JBIC Journal of Biological Inorganic Chemistry

, Volume 23, Issue 7, pp 1057–1070 | Cite as

The Asp99–Arg188 salt bridge of the Pseudomonas aeruginosa HemO is critical in allowing conformational flexibility during catalysis

  • Geoffrey A. Heinzl
  • Weiliang Huang
  • Elizabeth Robinson
  • Fengtian Xue
  • Pierre Möenne-Loccoz
  • Angela Wilks
Original Paper
Part of the following topical collections:
  1. Alison Butler: Papers in Celebration of Her 2018 ACS Alfred Bader Award in Bioorganic or Bioinorganic Chemistry


The P. aeruginosa iron-regulated heme oxygenase (HemO) is required within the host for the utilization of heme as an iron source. As iron is essential for survival and virulence, HemO represents a novel antimicrobial target. We recently characterized small molecule inhibitors that bind to an allosteric site distant from the heme pocket, and further proposed binding at this site disrupts a nearby salt bridge between D99 and R188. Herein, through a combination of site-directed mutagenesis and hydrogen–deuterium exchange mass spectrometry (HDX-MS), we determined that the disruption of the D99–R188 salt bridge leads to significant decrease in conformational flexibility within the distal and proximal helices that form the heme-binding site. The RR spectra of the resting state Fe(III) and reduced Fe(II)-deoxy heme-HemO D99A, R188A and D99/R188A complexes are virtually identical to those of wild-type HemO, indicating no significant change in the heme environment. Furthermore, mutation of D99 or R188 leads to a modest decrease in the stability of the Fe(II)-O2 heme complex. Despite this slight difference in Fe(II)-O2 stability, we observe complete loss of enzymatic activity. We conclude the loss of activity is a result of decreased conformational flexibility in helices previously shown to be critical in accommodating variation in the distal ligand and the resulting chemical intermediates generated during catalysis. Furthermore, this newly identified allosteric binding site on HemO represents a novel alternative drug-design strategy to that of competitive inhibition at the active site or via direct coordination of ligands to the heme iron.

Graphical abstract


Pseudomonas aeruginosa Heme oxygenase Biliverdin Oxygen activation Protein dynamics 



Circular dichroism


Hydrogen–deuterium exchange mass spectrometry


Resonance Raman spectroscopy


Double mutant




Pseudomonas aeruginosa iron-regulated heme oxygenase


Biliverdin IXα selective mutant of P. aeruginosa iron-regulated heme oxygenase


(4-(2-hydroxyethyl)-1-piperazineethanesulfonic acid


Heme oxygenase


Human heme oxygenase 1


Human heme oxygenase 2


Hydrogen–deuterium exchange mass spectrometry


Infrared fluorescent protein


Isopropyl β-d-thiogalactopyranoside


Iron-regulated surface-determinant protein G


Iron-regulated surface-determinant protein I




Mycobacterium heme utilization degrader


Nickel–nitriloacetic acid


Phenylmethanesulfonyl fluoride


Rat heme oxygenase 1


Sodium dodecyl sulfate-polyacrylamide gel electrophoresis


Wild type


Canonical heme oxygenases (HO) catalyze the oxidative cleavage of heme to biliverdin (BVIX), and have been identified in animals [1, 2, 3, 4, 5, 6] plants [7, 8], and bacteria [9, 10, 11, 12], where they play a central role in iron recycling [13], antioxidant defense [14], signaling [15], iron acquisition [9, 10, 11] and synthesis of light-sensing bilins [12, 16], respectively. While non-canonical HO’s have recently been identified in Gram-positive pathogens, the overall structure and oxidative cleavage of heme is distinct from that of the canonical HO’s [17, 18, 19]. The canonical HO’s in Gram-negative pathogens and higher eukaryotes display an overall similar helical fold, where heme is bound between the proximal (I and II) and distal helices (VII and VIII), respectively (Fig. 1a) [20, 21].
Fig. 1

Structure of the P. aeruginosa HemO in the resting ferric state (PDB:1SK7). a The heme is held between proximal helices (I and II) and distal helices (VII and VIII) with the Gly–Gly kink connecting helix VII and VIII. Heme is ligated through proximal His-26 shown in blue; B. Rotation 90° showing a side view with the D99–R188 salt bridge between the loop connecting helix VII and VIII and helix XII shown in stick format

Over the past two decades, many groups have contributed to the mechanistic understanding of oxidative heme cleavage to biliverdin IXα (BVIXα) with the concomitant release of Fe3+ and carbon monoxide (CO(g)) [22, 23, 24, 25, 26, 27, 28, 29, 30, 31]. Through these studies, several reaction intermediates have been spectroscopically and structurally characterized, including the reduced oxy (Fe2+-O2) species, the ferric hydroperoxide (Fe3+-OOH) intermediate, α-meso-hydroxyheme, α-verdoheme, and the terminal product biliverdin [32, 33, 34, 35].

In contrast to all other canonical HO’s, the oxidative cleavage of heme by the Pseudomonas aeruginosa HemO leads to the formation of BVIXβ and BVIXδ [9]. The altered regioselectivity was shown to be a consequence of a 90o in-plane rotation of the heme within the binding pocket, mediated by alternate interactions of the heme propionates with the protein scaffold [21]. However, as for all HO’s, a highly ordered network of water molecules within the heme-binding pocket is required for catalytic activity [20, 21, 36]. The structural water network is critical in providing the hydrogen bond network required for stabilizing the Fe(II)-O2 ligand, and in providing the proton relay in generating the activated Fe(III)-OOH intermediate [37, 38, 39]. Previous H/D-NMR studies of the HemO WT Fe(III)-CN, Fe(III)N3 [40] and HemO R80L Fe(III)-CN and Fe(II)-CO complexes [41] concluded the structural water network is integral to long-range communication and conformational flexibility in helices distant from the active site. The authors proposed such conformational plasticity is required for accommodating changes in axial ligand coordination and chemical intermediates during catalysis. Therefore, disruption of conformational flexibility and/or the structural water network offers an alternative drug-design strategy in addition to directly competing with heme binding at the active site, or by coordination to the heme iron [42, 43].

In keeping with this hypothesis, we have recently shown by HDX-MS that a series of iminoguanidine compounds inhibit HemO by binding to a newly discovered site on the back side of the protein rather than through competitive inhibition at the active site [44]. In these studies, Site-Identification by Ligand Competitive Saturation (SILCS) analysis [45] highlighted a binding site in close proximity to R188, which forms a salt bridge with D99 connecting helix XII with the loop connecting helices VI and VII on the back side of HemO [44]. Interestingly, these helices have been implicated in long-range motions required to accommodate changes in the distal ligand and chemical intermediates generated during catalysis. While computational and biophysical characterization of the ligand–protein interaction identified the binding site of the iminoguanidine inhibitors, the mechanism of HemO inhibition has not been elucidated [44]. We hypothesized based on the proximity of the iminoguanidine inhibitor binding site to the D99–R188 salt bridge that disruption of the salt bridge on binding may result in changes in conformational flexibility and/or a disruption of the structural water network. To further probe this hypothesis, we undertook a site-directed mutagenesis approach creating the D99A, R188A and D99A/R118A HemO mutant proteins. Biochemical and spectroscopic analysis of the D99 or R188 to Ala mutant proteins show that despite retention of the overall structural fold, the D99A and R188A mutants show significant changes in conformational flexibility within distal helices and those involved in direct hydrogen bonding interactions with the active site structural water network. Furthermore, the decrease in conformational dynamics of the active site helices in the D99 and R188 mutant proteins results in a complete loss of enzyme activity. The data are consistent with previous H/D-NMR studies that concluded long-range communication of the active site hydrogen bond network with helices distant from the active site is critical for conformational flexibility during catalysis [40, 41]. Furthermore, the current studies provide a plausible mechanism of action for the iminoguanidine inhibitors, and offer an attractive alternative approach to competitive inhibition of heme binding at the active site of HemO.

Materials and methods

Bacterial strains

Genetic engineering and plasmid replication were performed in Escherichia coli strain DH5α (F endA1 glnV44 thi-1 recA1 relA1 gyrA96 deoR nupG purB20 φ80dlacZ ΔM15 Δ(lacZYA-argF) U16, hsdR17(r K m K + ), λ). HemO WT and mutant protein expressions were performed in E. coli strain BL21 (DE3) (B F ompT gal dcm lon hsdSB(r B m B ) λ (DE3 [lacI lacUV-T7p07 ind1 sam7 nin5])) [malB+]K-12(λs)).

Site-directed mutagenesis

HemO mutants were constructed in the pET21a vector used previously to express HemO [9, 44]. Primers were synthesized by Sigma and mutations were introduced by PCR using the QuikChange II XL Site-Directed Mutagenesis Kit (Agilent Technologies) according to the manufacturer’s instructions. Primers used to introduce the D99A mutation were 5′-GTACCGGAGGGCGCGCAGAGCGTGCGCG-3′ and 5′-CGCGCACGCTCTGCGCGCCCTCCGGTAC-3′; primers used to introduce the R188A mutation were 5′-CGCCAGCGATGCCTTCAATGCTTTCGGCGACC-3′ and 5′-GGTCGCCGAAAGCATTGAAGGCATCGCTGGCG-3′; the underlined codons represent the introduced mutations. Isolated plasmids (pET21a) containing the mutations were transformed into E. coli Dh5α cells for replication and into BL21 (DE3) cells for overexpression. The BVIXα regioselective mutant of HemO (HemOα) (N19 K/K34A/F117Y/K132A) incorporating the D99A mutation, R188A mutation, or both were constructed in a similar manner using the same primers with the gene cloned into the pBAD33 vector [46].

Protein expression and purification

The HemO WT and the respective D99 and R188 variants were expressed and purified as previously described [9, 44]. Briefly, an overnight culture of transformed BL21 (DE3) cells grown in Luria–Bertani broth (LB) and treated with 30 µg/mL kanamycin was used to inoculate 4 1-L cultures at OD600 = 0.05. Cultures were treated with 30 µg/mL kanamycin and incubated at 37 °C with shaking until OD600 = 0.6. Expression was induced with 1 mM isopropyl-d-thiogalactopyranoside (IPTG) and cells continued shaking at 25 °C for 4 h. Cells were pelleted at 12,000g for 12 min and pellets were stored at − 80 °C until purification. For purification, pellets were thawed on ice and lysed in 50 mL lysis buffer (50 mM tris, pH = 8.0, 50 mM NaCl, 100 µM phenylmethanesulfonyl fluoride (PMSF), lysozyme, DNAse I, protease inhibitor cocktail (complete mini, Roche Diagnostics)) for 1 h. Lysate was sonicated then centrifuged at 21,000g at 4 °C for 1 h. The resulting supernatant was applied to a Ni–NTA column pre-equilibrated with 50 mM Tris (pH = 8.0), 300 mM NaCl, 2 mM imidazole. Column was washed with 50 mM Tris (pH = 8.0), 300 mM NaCl, 20 mM imidazole, and IFP1.4-His6 was eluted with 50 mM Tris (pH = 8.0), 300 mM NaCl, 250 mM imidazole in 10 mL fractions. Fractions were analyzed by SDS-PAGE and fractions containing IFP1.4-His6 were pooled and exhaustively dialyzed against 20 mM Tris (pH = 8.0). SDS-PAGE and UV–Vis analyses determined that protein was sufficiently pure and concentrated for use. Expressions typically yielded 1.8 mg protein per 1-L culture.

UV–Vis spectroscopy

Purified HemO (WT or mutants) was incubated with 1.5 equivalents of heme from freshly prepared hemin solution in 0.1 M NaOH brought up with 20 mM sodium citrate buffer (pH = 7.0). Excess hemin was removed by passage of the holo-HemO protein over a Q-Sepharose column (1.5 cm × 1.0 cm) and the holo-HemO eluted with 250 mM NaCl in 20 mM sodium citrate buffer (pH = 7.0). Holo-proteins were diluted to 10 µM (using ε280 = 13,980 M−1 cm−1 for apo-HemO) into 20 mM sodium citrate buffer (pH = 7.0) and absorption spectra were recorded on an Agilent Cary 300 UV–Vis spectrophotometer. Extinction coefficients were determined using the pyridine hemochrome method (ε418 = 170 mM−1 cm−1; ε525 = 17.5 mM−1 cm−1; ε555 = 34.4 mM−1 cm−1) [47]. For heme titration experiments, proteins were diluted to 15 µM in 20 mM sodium citrate buffer (pH = 7.0). Hemin was prepared as described above and added in 0.5 μM increments to final concentrations of 25 µM. Heme binding was analyzed by the increase in the Soret band (405 nm). The ferrous oxy (Fe(II)-O2) and CO (Fe(II)-CO) HemO complexes were obtained following saturation with CO(g) and addition of excess sodium dithionite. The absorption spectrum of the Fe(II)-CO HemO complex was obtained following dilution in CO(g) saturated 20 mM sodium citrate buffer (pH = 7.0). Fe(II)-O2 HemO complexes were generated by passage over a Sephadex G-25 desalting column (1.5 cm × 5 cm) to remove excess reductant. The Fe(II)-O2 HemO complexes were analyzed immediately following elution in 20 mM sodium citrate buffer (pH = 7.0). The decay and stability of the Fe(II)-O2 HemO D99A and R188A complexes were monitored by recording the spectra at 1 min intervals for 60 min. The rate of decay was analyzed by fitting the visible α-band at 541 nm to an exponential decay model and the half-life was determined using initial spectra as 100% oxy-complex and final spectra as 0% oxy-complex.

Circular dichroism

Purified apo- and holo-HemO (WT or mutants) was buffer exchanged into 50 mM sodium phosphate (pH = 7.2) and diluted to 2.5 µM (using ε280 = 13,980 M−1 cm−1 for apo-HemO) in the same buffer. Circular dichroism (CD) spectra were recorded on a JASCO J-810 spectropolarimeter using a JASCO PFD-425S temperature controller. Spectra were recorded at 25 °C from 180 nm to 260 nm in triplicate, and then spectra were averaged and smoothed. Each protein was analyzed three times. For determination of melting temperature (Tm), temperature was increased from 15 to 90 °C at a rate of 1 °C per min, and proteins were monitored at 222 nm. Tm was determined by analysis of linear range of unfolding and solved for 50% folded, using 15 °C as 100% folded and 90 °C as 0% folded.

Fluorescence spectrophotometry

Heme-binding affinities of the WT and mutant HemO mutants were determined by fluorescence quenching as previously described. Briefly, fluorescence experiments were performed on a Synergy H1 hybrid microplate reader using flat-bottom black 96-well plates. Protein concentration was 1 µM in 20 mM Tris–HCl (pH = 8.0). Freshly prepared heme was added to final concentrations of 0.05–100 µM. Solutions were excited at 295 nm and emission spectra were recorded from 300 to 400 nm. Decrease in maximum emission (332 nm) was fit to a one-site binding model as a function of increasing heme concentration.

Resonance Raman

RR spectra were obtained using a custom McPherson 2061/207 spectrograph (0.67 m with variable gratings) equipped with a Princeton Instruments liquid N2-cooled CCD detector (LN-1100 PB). Excitations at 407 and 442 nm were provided by a krypton laser (Innova 302, Coherent) and a helium–cadmium laser (Liconix 4240NB). Kaiser Optical edge filters were used to attenuate Rayleigh scattering. Spectra were collected at room temperature using a 90° scattering geometry on samples mounted on a reciprocating translation stage. Frequencies were calibrated relative to indene and CCl4 and are accurate to ± 1 cm−1; CCl4 was also used to check the polarization conditions. The integrity of the RR samples, before and after laser illumination, was confirmed by direct monitoring of their UV–Vis spectra in the Raman capillaries. All RR measurements were conducted on ~ 100 μM protein solutions in 20 mM Tris–HCl buffer (pH 8.0), 150 mM NaCl, and 10% glycerol. The reduced protein samples were prepared by addition of sodium dithionite (3 mM final concentration) under anaerobic conditions.

In vitro HemO enzyme activity measurements

Turnover of the holo-HemO WT or mutant proteins was performed in the presence of the NADPH-dependent ferredoxin reductase (pa-FPR) and NADPH as described previously [48]. Briefly, reactions were monitored spectrophotometrically in a 1-cm path length cuvette. The reactions contained 8–10 μM HemO, D99A or R188A mutants, 20 μM FPR in a final volume of 1 ml. Catalase (final concentration of 0.1 mg/ml) was used as a H2O2 scavenger. The reactions were initiated on addition of NADPH at a final concentration in the cuvette of 200 μM. Heme turnover was monitored between 300 and 800 nm with repeated scanning every 30 s for 30 min. To obtain the spectrum of iron free BVIX, the reaction was acidified with 3 N HCl (10 μl).

IFP1.4 in cell fluorescence activity

The fluorescence assays were performed as previously described with some minor modifications [44]. E. coli BL21 (DE3) were transformed with pBAD plasmid expressing either the HemOα or HemOα D99A, R188A, or D99/R188A mutants and pET28a expressing the bacterial phytochrome IFP1.4-His6. Protein expression was performed at 37 °C in LB media containing 25 µg/mL chloramphenicol and 30 µg/mL kanamycin to OD600 = 0.8. IFP1.4 was induced with IPTG (final concentration 1 mM) and cells expressed at 25 °C for 2 h. HemO (wild type or mutants) was then induced on addition of arabinose (final concentration 0.02%) and cells were aliquoted as 200 µL cultures in a black, clear-bottom 96-well plate. Cells were maintained at 25 °C with orbital shaking monitoring both OD600 and fluorescence emission on formation of the IFP1.4- BVIXα conjugate was then monitored at 700 nm following excitation at 630 nm. Fluorescence was monitored every 20 min over a 20 h period on a Synergy H1 hybrid microplate reader. Fluorescence change over time was corrected to account for differences in OD600 between samples. Negative controls were performed in the absence of arabinose.

Hydrogen–deuterium exchange mass spectrometry (HDX-MS)

Purified HemO mutants D99A and R188A were analyzed by hydrogen–deuterium exchange mass spectrometry as previously described for the wild-type HemO [44]. The heme concentration required for 95% saturation was estimated using a KD of 1 μM. Therefore, to account for heme exchange on dilution of the sample 100 Μm, apo-HemO was reconstituted with 114 μM heme during the sample preparation, and heme was added to the deuteration buffer to a final concentration of 21 μM. HemO protein samples were prepared by diluting the HemO protein to a final concentration of 100 μM with 20 mM HEPES buffer in H2O (pH 7.4). Deuterium exchange was initiated by diluting the HemO sample 50-fold with 20 mM HEPES buffer in D2O (pD 7.4, deuteration buffer) at 23 °C. 100 pmol of protein was removed from the reaction at 30 min, and the deuteration reaction immediately quenched by lowering the pH to 2.5 with ice-cold HCl. Quenched samples were frozen on dry ice prior to analysis. Following thawing, samples were immediately injected into a nanoACQUITY UPLC system with HDX manager (Waters). The protein was digested online at 10 °C using an Enzymate BEH Pepsin Column (2.1 × 30 mm, Waters) in 1 min. The digest was trapped and desalted online on an ACQUITY Vanguard BEH C18 pre-column (2.1 × 5 mm, Waters) at 0 °C for 4 min at a flow rate of 125 μl/min in 0.1% formic acid. Peptides were separated on an ACQUITY UPLC BEH C18 column (1.7 μm, 1 × 100 mm, Waters) at 0 °C by a 15 min linear acetonitrile gradient (5–50%) with 0.1% formic acid at a flow rate of 40 μl/min. The eluent was directed into the ion source of a coupled SYNAPT G2 HDMS mass spectrometer (Waters). Mass spectra were acquired in the MSE mode over the m/z range of 50–2000. Mass spectrometer parameters were as follows: electrospray ionization positive (ESI +) mode, capillary voltage 3 kV, collision energy 20–30 eV, sampling cone voltage 30 V, source temperature 80 °C, desolvation temperature 175 °C and desolvation gas flow 500 L/h. To generate a peptide list for ion search, 100 pmol of undeuterated protein in 2 mM HCl in H2O was injected. The undeuterated peptides were identified using Waters ProteinLynx Global Server software. The peptide list generated was imported into Waters DynamX software to guide the search of deuterated peptides, and the relative deuterium incorporation levels for each deuterated peptide were calculated using the time zero sample as reference.

Results and discussion

Spectroscopic characterization of the holo-HemO D99A, R188A and D99A/R188A complexes

HemO WT and the D99A, R188A andD99A/R188A mutants following purification and reconstitution with heme were analyzed by absorption and CD spectrometry. The CD spectra of the D99A, R188A and D99/R188A holo-HemO mutants show a similar overall helical fold as for the WT HemO (Fig. 2a). The melting temperature (Tm) for the apo- and holo-HemO mutant proteins as monitored by CD shows a slight decrease in stability compared to the WT (Table 1), presumably due to loss of the D99–R188 salt bridge between helix VII and helix XII (Fig. 1b). However, the D99A, R188A and D99/R188A HemO proteins bind heme with similar affinities to the WT (Table 1). Furthermore, the resting state ferric (Fe3+) complexes of the HemO mutants show similar overall spectra, with Soret bands at 404 nm compared with 405 nm for the WT, and visible bands consistent with a six-coordinate high-spin heme (Fig. 2b).
Fig. 2

CD and absorption spectra of the HemO WT and D99A, R188A and D99A/R188A mutant proteins. a CD spectra of the apo-HemO proteins. b Absorption spectra of the holo-HemO proteins. c CD spectra of the holo-HemO proteins. CD spectra were recorded with 2.5 μM apo- or holo-HemO proteins in 50 mM potassium phosphate buffer (pH = 7.2) and absorption spectra in 20 mM sodium citrate (pH 7.0) at 25 °C as described in the “Materials and Methods

Table 1

Heme-binding constants and melting temperatures of HemO WT and D99A, R188A and D99A/R188A apo- and holo-HemO proteins. aStandard deviation, *p < 0.05 vs WT, **p < 0.005 vs WT

HemO mutant

Heme-binding affinity KD (μM)

Apo-HemO Tm (°C)

Holo-HemO Tm (°C)


0.87 ± 0.17a

56.6 ± 1.2a

63.6 ± 1.7


0.83 ± 0.21

53.2 ± 1.1*

59.7 ± 1.2*


0.79 ± 0.22

49.6 ± 1.0**

54.5 ± 1.9**


0.99 ± 0.21

51.7 ± 0.6**

60.2 ± 0.8*

To characterize further the coordination and spin states of the D99–R188 salt bridge mutants, high-frequency RR spectra of the ferric and ferrous proteins were obtained using Soret excitation (Fig. 3). These spectral data are remarkably similar to those of the WT protein, with oxidation marker band ν4 at 1376 and 1354 cm−1 in the ferric and ferrous proteins, respectively. In the ferric proteins, the ν3 and ν2 spin-state marker bands show two sets of frequencies that are indicative of a six-coordinate low-spin/high-spin mixtures consistent with occupancy of the distal coordination site by an aqua/hydroxy ligand. After reduction with excess dithionite, all proteins adopt a pure five-coordinate high-spin configuration with unique ν3 and ν10 modes at 1469 and 1602 cm−1. These high-frequency RR spectra also identify the vinyl groups stretching vibrations νC=C, which are observed at 1624 and 1620 cm−1 in the ferric and ferrous proteins, respectively. The characteristics of the proximal histidine coordination to the Fe(II)-deoxy heme were also monitored by the acquisition of low-frequency RR spectra with a 442 nm excitation known to specifically enhance the FeII-NHis stretching mode in high-spin heme proteins. All spectra show an intense band at 220 cm−1 readily assigned to the ν(FeII-NHis) mode (Fig. 4). Deformation modes from the peripheral propionate and vinyl modes observed between 350 and 440 cm−1 are also unchanged by the salt-bridge mutations (Fig. 4). Taken together, the spectroscopic data indicate that mutation of D99 and/or R188 has no effect on the overall structure or heme-binding properties of the protein.
Fig. 3

High-frequency region of the RR spectra of HemO WT and D99A, R188A and D99A/R188A mutant proteins in the ferric (left panel) and ferrous (right panel) states obtained with a 407 nm excitation. Spectra are listed from the upper spectrum for HemO WT, followed by D99A, R188A and D99A/R188A

Fig. 4

Low-frequency region of the RR spectra of HemO WT and D99A, R188A and D99A/R188A mutant ferrous proteins obtained with a 442 nm excitation

Stability of the reduced Fe(II)-O2 holo-HemO D99A, R188A and D99A/R188A complexes

The HemO wild-type Fe(II)-CO and Fe(II)-O2 complexes on reduction with sodium dithionite were isolated as previously reported [9], and the respective Soret and visible bands are listed in Table 2. HemO mutants D99A, R188A, and D99A/R188A all form Fe(II)-CO and Fe(II)-O2 complexes, albeit with slightly altered Soret and visible bands. The Fe(II)-O2 complexes of all three mutants showed blue-shifted Soret bands compared to the WT protein, suggesting that the distal O2 ligand is experiencing minimal perturbation of the heme environment as a result of the mutations. The visible bands of the Fe(II)-CO and Fe(II)-O2 complexes of the D99 and R188 mutants are also slightly shifted consistent with no significant changes in heme coordination or spin state.
Table 2

Spectroscopic properties of the holo-HemO WT, D99A, R188A and (Fe(III) and D99A/R188A resting state Fe(II) CO- and O2-complexes

Protein complex

Soret band (nm)

Q-bands (nm)

Q-band ratio













540, 570




540, 570




539, 570




539, 570





541, 577




541, 576




541, 576




540, 575


However, despite the fact the overall structural fold and heme environment of the D99A and R188A and D99/R188A mutants are similar to WT HemO, the stability of the Fe(II)-O2 complexes in all three mutants was slightly decreased. Following generation of the Fe(II)-O2 complexes and removal of excess reductant the decay of the Fe(II)-O2 complex was monitored. In the absence of reductant, over time the Soret and visible bands shift back to the Fe(III) resting state. However, the rate of decay of the WT HemO Fe(II)-O2 complex (12.6 min) (Fig. 5a) was modestly slower than the decay observed for the three mutants, with half-lives of 7.6 min, 6.3 min, and 5.2 min for D99A, R188A, and D99A/R188A, respectively (Fig. 5b–d). The increased rate of decay of the Fe(II)-O2 holo-HemO mutant complexes is indicative of a decrease in the stability of the bound O2 ligand. Previous NMR studies have shown structural elements surrounding the heme pocket undergo changes in conformational freedom as a function of the ligand allowing flexibility of the active site hydrogen bonding network to accommodate changes in the ligand [41]. As we observe no apparent changes per se in the heme-binding environment of the D99A, R188A and D99/R188A mutants based on the RR, we hypothesized the decreased stabilization of the Fe(II)-O2 ligand may be result of changes in the conformational flexibility of the protein.
Fig. 5

Decay of the Fe(II) O2-complexes for HemO wild-type and mutant proteins. Visible bands of the Fe(II)-O2 HemO complexes a WT, b D99A, c R188A, and d D99A/R188A mutants were observed for 1 h following complex formation. The decrease in absorbance of both the α-band (577 nm) and β-band (541 nm) of the Fe(II)-O2 holo-HemO complexes was followed by absorption spectroscopy (left panel). The rate of decay at 577 and 541 nm was plotted for each protein (right panel). Samples were prepared as described in “Materials and Methods

Structural analysis of HemO D99A and R188A mutants by hydrogen–deuterium exchange mass spectrometry (HDX-MS)

To determine the effect of disrupting the D99–R188 salt bridge on protein conformational flexibility, we performed HDX-MS on the holo-HemO D99A and R188A mutants. The peptides obtained following deuterium exchange and pepsin digestion showed greater than 95% coverage for the HemO D99 and R188 proteins (Fig S1) as previously reported for HemO WT [44]. Analysis of the HDX at 30 min showed that in contrast to the wild-type HemO, both D99A and R188A mutants show a decreased deuterium incorporation in the proximal (I and II) and distal helices (VII and VIII) of the heme-binding pocket (Figs. 6, 7). Specifically, in the D99A HemO mutant residues, 12–18 of proximal helix I and residues 117–128 of distal helices VII and VIII showed decreased exchange (Fig. 6a, b). Similarly, the R188A HemO mutant shows decreased exchange in proximal helices I and II (residues 12–43) and in distal helices VII–XI (residues 117–161) (Fig. 7a, b). Interestingly, the pattern of decreased flexibility in the D99 and R188 HemO mutant proteins map to the same regions reported to show the greatest flexibility by H/D-NMR when accommodating changes in the distal ligand during catalysis [40, 41]. On the backside of HemO the loop between helices VI and VII which contains D99 also shows decreased exchange in both mutants (as evidenced by residues 67–109 and 69–107 for D99A and R188A, respectively). In contrast to the decreased exchange in the proximal and distal helices, an increase in deuterium incorporation is seen in helices III (residues 55–62 and 54–62 of D99A and R188A, respectively) and XII (residues 182–192 and 175–192 in D99A and R188A, respectively). Overall, the global reduction in deuterium incorporation in helices around the distal pocket is more pronounced in the R188A mutant. However, a similar pattern of decreased conformational flexibility is observed in both holo-HemO D99A and R188A in helices VI, VII and VIII and increased flexibility in helix III. Interestingly, these helices all contribute at least one direct hydrogen bond to a structural water molecule within the active site.
Fig. 6

HDX-MS of the resting state Fe(III) holo-HemO D99A. Deuterium incorporation was compared to HemO wild-type and region showing the greatest change in deuterium uptake are color coded and mapped on the HemO crystal structure (PDB 1SK7). Peptide regions with increased deuterium incorporation are colored in red, while decreased exchange is colored in blue. a Holo-HemO viewed from the front with the proximal H26, D99 and R188 shown in stick format. b Holo-HemO viewed from the back. c Relative deuterium uptake for helices showing the most significant changes in deuterium uptake at 30 min (Student’s t test p, < 0.05, n = 3). Samples were prepared and analyzed as described in the “Materials and Methods

Fig. 7

HDX-MS of the resting state Fe(III) holo-HemO R188A. Deuterium incorporation was compared to HemO wild-type and region showing the greatest change in deuterium uptake are color coded and mapped on the HemO crystal structure (PDB 1SK7). Peptide regions with increased deuterium incorporation are colored in red, while decreased exchange is colored in blue. a Holo-HemO viewed from the front with the proximal H26, D99 and R188 shown in stick format. b Holo-HemO viewed from the back. c Relative deuterium uptake for helices showing the most significant changes in deuterium uptake at 30 min (Student’s t test p, < 0.05, n = 3). Samples were prepared and analyzed as described in the “Materials and Methods

Enzymatic activity of the HemO D99A, R188A, and D99A/R188A mutants

The rate limiting step in heme degradation by HO is the release of product. In mammalian HO’s this is circumvented by coupling BVIXα reduction to bilirubin IXα (BRIXα) through biliverdin reductase [2, 6]. The turnover of the bacterial holo-HemO complex results in Fe(III)-BVIX bound in the active site, which in the absence of further reduction of the Fe(III) or the presence of a BVIX-binding protein does not allow for multiple enzyme turnover in vitro [9, 10, 11]. Furthermore, the final in vitro product of heme degradation by the bacterial HemO enzymes is not released to the eukaryotic BVIX reductase enzymes.

Compounding the issue of measuring enzymatic heme degradation in vitro is the use of chemical reductants such as ascorbate as the electron donor, which can result in reduction of O2 to non-coordinated H2O2 leading to non-enzymatic coupled oxidation of heme [49, 50]. Therefore, the D99A, R188A and D99A/R188A mutants in which the Fe(II)-O2 intermediate is destabilized may lead to the rapid formation of H2O2 and non-enzymatic heme degradation. To circumvent this complication, we performed in vitro single turnover enzyme assays utilizing the HemO-specific redox partner NADPH ferredoxin reductase (FPR) in the presence of catalase to consume H2O2 produced as a consequence of enzymatic uncoupling of FPR and HemO in vitro [48]. As shown in Fig. 8a, initiation of the reaction with NADPH results in a shift in the Soret from 406 to 413 nm with a concomitant increase in bands at 541 and 577 nm on formation of the Fe(II)-O2 heme complex (Fig. 8a; inset). The formation of the Fe(III)-BVIX complex is indicated by the decrease in the visible bands over time and a shift of the Soret peak back to 406 nm. Acidification of the reaction to release the iron from the BVIX complex results in broad bands at ~ 680 nm and 380 nm typical of BVIX. Interestingly, the D99A and R188A mutants on initiation of the reaction with NADPH result in a shift in the Soret toward 411 nm and a subsequent decay in the intensity of the Soret band (Fig. 8b, c). However, in contrast to the turnover of the holo-HemO WT complex, we do not observe formation of a stable Fe(II)-O2 complex as evident by the reduced intensity of the visible bands at 541 and 576 nm (Fig. 8b, c; inset). This is consistent with the previous data where the holo-HemO D99A and R188A Fe(II)-O2 complexes auto-oxidized more rapidly than the HemO WT Fe(II)-O2 complex (Fig. 5). Despite the presence of catalase in the assays for both the D99A and R188A holo-HemO complexes, we observe a decrease in the intensity of Soret and visible bands over time consistent with heme degradation (Fig. 8b, c). However, on acidification of the D99A and R188A holo-HemO reactions rather than the broad featureless band around 680 nm consistent with formation of biliverdin, we noted a peak at 660 nm more indicative of verdoheme [51]. This is consistent with the fact that in contrast to HemO WT the HemO D99A, R188A and D99A/R188A mutants in E. coli do not give rise to the green pigmented cell pellets indicative of BVIXβ and BVIXδ (Fig. 9a). Therefore, we attribute the degradation of heme in the reconstituted in vitro assay to the autoxidation of the unstable Fe(II)-O2 complex generating H2O2 within the heme active site leading to coupled oxidation of heme to verdoheme. To determine if heme degradation in the D99A and R188A mutants is a consequence of non-specific coupled oxidation of heme to verdoheme, we employed a cell-based assay recently developed in our laboratory that takes advantage of the fact expression of HemO in E. coli leads to the production of endogenous BVIXα, that can be coupled to a fluorescence readout on co-expression of the BVIXα-dependent bacterial phytochrome IFP1.4 [44, 52, 53, 54]. As the in vivo assay relies on the increased fluorescence properties of the natural IFP1.4 ligand BVIXα, this required utilizing the previously characterized HemOα variant that switch the regioselectivity to yield BVIXα through an in-plane rotation of the heme [55]. The in-plane rotation placing the α-meso-carbon in position for oxidative cleavage does not alter the enzymatic activity of the HemOα variant compared to HemO WT [55]. Similar to the HemO WT, the expression of the HemOα variant in E. coli gives rise to green pigmented cells as a consequence of BVIXα accumulation, whereas the corresponding HemOα D99 and R188 mutants are colorless (Fig. 9a). We further confirmed the inactivity of the HemOα D99 and R188 mutants in the previously developed in cell E. coli assay expressing both HemOα and the BVIXα-dependent bacterial phytochrome IFP1.4 [44]. As expected for the HemOα variant we observe an increase in fluorescence over time as the IFP1.4-BVIXα chromophore accumulates in the bacterial cells (Fig. 9b). However, the D99A, R188A, or D99A/R188A mutants in the HemOα background show no increase in fluorescence over time (Fig. 9b). Taken together the data suggest the decreased conformational dynamics of the D99A, R188A and D99A/R188A HemO mutants does not allow for the accommodation of changes in ligand coordination and chemical structure during catalysis. Furthermore, in the in vitro activity assay, the decrease in stability of the Fe(II)-O2 is sufficient to lead to non-enzymatic heme degradation as a result of the generation of H2O2 within the active site.
Fig. 8

In vitro turnover of a holo-HemO WT, b D99A, and c R188A proteins. Spectral changes on addition of 200 μM NADPH to a solution containing HemO (8–10 μM), FPR (20 μM) in 20 mM Tris–HCl pH 7.4. The inset (×10) shows the increase in intensity of the visible bands on formation of the Fe(II)-O2 complex which subsequently decreases on formation of the Fe(III)-BVIX complex. Following completion of the assay acidification with 3 N HCl releases BVIX (dashed line) as judged by the band at 680 nm (a) or a spectrum more typical of verdoheme with a peak at 655 nm (b, c). See “Materials and Methods” for more details

Fig. 9

a Cell pellets of HemO WT and HemOα and their corresponding D99A, R188A, D99A/R188A mutants harvested at 6 h post-induction with 1 mM IPTG. The HemO WT cell pellet shows less intense green pigmentation than the HemOα variant as BVIXβ and BVIXδ are more readily excreted into the extracellular medium than BVIXα. b HemO activity as monitored by fluorescence of the IFP1.4-BVIXα chromophore in cells expressing HemOα, or the corresponding D99A, R188A and D99A/R188A mutants. Following induction of HemOα and IFP1.4 cultures were incubated at 25 °C with shaking and the OD600 and fluorescence of the IFP1.4-BVIXα complex (ex. 630 nm, em. 700 nm) was monitored over 20 h on a Synergy H1 hybrid microplate reader as described in the “Materials and Methods


The P. aeruginosa HemO enzyme is essential for heme utilization within the host, and together with the emerging role of BVIXβ in adaptation and virulence, represents a novel drug target [55, 56, 57]. We recently identified of a series of iminoguanidine inhibitors that bind on the back side of HemO close to a salt bridge formed between D99 and R188. However, the mechanism of enzyme inhibition by the iminoguanidine ligands is not known. Interestingly, D99 is within a loop connecting helix VI to distal helix VII which forms the heme-binding site (Fig. 1a). Furthermore, distal helix VII is critical for maintenance of the structural integrity of the active water network required for catalysis [21]. We previously hypothesized the mechanism of inhibition may be a result of disrupting the D99–R188 salt bridge, leading to long-range allosteric effects on distal helix VII. To further probe if the iminoguanidine inhibitors are potentially disrupting the D99–R188 salt bridge we constructed the D99A, R188A and D99/R188A site-directed mutants. Analysis of the secondary structure of the apo and holo-HemO D99 and R188 proteins by CD shows no significant differences from the WT (Fig. 2a) [55]. As previously reported, the secondary structure of the HemOα variant is indistinguishable from that of the WT [55]. Analysis of the D99A, R188A, and D99A/R188A mutants by CD also showed similar profiles to the HemOα variant (data not shown). Furthermore, the heme-binding affinities and absorption spectra of the resting state holo-HemO mutant proteins are indistinguishable from that of the WT HemO (Fig. 2b and Table 1). A more detailed analysis of the heme environment of the holo-HemO D99A, R188A and D99/R188A mutants by RR indicated that like the resting state holo-HemO WT, the heme is six-coordinate high-spin with a water molecule as the sixth ligand to the heme. The conserved RR spectra in the resting state Fe(III) and reduced Fe(II)-deoxy D99A, R188A and D99/R188A HemO mutant complexes are inconsistent with a significant disruption of the active site structure. Furthermore, the rate of decay of the Fe(II)-O2 complex for the HemO D99A, R188A and D99/R188A mutants was only slightly increased over that of WT. However, despite the fact the heme environment and the stability of the initial Fe(II)-O2 complex were not significantly affected in the HemO D99A, R188A and D99/R188A mutants, they are catalytically inactive (Figs. 8 and 9).

Interestingly, previous HDX-MS analysis of HemO in the presence of iminoguanidine inhibitors shown to bind in close proximity to the D99–R188 salt bridge induced significant changes in conformational flexibility within structural elements surrounding the active site [44]. Similarly, our current HDX analysis on mutation of either D99 or R188 shows similar changes in conformational dynamics within the same structural elements as observed with the inhibitors (Figs. 6 and 7). Specifically, we see increased conformational flexibility in helix III and XII and decreased disorder in helices VII and VIII that surround the heme active site. Our data are also consistent with previous HDX-NMR studies of the P. aeruginosa HemO WT protein in which the authors concluded the conformational dynamics of the protein are critical in allowing the changes in the coordinated ligand as well as the electronic and structural changes to the heme during catalytic turnover [40, 41]. Specifically, the authors show the Fe(III)-N3 ligand as a mimic of the activated Fe(III)-OOH intermediate induced chemical shift perturbations throughout structural elements surrounding the heme, with the greatest changes being observed within proximal helix II, distal helix VII, as well long-range effects in helices V, VI, X and the loop preceding helix X (Fig. 1A). Subsequent studies on the HemO R80L mutant which disrupts the structural water network also showed an overall increase in HD exchange and conformational dynamics in these same structural elements [41]. Interestingly, the structural elements shown to undergo conformational flexibility in accommodating ligand changes by H/D-NMR in HemO are those also shown to be affected in the current HDX-MS analysis of the salt-bridge mutants (particularly those in proximal helix II, distal helix V, VI and VII) (Figs. 6 and 7). Our current data when taken in the context of previous studies suggest the D99–R188 salt bridge is critical in modulating the conformational dynamics and long-range effects on the heme active site required for catalysis. Furthermore, the current studies provide a rationale for the mechanism of inhibition by the previously characterized allosteric iminoguanidine inhibitors [44]. We propose that the allosteric HemO inhibitors that target protein conformation and dynamics rather than competitive inhibition at the heme active site offer an alternative and complementary approach to the design of novel antimicrobial agents targeting P. aeruginosa.



The authors would like to thank Bennett Giardina for technical advice and assistance with the IFP in cell activity assays.

Author contributions

GH generated, purified and characterized the D99 and R188 mutants in the WT and HemOα background. WH performed all of the HDX-MS experiments. ER performed the in vitro activity assays. PML performed and interpreted the resonance Raman experiments. GH, WH, AW and PML wrote the manuscript. All authors contributed to final editing of the manuscript and have given approval to the final version of the manuscript.


This research was funded in part by pre-doctoral fellowships from the ACS Division of Medicinal Chemistry and the American Foundation for Pharmaceutical Education to Geoffrey Heinzl; NIH Grant T32GM066706; and NIH Grant AI102883 to Angela Wilks.

Supplementary material

775_2018_1609_MOESM1_ESM.pdf (200 kb)
Supplementary material 1 (PDF 200 kb)


  1. 1.
    Wilks A, Black SM, Miller WL, Ortiz de Montellano PR (1995) Biochemistry 34:4421–4427CrossRefGoogle Scholar
  2. 2.
    Wilks A, Ortiz de Montellano PR (1993) J Biol Chem 268:22357–22362PubMedGoogle Scholar
  3. 3.
    Yoshida T, Kikuchi G (1977) J Biochem (Tokyo) 81:265–268CrossRefGoogle Scholar
  4. 4.
    Yoshida T, Kikuchi G (1979) J Biol Chem 254:4487–4491PubMedGoogle Scholar
  5. 5.
    Yoshida T, Takahashi S, Kikuchi G (1974) J Biochem (Tokyo) 75:1187–1191CrossRefGoogle Scholar
  6. 6.
    Yoshinaga T, Sassa S, Kappas A (1982) J Biol Chem 257:7778–7785PubMedGoogle Scholar
  7. 7.
    Beale SI, Cornejo J (1984) Arch Biochem Biophys 235:371–384CrossRefGoogle Scholar
  8. 8.
    Cornejo J, Willows RD, Beale SI (1998) Plant J Cell Mol Biol 15:99–107CrossRefGoogle Scholar
  9. 9.
    Ratliff M, Zhu W, Deshmukh R, Wilks A, Stojiljkovic I (2001) J Bacteriol 183:6394–6403CrossRefGoogle Scholar
  10. 10.
    Wilks A, Schmitt MP (1998) J Biol Chem 273:837–841CrossRefGoogle Scholar
  11. 11.
    Zhu W, Wilks A, Stojiljkovic I (2000) J Bacteriol 182:6783–6790CrossRefGoogle Scholar
  12. 12.
    Wegele R, Tasler R, Zeng Y, Rivera M, Frankenberg-Dinkel N (2004) J Biol Chem 279:45791–45802CrossRefGoogle Scholar
  13. 13.
    Poss KD, Tonegawa S (1997) Proc Natl Acad Sci USA 94:10919–10924CrossRefGoogle Scholar
  14. 14.
    Poss KD, Tonegawa S (1997) Proc Natl Acad Sci USA 94:10925–10930CrossRefGoogle Scholar
  15. 15.
    Dennery PA (2014) Antioxid Redox Signal 20:1743–1753CrossRefGoogle Scholar
  16. 16.
    Davis SJ, Vener AV, Vierstra RD (1999) Science 286:2517–2520CrossRefGoogle Scholar
  17. 17.
    Matsui T, Nambu S, Ono Y, Goulding CW, Tsumoto K, Ikeda-Saito M (2013) Biochemistry 52:3025–3027CrossRefGoogle Scholar
  18. 18.
    Nambu S, Matsui T, Goulding CW, Takahashi S, Ikeda-Saito M (2013) J Biol Chem 288:10101–10109CrossRefGoogle Scholar
  19. 19.
    Skaar EP, Gaspar AH, Schneewind O (2004) J Biol Chem 279:436–443CrossRefGoogle Scholar
  20. 20.
    Schuller DJ, Zhu W, Stojiljkovic I, Wilks A, Poulos TL (2001) Biochemistry 40:11552–11558CrossRefGoogle Scholar
  21. 21.
    Friedman J, Lad L, Li H, Wilks A, Poulos TL (2004) Biochemistry 43:5239–5245CrossRefGoogle Scholar
  22. 22.
    Davydov R, Kofman V, Fujii H, Yoshida T, Ikeda-Saito M, Hoffman BM (2002) J Am Chem Soc 124:1798–1808CrossRefGoogle Scholar
  23. 23.
    Davydov R, Matsui T, Fujii H, Ikeda-Saito M, Hoffman BM (2003) J Am Chem Soc 125:16208–16209CrossRefGoogle Scholar
  24. 24.
    Davydov RM, Yoshida T, Ikeda-Saito M, Hoffman BM (1999) J Am Chem Soc 121:10656–10657CrossRefGoogle Scholar
  25. 25.
    Hernandez G, Wilks A, Paolesse R, Smith KM, Ortiz de Montellano PR, La Mar GN (1994) Biochemistry 33:6631–6641CrossRefGoogle Scholar
  26. 26.
    Liu Y, Moenne-Loccoz P, Loehr TM, Ortiz de Montellano PR (1997) J Biol Chem 272:6909–6917CrossRefGoogle Scholar
  27. 27.
    Liu Y, Ortiz de Montellano PR (2000) J Biol Chem 275:5297–5307CrossRefGoogle Scholar
  28. 28.
    Sun J, Wilks A, Ortiz de Montellano PR, Loehr TM (1993) Biochemistry 32:14151–14157CrossRefGoogle Scholar
  29. 29.
    Takahashi S, Ishikawa K, Takeuchi E, Ikeda-Saito M, Yoshida T, Rousseau DL (1995) J Am Chem Soc 117:6002–6006CrossRefGoogle Scholar
  30. 30.
    Takahashi S, Matera KM, Fujii H, Zhou H, Ishikawa K, Yoshida T, Ikeda-Saito M, Rousseau DL (1997) Biochemistry 36:1402–1410CrossRefGoogle Scholar
  31. 31.
    Wilks A, Torpey J, Ortiz de Montellano PR (1994) J Biol Chem 269:29553–29556PubMedGoogle Scholar
  32. 32.
    Unno M, Matsui T, Ikeda-Saito M (2012) J Inorg Biochem 113:102–109CrossRefGoogle Scholar
  33. 33.
    Lai W, Chen H, Matsui T, Omori K, Unno M, Ikeda-Saito M, Shaik S (2010) J Am Chem Soc 132:12960–12970CrossRefGoogle Scholar
  34. 34.
    Lad L, Ortiz de Montellano PR, Poulos TL (2004) J Inorg Biochem 98:1686–1695CrossRefGoogle Scholar
  35. 35.
    Lad L, Friedman J, Li H, Bhaskar B, Ortiz de Montellano PR, Poulos TL (2004) Biochemistry 43:3793–3801CrossRefGoogle Scholar
  36. 36.
    Schuller DJ, Wilks A, Ortiz de Montellano PR, Poulos TL (1999) Nat Struct Biol 6:860–867CrossRefGoogle Scholar
  37. 37.
    Matsui T, Furukawa M, Unno M, Tomita T, Ikeda-Saito M (2004) J Biol Chem 280:2981–2989CrossRefGoogle Scholar
  38. 38.
    Liu Y, Lightning LK, Huang H, Moenne-Loccoz P, Schuller DJ, Poulos TL, Loehr TM, de Montellano PRO (2000) J Biol Chem 275:34501–34507CrossRefGoogle Scholar
  39. 39.
    Lightning LK, Huang H, Moenne-Loccoz P, Loehr TM, Schuller DJ, Poulos TL, de Montellano PR (2001) J Biol Chem 276:10612–10619CrossRefGoogle Scholar
  40. 40.
    Rodriguez JC, Wilks A, Rivera M (2006) Biochemistry 45:4578–4592CrossRefGoogle Scholar
  41. 41.
    Rodriguez JC, Zeng Y, Wilks A, Rivera M (2007) J Am Chem Soc 129:11730–11742CrossRefGoogle Scholar
  42. 42.
    Hom K, Heinzl GA, Eakanunkul S, Lopes PE, Xue F, Mackerell AD Jr, Wilks A (2013) J Med Chem 56:2097-heCrossRefGoogle Scholar
  43. 43.
    Furci LM, Lopes P, Eakanunkul S, Zhong S, MacKerell AD Jr, Wilks A (2007) J Med Chem 50:3804–3813CrossRefGoogle Scholar
  44. 44.
    Heinzl GA, Huang W, Yu W, Giardina BJ, Zhou Y, MacKerell AD Jr, Wilks A, Xue F (2016) J Med Chem 59:6929–6942CrossRefGoogle Scholar
  45. 45.
    Raman EP, Yu W, Lakkaraju SK, MacKerell AD Jr (2013) J Chem Inf Model 53:3384–3398CrossRefGoogle Scholar
  46. 46.
    Guzman LM, Belin D, Carson MJ, Beckwith J (1995) J Bacteriol 177:4121–4130CrossRefGoogle Scholar
  47. 47.
    Fuhrop JH, Smith KM (eds) (1975) Porphyrins and metalloporphyrins. Elsevier, Amsterdam, pp 804–807Google Scholar
  48. 48.
    Wang A, Zeng Y, Han H, Weeratunga S, Morgan BN, Moenne-Loccoz P, Schonbrunn E, Rivera M (2007) Biochemistry 46:12198–12211CrossRefGoogle Scholar
  49. 49.
    Avila L, Huang HW, Damaso CO, Lu S, Moenne-Loccoz P, Rivera M (2003) J Am Chem Soc 125:4103–4110CrossRefGoogle Scholar
  50. 50.
    Rodriguez JC, Rivera M (1998) Biochemistry 37:13082–13090CrossRefGoogle Scholar
  51. 51.
    Damaso CO, Bunce RA, Barybin MV, Wilks A, Rivera M (2005) J Am Chem Soc 127:17582–17583CrossRefGoogle Scholar
  52. 52.
    Shu X, Royant A, Lin MZ, Aguilera TA, Lev-Ram V, Steinbach PA, Tsien RY (2009) Science 324:804–807CrossRefGoogle Scholar
  53. 53.
    Filonov GS, Piatkevich KD, Ting LM, Zhang J, Kim K, Verkhusha VV (2011) Nat Biotechnol 29:757–761CrossRefGoogle Scholar
  54. 54.
    Sigala PA, Crowley JR, Hsieh S, Henderson JP, Goldberg DE (2012) J Biol Chem 287:37793–37807CrossRefGoogle Scholar
  55. 55.
    Mourino S, Giardina BJ, Reyes-Caballero H, Wilks A (2016) J Biol Chem 291:20503–20515CrossRefGoogle Scholar
  56. 56.
    Barker KD, Barkovits K, Wilks A (2012) J Biol Chem 287:18342–18350CrossRefGoogle Scholar
  57. 57.
    O’Neill MJ, Wilks A (2013) ACS Chem Biol 8:1794–1802CrossRefGoogle Scholar

Copyright information

© SBIC 2018

Authors and Affiliations

  1. 1.Department of Pharmaceutical Sciences, School of PharmacyUniversity of Maryland, BaltimoreBaltimoreUSA
  2. 2.Division of Environmental and Biomolecular Systems, School of MedicineOregon Health and Science UniversityPortlandUSA
  3. 3.Laboratory of Applied Biochemistry, Division of Biotechnology Products Research and Review IIIOffice of Biotechnology Products, Office of Pharmaceutical Quality, Center for Drug Evaluation and Research, Food and Drug AdministrationSilver SpringUSA

Personalised recommendations