Nanoelectrospray Ionization Mass Spectrometric Study of

May 16, 2013 - Nanoelectrospray Ionization Mass Spectrometric Study of Mycobacterium tuberculosis CYP121–Ligand Interactions ... E-mail: [email protected]...
0 downloads 8 Views 2MB Size
Download PDF
Article pubs.acs.org/ac

Nanoelectrospray Ionization Mass Spectrometric Study of Mycobacterium tuberculosis CYP121−Ligand Interactions Katie M. Duffell,† Sean A. Hudson,† Kirsty J. McLean,‡ Andrew W. Munro,‡ Chris Abell,† and Dijana Matak-Vinković*,† †

Department of Chemistry, University of Cambridge, Lensfield Road, Cambridge CB2 1EW, United Kingdom Manchester Institute of Biotechnology, Faculty of Life Sciences, University of Manchester, 131 Princess Street, Manchester M1 7DN, United Kingdom

‡

S Supporting Information *

ABSTRACT: Nondenaturing nanoelectrospray ionization mass spectrometry (nanoESI MS) of intact protein complexes was used to study CYP121, one of the 20 cytochrome P450s in Mycobacterium tuberculosis (Mtb) and an enzyme that is essential for bacterial viability. The results shed new light on both ligand-free and ligand-bound states of CYP121. Isolated unbound CYP121 is a predominantly dimeric protein, with a minor monomeric form present. High affinity azoles cause the dissociation of dimeric CYP121 into monomer, whereas weaker azole binders induce partial dimer dissociation or do not significantly destabilize the dimer. Complexes of CYP121 with azoles were poorly detected by nanoESI MS, indicating kinetically labile complexes that are easily prone to gas-phase dissociation. Unlike with the azoles, CYP121 forms a stable complex with its natural substrate cYY that does not undergo gasphase dissociation. In addition, a series of potential ligands from fragment-based studies were used as a test for nanoESI MS work against CYP121. Most of these ligands formed stable complexes with CYP121, and their binding did not promote dimer dissociation. On the basis of binding to the monomer and/or CYP121 dimer it was possible to determine the relative order of their CYP121 binding affinities. The top nanoESI MS screening hit was confirmed by heme absorbance shift assay to have a Kd of 40 μM.

T

he human pathogen Mycobacterium tuberculosis (Mtb)1 causes tuberculosis (TB), a chronic, infectious disease responsible for major worldwide human mortality, and one which presents a huge challenge for drug discovery and development.2 To date there has been limited success in developing new drugs, for which there is a real need as a consequence of (i) the emergence of drug and multidrugresistant Mtb strains,3 and (ii) the synergy of TB with the human immune deficiency virus (HIV). The best characterized Mtb strain, H37Rv,4 has 20 genes encoding cytochrome P450 (CYP) enzymes.5 This high CYP gene “density” in the Mtb genome suggests important roles for P450 enzymes, and indeed, a number have been shown to be essential for bacterial viability or host infection6,7 and are potential novel drug targets. The molecular basis of the P450−ligand interactions, as well as the relationship between the structure and binding selectivity and affinity, is of crucial importance, and could facilitate the design of novel therapeutic agents to treat TB. The antifungal azoles and triazole drugs are potent P450 inhibitors, targeting mainly the fungal sterol demethylase (CYP51). Newer generations of azoles do not interact as strongly with human P450s as did earlier azole drugs, and many can be applied systemically as well as topically. Importantly, © 2013 American Chemical Society

various azoles were shown to act as effective inhibitors of cell growth in mycobacteria.8 More recent studies reveal that econazole and clotrimazole have strong antimycobacterial potential against latent Mtb under in vitro conditions, while econazole shows antitubercular activity in mice.9 We have concentrated our study on the Mtb P450 isoform CYP121, which has been shown to be essential for Mtb viability in vitro,10 analyzing its interactions with azoles, with the natural substrate, and finally with 10 novel compounds that were rationally selected on the basis of our recent fragment-based program against CYP121.11 In 2009 Gondry and co-workers identified cyclodipeptide (CDP) synthases as a new class of enzymes. Rv2275, the gene adjacent to CYP121, was found to encode a CDP synthase responsible for producing cyclo(L-TryL-Tyr) (cYY), which they established was a substrate for CYP121.12 Crystal structures of CYP121 in the unbound form, in complex with the azole inhibitor fluconazole and with cYY, have been reported.13 The crystal structure of the unbound CYP121 revealed a highly constrained active site. Major Received: January 23, 2013 Accepted: May 16, 2013 Published: May 16, 2013 5707

dx.doi.org/10.1021/ac400236z | Anal. Chem. 2013, 85, 5707−5714

Analytical Chemistry

Article

such as nanoESI, it is possible to transfer noncovalent complexes from solution to the vacuum of the mass spectrometer and detect the intact complex. This technique has several potential advantages that make it a valuable addition to the arsenal of available binding assays. These include sensitivity, selectivity, and the fact that procedures do not require immobilization or labeling and allow simultaneous measurements of multiple binding equilibria, and require only microgram quantities of the target protein.25 However, the binding measurements can be affected, depending on the experimental conditions and the nature of interactions, by the occurrence of nonspecific protein−ligand binding during the ESI process, or by gas phase dissociation of the complex ion.24,26,27 In this work we aimed to detect, confirm, and characterize the oligomerization state, binding stoichiometries, and interactions of CYP121 with azoles (well-known P450 inhibitors, cYY), the natural substrate of CYP121, and 10 novel test compounds, using nanoESI MS.

reorganization of CYP121 structure was predicted to be necessary to facilitate direct coordination of the P450 heme iron by azole drugs.14 Unexpectedly, the fluconazole-bound form revealed two possible binding modes, either via a bridging water molecule or directly via a triazole nitrogen to the heme iron. The cYY substrate binds in the active site in regions overlapping with those occupied by the ligand in the fluconazole-bound structure,13,14 with cYY interacting with the heme via two hydrogen bonded water molecules, one of which is the sixth ligand to the heme iron.13 Recently, a fragment-based approach was used to study CYP121 inhibition, and crystal structures of CYP121 in complex with fragments and elaborated compounds were successfully determined.11 Novel inhibitors were developed that either coordinated the heme iron directly, or else bound in the active site distal from the immediate heme coordination sphere. The binding of the novel ligands did not cause any conformational changes in CYP121 as observed by crystallography. Notably, a crystal structure showing direct binding of econazole to heme iron in another Mtb P450 (CYP130) has also been determined.15 This crystal structure shows only the direct coordination of the azole to the heme iron, but the oligomerization states of the econazole-bound and unbound enzymes are different. The unbound form of CYP130 is dimeric with a “closed” conformation determined by the position of the BC loop and the F and G helices. These are segments of the protein structure important for active site access, and in CYP130 envelop the inhibitor. Conformational changes caused by “open−closed” transitions reshape the protein surface and are important to enable substrate/inhibitor access and to stabilize their binding in the active site. The binding of azoles to CYP121 was further characterized using optical titrations for determination of the dissociation binding constants.16 P450 optical titrations for the binding of inhibitors (azoles) to the heme iron follow the induced absorption red shift in the major (Soret) band of the heme spectrum. In the case of substrates, a Soret blue shift occurs due to removal of the distal water ligand to the heme iron and in order to prime the enzyme for catalysis. The dissociation constants (Kd values) are computed from equilibrium titration of such ligands with the P450, and by plotting the induced absorption change versus the applied ligand concentration. The resultant data plot (induced absorption change versus ligand concentration) is fitted using a single site binding function to determine the Kd. The binding of clotrimazole, econazole, and miconazole to CYP121 is very tight, with Kd values for the highest affinity azoles estimated at 24 nM (econazole), 73 nM (clotrimazole), and 136 nM (miconazole), while ketoconazole (3.4 μM) and fluconazole (8.6 μM) are bound more weakly. In addition to optical titration16 and the heme absorbance shift assay, a variety of analytical methods have been used to study Mtb P450−azole interactions in vitro including X-ray crystallography,10,14,17−19 EPR spectroscopy,16 isothermal titration calorimetry,15 and molecular docking.20 In general these analytical methods, together with surface plasmon resonance spectroscopy21 and NMR spectroscopy, are powerful techniques for screening ligands against target proteins. Only recently have nanoelectrospray ionization mass spectrometry (nanoESI MS) assays, based on the direct detection and quantification of the relative abundance of free and ligand-bound protein ions, emerged as promising complementary techniques for evaluating the binding stoichiometries and affinities of protein−ligand complexes in solution.22−24 Using soft ionization techniques,

■

EXPERIMENTAL SECTION All experiments with CYP121 were performed using the Nterminal His6-tagged forms unless otherwise specified. Protein masses (Table S1, Supporting Information) were confirmed by mass spectrometry/proteomics using an Applied Biosystems QSTAR nanoESI QTOF mass spectrometer (Applied Biosystems, CA). Azoles (Supporting Information, Table S2) were obtained from Sigma-Aldrich Company Ltd. (Dorset, U.K.), and potential CYP121 binders from the DuPont compound library collection (DuPont Crop Protection, Stine-Haskell Research Centre, Delaware). All other chemicals were of analytical grade and obtained from Sigma-Aldrich. Heme absorbance shift assays were performed as previously described, and involved the titration of azole drugs, cYY, and other ligands against CYP121 using a Cary 50 UV−vis spectrophotometer at 25 °C in 100 mM Tris buffer, pH 7.5. Typically, ∼2.5 μM CYP121 was used for titrations, with aliquots of concentrated ligands/substrate added in DMSO to a final volume of not more than 1% of the overall volume. Spectra were recorded before ligand addition and at each subsequent stage in the titration. Following completion of the titration, difference spectra were generated by subtraction of the ligand-free CYP121 spectrum from each of the subsequent spectra. Absorbance maximum (Amax) and minimum (Amin) wavelengths were identified for each spectral set, and induced absorption change (ΔAbs) at each point in the titration computed as the difference between Amax minus Amin, using the same wavelength pair throughout an individual titration. The Amax values were then plotted versus the relevant [ligand] and data fitted using a hyperbolic function for weaker binding ligands (i.e., Kd greater than ∼20 μM), or using the Morrison equation28 for tight binding ligands (in cases where the Kd was much lower) to determine the Kd value in each case.11 Samples for nanoESI MS were prepared by dilution of the original stock solution, Supporting Information, Table S1, and buffer exchange to 200 mM ammonium acetate, prepared from a 7.5 M solution, and filtered using a Millex-GS syringe-driven filter unit, 0.22 μM pore size (Millipore Corporation, Bedford MA), at pH 7 using Micro Bio-Spin 6 chromatography columns, MW exclusion limit 6 kDa (Bio-Rad). For the preparation of protein−ligand complexes, proteins were mixed with an excess of ligand dissolved in 100% DMSO (typically 5708

dx.doi.org/10.1021/ac400236z | Anal. Chem. 2013, 85, 5707−5714

Analytical Chemistry

Article

assemblies of the remaining six, which are CYP121 ligand-free, single site mutant structures (PDB codes: 3CXY, 3CXZ, 3CYO, 3CY1, 3CXV, and 3CXX) were generated by PISA software and assigned as dimeric. These computational predictions are in agreement with results we obtained by nanoESI MS of intact CYP121. CYP121 is not the only Mtb P450 for which there is evidence for the enzyme being in a dimeric form. CYP130 has been crystallized in the ligand-free form as a monomer and in an econazole-bound form as a dimer.15 As discussed above, the monomeric ligand-free form was in a relatively “open” conformation with respect to active site access from the exterior of the P450, while the dimeric CYP130 in complex with econazole was in a more “closed” conformation as a consequence of relocation of BC and FG helical segments of the protein to interact with the bound ligand. Interaction of CYP121 with Azole Drugs. The azoles clotrimazole (CYP121 Kd = 0.07 ± 0.01 μM), miconazole (0.14 ± 0.02 μM), econazole (0.025 ± 0.01 μM), ketoconazole (3.4 ± 0.3 μM), fluconazole (8.6 ± 0.2 μM), and itraconazole (13.4 ± 0.9 μM) are well-known binders to different P450 enzymes, including those from Mtb. To verify the interaction of CYP121 with these compounds by native nanoESI-MS, CYP121 was directly mixed with azoles in the form of stock solutions in 100% DMSO. The influence of DMSO on the quality of the spectra was first analyzed (Supporting Information, Figure S3). The presence of DMSO did not have any impact on the measured masses, although a notable difference was observed in which charge states were populated, with a shift of peaks in the range of m/z. It is known from previous studies that addition of DMSO can cause significant changes in protein behavior even at low amounts.30,31 The addition of DMSO may influence the stability of proteins and/or change their binding properties. Low DMSO concentrations influence the ionization process in electrospray ionization mass spectrometry (ESI-MS), resulting in a loss of the protein ion signal intensity as well as a decrease in the basic sites available for protonation on the protein. At 5% DMSO, peaks in the range from m/z 3700 to 4700 are assigned to monomeric CYP121 and those from 5200 to 6200 to the dimer, in comparison to those in the ranges 3200−4200 and 4200−5800 using 0% DMSO. The mass spectra recorded when 25 and 50 μM clotrimazole was added to a 2.5 μM solution of CYP121 are shown in the Supporting Information Figure S4a. Surprisingly, addition of clotrimazole was observed to result in the conversion of the dimer to the monomer. When clotrimazole was added at 10 times higher concentration than CYP121, only a small amount of dimer was detected, and the monomer became the predominant species. When CYP121 was mixed with a 20 times higher concentration of clotrimazole, the dimer completely dissociated into monomer. In order to confirm absence of the CYP121 dimer, spectra of the CYP121 without and with clotrimazole were recorded with spraying conditions with lower voltages than those initially used (Figure S4b). It was found that the dimer still dissociated in the same way. Mass spectra were subsequently recorded for addition of econazole, fluconazole, itraconazole, ketoconazole, and miconazole at both 25 μM (Supporting Information, Figure S5) and 50 μM (Figure 1) concentrations. Analogously to clotrimazole, miconazole caused almost complete dimer dissociation, while ketoconazole and econazole induced partial dimer dissociation, and fluconazole and itraconazole had little effect on the observed spectra. A possible explanation is that the more strongly binding azoles bind at the active site of CYP121 and

1:10 molar ratio) to produce 5% (v/v) solutions of ligand in protein. Mass spectra were recorded on a Synapt HDMS instrument (Waters UK Ltd., Manchester, U.K.). To perform nano-ESI, capillaries were prepared in-house from borosilicate glass tubes of 1 mm outer diameter and 0.50 mm inner diameter (Harvard Apparatus, Kent) using a P-2000 micropipet puller (Sutter Instrument Company, Intracel LTD, Hertfordshire, U.K.) and gold-coated using an Edwards Coating System E306A. The capillary tips were cut under a stereo microscope to give inner diameters of 1−5 μm, and 2.5 μL of protein solution was loaded for sampling.29 Given below are instrumental conditions used for measurements carried out in a positive ion mode. A capillary voltage of 1.70 kV was applied to perform nanoESI. A cone voltage of 80 V was used, and the source temperature was maintained at 20 °C. Other voltages important for ion transmission were trap collision energy 12.0−30.0 V and transfer collision energy 12.0−25.0 V with trap and transfer pressure 5.27 × 10−2 mbar, IMS pressure 5.02 × 10−1 mbar, TOF analyzer pressure 1.17 × 10−6 mbar. External calibration of the spectra was achieved using cesium iodide at 100 mg mL−1 in water. Data acquisition and processing were performed using Micromass MassLynx 4.1.

■

RESULTS AND DISCUSSION By using nanoESI MS to study small molecules binding to the intact Mtb CYP121 P450, it was possible to identify the oligomerization state of the protein in conditions near physiological pH, in ammonium acetate buffer, pH 7.0, and to observe ligand binding. Contrary to the expectation that all Mtb P450 isoforms are monomeric in the unbound form, it was found that CYP121 is predominantly homodimeric in aqueous, buffered solution. In order to study the oligomerization state of CYP121, a series of mass spectra were recorded at decreasing concentrations of the protein (Supporting Information, Figure S1). The signals corresponding to the dimer and monomer remained constant while the series of weak peaks in the range m/z 6100−7000, assigned to tetrameric CYP121, decreased with decreasing protein concentration and finally disappeared at 2.5 μM CYP121. This behavior is typical of nonspecific interactions and oligomers that do not have any physiological relevance.29 For comparison, two Mtb P450 isoforms, CYP125 and CYP126, were tested, and both were exclusively monomeric (Supporting Information, Figure S2a,b), under identical solution and spraying conditions as those for which CYP121 was recorded. The mass spectrum of CYP121 has major peaks in the m/z range from 4200 to 5800, assigned to the dimer, with the molecular weight of 89 938 ± 23 Da, while weak intensity peaks, in the range from m/z 3200 to 4200, are assigned to monomeric CYP121 with the molecular weight of 44 962 ± 4 Da. Although N-terminal His6-tagged CYP121 was used as standard, we also tested untagged CYP121 with MW 43741 Da and showed that it was also predominantly dimeric (Supporting Information, Figure S2c). There are 19 crystal structures of CYP121 deposited in the RCSB Protein Data Bank (PDB), that are of a combination of wild type, active site single mutants, and CYP121 in complex with small molecules. According to the deposited data, four of the ligand-free, wild type CYP121 crystal structures (PDB codes: 3G5H, 3G5F, 1N4O, and 2IJ5) are biological monomers as assigned by the authors, as are the complexes of CYP121 with fluconazole, iodopyrazole, and five small molecules obtained from fragment-based studies. However, the biological 5709

dx.doi.org/10.1021/ac400236z | Anal. Chem. 2013, 85, 5707−5714

Analytical Chemistry

Article

fluconazole14 showed two binding modes for the azole: either with its triazole nitrogen bound directly to the heme iron, or with the triazole bound via a bridging water molecule (i.e., via the natural sixth ligand to the iron). No gross conformational changes were observed in the fluconazole complex, likely indicative of a lower energetic cost associated with a binding mode in which indirect coordination (via water) to the heme iron occurs, as opposed to more substantial structural reorganization required to enable a different binding mode with exclusive direct coordination of the CYP121 heme iron. Econazole, clotrimazole, and miconazole probably bind the CYP121 heme directly, as observed in the CYP130−econazole complex, and cause conformational changes that result in dissociation of the dimer. Such a hypothesis is in agreement with previously tested CYP121−azole binding results14,15 although the dimeric state of CYP121 was not previously examined. The possibility that the hydrophobic azoles, which were added at an excess of 10−20-fold over the P450, would bind to the surface of the protein nonspecifically and destabilize the dimer, cannot be completely eliminated. Such nonspecifically bound ligand on the surface of the protein would be expected to detach in the vacuum. Moreover, there are no meaningful data for azole drugs binding to superficial sites on the outside of proteins, mainly since the hydrophobic actives site cavities of the P450s show much greater affinity for this type of drug. Binding of the Natural Substrate cYY to CYP121. Cyclo-L-Tyr-L-Tyr (cYY) is the natural substrate of CYP121. CYP121 catalyzes oxidative coupling of the tyrosyl side chains to form the product mycocyclosin. Optical titrations of CYP121 with cYY show that the P450 heme iron undergoes a low-spin to high-spin shift in its ferric state, usually associated with a substrate-dependent displacement of a water molecule as the sixth ligand to the heme iron. However, in the CYP121-cYY crystal structure,13 the cYY instead forms a hydrogen bonding network with surrounding residues, and interacts via a tyrosyl hydroxyl group with a water molecule, which in turn is hydrogen bonded to a water molecule retained as the sixth ligand to the heme iron.13 Thus, the cYY-bound structure is consistent with a low-spin heme iron form, and further structural reorganization must occur to enable displacement of the sixth ligand and binding of an oxygen molecule. However, the cYY binding mode observed is comparable to the predominant binding mode of fluconazole with CYP121, in that both stabilize the binding of the sixth ligand water.14 Neither cYY nor fluconazole cause significant CYP121 dimer dissociation (Figure 2). The complex of CYP121 with fluconazole was not detected by nanoESI MS (Figure 1) while the CYP121-cYY complex is clearly present. In this experiment 250 μM cYY was added to a 15 μM solution of CYP121. Weak peaks in the range from m/z 3700 to 5200 correspond to the monomer and those from 5200 to 6700 to the dimer. Peaks at the right shoulder of those corresponding to free dimer indicate the presence of cYY-bound CYP121. The recorded mass of the dimer-bound form suggests a combination of singly and doubly bound species. The spectral complexity, with several species present, including dimer in unbound and partially resolved bound form, as well as unresolved bound from unbound monomer, made it difficult to quantify the association constant by nanoESI MS. CYP121 Complexes with Novel Compounds. The first successful application of fragment-based approaches to the development of ligands for a cytochrome P450 was described

Figure 1. NanoESI mass spectra of 2.5 μM CYP121 with 50 μM azoles: 2.5 μM CYP121 was mixed with 5% DMSO and 50 μM clotrimazole, econazole, fluconazole, itraconazole, ketoconazole, and miconazole in 200 mM ammonium acetate (pH 7). Peaks in the range m/z 3700−5200 were assigned to monomer (Mw = 44 995 ± 10 Da) and in the range m/z 5200−6600 to dimer (Mw = 90066 ± 34 Da). Singly dashed lines represent calculated m/z values for monomeric or dimeric CYP121 in complex with only one molecule of bound azole. Doubly dashed lines represent calculated m/z positions where expected peaks for dimeric CYP121 complexes with two bound azoles should appear.

cause a conformational change that is retained when the ligand dissociates during MS analysis, thus resulting in the change in oligomeric state. Although the addition of some azoles to CYP121 strongly destabilizes the dimer, causing its dissociation, the recorded mass spectra showed predominantly monomer in the unbound form (Figure 1). The gas-phase ions corresponding to the specific monomeric (CYP121 + 1 azole) or dimeric (CYP1212 + 1 or 2 azoles) complexes were detected only with ketoconazole (Figure 1 and Supporting Information Figure S5) and clotrimazole (Figure 1, Supporting Information Figures S4a,b and S5). The absence of complexes with other azoles indicates their dissociation during MS analysis.32,33 Previous studies of the M. tuberculosis CYP130 P450 indicated that the econazole-bound form crystallized as a dimer, with a noncrystallographic dimer interface involving the conformationally flexible P450 BC-loop and F/G-helix regions, and the N-terminal part of the I helix. The F/G helices and BCloop regions are differently organized in the ligand-free CYP130 monomer.15 A similar, but smaller, dimerization interface was also observed for the Streptomyces coelicolor CYP154C1.34 It is possible that similar regions are involved in forming a CYP121 dimer in its ligand-free state.15 Previous structural studies of CYP121 indicated that its active site cavity is remarkably rigid in comparison to the rest of the structure, and that ligand binding might require conformational change.19 However, crystal structures of CYP121 in complex with 5710

dx.doi.org/10.1021/ac400236z | Anal. Chem. 2013, 85, 5707−5714

Analytical Chemistry

Article

Figure 2. NanoESI mass spectra of 15 μM CYP121 with cYY: 15 μM CYP121 was mixed with 5% DMSO and 250 μM cYY in 200 mM ammonium acetate (pH 7). Peaks in the range m/z 3700−5200 were assigned to monomer (Mw = 44 996 ± 6 Da) and in the range m/z 5200−6700 to dimer (Mw = 89923 ± 4 Da) and to ligand-bound dimer (Mw = 90 464 ± 15 Da). The recorded mass of the dimericligand complex suggests a combination of singly and doubly cYYbound dimer. Mw (cYY) = 326.35 Da. Singly dashed lines represent calculated m/z values for monomeric or dimeric CYP121 in complex with only one molecule of cYY. Doubly dashed lines represent calculated m/z positions where expected peaks for dimeric CYP121 complexes with two bound cYYs should appear.

recently for Mtb CYP121.11 Following on from this work, 10 compounds were selected on the basis of their similarity to our previous merged fragments and X-ray crystal structures11 (Supporting Information, Table S3). They possess new vectors of chemical growth that could potentially be accommodated within the CYP121 active site to enhance binding.11 ZJ806, WF960, WF957, ZV155, Y3472, and QX858 maintain the 6aminoquinoline scaffold of the previous amino-heme coordinating fragment hit, 2-methylquinolin-6-amine, and the lead merged fragment, 4-(1H-1,2,4-triazol-1-yl)quinolin-6-amine), which bind to CYP121 with Kd values of 400 and 28 μM, respectively.11 Compounds UG415, MF518, QK755, TS960 explore new chemical space around the lead non-heme coordinating merged fragment, 4-(4-phenoxy-1H-pyrazol-3yl)benzene-1,3-diol, which has a Kd of 500 μM.11 Mass spectra were recorded for CYP121 with compounds with molecular weights in the range 190−443 Da, at 50 μM concentration (Figure 3). Very minor dimer dissociation was observed only for TS960 and ZJ806, and was more evident from the slightly higher intensity of monomeric species than from the decrease in the dimer intensities. This effect on the dimer dissociation was similar to that seen with fluconazole or itraconazole. However, in contrast to these azoles, in both cases the ligand-bound monomer and ligand-bound dimer were detected for the fragment-related molecules, suggesting kinetically more stable complexes and possibly a different mode of binding, distal from the heme iron. Compound QK755 does not cause dissociation, and only ligand-bound monomer was observed, even at a 50 μM ligand concentration, suggesting relatively poor binding affinity with CYP121. Due to the high spectral complexity, with contributions from free and bound monomer, and free, singly, and doubly bound dimer, calculation of association/dissociation constants was not feasible. An attempt was, however, made to rank the compounds to give a relative order of binding affinities. Apart from 4 compounds (ZV155, WF957, WF960, and MF518) that, according to our data, do not bind CYP121, the rest of 6

Figure 3. NanoESI mass spectra of 2.5 μM CYP121 with novel compounds (L): 2.5 μM CYP121 was incubated with 5% DMSO and 50 μM UG415, MF518, QK755, ZJ806, WF960, WF957, ZV155, Y3472, and TS960 in 200 mM ammonium acetate (pH 7). Peaks in the range m/z 3700−5200 were assigned to monomer (Mw = 44 976 ± 31 Da) and to ligand-bound monomer (MwQK755 = 45 320 ± 33 Da, MwUG415 = 45 404 ± 27 Da, MwZJ806 = 45 273 ± 15 Da, MwTS960 = 45 352 ± 6 Da, MwY3472 = 45 167 ± 42 Da); and in the range m/z 5200−6600 to dimer (Mw = 89991 ± 17 Da), singly bound dimer (MwUG415 = 90 420 ± 26 Da, MwZJ806 = 90 336 ± 34 Da, MwTS960 = 90 422 ± 26 Da, MwQX858 = 90 268 ± 445 Da), and doubly bound dimer (MwY3472 = 90 404 ± 66 Da). Singly dashed lines represent calculated m/z values for monomeric or dimeric CYP121 in complex with only one molecule of bound compound. Doubly dashed lines represent calculated m/z positions where expected peaks for dimeric CYP121 complexes with two bound compounds should appear.

compounds were ranked on the basis of observation. The best binder of CYP121, out of 10 screened compounds, was the 6aminoquinoline-containing compound Y3472. This compound binds both the monomer and the dimer, and both the monomer and the dimer are present in a fully bound form. Two compounds (TS960 and QX858) bind to the monomer and to the dimer. While in both cases monomers were present in unbound and predominantly bound form, dimers were detected only in the bound form. Masses for singly bound compounds were measured. In both cases, predicted positions for peaks with doubly bound compounds indicate their presence, although the resolution of spectra was not high enough to calculate their masses. Compound ZJ806 has more monomer in the unbound than in the bound form, while the dimer exists as unbound, and bound with one and two compounds. Compounds UG415 and QX755 are, according to our spectra, very similar in terms of binding affinity, and in both cases unbound monomer and unbound dimer are dominant species, 5711

dx.doi.org/10.1021/ac400236z | Anal. Chem. 2013, 85, 5707−5714

Analytical Chemistry

Article

by Y3472 (shown as a ΔAmax at 428 nm and ΔAmin at 409.5 nm in the difference spectra) is also indicative of the expected type II CYP binding via heme iron coordination.

although complexes with momomer and singly bound dimer were detected. In an additional experiment, a mass spectrum of CYP121 with 2.5 μM Y3472 (Figure 4) showed a shoulder of unbound

■

CONCLUSIONS In this work, nanoESI MS was used to study binding stoichiometries and affinities of the essential Mtb P450 isoform CYP121 with azoles, the natural substrate cYY, and some novel compounds related to molecules identified as binding CYP121 in fragment-based studies.11 Surprisingly, CYP121 is a predominantly dimeric protein in the unbound (ligand-free) form, although monomer is also present. The monomeric/ dimeric oligomerization distribution of azole-bound CYP121 changes according to the specific binding affinity of the particular azole used. The degree of dimer dissociation, induced by mixing azoles with CYP121, correlates very well with the determined rank order of azole affinities by heme absorbance shift assays. More precisely, stronger binders such as clotrimazole or miconazole cause the dimer to dissociate, while weak binders, such as fluconazole or itraconazole, do not. However, only complexes of CYP121 with two of the azoles were detected, indicating kinetically labile assemblies that are prone to complete gas-phase dissociation during the MS analysis. In contrast to azole binding, the natural cyclic dipeptide substrate of CYP121, cYY, does not alter the dimeric/ monomeric distribution, and the resulting complex was far less inclined to gas-phase dissociation. Dimeric CYP121 in the unbound form and in complex with cYY was detected. Ten novel compounds rationally selected as potential CYP121 binders/inhibitors were tested, of which six form complexes with CYP121 that can be detected by nanoESI MS. Moreover, these compounds do not influence the dimeric/monomeric distribution of CYP121. Y3472 seems to bind very well to both monomeric and dimeric CYP121, even at stoichiometric amounts that suggested relatively high binding affinity. This was confirmed by heme absorbance shift assay, giving a Kd of 40 μM. Two out of the ten novel compounds tested, TS960 and ZJ806, caused minor dimer dissociation, and bound monomer was clearly resolved from unbound monomer. If we take into account that sometimes small differences afforded by binding of ligands to protein assemblies cannot be detected by “native” MS, as minimal activation of the assembly is sufficient to dissociate the ligand,35 the clear detection of CYP121-fragment complexes indicates tight binding and stability. NanoESI MS results provide valuable information, not only about the oligomerization state of CYP121, but also about its binding affinity and binding mode, both of which are parameters that vary according to ligand used. The summary of the effects caused by binding small molecule ligands such as azoles, cYY, and novel compounds (fragments) to CYP121 are presented in the Scheme 1. CYP121 is an essential P450 for Mtb viability and potential antitubercular drug target against the human pathogen. This novel application of nanoESI MS could be further developed for determining binding affinity based on protein mass and monomer/dimer equilibrium. It may thus provide an important tool for screening inhibitor interactions in the case of CYP121 (and other P450s), and for other macromolecular systems where stable ligand-bound states can be isolated, or where drugs target protein−protein interactions.36−39

Figure 4. NanoESI mass spectra of 2.5 μM CYP121 with 50 μM and 2.5 μM Y3472 (L): 2.5 μM CYP121 was incubated with 5% DMSO and 50 μM and 2.5 μM Y3472 in 200 mM ammonium acetate (pH 7). Peaks in the range m/z 3700−4800 were assigned to monomer (Mw = 44 976 ± 31 Da) and ligand-bound monomer (Mw = 45 167 ± 42 Da); and in the m/z range 5200−6600 to dimer (Mw = 89 991 ± 42 Da), singly bound dimer (Mw = 90 277 ± 22 Da), and doubly bound dimer (Mw = 90 404 ± 66 Da). In the presence of 50 μM Y3472, no free CYP121 dimer was observed. Singly dashed lines represent calculated m/z values for dimeric CYP121 in complex with only one molecule of Y3472 (middle spectrum). Doubly dashed lines represent the calculated m/z position where expected peaks for dimeric CYP121 complex with two bound Y3472 molecules should appear (top spectrum).

dimer, which disappeared when ligand concentration was at 50 μM, together with singly ligand-bound dimer. The fact that CYP121 is in the singly ligand-bound form at stoichiometric concentrations of protein and ligand indicates a relatively high binding affinity for Y3472. Binding was subsequently confirmed using a heme absorbance shift assay, giving a Kd of 40 μM (Figure 5), and was weaker than expected on the basis of MS data. The red shift in the heme Soret absorbance band induced

Figure 5. Absorbance difference spectra for CYP121 (5 μM) titrated with various concentrations of Y3472. Main figure shows the effects of increasing concentrations of Y3472 on the heme spectrum of CYP121, presented as type II difference spectra. The inset plot shows the change in heme absorbance (quantified as ΔAmax − ΔAmin) as a function of Y3472 concentration (concentration−response curve, □) with a one-site binding equilibrium model fitted to calculate the Kd (line). A Kd value of 40 ± 3 μM was determined for Y3472. 5712

dx.doi.org/10.1021/ac400236z | Anal. Chem. 2013, 85, 5707−5714

Analytical Chemistry

Article

Scheme 1. CYP121 in Interaction with Azoles, the Natural Substrate cYY, and Some Novel Compounds (L) from FragmentBased Studies Showing Distinct Behaviora

a High affinity azoles cause conformational changes that lead to the dissociation of the CYP121 dimer into monomers. The same effect was not observed when CYP121 interacts with the natural cYY substrate and most of the novel compounds. Monomeric and/or dimeric complexes of CYP121 with some novel compounds were detected, while binding of the natural substrate was detected only with the dimeric protein.

■

(2) Spigelman, M. K. J. Infect. Dis. 2007, 196 (Supplement 1), S28− S34. (3) Matsumoto, M.; Hashizume, H.; Tsubouchi, H.; Sasaki, H.; Itotani, M.; Kuroda, H.; Tomishige, T.; Kawasaki, M.; Komatsu, M. Curr. Top. Med. Chem. 2007, 7, 499−507. (4) Fleischmann, R. D.; Alland, D.; Eisen, J. A.; Carpenter, L.; White, O.; Peterson, J.; DeBoy, R.; Dodson, R.; Gwinn, M.; Haft, D.; Hickey, E.; Kolonay, J. F.; Nelson, W. C.; Umayam, L. A.; Ermolaeva, M.; Salzberg, S. L.; Delcher, A.; Utterback, T.; Weidman, J.; Khouri, H.; Gill, J.; Mikula, A.; Bishai, W.; Jacobs, J. W. R.; Venter, J. C.; Fraser, C. M. J. Bacteriol. 2002, 184 (19), 5479−5490. (5) McLean, K. J.; Clift, D.; Lewis, D. G.; Sabri, M.; Balding, P. R.; Sutcliffe, M. J.; Leys, D.; Munro, A. W. Trends Microbiol. 2006, 14 (5), 220−228. (6) McLean, K. J.; Belcher., J.; Driscoll, M. D.; Fernandez, C. C.; Le Van, D.; Bui, S.; Golovanova, M.; Munro, A. W. Future Med. Chem. 2010, 2 (8), 1339−1353. (7) Ouellet, H.; Johnston, J. B.; Ortiz de Montellano, P. R. Arch. Biochem. Biophys. 2010, 493 (1), 82−95. (8) McLean, K. J.; Marshall, K. R.; Richmond, A.; Hunter, I. S.; Fowler, K.; Kieser, T.; Gurcha, S. S.; Besra, G. S.; Munro, A. W. Microbiology 2002, 148 (10), 2937−2949. (9) Ahmad, Z.; Sharma, S.; Khuller, G. K. FEMS Microbiol. Lett. 2006, 258 (2), 200−203. (10) McLean, K. J.; Carroll, P.; Lewis, D. G.; Dunford, A. J.; Seward, H. E.; Neeli, R.; Cheesman, M. R.; Marsollier, L.; Douglas, P.; Smith, W. E.; Rosenkrands, I.; Cole, S. T.; Leys, D.; Parish, T.; Munro, A. W. J. Biol. Chem. 2008, 283 (48), 33406−33416. (11) Hudson, S. A.; McLean, K. J.; Surade, S.; Yang, Y.-Q.; Leys, D.; Ciulli, A.; Munro, A. W.; Abell, C. Angew. Chem. Int. Ed. 2012, 51 (37), 9311−9316. (12) Gondry, M.; Sauguet, L.; Belin, P.; Thai, R.; Amouroux, R.; Tellier, C.; Tuphile, K.; Jacquet, M.; Braud, S.; Courcon, M.; Masson, C.; Dubois, S.; Lautru, S.; Lecoq, A.; Hashimoto, S.-i.; Genet, R.; Pernodet, J.-L. Nat. Chem. Biol. 2009, 5 (6), 414−420. (13) Belin, P.; Le Du, M. H. l. n.; Fielding, A.; Lequin, O.; Jacquet, M. l.; Charbonnier, J.-B.; Lecoq, A.; Thai, R.; Courçon, M.; Masson, C. d.; Dugave, C.; Genet, R.; Pernodet, J.-L.; Gondry, M. Proc. Natl. Acad. Sci. U.S.A. 2009, 106 (18), 7426−7431. (14) Seward, H. E.; Roujeinikova, A.; McLean, K. J.; Munro, A. W.; Leys, D. J. Biol. Chem. 2006, 281 (51), 39437−39443.

ASSOCIATED CONTENT

S Supporting Information *

Additional figures and tables. This material is available free of charge via the Internet at http://pubs.acs.org.

■

AUTHOR INFORMATION

Corresponding Author

*Phone: 00441223763174. E-mail: [email protected]. Author Contributions

The manuscript was written through contributions of all authors. Notes

The authors declare no competing financial interest.

■

ACKNOWLEDGMENTS The authors would like to thank Dr. Anthony Coyne for helpful discussions and reading of the manuscript. Compounds UG415, MF518, QK755, ZJ806, WF960, WF957, ZV155, Y3472, and TS960 were from the DuPont HTS compound library and were kindly provided by Dr. Robert Pasteris (DuPont Crop Protection). We acknowledge funding from the BBSRC (Grants BB/I019227/1 and BB/I019669/1 to C.A. and A.W.M., respectively) and the Wellcome Trust under the Seeding Drug Discovery Initiative. S.A.H. was supported by a Sir Mark Oliphant Cambridge Australia Scholarship awarded by the Cambridge Commonwealth Trust & Cambridge Overseas Trust, University of Cambridge.

■

REFERENCES

(1) Cole, S. T.; Brosch, R.; Parkhill, J.; Garnier, T.; Churcher, C.; Harris, D.; Gordon, S. V.; Eiglmeier, K.; Gas, S.; Barry, C. E.; Tekaia, F.; Badcock, K.; Basham, D.; Brown, D.; Chillingworth, T.; Connor, R.; Davies, R.; Devlin, K.; Feltwell, T.; Gentles, S.; Hamlin, N.; Holroyd, S.; Hornsby, T.; Jagels, K.; Krogh, A.; McLean, J.; Moule, S.; Murphy, L.; Oliver, K.; Osborne, J.; Quail, M. A.; Rajandream, M. A.; Rogers, J.; Rutter, S.; Seeger, K.; Skelton, J.; Squares, R.; Squares, S.; Sulston, J. E.; Taylor, K.; Whitehead, S.; Barrell, B. G. Nature 1998, 393 (6685), 537−544. 5713

dx.doi.org/10.1021/ac400236z | Anal. Chem. 2013, 85, 5707−5714

Analytical Chemistry

Article

(15) Ouellet, H.; Podust, L. M.; de Montellano, P. R. O. J. Biol. Chem. 2008, 283 (8), 5069−5080. (16) McLean, K. J.; Cheesman, M. R.; Rivers, S. L.; Richmond, A.; Leys, D.; Chapman, S. K.; Reid, G. A.; Price, N. C.; Kelly, S. M.; Clarkson, J.; Smith, W. E.; Munro, A. W. J. Inorg. Biochem. 2002, 91 (4), 527−541. (17) Hudson, S. A.; McLean., K. J.; Munro, A. W.; Abell, C. Biochem. Soc. Trans. 2012, 40, 573−579. (18) McLean, K. J.; Lafite, P.; Levy, C.; Cheesman, M. R.; Mast, N.; Pikuleva, I. A.; Leys, D.; Munro, A. W. J. Biol. Chem. 2009, 284 (51), 35524−35533. (19) Leys, D.; Mowat, C. G.; McLean, K. J.; Richmond, A.; Chapman, S. K.; Walkinshaw, M. D.; Munro, A. W. J. Biol. Chem. 2003, 278 (7), 5141−5147. (20) Sundaramurthi, J. C.; Kumar, S.; Silambuchelvi, K.; Hanna, L. E. Bioinformation 2011, 7 (3), 130−3. (21) Pearson, J. T.; Hill, J. J.; Swank, J.; Isoherranen, N.; Kunze, K. L.; Atkins, W. M. Biochemistry 2006, 45 (20), 6341−6353. (22) Daniel, J. r. M.; Friess, S. D.; Rajagopalan, S.; Wendt, S.; Zenobi, R. Int. J. Mass Spectrom. 2002, 216 (1), 1−27. (23) Heck, A. J. R.; van den Heuvel, R. H. H. Mass Spectrom. Rev. 2004, 23 (5), 368−389. (24) El-Hawiet, A.; Kitova, E. N.; Liu, L.; Klassen, J. S. J. Am. Soc. Mass Spectrom. 2010, 21 (11), 1893−1899. (25) El-Haweit, A.; Shoemaker, G. K.; Daneshfar, R.; Kitova, E. N.; Klassen, J. S. Anal. Chem. 2011, 84 (1), 50−58. (26) Sun, N.; Soya, N.; Kitova, E.; Klassen, J. J. Am. Soc. Mass Spectrom. 2009, 21 (3), 472−481. (27) El-Hawiet, A.; Kitova, E. N.; Arutyunov, D.; Simpson, D. J.; Szymanski, C. M.; Klassen, J. S. Anal. Chem. 2012, 84 (9), 3867−3870. (28) Morrison, J. F. Biochim. Biophys. Acta, Enzymol. 1969, 185 (2), 269−286. (29) Hernandez, H.; Robinson, C. V. Nat. Protoc. 2007, 2 (3), 715− 726. (30) Lemaire, D.; Marie, G. r.; Serani, L.; Laprévote, O. Anal. Chem. 2001, 73 (8), 1699−1706. (31) Tjernberg, A.; Markova, N.; Griffiths, W. J.; Hallén, D. J. Biomol. Screening 2006, 11 (2), 131−137. (32) Wang, W.; Kitova, E. N.; Klassen, J. S. Anal. Chem. 2003, 75 (19), 4945−4955. (33) Sun, J.; Kitova, E. N.; Klassen, J. S. Anal. Chem. 2006, 79 (2), 416−425. (34) Podust, L. M.; Kim, Y.; Arase, M.; Neely, B. A.; Beck, B. J.; Bach, H.; Sherman, D. H.; Lamb, D. C.; Kelly, S. L.; Waterman, M. R. J. Biol. Chem. 2003, 278 (14), 12214−12221. (35) Hyung, S.-J.; Robinson, C. V.; Ruotolo, B. T. Chem. Biol. 2009, 16 (4), 382−390. (36) Pacholarz, K. J.; Garlish, R. A.; Taylor, R. J.; Barran, P. E. Chem. Soc. Rev. 2012, 41 (11), 4335−4355. (37) Wortmann, A.; Jecklin, M. C.; Touboul, D.; Badertscher, M.; Zenobi, R. J. Mass Spectrom. 2008, 43 (5), 600−608. (38) Krishnaswamy, S. R.; Williams, E. R.; Kirsch, J. F. Protein Sci. 2006, 15 (6), 1465−1475. (39) Kool, J.; Jonker, N.; Irth, H.; Niessen, W. Anal. Bioanal. Chem. 2011, 401 (4), 1109−1125.

5714

dx.doi.org/10.1021/ac400236z | Anal. Chem. 2013, 85, 5707−5714