Homology Modeling and Molecular Dynamics Simulation Combined

Nov 23, 2017 - A combination of molecular dynamics (MD) simulations and X-ray scattering (SAXS) has emerged as the approach of choice for studying pro...
0 downloads 7 Views 7MB Size
Article Cite This: J. Phys. Chem. B 2017, 121, 11229−11240

pubs.acs.org/JPCB

Homology Modeling and Molecular Dynamics Simulation Combined with X‑ray Solution Scattering Defining Protein Structures of Thromboxane and Prostacyclin Synthases Hsiao-Ching Yang,*,† Cheng-Han Yang,† Ming-Yi Huang,† Jyh-Feng Lu,‡ Jinn-Shyan Wang,‡ Yi-Qi Yeh,§ and U-Ser Jeng§,∥ †

Department of Chemistry, Fu Jen Catholic University, New Taipei City 24205, Taiwan School of Medicine, Fu Jen Catholic University, New Taipei City 24205, Taiwan § National Synchrotron Radiation Research Center, Hsinchu Science Park, Hsinchu 30076, Taiwan ∥ Department of Chemical Engineering, National Tsing Hua University, Hsinchu 30013, Taiwan ‡

S Supporting Information *

ABSTRACT: A combination of molecular dynamics (MD) simulations and X-ray scattering (SAXS) has emerged as the approach of choice for studying protein structures and dynamics in solution. This approach has potential applications for membrane proteins that neither are soluble nor form crystals easily. We explore the watercoupled dynamic structures of thromboxane synthase (TXAS) and prostacyclin synthase (PGIS) from scanning HPLC−SAXS measurements combined with MD ensemble analyses. Both proteins are heme-containing enzymes in the cytochrome P450 family, known as prostaglandin H2 (PGH2) isomerase, with counter-functions in regulation of platelet aggregation. Currently, the X-ray crystallographic structures of PGIS are available, but those for TXAS are not. The use of homology modeling of the TXAS structure with ns−μs explicit water solvation MD simulations allows much more accurate estimation of the configuration space with loop motion and origin of the protein behaviors in solution. In contrast to the stability of the conserved PGIS structure in solution, the pronounced TXAS flexibility has been revealed to have unstructured loop regions in connection with the characteristic P450 structural elements. The MD-derived and experimental-solution SAXS results are in excellent agreement. The significant protein internal motions, whole-molecule structures, and potential problems with protein folding, crystallization, and functionality are examined.



metabolism and detoxification.10 Data-mining methods have been used to find a P450 gene superfamily in which there have seven main gene clusters with 57 functional P450 genes and 46 pseudogenes in the human.11 Several hundred P450 crystal structures have been determined, and on the basis of sequence alignments, the overall fold is maintained in all P450s. They have a conserved buried Cys-heme pocket, which catalyzes the C−H bond hydroxylation in the use of various organic compounds as carbon sources and in the production of important natural products such as antibiotics, steroids, and fatty acids.12,13 In general, many of the drug-metabolizing P450s are not specific and are capable of hydroxylating a variety of diverse and unrelated compounds.12 However, those P450s involved in the production of important intermediates, such as in lipid fatty

INTRODUCTION

The function of a protein is determined by both its folding structure and its conformational dynamics. In solution, a folded protein is not a single structure but an ensemble of dynamic entities that are relevant to substrate binding and enzyme activity for efficient chemical and physiological functions.1−3 Although more than 100,000 protein crystallographic structures have been determined by X-ray diffraction, pending investigations have yet to address questions of greater difficulty in protein crystallization, such as that of membrane proteins.4−6 Nevertheless, even if the protein has a crystallographic structure, the static crystal structure cannot represent the native functional structure in hydration in many cases. Recent advances have provided more convincing evidence that probing the protein hydrodynamics may pave a way to understanding the underlying mechanisms in the activated biofunctionality, such as cytochrome P450 reactions.7−9 P450s refer to a group of enzymes that catalyze the mono-oxygenation of various organic molecules. They are best known for their role in drug © 2017 American Chemical Society

Received: August 19, 2017 Revised: November 23, 2017 Published: November 23, 2017 11229

DOI: 10.1021/acs.jpcb.7b08299 J. Phys. Chem. B 2017, 121, 11229−11240

Article

The Journal of Physical Chemistry B Scheme 1. Isomerization of Prostaglandin H2 (PGH2) by Prostacyclin Synthase (PGIS) and Its Counterpart Enzyme Thromboxane Synthase (TXAS)

associated with a variety of cardiovascular and pulmonary diseases.20 Unfortunately, mainly due to the lack of an X-ray crystallographic structure, information relevant to the structure−functionality relationship of TXAS is still rather limited. Alternatively, molecular modeling and spectral binding assays have been implemented in predicting the active site and its mechanism of metabolism.21−23 When implemented in combination with other data, such as intermolecular ligand binding, optical probes are also extremely useful in the determination of molecular complexes and protein microenvironments.8,9,24,25 Evidently, these models require the construction of a three-dimensional representation of TXAS in order to probe the nature and the structural properties responsible for its specific functionality.8,9 The protein interaction sites/heme active site, which can be determined from the protein model, have been further scrutinized by optical probes in a bid to enhance their predictive probability.24,25 Nevertheless, these models have to be verified by either crystallization or site-directed mutagenesis experiments, in which the increasing number of P450 crystal structures would facilitate the homology modeling efforts. Therefore, assessment of the quality of data and structures is of utmost importance, particularly when automation protocols are employed to assess the consistency of the data directly by information theoretical methods.26−29 In this study, a combination of all-atom molecular dynamics simulations and scanning small-angle X-ray scattering (SAXS)/ UV−vis absorption measurements along the elution path of high-performance liquid chromatography (HPLC)30,31 was performed to probe the structures and dynamical behaviors for the counter-functioning proteins of TXAS and PGIS in solution. Notably, on synchrotrons, the intense X-ray beam may cause radiation damage to the protein samples.27 Here we show that the scanning HPLC−SAXS approach can effectively exclude scattering containment from coexisting oligomer species and collect SAXS of individual monomer protein

acid and steroid metabolism, are highly specific. A particularly challenging problem is to understand how P450s adapt to accommodate different substrates, given the restriction that the same fold is maintained in all P450s, as well as the flexibility in solution that allows the loop open−close motion,12 which permits substrates to enter and products to leave. Here we explore the prostaglandin H2 (PGH2) isomerases, Scheme 1, thromboxane synthase (TXAS) and prostacyclin synthase (PGIS). These two proteins provide dramatic examples of the protein duality of restriction and flexibility of ranges of motion available to P450 catalytic and product fidelity, with counterfunctions in regulation of platelet aggregation and cardiovascular homeostasis.14,15 Unlike other microsomal P450s, which catalyze the NAD(P)H and O2-dependent mono-oxygenation reactions, a distinctive feature of PGIS and TXAS is that they do not need molecular oxygen, reductase, or any other external electron donor of NAD(P)H.16 TXAS not only catalyzes the isomerization reaction of PGH2, yielding the unstable proaggregatory and vasoconstrictive agent, thromboxane A2 (TXA2), but also alternatively performs a fragmentation reaction of PGH2, producing 12-L-hydroxy-5,8,10-heptadecatrienoic acid (HHT) and malondial-dehyde (MDA), at a molar ratio of 1:1:1 for TXA2:HHT:MDA.16 This is in sharp contrast to its counter enzyme of prostacyclin synthase (PGIS), which presents high product fidelity of isomerization of PGH2 into prostacyclin I2 (PGI2) as an autocrine and/or paracrine lipid mediator.16 Crystallographic analysis, together with molecular dynamics simulations and resonance Raman spectra, have previously revealed that PGIS presents a ligand-specific heme conformational change to accommodate the substrate binding, but a similar behavior is not observed for TXAS.7 To gain in-depth insight into PGH2 metabolized by P450s, a number of groups have demonstrated the role of TXAS in the platelet function and vascular biology.17−19 The characteristics and balance of these products are believed to be crucial for evaluating the risks 11230

DOI: 10.1021/acs.jpcb.7b08299 J. Phys. Chem. B 2017, 121, 11229−11240

Article

The Journal of Physical Chemistry B

revealing a rich conformational landscape for analysis (rootmean-square fluctuation (RMSF) and 2D-root-mean-square displacement (RMSD), see Supporting Information). Protein Expression and Purification. Recombinant human TXAS and PGIS were expressed and purified.36 In brief, the constructs for TXAS(PGIS) recombination were designed by replacing the first 28 (17) amino acid residues in the N-terminal transmembrane domain with the hydrophilic sequence, MAKKTSS, for TXAS (PGIS) cDNA. A fourhistidine tag was added to the C-terminus, transforming into BL21/(DE3)pLys for TXAS (PGIS) to facilitate protein purification. Bacteria transformed with the TXAS (PGIS) expression vector were grown overnight in 2YT(LB) medium containing 100 μg/mL ampicillin. An overnight culture was inoculated at a 1:40 ratio into medium. Bacteria were grown at 37 °C in a shaker at 225 rpm until the A600 was between 0.4 and 0.5, and δ-ALA was added to a final concentration of 0.25 mM. When the absorbance was between 0.8 and 0.9, expression of protein was induced by addition of 1 mM IPTG, and the culture was continued for 18−20 h at 200 rpm and 30 °C before cells were harvested by centrifugation at 4000 rpm and then frozen at −80 °C. Frozen cell pellets were thawed and resuspended in sodium phosphate buffer (50 mM NaPi, pH 7.5 containing 10% glycerol, 0.1 M NaCl, 50 μg/ml DNase, 2 mM MgCl2, and 20 mg Lysozyme) at a buffer solution to cell pellet ratio of 4:1 (v:w). The cells were lysed by sonicator with 20 cycles of 4 s burst/4 s cooling. To solubilize protein, NP40 was added in 3% (final concentration of 25 mM OG for PGIS), and the homogenate was stirred overnight at 4 °C. The target proteins containing supernatant were collected by centrifugation at 34 000 rpm and 4 °C for 1 h, and then, the crude solubilized enzyme was applied to a Ni-NTA column, followed by a ion-exchange chromatography column, to maximize the purity as previously described. The concentration and purity of the resulting enzymes were determined by the absorption ratio of the Soret band (TXAS, 420 nm; PGIS, 418 nm) and tryptophan (280 nm). Scanning HPLC/SAXS and Data Analysis. X-ray solution scattering measurements were performed at the BL23A SWAXS endstation of the National Synchrotron Radiation Research Center (NSRRC), Taiwan. Using 15 keV X-rays (wavelength λ = 0.8266 Å) with sample-to-detector distances of 1254 and 3404 mm. The scattering vector q is defined by 4πλ−1 sin θ with scattering angle 2θ that covered from 0.007 to 0.65 Å−1. Data were corrected for electronic noise, sample transmission, and detector sensitivity, followed by scaling to the absolute intensity I(q) in units of cm−1 (scattering cross section per unit volume) via the calibrated scattering intensity from cytochrome c. Through the online size-exclusion highperformance liquid chromatographic (SE-HPLC) system,30 the monomer protein structures with dynamic features for both protein samples in solution were measured. A slow HPLC elution rate was used for improved SAXS data collection, with simultaneous UV−vis monitoring (200−800 nm in wavelength). Data were evaluated for radiation damage, background subtraction quality, and sample concentration effects. Subsequently, the well-overlapped SAXS profiles collected over the sample elution peak of HPLC were integrated to improve the data statistics. Zero-angle scattering intensities (Io) and the corresponding Rg values were extracted from SAXS data using the Guinier approximation. SAXS data evaluation and model fittings were performed using the ATSAS package.40 CRYSOL was used in the SAXS data fitting with the available crystal

structures (see Figure 2a), as well as limit radiation damage via flowing of the sample during data collection (the online system, see Figure S5).30 Our MD simulation strategy contains three consecutive steps: (i) perform long-time explicit water-coupled molecular dynamics simulations to collect the time-dependent dynamical trajectory of a given protein, (ii) demonstrate the MD trajectory ensemble analysis with configuration space to derive the simulation-derived SAXS spectra, and (iii) clarify the protein internal motions to affect the protein fold restriction and flexibility to examine the experimental-solution SAXS results. Comparison of the crystallographic structures, homology modeling, explicit water MD simulations, and SAXS results reveals the significant differences in structural features of the objective proteins of PGIS and TXAS, including the size and shape as well as the flexibility, and correlations to the countering biofunctionalities.



MATERIALS AND METHODS Full computational methodology, growth, and induction of bacteria and X-ray scattering profile analyses are included in the Supporting Information. Homology Modeling and Molecular Dynamics Simulation. Seminal crystallography work allowed us to perform the homology sequence and structural alignment analyses for the relevant P450s, and to construct a 3D structure of the pending TXAS from homology modeling. These P450s crystal structures (we have aligned 324 solved structures in protein data bank (PDB)) show the highly conserved secondary structure features and distinct substrate specificities. The newly crystallized CYP3A4 has been already investigated as an alternative template to model P450s for drug discovery applications.32 Analyzing the CYP3A4 structure with human TXAS provides a basis for homology studies. It is highly homologous in identity of 34.9%, in similarity of 61.1% and for understanding some facts and drawbacks of the modeling techniques used earlier for in P450cam.33 All molecular dynamics simulations were conducted using the Amber 14 package.34 Starting coordinates were taken from two X-ray crystallography structures of PGIS (PDB: 2IAG35 and 3B9936) and the homology structure of TXAS. Simulations were based on a force field that extends the improved side-chain torsion potentials of the Amber ff14SB37 protein force field, and a set of GAFF38 parameters was adopted for the description of the heme. The heme Mulliken charges were added to the force field, and the residue Mulliken charges were based on the amino acid libraries in the Amber 14 package. Periodic boundary conditions were imposed for solvent−solute systems of PGIS/TXAS in parallelepiped boxes. The box lengths of PGIS, 85.33 × 83.31 × 85.90 Å3, contained 445 amino acids with 17 000 TIP4P-Ew39 water molecules. Similarly, the box lengths of TXAS, 86.64 × 96.48 × 95.59 Å3, contained 500 amino acids with 24 500 TIP4P-Ew water molecules. The system then underwent a 10 ns annealing NPT ensemble with equilibrated steps from 0 to 300 K under a constant pressure of 1.0 bar. A Langevin thermostat was used to maintain the system temperature by controlling the collision frequency at 1 ps−1 to the target temperature 300 K. MD simulations were carried out in the canonical ensemble (NVT) with the Langevin thermostat to maintain the system temperature. The SHAKE algorithm was implemented to constrain the covalent bond involving hydrogen atoms. Numerical integration was performed with a time-step of 1 fs for all MD simulations. We performed ∼1 μs MD simulations for both protein systems, 11231

DOI: 10.1021/acs.jpcb.7b08299 J. Phys. Chem. B 2017, 121, 11229−11240

Article

The Journal of Physical Chemistry B

Figure 1. Schematic diagram of P450s in the sequence alignments with structure folds to PGIS (PDB: 3B99) and TXAS homology structures. The partial sequences shown for the motifs, the F, G, I, and relevant helices, are labeled, and the connection loops are highlighted. These helices slide over the surface of the I helix, which leads to an open−close motion of the access channel leading to the active site of the heme, with stick figures in violet.

structures and homology structure as well as those dynamical conformations of all-atom MD simulations. The value of χ2 provides the discrepancy between the experimental data and calculated curves. The χ2 equation is defined as eq 1: χ2 =

1 N

⎡ Iexp(si) − Is(si) ⎤2 ⎥ ∑ ⎢⎢ ⎥⎦ σ(si) ⎣

ρ(S1 , S2) ⎧ ⎡ 1 ⎪1 ⎢ =⎨ 2 ⎪ ⎩ 2 ⎢⎣ N1d 2

N1

∑ ρ2 (S1i , S2) + i=1

1 N2d12

N2

⎤⎫ ⎪

i=1

⎥⎦⎪ ⎭

∑ ρ2 (S2i , S1)⎥⎬ (2)

where Ni is the number of points in Si and the fineness di is the average distance between the neighbor points in Si.

(1)



where N is the number of experimental points, Iexp(si) is the experiment curve, Is(si) is structure/homology/MD conformations simulation-derived-SAXS curves, and σ(si) is the experiment errors. We also reconstructed the protein envelopes of the samples using ab initio modeling of GASBOR, thereby building a lowresolution structure from the experimental SAXS profile to validate the model against the available higher-resolution structures from complementary methods such as X-ray crystallography and MD structures. The value of the normalized spatial discrepancy (NSD) provides a quantitative estimate of the similarity between the envelope and MD structures. The NSD equation is defined as eq 2

RESULTS AND DISCUSSION Homology TXAS Construction and Functionally Significant Helices with Loops of P450s. By homology sequence alignment of these objective P450s, Figure 1, this reveals the general feature of those secondary structures and determines relative functional motifs (details, see Figure S1). The arrangement of the seven α-helixes and three β-sheet motifs is generally correctly represented, and significant differences were observed in the relative orientations and shifts of the helices with regard to the center of the heme-ligation cysteine residue. PGIS exhibits a highly conserved structure fold feature, as maintained in P450s. Most structural deviations were 11232

DOI: 10.1021/acs.jpcb.7b08299 J. Phys. Chem. B 2017, 121, 11229−11240

Article

The Journal of Physical Chemistry B

Figure 2. Comparison of (a) the SAXS profiles via the size-exclusion column (SEC) elution and bypass modes. (b) Radiation damage on the cumulative time SAXS profiles of the PGIS monomeric protein sample. Radii of gyration (Rg) and I(0) profiles extracted from the SAXS measurements of well-dissolved proteins of (c) 64 μM PGIS and (d) 176 μM TXAS. The HPLC elution rate control was adopted for SAXS data collection time and further verified along IUV (280/420 nm wavelength). The inset shows the linear fitting via the Guinier approximation, which reveals the sizes of the monomer structures of PGIS and TXAS, respectively. Also, (e) the Kratky−Porod plot reveals the protein folded well as the globular protein. (f) Comparison of the pair-distance p(r) plots indicates the asymmetric distribution of the TXAS structure as well as the suggestion for the homology results.

thermophilic P450, CYP119.45 The structure of this P450 has been solved with different types of inhibitors bound to the heme iron.46 A rather dramatic unfolding of the F helix and large movements of the F/G loop were observed, further supporting the view that the F/G loop and associated elements of the structure can shape themselves around active site ligands. Although quite speculative, it is intriguing to consider the possibility of such a different reshaping of the active site and internal channels taking place in PGIS and TXAS via changes in those resulting F/G and H/I loops.12 Therefore, the main objective of this study was to validate this simulation-derived TXAS structure with those connection loop structures and the protein fold and conformational flexibility, as compared to its counter protein of PGIS. It is of great importance to elucidate the connections of these structural features to the mechanism at the molecular level, which underlies the versatile but specific PGH2 metabolic capacities of TXAS and PGIS. Experimental Validation of the PGIS and TXAS Structures in Solution. In Figure 2a, we show the elution SAXS profile of PGIS compared with a simple static measurement and prove that the scanning HPLC−SAXS approach can effectively limit radiation damage via flowing of the sample during data collection. Also, Figure 2b elaborates the actual cumulative time SAXS profiles of individual monomer protein structures and reveals that the intense Xray beam indeed causes radiation damage to the protein sample. Figure 2c,d shows the zero-angle intensity I(0) and the radius

observed in TXAS, in the form of breaks of helices F, G, H, I and the significant long loop with ∼42 residues, which connects helices H and I, as compared with ∼5 residues forming the loop structure in CYP3A4 (PDB: 1TQN).41 A combination of homology modeling based on the known homologous P450s structures and explicit water molecular dynamics studies was used to optimize the resulting homology structure, including its backbone conformations, the residue interactions, contact, and local environments target for an acceptable TXAS structure.8 Nevertheless, the unexpected long disordered loop feature in TXAS led us to address how this unstructured loop may affect the relevant conformational transitions and structural reshaping in solution. Previous knowledge of the loop motions in P450s was mostly derived from the bacterial fatty acid mono-oxygenase P450 BM342 and, more recently, P450 3A4,32 cam,33 and 2D6.43 The substrate-bound and substrate-free crystal structures of P450s show that the F and G helices and the connecting loop (Figure 1) are flexible and undergo the slide conformational change over the I and L helices. Such a conformational change leads to the open−close motion that permits substrates to enter and products to leave.12 It might be anticipated that TXAS and PGIS will experience such motion. Moreover, in some P450 crystallographic structures, the electron density for the F/G loop is visible only in the substrate-bound form,44 suggesting that the F/G loop may be able to shape itself around the substrate. The one clear example of such reshaping is in the 11233

DOI: 10.1021/acs.jpcb.7b08299 J. Phys. Chem. B 2017, 121, 11229−11240

Article

The Journal of Physical Chemistry B

Figure 3. Representative SAXS data of protein at high and low concentrations. The data are, respectively, fitted using the CRYSOL (dash curves) models. (a) The PGIS merged data (64 μM merged with 316 μM, colored in purple) are fitted with PDB 3B99. (b) The TXAS merged data (176 μM merged 234 μM, colored in deep blue) are fitted with the homology structure. (c) Results of the model fitted by an elliptical cylinder (yellow line with blue bead) indicate that the H/I loop of TXAS lies along the major axis and the results of the larger aspect ratio. Corresponding GABOR envelopes (blue mesh) and average shape are shown in the gray surface. Superimposing the average shape with MD structure (color ribbon) shows the high structural similarity of PGIS and TXAS.

Table 1. Comparison of the Form Factor Characteristics of PGIS and TXAS Proteins in Experiment Structures and Simulation Structures PGIS expt Rg (Å) minor (Å) major (Å) n/ma length (Å)

25.1 23.5 38.5 1.6 44.3

± ± ± ± ±

0.1 0.4 2.4 0.1 0.6

TXAS

MD-derived

crystal (2IAG/3B99)

24.6 ± 0.1 24.0−25.5 34.0−35.0 1.3−1.4 37.0−39.0

23.4/23.8 26.4/26.0 35.6/33.8 1.4/1.3 40.0/35.5

expt 32.2 26.5 52.8 2.0 41.9

± ± ± ± ±

0.9 3.3 1.2 0.2 4.4

MD-derived

homology

28.8 ± 0.1 (29.8 ± 0.1)b 24.0−26.0 39.0−42.0 1.6−1.7 37.0−40.0

24.4 25.9 36.9 1.4 37.2

a n/m is the major/minor axis of the elliptical cylinder model. bConsider the oligomer−TXAS cobinding structures (with oligomer length of 10 residues).

using the relation qcRg ∼ 31/2.48 The similar shape profile with a relative center shift to qc ∼ 0.055 A−1 for the TXAS solution implies the extension structure, with an estimated Rg of ∼32 Å. The Kratky plots exhibit that both PGIS and TXAS retained the globular structure rather than the unfolded chain under the experimental conditions. This finding supports the presence of natively extended TXAS conformations. Also, the pair-distance functions p(r) Fourier transformed from the SAXS data in Figure 2f reveal a maximal-distance Dmax = 100 Å for TAXS, larger than that of PGIS of 70 Å, implying a more extended conformation of TXAS compared to the globular structure of PGIS. Apparently, with simultaneous scanning SAXS and UV− vis absorption,30 we can capture the monomer protein structure

of gyration Rg values, extracted using the Guinier approximation47 from the SAXS data measured along the HPLC elution path for PGIS and TXAS, respectively. These two enzymes were purified to homogeneity as 56.63 and 59.02 kDa hemoproteins, respectively, with spectroscopic characteristics of dual absorption at 280 and 420 nm. Although the TXAS sample revealed two minor peaks with the aggregated proteins (left side) and fragments (right side), the elution allows an SAXS objective measure for the major peak region of protein monomers. Figure 2e illustrates the scanning HPLC−SAXS data in the Kratky−Porod representation, in which the bellshaped profile centered at qc, ∼0.065 A−1 for the PGIS solution, reveals a globular structure with an estimated Rg of ∼25 Å by 11234

DOI: 10.1021/acs.jpcb.7b08299 J. Phys. Chem. B 2017, 121, 11229−11240

Article

The Journal of Physical Chemistry B

Figure 4. Comparison of the root-mean-square fluctuations (RMSF) of the amino acid residues for PGIS and TXAS for describing the hydrodynamic conformational transitions, as the all-atom MD-snapshot overlay displaying the protein internal motions compared to the GASBOR envelopes.

models were generated by GASBOR; 25 models were selected with the χ2 value equal to or less than 1, Figure S8, and these chain models had very similar characteristics with an average normalized spatial discrepancy (NSD) value of 0.95 for PGIS, compared with the high NSD value of 1.51 for TXAS to support the significant loop flexibility. Figure 3c presents the crystallographic structure of PGIS which overlaps decently with the corresponding GASBOR envelope with an NSD value of ∼1.5 (see Figure S9 for the top view). It was found that the GASBOR fitted structures of TXAS agree well with the homology MD structures, with the overlapped rigid core, the small flexible loop (∼15 residues in the E/F loop and 4 His-tags attached C-terminal), and the significantly H/I flexible subdomain coil (∼42 residues). These loop motions reflect not only the protein reshaping in water but also the internal channel regulatory interface to the buried heme site (see Figure 3c). The yield interface difference is associated with a change in the substrate accessible volume of around 430 Å3 for TXAS, varying from 1220 to 1650 Å3, as compared to 70 Å3 for PGIS, varying from 1160 to 1230 Å3 (according to the modeling software,51,52 cf., Figure 3c and Figure S10). Moreover, water molecules were found to reside within the identified channels (see Figure S11). Molecular dynamics simulations in a water box were sufficient to observe the movement of water molecules in the channels and the exchange with the heme site (see Figure S12). This supports the finding that the protein−water interactions via the channels contribute a strong component to the protein flexibility in solution, such that the channel volume change reports largely on variations of the TXAS protein structures (Figure 3c). Because proteins are inherently flexible systems displaying a broad range of different types of conformations in solution, strictly speaking, the MD trajectory ensemble provides a means to increase the accuracy of configuration space for determining protein conformational variations and the assessment of experimental-solution SAXS results. Several pioneering studies53−56 have addressed that χ2 and NSD could be used to assess models of conformational diversity in X-ray crystallography comparing to solution scattering structures. Some of these studies used ensemble optimization method (EOM),57 minimal ensemble search (MES),58 and basis-set supported SAXS (BSS-SAXS)59 to experimentally validate predictions of

with in situ shape information on the objective proteins of PGIS and TXAS under the native solution environment. In all these circumstances, SAXS data revealed rather large Rg values and expanded p(r) functions with larger Dmax values for TXAS than PGIS, which is consistent with the analyses of the PGIS crystal structures and the TXAS homology structures. Figure 3 illustrates the scanning HPLC−SAXS data of PGIS (3a) and TXAS (3b) for merged concentrations, and modeling49 of experimental SAXS profiles allows for protein shape reconstruction as an illustration of the target PGIS and TXAS protein structures described by the structures clustered visually into elliptical cylinder models (3c). Table 1 elaborates those experiment-derived characteristics, for which the long axes are about 50 and 38 Å for TXAS and PGIS, respectively. For the form factor comparison of the experimental, crystallographic, and MD simulation structures, the relatively larger aspect ratio of TXAS compared to that of PGIS, i.e., 2.0 to 1.6, indicates the larger major axis, which may come from the flexible loop structure of TXAS. Moreover, for PGIS, the similarity of both the dynamical structure in solution and the crystal form, with only 5% length variation between the axis ratio of major/minor, indicates that the PGIS protein has a conserved-chain folding structure in solution, as evidenced by X-ray crystallography. This reflects the stability of PGIS in solution with the observation of fast/easy crystallization behavior in PGIS protein. On the other hand, the axis ratio variation (ca. 23%) of the TXAS protein strongly indicates that the flexibility of dynamical structures varies between the homology and MD simulation structures. The agreement is substantial indicating there is room for improvement in modeling the anisotropic conformational distribution, such as consideration of the dynamic contribution from the flexible loop motions in solution. Functional Loop Motions Coupled to the Channel Interface Change. To reveal the corresponding structural changes of the proteins in solution, we analyzed the SAXS data to obtain the protein envelopes of the samples using ab initio modeling of GASBOR. This method allowed the construction of a low-resolution structure from the experimental SAXS profile to validate the model against the available higherresolution structures from complementary methods such as Xray crystallography and MD structures.50 More than 100 11235

DOI: 10.1021/acs.jpcb.7b08299 J. Phys. Chem. B 2017, 121, 11229−11240

Article

The Journal of Physical Chemistry B

Figure 5. Comparison of (a) 2D-RMSD of PGIS-GH loop and TXAS-HI-loop motions. The root-mean-square deviation of every conformation to all other conformations as a function of time of the specific loop atoms during a 100 ns MD trajectory in PGIS and TXAS, consisting of (b) several metastable states as the snapshots overlay displaying the conformation cluster into flexible regions whose loop motions are highly correlated, and (c) then gradually adding sequential time points to the scatter plot in the space of the dominant components for PGIS and TXAS, respectively.

loop). For a quantitative comparison, these diffusive coordinate motions can be described as cooperative subdiffusion originating from the loops connecting these helices, as summarized in Figure S2, which displays the diffusion coefficient analysis of those connection loop motions. The protein structures are dictated by these internal motions rather than overall rotations and translations of the protein. To gain insight into the functional significance of the states, the snapshot overlay was created to visualize the loop motions along the trajectory (Figure 4). Especially prominent is a follow-up hinge motion of the HI-loop against the core region in TXAS, as illustrated in Figure 5: the conformation cluster in flexible regions whose motions are highly correlated, consisting of at least six metastable states as the snapshot overlay displaying the corresponding loop motions. The large-scale molecular organization apparently exhibits an association with the disordered H/I loop domain, with coil-to-stretch motion passing through the solution, where the rotation of peptide bonds involves a collective reorientation of the interconnected H/I loop forming the period motions of the TXAS flexibility. These motions would likely modulate the diffusive coordinate interactions of water molecules putatively involved in the protein internal channel hydration (see Figure 3c and Figure S12), and an opening of helix G away from helix F.

correlated motions from molecular dynamics (MD) simulations. Notably, recent work60 demonstrated that an energy restraint ensemble on the coarse-graining (CG) models is necessary to avoid overinterpretation of experimental SAXS data by spurious conformational representations. Naturally, the protein structure dictates the features of internal loop motions collectively on the nanosecond time scale.24,25,61 In MD simulations, periodic boundary conditions and long-range hydrodynamic interactions can introduce an effective coupling between the protein and explicit water solvent, and their diffusive coordinate motions. In the present study, the explicit water-coupled protein dynamics with the ∼1 μs trajectory demonstrates a rich conformational landscape of adequate sampling for the validation of MD conformations of solution Xray scattering. To analyze the trajectory, we performed conformation component analysis (RMSF and 2D-RMSD, Supporting Information methods details) to reveal insights into the origin of structure flexibility and correlated motions in proteins. Figure 4 demonstrates the root-mean-square fluctuation (RMSF) analysis for the MD ensemble of 100 ns trajectory, revealing the MD conformations involved in substantial motions in the loop regions (by noting of TXAS HI loop region comparing to PGIS with little motion in region of GH11236

DOI: 10.1021/acs.jpcb.7b08299 J. Phys. Chem. B 2017, 121, 11229−11240

Article

The Journal of Physical Chemistry B

Table 2. Fitting Quality of Experimental-Solution SAXS with MD Ensemble Structures from Metastable States by Noting of Loop Motions that Occurred between the Structures at Different RMSD Levels merged conc MD ensemblea (Δt)

RMSD (Å−1)

low conc χ2 q = 0.01−0.28 Å−1

χ2 q = 0.01−0.3 Å−1

χ2 q = 0.04−0.3 Å−1

PGIS 5 ns 13 ns 20 ns 25 ns 73 ns 92 ns av cryst structb 25 ns 39 ns 53 ns 63 ns 87 ns 100 ns av homology struct

2.92 3.25 3.10 3.35 4.43 3.49

8.63 10.78 13.12 14.68 16.04 15.47

2.73 2.63 2.38 2.51 2.52 2.72 2.60 2.62 5.31 5.00 6.08 5.01 4.29 4.73 5.03 5.73

6.19 5.92 4.89 5.33 5.28 6.31 5.57 5.95 TXASc (2.11) (2.16) (2.70) (2.48) (1.96) (1.97) (2.17)

9.25 9.72 12.77 11.40 7.57 8.15 9.39 27.07

4.30 4.15 3.35 3.63 3.58 4.54 3.80 4.13 (2.76) (2.97) (4.81) (4.23) (2.61) (2.59) (3.09)

5.74 6.22 8.09 7.46 4.75 5.08 5.78 15.72

(1.61) (1.79) (2.97) (2.94) (1.77) (1.75) (1.89)

a The constructed MD ensemble contains the metastable states along the elaborated 100 ns trajectory extracted from the ∼1 μs simulation; regarding different scattering regions and concentrations, χ2 evaluates the scatter for each of the metastable state in the space of 10 conformations in nanosecond resolution to experimental-solution SAXS. bCrystal structure of PGIS (PDB ID: 2IAG). cConsider both native TXAS structures and oligomer−TXAS cobinding structures (with oligomer length of 10 residues).

Experimental Validation of Protein Flexibility in Solution. To assess the reproducibility of X-ray scattering calculations, we gradually added sequential time points to a scatter profile in the space of the time-dependent loop motion components. We identified at least six metastable states by noting when loop motions occurred between the structures (Figure 5). The time between the transitions varies considerably from 25 ns (Figure 5b, orange track) to 100 ns (Figure 5b, yellow track); notably, these times are longer than the 10 ns duration of the previous MD simulation of TXAS.9 The 53 ns track is visited twice: the first time for 53 ns, and the second time for 184 ns (see Figure S13). The loop motion spends nearly the duration of about 100 ns; visualization in three dimensions revealed that this loop motion cycle has fine structures of substates that lie close together in the space of the dominant components. Therefore, the experimental SAXS profile of a given protein is assumed to be derived from a number of coexisting conformational states in solution, as shown in Figure 5. These ensemble conformations can yield different correlations in structure variations and yield the different χ2 to the solution scatter profile. Table 2 elaborates the values of derived chi-square (χ2) for the all-atom crystallographic structure/homology/MD-derived-SAXS scatters to the experimental-solution SAXS results for PGIS (left panel of Figure 5c) and TXAS (right panel of Figure 5c), using the CRYSOL module. In comparison to the χ2 values of ca. 5.95 and 27.07 for PGIS crystallography and TXAS homology structures, those χ2 values obtained from the MD trajectory subensemble show a substantial agreement with the solutionmeasured SAXS, ca. 4.9−6.3 and 7.6−12.8 for PGIS and TXAS. Because the statistic χ2 value describes the global goodness-offit of the atomic resolution conformation to the experimentalsolution SAXS data, it is possible to obtain a range of χ2 values which depend on the sample conditions and measurement

error (noisy) as well as the reliability of modeling conformations. Our SAXS data show the related low errors (∼3% noise) that may overestimate the χ2 value, as pointed out in the previous studies,53,54,62 in comparison to the evaluation at low concentrations (∼10% noise). Moreover, Table 2 illustrates the different statistical region fitting of the MD ensemble scattering to the solution SAXS, which makes it clear to see that the greater χ2 value mostly comes from the low q deviations either in PGIS or TXAS. This says that our protein solution samples may undergo exposure to synchrotron radiation damage (see Figure 2b) or binding with some residue oligomers. This also reflects on the corresponding Rg value larger than the MD-derived value (see Table 1) since the protein size, radiation damage, and impurity are inherently linked. In particular, for the TXAS measure, the discrepancy is supposed to be coming from the protein instability and flexibility in solution as manifested above. Nevertheless, most important of all is the resolution of the MD-based trajectory, which helps us assess the experimental scatter reproducibility by the reasonable conformations for each χ2 value about the origin. Of course, an ensemble of 10−100 ns resolved structures seems far too small to explain the conformational behavior of a flexible protein in solution, but these subpopulations appear sufficient to describe the protein hydrodynamic scattering, as revealed in Figure 5c, where specific scattering contributions from the main core and disordered loop dynamical structures are observed. Implication of Protein Flexibility to Catalysis. Such protein internal diffusion motion supports the explanation of a clear dependence of substructure loop diffusion on the environment of the construction. Both resting proteins of PGIS and TXAS were observed to experience the open−close motion around the F and G helices over the I helix, and trajectory analyses illustrated the rather wide ranges of motion 11237

DOI: 10.1021/acs.jpcb.7b08299 J. Phys. Chem. B 2017, 121, 11229−11240

Article

The Journal of Physical Chemistry B available to TXAS via the cooperation of significant H/I loop movements. Such motions for the helices and loops were observed to be able to reshape the internal substructures of the heme pocket and channels for the resting, substrate-bound, or membrane-bound states. A rather dramatic movement of the H/I loop was observed in TXAS, with breaks in the F, G, H, and I helices providing flexibility that opens up a new access channel to the heme pocket (see Figure S11). In effect, such a wide-open conformation may nicely illustrate a rather flexible PGH2 bound to the heme iron in TXAS. This prevailing view further supports the view that these associated structural elements can reshape themselves around active site ligands, as introduced above. TXAS catalyzes the isomerization reaction of PGH2, yielding not only TXA2, but also products of HHT and MDA. This is in sharp contrast to PGIS, which presents a high product fidelity of isomerization of PGH2 into prostacyclin PGI2.19 As revealed in this study, in the rigorous protein structure in solution, the connecting segment between the F, G, H, and I helices in PGIS is ordered and helical with relatively slow coordinate motion (Figures 4 and 5). Large conformational changes are due primarily to protein internal motions of the connection loops between F, G, H, and I helices. The current TXAS/PGIS structures very likely represent the extreme paradigm of protein internal motion for flexibility to explain how P450s adapt to accommodate different isomerizations of PGH2, given the restriction that the same fold is maintained in all P450s.



motions, as observed in TXAS. Also these motions reflect the change of the internal channel regulatory interface, which may furnish different open−close stages for substrate PGH2 binding. 5. The development of realistic ensemble models of unstructured states of proteins has been an important subject of research for many years. In this study, this challenge has been addressed by the use of ensembles of reliable MD conformations to deconvolute the experimental SAXS data, which represent average values for the MD trajectory ensemble of configuration space.

ASSOCIATED CONTENT

S Supporting Information *

The Supporting Information is available free of charge on the ACS Publications website at DOI: 10.1021/acs.jpcb.7b08299. Computational methodology, growth and induction of bacteria and X-ray scattering profile analyses (PDF)



AUTHOR INFORMATION

Corresponding Author

*E-mail: [email protected]. Phone: +886-2-29052534. ORCID

Hsiao-Ching Yang: 0000-0002-7970-1357 Yi-Qi Yeh: 0000-0002-2387-3600 U-Ser Jeng: 0000-0002-2247-5061



CONCLUSION In summary, scientists have expended tremendous efforts to probe protein hydrodynamics to develop understanding of the underlying mechanisms in activated biofunctionality and the relevant enzymatic reactions. In this study, we have demonstrated that HPLC−SAXS/UV−vis combined with explicit water MD simulations is a promising way to probe the in situ structures and dynamical behaviors of the counterfunctioning proteins of TXAS and PGIS in solution. Such an approach reveals the reliable structure of TXAS, even in the absence of a crystal structure. Several points are highlighted as follows. 1. Both target proteins have inherent flexibility in solution. In particular, TXAS displays a broad range of conformations in solution; therefore, a single model cannot represent the intrinsically flexible protein, and alternative MD trajectories are required. 2. PGIS exhibits the highly conserved P450 folding feature, which contributes to its stability in solution as compared to that of the large disordered loop structure observed in TXAS, which contributes to its dramatic flexibility with structural transitions in solution. Also, the behaviors reflect the SAXS-resolved cylinder shape characteristics, with the major axis of TXAS, ∼50 Å, being larger than that of PGIS, ∼38 Å. 3. Specific contributions from the main core and disordered loop dynamical structures were noted. The SAXS profile of TXAS suggests coexisting conformational states in solution, as also evidenced in MD trajectory ensemble analyses. 4. Most importantly, proteins feature a large spectrum of motional modes in solution, encompassing fluctuations around the core structure and large-scale molecular reorganizations via inherent disordered flexible loop

Notes

The authors declare no competing financial interest.



ACKNOWLEDGMENTS This work was supported by Grant MOST 105-2119-M-030002-MY2 from the Ministry of Science and Technology of Taiwan. We also thank the BL23A1 National Synchrotron Radiation Research Center of Taiwan for providing beam time and technical support.



REFERENCES

(1) Papaleo, E.; Saladino, G.; Lambrughi, M.; Lindorff-Larsen, K.; Gervasio, F. L.; Nussinov, R. The role of protein loops and linkers in conformational dynamics and allostery. Chem. Rev. 2016, 116 (11), 6391−6423. (2) Cho, H. S.; Dashdorj, N.; Schotte, F.; Graber, T.; Henning, R.; Anfinrud, P. Protein structural dynamics in solution unveiled via 100ps time-resolved X-ray scattering. Proc. Natl. Acad. Sci. U. S. A. 2010, 107 (16), 7281−7286. (3) Putnam, C. D.; Hammel, M.; Hura, G. L.; Tainer, J. A. X-ray solution scattering (SAXS) combined with crystallography and computation: defining accurate macromolecular structures, conformations and assemblies in solution. Q. Rev. Biophys. 2007, 40 (3), 191− 285. (4) Caffrey, M. Membrane protein crystallization. J. Struct. Biol. 2003, 142 (1), 108−132. (5) Meisburger, S. P.; Thomas, W. C.; Watkins, M. B.; Ando, N. Xray scattering studies of protein structural dynamics. Chem. Rev. 2017, 117 (12), 7615−7672. (6) Carpenter, E. P.; Beis, K.; Cameron, A. D.; Iwata, S. Overcoming the challenges of membrane protein crystallography. Curr. Opin. Struct. Biol. 2008, 18 (5), 581−586. (7) Chao, W. C.; Lu, J. F.; Wang, J. S.; Yang, H. C.; Chen, H. H.; Lan, Y. K.; Yu, Y. C.; Chou, P. T.; Wang, L. H. Probing the interaction between prostacyclin synthase and prostaglandin H2 analogues or inhibitors via a combination of resonance Raman spectroscopy and 11238

DOI: 10.1021/acs.jpcb.7b08299 J. Phys. Chem. B 2017, 121, 11229−11240

Article

The Journal of Physical Chemistry B molecular dynamics simulation approaches. J. Am. Chem. Soc. 2011, 133 (46), 18870−9. (8) Chao, W.-C.; Lu, J.-F.; Wang, J.-S.; Yang, H.-C.; Pan, T.-A.; Chou, S. C.-W.; Wang, L.-H.; Chou, P.-T. Probing ligand binding to thromboxane synthase. Biochemistry 2013, 52 (6), 1113−1121. (9) Shen, J.-Y.; Chao, W.-C.; Liu, C.; Pan, H.-A.; Yang, H.-C.; Chen, C.-L.; Lan, Y.-K.; Lin, L.-J.; Wang, J.-S.; Lu, J.-F.; et al. Probing water micro-solvation in proteins by water catalysed proton-transfer tautomerism. Nat. Commun. 2013, 4, 2611. (10) Coleman, R. A.; Humphrey, P. P.; Kennedy, I.; Levy, G. P.; Lumley, P. Comparison of the actions of U-46619, a prostaglandin H2analogue, with those of prostaglandin H2 and thromboxane A2 on some isolated smooth muscle preparations. Br. J. Pharmacol. 1981, 73 (3), 773−778. (11) Nelson, D. R. The Cytochrome P450 homepage. Hum. Genomics 2009, 4 (1), 59−65. (12) Poulos, T. L. Cytochrome P450 flexibility. Proc. Natl. Acad. Sci. U. S. A. 2003, 100 (23), 13121−13122. (13) Ortiz de Montellano, P. R. Hydrocarbon hydroxylation by cytochrome P450 enzymes. Chem. Rev. 2010, 110 (2), 932. (14) Collins, P. W.; Djuric, S. W. Synthesis of therapeutically useful prostaglandin and prostacyclin analogs. Chem. Rev. 1993, 93 (4), 1533−1564. (15) Magri, D. C.; Workentin, M. S. A radical-anion chain mechanism following dissociative electron transfer reduction of the model prostaglandin endoperoxide, 1,4-diphenyl-2,3dioxabicyclo[2.2.1]heptane. Org. Biomol. Chem. 2008, 6 (18), 3354− 3361. (16) Hecker, M.; Ullrich, V. On the mechanism of prostacyclin and thromboxane A2 biosynthesis. J. Biol. Chem. 1989, 264 (1), 141−50. (17) Pradono, P.; Tazawa, R.; Maemondo, M.; Tanaka, M.; Usui, K.; Saijo, Y.; Hagiwara, K.; Nukiwa, T. Gene transfer of thromboxane A2 synthase and prostaglandin I2 synthase antithetically altered tumor angiogenesis and tumor growth. Cancer Res. 2002, 62 (1), 63−66. (18) Golino, P.; Rosolowsky, M.; Yao, S. K.; McNatt, J.; De Clerck, F.; Buja, L. M.; Willerson, J. T. Endogenous prostaglandin endoperoxides and prostacyclin modulate the thrombolytic activity of tissue plasminogen activator. Effects of simultaneous inhibition of thromboxane A2 synthase and blockade of thromboxane A2/ prostaglandin H2 receptors in a canine model of coronary thrombosis. J. Clin. Invest. 1990, 86 (4), 1095−1102. (19) Funk, C. D. Prostaglandins and leukotrienes: advances in eicosanoid biology. Science 2001, 294 (5548), 1871−1875. (20) Vanhoutte, P. M.; Shimokawa, H.; Feletou, M.; Tang, E. H. C. Endothelial dysfunction and vascular disease − a 30th anniversary update. Acta Physiol. 2017, 219 (1), 22−96. (21) Young, T.; Abel, R.; Kim, B.; Berne, B. J.; Friesner, R. A. Motifs for molecular recognition exploiting hydrophobic enclosure in proteinligand binding. Proc. Natl. Acad. Sci. U. S. A. 2007, 104 (3), 808−813. (22) Ball, P. Water as an active constituent in cell biology. Chem. Rev. 2008, 108 (1), 74−108. (23) Snyder, P. W.; Mecinović, J.; Moustakas, D. T.; Thomas, S. W.; Harder, M.; Mack, E. T.; Lockett, M. R.; Héroux, A.; Sherman, W.; Whitesides, G. M. Mechanism of the hydrophobic effect in the biomolecular recognition of arylsulfonamides by carbonic anhydrase. Proc. Natl. Acad. Sci. U. S. A. 2011, 108 (44), 17889−17894. (24) Chao, W.-C.; Shen, J.-Y.; Yang, C.-H.; Lan, Y.-K.; Yuan, J.-H.; Lin, L.-J.; Yang, H.-C.; Lu, J.-F.; Wang, J.-S.; Wee, K.; et al. The in situ tryptophan analogue probes the conformational dynamics in asparaginase isozymes. Biophys. J. 2016, 110 (8), 1732−1743. (25) Chao, W.-C.; Shen, J.-Y.; Lu, J.-F.; Wang, J.-S.; Yang, H.-C.; Wee, K.; Lin, L.-J.; Kuo, Y.-C.; Yang, C.-H.; Weng, S.-H.; et al. Probing water environment of trp59 in ribonuclease T1: insight of the structure−water network relationship. J. Phys. Chem. B 2015, 119 (6), 2157−2167. (26) Boldon, L.; Laliberte, F.; Liu, L. Review of the fundamental theories behind small angle X-ray scattering, molecular dynamics simulations, and relevant integrated application. Nano Rev. 2015, 6, 25661.

(27) Carugo, O.; Carugo, K. D. When X-rays modify the protein structure: radiation damage at work. Trends Biochem. Sci. 2005, 30 (4), 213−219. (28) Pabst, G.; Kučerka, N.; Nieh, M. P.; Rheinstädter, M. C.; Katsaras, J. Applications of neutron and X-ray scattering to the study of biologically relevant model membranes. Chem. Phys. Lipids 2010, 163 (6), 460−479. (29) Klauda, J. B.; Kučerka, N.; Brooks, B. R.; Pastor, R. W.; Nagle, J. F. Simulation-based methods for interpreting X-ray data from lipid bilayers. Biophys. J. 2006, 90 (8), 2796−2807. (30) Yeh, Y.-Q.; Liao, K.-F.; Shih, O.; Shiu, Y.-J.; Wu, W.-R.; Su, C.-J.; Lin, P.-C.; Jeng, U. S. Probing the acid-induced packing structure changes of the molten globule domains of a protein near equilibrium unfolding. J. Phys. Chem. Lett. 2017, 8 (2), 470−477. (31) Brookes, E.; Vachette, P.; Rocco, M.; Pérez, J. US-SOMO HPLC-SAXS module: dealing with capillary fouling and extraction of pure component patterns from poorly resolved SEC-SAXS data. J. Appl. Crystallogr. 2016, 49 (5), 1827−1841. (32) Yano, J. K.; Wester, M. R.; Schoch, G. A.; Griffin, K. J.; Stout, C. D.; Johnson, E. F. The structure of human microsomal cytochrome P450 3A4 determined by X-ray crystallography to 2.05-Å resolution. J. Biol. Chem. 2004, 279 (37), 38091−38094. (33) Meilleur, F.; Dauvergne, M.-T.; Schlichting, I.; Myles, D. A. Production and X-ray crystallographic analysis of fully deuterated cytochrome P450cam. Acta Crystallogr., Sect. D: Biol. Crystallogr. 2005, 61 (5), 539−544. (34) Case, D. A.; Betz, R. M.; Cerutti, D. S.; Cheatham, T. E., III; Darden, T. A.; Duke, R. E.; Giese, T. J.; Gohlke, H.; Homeyer, N.; et al. AMBER 2015; University of California: San Francisco, 2015. (35) Chiang, C.-W.; Yeh, H.-C.; Wang, L.-H.; Chan, N.-L. Crystal structure of the human prostacyclin synthase. J. Mol. Biol. 2006, 364 (3), 266−274. (36) Li, Y.-C.; Chiang, C.-W.; Yeh, H.-C.; Hsu, P.-Y.; Whitby, F. G.; Wang, L.-H.; Chan, N.-L. Structures of prostacyclin synthase and its complexes with substrate analog and inhibitor reveal a ligand-specific heme conformation change. J. Biol. Chem. 2008, 283, 2917−2926. (37) Maier, J. A.; Martinez, C.; Kasavajhala, K.; Wickstrom, L.; Hauser, K. E.; Simmerling, C. ff14SB: Improving the accuracy of protein side chain and backbone parameters from ff99SB. J. Chem. Theory Comput. 2015, 11 (8), 3696−3713. (38) Wang, J.; Wolf, R. M.; Caldwell, J. W.; Kollman, P. A.; Case, D. A. Development and testing of a general amber force field. J. Comput. Chem. 2004, 25 (9), 1157−1174. (39) Horn, H. W.; Swope, W. C.; Pitera, J. W.; Madura, J. D.; Dick, T. J.; Hura, G. L.; Head-Gordon, T. Development of an improved four-site water model for biomolecular simulations: TIP4P-Ew. J. Chem. Phys. 2004, 120 (20), 9665−9678. (40) Franke, D.; Petoukhov, M. V.; Konarev, P. V.; Panjkovich, A.; Tuukkanen, A.; Mertens, H. D. T.; Kikhney, A. G.; Hajizadeh, N. R.; Franklin, J. M.; et al. ATSAS 2.8: a comprehensive data analysis suite for small-angle scattering from macromolecular solutions. J. Appl. Crystallogr. 2017, 50 (4), 1212−1225. (41) Williams, P. A.; Cosme, J.; Vinković, D. M.; Ward, A.; Angove, H. C.; Day, P. J.; Vonrhein, C.; Tickle, I. J.; Jhoti, H. Crystal structures of human cytochrome P450 3A4 bound to metyrapone and progesterone. Science 2004, 305 (5684), 683−686. (42) Girvan, H. M.; Seward, H. E.; Toogood, H. S.; Cheesman, M. R.; Leys, D.; Munro, A. W. Structural and spectroscopic characterization of P450 BM3 mutants with unprecedented P450 heme iron ligand sets: new heme ligation states influence conformational equilibria in P450 BM3. J. Biol. Chem. 2007, 282 (1), 564−572. (43) Rowland, P.; Blaney, F. E.; Smyth, M. G.; Jones, J. J.; Leydon, V. R.; Oxbrow, A. K.; Lewis, C. J.; Tennant, M. G.; Modi, S.; Eggleston, D. S.; et al. Crystal structure of human cytochrome P450 2D6. J. Biol. Chem. 2006, 281 (11), 7614−7622. (44) Wester, M. R.; Johnson, E. F.; Marques-Soares, C.; Dijols, S.; Dansette, P. M.; Mansuy, D.; Stout, C. D. Structure of mammalian cytochrome P450 2C5 complexed with diclofenac at 2.1 Å resolution: 11239

DOI: 10.1021/acs.jpcb.7b08299 J. Phys. Chem. B 2017, 121, 11229−11240

Article

The Journal of Physical Chemistry B evidence for an induced fit model of substrate binding. Biochemistry 2003, 42 (31), 9335−9345. (45) Park, S.-Y.; Yamane, K.; Adachi, S.-i.; Shiro, Y.; Weiss, K. E.; Maves, S. A.; Sligar, S. G. Thermophilic cytochrome P450 (CYP119) from sulfolobus solfataricus: high resolution structure and functional properties. J. Inorg. Biochem. 2002, 91 (4), 491−501. (46) Yano, J. K.; Koo, L. S.; Schuller, D. J.; Li, H.; Ortiz de Montellano, P. R.; Poulos, T. L. Crystal Structure of a Thermophilic cytochrome P450 from the archaeon sulfolobus solfataricus. J. Biol. Chem. 2000, 275 (40), 31086−31092. (47) Jeffries, C. M.; Graewert, M. A.; Blanchet, C. E.; Langley, D. B.; Whitten, A. E.; Svergun, D. I. Preparing monodisperse macromolecular samples for successful biological small-angle X-ray and neutron scattering experiments. Nat. Protoc. 2016, 11 (11), 2122−2153. (48) Receveur-Bréchot, V.; Durand, D. How random are intrinsically disordered proteins? a small angle scattering perspective. Curr. Protein Pept. Sci. 2012, 13 (1), 55−75. (49) Kline, S. Reduction and analysis of SANS and USANS data using IGOR Pro. J. Appl. Crystallogr. 2006, 39 (6), 895−900. (50) Koch, M. H. J.; Vachette, P.; Svergun, D. I. Small-angle scattering: a view on the properties, structures and structural changes of biological macromolecules in solution. Q. Rev. Biophys. 2003, 36 (2), 147−227. (51) Sehnal, D.; Svobodová Vařeková, R.; Berka, K.; Pravda, L.; Navrátilová, V.; Banás,̌ P.; Ionescu, C.-M.; Otyepka, M.; Koča, J. MOLE 2.0: advanced approach for analysis of biomacromolecular channels. J. Cheminf. 2013, 5 (1), 39. (52) Kozlikova, B.; Sebestova, E.; Sustr, V.; Brezovsky, J.; Strnad, O.; Daniel, L.; Bednar, D.; Pavelka, A.; Manak, M.; Bezdeka, M.; et al. CAVER Analyst 1.0: graphic tool for interactive visualization and analysis of tunnels and channels in protein structures. Bioinformatics 2014, 30 (18), 2684−2685. (53) Rambo, R. P.; Tainer, J. A. Accurate assessment of mass, models and resolution by small-angle scattering. Nature 2013, 496 (7446), 477−481. (54) Svergun, D.; Barberato, C.; Koch, M. H. J. CRYSOL - a program to evaluate X-ray solution scattering of biological macromolecules from atomic coordinates. J. Appl. Crystallogr. 1995, 28 (6), 768−773. (55) Svergun, D. I.; Petoukhov, M. V.; Koch, M. H. Determination of domain structure of proteins from X-ray solution scattering. Biophys. J. 2001, 80 (6), 2946−2953. (56) Wall, M. E.; Van Benschoten, A. H.; Sauter, N. K.; Adams, P. D.; Fraser, J. S.; Terwilliger, T. C. Conformational dynamics of a crystalline protein from microsecond-scale molecular dynamics simulations and diffuse X-ray scattering. Proc. Natl. Acad. Sci. U. S. A. 2014, 111 (50), 17887−17892. (57) Tria, G.; Mertens, H. D. T.; Kachala, M.; Svergun, D. I. Advanced ensemble modelling of flexible macromolecules using X-ray solution scattering. IUCrJ 2015, 2 (2), 207−217. (58) Pelikan, M.; Hura, G. L.; Hammel, M. Structure and flexibility within proteins as identified through small angle X-ray scattering. Gen. Physiol. Biophys. 2009, 28 (2), 174−189. (59) Yang, S.; Blachowicz, L.; Makowski, L.; Roux, B. Multidomain assembled states of Hck tyrosine kinase in solution. Proc. Natl. Acad. Sci. U. S. A. 2010, 107 (36), 15757−15762. (60) Zhu, G.; Saw, W. G.; Nalaparaju, A.; Grüber, G.; Lu, L. Coarsegrained molecular modeling of the solution structure ensemble of dengue virus nonstructural protein 5 with small-angle X-ray scattering intensity. J. Phys. Chem. B 2017, 121 (10), 2252−2264. (61) Henzler-Wildman, K.; Kern, D. Dynamic personalities of proteins. Nature 2007, 450 (7172), 964−972. (62) Jacques, D. A.; Guss, J. M.; Svergun, D. I.; Trewhella, J. Publication guidelines for structural modelling of small-angle scattering data from biomolecules in solution. Acta Crystallogr., Sect. D: Biol. Crystallogr. 2012, 68 (6), 620−626.

11240

DOI: 10.1021/acs.jpcb.7b08299 J. Phys. Chem. B 2017, 121, 11229−11240