic activity, since a constructed S95A variant of SimC7 intermediate, which would be unfavorable in the hydrophobic shows near wild-type enzymatic activity (Schafer ¨ et al. 2016). Note that one face of the pocket is formed by the NADP cofactor itself. In the natural SimC7 sub- active site, and possibly also by the increased puckering of the strate, R = H; in the substrate used here, R includes the deoxysugar, tetraene angucyclinone ring system that would occur when the C7 keto linker and the aminocoumarin. For clarity, all hydrogens have been omitted. group is reduced to a hydroxyl. Finally, although SD8 itself is not viable as a clinical antibi- In both gyrase and SimR, there are a substantial number of otic, due at least in part to its poor penetration into bacteria, the way in which it inhibits DNA gyrase is unique. It therefore has interactions with the terminal aminocoumarin and angucycli- none groups, which are bound by separate subunits; addition- the potential to guide the development of new, clinically rele- vant compounds acting against this enzyme, and the detailed ally, there are a handful of contacts involving the linker region. Figure 11. Comparison of the canonical SDR ketoreduction mechanism with the novel SimC7 reaction mechanism. (a) In classical SDR proteins, the conserved active site tyrosine serves as a central acid-base catalyst that donates a proton to the substrate. The adjacent lysine residue lowers the pKa of the tyrosine hydroxyl group and often contributes directly to the proton relay mechanism; the hydroxyl group of the serine stabilizes and polarizes the carbonyl group of the substrate. (b) SimC7 has an atypical catalytic triad consisting of Ser95, Ile108 and His112. In the first step of the SimC7 mechanism, the C-7 carbonyl group of the angucyclinon e is reduced by transfer of the 4-pro-S hydride of the cofactor onto the C-7 carbon of the substrate. This transfer from below the C-ring results in the characteristic 7-S-stereochemistry of SD8. Ketoreduction at position C-7 is completed by intramolecular proton transfer from the neighboring C-8 hydroxyl group of the angucyclinone; the resultant negative charge on C-8 is stabilized by the adjacent aromatic ring system. In the second step, the C-8 phenolate intermediate regains a proton from bulk water after leaving the substrate binding pocket. In the natural SimC7 substrate, R = H; in the substrate used here, R includes the deoxysugar, tetraene linker and the aminocoumarin. Note that there are no water molecules in the active site pocket that could contribute to the reaction mechanism. In the ternary complex, the nearest ˚ ˚ water to O-7 of the angucyclinone is ∼5.5 A away, and the nearest water to O-8 is ∼4.9 A away. Because of steric constraints within the pocket, neither could approach the substrate oxygen atoms without either a repositioning of the substrate or a conformational change in the protein. Downloaded from https://academic.oup.com/femsre/article/42/1/fux055/4604775 by DeepDyve user on 13 July 2022 Buttner et al. 111 structural information available should potentiate such devel- Dall’Acqua W, Carter P. Substrate-assisted catalysis: molecular opment. basis and biological significance. Protein Sci 2000;9:1–9. Edwards MJ, Flatman RH, Mitchenall LA et al. Crystallization and preliminary X-ray analysis of a complex formed between the ACKNOWLEDGEMENTS antibiotic simocyclinone D8 and the DNA breakage-reunion domain of Escherichia coli DNA gyrase. Acta Crystallogr Sect F We thank Hans-Peter Fiedler for providing simocyclinone D8. Struct Biol Cryst Commun 2009a;F65:846–8. Edwards MJ, Flatman RH, Mitchenall LA et al. A crystal struc- FUNDING ture of the bifunctional antibiotic simocyclinone D8, bound to DNA gyrase. Science 2009b;326:1415–8. This work was funded by a BBSRC studentship to M.S., Edwards MJ, Williams MA, Maxwell A et al. Mass spectrom- BBSRC grant BB/I002197/1 to M.J.B. and D.M.L., BBSRC grant etry reveals that the antibiotic simocyclinone D8 binds BB/I002049/1 to A.M. and D.M.L. and by BBSRC Institute Strategic to DNA gyrase in a “bent-over” conformation: evidence Programme Grant BB/J004561/1 to the John Innes Centre. of positive cooperativity in binding. Biochemistry 2011;50: Conflict of interest. None declared. 3432–40. Flatman RH, Howells AJ, Heide L et al. Simocyclinone D8, an in- hibitor of DNA gyrase with a novel mode of action. Antimicrob REFERENCES Agents Chemother 2005;49:1093–100. Ahn SK, Tahlan K, Yu Z et al. Investigation of transcription re- Forterre P, Gribaldo S, Gadelle D et al. Origin and evolution of pression and small-molecule responsiveness by TetR-like DNA topoisomerases. Biochimie 2007;89:427–46. transcription factors using a heterologous Escherichia coli- Galm U, Schimana J, Fiedler HP et al. Cloning and analysis of the based assay. J Bacteriol 2007;189:6655–64. simocyclinone biosynthetic gene cluster of Streptomyces an- Alt S, Mitchenall LA, Maxwell A et al. Inhibition of DNA gy- tibioticus Tu 6040. Arch Microbiol 2002;178:102–14. rase and DNA topoisomerase IV of Staphylococcus aureus and Harms A, Stanger FV, Scheu PD et al. Adenylylation of gyrase and Escherichia coli by aminocoumarin antibiotics. J Antimicrob Topo IV by FicT toxins disrupts bacterial DNA topology. Cell Chemother 2011;66:2061–9. Rep 2015;12:1497–507. Anderle C, Hennig S, Kammerer B et al. Improved mutasynthetic Hearnshaw SJ, Edwards MJ, Stevenson CE et al. Anew crys- approaches for the production of modified aminocoumarin tal structure of the bifunctional antibiotic simocyclinone D8 antibiotics. Chem Biol 2007b;14:955–67. bound to DNA gyrase gives fresh insight into the mechanism Anderle C, Li SM, Kammerer B et al. New aminocoumarin antibi- of inhibition. J Mol Biol 2014;426:2023–33. otics derived from 4-hydroxycinnamic acid are formed after Heide L. Genetic engineering of antibiotic biosynthesis for heterologous expression of a modified clorobiocin biosyn- the generation of new aminocoumarins. Biotechnol Adv thetic gene cluster. J Antibiot (Tokyo) 2007a;60:504–10. 2009;27:1006–14. Austin MJ, Hearnshaw SJ, Mitchenall LA et al. A natural product Heide L. New aminocoumarin antibiotics as gyrase inhibitors. Int inspired fragment-based approach towards the development J Med Microbiol 2014;304:31–36. of novel anti-bacterial agents. MedChemComm 2016;7:1387– Heide L, Gust B, Anderle C et al. Combinatorial biosynthesis, metabolic engineering and mutasynthesis for the genera- Bax BD, Chan PF, Eggleston DS et al. Type IIA topoisomerase tion of new aminocoumarin antibiotics. Curr Top Med Chem inhibition by a new class of antibacterial agents. Nature 2008;8:667–79. 2010;466:935–40. Holzenkampfer M, Walker M, Zeeck A et al. Simocyclinones, Bilyk O, Brotz E, Tokovenko B et al. New Simocyclinones: sur- novel cytostatic angucyclinone antibiotics produced by Strep- prising evolutionary and biosynthetic insights. ACS Chem Biol tomyces antibioticus Tu 6040 II. Structure elucidation and 2016;11:241–50. biosynthesis. J Antibiot (Tokyo) 2002;55:301–7. Boucher HW, Talbot GH, Bradley JS et al. Bad bugs, no drugs: no Hopwood DA. How do antibiotic-producing bacteria ensure their ESKAPE! An update from the Infectious Diseases Society of self-resistance before antibiotic biosynthesis incapacitates America. Clin Infect Dis 2009;48:1–12. them? Mol Microbiol 2007;63:937–40. Bush K, Courvalin P, Dantas G et al. Tackling antibiotic resistance. Itou H, Watanabe N, Yao M et al. Crystal structures of the mul- Nat Rev Microbiol 2011;9:894–6. tidrug binding repressor Corynebacterium glutamicum CgmR Bush NG, Evans-Roberts K, Maxwell A. DNA topoiso- in complex with inducers and with an operator. J Mol Biol merases. EcoSal Plus 2015;6. http://www.ncbi.nlm.nih.gov/ 2010;403:174–84. pubmed/26435256. Jiang H, Hutchinson CR. Feedback regulation of doxoru- Chopra I, Roberts M. Tetracycline antibiotics: mode of action, ap- bicin biosynthesis in Streptomyces peucetius. Res Microbiol plications, molecular biology, and epidemiology of bacterial 2006;157:666–74. resistance. Microbiol Mol Biol Rev 2001;65:232–60; second page, Kallberg Y, Oppermann U, Persson B. Classification of the short- tableofcontents. chain dehydrogenase/reductase superfamily using hidden Clardy J, Fischbach MA, Walsh CT. New antibiotics from bacterial Markov models. FEBS J 2010;277:2375–86. natural products. Nat Biotechnol 2006;24:1541–50. Kavanagh KL, Jornvall H, Persson B et al. Medium- and short- Collin F, Karkare S, Maxwell A. Exploiting bacterial DNA gyrase chain dehydrogenase/reductase gene and protein families: as a drug target: current state and perspectives. Appl Microbiol the SDR superfamily: functional and structural diversity Biotechnol 2011;92:479–97. within a family of metabolic and regulatory enzymes. Cell Mol Cundliffe E. How antibiotic-producing organisms avoid suicide. Life Sci 2008;65:3895–906. Annu Rev Microbiol 1989;43:207–33. Laponogov I, Pan XS, Veselkov DA et al. Structural basis of gate- Cuthbertson L, Nodwell JR. The TetR family of regulators. Micro- DNA breakage and resealing by type II topoisomerases. PLoS biol Mol Biol Rev 2013;77:440–75. ONE 2010;5:e11338. Downloaded from https://academic.oup.com/femsre/article/42/1/fux055/4604775 by DeepDyve user on 13 July 2022 112 FEMS Microbiology Reviews, 2018, Vol. 42, No. 1 Laponogov I, Sohi MK, Veselkov DA et al. Structural insight into Reichheld SE, Yu Z, Davidson AR. The induction of folding coop- the quinolone-DNA cleavage complex of type IIA topoiso- erativity by ligand binding drives the allosteric response of merases. Nat Struct Mol Biol 2009;16:667–9. tetracycline repressor. P Natl Acad Sci USA 2009;106:22263–8. Le TB, Fiedler HP, den Hengst CD et al. Coupling of the biosyn- Richter SN, Frasson I, Palumbo M et al. Simocyclinone D8 turns thesis and export of the DNA gyrase inhibitor simocyclinone on against Gram-negative bacteria in a clinical setting. Bioorg in Streptomyces antibioticus. Mol Microbiol 2009;72:1462–74. Med Chem Lett 2010;20:1202–4. Le TB, Schumacher MA, Lawson DM et al. The crystal structure Sadiq AA, Patel MR, Jacobson BA et al. Anti-proliferative effects of of the TetR family transcriptional repressor SimR bound to simocyclinone D8 (SD8), a novel catalytic inhibitor of topoi- DNA and the role of a flexible N-terminal extension in minor somerase II. Invest New Drugs. 2009. groove binding. Nucleic Acids Res 2011a;39:9433–47. Schafer M, Le TB, Hearnshaw SJ et al. SimC7 Is a Novel NAD(P)H- Le TB, Stevenson CE, Fiedler HP et al. Structures of the TetR- dependent ketoreductase essential for the antibiotic activ- like simocyclinone efflux pump repressor, SimR, and the ity of the DNA gyrase inhibitor simocyclinone. J Mol Biol mechanism of ligand-mediated derepression. J Mol Biol 2015;427:2192–204. 2011b;408:40–56. Schafer ¨ M, Stevenson CE, Wilkinson B et al. Substrate-assisted Luft T, Li SM, Scheible H et al. Overexpression, purification and catalysis in polyketide reduction proceeds via a phenolate characterization of SimL, an amide synthetase involved in intermediate. Cell Chem Biol 2016;23:1091–7. simocyclinone biosynthesis. Arch Microbiol 2005;183:277–85. Schimana J, Fiedler HP, Groth I et al. Simocyclinones, novel cy- Maxwell A, Lawson DM. The ATP-binding site of type II topoiso- tostatic angucyclinone antibiotics produced by Streptomyces merases as a target for antibacterial drugs. Curr Top Med Chem antibioticus Tu 6040. I. Taxonomy, fermentation, isolation and 2003;3:283–303. biological activities. J Antibiot (Tokyo) 2000;53:779–87. McNicholas S, Potterton E, Wilson KS et al. Presenting your struc- Schimana J, Walker M, Zeeck A et al. Simocyclinones: diversity of tures: the CCP4mg molecular-graphics software. Acta Crystal- metabolites is dependent on fermentation conditions. JInd logr D Biol Crystallogr 2011;67:386–94. Microbiol Biotechnol 2001;27:144–8. Miller DJ, Zhang YM, Subramanian C et al. Structural basis for the Schumacher MA, Miller MC, Grkovic S et al. Structural mecha- transcriptional regulation of membrane lipid homeostasis. nisms of QacR induction and multidrug recognition. Science Nat Struct Mol Biol 2010;17:971–5. 2001;294:2158–63. Morais Cabral JH, Jackson AP, Smith CV et al. Crystal struc- Schumacher MA, Miller MC, Grkovic S et al. Structural basis for ture of the breakage-reunion domain of DNA gyrase. Nature cooperative DNA binding by two dimers of the multidrug- 1997;388:903–6. binding protein QacR. EMBO J 2002;21:1210–8. Oppegard LM, Hamann BL, Streck KR et al. In vivo and in vitro Sissi C, Vazquez E, Chemello A et al. Mapping simocyclinone D8 patterns of the activity of simocyclinone D8, an angucy- interaction with DNA gyrase: evidence for a new binding site clinone antibiotic from Streptomyces antibioticus. Antimicrob on GyrB. Antimicrob Agents Chemother 2010;54:213–20. Agents Chemother 2009;53:2110–9. Tahlan K, Ahn SK, Sing A et al. Initiation of actinorhodin export Orth P, Saenger W, Hinrichs W. Tetracycline-chelated Mg2+ ion in Streptomyces coelicolor. Mol Microbiol 2007;63:951–61. initiates helix unwinding in Tet repressor induction. Biochem- Tahlan K, Yu Z, Xu Y et al. Ligand recognition by ActR, a TetR-like istry 1999;38:191–8. regulator of actinorhodin export. J Mol Biol 2008;383:753–61. Orth P, Schnappinger D, Hillen W et al. Structural basis of Theobald U, Schimana J, Fiedler HP. Microbial growth and pro- gene regulation by the tetracycline inducible Tet repressor- duction kinetics of Streptomyces antibioticus Tu 6040. Antonie operator system. Nat Struct Biol 2000;7:215–9. Van Leeuwenhoek 2000;78:307–13. Ostash I, Ostash B, Luzhetskyy A et al. Coordination of export Trefzer A, Pelzer S, Schimana J et al. Biosynthetic gene cluster and glycosylation of landomycins in Streptomyces cyanogenus of simocyclinone, a natural multihybrid antibiotic. Antimicrob S136. FEMS Microbiol Lett. 2008;285:195–202. Agents Chemother 2002;46:1174–82. Otten SL, Ferguson J, Hutchinson CR. Regulation of daunorubicin Verghese J, Nguyen T, Oppegard LM et al. Flavone-based ana- production in Streptomyces peucetius by the dnrR2 locus. JBac- logues inspired by the natural product simocyclinone D8 as teriol 1995;177:1216–24. DNA gyrase inhibitors. Bioorg Med Chem Lett 2013;23:5874–7. Pacholec M, Freel Meyers CL, Oberthur M et al. Characterization Vos SM, Tretter EM, Schmidt BH et al. All tangled up: how cells of the aminocoumarin ligase SimL from the simocyclinone direct, manage and exploit topoisomerase function. Nat Rev pathway and tandem incubation with NovM,P,N from the Mol Cell Biol 2011;12:827–41. novobiocin pathway. Biochemistry 2005;44:4949–56. Wang JC. DNA topoisomerases. Annu Rev Biochem 1996;65:635–92. Persson B, Kallberg Y. Classification and nomenclature of the su- Wang JC. Cellular roles of DNA topoisomerases: a molecular per- perfamily of short-chain dehydrogenases/reductases (SDRs). spective. Nat Rev Mol Cell Biol 2002;3:430–40. Chem Biol Interact 2013;202:111–5. Willems AR, Tahlan K, Taguchi T et al. Crystal structures of the Pommier Y, Leo E, Zhang H et al. DNA topoisomerases and their Streptomyces coelicolor TetR-like protein ActR alone and in poisoning by anticancer and antibacterial drugs. Chem Biol complex with actinorhodin or the actinorhodin biosynthetic 2010;17:421–33. precursor (S)-DNPA. J Mol Biol 2008;376:1377–87. Ramos JL, Martinez-Bueno M, Molina-Henares AJ et al. The TetR Yu Z, Reichheld SE, Savchenko A et al. A comprehensive analysis family of transcriptional repressors. Microbiol Mol Biol Rev of structural and sequence conservation in the TetR family 2005;69:326–56. transcriptional regulators. J Mol Biol 2010;400:847–64.

Journal

FEMS Microbiology Reviews – Oxford University Press

Published: Jan 1, 2018

Keywords: dna gyrase; ligands; streptomyces; antibiotics; biosynthesis; enzymes; dna; binding (molecular function); transcription factor

There are no references for this article.