molecules Review
Neutron Crystallography for the Study of Hydrogen Bonds in Macromolecules Esko Oksanen 1,2 , Julian C.-H. Chen 3 and Suzanne Zoë Fisher 1,4, * 1 2 3 4
*
Science Directorate, European Spallation Source ERIC, Tunavägen 24, 22100 Lund, Sweden;
[email protected] Department of Biochemistry and Structural Biology, Lund University, Sölvegatan 39, 22362 Lund, Sweden Bioscience Division, Los Alamos National Laboratory, Los Alamos, NM 87545, USA;
[email protected] Department of Biology, Lund University, Sölvegatan 35, 22362 Lund, Sweden Correspondence:
[email protected]; Tel.: +46-72-179-2250
Academic Editor: Steve Scheiner Received: 12 March 2017; Accepted: 1 April 2017; Published: 7 April 2017
Abstract: The hydrogen bond (H bond) is one of the most important interactions that form the foundation of secondary and tertiary protein structure. Beyond holding protein structures together, H bonds are also intimately involved in solvent coordination, ligand binding, and enzyme catalysis. The H bond by definition involves the light atom, H, and it is very difficult to study directly, especially with X-ray crystallographic techniques, due to the poor scattering power of H atoms. Neutron protein crystallography provides a powerful, complementary tool that can give unambiguous information to structural biologists on solvent organization and coordination, the electrostatics of ligand binding, the protonation states of amino acid side chains and catalytic water species. The method is complementary to X-ray crystallography and the dynamic data obtainable with NMR spectroscopy. Also, as it gives explicit H atom positions, it can be very valuable to computational chemistry where exact knowledge of protonation and solvent orientation can make a large difference in modeling. This article gives general information about neutron crystallography and shows specific examples of how the method has contributed to structural biology, structure-based drug design; and the understanding of fundamental questions of reaction mechanisms. Keywords: hydrogen bond; neutron crystallography; structural enzymology; nuclear density maps; solvent networks
1. Hydrogen Bonds in Protein Structures A hydrogen bond (H bond) is established through the sharing of an H atom that is covalently attached to an electronegative donor (D), with a different electronegative acceptor atom (A), and can be written as D-H . . . A. H bonds are one of the major drivers that participate in the correct folding and organization of secondary and tertiary structural elements in proteins. In addition to their fundamental role in structure, H bonds are also intimately involved in protein:ligand binding interactions and also form the basis for most types of enzyme-catalyzed reactions [1]. Proteins contain many different kinds of H bond donors and acceptors, from the backbone amide and carbonyl groups, solvent, to the side chains of most amino acids residues. H bonds are directional, which is the driving force behind the specificity of molecular recognition processes in ligand binding [2]. Studies in the 90s aimed at investigating the H bond frequency between H bond donors/acceptors and have been carried out on a representative set of protein crystal structures available at the time. It was observed that most donors and acceptors engage in H bonds, with only a small percentage remaining as unsatisfied atoms. Many amino acid side chains can and do make multiple H bonds, e.g.,
Molecules 2017, 22, 596; doi:10.3390/molecules22040596
www.mdpi.com/journal/molecules
Molecules 2017, 22, 596
2 of 26
Lys, Arg, Asp, while others do not. For protein structure refinement and prediction calculations this is an important consideration to include [3]. From a theoretical chemistry point of view, the H bond interaction energy can essentially be described as being composed of a combination of covalent and weak interactions, e.g., van der Waals forces [2,4]. The energy associated with an H bond is 2–10 kcal/mol. While this may seem insignificant it is important to remember that these bonds are flexible and dynamic in that they are constantly broken and reformed under a range of temperatures and physiological conditions [2]. There is quite a large variation in H bond properties as observed in protein structures, especially in bond angles when comparing secondary structure elements to side chain characteristics. The D . . . A distances are typically between 1.7 and 2.4 Å, and the angle at the H atom can be anywhere from 130◦ to 170◦ , compared to ~150◦ at the acceptor [5]. As H bonds are difficult to directly observe, they are generally inferred when the donor-acceptor distance is less than ~3.5 Å, and the angles at the donor/acceptor are >90◦ [1]. Simple H atom sharing between two electronegative atoms represents the most common type of H bond, however, it is now accepted that there are additional types of weaker H bonds; e.g., C-H groups as donors, N-H . . . S, and C-H . . . π interactions with aromatic amino acid residues as acceptors [1,6]. A study conducted using NMR spectroscopy, DFT calculation, and data mining of high resolution protein crystal structures revealed that C-H . . . π type interactions are in fact quite common in proteins. These interactions are H bond-like in nature and occur between the methyl groups of Ile, Leu, and Val and the delocalized π electrons in aromatic residues. The results showed an average of four of these for every one hundred residues and are a functionally important part of protein structures [7]. H bonds in macromolecules are challenging to observe and classify, accordingly the criteria for identifying them in proteins are quite broad, as explained above. Another type of H bond that is commonly found in protein structures is the bifurcated H bond. In this case the slightly positively charged H atom is delocalized between two acceptor atoms. These are most often seen between Ser/Thr side chains with Asp/Glu as acceptors [8]. This indicates that H bonds cannot be identified only based on geometric criteria and that some additional structural or chemical knowledge is needed to fully understand all the H bonds in a protein. In recent years, bifurcated H bonds have also been directly observed in neutron protein crystal structures between Tyr and solvent, and Lys and Thr [9,10]. H atoms are abundant in macromolecules, comprising roughly half of the elements that proteins are composed of (C, H, N, O, S). However, despite their abundance and structural-functional importance, including but not limited to H bonding, electrostatic interactions, solvent coordination and solvent activation, H atoms and their interactions are very difficult to directly observe. This is especially true for X-ray crystallographic studies where information regarding H bonds are usually derived based on their heavy atom neighbor positions and the distances between them. The direct determination of H atom positions in protein crystal structures is restricted mainly due to their low X-ray scattering power, their inherent mobile nature, and limited crystallographic resolution [11]. 2. Neutron Crystallography and Hydrogen Atoms Neutron protein crystallography (NPX) is well-suited to the determination of the three-dimensional structures of proteins due to their sensitivity to the light atom H and lack of radiation damage at room temperature. NPX is conceptually very similar to standard X-ray crystallography, and crystallographic resolution is understood in the same way. The main difference between the methods lies in how neutrons are scattered from atoms. X-rays are scattered from the clouds of electrons around atomic nuclei, and the magnitude depends on the atomic Z number leading to heavy atoms scattering X-rays better than light atoms (Table 1). In practice this most often means that H atoms are invisible, except for ultra-high resolution structures that are exceptionally well-ordered, i.e., display low crystallographic B factors. Neutrons are scattered from atomic nuclei and the magnitude is independent of the Z number. Table 1 shows neutron scattering lengths for the most common atom
Molecules 2017, 22, 596
3 of 26
types found in proteins. Coherent scattering lengths of neutrons are very similar in magnitude for C, N, 1 H (and 2 H or D), and O, making it as easy to observe C as H (or D) [11,12]. Table 1. Neutron and X-ray scattering lengths for atom types commonly found in proteins § .
Isotope
Z
Neutron Scattering Length (10−13 cm)
Incoherent Cross Section σi (10−28 m2 )
X-ray Scattering Length (10−13 cm), sin θ = 0
X-ray Scattering Length (10−13 cm), (sin θ)/λ = 0.5 Å−1
1H
1 1 6 7 8 16
−3.74 6.67 6.65 9.37 5.80 2.80
80.27 2.05 0.00 0.50 0.00 0.00
2.80 2.80 16.90 19.70 22.50 45.00
0.20 0.20 4.80 5.30 6.20 1.90
2H
(D) 12 C 14 N 16 O 32 S
§
Taken from [11,13].
In practice this means that even at medium (2.0–2.5 Å crystallographic resolution) it is routine to readily observe the three dimensional positions of H atoms. By extension this means that neutrons are also used to study H bonds [14]. The isotope of 1 H, deuterium (2 H or D), has superior neutron scattering characteristics, making it optimal to replace H with D whenever possible (Table 1). The presence of 1 H presents several practical problems for crystallographic data collection. The large incoherent cross section (Table 1) contributes to background as the incoherently scattered neutrons do not contribute to measured Bragg reflections. In addition, the coherent scattering length is negative and leads to troughs instead of peaks and can produce cancellation of positive peaks of adjacent atoms, complicating and confusing analysis of nuclear density maps when 1 H is present. These effects can be minimized or even eliminated by D exchange (partial deuteration) or expression of the target protein under fully deuterated conditions (perdeuteration). Deuteration in general leads to higher quality maps and enables less ambiguous data analysis and structure interpretation [11,13]. Out of all H atoms in a protein, ~25% are titratable and can be exchanged by vapor H/D exchange or by soaking crystals in D2 O-containing solutions. These include H found in the bulk solvent, ordered solvent, polar and charged amino acid side chains, and the protein backbone. Neutron structures derived from H/D exchanged proteins represent the vast majority of deposited structures in the PDB [11]. The remaining 75% of H atoms are non-exchangeable, typically found attached to carbons and have to be incorporated as D during protein expression. This is routinely done by growing E. coli cultures for protein expression in deuterated minimal media, such as M9, composed of stock solutions and components dissolved in D2 O [11]. For the highest level D incorporation (~99%) it is best to use a perdeuterated carbon source, e.g., perdeuterated glycerol or glucose. This is very costly however and ~85% incorporation can be achieved by using deuterated M9 minimal media but supplementing with an unlabeled carbon source [15]. Cultures can be adapted in progressively increasing amounts of D2 O or grown to high densities and switched to deuterated media just prior to induction of protein expression. All proteins are then extracted and purified as usual, under hydrogenous conditions, with lost D atoms back-exchanged at a later point. Depending on the strategy chosen the resulting protein will have varying degrees and distribution of D incorporation. While perdeuteration is preferable from a theoretical point of view, there are practical drawbacks that make it difficult to achieve in practice and the low numbers of neutron structures determined from perdeuterated proteins reflects this. Although perdeuterated proteins are nearly identical in structure to their hydrogenated counterparts, the proteins are often found to have altered chemical properties, such as reduced solubility and stability. Deuterated cultures also grow slowly, and perdeuterated protein yields are generally lower. Perdeuteration also introduces the need to fine-screen or rescreen crystallization conditions under perdeuterated conditions [16]. As there is no radiation damage with neutrons used for protein crystallography applications, it is routine to collect data at room temperature. This feature makes sample preparation for data collection simple as crystals can be mounted in quartz capillaries and subjected to H/D exchange in
Molecules 2017, 22, 596
4 of 26
the capillary prior to exposure to neutrons. To achieve this, crystals are mounted in capillaries4 and Molecules 2017, 22, 596 of 25 a liquid plug of deuterated mother liquor is injected on either side of the crystal prior to sealing the capillary. capillary. The The pitfalls pitfalls of of finding finding protective protective cryogenic cryogenic solutions solutions and and the the actual actual freezing freezing of of crystals crystals are are also avoided [11,16]. also avoided [11,16]. Globally Globally there there are are only only aa handful handful of of instruments instruments available available that that are are specifically specifically optimized optimized for for macromolecular there areare currently twotwo instruments, bothboth located at Oak macromolecular crystallography. crystallography.InInthe theUSA USA there currently instruments, located at Ridge National Laboratory (ORNL, Oak Ridge, TN, USA). The IMAGINE instrument is located at the Oak Ridge National Laboratory (ORNL, Oak Ridge, TN, USA). The IMAGINE instrument is located HFIR source, whilewhile MaNDi is found at theatSpallation Neutron Source. In Japan there is the at the reactor HFIR reactor source, MaNDi is found the Spallation Neutron Source. In Japan there is iBIX instrument at the Japan Proton Accelerator Research Complex (J-PARC, Tokai, Japan). In Europe the iBIX instrument at the Japan Proton Accelerator Research Complex (J-PARC, Tokai, Japan). In there arethere currently two instruments operating at reactor neutron sources (Figure 1). Depending on Europe are currently two instruments operating at reactor neutron sources (Figure 1). Depending the source andand how thethe instrument is is setsetup, on the source how instrument up,data datacan canbe becollected collectedwith with monochromated monochromated (single (single wavelength) method or with a white Laue (multi-wavelength) method. The LADI-III (ILL, Grenoble, wavelength) method or with a white Laue (multi-wavelength) method. The LADI-III (ILL, Grenoble, France) France) instrument instrument uses uses aa quasi-Laue quasi-Laue spectrum, spectrum, while while BIODIFF BIODIFF (MLZ, (MLZ, Munich, Munich, Germany) Germany) uses uses aa monochromated beam. However However data data is is collected, collected, the the resulting resulting neutron neutron diffraction diffraction data data sets monochromated beam. sets from from different instruments are remarkably comparable and can be refined using standard crystallographic different instruments are remarkably comparable and can be refined using standard crystallographic software and analysis oneone cancan readily observe the software packages. packages. After Aftercareful carefulcrystallographic crystallographicrefinement refinement and analysis readily observe three-dimensional structural details of H/D atoms, H bonds, water molecules and their interactions, the three-dimensional structural details of H/D atoms, H bonds, water molecules and their the charged state of aminostate acidsofresidues and their interactions, and the details and of ligand binding. interactions, the charged amino acids residues and their interactions, the details of All this binding. information essential for the study of enzyme andenzyme drug binding, making neutron ligand Allisthis information is essential for mechanism the study of mechanism and drug protein a powerful tool in structural biology [17,18]. binding,crystallography making neutron protein crystallography a powerful tool in structural biology [17,18].
Figure 1.1.Neutron protein crystallography instrumentation. (Left) photograph of the LADI-III Quasi-Laue Figure Neutron protein crystallography instrumentation. (Left) photograph of the LADI-III diffractometer, located at the Institut Laue-Langevin, Grenoble FR [19]; (Right) photograph of the Quasi-Laue diffractometer, located at the Institut Laue-Langevin, Grenoble FR [19]; (Right) photograph BIODIFF monochromatic diffractometer, located at Heinz Maier-Leibnitz Zentrum, Munich of the BIODIFF monochromatic diffractometer, located at Heinz Maier-Leibnitz Zentrum, Munich DE. DE. Re-used with with permission permission from from the the photographer, photographer, Bernhard BernhardLudewig. Ludewig. Re-used
In general the technique is still underutilized, with only ~100 PDB entries resulting from neutron In general the technique is still underutilized, with only ~100 PDB entries resulting from neutron diffraction studies. This is largely due to three major bottlenecks: (1) neutron scattering is a flux diffraction studies. This is largely due to three major bottlenecks: (1) neutron scattering is a flux limited technique and accordingly requires long data collections times, ~5–20 days depending on the limited technique and accordingly requires long data collections times, ~5–20 days depending on the instrument; (2) quite large crystal volumes are required for most instruments (>0.5 mm33 on average); instrument; (2) quite large crystal volumes are required for most instruments (>0.5 mm on average); and (3) there are only a handful of instruments available worldwide to external users. Advances in and (3) there are only a handful of instruments available worldwide to external users. Advances in sources, beamlines, and detector technologies are easing the way forward with much shorter data sources, beamlines, and detector technologies are easing the way forward with much shorter data collection times from smaller crystals [11,16–18]. collection times from smaller crystals [11,16–18]. This review presents a number of topical examples of different kinds of structural and functional This review presents a number of topical examples of different kinds of structural and functional information that can be obtained with neutron crystallography, specifically how they relate to the information that can be obtained with neutron crystallography, specifically how they relate to the study study and importance of H bonds. Examples included are meant to span a broad area in structural and importance of H bonds. Examples included are meant to span a broad area in structural biology biology and include systems where ordered solvent and H bonded solvent network are intimately linked to catalysis. Fluorescent and chromoproteins are also shown as excellent cases where optical properties are solely determined by changes in H bonding around a chromophore. Several enzymology examples are discussed in a later section, highlighting the importance of knowledge of
Molecules 2017, 22, 596
5 of 26
and include systems where ordered solvent and H bonded solvent network are intimately linked to catalysis. Fluorescent and chromoproteins are also shown as excellent cases where optical properties 22, 596 5 of 25 areMolecules solely 2017, determined by changes in H bonding around a chromophore. Several enzymology examples are discussed in a later section, highlighting the importance of knowledge of H bonds and being able H bonds and being able to determine how they change during catalysis. Due to the strong scattering to determine how they change during catalysis. Due to the strong scattering of D, neutrons have also of D, neutrons have also been used to “see” exotic solvent species such as the elusive hydronium ion been used to “see” exotic solvent species such as the elusive hydronium ion (H O+ or D3 O+ ) an and (H3O+ or D3O+) and protons (H+ or D+ containing no electrons and therefore invisible3 with X-rays), + or D+ containing no electrons and therefore invisible with X-rays), an important proposed protons (H important proposed partner in certain kinds of enzyme-catalyzed reactions. Concluding remarks partner certain kinds ofperspective enzyme-catalyzed reactions. Concluding remarks offer a forward-looking offer ainforward-looking on the place of neutrons in current-day structure biology. perspective on the place of neutrons in current-day structure biology. 3. Solvent Networks and H Bonding to Proteins 3. Solvent Networks and H Bonding to Proteins 3.1. Ultra High Resolution Neutron Crystallography 3.1. Ultra High Resolution Neutron Crystallography Ultra-high resolution (1.1 Å) neutron diffraction data was collected on a large crystal of the small Ultra-high resolution (1.1 Å) neutron diffraction data was collected on a large crystal of the small protein crambin (46 residues, 4.7 kDa), isolated from seeds of Crambe abyssinica. The crystal had been protein crambin (46 residues, 4.7 kDa), isolated from seeds Crambe abyssinica. The high crystal had been exchanged against fully deuterated mother liquor prior to of data collection. The very resolution exchanged against fully deuterated mother liquor prior to data collection. The very high resolution of the data allowed for a nearly complete inventory (~95%) of all H and D atoms in the protein [20]. of Having the dataunambiguous allowed for astructural nearly complete inventory (~95%) of all and D atoms in the protein [20]. positions for the light atoms alsoHallowed for easy interpretation Having unambiguous structural positions for the light atoms also allowed for easy interpretation and and orientation of solvent molecules, with the ability to directly observe H bonding networks orientation of solvent molecules, with the ability to directly observe H bonding networks (Figure 2). (Figure 2).
Figure 2. Ultra-high resolution X-ray and neutron crystallographic H/D exchanged Figure 2. Ultra-high resolution X-ray and neutron crystallographic studiesstudies of H/Dofexchanged crambin, crambin, showing (A) details of possible H bonds between five ordered water molecules in crambin showing (A) details of possible H bonds between five ordered water molecules in crambin as observed as observed in the X-ray crystal structure; and (B) observed H bonds between water molecules as in the X-ray crystal structure; and (B) observed H bonds between water molecules as seen in the seen in the neutron crystal structure. Figure reproduced from [20]. neutron crystal structure. Figure reproduced from [20].
Even with the highest resolution X-ray structures, it remains difficult to visualize H atoms due withZthe highest resolution X-rayatoms, structures, to visualize H atoms due to Even their low and proximity to heavier and it asremains such, it difficult is difficult to accurately model Hto their low Z and proximity to heavier atoms, and as such, it is difficult to accurately model H bonding bonding networks. In the case of crambin there is a 0.48 Å resolution crystal structure of crambin networks. crambin is a recently 0.48 Å resolution crystal structure of crambin available availableIn inthe the case PDBof (3NIR), andthere also the collected 0.38 Å X-ray diffraction data set [21,22].in theEven PDBin(3NIR), and also the recently X-ray diffraction set [21,22]. Even these extraordinary cases it collected was still 0.38 onlyÅpossible to observedata around two-thirds ofin allthese H extraordinary still only possible to observe two-thirds H atoms in the atoms in thecases Fo − Fitc was difference maps. Inferring H atomaround positions and the of H all bonds they may beFo with maps. can beInferring misleading, as exemplified thethe study of crambin (Figure 2A,B). There − involved Fc difference H atom positions in and H bonds they may be involved withare can ordered water molecules in seen ultra-high resolution X-ray2A,B). crystalThere structures of crambin befive misleading, as exemplified theinstudy of crambin (Figure are five ordered with water very well-defined oxygen atom positions but crystal no definitive information on location of the H (or D) molecules seen in ultra-high resolution X-ray structures of crambin with very well-defined atoms. The dashed lines in Figure 2A indicate putative H bonds between these waters. Further oxygen atom positions but no definitive information on location of the H (or D) atoms. The dashed inspection of 2A the indicate nuclear density maps showed that there arewaters. in fact only two inspection H bonds between these lines in Figure putative H bonds between these Further of the nuclear waters (Figure 2B), and not five as can be interpreted based on the X-ray data alone. By taking density maps showed that there are in fact only two H bonds between these waters (Figure 2B), and ofbe theinterpreted high resolution the data, together with the scattering exhibited by notadvantage five as can basedofon the X-ray data alone. By strong taking neutron advantage of the high resolution D atoms, it was possible to model the anisotropic vibrational behavior of a selected number of D of the data, together with the strong neutron scattering exhibited by D atoms, it was possible to model atoms in crambin. Qualitatively, the D atoms were on average more anisotropic than their bonded, the anisotropic vibrational behavior of a selected number of D atoms in crambin. Qualitatively, the D neighboring heavy atom [20]. atoms were on average more anisotropic than their bonded, neighboring heavy atom [20]. 3.2. H Bonded Water Networks in Proton Transfer Human carbonic anhydrase II (HCA II) is one of 15 expressed isoforms found in humans. HCA II (261 residues, ~29 kDa) is a Zn-dependent enzyme that catalyzes the reversible reaction of carbon
Molecules 2017, 22, 596
6 of 26
3.2. H Bonded Water Networks in Proton Transfer Human carbonic anhydrase II (HCA II) is one of 15 expressed isoforms found in humans. HCA II (261 residues, ~29 kDa) is a Zn-dependent enzyme that catalyzes the reversible reaction of carbon dioxide hydration to form bicarbonate and a proton. In general, HCAs are found throughout the human Molecules 2017, 22, 596 6 of 25 body, from blood to the brain, and are involved in many and diverse physiological reactions [23]. The HCA active site has bicarbonate been studied and defined decades of biochemical dioxideIIhydration to form and aextensively proton. In general, HCAs areby found throughout the human and body, from blood to the brain, and are involved in many and diverse physiological reactions biophysical characterization, including X-ray crystallography, mutagenesis, NMR, and kinetics.[23]. HCA II HCA II active site has studied defined by decades of biochemical and is oneThe of the fastest enzymes withbeen a kcat of 106 extensively and kcat /Kand approaching the diffusion limit. The reaction M biophysical characterization, X-ray crystallography, mutagenesis, NMR,dioxide and kinetics. follows a ping-pong mechanismincluding and is composed of two half reactions: carbon hydration HCA II is one of the fastest enzymes with a kcat of 106 and kcat/KM approaching the diffusion limit. The followed by a proton transfer step. The rate-limiting proton transfer portion of the catalysis is reaction follows a ping-pong mechanism and is composed of two half reactions: carbon dioxide mediated by a very well-ordered H-bonded water network that spans over ~8 Å and connects catalytic hydration followed by a proton transfer step. The rate-limiting proton transfer portion of the (Zn-coordinated) water to the bulk solvent via a proton shuttling residue, His64. The active site catalysis is mediated by a very well-ordered H-bonded water network that spans over ~8 Å and cavity is lined with several hydrophilic amino acid residues for maintaining properHis64. active site connects catalytic (Zn-coordinated) water to the bulk solventcrucial via a proton shuttling residue, chemistry, geometry, and for coordinating the solvent network. The result is a well-tuned environment The active site cavity is lined with several hydrophilic amino acid residues crucial for maintaining that allows and efficientgeometry, proton transfer proper fast active sitevery chemistry, and for [23]. coordinating the solvent network. The result is a well-tunedstudies environment thatIIallows fastpH and (pH very 10) efficient proton [23]. details of the ordered Neutron of HCA at high showed thetransfer H bonding Neutron studies of HCA at high pH (pHnon-physiological 10) showed the H bonding details of the the organization ordered solvent network. Consistent withIIthe decidedly pH of the study, solvent network. Consistent with the decidedly non-physiological pH of the study, the organization and H bonding in the solvent network was such that it could not support proton transfer between and H bonding in the solvent network was such that it could not support proton transfer between catalytic water and the proton shuttling residue His64. In addition it was observed that the side chain catalytic water and the proton shuttling residue His64. In addition it was observed that the side of Tyr7, one of the residues that line the active site and coordinate the water network, was deprotonated chain of Tyr7, one of the residues that line the active site and coordinate the water network, was (Figure 3a) [24]. This was3a) surprising, evensurprising, though the pH of thethe crystallization was over 10, deprotonated (Figure [24]. This was even though pH of the crystallization wasas the theoretical pK of Tyr in solution is ~11. Tyr7 was the only deprotonated Tyr and all seven other over 10, asa the theoretical pKa of Tyr in solution is ~11. Tyr7 was the only deprotonated Tyr and all Tyr residues were as expected. seven otherneutral, Tyr residues were neutral, as expected.
Figure 3. Three different statesofofTyr7 Tyr7in inhuman human carbonic II. II. (a) (a) Tyr7 is deprotonated at Figure 3. Three different states carbonicanhydrase anhydrase Tyr7 is deprotonated at high pH; (b) Tyr7 is neutral at neutral pH but involved in a bifurcated H bond; (c) at pH 6.0 Tyr7 is high pH; (b) Tyr7 is neutral at neutral pH but involved in a bifurcated H bond; (c) at pH 6.0 Tyr7 is still still neutral and involved in a conventional H bond, participating as both acceptor and donor. 2Fo − Fc neutral and involved in a conventional H bond, participating as both acceptor and donor. 2Fo − Fc nuclear density maps are shown in magenta and are contoured at 1.3σ level. PDB IDs used (a) 3kkx, nuclear density maps are shown in magenta and are contoured at 1.3σ level. PDB IDs used (a) 3kkx, (b) 3tmj, and (c) 4y0j. (b) 3tmj, and (c) 4y0j.
Further neutron structures were solved at pH 7.8 and 6.0 to further probe the solvent network and changes in protonation Tyr7solved (Figureat3b,c) The 7.8 neutron structure showed a Further neutron structuresofwere pH [9,24]. 7.8 and 6.0pH to further probe the solvent network rearrangement of the solvent network, with all the waters now connected by H bonds, and neutral and changes in protonation of Tyr7 (Figure 3b,c) [9,24]. The pH 7.8 neutron structure showed a His64 poised to accept a proton. This pH “switch” indicated the pathway that an excess H+ can use to rearrangement of the solvent network, with all the waters now connected by H bonds, and neutral leave the active site. At pH 7.8 Tyr7 was neutral and involved in a bifurcated H bond with two His64 poised to accept a proton. This pH “switch” indicated the pathway that an excess H+ can use solvent molecules (Figure 3b). At pH 6.0 the water network was unchanged compared to the neutral to leave the active At pH 7.8 neutral involved in aconventional bifurcated H bond with two form, but Tyr7 site. had changed howTyr7 it Hwas bonds to the and solvent to a more configuration solvent molecules (Figure 3b). At pH 6.0 the water network was unchanged compared (Figure 3c). Another key observation was that of the proton shuttling residue, His64. At to pHthe 6.0neutral it form,was butdirectly Tyr7 had changed how it H bonds to the solvent to a more conventional configuration observed that His64 was charged (doubly protonated) and flipped away from the active (Figure key aobservation was that with of the proton shuttling residue, His64. At pH site3c). and Another involved with π-stacking interaction Trp5. Comparison of the low pH form with the 6.0 it high resolution X-ray crystal of HCA II in complex with substrate carbon dioxide, wasactive was directly observed that His64structure was charged (doubly protonated) and flipped away fromitthe seeninvolved that the active appeared interaction similar in solvent positions and amino of acid orientations. site and with sites a π-stacking with Trp5. Comparison theresidue low pH form with the This suggests that protonating the active site by lowering the pH provides a glimpse of how the active site is poised to accept substrate.
Molecules 2017, 22, 596
7 of 26
high resolution X-ray crystal structure of HCA II in complex with substrate carbon dioxide, it was seen that the active sites appeared similar in solvent positions and amino acid residue orientations. This suggests that protonating the active site by lowering the pH provides a glimpse of how the active site is poised to accept substrate. The first neutron structural studies were followed up with NMR titration studies of 13 C-Tyr labeled HCA II as a function of pH. It was possible to assign chemical shifts for all eight Tyr residues and the data confirmed the observation of a perturbed pKa for Tyr7. Consistent with the crystallographic data Tyr7 had a severely perturbed pKa at ~7, compared to ~10–12 for the seven other Tyr residues. The combined results of the high pH (pH 10), neutral pH (pH 7.8) and low pH (pH 6) forms revealed important details about proton transfer, amino acid residues protonation and H bonding, and how ordered solvent networks can respond and adapt to changing conditions [9]. These structural features could never be observed with X-ray crystallography alone as the ~1 Å X-ray structure did not indicate any explicit H atom positions, even on well-ordered residues. This study used pH as a variable parameter to mimic events that could occur during catalysis. Analysis of these data suggests an alternate catalytic model for how HCA II works. With a pKa in the physiological range Tyr7 may very well be part of catalysis under normal pH conditions. 3.3. Chromoprotein Dathail and H Bonding Fluorescent proteins (FP) occur naturally and display different fluorescent characteristics and are widely used in biological research as probes and reporters. Protein engineering, mutagenesis, and directed evolution have dramatically increased the variety and spectral properties of FPs. Green fluorescent protein (GFP) has a β-barrel structure made of 11 β strands with the chromophore located in a central pocket. The chromophore is formed after expression by the cyclization of 3 amino acids to form p-hydroxybenzylidenimidazolinone. The pocket around the chromophore is lined with amino acid side chains and ordered water molecules; mutagenesis can dramatically affect the spectral properties of variants. Some of these chromoproteins can photoswitch between fluorescent and non-fluorescent states by simple illumination by a specific wavelength of light. From X-ray analysis it is thought that switching from the fluorescent state is related to cis-to-trans isomerization of the chromophore, with cis corresponding to the ground state (fluorescent). Due to the structural rigidity of FPs it has been suggested that H bonds and related chemical environment around the chromophore play a fundamental role [25]. Neutron, X-ray and spectroscopic studies of a newly engineered FP, Dathail, revealed some interesting and unusual photoswitching behavior. Dathail has very high temperature stability, is less prone to aggregation, and readily crystallizes. Even more interesting is that the crystals tolerate photoswitching, allowing data to be collected of the “dark” (ground) and “light’ (metastable) states. Dathail is negative switching, with highest fluorescence in the ground state that is turned off when exposed to 497 nm light. Fluorescence emission spectra also showed these two states. The protein returns to the dark state through a very slow thermal relaxation process which takes several hours at 10 ◦ C [26]. X-ray crystal structures of the dark state showed that the chromophore is very well ordered in the cis conformation. There was a dramatic difference in the light state where the chromophore, surrounding amino acid residues, and solvent were very mobile and disordered in this form. X-ray structures alone did not provide insights into the photoswitching behavior of Dathail. The structure of the Dathail in the ground state is remarkably similar to those of other FPs, especially with regards to the chromophore region. The observed structural configuration is associated with fluorescence in other chromoproteins, however, Dathail is non-fluorescent in this form. As the differences seem to lie in H bonding patterns, neutron crystallographic data were collected from a large 10 mm3 H/D exchanged crystal of the dark (ground) state [26]. The fact that the ground-state chromophore of Dathail is clearly protonated and stable with the chromophore in the cis conformation challenges the consensus view that a stable, co-planar chromophore in the cis isomer leads to stable fluorescence, given that Dathail is essentially non-fluorescent.
Molecules 2017, 22, 596
8 of 26
Neutron studies showed that the chromophore is protonated in the ground state and showed further details of water molecules, H bonds, and protonation states of certain residues in the chromophore pocket. Furthermore, there is an H bonded water network from the chromophore to the bulk solvent that may be involved with the transfer of excited state protons out of the pocket during photo-activation and helps to explain the lack of fluorescence in the ground state. The precise details Molecules 2017, 22, 596 of the water network around the chromophore are unique to Dathail. 8 of 25 This is due to the presence of a Ser to Pro mutation at position 142. The waters mediate the de- and The of precise details of the water around the chromophore are are unique to Dathail. This to is the Ser re-protonation the chromophore. In network other FPs the chromophores thought to H bond due to the presence of a Ser to Pro mutation at position 142. The waters mediate the de- and (or Thr) residue in this position, making it difficult to protonate. From these neutron studies it was also re-protonation of the chromophore. In other FPs the chromophores are thought to H bond to the Ser seen that is neutral the ground state. His193 an H bond acceptor from studies a waterit molecule in (orHis193 Thr) residue in thisinposition, making it difficult to is protonate. From these neutron was the network, andthat is in turn ais donor A Gly atHis193 position 144Haffects how His193 coordinated, also seen His193 neutralto inGlu211. the ground state. is an bond acceptor from ais water especially compared to otherand chromoproteins there is aatGlu at this imidazole molecule in the network, is in turn a donorwhere to Glu211. A Gly position 144position. affects howThe His193 coordinated, especially compared to other chromoproteins where is chromoproteins a Glu at this position. ring of is His193 is π-stacked with the chromophore itself, similar tothere other (Figure 4). The imidazole ring of His193 is π-stacked with the chromophore itself, similar to other chromoproteins However, the loss of a H bond due to a Glu to Gly mutation makes His193 more dynamic and less able (Figure 4). However, the loss of a H bond due to a Glu to Gly mutation makes His193 more dynamic to order the chromophore. In this study neutrons provided valuable insights into the chromophore and less able to order the chromophore. In this study neutrons provided valuable insights into the pocket organization, the role of solvent and helpedand identify H bonding interactions chromophore pocket organization, the molecules, role of solvent molecules, helped key identify key H bonding that areinteractions at the heart ofare this behavior [26]. that at unique the heartphotoswitching of this unique photoswitching behavior [26].
4. The chromophore pocket of Dathail shows interactions to amino acidto residues andacid organization Figure Figure 4. The chromophore pocket of Dathail shows interactions amino residues and and orientation of key water molecules. generated fromgenerated PDB 4ebj [26]. organization and orientation of key water Figure molecules. Figure from PDB 4ebj [26].
4. Unusual H Bonds and Solvent Species as Observed in Neutron Crystal Structures
4. Unusual H Bonds and Solvent Species as Observed in Neutron Crystal Structures 4.1. Photoactive Yellow Protein: Short H Bonds vs. Low-Barrier H Bonds
4.1. Photoactive Yellow Protein: Short H Bonds vs. Low-Barrier H Bonds
Photoactive yellow protein (PYP) is a soluble 14 kDa protein from Halorhodospira halophila that absorbs light via itsprotein para-coumaric (pCA) chromophore. The absorption spectra for PYP is Photoactive yellow (PYP) acid is a soluble 14 kDa protein from Halorhodospira halophila that equivalent to the negative photo-dependent response of the organism, implicating PYP in its absorbs light via its para-coumaric acid (pCA) chromophore. The absorption spectra for PYP is biological behavior. PYP has a reversible photocycle during which pCA undergoes trans-to-cis equivalent to the negative photo-dependent response of the organism, implicating PYP in its biological isomerization, causing remodeling of the H bonded network and proton transfer processes. behavior. PYP has aofreversible photocycle which pCA undergoesregions trans-to-cis isomerization, Isomerization pCA culminates in largeduring structural changes of N-terminal of the protein. causingThis remodeling of the H bonded network and proton transfer processes. Isomerization means that H bonding is essential for photocycle kinetics in the protein. In the ground state of pCA culminates structural changes of N-terminal ofoxygen the protein. This meansTyr42 that and H bonding pCAin is large stabilized by two H bonds between the pCAregions phenolate atom and residues Glu46. High-resolution X-ray structural studies revealed these bonds to be rather short at 2.49 Å and is essential for photocycle kinetics in the protein. In the ground state pCA is stabilized by two H 2.58 Å, respectively. It was suggested thatatom these and bonds represent short and H bonds (SHB) [27]. bonds between the pCA phenolate oxygen residues Tyr42 Glu46. High-resolution X-ray As the H bonds become shorter there is a lengthening of the donor O-H covalent bond. The barrier structural studies revealed these bonds to be rather short at 2.49 Å and 2.58 Å, respectively. It was decreases as the disorder over two protonation sites collapse to a centered single-well potential, the suggested that these bonds short HLBHB bondsthe(SHB) so-called low-barrier H represent bond (LBHB). In an proton[27]. is shared between donor and acceptor Asatoms, the H making bonds become shorter there is a lengthening of the donor O-H covalent bond. The barrier it behave like a covalent bond [28]. crystallography was used to study PYPcollapse in ground in ordersingle-well to investigatepotential, these decreases asNeutron the disorder over two protonation sites tostate a centered the short H bonds and elucidate H atom in the the area of the core [28,29]. The firstacceptor so-called low-barrier H to bond (LBHB). In positions an LBHB proton ishydrophobic shared between donor and neutron crystal structure of PYP was determined to 2.5 Å resolution, from a ~0.8 mm3 crystal that atoms, making it behave like a covalent bond [28]. was H/D exchanged. From this first structure it was possible to elucidate some of the details around the pCA pocket, with pCA being deprotonated and the phenolate oxygen accepting H bonds from Glu46 and Tyr42. Tyr42 in turn is H bonded to Arg52. The observed H bond lengths to pCA were
Molecules 2017, 22, 596
9 of 26
Neutron crystallography was used to study PYP in ground state in order to investigate these short H bonds and to elucidate H atom positions in the area of the hydrophobic core [28,29]. The first neutron crystal structure of PYP was determined to 2.5 Å resolution, from a ~0.8 mm3 crystal that was H/D exchanged. From this first structure it was possible to elucidate some of the details around the pCA pocket, with pCA being deprotonated and the phenolate oxygen accepting H bonds from Glu46 Molecules 2017, 22, 596 of 25 and Tyr42. Tyr42 in turn is H bonded to Arg52. The observed H bond lengths to pCA were 9consistent with the high-resolution X-ray work, remarkable given the moderate resolution (2.5 Å) of the neutron consistent with the high-resolution X-ray work, remarkable given the moderate resolution (2.5 Å) of structure (Figure 5) [29]. (Figure 5) [29]. the neutron structure
Figure 5. Nuclear density maps around the pCA of PYP. (Left) H bonds around the pCA in PYP as Figure 5. Nuclear density maps around the pCA of PYP. (Left) H bonds around the pCA in PYP as observed in the 2.52.5 Å structure withpermission permission from International of observed in the Å structure[29]. [29].Reproduced Reproduced with from the the International UnionUnion of Crystallography; (Right) High resolution neutron studies to 1.5 Å enable accurate determination of Crystallography; (Right) High resolution neutron studies to 1.5 Å enable accurate determination of H/DH/D atom positions and assignment of LBHB in PYP. Figure reproduced from [28]. atom positions and assignment of LBHB in PYP. Figure reproduced from [28].
In 2009 a much higher resolution neutron structure of PYP was reported, to 1.5 Å resolution. In 2009 a much higher resolution neutron structure of PYP was reported, to 1.5 Å resolution. Now it was possible to fully investigate the short H bonds around pCA to see if they could Now it was possible to fully investigate the short H bonds around pCA to see if they could correspond correspond to LBHB [28]. From detailed investigations of omit Fo − Fc and 2Fo − Fc nuclear density to LBHB From detailed of omit 2Fao LBHB, − Fc nuclear density mapsHit was o − Fcould c andbe maps[28]. it was found that theinvestigations SHB between Glu46 andFpCA while the Tyr42-pCA found thatwas thebetter SHB between andionic pCAHcould be a LBHB, while the Tyr42-pCA H bond was better bond classified Glu46 as a short bond (SIHB). classifiedItas short ionic H bond (SIHB). is athought that the deprotonated/charged pCA represents an isolated charge in the otherwise hydrophobic ofdeprotonated/charged PYP with Arg52 acts aspCA a counter-ion. in charge the high-resolution It is thoughtpocket that the representsHowever, an isolated in the otherwise neutron structure this Arg was neutral. It was postulated that the SIHB and LBHB compensate for hydrophobic pocket of PYP with Arg52 acts as a counter-ion. However, in the high-resolution neutron this and provide structural stability PYP. LBHBsthat havethe been implicated in other systems forfor being structure this Arg was neutral. It wastopostulated SIHB and LBHB compensate this and part of catalysis and/or stabilization of intermediates. In this case the proposed LBHB may stabilize provide structural stability to PYP. LBHBs have been implicated in other systems for being part of an isolated charge and participate in proton transfer during the photocycle of PYP [28]. Comparison catalysis and/or stabilization of intermediates. In this case the proposed LBHB may stabilize an of the two studies also shows, unsurprisingly, that higher resolution neutron crystallographic data isolated charge and participate in proton transfer during the photocycle of PYP [28]. Comparison of allows for more accurate and detailed studies.
the two studies also shows, unsurprisingly, that higher resolution neutron crystallographic data allows for more accurate and detailed studies. 4.2. Unusual H Bonds as Seen with Neutrons Recent neutron structures have revealed a wide variety of H bonding details, including LBHBs, 4.2. Unusual H Bonds as Seen with Neutrons
SIHBs, and resonance-assisted H bonds. Despite over 40 years of structural, mechanistic, and computational work, there remains much atowide understand this fundamental, yet varied Recent neutron structures have revealed variety about of H bonding details, including LBHBs, interaction, found in protein structures and enzyme SIHBs, and resonance-assisted H bonds. Despitemechanisms. over 40 years of structural, mechanistic, and To study whether LBHB was involved in catalysis,about neutron studies were carried on a interaction, serine computational work, therea remains much to understand this fundamental, yetout varied protease, porcine pancreatic elastase (PPE), in complex with a peptidic inhibitor FR130180 [30]. found in protein structures and enzyme mechanisms. The residues His57-Asp102-Ser195 compose the catalytic triad and are conserved in the serine To study whether a LBHB was involved in catalysis, neutron studies were carried out on a proteases. In the first step of catalysis, the -OD1 of Ser195 makes a nucleophilic attack on a substrate serine protease, porcine pancreatic elastase (PPE), in complex with a peptidic inhibitor FR130180 [30]. carbonyl group. Next, a tetrahedral intermediate is formed through a covalent bond to Ser195 and The residues His57-Asp102-Ser195 the catalytic triad and conserved in theelectrostatic serine proteases. the substrate carbonyl group. compose The tetrahedral intermediate is are stabilized through In the first step that of catalysis, -OD1 Ser195amides makesof a nucleophilic attackTogether on a substrate carbonyl interactions involve Hthe bonds to of backbone Ser195 and Gly193. this forms a group. Next, aoxyanion tetrahedral intermediate is formed through a covalent bond toa Ser195 and the side substrate proposed hole. Ser195 is activated for nucleophilic attack through SSHB between chainsgroup. of His57The andtetrahedral Asp102 [31].intermediate The tetrahedralisintermediate is resolved through proton transfer that carbonyl stabilized through electrostatic interactions from His57 to the leaving group, leaving an acyl-enzyme intermediate. The H bond between Ser195, involve H bonds to backbone amides of Ser195 and Gly193. Together this forms a proposed oxyanion His57, and Asp102 has been proposed to be a LBHB and has been investigated with high resolution X-ray crystallography and NMR spectroscopy, in different types of proteases such as trypsin and subtilisin. The studies were inconclusive and in some cases instead showed a SIHB where the short
Molecules 2017, 22, 596
10 of 26
hole. Ser195 is activated for nucleophilic attack through a SSHB between side chains of His57 and Asp102 [31]. The tetrahedral intermediate is resolved through proton transfer from His57 to the leaving group, leaving an acyl-enzyme intermediate. The H bond between Ser195, His57, and Asp102 has been proposed to be a LBHB and has been investigated with high resolution X-ray crystallography and NMR spectroscopy, in different types of proteases such as trypsin and subtilisin. The studies were inconclusive and in some cases instead showed a SIHB where the short H bond (~2.6 Å) is Molecules 2017, 22, 596 10 of 25 observed. However, the shared H atom remains with the donor, and is not found equidistant between donor/acceptor in aHowever, LBHB [32,33]. H bond (~2.6as Å)expected is observed. the shared H atom remains with the donor, and is not found FR130180 to donor/acceptor PPE with theassame tetrahedral transition equidistantbinds between expected in a LBHB [32,33]. state intermediate as seen with binds to PPEmaps with show the same transition state intermediate as seen in with substrate.FR130180 The nuclear density thattetrahedral His57 is protonated (charged) and involved a short substrate. The nuclear density maps show that His57 is protonated (charged) and involved in a short H bond with Asp102, however the shared D density is clearly located on -ND1 of His57, and is most Asp102, however the shared D densitythe is clearly located on -ND1 His57, and is is most likelyHa bond SIHB.with In addition the neutron data showed carboxyl oxygen (O32)ofof FR130180 present likely a SIHB. In addition the neutron data showed the carboxyl oxygen (O32) of FR130180 is present as an oxygen anion (oxyanion) and was the first oxyanion in a tetrahedral intermediate observed at an as an oxygen anion (oxyanion) and was the first oxyanion in a tetrahedral intermediate observed at oxyanion hole. At a substrate binding subsite, S4, a π-cation interaction is observed between Arg217 an oxyanion hole. At a substrate binding subsite, S4, a π-cation interaction is observed between and the benzoyl ring of FR130180. Finally, there is an H bond between the –OH group of Thr175 and Arg217 and the benzoyl ring of FR130180. Finally, there is an H bond between the –OH group of the carboxyl group benzoic acidofinbenzoic the inhibitor. In this interaction, the carboxyl is protonated Thr175 and the of carboxyl group acid in the inhibitor. In this interaction, thegroup carboxyl group and the O-D distance is longer than expected (1.21 vs. 1.02 Å), while being only 1.41 Å is protonated and the O-D distance is longer than expected (1.21 vs. 1.02 Å), while being onlyaway 1.41 Åfrom the -OD1 Thr175. This is consistent with another of very strong H bond, awayof from the -OD1 ofconfiguration Thr175. This configuration is consistent with type another type of very strong H the bond, assisted the resonance assisted H bond (RAHB) [30]. resonance H bond (RAHB) [30]. The ultra-high resolution (1.1Å)Å)ofofthe thecrambin crambin neutron made it possible to clearly The ultra-high resolution (1.1 neutronstructure structure made it possible to clearly resolve H atom positions and the degree of exchange with D [20]. More accurately than in other resolve H atom positions and the degree of exchange with D [20]. More accurately than in other neutron structures resolved to date, the relative occupancies of H/D at exchangeable positions could neutron structures resolved to date, the relative occupancies of H/D at exchangeable positions could be refined, and solvent molecules and side chains could be identified and oriented. Of note were two be refined, and solvent molecules and side chains could be identified and oriented. Of note were two particular regions of the protein with unexpected characteristics. In Gly31, the two Hs appear to particular regions of the protein with unexpected characteristics. In Gly31, the two Hs appear to have have different degrees of exchange, based on the observed nuclear density for the two H atoms different degrees (Figure 6). of exchange, based on the observed nuclear density for the two H atoms (Figure 6).
Figure 6. Nuclear 2Fo − Fc density for crambin residue Gly31, with negative (pink) density for H and
Figure 6. Nuclear 2Fo − Fc density for crambin residue Gly31, with negative (pink) density for H and positive (green) for other atoms, showing different degrees of H/D exchange for the two alpha positive (green) for other atoms, showing different degrees of H/D exchange for the two alpha carbons. carbons. Figure reproduced from [20]. Figure reproduced from [20].
One interpretation of this is that there was partial exchange of the Cα-H against the deuterated solvent, leading to partial possibility of Cα-H…O H bonding. can be One interpretation of thisDisoccupancy, that there and was the partial exchange of the Cα-H againstThis the deuterated attributed to differing chemical environments in the crystal, corroborated by solution state NMR studies. solvent, leading to partial D occupancy, and the possibility of Cα-H . . . O H bonding. This can be two Cα-H atoms inenvironments Gly31 were observed to havecorroborated distinct NMR chemical shifts studies. of attributedThe to differing chemical in the crystal, byproton solution state NMR 3.96 and 5.56 ppm in solution, reflecting different chemical environments for the two alpha The two Cα-H atoms in Gly31 were observed to have distinct NMR proton chemical shifts of 3.96 hydrogens [34]. Since Gly does not have a side chain, they are relatively strong carbon acids (pKa ~ 22), and 5.56 ppm in solution, reflecting different chemical environments for the two alpha hydrogens [34]. comparable to the side chain hydroxyl groups of Ser and Thr (pKa ~ 18–20). Since Gly Another does notseries haveofapotential side chain, they are relatively carbon (pKregion comparable a ~22),of Cα-H…O H bonds werestrong observed in theacids β-strand crambin, to the side hydroxyl groupsoxygen of Ser of and Thrand (pKthe a ~18–20). e.g.,chain between the carbonyl Thr1 Cα-H of Ile34. These observations add to the Another series of potential Cα-H . . . O H bonds were in the β-strand of crambin, evidence that these weak, but numerous, stabilizing H observed bond interactions serve toregion contribute to e.g., between the carbonyl oxygen of Thr1 and the Cα-H of Ile34. These observations add to the secondary and tertiary structure stability. Secondly, Tyr44 in crambin is a well-ordered surface residue is solvated by anumerous, network of stabilizing waters on the one side, and makesserve a vantoder Waals to evidence thatthat these weak, but H bond interactions contribute interaction with Ile33 on the other side. As shown in Figure 7, one of these ordered water molecules, W1023, has its D1 atom pointing towards the center of the aromatic ring, indicating a potential O-H…π H bonding interaction [20]. With improvements in X-ray and neutron instrumentation and
Molecules 2017, 22, 596
11 of 26
secondary and tertiary structure stability. Secondly, Tyr44 in crambin is a well-ordered surface residue that is solvated by a network of waters on the one side, and makes a van der Waals interaction with Ile33 on the other side. As shown in Figure 7, one of these ordered water molecules, W1023, has its D1 atom2017, pointing Molecules 22, 596 towards the center of the aromatic ring, indicating a potential O-H 11 of. .25. π H bonding interaction [20]. With improvements in X-ray and neutron instrumentation and crystallization crystallization ultra high-resolution crystallography willHallow for H bonding interactions to approaches, ultra approaches, high-resolution crystallography will allow for bonding interactions to be studied be studied in greater detail and to better accuracy. in greater detail and to better accuracy.
Figure 7. Crambin residue Tyr44 is making van der Waals contact with Ile33 on one side of the Figure 7. Crambin residue Tyr44 is making van der Waals contact with Ile33 on one side of the aromatic aromatic ring and is solvated on the other side. W1023 interacts with the π-system of residue Tyr44. ring and is solvated on the other side. W1023 interacts with the π-system of residue Tyr44. D1 of W1023 D1 of W1023 is directly pointing Gly31 and different degrees of H/D exchange. Figure reproduced is directly pointing Gly31 and different degrees of H/D exchange. Figure reproduced from [20]. from [20].
−−), and Hydronium Ions (H +O+ ) in Proteins +), Hydroxides 4.3. Protons (H+ ), (OH 4.3. Protons (HHydroxides (OH ), Hydronium Ions (H3O3) in Proteins + important Water (H2O), protons ), hydroxide(OH (OH−−)) and (H(H 3O+)O players Water (H2 O), protons (H+(H ),+hydroxide andhydronium hydroniumions ions important players 3 are) are in many chemical and biochemical processes. They may be involved in a variety of different enzyme in many chemical and biochemical processes. They may be involved in a variety of different enzyme + catalytic processes difficult,ororininthe thecase case of of H to to directly observe usingusing X-rays. catalytic processes butbut areare difficult, H+, ,impossible impossible directly observe X-rays. In electron density maps derived from X-ray crystallography, most of these hydrated proton forms In electron density maps derived from X-ray crystallography, most of these hydrated proton forms are indistinguishable from each other as only the oxygen peak of these species are visible in electron are indistinguishable from each other as only the oxygen peak of these species are visible in electron density maps. Neutrons provide the appropriate tool to “see” the light atoms in solvent and make it density maps. provide the appropriate tool towater. “see” the light atoms in solvent and make it possible to Neutrons discern exotic solvent species from ordinary possible to discern exotic solvent ordinary Structure-function studies species of xylosefrom isomerase (XI) water. using neutrons made it possible to observe Structure-function studies of xylose (XI)inusing neutronssystem made it[35,36]. possible observe the first hydronium, hydroxyl, and isomerase proton ions a biological XI tofrom the first hydronium, hydroxyl, andkDa proton ions in a biological system [35,36]. from Streptomyces Streptomyces rubiginosus is a 172 homotetramer with 2 metal sites that bind XI divalent cations, usuallyismagnesium, in the active site. XI catalyzes the interconversion of D-glucose D-fructose, or rubiginosus a 172 kDa homotetramer with 2 metal sites that bind divalent cations, to usually magnesium, D -xylose to D -xylulose, and is of considerable industrial interest for production of sweeteners and in in the active site. XI catalyzes the interconversion of D-glucose to D-fructose, or D-xylose to D-xylulose, the biofuel field [37]. Metal site M1 is coordinated by a number of waters, and Glu and Asp amino and is of considerable industrial interest for production of sweeteners and in the biofuel field [37]. acid residues. Metal site M2 is coordinated by His, Asp, Glu side chains and a catalytic water Metal site M1 is coordinated by a number of waters, and Glu and Asp amino acid residues. Metal molecule. During catalysis there are multiple steps: substrate binding and coordination adjacent to site M2 is coordinated by His, Asp, Glu side chains and a catalytic water molecule. During catalysis the two metal sites, sugar ring opening, movement of a H from sugar C2 to sugar C1 through a thereproposed are multiple steps: substrate binding coordination adjacent to theThe twoprocess metal sites, sugar ring hydride shift or donation of aand hydride anion, and ring closure. culminates opening, movement of a H from sugar C2 to sugar C1 through a proposed hydride shift or with the product leaving the active site. These steps necessarily involve complex changes in waterdonation and of a hydride anion, ring closure. process culminates theprotonation/deprotonation product leaving the active amino side chainand positions, as well The as movement of H atoms with through of site. active site involve residues complex [35,38]. changes in water and amino side chain positions, as well as Thesewater stepsand necessarily Using combination of protonation/deprotonation different metal substitutions and by soaking in substrate or product movement of Haatoms through of water and active site residues [35,38]. molecules, it was possible to trap various stages occur during catalysis or by product XI. Using a combination of different metalintermediate substitutions and that by soaking in substrate In one such structure of the natural magnesium-containing enzyme in complex with the product molecules, it was possible to trap various intermediate stages that occur during catalysis by XI. In one xylulose, a hydroxyl anion was observed in place of the catalytic water at the M2 site. This observation such structure of the natural magnesium-containing enzyme in complex with the product xylulose, supports the idea that the catalytic water molecule donated a proton during the isomerization step of a hydroxyl anion was observed in place of the catalytic water at the M2 site. This observation supports catalysis and gave important clues as to the molecular events that occur upon isomerization [35]. the idea that the catalytic water molecule a proton during of catalysis Subsequent neutron studies lookeddonated at metal free XI at both pHthe 5.9 isomerization and 7.7 to help step understand and gave important clues as to the molecular events that occur upon isomerization [35]. why the enzyme is inactive at low pH. XI is most active at ~pH 8, however for industrial applications
it is desirable to engineer the enzyme to also be active at lower pH. It is known that at low pH the metals from M1 and M2 are expelled due to electrostatic changes in the area and this was also
Molecules 2017, 22, 596
12 of 26
Subsequent neutron studies looked at metal free XI at both pH 5.9 and 7.7 to help understand why the enzyme is inactive at low pH. XI is most active at ~pH 8, however for industrial applications it is desirable to engineer the enzyme to also be active at lower pH. It is known that at low pH the metals from Molecules M1 and2017, M222,are 596expelled due to electrostatic changes in the area and this was also observed 12 of 25in the neutron crystal structures. Neutron studies were conducted to study how the active site looks in the observed in the neutron crystal structures. studies were as conducted to study how the active absence of metal, and how the charged stateNeutron of residues change a function of pH. site looks in metal, andathow residues change as a function of pH. by a Analysis of the theabsence neutronofstructure pH the 7.7 charged showedstate thatof the metal at position M1 is replaced Analysis of the neutron structure at pH 7.7 showed that the metal at position M1 is replaced by for hydronium molecule (Figure 8) [36]. The hydronium ion templates the appropriate size and shape a hydronium molecule (Figure 8) [36]. The hydronium ion templates the appropriate size and shape metal binding at M1 and is coordinated by Glu181, Glu217, Asp245, and Asp287. In the pH 5.9 structure for metal binding at M1 and is coordinated by Glu181, Glu217, Asp245, and Asp287. In the pH 5.9 the hydronium ion has been dehydrated to leave a proton only in the M1 site. The proton is involved structure the hydronium ion has been dehydrated to leave a proton only in the M1 site. The proton is in a very shortintrifurcated bond to coordinating residues and located Glu217. At low pH involved a very shortHtrifurcated H bond to coordinating residues andclosest locatedto closest to Glu217. the amino acid residues have collapsed around the site and does not have the proper geometry At low pH the amino acid residues have collapsed around the site and does not have the proper to accept an incoming metal ligand. Soaking in solutions metal salts were also unsuccessful, providing geometry to accept an incoming metal ligand. Soaking inofsolutions of metal salts were also unsuccessful, providing further lowof pH form of M1 is not bind metal, explaining further evidence thatevidence the lowthat pHthe form M1 is not able to able bindtometal, thus thus explaining thethe lack of lackofofXI activity XI at low pH [37]. activity at lowofpH [37].
Figure Atomicdetails details of of metal metal binding at two different pHs. pHs. Figure 8. 8.Atomic binding site siteM1 M1using usingneutrons neutrons at two different + located in metal site M1 at pH 7.7 (from PDB ID 3kcj); 3O+ (Left) Coordination of hydronium ion D (Left) Coordination of hydronium ion D3 O located in metal site M1 at pH 7.7 (from PDB ID 3kcj); + located in metal site M1 at pH 5.9 (from PDB ID 3qza). (Right) Coordination of proton (Right) Coordination of proton D+Dlocated in metal site M1 at pH 5.9 (from PDB ID 3qza).
4.4. The Role of D3O+ Ions in Redox Processes
4.4. The Role of D3 O+ Ions in Redox Processes
Rubredoxins are highly thermostable, small, mononuclear iron-sulfur cluster proteins found in
Rubredoxins areand highly thermostable, mononuclear proteins both prokaryotes eukaryotes. As redoxsmall, proteins they serve as iron-sulfur an importantcluster model system forfound the in both study prokaryotes and eukaryotes. As redox proteins they serve as an important model system of proton transfer in catalysis. Iron-sulfur proteins vary greatly in their redox potentials, for the study of proton transfer in catalysis. Iron-sulfur vary greatly in their redox potentials, reflecting differences in solvation, electrostatics, and proteins H bonding. Neutron studies were initiated to obtaindifferences high resolution structureselectrostatics, of both oxidized on rubredoxin, in order to to reflecting in solvation, andand H reduced bonding.forms Neutron studies were initiated study water molecules, H bonding, and to investigate the charged state of chemical groups in the obtain high resolution structures of both oxidized and reduced forms on rubredoxin, in order to study as a function of oxidation [39]. waterprotein molecules, H bonding, and tostate investigate the charged state of chemical groups in the protein as Inspection of the ~1.3 and 1.4 Å resolution nuclear density maps revealed several interesting a function of oxidation state [39]. structural features involved with protein stability in both oxidized and reduced forms of the protein. Inspection of the ~1.3 and 1.4 Å resolution nuclear density maps revealed several interesting Four hydronium ions were observed in both oxidized and reduced rubredoxin structures. The first structural features involved with protein stability in both oxidized and reduced forms of the protein. molecule, HYD1, was located near Leu51 amide group and involved with amide-imide Four tautomerization. hydronium ions were observed in both oxidized and reduced rubredoxin structures. The first The second one, HYD2, is found by Pro44-Pro19-Ser46. HYD2 is ~10 Å away from molecule, HYD1, was located Leu51 amide groupHand involved with amide-imide tautomerization. the iron-sulfur cluster andnear forms water-mediated bonds to iron-sulfur coordinating Cys and The second HYD2, isInfound by HYD2 Pro44-Pro19-Ser46. is H ~10 Å away from the iron-sulfur adjacentone, Ser residues. this way is at the center HYD2 of a tight bonded group that connects water to the metal cluster coordinating residues. HYD3 also found near the iron-sulfur cluster (~9 Å), cluster and forms water-mediated H bonds to iron-sulfur coordinating Cys and adjacent Ser residues. but near Pro25-Ser24. The last one, HYD4, is H bonded to Ala16 carbonyl group. Occupancy In this way HYD2 is at the center of a tight H bonded group that connects water to the metal cluster refinement of one of the D atoms water/hydronium equilibrium. coordinating residues. HYD3 also indicates found near the iron-sulfur cluster (~9 Å), but near Pro25-Ser24. In summary then, three of the four hydronium ions are observed close to the protein main chain The last one, HYD4, is H bonded to Ala16 carbonyl group. Occupancy refinement of one of the D and another one forms the core of an intricate H bonded water network. One of these hydronium atoms indicates water/hydronium equilibrium. ions is located near main chain amide of Leu51 and is probably involved with redox driven In summary then, three of the four hydronium ions are observed close to the protein main chain tautomerization of amide to imide forms of the residue. In addition to the several hydronium ions and another one forms the core of an intricate H bonded water network.shifts Oneobserved, of these hydronium seen in rubredoxin, there were two additional important protonation in residues ions Glu47, Asp13, and Asp15. Glu47 was seen to be unprotonated in the reduced form while being
Molecules 2017, 22, 596
13 of 26
is located near main chain amide of Leu51 and is probably involved with redox driven tautomerization of amide to imide forms of the residue. In addition to the several hydronium ions seen in rubredoxin, there were two additional important protonation shifts observed, in residues Glu47, Asp13, and Asp15. Glu47 was seen to be unprotonated in the reduced form while being protonated in the oxidized form. This shifting D atom is also associated with HYD2. Asp13 on the other hand is unprotonated in the oxidized form while Asp15 is protonated. The observation of four hydronium ions and three shifting residues in close proximity to the iron-sulfur cluster suggest that these entities and surrounding solvent play important roles in charge transfer processes that occur in the change between oxidization and reduction in rubredoxin and associated redox processes [39]. With the growing use of neutron protein crystallography, it can be expected that more of these unusual solvent types will be directly observed and shed light on intricate and complex protein functions. 5. Structural Enzymology and H Bonds Hydrogen atoms are central to enzyme mechanisms and catalysis often relies on the movement of H atoms, either through proton shuttling or hopping or through the breaking/making of H bonds. There are many examples of neutron crystallography helping to elucidate enzyme mechanistic questions and all cannot be covered here. In this section a few examples are discussed where using neutrons made a significant contribution with regards to hydrogen atoms and their interactions. 5.1. Cytochrome C Peroxidase Heme peroxidases are widespread in nature and support a wide range of biological functions. They employ peroxide to oxidize a number of different substrates. Catalysis by heme peroxidases require the formation of two transient, highly oxidized ferryl (Fe(IV)) intermediates, the so-called Compounds I and II. These intermediates are applied widely in other oxygen-dependent catalytic heme-containing enzymes, including the cytochrome P450s, nitric oxide synthases, terminal oxidases and the heme dioxygenases [40]. It is important to understand the protonation state around the heme as this determines the reactivity and involvement of the enzyme. Compound I and Compound II form sequentially during catalysis and represent different Fe-IV (ferryl) species. They are different in their oxidation state, specifically the porphyrin ring. It is thought that the species are Fe(IV)=O or Fe(IV)-OH and it has been challenging to characterize the identities definitively [40,41]. Photo-reduction of the iron has been observed in both Raman spectroscopy and X-ray crystallography, rendering the interpretation of the oxidation states derived from these methods unreliable. Neutron crystallography can provide the high resolution, unambiguous structural information required to characterize the iron oxidation state, without inducing photo-reduction in the samples [40,41]. To unravel the mechanism of the peroxidases, two related heme peroxidases have been studied with neutron crystallography. Cytochrome c peroxidase (Ccp) oxidizes the small protein cytochrome c by hydrogen peroxide, and the mechanistically and structurally related ascorbate peroxidase (APX), the oxidation of ascorbate [40,41]. Cytochrome c peroxidase is a well-studied model system for heme-containing oxidases, which play a key role in many metabolic processes. Various species along the reaction path can be characterized by spectroscopic methods and trapped by cryo-cooling for crystallographic structure determination. However, the protonation states—which are crucial for understanding the electronic structure—are not easy to infer even with the combination of spectroscopy and crystallography. The structure from neutron diffraction study of the ferric state of the heme, determined at 2.4 Å resolution, showed a heavy water molecule (i.e., D2 O) bound at the start of the reaction (Figure 9, left). Crystals were then treated with peroxide, known to induce formation of Compound I. The treated crystals were cryo-cooled to trap the intermediate and studied with neutrons. The neutron structure, determined to 2.5 Å resolution, clearly shows the oxygen to be unprotonated, consistent with the H bonding structure of the active site (Figure 9, right). The ferryl iron is then double bonded to an
Molecules 2017, 22, 596
14 of 26
oxygen (Fe(IV)=O). The Trp 191 radical is protonated, so the radical species is a (protonated) π-cation radical [41]. 2017, 22, 596 Molecules 14 of 25
Figure 9. Neutron crystal structures peroxidaseinin the ferric Compound I form. Figure 9. Neutron crystal structuresofofcytochrome cytochrome cc peroxidase the ferric andand Compound I form. 2 O, is bound to the heme; (Right) nuclear (Left) nuclear density maps show that a heavy water, D (Left) nuclear density maps show that a heavy water, D2 O, is bound to the heme; (Right) nuclear density maps show oxygen boundtotothe the heme heme in I intermediate form. Figure density maps show thatthat oxygen is is bound in the theCompound Compound I intermediate form. Figure reproduced from [41]. reproduced from [41].
Another recent neutron crystallographic study focused on determining the oxidation and
Another recent crystallographic focused on determining the oxidation coordination state ofneutron the Compound II intermediate study in APX [40]. APX catalyzes the peroxide-dependent and oxidation coordination state of II intermediate in II APX [40]. APX of ascorbate by the usingCompound both Compound I and Compound intermediates. APXcatalyzes and Ccp the are highly conserved and Ccp has servedby as ausing model system for mechanistic studies in heme enzyme peroxide-dependent oxidation of ascorbate both Compound I and Compound II intermediates. catalysis. The experimental challenges of studying Compound II are more easily overcome in APX APX and Ccp are highly conserved and Ccp has served as a model system for mechanistic studies compared to Ccp. In APX Compound I appears as a ferryl heme and a porphyrin π-cation radicaleasily in heme enzyme catalysis. The experimental challenges of studying Compound II are more and is readily distinguishable from Compound II, which contains only a ferryl species. This is overcome in APX compared to Ccp. In APX Compound I appears as a ferryl heme and a porphyrin different from Ccp where Compound I is a ferryl heme and a Trp radical. Careful X-ray analysis and π-cation radical and is readily distinguishable from Compound II, which contains only a ferryl species. microspectrophotometry, along with neutron crystallography, revealed the heme ligand to be This ishydroxyl. differentAtfrom Ccp where Compound I is a ferryl heme and a Trp radical. Careful X-ray analysis the resolution limits of the crystallographic studies, it is not justified to assert that the and microspectrophotometry, along with revealed theanalysis heme ligand refined distance (1.88 Å) corresponds to aneutron hydroxylcrystallography, bound to the heme. However, of omitto be hydroxyl. At the thetocrystallographic it is notIIjustified assert that maps and lackresolution of residual limits densityofled the determinationstudies, that Compound in APX to is indeed the refined distance (1.88 Å)these corresponds to anew hydroxyl the heme. However, analysis of Fe(IV)-OH [40]. Together results offer insightsbound on the to mechanism of heme peroxidases, the entire family of the enzymes. Importantly, the first time, they is offer omit with mapsimplications and lack of for residual density led to determination thatfor Compound II in APX indeed unambiguous insights into the coordination and protonation state around the heme groups where Fe(IV)-OH [40]. Together these results offer new insights on the mechanism of heme peroxidases, with other techniques have been unsuitable. implications for the entire family of enzymes. Importantly, for the first time, they offer unambiguous insights into the coordination and protonation state around the heme groups where other techniques 5.2. Urate Oxidase have been unsuitable. Urate oxidase (EC 1.7.3.3) is an enzyme involved in purine metabolism. It is a co-factorless oxidase that catalyzes 5.2. Urate Oxidase the reaction between uric acid and molecular oxygen to form 5-hydroxy-isourate (5-HIU), which in turn is hydrolyzed to allantoin.
UrateThe oxidase 1.7.3.3) is an involved in purine in metabolism. It ispredominant a co-factorless reaction(EC between oxygen andenzyme a urate dianion is spontaneous solution, but the oxidase that catalyzes the reaction between uric acid and molecular oxygen to form 5-hydroxy-isourate protonation state of uric acid in solution is a monoanion with N3 deprotonated (Figure 10) [42]. (5-HIU), whichthe in principal turn is hydrolyzed to allantoin. Therefore role of the enzyme is to deprotonate the urate substrate in proximity of the oxygen. Numerous X-ray structures determined with inhibitors that mimic as well The reaction between oxygen and awere urate dianion isboth spontaneous in solution, but urate the predominant as oxygen at atomic The results difficult towith reconcile with the biochemical protonation state of uricresolution. acid in solution is awere monoanion N3 deprotonated (Figuredata, 10) [42]. since the site of the next deprotonation in solution, N7, is H bonded to a main chain amide and Therefore the principal role of the enzyme is to deprotonate the urate substrate in proximitynoof the feasible general bases be found were in the determined vicinity. The neutron structure at 2.35that Å resolution (Figure oxygen. Numerous X-raycan structures both with inhibitors mimic urate as11) well as showed that in fact the species in the active site was the enol tautomer of urate, 8-hydroxyxanthine. oxygen at atomic resolution. The results were difficult to reconcile with the biochemical data, since the site of the next deprotonation in solution, N7, is H bonded to a main chain amide and no feasible general bases can be found in the vicinity. The neutron structure at 2.35 Å resolution (Figure 11) showed that in fact the species in the active site was the enol tautomer of urate, 8-hydroxyxanthine.
Molecules 2017, 22, 596 Molecules 2017, 22, 596 Molecules 2017, 22, 596
15 of 26 15 of 25 15 of 25
Figure 10. The reaction catalyzed by urate oxidase. Figure reproduced from [42]. [42]. Figure oxidase.Figure Figurereproduced reproducedfrom from Figure10. 10.The Thereaction reactioncatalyzed catalyzed by by urate urate oxidase. [42].
Figure The 2.35ÅÅresolution resolution2mF 2mFoo-DF -DFcc nuclear nuclear density with Figure 11.11. The 2.35 density map mapof ofurate urateoxidase oxidase(A) (A)compared compared with Figure 11. The 2.35 Å resolution 2mFo -DFc nuclear density map of urate oxidase (A) compared with 1.05 X-raymap map(B); (B);(C) (C)aaschematic schematic representation representation of substrate. thethe 1.05 ÅÅ X-ray of the thesurroundings surroundingsofofthe theurate urate substrate. the 1.05 Å X-ray map (B); (C) a schematic representation of the surroundings of the urate substrate. Panels A and reproducedfrom from[42]. [42]. Panels and BB reproduced Panels A A and B reproduced from [42].
It was shownbybyquantum quantumchemical chemical calculations calculations that unstable It was shown that despite despitethe theenol enoltautomer tautomerbeing being unstable It was shown by quantum chemical calculations that despite theclearly enol tautomer being unstable in in solution, the active site can stabilize it. Interestingly, the O-H was pointing away from thethe in solution, the active site can stabilize it. Interestingly, the O-H was clearly pointing away from solution, the active site can stabilize it. Interestingly, the O-H was clearly pointing away from theThere purine purine plane, even though there was no H bonding partner observed to explain this behavior. purine plane, even though there was no H bonding partner observed to explain this behavior. There plane, though there was no H bonding partner observed explain this behavior. There was even however enough empty space for a water molecule, even though none were observed in the was however enough empty space for a water molecule, even to though none were observed inwas the X-ray maps. quantum calculations wereinperformed to however enoughFurther empty space for amechanics/molecular water molecule, even mechanics though none were observed the X-ray maps. X-ray maps. Further quantum mechanics/molecular mechanics calculations were performed to understand this. It was not possible to mechanics obtain good geometry consistent with thetohigh-resolution Further quantum were performed this. understand this. Itmechanics/molecular was not possible to obtain goodcalculations geometry consistent with the understand high-resolution X-ray structure unless a water molecule was placed in proper H bonding geometry with respect to It was structure not possible to obtain good geometry consistent the high-resolution X-ray structure unless X-ray unless a water molecule was placed inwith proper H bonding geometry with respect to the O-H. This illustrates the in synergies between crystallography andrespect computational chemistry where a water molecule was placed proper H bonding geometry with to the O-H. This illustrates the O-H. This illustrates the synergies between crystallography and computational chemistry where methods can predict and or fill in details when crystalwhere structures give incomplete thecomputational synergies between crystallography computational chemistry computational methods computational methods can predict or fill in detailsinwhen crystal instructures give incomplete information. A water molecule was indeed observed this position an X-ray structure from can predict orAfill in details whenwas crystal structures giveinincomplete information. A water molecule information. water molecule indeed observed this position in an X-ray structure from anaerobically grown crystals [43]. was indeed observed in this position in an X-ray structure from anaerobically grown crystals [43]. anaerobically grown crystals [43]. This result allowed a more plausible mechanism to be postulated where the deprotonation This result allowed a amore plausible mechanism to be postulated wherewhere the deprotonation occurs This result allowed more plausible mechanism be postulated deprotonation occurs on the oxygen instead of the nitrogen. One of the to unsolved questions in thethe mechanism was on the oxygen instead of the nitrogen. One of the unsolved questions in the mechanism was always occurs on the oxygen instead of the nitrogen. One of the unsolved questions in the mechanism was always how an H atom from the N7 position can reach the oxygen site above the ring plane. how an H atom from the N7 position can reach the oxygen site above the ring plane. A surprising always how anobservation H atom from N7 position oxygennetwork site above the ring A surprising in thethe neutron structurecan was reach that anthe H bonded extended fromplane. N7 observation in the neutroninstructure wasstructure that an− Hwas bonded network extended fromextended N7 all the wayN7 to A surprising the(that neutron that an H bonded network from all the way observation to the oxygen site contains Cl in this structure), and the H atoms were disordered − in this structure), and the H atoms were disordered (Figure 12). the oxygen site (that contains Cl 12). all(Figure the way to the oxygen site (that contains Cl− in this structure), and the H atoms were disordered Nevertheless the oxygen atoms atomswere werewell wellordered. ordered.This This Nevertheless theX-ray X-raystructure structure showed showed that that the the oxygen is is (Figure 12). indicative of of a proton chain that can easily pass the proton through three water molecules, a aLys indicative a proton relay chain that can easily pass the proton through three water molecules, Nevertheless therelay X-ray structure showed that the oxygen atoms were well ordered. This is and a Ser side chain. The existence of such a relay chain also explains how the oxygen site can Lys and aofSer side chain. The existence of such a relay also explains howthree the oxygen site can be be indicative a proton relay chain that can easily passchain the proton through water molecules, a provided with a second proton to eventually form H O . 2 2 provided with a second proton to eventually form H 2 O 2 . Lys and a Ser side chain. The existence of such a relay chain also explains how the oxygen site can be
provided with a second proton to eventually form H2O2.
Molecules 2017, 22, 596
16 of 26
Molecules 2017, 22, 596
16 of 25
Figure 12. The postulated proton relay chain as seen in the 2mFo-DFc nuclear density map. Figure Figure 12. The postulated proton relay chain as seen in the 2mFo -DFc nuclear density map. Figure reproduced from from [42]. [42]. reproduced
5.3. β-Lactamase 5.3. β-Lactamase Bacteria have developed antibiotic resistance to β–lactam antibiotics by expressing an enzyme, Bacteria have developed antibiotic resistance to β–lactam antibiotics by expressing an enzyme, β-lactamase, which breaks the amide bond within the lactam ring and renders the compound ineffective. β-lactamase, which breaks the amide bond within the lactam ring and renders the compound This class of antibiotics inhibits cell wall synthesis by binding to bacterial penicillin-binding proteins ineffective. This class of antibiotics inhibits cell wall synthesis by binding to bacterial penicillin-binding (PBPs). The β-lactamases are divided into classes A through D. All but class B use a serine-reactive proteins (PBPs). The β-lactamases are divided into classes A through D. All but class B use a mechanism. The class A CTX-M extended-spectrum-β-lactamase (ESBL) has high hydrolytic activity serine-reactive mechanism. The class A CTX-M extended-spectrum-β-lactamase (ESBL) has high against first, second, and third generation cephalosporins and monobactams. The enzymes work by hydrolytic activity against first, second, and third generation cephalosporins and monobactams. using a general base mechanism to break the amide bond within the β-lactam ring [44]. One class A The enzymes work by using a general base mechanism to break the amide bond within the β-lactam enzyme, CTX-M-44, was chosen for neutron studies to further investigate the role of active site ring [44]. One class A enzyme, CTX-M-44, was chosen for neutron studies to further investigate Glu166, Ser70, and Lys73 in catalysis. CTX-M-44 is composed of 262 residues and has two conserved the role of active site Glu166, Ser70, and Lys73 in catalysis. CTX-M-44 is composed of 262 residues domains, with the active site located at the interface between the domains. It has been proposed that and has two conserved domains, with the active site located at the interface between the domains. Glu166 acts as a general catalytic base in the acylation part of the reaction through deprotonation of It has been proposed that Glu166 acts as a general catalytic base in the acylation part of the reaction Ser70. Ser70 then attacks the carbonyl carbon of the β-lactam ring. Glu166 is also thought to be through deprotonation of Ser70. Ser70 then attacks the carbonyl carbon of the β-lactam ring. Glu166 involved in the second part de-acylation part of the reaction by activating a hydrolytic water is also thought to be involved in the second part de-acylation part of the reaction by activating a molecule. Interestingly, the CTX-M-44 Glu166Ala (E166A) mutant is still able to degrade β-lactams, hydrolytic water molecule. Interestingly, the CTX-M-44 Glu166Ala (E166A) mutant is still able to but with the rates several orders of magnitude slower [45]. This suggests there is more than one way to degrade β-lactams, but with the rates several orders of magnitude slower [45]. This suggests there is degrade β-lactams, including some that have been proposed to use Lys73 [46]. Neutron studies were more than one way to degrade β-lactams, including some that have been proposed to use Lys73 [46]. conducted on the E166A mutant as it degrades the β-lactam slow enough to allow crystallographic Neutron studies were conducted on the E166A mutant as it degrades the β-lactam slow enough to investigation of the acyl-intermediate. Studies were done on both the ligand-free enzyme and in allow crystallographic investigation of the acyl-intermediate. Studies were done on both the ligand-free complex with cefotaxime. The data on the complex revealed that upon inhibitor binding, an enzyme and in complex with cefotaxime. The data on the complex revealed that upon inhibitor binding, oxyanion hole is formed involving the backbone amides of Ser237 and Ser70 with stabilizing H an oxyanion hole is formed involving the backbone amides of Ser237 and Ser70 with stabilizing H bonds to the carbonyl oxygen of cefotaxime. This interaction serves to stabilize the negative charge that is bonds to the carbonyl oxygen of cefotaxime. This interaction serves to stabilize the negative charge that formed in the tetrahedral intermediates during the acylation and deacylation stages of hydrolysis. is formed in the tetrahedral intermediates during the acylation and deacylation stages of hydrolysis. Lys73 is highly conserved in class A β-lactamases and has been thought to act as general base, Lys73 is highly conserved in class A β-lactamases and has been thought to act as general base, in competition with Glu166, by abstracting a proton from Ser70. The ability of E166A mutant to still in competition with Glu166, by abstracting a proton from Ser70. The ability of E166A mutant to produce the acyl-enzyme complexes, which supports the idea that Lys73 is involved, albeit less still produce the acyl-enzyme complexes, which supports the idea that Lys73 is involved, albeit less effectively than Glu166. In the ligand-free neutron structure, it is seen that Lys73 is charged (-D3+) effectively than Glu166. In the ligand-free neutron structure, it is seen that Lys73 is charged (-D3 + ) and and occupies a single conformation that is identical to that observed in the form where Glu166 is occupies a single conformation that is identical to that observed in the form where Glu166 is present. present. In the cefotaxime-complex structures however, Lys73 is observed in two alternative In the cefotaxime-complex structures however, Lys73 is observed in two alternative conformations conformations that refine to near equal occupancy, presumably due to different charged states that refine to near equal occupancy, presumably due to different charged states (Figure 13). These (Figure 13). These observations together support a role for Ly73 in proton transfer and represents a observations together support a role for Ly73 in proton transfer and represents a viable alternative viable alternative pathway when Glu166 is mutated [45]. pathway when Glu166 is mutated [45].
Molecules 2017, 22, 596
17 of 26
Molecules 2017, 22, 596
Molecules 2017, 22, 596
17 of 25
17 of 25
Figure 13. Proposed catalytic residuesLys73 Lys73in in class class A—lactamase actact as aasgeneral base.base. In theIn the Figure 13. Proposed catalytic residues A—lactamasecould could a general acyl-enzyme complex, Lys73 can be modeled in two conformations. The nuclear density map is acyl-enzyme complex, Lys73 can be modeled in two conformations. The nuclear density map is shown shown in pink mesh and is contoured at 0.8σ. Reprinted (adapted) with permission from [45]. in pink mesh and is contoured at 0.8σ. Reprinted (adapted) with permission from [45]. Copyright 2016 Copyright 2016 American Chemical Society. American Chemical Society. Figure 13. Proposed catalytic residues Lys73 in class A—lactamase could act as a general base. In the
5.4. Diisopropyl Fluorophosphatase acyl-enzyme complex, Lys73(DFPase) can be modeled in two conformations. The nuclear density map is
5.4. Diisopropyl Fluorophosphatase shown in pink mesh and(DFPase) is contoured
at 0.8σ. Reprinted with permission [45]. in the The enzyme diisopropyl fluorophosphatase (DFPase; (adapted) 314 residues, 35 kDa),from found Copyright 2016 American Chemical Society. Mediterranean squid Loligo vulgaris, is a calcium-dependent that can hydrolyze The enzyme diisopropyl fluorophosphatase (DFPase;enzyme 314 residues, 35 kDa),a variety found of in the toxic organophosphorus compounds through hydrolysis of a P-F bond. The compounds include the Mediterranean squidFluorophosphatase Loligo vulgaris,(DFPase) is a calcium-dependent enzyme that can hydrolyze a variety 5.4. Diisopropyl nerve agents Sarin, Soman, and cyclosarin, and hydrolysis the pesticideofdiisopropyl fluorophosphate (DFP). of toxic organophosphorus compounds through a P-F bond. The compounds include The enzyme diisopropylinhibitors fluorophosphatase (DFPase; 314 and residues, kDa), found in the These agents act as irreversible of acetylcholinesterase serine35proteases. The neutron the nerveMediterranean agents Sarin, Soman, cyclosarin, and the pesticide diisopropyl fluorophosphate squid Loligoand vulgaris, a calcium-dependent enzyme thatstructures can hydrolyze a variety and of (DFP). structure of DFPase, together with a is variety of high-resolution X-ray of wild-type These agents act as irreversible inhibitors of acetylcholinesterase andThe serine proteases. Thethe neutron toxic organophosphorus compounds through hydrolysis of a P-F bond. compounds include site-directed mutants, and enzymology have greatly clarified the mechanism [47,48]. agents Sarin, Soman, and acyclosarin, andhigh-resolution the 18pesticide diisopropyl fluorophosphate (DFP). and structurenerve of DFPase, together with variety of X-ray structures of wild-type Single- and multiple-turnover studies utilizing an O substrate, together with mass spectroscopy, These agents actand as irreversible inhibitors of acetylcholinesterase serine proteases. The 18neutron site-directed enzymology have greatly clarified theand mechanism [47,48]. 16O-containing showed amutants, product under single-turnover conditions, and incorporation of O under structure of DFPase, together with a variety of high-resolution X-ray structures of wild-type and 18 Singlemultiple-turnover studies utilizing an O the substrate, with mass spectroscopy, multipleand turnover conditions. After multiple turnovers, enzymetogether was digested to determine site-directed mutants, and enzymology have greatly clarified the mechanism [47,48]. 1618 showed a O-containing product under single-turnover conditions, and incorporation of 18 O under where SingleO hadand been incorporated studies and results showed that the together proteolytic in question multiple-turnover utilizing an 18O substrate, withpeptide mass spectroscopy, contained the residue Asp229. These labeling demonstrated that multiple turnover conditions. multiple turnovers, the enzymeexperiments wasincorporation digested toof determine where 18O under showed a 16calcium-coordinating O-containing After product under single-turnover conditions, and 18 O had the multiple reaction proceeded through aAfter direct attack that ofturnovers, anthe amino acidenzyme on the substrate, andto notdetermine through a the been incorporated and results showed proteolytic peptide in question contained turnover conditions. multiple the was digested 18O had metal-activated water as previously thought. In the neutron structure, Asp229 was found to be where been incorporated and results showed that the proteolytic peptide in question calcium-coordinating residue Asp229. These labeling experiments demonstrated that the reaction deprotonated, consistent with the proposed (Figure 14A)experiments [49]. contained the calcium-coordinating residuemechanism Asp229. These labeling demonstrated that
proceeded through a direct attack of an amino acid on the substrate, and not through a metal-activated the reaction proceeded through a direct attack of an amino acid on the substrate, and not through a water as previously thought. In the neutron structure, Asp229 was found to be deprotonated, consistent metal-activated water as previously thought. In the neutron structure, Asp229 was found to be with the deprotonated, proposed mechanism (Figure 14A) [49]. consistent with the proposed mechanism (Figure 14A) [49].
Figure 14. DFPase active site as observed in the neutron crystal structure. (A) 2Fo − Fc nuclear density (gray) shows a deprotonated Asp229; (B) Detail of the calcium-binding environment, showing the Fo −Figure Fc omit14. density (gray) for the two D atoms inneutron W33. This establishes the identity of W33 density as water site as observed in the crystal structure. 2Fo2F − Fc − nuclear Figure 14. DFPaseDFPase activeactive site as observed in the neutron crystal structure.(A)(A) Fc nuclear density o and(gray) not hydroxide. Details surrounding calcium and associated water molecule shows a deprotonated Asp229; the (B) active Detail site of the calcium-binding environment, showingenable the (gray)mechanistic shows a deprotonated Asp229; (B)from Detail of the calcium-binding environment, showing the studies. Figure [47]. Fo − Fc omit density (gray)reproduced for the two D atoms in W33. This establishes the identity of W33 as water
Fo − Fc omit density (gray) for the two D atoms in W33. This establishes the identity of W33 as water and not hydroxide. Details surrounding the active site calcium and associated water molecule enable and not DetailsFigure surrounding thefrom calcium andwith associated water molecule enable Thehydroxide. catalytic ion isreproduced involved inactive 7-fold coordination, a solvent molecule bound in mechanisticcalcium studies. [47].site mechanistic studies. Figure reproduced from [47]. the active site pocket. While X-ray structures were not able to discern the species of the solvent The calcium ion revealed is involved in 7-fold coordination, with solvent moleculeascertained bound in molecule, thecatalytic neutron structure it to be a water molecule, anda not a hydroxyl, the active site pocket. While X-ray structures were maps not able to discern the species of the solvent by the elongated nuclear density and F o − Fc omit (Figure 14B) [47]. This unambiguously The catalytic calcium ion is involved in 7-fold coordination, with a solvent molecule bound in molecule, the neutron structure revealed it to be a water molecule, and not a hydroxyl, ascertained
the active pocket. While to discern species of the solvent by site the elongated nuclear X-ray densitystructures and Fo − Fcwere omit not mapsable (Figure 14B) [47].the This unambiguously molecule, the neutron structure revealed it to be a water molecule, and not a hydroxyl, ascertained by the elongated nuclear density and Fo − Fc omit maps (Figure 14B) [47]. This unambiguously showed
Molecules 2017, 22, 596 Molecules 2017, 22, 596
18 of 26 18 of 25
that the mechanism did not involve direct water activation by the calcium ion, as had been previously showed by that the mechanism didbut notwas involve directwith waterthe activation bythe therole calcium ion, as had Of been suggested other investigators, consistent Asp229 in as nucleophile. note previously suggested by other investigators, but was consistent with the Asp229 in the role as was the orientation of the bound water (D2 O) molecule in relation to the position of the catalytic nucleophile. Of anote was thedistance orientation of Å) thebetween bound water 2O)Dmolecule relation thethe calcium ion, with very close (2.08 one of(D the atoms ofinthe water to and position the a catalytic ion, a very close distance (2.08 Å) between one of the D atoms calcium ion,ofand Ca-O-Dcalcium angle of 53with degrees. of the water and the calcium ion, and a Ca-O-D angle of 53 degrees. 5.5. Fungal Lytic Polysaccharide Monooxygenase 5.5. Fungal Lytic Polysaccharide Monooxygenase Fungal lytic polysaccharide monooxygenases (LPMO) are Cu-containing metalloenzymes Fungal lyticdegradation polysaccharide monooxygenases (LPMO) are Cu-containing metalloenzymes involved with the of polysaccharides through the insertion of one O into the glycosidic involved with the degradation of polysaccharides through the insertion of one O into glycosidic bond carbon, leading to chain cleavage. There is significant interest in the mechanism ofthe these enzymes bond carbon, leading to chain cleavage. There is significant interest in the mechanism of these as there is potential application in cellulose degradation for biofuel production. The general mechanism enzymes as there is potential application in cellulose degradation for biofuel production. The starts with a single electron reducing the resting state Cu(II) to Cu(I). Cu(I) has a high affinity for general mechanism starts with a single electron reducing the resting state Cu(II) to Cu(I). Cu(I) has a molecular oxygen (O2 ). Oxygen binding to Cu(I) quickly regenerates Cu(II) through oxidation with high affinity for molecular oxygen (O2). Oxygen binding to Cu(I) quickly regenerates Cu(II) through the formation of Cu-superoxide. From this intermediate there are multiple possible pathways that oxidation with the formation of Cu-superoxide. From this intermediate there are multiple possible may involve superoxide, hydroperoxyl, or oxyl that are responsible for abstracting an H atom from the pathways that may involve superoxide, hydroperoxyl, or oxyl that are responsible for abstracting an glycosidic the absence of substrate the superoxide species can leave Cu(II) H atom carbon. from theInglycosidic carbon. In the absence of substrate the superoxide speciesand cangenerate leave peroxide. There are many uncertainties in many how these enzymesinactivate oxygen and the mechanism Cu(II) and generate peroxide. There are uncertainties how these enzymes activate oxygen of peroxide no substrate is present and theproduction mechanismwhen of peroxide production when [50,51]. no substrate is present [50,51]. Neurospora crassa LPMO is a 223 residue (23.3 enzyme withCu one per monomer. Neurospora crassa LPMO is a 223 residue (23.3 kDa)kDa) enzyme with one perCu monomer. High High resolution crystallography DFT calculations were supplemented with neutron resolution X-rayX-ray crystallography and DFT and calculations were supplemented with neutron crystallography crystallography in order to better understand mono-Cu activation of molecular oxygen and in order to better understand mono-Cu activation of molecular oxygen and H2O2 production in H the 2 O2 production the absence of substrate. absence ofinsubstrate. X-ray crystal structuresofofthe theresting restingCu(II) Cu(II) state state showed showed that a conserved X-ray crystal structures thatCu Cuisiscoordinated coordinatedbyby a conserved “His brace”made madeofof the the backbone backbone amide of His1 and and the imidazole of His84. The “His brace” amideand andimidazole imidazole of His1 the imidazole of His84. hydroxyl group of Tyr168 is pointing towards with Cu site with the remaining axial and equatorial The hydroxyl group of Tyr168 is pointing towards with Cu site with the remaining axial and equatorial coordination positions occupiedby bytwo twowaters, waters,H H22O Oax ax and eq (Figure 15) [51]. coordination positions occupied andHH2O 2 Oeq (Figure 15) [51].
Figure Cu-binding site coordination and coordination in different structures offrom LPMO from Figure 15. 15. Cu-binding site and in different crystal crystal structures of LPMO Neurospora 2− ) bound; bound; (b)structure X-ray Neurospora crassa. (a) X-ray structure of crystal with peroxo ascorbate, (O22−)(b) crassa. (a) X-ray structure of crystal treated withtreated ascorbate, (O2peroxo X-ray structure of crystal withmolecular ascorbate, oxygen molecular bound in “pre-bound” position; of crystal treated withtreated ascorbate, (O2oxygen ) bound(Oin2) “pre-bound” position; (c) neutron (c) neutron structure with His157 modeled as neutral and pointing at “pre-bound” position. Figure structure with His157 modeled as neutral and pointing at “pre-bound” position. Figure reproduced reproduced with permission from [51]. with permission from [51].
The coordination sphere for Cu indicates the presence of Cu(II), despite the possibility of The coordination for Cu indicates Cu(II), despite the presence possibility photoreduction uponsphere X-ray exposure. Treatingthe thepresence crystals of with ascorbate in the of of photoreduction upon prior X-raytoexposure. Treating the crystals ascorbateofinthe theCu(II)-dioxo presence of atmospheric oxygen freezing leads to reduction to Cu(I)with and formation 2− atmospheric oxygen prior to freezing leads to reduction to Cu(I) and formation of the Cu(II)-dioxo complex. The crystal structure shows that H2Oeq is displaced by a peroxo (O2 ) species in the equatorial complex. crystal structure shows that H2 Oeq isisdisplaced a peroxo (O2 2− ) species position The (bond length ~1.44 Å). This configuration consistentbywith displacement by waterintothe equatorial position in (bond length ~1.44 Å). ThisInconfiguration is consistentsymmetry with displacement by water release peroxide the absence of substrate. the non-crystallographic related monomer molecular oxygeninwas instead observed, adjacent H2Oax. The molecular oxygen in this position to release peroxide the absence of substrate. In the to non-crystallographic symmetry related monomer was proposed be in a “pre-bound” and not as wasoxygen seen ininthe adjacent molecular oxygentowas instead observed,position adjacent to H The molecular this position ax . activated 2 Oyet molecule [51].
Molecules 2017, 22, 596
19 of 26
was proposed to be in a “pre-bound” position and not yet activated as was seen in the adjacent molecule [51]. The neutron crystal structure of the resting state showed a nearby His157 to be neutral. At lower pH this His157 would be protonation and positioned to coordinate molecular oxygen in this “pre-bound” location. This implies that His157 could possibly promote oxygen binding and activation. DFT calculations were performed on different His157 protonation states and conformations and showed a strong thermodynamically favored addition of oxygen when His157 is charged. Overall these studies revealed additional details about oxygen binding and activation not considered before and the observation of “pre-bound” molecular oxygen indicate a mechanism for H2 O2 formation in the absence of substrate [51]. 6. Ligand Binding and Substrate Specificity through H Bonds Inhibitor binding to target proteins or enzymes can involve H atoms in different ways: through direct H bonding, water-mediated H bonding, C-H . . . π interactions, hydrophobic interactions, and electrostatic interactions (e.g., deprotonated Lys or His residues). In addition to the role of H atoms in the protein and solvent, the ligand’s own protonation state will have a large effect on solubility and binding properties. While X-ray crystallography has been a powerful tool in ligand binding studies and for rational drug design, it has limitations on the details of the information it can provide. Neutrons are severely underutilized in this field and can fill the knowledge gap by providing atomic information on ligand protonation and detailed information on how they bind to the target protein. 6.1. Rational Drug Design Using Neutrons as a Guide HIV protease (PR) is an aspartic protease and an integral part of the viral life cycle. PR cleaves the pre-cursor proteins Gag and Gag-Pol into functional, catalytic enzymes involved in viral replication and the structure of the viral capsid. Upon PR inhibition, the resulting virus particles are immature and non-infectious. As such PR is an important drug target for suppression and control of HIV infection and progression to AIDS. PR serves as a standout example of successful structure-based drug design efforts using X-ray crystallography. Since the first reported PR crystal structure, many PR inhibitors (PI) have been developed and are in use today [15]. Neutron studies of PR:PI complexes can give useful information to guide next generation structure-based drug design. Such analysis can reveal unfilled pockets in the active site, the role of water-mediated PI binding, and areas where H-bonding can be introduced, or removed, by chemical modification of new ligands. PR is a homodimer with each monomer containing 99 amino acids. The PR active site is made up of a conserved triad, Asp25-Thr26-Gly27 and Asp250 -Thr260 -Gly270 from the adjacent monomer. The two Asp residues from opposite monomers point at each other and substrate (peptides/proteins or inhibitors) binds between the catalytic Asp residues and the flexible flap regions. The first neutron structure of HIV PR in complex with a clinical drug, amprenavir (APV), was compared to another HIV PR neutron structure in complex with a non-clinical inhibitor KNI-272 [15,52]. In the APV complex neutron structure Asp25 is protonated and donates an H bond to the -OH group of APV. In contrast, Asp250 accepts an H bond from the same -OH. It is interesting that overall the distribution of observed H bonds in the active site varies from the PR:KNI-272 structure. This shows that the location of H or D atoms on catalytic Asp residues, as well as the identity of H bond donors/acceptors, can be influenced strongly by the structure and chemistry of the PI. Previous studies on ultra-high resolution X-ray structures enabled identification of a number of proposed H bonds between the ligand and PR [53]. These interactions are expected to enhance, or even enable, binding interactions. When the nuclear density maps for the PR:APV complex were studied, the inventory of actual H bonds were quite different from those inferred from the ultra-high resolution X-ray studies. In fact, some of the details seen in the neutron structure contradicted some of assigned H bonds and their relative importance in determining ligand binding. The neutron data shows only two strong H bonds. These are found between APV and catalytic Asp residues, and between the
Molecules 2017, 22, 596
20 of 26
tetrahydrofuran moiety and backbone amides of Asp29 and Asp30. There is also a water-mediated H bond between Asp300 and the flap regions. The result is three total H bonds and not four as was deduced from the X-ray data alone [15]. Previous 13 C-NMR data showed that the catalytic Asp residues have mismatched pKa , and that this affects ionization and subsequent H bond potential. By direct investigation of the protonation states of Asp25 and Asp250 , the nuclear density maps support the NMR data in that only Asp25 is protonated [54]. Binding interactions can be improved by ligand derivatization to match the pKa ’s. This can be achieved by incorporating a strong electron-accepting group in the APV molecule. Another strategy is to include chemical groups on the PI to directly engage PR in H bonds and minimize water-mediated H bonding. Such a strategy will also improve the enthalpy of binding. In general the results from neutron studies highlight some issues with structure-based drug design. The overarching one is that X-ray data alone can be misleading from a rational drug design view. Neutrons can fill this gap by giving information on ionization/protonation state of residues, water, and H bonds involved. The data can also be used to estimate the importance of interactions based on distances and angles between donors/acceptors. Another issue is that the chemical and physical basis as to why some drugs bind better than others may be very subtle [15]. For these reasons it is important to use multiple methods or approaches to ligand/inhibitor binding interactions. Neutrons clearly have an important future role in structure-based drug design and this may be one of the best applications of the technique in structural biology. With new sources and improved instruments the practical feasibility of using NPX for drug design studies will only increase. 6.2. Ordered Water Network and H Bonds Provide a Template and Specificity for Substrate Binding Carbohydrate binding modules (CBMs) are domains that are part of carbohydrate-metabolizing enzymes. They confer substrate specificity and increase enzymatic throughput by aiding enzymatic recognition of substrates and also by tethering the catalytic part of the enzyme close to the substrate [55]. Some CBMs only bind to a particular carbohydrate while others display more promiscuous binding and understanding the details will be useful in designing enzyme with specific properties, especially for industrial application (e.g., biorefinery processes). Neutron protein crystallography is a great tool for the investigation of substrate specificity. For example, it yields important clues in the case of one such engineered CBM, X-2 L110F, from a xylanase that is expressed in Rhodothermus marinus [56]. Wild type X-2 specifically binds to xylan. Mutation of a single residue Leu110 to a Phe (X-2 L110F) changes the substrate specificity to bind not only xylan, but also xyloglucan, and β-glucan. For the neutron studies the CBM was complexed to a branched heptasaccharide (XXXG) to probe the interactions to substrate that enable broader binding to more substrates. Even though H atoms are known to be important mediators of CBM binding to carbohydrates, there is little direct evidence of their involvement in H bonds, water mediated binding, π-stacking, and other electrostatic interactions. XXXG is a heptasaccharide that contains a chain of four glucose molecules and three branched xylose molecules. Two of the glucose molecules in XXXG are involved in H bonds and a π-stacking interaction to two Phe residues while the other two are not engaged in any strong or visible interactions. An H bond was observed between Tyr149 and one of the branching xyloses, XYS1173, and explains why mutation of this residue disrupts binding ability to branched sugars. XYS1173 is also involved in two additional H bonds to His146 and Arg115 and in a π-stacking interaction with Pro68. Neutron data shows that Arg115 is oriented to donate the H bond through a network of interactions that results in optimal positioning for this interaction. The neutron studies revealed the interactions surrounding branched xylose XYS1173 and support the inability of the L110F mutant to bind sugars that are further substituted at this position (e.g., XLLG) [57]. High resolution X-ray crystallographic studies of wild type X-2 and apo X-2 L110F show a number of ordered water molecules engaged in a network occupy the ligand binding cleft [58]. At least ten of these waters are displaced upon ligand binding giving an entropic enhancement to binding. In addition, the oxygen atoms of these waters are precisely replaced by sugar oxygen atoms upon
Molecules 2017, 22, 596
21 of 26
binding, suggesting that waters serve as a chemical template for ligand binding. Xylan and xyloglucan Molecules 2017, 22, 596 21 of 25 displace a similar number of waters, but XXXG removes an additional three water molecules and causes compensatory amino acid side chain movements. Three conserved water molecules that these waters and the H bonds to the protein serve to orient the residues to be in optimal H bonding form an H bonded network were observed between BGC1171 and several ligand binding residues. configurations. Together these waters and the H bonds to the protein serve to orient the residues to be in optimal H This work highlighted several important considerations in carbohydrate binding interactions. bonding configurations. Mainly it is proper positioning of interacting amino acid side chains and H bonds to ordered water This work highlighted several important considerations in carbohydrate binding interactions. that give clues as to how a carbohydrate ligand may bind. These factors are also involved in Mainly it is proper positioning of interacting amino acid side chains and H bonds to ordered water that understanding substrate specificity and could be exploited to design CBM for specific carbohydrate give clues as to how a carbohydrate ligand may bind. These factors are also involved in understanding recognition [57]. substrate specificity and could be exploited to design CBM for specific carbohydrate recognition [57].
6.3. Water Displacement, Drug Ionization, and H Bonds in Mediating Drug Binding 6.3. Water Displacement, Drug Ionization, and H Bonds in Mediating Drug Binding Human carbonic anhydrases are expressed in diverse and cellthroughout types throughout Human carbonic anhydrases are expressed in diverse tissuestissues and cell types the body.the body. Due to the fundamental nature of the catalyzed reaction, carbon dioxide hydration, CAs are Due to the fundamental nature of the catalyzed reaction, carbon dioxide hydration, CAs are involved involved in a number of physiological processes. They are at the center of many pH homeostasis in a number of physiological processes. They are at the center of many pH homeostasis processes, processes, gluconeogenesis, bone remodeling, cerebrospinal fluid production and so on. One human gluconeogenesis, bone remodeling, cerebrospinal fluid production and so on. One human isoform, isoform, HCA IX, has been found overexpressed in numerous cancer cells and are now an attractive HCA IX, has been found overexpressed in numerous cancer cells and are now an attractive drug target drug target for cancer detection, imaging, staging, and treatment [59,60]. Carbonic anhydrase for cancer detection, imaging, staging, and treatment [59,60]. Carbonic anhydrase inhibitors (CAIs) inhibitors (CAIs) have been prescribed for decades as diuretics and for the treatment of glaucoma. have been prescribed for decades as diuretics and for the treatment of glaucoma. Today there are over Today there are over twenty CAI in clinical use. In recent years their use has been expanded for use twenty CAI in clinical use. In recent years their use has been expanded for use as anti-epilepsy and as anti-epilepsy and anti-obesity drugs. Clinically used CAIs have a common motif, a sulfonamide anti-obesity drugs. Clinically used CAIs have a common motif, a sulfonamide group that coordinates group that coordinates directly to the active site Zn+2, displacing a critical catalytic water molecule in directly to the active site Zn+2 , displacing a critical catalytic water molecule in the process [60,61]. the process [60,61]. HCAs share a high degree of sequence and structural identity, making it very HCAs share a high degree of sequence and structural identity, making it very challenging to design challenging to design inhibitors that display isoform selectivity. In the case of HCA IX, it would be inhibitors that display isoform selectivity. In the case of HCA IX, it would be more clinically effective to more clinically effective to have an inhibitor specific for HCA IX to minimize binding to the other have an inhibitor specific for HCA IX to minimize binding to the other HCAs with important, central HCAs with important, central physiological functions. Acetazolamide (AZM) is a classic CAI with a physiological functions. Acetazolamide (AZM) is a classic CAI with a Ki of 12 nM for HCA II. AZM Ki of 12 nM for HCA II. AZM has three possible protonation states in solution. It was unknown has three possible protonation states in solution. It was unknown which form binds to CA under which form binds to CA under physiological pH conditions. High-resolution X-ray crystallographic physiological pH conditions. High-resolution X-ray crystallographic studies did not reveal the exact studies did not reveal the exact nature of the molecule in terms of ionization state or how specifically nature of the molecule in terms of ionization state or how specifically it binds to the active site of it binds to the active site of HCA II. HCA II.
Figure Comparison of AZM MZM binding water displacement as observed in neutron Figure 16. 16. Comparison of AZM andand MZM binding andand water displacement as observed in neutron crystallographic studies. (A) AZM complex; (B) MZM complex [60]. Figure reproduced, with crystallographic studies. (A) AZM complex; (B) MZM complex [60]. Figure reproduced, with permission from International Union of Crystallography from permission from the the International Union of Crystallography from [60].[60].
A 2.0 Å neutron structure of HCA in complex AZM revealed in complex this complex A 2.0 Å neutron structure of HCA II inIIcomplex withwith AZM revealed that that AZMAZM in this is is anionic with the unprotonated sulfonamide group coordinated to the active site zinc. In contrast anionic with the unprotonated sulfonamide group coordinated to the active site zinc. In contrast to the to the negatively sulfonamide group, the acetamido group is protonated with clearly visible negatively chargedcharged sulfonamide group, the acetamido group is protonated with clearly visible nuclear nuclear density for the exchanged D atom at this position (Figure 16A). The single D atom on the density for the exchanged D atom at this position (Figure 16A). The single D atom on the sulfonamide sulfonamide group is engaged as H bond donor to Thr199. Thr200 acts as a bifurcated H bond donor group is engaged as H bond donor to Thr199. Thr200 acts as a bifurcated H bond donor to an ordered to anmolecule ordered and solvent molecule and the backbone of Pro201. In this way the solventtomolecule solvent the backbone of Pro201. In this way the solvent molecule is poised act as H is poised to act as H bond acceptor from the acetamido group of AZM. A total of four water molecules are displaced upon AZM binding and this will lead to entropic gains in drug binding. Finally the data also showed two C-H…π interactions. The first is between the AZM thiadiazole ring and Leu198, the other between the AZM methyl group and Pro201 itself [59,60].
Molecules 2017, 22, 596
22 of 26
bond acceptor from the acetamido group of AZM. A total of four water molecules are displaced upon AZM binding and this will lead to entropic gains in drug binding. Finally the data also showed two C-H . . . π interactions. The first is between the AZM thiadiazole ring and Leu198, the other between the AZM methyl group and Pro201 itself [59,60]. Another neutron study investigated the drug methazolamide (MZM) bound to HCA II [60]. MZM is a methyl derivative of AZM and has superior properties in terms of shelf life, stability, and is less water-soluble than AZM, making it better suited for use in glaucoma treatment. Unsurprisingly the data showed that MZM was also unprotonated on the sulfonamide moiety and has similar coordination to the zinc as AZM (Figure 16B). MZM displaced a total of seven water molecules, compared to four by AZM. Three of the water molecules are common between the two structures. This means that in the AZM complex there are still a number of ordered, H bonded water molecules that are absent in the room temperature MZM complex. In the MZM situation there is an enthalpic trade-off that involves the breaking of so many H bonds to expel the waters, this balances the entropic gain of releasing water to the bulk solvent. This explains why the Ki for HCA II of MZM is still very similar as AZM (i.e., ~10 nM). The added methyl group on the thiadiazole group of MZM disrupts the water-mediated H bond that was observed between AZM and Thr200 and Pro201 [59,60]. Detailed knowledge of H bonds and water networks in ligand binding is valuable and neutron crystallography will have increasing importance in rational drug design that can exploit this knowledge. Having quantitative information on H bonds can be used for example to modify or derivatize compounds to introduce compensatory interactions or to produce a desired ionization form of the ligand. 7. Conclusions This review highlights specific areas of structural enzymology and general protein biochemistry where neutrons have made a large impact. Due to their unique nature and interactions with biological materials, neutrons have the ability to give high resolution, detailed, unambiguous information on H atoms and their all-important interaction, the H bond. Neutron crystallography has provided the opportunity to “see” into the heart of the enzyme active sites or ligand binding regions, and often brings new insights. Neutrons have contributed direct evidence of amino acid involvement in catalysis or drug binding, how waters are organized and oriented, and even how some waters are activated to perform catalysis. Arguably the most important thing neutrons can show us is the H bond and how it ties the protein, solvent, and/or ligand together. As illustrated in this review, neutron crystallography complements results from NMR spectroscopy, X-ray crystallography, modeling, and computational chemistry. In recent years, significant advances have been made in the fields of DFT calculations and 1 H-NMR spectroscopy for the study of H bonds [62,63]. It can be expected that theoretical and experimental information gained from all these methods will impact how protein structures are refined and analyzed. Neutron crystallography is under-utilized by the general structural biology community due to perceived issues with sample preparation and acquiring beam time. With current molecular biology tools and equipment for protein production, purification and crystallization, there are now far fewer technical limitations to preparing sufficient material to grow large enough crystals for neutron diffraction experiments. Current instruments routinely collect high quality data from crystals that are 0.1–1 mm3 and smaller in volume, a range that is readily attainable for many systems. The low number of instruments and highly competitive nature of being awarded beam time, especially compared to macromolecular X-ray beamlines, represents the true bottleneck for current and new users. With the advent of the new European Spallation Source (Lund, Sweden), and the planned Macromolecular Diffractometer NMX there, data collection and sample requirements will dramatically improve for the user community. Acknowledgments: SZF and EO are partially funded by Interreg ESS & MAX IV Cross Border Science and Society and SZF by the SINE2020 grant (under Horizon 2020 grant agreement No. 654000).
Molecules 2017, 22, 596
23 of 26
Conflicts of Interest: The authors declare no conflict of interest.
References 1. 2. 3. 4. 5. 6.
7. 8. 9.
10.
11. 12. 13. 14. 15.
16.
17. 18. 19.
20.
Hubbard, R.E.; Haider, M.K. Hydrogen bonds in proteins: Role and strength. In Encyclopedia of Life Science; John Wiley & Sons Ltd.: Chichester, UK, 2010. Martin, T.W.; Derewenda, S. The name is bond—H bond. Nat. Struct. Biol. 1999, 6, 403–405. [CrossRef] [PubMed] McDonald, I.K.; Thornton, J.M. Satisfyin hydrogen bonding potential in proteins. J. Mol. Biol. 1994, 238, 777–793. [CrossRef] [PubMed] Tuttle, T.; Grafenstein, J.; Wu, A.; Kraka, E.; Cremer, D. Analysis of the NMR spin-spin coupling mechanism across a Hbond: Nature of the H-bond in proteins. J. Phys. Chem. 2004, 108, 1115–1129. [CrossRef] Baker, E.N.; Hubbard, R.E. Hydrogen bonding in globular proteins. Prog. Biophys. Mol. Biol. 1984, 44, 97–179. [CrossRef] Kumar, M.; Balaji, P.V. C-H . . . pi interactions in proteins: Prevalence, pattern of occurrence, residue propensities, location, and contribution to protein stability. J. Mol. Model. 2014, 20, 2136–2150. [CrossRef] [PubMed] Plevin, M.J.; Bryce, D.L.; Boisbouvier, J. Direct detection of CH/π interactions in proteins. Nat. Chem. 2010, 2, 466–471. [CrossRef] [PubMed] Torshin, I.Y.; Weber, I.T.; Harrison, R.W. Geometric criteria of hydrogen bonds in proteins and identification of “bifurcated” hydrogen bonds. Prot. Eng. 2002, 15, 359–363. [CrossRef] Michalczyk, R.; Unkefer, C.J.; Bacik, J.P.; Schrader, T.E.; Ostermann, A.; Kovalevsky, A.Y.; McKenna, R.; Fisher, S.Z. Joint neutron crystallographic and NMR solution studies of Tyr residue ionization and hydrogen bonding: Implications for enzyme-mediated proton transfer. Proc. Natl. Acad. Sci. USA 2015, 112, 5673–5678. [CrossRef] [PubMed] Yokoyama, T.; Mizuguchi, M.; Ostermann, A.; Kusaka, K.; Niimura, N.; Schrader, T.E.; Tanaka, I. Protonation State and Hydration of Bisphosphonate Bound to Farnesyl Pyrophosphate Synthase. J. Med. Chem. 2015, 58, 7549–7556. [CrossRef] [PubMed] Blakeley, M.P.; Hasnain, S.S.; Antonyuk, S.V. Sub-atomic resolution X-ray crystallography and neutron crystallography: Promise, challenges and potential. IUCrJ 2015, 2, 464–474. [CrossRef] [PubMed] O’Dell, W.B.; Bodenheimer, A.M.; Meilleur, F. Neutron protein crystallography: A complementary tool for locating hydrogens in proteins. Arch. Biochem. Biophys. 2016, 602, 48–60. [CrossRef] [PubMed] NIST Website for Neutron Scattering Lengths. Available online: http://www.ncnr.nist.gov/resources/nlengths/ (accessed on 17 January 2017). Niimura, N.; Bau, R. Neutron protein crystallography: Beyond the folding structure of biological macromolecules. Acta Crystallogr. A 2008, 64, 12–22. [CrossRef] [PubMed] Weber, I.T.; Waltman, M.J.; Mustyakimov, M.; Blakeley, M.P.; Keen, D.A.; Ghosh, A.K.; Langan, P.; Kovalevsky, A.Y. Joint X-ray/neutron crystallographic study of HIV-1 protease with clinical inhibitor amprenavir: Insights for drug design. J. Med. Chem. 2013, 56, 5631–5635. [CrossRef] [PubMed] Fisher, S.J.; Blakeley, M.P.; Howard, E.I.; Petit-Haertlein, I.; Hartlein, M.; Mitschler, A.; Cousido-Siah, A.; Salvay, A.G.; Popov, A.; Muller-Dieckman, C.; et al. Perdeuteration: Improved visualization of solvent structure in neutron macromolecular crystallography. Acta Crystallogr. D 2014, 70, 3266–3272. [CrossRef] [PubMed] Langan, P.; Chen, J.C.-H. Seeing the chemistry in biology with neutron crystallography. Phys. Chem. Chem. Phys. 2013, 15, 13705–13712. [CrossRef] [PubMed] Chen, J.C.-H.; Unkefer, C.J. Fifteen years of the Protein Crystallography Station: The coming of age of macromolecular neutron crystallography. IUCrJ 2017, 4, 72–86. [CrossRef] [PubMed] Blakeley, M.P.; Teixeira, S.C.; Petit-Haertlein, I.; Hazemann, I.; Mitschler, A.; Haertlein, M.; Howard, E.; Podjarny, A.D. Neutron macromolecular crystallography with LADI-III. Acta Crystallogr. D 2010, 66, 1198–1205. [CrossRef] [PubMed] Chen, J.C.-H.; Hanson, B.L.; Fisher, S.Z.; Langan, P.; Kovalevsky, A.Y. Direct observation of hydrogen atom dynamics and interactions by ultrahigh resolution neutron protein crystallography. Proc. Natl. Acad. Sci. USA 2012, 109, 15301–15306. [CrossRef] [PubMed]
Molecules 2017, 22, 596
21. 22.
23. 24.
25. 26.
27. 28.
29.
30.
31. 32. 33.
34.
35.
36.
37. 38.
39.
24 of 26
Schmidt, A.; Teeter, M.; Weckert, E.; Lamzin, V.S. Crystal structure of small protein crambin at 0.48 Å resolution. Acta Crystallogr. F 2011, 67, 424–428. [CrossRef] [PubMed] Rosenbaum, G.; Ginell, S.L.; Chen, J.C.-H. Energy optimization of a regular macromolecular crystallography beamline for ultra-high-resolution crystallography. J. Synchrotron Radiat. 2015, 22, 172–174. [CrossRef] [PubMed] Silverman, D.N.; McKenna, R. Solvent-mediated proton transfer in catalysis by carbonic anhydrase. Acc. Chem. Res. 2007, 40, 669–675. [CrossRef] [PubMed] Fisher, Z.; Kovalevsky, A.Y.; Mustyakimov, M.; Silverman, D.N.; McKenna, R.; Langan, P. Neutron structure of human carbonic anhydrase II: A hydrogen-bonded water network “switch” is observed between pH 7.8 and 10.0. Biochemistry 2011, 50, 9421–9423. [CrossRef] [PubMed] Dedecker, P.; de Schryver, F.C.; Hofkens, J. Fluorescent proteins: Shine on, you crazy diamond. J. Am. Chem. Soc. 2013, 135, 2387–2402. [CrossRef] [PubMed] Langan, P.S.; Close, D.W.; Coates, L.; Rocha, R.C.; Ghosh, K.; Kiss, C.; Waldo, G.; Freyer, J.; Kovalevsky, A.; Bradbury, A.R. Evolution and characterization of a new reversibly photoswitching chromogenic protein, Dathail. J. Mol. Biol. 2016, 428, 1776–1789. [CrossRef] [PubMed] Anderson, S.; Crosson, S.; Moffat, K. Short hydrogen bonds in photoactive yellow protein. Acta Crystallogr. D 2004, 60, 1008–1016. [CrossRef] [PubMed] Yamaguchi, S.; Kamikubo, H.; Kurihara, K.; Kuroki, R.; Niimura, N.; Shimizu, N.; Yamazaki, Y.; Kataoka, M. Low-barrier hydrogen bond in photoactive yellow protein. Proc. Natl. Acad. Sci. USA 2009, 106, 440–444. [CrossRef] [PubMed] Fisher, S.Z.; Anderson, K.; Henning, R.; Moffat, K.; Langan, P.; Thiyagarajan, P.; Schultz, A.J. Neutron and X-ray structural studies of short hydrogen bonds in photoactive yellow protein (PYP). Acta Crystallogr. D 2007, 63, 1178–1184. [CrossRef] [PubMed] Tamada, T.; Kinoshita, T.; Kurihara, K.; Adachi, M.; Ohhara, T.; Imai, K.; Kuroki, R.; Tada, T. Combined high-resolution neutron and X-ray analysis of inhibited elastase confirms the active site oxyanion hole but rules against a low-barrier hydrogen bond. J. Am. Chem. Soc. 2009, 131, 11033–11040. [CrossRef] [PubMed] Cleland, W.W.; Kreevoy, M.M. Low-barrier hydrogen bonds and enzymic catalysis. Science 1994, 264, 1887–1890. [CrossRef] [PubMed] Kuhn, P.; Knapp, M.; Solitis, S.M.; Ganshaw, G.; Thoene, M.; Bott, M. The 0.78 Å structure of a serine protease: Bacillus lentus subtilisin. Biochemistry 1998, 37, 13446–13452. [CrossRef] [PubMed] Schmidt, A.; Jelsch, C.; Ostergaard, P.; Rypniewski, W.; Lamzin, V.S. Trypsin revisited: Crystallography at sub-atomic resolution and quantum chemistry revealing details of catalysis. J. Biol. Chem. 2003, 278, 43357–43362. [CrossRef] [PubMed] Xu, Y.; Wu, J.; Gorenstein, D.; Braun, W. Automated 2D NOESY assignment and structure calculation of crambin(S22/I25) with the self-correcting distance geometry based NOAH/DIAMOD programs. J. Magn. Reson. 1999, 136, 76–85. [CrossRef] [PubMed] Kovalevsky, A.Y.; Hanson, L.; Fisher, S.Z.; Mustyakimov, M.; Mason, S.A.; Forsyth, V.T.; Blakeley, M.P.; Keen, D.A.; Wagner, T.; Carrell, H.L.; et al. Metal ion roles and the movement of hydrogen during reaction catalyzed by D-xylose isomerase: A joint X-ray and neutron diffraction study. Structure 2010, 18, 688–699. [CrossRef] [PubMed] Kovalevsky, A.Y.; Hanson, B.L.; Mason, S.A.; Yoshida, T.; Fisher, S.Z.; Mustyakimov, M.; Forsyth, V.T.; Blakeley, M.P.; Keen, D.A.; Langan, P. Identification of the elusive hydronium ion exchanging roles with a proton in an enzyme at lower pH values. Angew. Chem. Int. Ed. Engl. 2011, 50, 7520–7523. [CrossRef] [PubMed] Asboth, B.; Naray-Szabo, G. Mechanism of action of D-xylose isomerase. Curr. Protein Pept. Sci. 2000, 1, 237–254. [CrossRef] [PubMed] Carrell, H.L.; Glusker, J.P.; Burger, P.V.; Manfre, F.; Tritsch, D.; Biellmann, J.F. X-ray analysis of D-xylose isomerase at 1.9 Å: Native enzyme in complex with substrate and with a mechanism-designed inactivator. Proc. Natl. Acad. Sci. USA 1989, 86, 4440–4444. [CrossRef] [PubMed] Cuypers, M.G.; Mason, S.A.; Blakeley, M.P.; Mitchell, E.P.; Haertlein, M.; Forsyth, V.T. Near-atomic resolution neutron crystallography on perdeuterated Pyrococcus furiosus rubredoxin: Implication of hydronium ions and protonation state equilibria in redox changes. Angew. Chem. 2013, 52, 1022–1025. [CrossRef] [PubMed]
Molecules 2017, 22, 596
40.
41.
42.
43.
44.
45.
46.
47.
48.
49.
50.
51. 52.
53.
54.
55.
25 of 26
Kwon, H.; Basran, J.; Casadei, C.M.; Fielding, A.J.; Schrader, T.E.; Ostermann, A.; Devos, J.M.; Aller, P.; Blakeley, M.P.; et al. Direct visualization of a Fe(IV)-OH intermediate in a heme enzyme. Nat. Commun. 2016, 7, 13445. [CrossRef] [PubMed] Casadei, C.M.; Gumiero, A.; Metcalfe, C.L.; Murphy, E.J.; Basran, J.; Concilio, M.G.; Teixeira, S.C.; Schrader, T.E.; Fielding, A.J.; Ostermann, A.; et al. Heme enzymes. Neutron cryo-crystallography captures the protonation state of ferryl heme in a peroxidase. Science 2014, 345, 193–197. [CrossRef] [PubMed] Oksanen, E.; Blakeley, M.P.; El-Hajji, M.; Ryde, U.; Budayova-Spano, M. The neutron structure of urate oxidase resolves a long-standing mechanistic conundrum and reveals unexpected changes in protonation. PLoS ONE 2014, 9, e86651. [CrossRef] [PubMed] Gabison, L.; Chopard, C.; Colloc’h, N.; Peyrot, F.; Castro, B.; Hajji, M.E.; Altarsha, M.; Monard, G.; Chiadmi, M.; Prangé, T. X-ray, ESR, and quantum mechanics studies unravel a spin well in the cofactor-less urate oxidase. Proteins 2011, 79, 1964–1976. [CrossRef] [PubMed] Chen, Y.; Delmas, J.; Sirot, J.; Shoichet, B.; Bonnet, R. Atomic resolution structures of the CTX-M β-lactamases: Extended spectrum activities form increased mobility and decreased stability. J. Mol. Biol. 2005, 348, 349–362. [CrossRef] [PubMed] Vandavasi, V.G.; Weiss, K.L.; Cooper, J.B.; Erskine, P.T.; Tomanicek, S.J.; Ostermann, A.; Schrader, T.E.; Ginell, S.L.; Coates, L. Exploring the Mechanism of β-Lactam Ring Protonation in the Class A β-lactamase Acylation Mechanism Using Neutron and X-ray Crystallography. J. Med. Chem. 2016, 59, 474–479. [CrossRef] [PubMed] Strynadka, N.C.; Adachi, H.; Jensen, S.E.; Johns, K.; Sielecki, A.; Betzel, C.; Sutoh, K.; James, M.N. Molecular structure of the acyl-enzyme intermediate in beta-lactam hydrolysis at 1.7 A resolution. Nature 1992, 359, 700–705. [CrossRef] [PubMed] Blum, M.-M.; Mustyakimov, M.; Ruterjans, H.; Kehe, K.; Schoenborn, B.P.; Langan, P.; Chen, J.C.-H. Rapid determination of hydrogen positions and protonation states of diisopropyl fluorophosphatase by joint neutron and X-ray diffraction refinement. Proc. Natl. Acad. Sci. USA 2009, 106, 713–718. [CrossRef] [PubMed] Blum, M.-M.; Lohr, F.; Richardt, A.; Ruterjans, H.; Chen, J.C.-H. Binding of a designed substrate analogue to diisopropyl fluorophosphatase: Implications for the phosphotriesterase mechanism. J. Am. Chem. Soc. 2006, 128, 12750–12757. [CrossRef] [PubMed] Chen, J.C.-H.; Mustyakimov, M.; Schoenborn, B.P.; Langan, P.; Blum, M.M. Neutron structure and mechanistic studies of diisopropyl fluorophosphatase (DFPase). Acta Crystallogr. D 2010, 66, 1131–1138. [CrossRef] [PubMed] Kjaergaard, C.H.; Qayyum, M.F.; Wong, S.D.; Xu, F.; Hemsworth, G.R.; Walton, D.J.; Young, N.A.; Davies, G.J.; Walton, P.H.; Johansen, K.S.; et al. Spectroscopic and computational insight into the activation of O2 by the mononuclear Cu center in polysaccharide monooxygenases. Proc. Natl. Acad. Sci. USA 2014, 111, 8797–8802. [CrossRef] [PubMed] O’Dell, W.B.; Agarwal, P.K.; Meilleur, F. Oxygen Activation at the Active Site of a Fungal Lytic Polysaccharide Monooxygenase. Angew. Chem. Int. Ed. 2017, 56, 767–770. [CrossRef] [PubMed] Adachi, M.; Ohhara, T.; Kurihara, K.; Tamada, T.; Honjo, E.; Okazaki, N.; Arai, S.; Shoyama, Y.; Kimura, K.; Matsumura, H.; et al. Structure of HIV-1 protease in complex with potent inhibitor KNI-272 determined by high-resolution X-ray and neutron crystallography. Proc. Natl. Acad. Sci. USA 2009, 106, 4641–4646. [CrossRef] [PubMed] Kovalevsky, A.Y.; Liu, F.; Leshchenko, S.; Ghosh, A.K.; Louis, J.M.; Harrison, R.W.; Weber, I.T. Ultrahigh resolution crystal structure of HIV-1 protease mutant reveals two binding sites for clinical inhibitor TMC114. J. Mol. Biol. 2006, 363, 161–173. [CrossRef] [PubMed] Wang, Y.X.; Freedberg, D.I.; Yamazaki, T.; Wingfield, P.T.; Stahl, S.J.; Kaufman, J.D.; Kiso, Y.; Torchia, D.A. Solution NMR evidence that the HIV-1 protease catalytic aspartyl groups have different ionization states in the complex formed with the asymmetric drug KNI-272. Biochemistry 1996, 35, 9945–9950. [CrossRef] [PubMed] Hervé, C.; Rogowski, A.; Blake, A.W.; Marcus, S.E.; Gilbert, H.J.; Knox, J.P. Carbohydrate-binding modules promote the enzymatic deconstruction of intact plant cell walls by targeting and proximity effects. Proc. Natl. Acad. Sci. USA 2010, 107, 15293–15298. [CrossRef] [PubMed]
Molecules 2017, 22, 596
56.
57.
58.
59.
60.
61. 62. 63.
26 of 26
Karlsson, E.N.; Bartonek-Roxå, E.; Holst, O. Cloning and sequence of a thermostable multidomain xylanase from the bacterium Rhodothermus marinus. Biochim. Biophys. Acta Gene Struct. Expr. 1997, 1353, 118–124. [CrossRef] Fisher, S.Z.; von Schantz, L.; Håkansson, M.; Logan, D.T.; Ohlin, M. Neutron crystallographic studies reveal hydrogen bond and water-mediated interactions between a carbohydrate-binding module and its bound carbohydrate ligand. Biochemistry 2015, 54, 6435–6438. [CrossRef] [PubMed] Von Schantz, L.; Håkansson, M.; Logan, D.T.; Walse, B.; Österlin, J.; Karlsson, E.N.; Ohlin, M. Structural basis for carbohydrate-binding specificity—A comparative assessment of two engineered carbohydrate-binding modules. Glycobiology 2012, 22, 948–961. [CrossRef] [PubMed] Fisher, S.Z.; Aggarwal, M.; Kovalevsky, A.Y.; Silverman, D.N.; McKenna, R. Neutron diffraction of acetazolamide-bound human carbonic anhydrase II reveals atomic details of drug binding. J. Am. Chem. Soc. 2012, 134, 14726–14729. [CrossRef] [PubMed] Aggarwal, M.; Kovalevsky, A.Y.; Velazquez, H.; Fisher, S.Z.; Smith, J.C.; McKenna, R. Neutron structure of human carbonic anhydrase II in complex with methazolamide: Mapping the solvent and hydrogen-bonding patterns of an effective clinical drug. IUCrJ 2016, 3, 319–325. [CrossRef] Supuran, C.T.; Scozzafava, A.; Conway, J. Carbonic Anhydrase—Its Inhibitors and Activators; CRC Press: Boca Raton, FL, USA, 2004; Volume 1. Deringer, V.L.; Hoepfner, V.; Dronskowski, R. Accurate hydrogen positions in organic crystals: Assessing a quantum-chemical aide. Cryst. Growth Des. 2012, 12, 1014–1021. [CrossRef] Siskos, M.G.; Choudhary, M.I.; Gerothanassis, I.P. Hydrogen Atomic Positions of O–H . . . O Hydrogen Bonds in Solution and in the Solid State: The Synergy of Quantum Chemical Calculations with 1H-NMR Chemical Shifts and X-ray Diffraction Methods. Molecules 2017, 22, 415. [CrossRef] [PubMed] © 2017 by the authors. Licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY) license (http://creativecommons.org/licenses/by/4.0/).
