Structural Evidence for Rifampicin Monooxygenase InactivatingRifampicin by Cleaving Its Ansa-BridgeLi-Kai Liu,† Yumin Dai,§ Heba Abdelwahab,§ Pablo Sobrado,*,§ and John J. Tanner*,†,‡

†Department of Biochemistry and ‡Department of Chemistry, University of MissouriColumbia, Columbia, Missouri 65211, UnitedStates§Department of Biochemistry, Virginia Tech, Blacksburg, Virginia 24061, United States

*S Supporting Information

ABSTRACT: Rifampicin monooxygenase (RIFMO) de-creases the potency of rifampicin (RIF) by converting it tooxidative products. Further decomposition of RIF has beenobserved in bacteria producing RIFMO and contributes toRIFMO-mediated drug resistance. Here we report the firstcrystal structure of RIFMO in complex with thehydroxylated RIF product. The 2.10 Å resolution structurereveals a breach of the ansa aliphatic chain of RIF betweennaphthoquinone C2 and amide N1. Our data suggest thatRIFMO catalyzes the hydroxylation of RIF at the C2 atomfollowed by cleavage of the ansa linkage, which leads toinactivation of the antibiotic by preventing key contactswith the RNA polymerase target.

Microorganisms have evolved sophisticated strategies toevade antibiotics, such as efflux pumps, mutagenesis of

proteins targeted by antibiotics, and enzyme-catalyzed chemicalmodification of antibiotics. Among the latter category, theenzyme rifampicin monooxygenase (RIFMO, also known asRox) has been identified as a flavin-dependent monooxygenasethat modifies rifampicin (RIF) (Figure 1A), an antibiotic usedin combination therapy to treat mycobacterial and non-mycobacterial infections.1,2 The gene encoding RIFMO ispresent in a variety of environmental bacteria, notably Nocardia,Streptomyces, and Rhodococcus species.3−6 RIFMO is one of twoknown flavin monooxygenases that inactivate antibiotics, theother being TetX, which acts on tetracycline and has a proteinfold different from that of RIFMO.7−9

The mechanism of RIFMO is an area of active investigation.RIFMO (from Nocardia farcinica) was initially characterized asan NADPH-dependent N-monooxygenase proposed to hydrox-ylate RIF at the N2′ atom to produce 2′-N-hydroxy-4-oxo-RIF.This hydroxylation was proposed to lead to subsequent RIFdecomposition.3,4 However, direct structural evidence for thismodification and degradation of RIF has remained elusive.Further, it was unclear how the addition of a hydroxyl group toN2′ compromises the binding of RIF to its target, the β-subunitin RNA polymerase. Recently, Koteva et al. showed thatRIFMO from Streptomyces venezuelae, in fact, catalyzes adifferent reaction involving monooxygenation of position 2 ofthe naphthyl group, followed by linearization of the antibiotic(Figure 1A).5 Direct structural evidence for the linearizedproduct, however, is lacking.

Herein, we report the first crystal structure of RIFMOcomplexed with its product. The structure provides directevidence for the Koteva et al. linearization mechanism. We alsoprovide nuclear magnetic resonance (NMR) analysis furthersupporting this mechanism.A 2.10 Å resolution crystal structure of RIFMO from N.

farcinica (74% identical to the S. venezuelae enzyme) wasdetermined from a crystal prepared in the presence of RIF andNADPH [Table S1, Protein Data Bank (PDB) entry 6C7S].The electron density maps clearly show a form of RIF bound inthe active site (Figure 1C). As in our previous structure ofRIFMO complexed with RIF,10 strong electron density placesthe naphthoquinone ring system next to the FAD N5 atomedge of the flavin isoalloxazine (Figure 1B). However, the

Received: February 16, 2018Revised: March 22, 2018Published: March 26, 2018

Figure 1. Structural evidence for cleavage of the RIF ansa-bridge byRIFMO. (A) Reaction catalyzed by RIFMO. (B) Pose of RIF-OH inthe active site. (C) Electron density for RIF-OH. The top panel showsthe final 2FO − FC map contoured at 1.0σ. The bottom panel shows asimulated annealing FO − FC omit map contoured at 2.5σ.

Communication

pubs.acs.org/biochemistryCite This: Biochemistry 2018, 57, 2065−2068

© 2018 American Chemical Society 2065 DOI: 10.1021/acs.biochem.8b00190Biochemistry 2018, 57, 2065−2068

density suggested that the bond between C2 of thenaphthoquinone and N1 of the ansa-bridge was broken,resulting in a gap of 7.0 Å between these two atoms (Figure1C). The maps further indicated an atom bound to C2, whichwas modeled as O and presumed to be OH of the product 2-hydroxyl-rifampicin (RIF-OH) based on NMR analysispresented below. Assuming the bound species represents theproduct of RIFMO catalysis, the structure suggests that RIFMOconverts RIF from a macrocyclic compound into a linearmolecule, consistent with the mechanism of Koteva et al.5

Transformation of the substrate into the product results inprofound conformational changes (Figure 2A). No longer

constrained by the covalent bond with C2, the N1 atom shiftsby 5.7 Å, creating a gap of 7.0 Å between the formerly bondedatoms. The rupture of the C2−N1 bond allows the ansa-bridgeto relax, with atoms of the bridge moving by 7−8 Å. Theprotein structure also changes during catalysis. In particular,upon product formation, Arg43 moves into the space vacatedby the ansa-bridge (Figure 2A).Previously, we have shown by HPLC and spectral analyses

that the reaction of RIFMO with RIF occurs via the formationof an intermediate (P*), which converts over time to the finalRIF-OH product.4 We reanalyzed the 1H NMR assignments ofRIF-OH and compared it to RIF. We also acquired electrosprayionization mass spectrometry, 1H−13C HMBC, and 13C NMRdata for P* and RIF-OH for further characterization. Analysisof the mass showed that the intermediate and final productmost likely are tautomers, as an ion with an m/z value of 839.5consistent with the addition of a hydroxyl to RIF was observed

for both samples (Figure S1). The 1H NMR spectra of RIF-OHfrom our previous work4 helped us identify signals of amideprotons (NH) in the aromatic region, which were split into twopeaks at 7.07 (1H) and 7.23 (1H) ppm. The integration ofboth peaks is one, clearly suggesting the presence of twodistinct NH protons in RIF-OH (Figure S2). To confirm thelocation of these two NH protons, 1H−13C HMBC experi-ments were conducted. From the two-dimensional NMRspectra, strong correlations between NH and C15, as well asbetween NH and C16, were observed, while there is nocorrelation displayed between NH and C2, indicating these twoprotons were both primary amide protons located the end ofthe chain system (Figure S3). The observation of twomagnetically inequivalent protons displaying two distinctchemical shifts is consistent with attachment to an amide group.In addition, we made chemical shift assignments for 1H

NMR of P* (Figure S2C). Interestingly, the 1H NMR spectraof P* and RIF-OH are very similar, differing only in hydroxylprotons located at 3.0−4.0 ppm. The P* 1H NMR spectrumalso shows two peaks corresponding to the two single NHprotons at 7.07 (1H) and 7.24 (1H) ppm, indicating that theC2−N1 bond is also cleaved in P*.We also acquired 13C NMR spectra of RIF-OH to look for a

change at C2. Inspection of the 13C NMR data revealed majordifferences between RIF and RIF-OH located in C1−C10,where the naphthoquinol was converted to naphthoquinone(Table S2). For RIF-OH, the chemical shift of C2 was at 176.3ppm, indicating C2 is an oxygen-bearing carbon, while for RIF,nitrogen-bearing C2 is assigned to 115.5 ppm, which wasconsistent with previously reported C NMR data of RIF.3 Theshift of the carbon signal of C2 from 115.5 to 176.3 ppm servedas a direct evidence of cleavage of the bond between C2 andN1. The rest of the signals involved in the naphthoquinonesystem all have shifts affected by the conversion fromnaphthoquinol to naphthoquinone. In particular, in thestructure of RIF, both C1 and C4 were connected to thehydroxyl group; however, in RIF-OH, both carbons wereconverted from hydrogenated carbons to carbonyl, as suggestedby the observation of the signals downfield-shifted from 148.6to 189.5 ppm for C1 and from 145.4 to 189.2 ppm for C4.Overall, the NMR analysis indicates that P* and RIF-OH arelinearized RIFMO reaction products, consistent with thereaction described by Koteva et al. for the S. venezuelaeenzyme.5

Taken together, our crystal structure, mass spectra, andNMR data are consistent with the mechanism of Koteva et al.5

(Scheme 1). During the reaction, the C1 hydroxyl isdeprotonated to facilitate nucleophilic attack at the distalhydroxyl of the reactive RIFMO-C4a-hydroperoxyflavin. The

Figure 2. Comparison of the substrate and product complexes ofRIFMO. (A) Superposition of RIFMO complexed with the substrateRIF (cyan, PDB entry 5KOX) and the product RIF-OH (white, PDBentry 6C7S). The arrows indicate the conformational changes thatoccur as a result of catalysis. (B) Structure of RIFMO complexed withRIF, highlighting interactions between the FAD and RIF (PDB entry5KOX).

Scheme 1. Mechanism of RIFMO5

Biochemistry Communication

DOI: 10.1021/acs.biochem.8b00190Biochemistry 2018, 57, 2065−2068

2066

C2 hydroxyl of the tetrahedral intermediate thus generated issubject to deamination of the ansa amide from naphthoquinoneand results in cleavage of the C2−N1 bond. This generates asemiquinone P* with a carbonyl at C2, which subsequentlyequilibrates to the final product RIF-OH. Deprotonation of theC1 hydroxyl in the initial step of the mechanism can befacilitated by the hydrogen bond ring system incorporating C1and C811 and the donation of a hydrogen bond from the N5atom of the reduced FAD to O1 of RIF (Figure 2B). Manyrifamycin derivatives have the conserved naphthoquinone C1−C4 moiety. Therefore, other rifamycins that are capable ofgenerating the C1 anion could be substrates for RIFMO.The linearization reaction is consistent with the structure of

RIFMO complexed with the RIF, which we recently reported.10

In the RIFMO−RIF complex, the FAD C4a atom is only 4.7 Åfrom the new proposed site of hydroxylation on RIF, C2(Figure 2B). Thus, the structure is consistent with transfer of ahydroxyl group from the RIFMO-C4a-hydroperoxyflavin to C2of RIF. In contrast, the originally proposed site ofhydroxylation, N2′, is 7.7 Å from C4a and shielded by anactive site loop. Thus, N2′ is poorly positioned to react withC4a-hydroperoxyflavin. We previously invoked a conforma-tional change in the protein and/or substrate to rationalize theapparent discrepancy between the structure and the mecha-nism. The linearization reaction solves this conundrum.The linearization reaction also provides a plausible and direct

explanation for how RIFMO compromises the antibioticactivity of RIF. The binding site for RIF on RNA polymeraseis complementary to the oval shape of the macrocycle, and theansa-bridge makes extensive contacts with the enzyme (Figure3).12 For example, the ansa-bridge contacts 12 residues of RNA

polymerase (4 Å cutoff). Of the 42 atoms of RNA polymerasethat contact RIF, 36 contact the ansa-bridge. This binding poseis consistent with studies showing that the two hydroxyl groupsof the ansa-bridge are critical for the activity of RIF;13 notethese two hydroxyls (O21 and O23) are directed toward thepolymerase (Figure 3). It seems unlikely that the linearized RIFobserved here could adopt the oval conformation of themacrocycle required to make such an intimate and specificassociation with RNA polymerase. To do so would require asubstantial entropy penalty, which would reduce the bindingaffinity. Furthermore, the originally proposed site of hydrox-ylation, N2′, does not contact RNA polymerase, and it seemsunlikely that the addition of a hydroxyl group at this positionwould substantially decrease the binding affinity (Figure 3). Insummary, linearization of the ansa-bridge by RIFMO is a moreplausible mechanism of inactivation of RIF than hydroxylationof N2′.

In conclusion, we presented the first direct structuralevidence showing the product of the RIFMO reaction. Thestructure confirms that RIFMO is not an N-hydroxylatingenzyme as initially thought. Rather, it is a C-hydroxylatingmonooxygenase that cleaves the ansa-bridge of rifampicin,compromising the binding affinity of the drug for RNApolymerase.

■ ASSOCIATED CONTENT*S Supporting InformationThe Supporting Information is available free of charge on theACS Publications website at DOI: 10.1021/acs.bio-chem.8b00190.

Experimental Procedures, Tables S1 and S2, and FiguresS1−S3 (PDF)

■ AUTHOR INFORMATIONCorresponding Authors*E-mail: psobrado@vt.edu.*E-mail: tannerjj@missouri.edu.ORCIDYumin Dai: 0000-0002-9228-3671Pablo Sobrado: 0000-0003-1494-5382John J. Tanner: 0000-0001-8314-113XFundingThis work was supported by National Science FoundationGrants CHE-1506206 and MCB 1021384.NotesThe authors declare no competing financial interest.

■ ACKNOWLEDGMENTSThe authors thank Dr. Jay Nix for help with X-ray diffractiondata collection at the Advanced Light Source and Dr. Wei G.Wycoff for providing comments on the NMR data. TheAdvanced Light Source is supported by the Director, Office ofScience, Office of Basic Energy Sciences, of the U.S.Department of Energy under Contract DE-AC02-05CH11231.

■ REFERENCES(1) Forrest, G. N., and Tamura, K. (2010) Rifampin combinationtherapy for nonmycobacterial infections. Clin. Microbiol. Rev. 23 (1),14−34.(2) Lee, C.-Y., Huang, C.-H., Lu, P.-L., Ko, W.-C., Chen, Y.-H., andHsueh, P.-R. (2017) Role of rifampin for the treatment of bacterialinfections other than mycobacteriosis. J. Infect. 75 (5), 395−408.(3) Hoshino, Y., Fujii, S., Shinonaga, H., Arai, K., Saito, F., Fukai, T.,Satoh, H., Miyazaki, Y., and Ishikawa, J. (2010) Monooxygenation ofrifampicin catalyzed by the rox gene product of Nocardia farcinica:structure elucidation, gene identification and role in drug resistance. J.Antibiot. 63 (1), 23−28.(4) Abdelwahab, H., Martin Del Campo, J. S., Dai, Y., Adly, C., El-Sohaimy, S., and Sobrado, P. (2016) Mechanism of RifampicinInactivation in Nocardia farcinica. PLoS One 11 (10), e0162578.(5) Koteva, K., Cox, G., Kelso, J. K., Surette, M. D., Zubyk, H. L.,Ejim, L., Stogios, P., Savchenko, A., Sorensen, D., and Wright, G. D.(2018) Rox, a Rifamycin Resistance Enzyme with an UnprecedentedMechanism of Action. Cell Chem. Biol., n/a.(6) Andersen, S. J., Quan, S., Gowan, B., and Dabbs, E. R. (1997)Monooxygenase-like sequence of a Rhodococcus equi gene conferringincreased resistance to rifampin by inactivating this antibiotic.Antimicrob. Agents Chemother. 41 (1), 218−21.(7) Park, J., Gasparrini, A. J., Reck, M. R., Symister, C. T., Elliott, J. L.,Vogel, J. P., Wencewicz, T. A., Dantas, G., and Tolia, N. H. (2017)

Figure 3. RIF-binding site of RNA polymerase (PDB entry 1I6V). Thered surface indicates atoms that contact RIF (4 Å cutoff).

Biochemistry Communication

DOI: 10.1021/acs.biochem.8b00190Biochemistry 2018, 57, 2065−2068

2067

Plasticity, dynamics, and inhibition of emerging tetracycline resistanceenzymes. Nat. Chem. Biol. 13 (7), 730−736.(8) Volkers, G., Palm, G. J., Weiss, M. S., Wright, G. D., and Hinrichs,W. (2011) Structural basis for a new tetracycline resistance mechanismrelying on the TetX monooxygenase. FEBS Lett. 585 (7), 1061−6.(9) Forsberg, K. J., Patel, S., Wencewicz, T. A., and Dantas, G. (2015)The Tetracycline Destructases: A Novel Family of Tetracycline-Inactivating Enzymes. Chem. Biol. 22 (7), 888−97.(10) Liu, L. K., Abdelwahab, H., Martin Del Campo, J. S., Mehra-Chaudhary, R., Sobrado, P., and Tanner, J. J. (2016) The Structure ofthe Antibiotic Deactivating, N-hydroxylating Rifampicin Monooxyge-nase. J. Biol. Chem. 291 (41), 21553−21562.(11) Howes, B. D., Guerrini, L., Sanchez Cortes, S., Marzocchi, M. P.,Garcia Ramos, J. V., and Smulevich, G. (2007) The influence of pHand anions on the adsorption mechanism of rifampicin on silvercolloids. J. Raman Spectrosc. 38 (7), 859−864.(12) Campbell, E. A., Korzheva, N., Mustaev, A., Murakami, K., Nair,S., Goldfarb, A., and Darst, S. A. (2001) Structural Mechanism forRifampicin Inhibition of Bacterial RNA Polymerase. Cell 104 (6),901−912.(13) Brufani, M., Cerrini, S., Fedeli, W., and Vaciago, A. (1974)Rifamycins: an insight into biological activity based on structuralinvestigations. J. Mol. Biol. 87 (3), 409−435.

Biochemistry Communication

DOI: 10.1021/acs.biochem.8b00190Biochemistry 2018, 57, 2065−2068

2068

S1

SUPPORTING INFORMATION Structural evidence for rifampicin monooxygenase inactivating rifampicin by cleaving its ansa-bridge

Li-Kai Liu,† Yumin Dai,‡ Heba Abdelwahab,‡ Pablo Sobrado,‡,* and John J. Tanner†,^,*

From the †Departments of Biochemistry and ^Chemistry, University of Missouri-Columbia, Columbia, Missouri 65211 and ‡Department of Biochemistry, Virginia Tech, Blacksburg, VA 24061

To whom correspondence should be addressed: John J. Tanner, Department of Biochemistry, University of Missouri-Columbia, Columbia, MO 65211, USA. Tel: (573) 884-1280; Email: tannerjj@missouri.edu; Pablo Sobrado, Department of Biochemistry, Virginia Tech, Blacksburg, VA 24061; Tel: (540) 231-9485; Email: psobrado@vt.edu.

Table of Contents Experimental Procedures

Crystallization and crystal soaking…………………..………………...…………... S2

X-ray diffraction data collection and structure determination………...…………… S2

Product isolation………...………………………...……………………………….. S3

Mass spectrometry….....…………………………………………………………… S4

NMR analysis of isolated products………………………...………………………. S5

Table S1. X-ray diffraction data collection and refinement statistics …………………..… S6

Table S2. 13C-NMR spectroscopic data for RIF-OH……………………..……..……….... S7

Figure S1. Mass analysis of the RIFMO reaction products……….………….................… S8

Figure S2. 1H-NMR of RIF and reaction products………. ……………………..………... S9

Figure S3. 1H-13C HMBC correlations of RIF-OH………..……………………………… S11

References…………………………………………………………………………………. S12

S2

Experimental Procedures

Crystallization and crystal soaking. The full-length RIFMO construct from Nocardia

farcinica was expressed in Escherichia coli and purified as described previously.1-2 Crystals of

RIFMO complexed with RIF-OH were obtained by co-crystallization in a PEG condition,

followed by crystal soaking. They were obtained by hanging-drop vapor diffusion using the

protein concentrated to 10 mg/ml in a buffer consisting of 25 mM Hepes at pH 7.5 and 100 mM

NaCl. A stock solution of 200 mM RIF in 100% dimethyl sulfoxide was prepared. Prior to

crystallization, the protein sample was supplemented first with RIF at a final concentration of 5

mM (2.5% dimethyl sulfoxide final concentration), then with NADPH at 100 mM final

concentration. The protein-ligand solution was mixed well and allowed to equilibrate for 30

minutes. The sample was centrifuged at the top speed for 5 minutes, and the supernatant was

used for setting up crystallization experiments. Crystallization drops were prepared by mixing

1.0 volume of the protein-ligand sample with 1.0 volume of the reservoir solution (24% (w/v)

PEG 3350, 200 mM magnesium chloride), and 0.2 volume of a 500x microseed stock made from

the same crystals. Yellow crystals grew to 40 × 40 × 200 µm3 over a period of a week at room

temperature. The crystals were then soaked for 1.5 hours in a reducing buffer containing 5 mM

RIF, 100 mM NADPH, and 100 mM dithionite before cryo-protecting with 15% glycerol and

flash-cooling in liquid N2.

X-ray diffraction data collection and structure determination. Diffraction data were

collected in shutterless mode from single crystals at beamline 4.2.2 at the Advanced Light

Source, Berkeley, CA. Each data set consisted of a wedge of 180° of data collected with a

Taurus-1 detector over a period of 180 seconds. Images were written to disk every 0.2 seconds,

so that the data set consisted of 900 images with an effective oscillation width of 0.2°. The data

S3

were indexed, integrated, and scaled with the XDS package.3 Similar to our published RIFMO

structures, the space group is P6522 with unit cell parameters of a = b = 81.9 Å, and c = 286.2 Å.

There is one RIFMO complex in the asymmetric unit. Intensities were converted to amplitudes

with Truncate.4 Data processing statistics are provided in Table S1.

Initial phases were determined by molecular replacement with the program PHASER.5 The

apo structure was used to solve the co-complex structure with RIF-OH bound (PDB ID 5KOW).

Iterative rounds of model building and restrained refinement were carried out with COOT6-7and

phenix.refine.8 The initial coordinates of RIF-OH were generated in JLigand.9 The restraint files

for both RIF-OH and FAD were generated in eLBOW.10 Electron density for the

methylpiperazine of RIF-OH was missing; therefore, these atoms were omitted from the model

deposited in the PDB. The absence of electron could reflect conformational disorder or

hydrolysis of the imine linkage of RIF to liberate 1-amino-4-methylpiperazine.11-12 We note that

the methylpiperazine of RIF is missing in other structures (PDB IDs 1SKX, 5KOX, and 6BRD).

The structure was validated with MolProbity.13

Product Isolation. The isolation of the final product of the RIFMO reaction as well as the

intermediate P* were described previously.2 The production and isolation of P* and RIF-OH

were achieved in two different reaction conditions. Both reactions were incubated in 100

mM sodium phosphate buffer, pH 7.5, in the presence of an NADPH-regenerating system.

For the P*-generating reaction, a 220-mL solution that consisted of 100 μM NADP+, 10

mM glucose-6-phosphate (G6P), and 500 units of glucose-6-phosphate dehydrogenase

(G6PDH) was incubated at 25°C for 45 min with occasional shaking to generate NADPH.

Next, 100 μM RIF was added to this mixture and the reaction was initiated by the

addition of 1 μM RIFMO and incubated for 10 min with shaking at 25°C. For the RIF-

S4

OH-generating reaction, a 20-mL solution containing 1 mM NADP+, 10 mM G6P, and

500 units of G6PDH was incubated at 25°C for 45 min with occasional shaking. Then, 1

mM RIF was added to this mixture and the reaction was initiated by the addition of 10

μM RIFMO and incubated for 2.5 h at 25°C with shaking. The progress of each reaction

was monitored by HPLC to validate the formation of only P* (retention time at 13.4 min)

or Rif-OH (at 22.1 min) (see Figure 5 in Abdelwahab et al.2 ). The reactions were then

quenched with equal volumes of chloroform. Upon phase separation, the lower layers

were separated and the solvent was completely evaporated. Finally, the residual products

were re-dissolved in methanol and used for subsequent purification and separation by

preparative thin-layer chromatography, using 0.5 mm thick silica gel 60F254 (MP

Biomedicals) and using the lower phase of the mixture [chloroform:methanol:H2O

(65:15:5, v/v/v)] as the mobile phase. The bands corresponding to P* and RIF-OH were

scraped from the TLC plates and extracted with chloroform under sonication and

centrifuged. The organic solvent was evaporated using a Rotavapor until complete

dryness and the dried samples were stored in dark at 4°C for further analysis.

Mass spectrometry. RIF and reaction product samples were initially dissolved in DMSO to

make the concentration of 10 mg/mL, and were further diluted with 50/50 MeOH/H2O

containing 0.1% formic acid to make the final concentration of 0.5 mg/mL. All analyses were

performed using an Agilent 1200 Series HPLC and Thermo TSQ Quantum mass spectrometry

(positive mode). 5 µL of each sample was loaded to the C18 column (Phenomenex C18 Luna,

3.5 µm dp, 150 x 2 mm) equilibrated with 95/5% A/B (A: water + 0.1% formic acid, B:

acetonitrile + 0.1% formic acid), and was eluted under the linear gradient program (95/5% to

S5

5/95% (A/B) over 10 min, isocratic at 5/95% for additional 5 min and equilibrated back to 95/5%

in 5 min), with the flow rate of 0.2 mL/min. Electrospray ionization was employed at 4500 volts

and a temperature of 300°C. Each mass spectroscopic data was acquired after running blank as

the negative control.

NMR analysis of isolated products. NMR analysis was performed on both the final purified

product RIF-OH as well as the intermediate P*. 2 mg of each sample were dissolved in DMSO-

d6 and the 1H-NMR, 13C-NMR as well as 1H-13C-HMBC experiments were conducted on a

Bruker Avance 600 instrument located in the Department of Chemistry, Virginia Tech. The

chemical shifts are given in d (ppm), and coupling constants (J) are reported in Hz.

S6

Table S1. X-ray diffraction data collection and refinement statisticsa

RIFMO-RIF-OHPDB ID code 6C7S Diffraction source ALS 4.2.2Space group P 65 2 2

Unit cell parameters (Å) a = 81.9, c = 286.2

Resolution range (Å) 63.54-2.10 (2.16-2.10)Total No. of reflections 636732 (31138)

No. of unique reflections 33890 (2345)

Completeness (%) 98.6 (86.6)Redundancy 18.8 (13.3)⟨I/σ(I)⟩ 25.3 (3.6) Rmerge 0.089 (0.669)Rmeas 0.094 (0.721)Rpim 0.029 (0.260)Mean CC1/2 0.999 (0.843)Wilson B factor (Å2) 26.2Rcryst 0.1999Rfree

b 0.2411No. of non-H atoms All 4076 Protein 3635 FAD 53 RIF-OH 53 Water 312RMSD bonds (Å) 0.002RMSD angles (°) 0.684Average B factors (Å2) All 30.7 Protein 30.2 FAD 29.7 RIF-OH 41.6 Water 33.7Ramachandran plotc Favoured (%) 97.1 Allowed (%) 2.9 Outliers 0.0Clashscorec (%-tile) 2.56 (99%)

Molprobity scorec (%-tile) 1.21 (100%)

Coordinate error (Å)d 0.24aValues for the outer resolution shell of data are given in parenthesis. bRandom 5% test set. cGenerated with MolProbity. dMaximum likelihood-based coordinate error estimate from phenix.refine.

S7

Table S2. 13C-NMR Spectroscopic Data for RIF-OH in DMSO-d6 (600 MHz)

RIF RIF-OH position δC

δC* δC δC

* 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 1’ 4’ 5’ 7’

148.6 115.5 117.1 145.4 98.5 171.8 101.0 184.4 114.3 117.2 184.4 108.6 22.0 7.5

166.3 132.2 132.0 127.7 138.2 38.4 72.1 32.9 76.4 38.1 73.6 40.1 76.5 117.6 142.8 20.6 17.8 11.4 8.7 9.0

169.5 20.7 55.7 137.6 47.7 51.2 42.1

148.5 117.2 115.5 145.4 101.0 171.8 98.5 184.2 114.2 117.0 184.2 108.5 22.0 7.4

166.2 132.2 132.0 127.7 138.1 38.4 72.1 32.9 76.1 38.1 73.6 40.1 76.4 117.6 142.8 20.6 17.8 11.4 8.7 9.0

169.5 20.7 55.6 137.6 47.9 51.2 42.1

189.5 176.3 113.6 189.2 114.4 170.3 112.5 168.4 109.3 126.5 190.3 105.8 20.5 7.7

169.6 130.3 131.9 126.2 140.3 40.4 73.2 34.0 76.0 36.6 72.7 40.1 77.2 111.8 141.4 20.6 16.3 10.8 9.6 10.0 168.8 21.8 54.9 154.0 52.9 54.1 45.1

179.2 174.1 107.8 185.1 114.8 170.7 113.0 169.1 108.7 135.7 190.2 105.7 20.5 7.5

170.9 130.5 132.4 126.3 141.0 40.4 73.6 34.0 75.9 36.3 72.6 39.9 76.9 111.7 141.3 20.7 16.6 10.9 9.9 10.2 169.8 21.0 55.2 n.o. 53.6 53.5 44.8

*Data published by Koteva et al.14

S8

Figure S1. Mass analysis of the RIFMO reaction products.

S9

Figure S2. 1H-NMR spectra of RIF and reaction products. (A) Overlaid 1H-NMR spectra of RIF, P*, and RIF-OH. (B) 1H-NMR of RIF (DMSO-d6, 500 MHz). (C) 1H-NMR of P* (DMSO-d6, 500 MHz). (D) 1H-NMR of RIF-OH (DMSO-d6, 500 MHz) from Abdelwahab, et al.2 A

B

DMSO-d6

S10

C

D

DMSO-d6

DMSO-d6

S11

Figure S3. 1H-13C HMBC correlations of RIF-OH.

S12

References

1. Liu, L. K.; Abdelwahab, H.; Martin Del Campo, J. S.; Mehra-Chaudhary, R.; Sobrado, P.; Tanner, J. J., The Structure of the Antibiotic Deactivating, N-hydroxylating Rifampicin Monooxygenase. J. Biol. Chem. 2016, 291 (41), 21553-21562.

2. Abdelwahab, H.; Martin Del Campo, J. S.; Dai, Y.; Adly, C.; El-Sohaimy, S.; Sobrado, P., Mechanism of Rifampicin Inactivation in Nocardia farcinica. PLoS One 2016, 11 (10), e0162578.

3. Kabsch, W., XDS. Acta Crystallogr. D Biol. Crystallogr. 2010, 66 (Pt 2), 125-32. 4. French, G. S.; Wilson, K. S., On the treatment of negative intensity observations. Acta Cryst.

1978, A34, 517-525. 5. McCoy, A. J.; Grosse-Kunstleve, R. W.; Adams, P. D.; Winn, M. D.; Storoni, L. C.; Read, R.

J., Phaser crystallographic software. J. Appl. Crystallogr. 2007, 40 (Pt 4), 658-674. 6. Emsley, P.; Cowtan, K., Coot: model-building tools for molecular graphics. Acta Cryst. 2004,

D60 (Pt 12 Pt 1), 2126-32. 7. Emsley, P.; Lohkamp, B.; Scott, W. G.; Cowtan, K., Features and development of Coot. Acta

Cryst. D Biol. Crystallogr. 2010, 66 (Pt 4), 486-501. 8. Afonine, P. V.; Grosse-Kunstleve, R. W.; Echols, N.; Headd, J. J.; Moriarty, N. W.;

Mustyakimov, M.; Terwilliger, T. C.; Urzhumtsev, A.; Zwart, P. H.; Adams, P. D., Towards automated crystallographic structure refinement with phenix.refine. Acta Crystallogr. D Biol. Crystallogr. 2012, 68 (Pt 4), 352-67.

9. Lebedev, A. A.; Young, P.; Isupov, M. N.; Moroz, O. V.; Vagin, A. A.; Murshudov, G. N., JLigand: a graphical tool for the CCP4 template-restraint library. Acta Crystallogr D Biol Crystallogr 2012, 68 (Pt 4), 431-40.

10. Moriarty, N. W.; Grosse-Kunstleve, R. W.; Adams, P. D., electronic Ligand Builder and Optimization Workbench (eLBOW): a tool for ligand coordinate and restraint generation. Acta Crystallogr. D Biol. Crystallogr. 2009, 65 (Pt 10), 1074-80.

11. Prankerd, R. J.; Walters, J. M.; Parnes, J. H., Kinetics for degradation of rifampicin, an azomethine-containing drug which exhibits reversible hydrolysis in acidic solutions. International Journal of Pharmaceutics 1992, 78 (1-3), 59-67.

12. Arca, H. Ç.; Mosquera-Giraldo, L. I.; Pereira, J. M.; Sriranganathan, N.; Taylor, L. S.; Edgar, K. J., Rifampin Stability and Solution Concentration Enhancement Through Amorphous Solid Dispersion in Cellulose ω-Carboxyalkanoate Matrices. Journal of pharmaceutical sciences 2018, 107 (1), 127-138.

13. Chen, V. B.; Arendall, W. B., 3rd; Headd, J. J.; Keedy, D. A.; Immormino, R. M.; Kapral, G. J.; Murray, L. W.; Richardson, J. S.; Richardson, D. C., MolProbity: all-atom structure validation for macromolecular crystallography. Acta Crystallogr. D Biol. Crystallogr. 2010, D66 (Pt 1), 12-21.

14. Koteva, K.; Cox, G.; Kelso, J. K.; Surette, M. D.; Zubyk, H. L.; Ejim, L.; Stogios, P.; Savchenko, A.; Sorensen, D.; Wright, G. D., Rox, a Rifamycin Resistance Enzyme with an Unprecedented Mechanism of Action. Cell Chem Biol 2018.