The Functional Relevance of π-hole Interactions with Nitrate Anions

(1)

MSc Chemistry

Molecular Sciences

Literature Thesis

The Functional Relevance of π-hole Interactions

with Nitrate Anions

by

Wouter Hageman

10529039

September 2019

12 EC

June – September 2019

Supervisor/Examiner:

2

nd

_Examiner:

dr. T.J. Mooibroek

dr. J.C. Slootweg

Van ‘t Hoff Institute of Molecular Sciences

(2)

(3)

3

Abstract

This review provides an overview of the broadness of the uses of σ- and π-hole interactions that have been reported over the years. These supramolecular interactions have been found to be the basis of applications regarding crystal engineering and structural integrities in biological systems. A description of these studies is given to gain insight on the applicability of the multiple types of σ- and π-hole interactions in chemistry and biology. In particular, the relevance of π-hole interactions between nitrate anions and carbonyl groups in proteins has been investigated. These interactions appeared to have a consistent structural relevance in three groups of enzymes: N-acetylglucosamine deacetylases, tyrosinases and kinases. It can also be speculated that the group of N-acetylglucosamine deacetylases possesses some functional relevance. In addition to the kinase group, one protein (3ezh) had a similar nitrate binding pocket, which showed that the π-hole interactions of the nitrate anion have influence towards the function of this enzyme that is involved in the Lipid A biosynthesis.

(4)

(5)

5

1. Introduction

The wide range of possibilities to discover new information on inter- and intramolecular interactions is one of the reasons causing increases attention towards the field of supramolecular chemistry. The relevance of the non-covalent forces can be observed in various biological, medicinal and chemical processes.1_{Understanding of these forces leads to better definition of}

systems, such as specific recognition, transport, regulation, crystal structure engineering and characterization. Among these forces, the electrostatic interactions have received much attention in recent studies, specifically the σ-hole and π-hole interactions.2,3_{The directional}

strength of these interactions and with that their relevance towards biological processes and synthetic design, is the reason for the substantial growth of interest into such systems. The identification of their relevance originates from studies that performed extensive analyses on structures from the Cambridge Structural Database (CSD) and the Protein Data Bank.2,4,5

Hydrogen bonding is considered to be a large contributor in structural networks that are held together by non-covalent interactions and can be viewed as the prototype for σ- and π-hole interactions. Additionally, the manner of interaction of both hydrogen bonding and σ-hole interactions are alike, as the donor/acceptor principles are equal, except for the exchange of the hydrogen atom with an atom that contains the σ-hole. The general view of these σ-hole bonding interaction can be described as D-X⸳⸳⸳A, where D stands for the donor atom, X for the atom with the σ-hole and A for the acceptor atom of the non-covalent bond. The donor atom of such interactions is defined as the atom connected to the atom with a positive electrostatic potential and receives electrons from the σ-hole acceptor, which is contrary to the electron transfers in organic donor/acceptor principles. This also means that the donor is the electrophilic moiety and the acceptor is the nucleophilic moiety. This also correlates to Lewis acidity and basicity that indicates the electron accepting or donating properties and with these interactions the donor acts as the Lewis acid and the acceptor as the Lewis base. In this interpretation the σ-hole is placed on the X-atom on the extension of the D-X plane. A σ-hole contains a region of positive electrostatic potential on an atom, for example iodine, that is normally considered to have a negative potential, caused by polarization. One example of σ-hole interactions is halogen bonding, in which the σ-hole is placed on a halogen atom that forms a non-covalent bond with an atom that has a negative electrostatic potential.4–9 The strength of the interactions generally increases going from chlorine to bromine to iodine. The large electronegative value of fluorine was found to neutralize the σ-hole, except for the cases where fluorine is bonded to atoms that

(8)

8 have equal electron-attracting

strength, such as F–CN or FOF. The stronger polarization of the smaller halogens works against this interaction, thus explaining the trend. Chlorine only shows a noticeable interaction when the other substituents on the donor atom have sufficient electron-withdrawing force to enhance the positive electrostatic potential on the chlorine. An example of this interaction is shown in Figure 1, where the σ-hole is shown as

the blue area on the iodine atom. The location of σ-holes coincides with the location of empty anti-bonding σ* orbitals, which is along the axis of the covalent bond with the donor atom. The σ* orbital can have a favorable interaction with any filled orbital that is near this orbital. An example of this is the interaction with a filled p-orbital of the acceptor atom and this can be viewed as a charge transfer from the acceptor p-orbital to the halogen donor D-X σ* orbital. This is confirmed with use of a molecular orbital (MO) theory scheme, which is shown in Figure 2. It becomes clear from this scheme that the p-orbital of the acceptor contributes its electrons into the orbital of the halogen bonded complex. A complete bonding analysis on this interaction was performed by Wolters and Bickelhaupt with density functional theory (DFT) calculations, which proved that these

interactions are not purely based on electrostatic factors, but also from orbital interactions, dispersion and repulsive Pauli exclusion terms.10

The electrostatic factors represent the Coulombic interaction that causes the

Figure 1: The molecular electrostatic potential of the CF3I molecule showing a

σ-hole on the I-atom in the C-I plane, having an energy of 34 kcal mol-1_{. The}

electrostatic potential goes from negative to positive with the transition from red to blue. The image was constructed with Spartan and the MEP was calculated with DFT method wB97X-D/6-31G*.

Figure 2: MO scheme of the orbital interaction in halogen bonding between the σ*

(9)

9 deformation and directionality of the interaction. The orbital interactions account for the charge transfer, which is responsible for the interaction between the occupied and the empty orbitals. The dispersion term corrects for the fluctuation of electrons caused that correspond to the dipole polarization of the interacting atoms. The Paul exclusion or repulsion is the term relating to steric repulsion and the destabilizing effect between the occupied orbitals of the reactants. The σ-hole interactions are dominated by electrostatic interactions but include a combination of all the terms. The view of halogen bonding can also be applied to chalcogen, pnictogen and tetrel bonding interactions, which belong groups IV, V and VI respectively. The atoms of these three groups can also have more than one σ-hole present due to higher covalencies of the atoms in those groups. Halogen, chalcogen, pnictogen and tetrel bonding types can be captured under the name σ-hole bonding.

A phenomenon similar to σ-hole interactions, named π-hole interactions, is an upcoming research subject in the field of supramolecular chemistry. The electrostatic contribution of these π-holes is the same as in σ-holes, namely a region of positive potential. However, this region is

not localized along the extension of a covalent bond, but usually on a central part of the molecular framework, which has a directional influence in a way that is perpendicular to this framework.11_{This directional influence originates from an empty or partially filled π- or π*}

orbitals, which steers the electron-rich component of the interaction in such a perpendicular fashion.2,3_{This influence is usually stronger for heavier atoms, which means that the}

electrostatic potential is larger when using more polarized donor atoms, for example selenium versus sulphur. This also occurs with phosphorous versus nitrogen shown in Figure 3, which clearly depicts a larger electrostatic potential of the π-hole. This interaction is mentioned by researchers under various names depending on the type of system. If the π-hole is localized on an electron-deficient aromatic system, it is usually referred to as lone-pair-π or anion-π interaction, depending on the electron-rich counterpart being neutral or anionic. The discovery of carbonyl interaction by Bürgi et al. was one of the earliest cases describing lone-pair-π interactions of nucleophiles with the carbon atom of a carbonyl group.12 These carbonyl

Figure 3: Molecular electrostatic potentials of pnictogen bond donors,

depicting the π-holes on top and N and P, with energies of 40 and 71 kcal mol

-1 _{respectively. The electrostatic potential goes from negative to positive with the}

transition from red to blue. The image was constructed with Spartan and the MEP was calculated with DFT method wB97X-D/6-31G*.

(10)

10 interactions were observed and used in biological and environmental systems, showing functional relevance in multiple cases, such as cooperative binding in secondary structures of proteins or as receptors used in transport agents.13–16 Although counterintuitive, under specific circumstances a π-hole can also appear in anions that are commonly used as electron donors, of which the nitrate anion might be one to account for. Proof has been found that NO3- can act as

π-hole donor and thus as a Lewis acid.17_{Since nitrate anions are often present in enzymes and}

proteins, this finding may suggest an undiscovered functional role in these structures. This work will present the research up until now that has been performed on σ-hole and π-hole interactions in various systems, followed by a detailed description and search towards the functional relevance of the π-hole interaction of nitrate anions in biological systems with use of current literature and crystal structures from the Brookhaven Protein Data Bank.18

2. Exploration of σ/π-hole interactions

Up until this day, large amounts of experimental and theoretical studies have been performed in order to elucidate the importance and occurrence of σ- and π-hole interactions. The following sections will illustrate examples to confirm the relevance of these interactions.

2.1. σ-Hole interactions

2.1.1. Halogen bonding

Halogen bonding has seen attention of many researchers in the field of biological and supramolecular chemistry due to their directional influence in crystal structures. One of the earlier researches towards this directional influence was carried out in 1986 by Ramasubbu et

al.19 Their work showed the non-covalent interaction of C-X (X = Cl, Br or I) bonds with electrophiles, nucleophiles and other C-X bonds. Metals were used as the electrophiles and Cambridge Crystallographic Data Base structures were compared on the basis of their angle with the halogen atom from the perspective of the C-X plane. This was accounted as a ‘side on’ approach, which was based on the average angle of 104°. The nucleophiles that they investigated, were mainly molecules that connected with an O or N atom and their average interaction angle was around 165°, which is nearly a ‘head on’ approach. This was one of the first indications of the σ-hole, since that approach is also ‘head on’, meaning in the plane of the C-X bond. The halogen-halogen interactions were then observed to be a combination of both

(11)

11 the previous discoveries, leading to the conclusion that these C-X⸳⸳⸳X-C interactions could be seen as an electrophile-nucleophile complex, since the average angle of the first halogen was around 180° and the average angle of the second halogen was around 90°, which are ‘head on’ and ‘side on’ approaches respectively. However, it was found that angles of 180° are rarely observed, since there is a statistical bias causing these angles to be off the preferred 180°. This bias was observed by Kroon et al. in the case of hydrogen bonds, but this is also the case for σ-hole donors.20

Such σ-hole interactions of a halogen atom with either an oxygen or nitrogen atom can also be applied in crystal engineering to create tectons by a self-assembling process. This idea was used by Guardigli et al., who used the halogen bonding of iodine in p-iodopentafluorobenzene together with diols, bisphenols and diamines to produces tectons with varying length, of which one example is given in Figure 4.21_{The iodine atoms in this molecule}

bind the ‘head on’ approach that was mentioned earlier due to the σ-hole located on the iodine itself.

There has also been theoretical research to uncover the binding strength and applicability of these σ-holes. One way this was done, was by performing Density Functional Theory (DFT) calculations to observe this interaction on the CF3I molecule, where the σ-hole

is located on the iodine atom, in the study of Romaniello et al.22 This interaction was investigated in adducts, where the acceptor molecules were (CH3)3N, (CH3)3P, (CH3)2S,

(CH3)3NO, (CH3)3PO and with the solvent DMSO. All adducts were found to have interactions

energies that were comparable to that of a regular hydrogen bond and the adducts with the strongest interactions appeared to be (CH3)3NO, which was bonded via the oxygen atom, with

Figure 4: The X-ray structure of a tecton connection showing the halogen bonding between the I and N atoms in the tecton. Color code: carbon = grey, hydrogen = white, nitrogen = blue, fluorine = red, iodine = purple.

(12)

12 an energy of -7.81

kcal/mol. This work showed that it is theoretically possible to obtain a variety of adducts that contain σ-hole

interactions strong enough to support such structures. Another DFT study focused on the electrostatic interactions in bromine and iodine trimers (RBr3 and RI3, with R = H, H3C, H2FC,

HF2C, F3C, CH2=CH, CH≡C, and Ph).23 In this work, they investigated the geometry, the

binding energies and the vibrational frequency shifts upon formation. All the trimers they investigated showed the ability to form trimers that bonded in the nearly ‘head on’ approach, also shown in Figure 5. It appeared that the binding energy of the iodine trimers were ~ 6 kcal mol-1 stronger relatively to the bromine trimers, but still able to from stable structures. No remarkable vibrational frequency shifts were obtained compared to the monomeric forms. The calculations showed that the electrostatic interaction of the halogen bonding is the dominant force that forms and retains the structure of the trimers.

To gain a better understanding of the impact of halogen bonding in biological systems, Auffinger et al. have modelled halogenated uridine and cytosine and compared this to halogenated methane.24 These models made clear that the σ-hole is strongest for the structure with a substituted iodine atom and also that there is an increase in the size and the strength of the σ-hole for the aromatic nucleobases, which could be an indication that this interaction might play a role in the structures of proteins or nucleic acids. After that, they searched through the PDB for complexes that contained a X⸳⸳⸳A (X = F, Cl, Br, I, A = O, N and S) with a distance shorter than their specific van der Waals distance. In this search, mainly structures with X⸳⸳⸳O occurrences were found and they showed a directional effect of the conformation of the structure they were found in. One of the examples suggest that they could be active in thyroxine and thyroxine derivatives, which is part of the hormonal system and functions as an essential part in the recognition of these hormones by cognate proteins. There was also a displacement of the ATP ligand that was steered by halogen bonding originating from a halogenated inhibitor, which indicates that there might be a role for this type of bonding in the design of drugs.

2.1.2. Chalcogen bonding

Figure 5: The halogen trimer with electrostatic mapping details to show the structure

(13)

13 Together with the rise of interest in halogen

bonding, other element groups were also observed to show similar non-covalent interactions that were viewed through the vision of σ-hole bonding. One of these groups is the chalcogen group, which contains oxygen, sulphur, selenium and tellurium. One of the first calculations towards proving the presence of σ-holes on these atoms were performed by Murray et

al.25 They started by calculating the electrostatic potential on the molecules R2Z, where R =

CH3, CH2F, F and CN and Z = O, S and Se. This resulted in two trends, one for the R groups

and one for the chalcogen groups. The R groups trend was CH3 < CH2F < F < CN, meaning

that CN had the largest influence on the positive potential of the σ-holes. The chalcogen trend was correlating to the polarizability of the atoms, which means that O was the least positive, followed by S and Se had the highest positive potential in every situation with a maximum potential of 46.9 kcal/mol, which is a clear indication of the σ-hole presence, which is shown in Figure 6. They also performed calculations to determine the interaction of the R2Z molecules,

where R = F and CN, with two nitrogen-based molecules, NH3 and HCN. The expected result,

judging from the electrostatic potentials, would be that CN-substituted molecules would have a stronger interaction, which is observed for the interaction with HCN and for OR2 with NH3,

but not for SR2 and SeR2 with NH3. This is likely due to a possible hydrogen bonding interaction

of the second fluorine atom with the H of NH3. This effect also explains why the interactions

are stronger for S and Se with NH3 than with HCN.

These apparent interactions can be used for many purposes, for instance to explain why dimethylsulfoxide (DMSO) and dimethylsulfon (DMSO2) have high solvating strength towards

not only polar molecules, but also towards aromatic compounds. Clark et al. performed a study to decipher the origin of this strength based on electrostatic potentials of the molecules itself compared to dimethylsulfide (DMS) and interactions energies in complexes of the DMSO and DMSO2 dimers and in complexes of DMSO and DMSO2 with water or acetone.26 Besides all

the positive and negative regions in the molecules, there was one clear σ-hole on the sulphur atom of DMSO and there were two on the sulphur atom of DMSO2, which implies that the

solvents contain a site susceptible to electronegative molecules, such as aromatic centers. The

Figure 6: The electrostatic potential mapping on (NC)2Se with the

two σ-holes showing in blue on the Se atom, having an energy of

46.9 kcal mol-1. The electrostatic potential goes from negative to

positive with the transition from red to blue. The image was constructed with Spartan and the MEP was calculated with DFT method wB97X-D/6-31G*.

(14)

14 modelling of the dimer complexes showed that the σ-holes on the sulphur atoms were able to exercise the directionality that was observed by the O-S⸳⸳⸳O angles, which were around 180° with the plane of the σ-hole. So, the combination of the multiple positive and negative regions in the solvent molecules together with the σ-holes on sulphur make for a highly effective solvation system towards polar as well as more aromatic non-polar compounds.

This concept can also be coupled to the fields of biochemistry and crystal engineering and this is shown in the work of Fanfrlík et al.27

In their study they performed a quantum chemical analysis on the σ-hole interactions of sulphur with a π-system that are induced by thiaborane complexes Ph-SB11, Cl-SB11 and

I-SB11. The electrostatic potential

mapping showed five σ-holes on the

sulphur atom of Ph-SB11 and a belt of positive potential on the sulphur atom of the others two

complexes, which indicates multiple options to take advantages of these regions. This is also displayed in Figure 7. They tested this by investigating several different stacking interactions of π-systems with these thiaboranes and this showed that the most favorable interactions were found in the case, where they speculated that a σ-hole was the cause of the stacking, going up to energies of –8.2 kcal mol-1_{. This means chalcogen interactions can be of significant value}

when designing protein-ligand interactions, since π-systems are occurring frequently in protein systems (phenylalanine) and in crystal engineering.

2.1.3. Pnictogen bonding

After the halogen and chalcogen groups, the next candidate in line is the pnictogen group, containing the atoms N, P, As, Sb and Bi. One of the early indications of σ-hole interactions in this group was given in the work of Frank et al.28 Their work discussed trichloride bismuth complexes that can form a large network with themselves, while being bound to a benzene ring in a η6 fashion. The latter is the first implication of a σ-hole, because of the electron-withdrawing Cl atom bound to the Bi center, possibly creating a positive electrostatic potential that is in the same plane as the η6 bonded benzene. The second implication can be observed in

Figure 7: The electrostatic potential mapping of Ph-SB11(left) and

Cl-SB11(right). The σ-hole regions are shown as the dark blue regions on

the sulphur atoms at the top of the molecules. The electrostatic potential

(15)

15 the complete crystal structure, showing multiple non-covalent interactions between bismuth and chloride atoms, which have a Cl-Bi⸳⸳⸳Cl close to 180°. The statistical bias on the 180° angle is also applying here. So, the σ-holes on the Bi-centers could be responsible for the general crystal structure that is observed. Similar observations were found in the Meshutkin crystals that were the same as the Bi-complexes, but with antimony instead of bismuth.29

A few decades later the characterization of the pnictogen bonding was performed by Del Bene et al., who investigated the non-covalent P⸳⸳⸳P bonding in a series of complexes with general formula (H2C=PX)2, with X = F, Cl,

OH, CN, NC, CCH, H, CH3 and BH2.30 The

electrostatic potential mapping for H2C=PF

clearly showed the presence of the σ-hole and the lone-pair on the phosphorous atom, observed in Figure 8. The σ-hole on phosphorous of one molecule and the lone-pair on phosphorous of the other molecule are the largest contributors in the interaction.They showed that these complexes had the possibility to form four different

conformations, of which three conformations were primarily stabilized by the P⸳⸳⸳P pnictogen bonding, proving the influential strength of the σ-hole interaction with the lone-pair of the complementary phosphorous atoms.

The group of Watt et al. focused their work on controlled self-assembly macrocycles, containing arene groups and pnictogen atoms.31_{The interaction between a As, Sb or Bi atom}

and the π-electrons of arene groups, such as naphthalene or stilbene, was found to stabilize the formation of multiple macrocycles that are linked together through these interactions and thus creating a controlled self-assembly process. The exchange of electron-deficient arene groups with more electron-rich arene groups, for instance hexafluorobenzene or s-triazine, will increase the strength of this interaction, which is in line with the general influence of the σ-hole interactions.32

In the work of Baiocco et al., an example of the structural and possible mechanistic influence of pnictogen bonding is represented, in which SbIII exerts a pnictogen-π interaction in

Figure 8: The electrostatic potential map of H2C=PF,

showing the σ-hole on P at the black dot, having an energy of

25 kcal mol-1_{, and the lone-pair of P at the light-blue dot. The}

electrostatic potential goes from negative to positive with the transition from red to blue. The image was constructed with Spartan and the MEP was calculated with DFT method wB97X-D/6-31G*.

(16)

16 trypanothione reductase.33 This enzyme can function as a drug that is currently used in the treatment of leishmaniasis, a parasitic disease causing ulcers all over the body. The SbIII was found to be present in the active site during the catalytic process, while being coordinated by one threonine residue and two cysteine residues, of which the latter ones have redox-active catalytic functions. The function of SbIII would be its coordination to the active site and thus a contribution to the tight binding of the inhibitor to the active site. So, besides presence of pnictogen bonding in theoretical work and synthetic methods, it might also have important functional influence in medicinal or biological chemistry.

2.1.4. Tetrel bonding

Although less explored, there have been extensive studies towards the σ-hole interactions in tetrel atoms (C, Si, Ge, Sn). One of these studies was performed by Murray et al., in which they obtained computational evidence that it is also possible to establish the presence of the directional effect of σ-hole interactions in these tetrel atoms.8 In their study, they confirm this with the positive electrostatic potential that is present in molecules, such as SiF4 or GeH4, which

can be compared to the halogen, chalcogen and pnictogen bonding interactions. Furthermore, they pointed out the existence of the directionality these interactions possess by investigating the interaction with the nitrogen atom in H3N and HCN and calculating the interaction angles

and energies of Si-N of Ge-N bonds. These results show significantly negative bonding energies and bonding angles of 180° along the axis of the σ-hole interaction, which suggest that these interactions are stable and directional.

A different broad study of these non-covalent tetrel bonding interactions was carried out by Bauza et al., who tried to give a glimpse of the possibilities that lie within this field.4_They

first show some crystal structures, in which a σ-hole interaction is present. The most interesting one of these was reported by Taylor et al. that created a stable compound consisting of eight equally distributed Si-F interactions, where a fluoride-anion is encapsulated between an octasiloxane cage.34_{Two of these compounds are also depicted in Figure 9. By searching}

through the CSD, they found that heavier tetrel atoms (Ge and Sn) are more likely to form hypervalent complexes compared to lighter tetrel atoms (C and Sn) that tend to prefer formation of non-covalent interactions. After these conclusions, they continued their investigation towards the directionality of the non-covalent interactions of carbon. From their results, they

(17)

17 observed a clear directional supramolecular interaction when the σ-holes on the carbon are well-exposed.

Figure 9: The X-ray crystal structures of a fluoride anion encapsulated into silica hexahedral cages.34

2.2. π-Hole interactions

2.2.1. Lone-pair-π bonding

Lone-pair-π bonding is an interaction that involves the lone-pair on electron-donating atoms, which contributes in an attractive interaction with an electron-deficient π-system, usually consisting of the conjugated structures of aromatic moieties. One of the earlier computational studies on this subject was performed by Alkorta et al., in which a possible interaction of the π-cloud of C6F6 with small electron donor atoms, such as HF, HCN, CNH and :CH2, was

investigated.35_{Their results showed an orbital overlap between the electropositive lone-pair on}

the F-, C-, N-atoms and the electron-deficient center of the C6F6 ring. In order to generalize

their observations, they performed a CSD-search based on the interaction distance and angle, which provided a multiplicity of the occurrence of the interaction with their systems, including other electron-donating moieties that showed similar behavior.

A while after the discovery of these interactions, a spark of interest was observed towards the behavior of such systems. For Egli and Sarkhel, this interest led to a study based on the lone-pair-π bonding of multiple oxygen-containing molecules, in particular that of carbonyl-based molecules.36_{Their research started off with a CSD-based analysis to determine}

the amount of hits that could indicate a lone-pair-π bonding interaction between the oxygen atom of X2C=O moieties and aromatic rings. By setting restraints on this system based on

distances and angles, they were able to obtain a significant amount of hits, which led to the conclusion that there was a clear angular preference towards the approach of the carbonyl. This angular preference was based on the mean deviation of 16.8° from the ideal approach of 120°.

(18)

18 After this, they also noticed a clear influence of this type of interaction between water and aromatic residues of proteins. They concluded that in rare cases the oxygen lone-pair interaction of the water molecule had a stronger stabilizing effect on the aromatic ring compared to the hydrogen-bonding interaction.

In a different extensive CSD-analysis, Mooibroek et

al. investigated the frequency

of lone-pair-π bonding presence in known crystal structures.37 In this analysis they looked into the amount of contacts between six

molecules, containing electron-rich atoms (halogens, oxygen and nitrogen), and thirteen aromatic molecules, where carbon atoms were substituted for nitrogen atoms. To obtain a proper dataset, restraints were set according to Figure 10, which consist of the distance from the electron-rich atom to the plane of the π-system and the deviation of the electron-rich atom from the plane perpendicular on the center of the π-system. Based on these factors, they managed to prove the rather common presence of lone-pair-π interaction in solid-state structures. A large amount of this presence is caused by solvent interaction that has a stabilizing effect on the structures. This research shows that this particular interaction should be a more considered interaction in crystal engineering and biochemistry, as it has a rather large appearance in such crystal structures.

Further exploration of databases was performed by Bartlett et al., in their case the exploration of the PDB, in order to determine not just these interactions in crystal structures, but specifically in the crystal structures of proteins.13_{Their work proved the significant}

prevalence of these interactions, which also revealed the stabilizing effect that the lone-pair-π bonding has on secondary structures of proteins. For instance, in α-helices these bonding interactions were found to cause the pyramidalization of the acceptor carbonyl group and the polarization of its π-electron cloud, leading to clear conformational stability. Moreover, they observed lone-pair-π bonding within the helix. They suggest that this bonding interaction facilitates the hydrogen bonding by aligning the helix in such a manner that it will reduce the conformational entropic and enthalpic penalties. Their work shows that these interactions are

(19)

19 likely to have a functional role in the stability and formation of the secondary structure of proteins.

One specific case in which such an interaction was observed, showed catalytic activity of methionine residues in enzyme-based reactions. In the study of Ranaghan et al., an enzyme complex is described that contains an arginine residue for substrate binding, a histidine residue acting as the putative base and a methionine residue to catalytically stabilize the histidine residue.38_{By performing mutagenesis experiments, which were based on substituting the}

methionine residue, they indicated that the methionine residue is important in catalysis and not in substrate binding. They also mentioned that the binding interaction of the sulphur atom in methionine with the ring of histidine is mostly due to the electrostatic influence of the sulphur atom. In this way, the transition state of a hydride transfer caused by the histidine is significantly stabilized, because of a conformational change resulting from the lone-pair-π interaction.

2.2.2. Anion-π bonding

Anion-π bonding is best described as a negatively charged atom having an interaction with the π-cloud of an electron-deficient aromatic conjugate. In order to determine the broadness of this interaction in such a conjugated system, Alkorta et al. presented their work on fully fluorinated six-membered rings (C6F6 and C5F5N) and fully fluorinated furan-type rings (C4F4O and

C4F4S).39 Their calculations were based on the molecular electrostatic potential (MEP) mapping

of the perfluoroaromatic moieties and the atoms in molecules (AIM) analysis. The MEP maps showed a positive electrostatic potential on all of the used aromatics, which are able to be the donor of the interaction. The AIM analysis showed a clear interaction pattern of the anions with the aromatic centers, designated by shortened atomic distances and minimum energies. They also suspect that there is a correlation between the strength of the aromaticity and the interaction energy of the complexes.

In a study of Quiñonero et al. a combination of crystallographic data and computational evidence was used to determine the presence of the anion-π interaction in systems with anions and 1,3,5-trinitrobenzene (TNB) and 1,3,5-trifluorobenzene (TFB).40_{In order to gain insight}

into this interaction, they make use of the quadrupole moment of the aromatics, which appeared to have a positive value above and below the ring and could thus interact with the anion. In their calculations they create a scenario in which the anion is centered on top of the aromatic part of

(20)

20 TNB or TFB, which can be viewed in

Figure 11. This way the interaction would be as strong as possible, because it will have the shortest distance. This showed that TNB creates systems with a strong and clear electrostatic interaction, however the systems with TFB were not mainly caused by an electrostatic term, but by an ion-induced polarization term.

These new findings led to continued research on this specific interaction. In the beginning, characterization of this interaction was mainly performed by theoretical studies, such as the work of Kim et al., which broadened the field by investigating a variety of anions and π-systems.41 They chose to look at olefinic, aromatic and heteroaromatic molecules as the donors and spherical, linear and trigonal planar anions as the acceptors. With help of their calculations, they were able to give a detailed energy decomposition on these anion-π systems. These energy decompositions led to the conclusion that the interaction present in these systems was mostly a contribution by a positive electrostatic potential and induction energy, the first suggesting the presence of the anion-π interaction. An additional indication of this interaction can be found in the vibrational energies of these systems, which showed a vibrational shift the moment they introduced the anion into the π-systems, caused by the supramolecular interaction between the orbital interaction of both the anion and the π-system. The magnitude of these shifts also follows the strength of the interaction energy, indicating a clear correlation between the stretching vibrations and the strength of the interaction.

Besides the theoretical studies, the anion-π interactions were also observed and even used in the synthesis of coordination polymers through supramolecular bonds formed with anion-π interactions. One example of this is the research carried out by Casellas et al., who managed to synthesize coordination polymers based on metal nitrates and a large N-donor ligand constructed with triazine and pyridine moieties.42 The coordination of these polymers

originate from the interaction between the anions that are present on the nitrates of the metals

Figure 11: The theoretical model of TNB with an anion (white circle on

top) with the critical points of both molecules indicated in between.40

Figure 12: Crystal structure of the coordination polymers with the dotted lines

(21)

21 and the π-systems of the triazine and pyridine moieties. They characterized this supramolecular coordination as the results of anion-π interactions between the anions of the nitrates with the triazines and pyridines, which is depicted with the dotted lines in Figure 12.

Apart from the presence of this interaction in multiple aromatic systems, crystal structures and to facilitate coordination polymers, it might also have some biological functionalities. Since ion transport is frequently occurring in biological systems, a ‘slide’ based on anion-π interaction could have a realistic impact in transporting anions. The work of Mareda

et al. focuses on this matter and have proposed a concept in which they describe the multi-ion

hopping mechanisms of anion transport.43_{A combination of anion speed and anion selectivity}

towards the channel walls with conjugated π-systems is required to create functional and efficient ion transport. The answer to this might lie in the use of a cooperative system that relies on the binding of multiple anions, which then could pass along the speed of the next incoming anion to the anion in the front. This means that the selectivity must be strong enough to create an anion-π based bond, but at the same time weak enough to be able to release the anion to pass it through the channel. Some refining of these ion channels needs to be done before it is proven that the interaction exist in these systems and a significant application can be developed. An extended review was also written to comprehend the use of these anion-π interactions in the transport of anions through π-acidic channels.44 Evidence that proved the functional existence of these interactions is such systems was given by further research of Dawson et al., who used tandem mass spectrometry and models to observe the anion-π interactions.45 These results showed that only the π-acidic channels could be the binding site of anions, which means that there is an electrostatic interaction inside these channels that regulates the transport.

It was found that anion-π interactions are also present in the active sites of certain enzymes. An example of this is reported by Estarellas et al., which includes relevant anion-π interactions that are involved in the degradation of uric acid.46 This study has selected some examples from the PDB, showing that these interactions in the active site of the urate oxidase enzyme have an influence on the inhibition of the enzymatic activity, which is depicted in Figure 13. The calculations on these structures show that the presence of the interactions is indeed relevant and

energetically favorable for their specific purposes. They state that the anion-π interaction of the

Figure 13: The active site of the UOX showing

(22)

22 cyanide anion with the π-system of either the substrate or the inhibitor has a stabilizing effect on either one, hence influences the inhibition of the substrate degradation.

2.2.3. Non-aromatic π-hole interactions

In the case of lone-pair-π and anion-π bonding interactions, there is a positive electrostatic potential on an aromatic conjugate, acting as the functional π-hole. However, it is also possible to have the π-hole reside on a moiety that is non-aromatic. In these cases, there is a local positive electrostatic potential on a single atom that is caused by one or multiple covalent π-bonds, which can in turn form a non-covalent supramolecular interaction with a molecule having a negative electrostatic charge, i.e. the oxygen atom in carbonyl moieties.

In order to determine the strength of such interactions, Guo et al. compared the π-hole interaction of F2CO and F2SiO with HCN to the hydrogen bonds between the same molecules.47

The π-hole of these molecules lies on the side of a carbon or silicon atom, relative to the molecular plane that has a double bond with the oxygen. These π-holes can have a supramolecular interaction with the negative potential of the nitrogen atom of HCN. The first detail that is stated, is the strength of the π-holes on both molecules, which is significantly larger for F2SiO than for F2CO. This also shows in the interaction energies when HCN is included

into the models, F2SiO shows an energy almost six times larger than that of F2CO. Compared

to the possible hydrogen bonds of the two molecule combinations, the π-hole interactions exert a higher energy than the hydrogen bonds between the hydrogen atom of HCN and the oxygen atom of the CO or SiO. However, the difference in energy again is significantly higher for F2SiO than for F2CO, 95.4 and 3.6 kJ mol-1 respectively. This means that there is a rather small

difference in interaction energies for the F2CO – HCN system, while the π-hole interaction of

the F2SiO is much larger than the corresponding hydrogen bonding. Interestingly, the large

energy of this specific π-hole interaction is indicating that it has the nature of a partially covalent interaction, while being a supramolecular interaction.

Using this electrostatic interaction, it is also possible to form heterodimeric and heterotrimeric complexes. The work of Azofra et al. presented their models of complexes built from SO3 and CO molecules.48 In these models, a π-hole is residing both on the top and the

bottom of the sulphur atom of SO3 and interacts with the negative potential of the lone-pairs on

either the carbon or oxygen atom of CO. The dimer has the most stable configuration when forming the S···C bond, because the negative potential of the carbon atom is stronger than that

(23)

23 of the oxygen atom. There are two possible options to form the heterotrimeric complexes, one with two CO molecules, the other with two SO3 molecules. With two CO molecules, the most

stable structure resembles the most stable dimeric structure, as the two CO molecules are both taking part in a π-hole interaction that forms two S···C bonds, compared to two S···O bonds. However, these complexes do exert some negative cooperativity, which was observed by the elongated S···X bonds. In the case of the SO3 dimers, there are more possible configurations as

the structure of SO3 allows for more geometric options. The most stable configurations are

formed when there is a cyclic formation, which has interactions between the π-hole on the sulphur atom of one SO3 and the negative potential on one of the oxygen atoms of the other

SO3, while also finishing the trimer with a S···C bond and a O···C bond. These cyclic modes

have significantly shortened bonds, which indicates that they exhibit positive cooperativity. This is the main reason that the cyclic configurations are more stable than their linear counterparts.

Anions are usually reacting as electron donors, and specifically nitrate anions are generally reactive due to their negative charge. However, they can also interact through the π-hole residing on top of the nitrogen atom and thus act as a Lewis acid. Bauzá et al. recognized this phenomenon and searched the CSD and PDB in order to highlight this interaction in known structures.17 These searches lead to the conclusion that nitrate anions can have a π-hole interaction with multiple moieties that contain electron rich atoms, including carbonyls, thiocarbonyls, water and sulphur. One example of the binding pocket of a nitrate anion inside of a protein is depicted in Figure 14, which shows that the nitrate is coordinated in between the two chains of the protein itself. This structure and other observations in the databases indicate the potential of π-hole interactions within such structures.

(24)

24

3. Relevance of π-hole interactions on nitrate anions in

proteins

3.1. The function of nitrate anions in biological systems

Since the work of Bauzá et al. has shown that nitrate anions are not just present in proteins but can also have their own function through π-hole interactions in these structures.17 In order to determine how important these findings are and to determine the broadness and role of these interactions in proteins are, it is necessary to identify how nitrate anions are functioning in biology in general. The presence of nitrate in biological systems has long been used as one of the main sources of inorganic nitrogen that functions as a key nutrient to promote the growth and development of plants.49 To provide nitrate to plants or other organisms, it is necessary to have a highly ordered distribution mechanism in place. This mechanism consists of defined nitrate transport and absorption pathways. Besides the value of being a nutrient, nitrate also plays a role in the signalling regulation that is used in the development and physiology of these organisms, which expresses itself in the metabolism and growth. A known nitrate sensor in plants was observed by Ho et al., who characterized the primary nitrate response of CHL1, a transporter gene in plants.50_{The CHL1 protein is connected to the protein kinase CIPK23,}

which is responsible for the phosphorylation of the CHL1 protein. The sensing mechanism of CHL1 is based on low nitrate concentration and causes said phosphorylation and could possibly be a general mechanism that applies to other organisms that have pathways for ion sensing. This means that CHL1 has a dual affinity towards the binding of nitrate and a phosphorylation switch that can be used to sense a wide range of nitrate concentrations. Since the protein is connected to a kinase, which are involved in energy transferring and protein activation/deactivation, nitrate signalling could be significant in regulation of cellular transport or protein regulation. Wang et al. also mentions other examples of the regulation and transporting functions in which nitrate plays a role.49 Their own study discusses the FIP1 protein, which is a polymerase that regulates nuclease activities. They proved that FIP1 is a main factor in nitrate signalling, expressing itself in the modulation of nitrate contents, regulation of nitrate transport, allocation and assimilation. This also means that FIP1, like other enzymes, is an enzyme that is dependent on a certain nitrate concentration.

Additionally, the use of nitrate and its reduction product nitrite have been exploited in the medicinal world to contribute to the promotion of human healthcare.51 Especially, the

(25)

25 contribution of nitrate in certain diets are believed to be the cause of lowering blood-pressure, the strengthening of gastric mucosal defence mechanism and the oxygen-cost reduction that is needed during exercise. The latter one in specific was discovered by Larsen et al., who performed a study that investigated the effect of a nitrate-based diet on metabolic and circulatory parameters during physical exercise.52 They concluded that supplementing nitrate in a diet has a positive effect on the oxygen cost during exercise. This effect was observed together with a lowered lactate concentration, which means that the body was producing its energy in a more efficient way. They believe that this effect is caused by the in vivo reduction of nitrate into nitrogen oxides that are active in energy production. Continued pharmacological research into the positive effects and exploitation of nitrate-driven pathways might lead to a new class of therapeutic treatments.

Furthermore, nitrate anions can have an indirect influence in the regulation of ion transport that is focused on other useful ions. Ion transport can be categorized in three options: uniport, symport and antiport. Uniport uses a channel in which a single ion is transported in one direction, whereas the other two require a second ion to be transported in the same or the opposite direction. In these cotransport situations, the second ion is usually required to even out the charge density differential that appears when charged molecules or atoms are transported through these channels or to trigger the target molecule by creating a charge difference between the two sides of the channel. Nitrate anions can be used as the second ion and thus have an indirect effect on the transport of necessary ions through membranes. For example, the treatment of illnesses that revolve around the disruption of anion transports, such as cystic fibrosis, can possibly be treated by designing new selective and efficient transporters. For these treatments to work, the deficient channels should be replaced by these synthetically designed transporters. The research that is carried out to solve this problem focuses their attention on creating compounds that can transport Cl-_{anions through the lipid bilayer membranes.}53,54

These studies have designed receptor that are tested on the transport properties of chloride anions and some of the antiport processes use nitrate as the counterion. The challenge here is creating transporter compounds that have optimized the absorption, distribution and excretion of chloride and nitrate anions, are consistent with the surrounding atmosphere and metabolism, and have proper toxicity conditions.

(26)

26

3.2. PDB search of π-hole interactions with nitrate anions

To determine the relevance of π-hole interactions that are induced by nitrate anions in protein structures, the PDB was used in order to find out how often π-hole interactions do occur in known structures. The search was based on multiple factors, most importantly the presence of nitrate anions that are within certain distances of electron rich moieties, such as C=O, C=S or OAr groups. The specified distances are the interaction distance between the nitrogen atom of nitrate and the interacting atom of the electron rich moieties, and the parallel displacement of the interacting atom compared to the plane directly above the nitrogen atom, which is perpendicular to the plane of the nitrate molecule itself. From these results, the ones with C=O groups were chosen for this report, since C=O groups had the most hits by far, which is also in line with the prevalence of C=O groups in proteins in general. To continue with narrowing down the search, both distances were combined to obtain results that consisted of an interaction distance lower than 4 Å and a parallel displacement lower than 0.5 Å. This resulted in 91 hits that satisfied all the conditions. Three types of proteins were selected as categories, because they had more hits than others, the types being N-acetylglucosamine deacetylases, tyrosinases and kinases. The protein structures, which were taken from the PDB, were analyzed using the program PyMol. In order to obtain knowledge of the structural position of the nitrate binding pockets in the proteins, tertiary structures were showed, highlighting the position of the nitrate binding pockets and the active sites of the enzymes. To determine the inside of the nitrate binding pockets itself, residues within 4 Å of the nitrate anion were depicted and measurements on the presumed π-hole interactions and hydrogen bonds were made to show geometrical influences. Overlaps of the tertiary structures and binding pockets were also obtained. The overlap of the tertiary structures was simply performed by overlaying multiple proteins of the same groups. The overlaps of the nitrate binding pockets were created by using a pair fitting overlay, in which the atoms of the nitrate anions and the oxygen atoms of the carbonyls were pair fitted on top of the corresponding atoms of the other binding pockets. In some cases, the inclusion of the carbonyl oxygen in the pair fitting process caused it to be unclear due to distortion compared to pair fitting only the atoms of the nitrate anion and it was chosen to only fit the atoms of the nitrate anions. Whether just the atoms of the nitrate anions were pair fitted or also the carbonyl oxygen, is mentioned in the captions of the corresponding figures.

(27)

27

3.2.1. N-acetylglucosamine deacetylases

The first category consists of proteins that belong to the group of N-acetylglucosamine deacetylases, which are a part of the family of hydrolases that are responsible for the catalysis of water-based cleavage reactions. N-acetylglucosamine deacetylase specifically catalyzes the reaction that converts N-acetyl-D-glucosamine into D-glucosamine and acetate. This category contained ten proteins that included a π-hole interaction of a C=O moiety with a nitrate anion. Four of these ten proteins were found to have a similar protein structure and binding pocket, which means a 40% overlap rate. These four hits are all present in the LpxC enzyme, which is an important part of the lipid A biosynthesis in the outer membrane of certain bacteria, because it is the catalyst in the deacetylation of UDP-3-O-(acyl)-N-acetylglucosamine, one of the key steps during the synthesis.55 Figure 15a shows the first example of these four proteins (refcode: 3p3e), with a zoom-in of the binding pocket that includes the residues in a range of 4 Å of the

Figure 15: a) The tertiary structure of the LpxC protein (blue circle: active site, yellow circle: nitrate binding pocket) with a

zoom-in of the nitrate binding pocket (residues ≤ 4 Å displayed), refcode: 3p3e.52_{The distances of the π-hole interaction and}

hydrogen bonds are depicted in Å b) The overlap of the tertiary structures of the four proteins (refcodes: 3p3e, 4lcf, 4lcg,

4lch) and a zoom-in of the overlap of the binding pockets fitted on the nitrate atoms and the carbonyl oxygen.52,53_{The average}

π-hole interaction distance is given. Color code: carbon = grey, hydrogen = white, nitrogen = blue, oxygen = red.

Ser-177

Phe-176

Phe-152

Ser-295

Ile-294

Pro-293

3.114

a)

b)

(28)

28 nitrate anion. It shows that the binding pocket of the nitrate anion is resided on the side of the protein and that the nitrate anion shows a geometry reminiscent of a π-hole interaction with a phenylalanine residue and is stabilized by the hydrogen bonds formed with residues from three parts of the protein and two water molecules. This can be seen by the residue numbers, which are 152, 176-177 and 293-295. Lee et al. describes that the active site of this protein is connected to the residues 191, 238-242 and 265 among other, which is in between the residues of the nitrate binding pocket. This indicates that the nitrate anion could be responsible for the specific tertiary structure of those parts of the protein and thus indirectly have a function in the way the active site is positioned and/or oriented. To determine whether this pattern repeats itself, the structure and binding pocket of this protein was overlapped with the three other LpxC enzymes (refcodes: 4lcf, 4lcg, 4lch), which is depicted in Figure 15b.56 From this, it can be observed that the four proteins are highly similar in their structures and nitrate binding pockets and that the average distance of the π-hole interaction is 3.114 Å. The van der Waals radius of a N-O interaction equals 3.07 Å. With a difference of 0.044 Å, it can be speculated that there is a non-covalent interaction between these residues. The similarity means that the position of this interaction is consistent for the four LpxC enzymes and will have the same influence in all four enzymes. The structural relevance of nitrate anions in these proteins is supported by the fact that in all four cases, a nitrate buffer was needed to obtain this specific structure, since they substituted i.e. an ammonium phosphate buffer for an ammonium nitrate buffer in the case of 3p3e.55 A possible functional relevance depends on whether or not the π-hole interaction of the nitrate anion with the C=O group has any influence on the active site or the deacetylation process.

3.2.2. Tyrosinases

The second category is that of the tyrosinases, which are part of the family of oxidoreductases that are essential components of the electron transfer reactions in organisms. The role of tyrosinases is to catalyze the hydroxylation of monophenol and the following oxidation, leading to the product quinone. The active part of these enzymes consists of a dinuclear copper center, taking care of the oxidation together with molecular oxygen. The quinone product is used in the synthesis of melanin, something that is produced in plants and human and controls the pigmentation. Most bacteria produce a similar pigment and the enzymes that regulates this process appeared to contain π-hole interactions of nitrate anions with C=O moieties. In this case, ten out of twelve proteins, equaling 83%, appeared to have equal structures and nitrate binding pockets. With the crystallization processes of these structures, a nitrate buffer was

(29)

29 necessary to obtain a specific structure with orthorhombic crystals and play a role in the formation of copper binding sites, indicating a functional relevance of the addition of nitrate in these proteins.57,58 One of the nitrate anions is also connected to the active center of the protein, but unfortunately is not the nitrate anion that takes part in a π-hole interaction. The nitrate anion that seems to have a π-hole interaction with a C=O group, based on the geometry of the binding pocket, and the structure of two monomers of the protein are depicted in Figure 16a (refcode: 1wx4). The tertiary structure shows that the nitrate binding pocket resides in between two monomers of the protein. The zoom-in of the binding pocket elucidates that the second monomer even forms a hydrogen bond with the nitrate anion through the lysine residue. The zoom-in also shows the geometrical indication of a π-hole interaction between the nitrate anion

Figure 16: a) The tertiary structure of two monomers of the tyrosinase protein (blue circle: active site, yellow circle: nitrate

binding pocket) with a zoom-in of the nitrate binding pocket (residues ≤ 4 Å displayed), refcode 1wx4.54_{The distances of the}

π-hole interaction and hydrogen bonds are depicted in Å b) The overlap of the tertiary structures of the ten proteins (refcodes: 1wx4, 2ahl, 2zmx, 2zmz, 2zwd, 2zwe, 2zwg, 3awt, 3aww, 3awx) and a zoom-in of the overlap of the binding pockets fitted on

the nitrate atoms and the carbonyl oxygen.54,55_{The average π-hole interaction distance is given. Color code: carbon = grey,}

hydrogen = white, nitrogen = blue, oxygen = red.

a)

b)

Asn-248

Leu-247

Asp-246

Thr-238

Gly-239

Lys-50

3.152

(30)

30 and the C=O group of the threonine residue and the stabilization of the nitrate anion with hydrogen bonds of the surrounding residues 238-248 and two water molecules. The active site consists of the residues 38-63 and 190-216, which are not near the nitrate binding pocket. This means that there is no direct influence of the π-hole interaction on the active site, and it seems that there is also no indirect influence through the structure and placement of the active site, since it appears to mainly connect the two monomers. However, this connection of the monomer does mean that the nitrate binding pocket, and also the π-hole interaction, has some structural relevance. To determine the consistency of this structural relevance, the structures and nitrate binding pockets of the other tyrosinases were overlapped (refcodes: 1wx4, 2ahl, 2zmx, 2zmz, 2zwd, 2zwe, 2zwg, 3awt, 3aww, 3awx), which is shown in Figure 16b. The second monomer was left out in this figure for clearness. The structural overlap of the proteins shows that they are highly similar, and this is confirmed by the nitrate binding pocket that also shows an average π-hole interaction of 3.152 Å. This average is 0.082 Å longer than the van der Waals radius, which still indicates of a weaker supramolecular interaction. It can be observed that there are just some slight differences in the displacements of the residues and the orientation of the water molecules. A consistency in 83% of the investigated tyrosinases indicates that there is a structural relevance in tyrosinases of this interaction caused by the nitrate anions.

3.2.3. Kinases

The last category is the kinases that belong to the family of transferases, which function as the transporters of all kinds of molecules through the metabolic systems of organisms. Kinases specifically focus on the catalysis of the phosphate transfer, which is the backbone of energy regulation. One bacterial kinase was already mentioned in the study of Bauzá et al.17_{, which}

discusses the presence of a double π-hole interaction of one nitrate anion with two C=O groups from the two chains of the histidine kinase NarX (refcode: 3ezh). The original work that describes this enzyme was written by Cheung and Hendrickson, who made a detailed description of the importance of the nitrate anion within this enzyme, but without the knowledge of π-hole interactions.59 They explain that the nitrate anion is a key factor in crystallizing the specific dimer that is active in the regulation of anaerobic respiration. By functioning as a nitrate sensor, it can bind nitrate and undergo conformational changes that result in a displacement of the helices and create new active sites that carry out the reactions needed for this regulation. Because the dimer is bound to the nitrate not only through hydrogen bonding, but also through

(31)

31 two π-hole interactions, it proves the structural and functional relevance the interaction with nitrate anions in this specific enzyme. While this enzyme is found in bacteria, the PDB search resulted in 25 hits of human kinases, all being checkpoint kinase 2. This specific kinase is involved in DNA repair and cell apoptosis and is linked to the suppression of tumors.60 22 of these 25 hits were found to be similar in nitrate binding pocket, which equals 88%. In all 22

Figure 17: a) The tertiary structure of the Checkpoint kinase 2 protein (blue circle: active site, yellow circle: nitrate binding

pocket) with a zoom-in of the nitrate binding pocket (residues ≤ 4 Å displayed), (refcode 2w7x).57_{The distances of the π-hole}

interaction and hydrogen bonds are depicted in Å b) The overlap of the tertiary structures of the 22 proteins (refcodes: 2cn5, 2w0j, 2w7x, 2wtd, 2wti, 2wtj, 2xbj, 2ycf, 2ycg, 2ycr, 2ycs, 2yiq,2yir, 2yit, 4a9t, 4a9u, 4bdb, 4bdf, 4bdg, 4bdh, 4bdi, 4bdj) and

a zoom-in of the overlap of the binding pockets fitted on the nitrate atoms.57-65 The average π-hole interaction distance is

given. Color code: carbon = grey, hydrogen = white, nitrogen = blue, oxygen = red.

Gly-403

Tyr-404

Asn-405

Leu-375

Arg-406

a)

b)

(32)

32 structures the crystallization procedure used a nitrate buffer solution to obtain the given structures. One of these structures and a zoom-in of the corresponding nitrate binding pocket is depicted in Figure 17a (refcode: 2w7x). This figure shows that the nitrate anion resides on one of the edges of the protein and that it is bound to one part of the protein chain, being residues 403-406. The shown geometry is reminiscent of a π-hole interaction between the nitrate anion and a glycine residue, and the nitrate is stabilized with multiple hydrogen bonds from the arginine and asparagine residues, but also from one water molecule. This resembles the nitrate binding pocket of the NarX enzyme, which also suggests a π-hole interaction with glycine residues and hydrogen bonding with arginine residues. The active site of the Checkpoint kinase 2 is formed by the residues 226, 234, 273, 299-308, 354, 367-370, which is about 30 residues away from the nitrate binding pocket. This indicates that the π-hole interactions most likely have no direct influence on the function of the enzyme. To further investigate the relevance of this interaction in the kinases, the overlap of the 22 protein structures and nitrate binding pockets is depicted in Figure 17b (refcodes: 2cn5, 2w0j, 2w7x, 2wtd, 2wti, 2wtj, 2xbj, 2ycf, 2ycg, 2ycr, 2ycs, 2yiq, 2yir, 2yit, 4a9t, 4a9u, 4bdb, 4bdf, 4bdg, 4bdh, 4bdi, 4bdj).61–68 The overlaps shows that all 22 of the proteins have almost equal tertiary structures, placing the nitrate binding pockets in the same place. It also shows that the binding pockets do show some more configurational differences, but are mostly in similar positions relative to the nitrate anions with an average π-hole interaction distance of 3.031 Å. This average distance is 0.039 Å shorter than the van der Waals radius, giving a strong sign of a non-covalent interaction taking place. Since there is a clear similarity in 22 of these structures, it can be concluded that the π-hole interaction of the nitrate anion with the C=O group of the glycine must have a structural relevance that forms part of the tertiary structure of the Checkpoint kinase 2 enzyme. There is no direct evidence that these interactions have a functional role in the function of the enzyme, since the active site is not near the nitrate binding pocket. However, since the binding pocket does show similarities to the binding pocket of protein 3ezh, where the π-hole interaction has a functional relevance on the activity of the enzyme, it could be speculated that the influence of the discussed π-hole interaction could be somewhat similar to the influence of the π-hole interaction in 3ezh. This speculation is supported by the fact that the nitrate binding pocket resides next to a rather flexible loop of the enzyme, which might be relevant in the approach of target molecules or formation of the active site.

(33)

33

4. Summary

This report showed a widespread variety of structures relying on σ- and π-hole interactions and that these interactions can be the basis of multiple applications in synthetic chemistry or functions in biological systems. Additionally, the relevance of π-hole interactions with nitrate anions in proteins was investigated by performing a database search of the PDB. This led to the discovery of three protein groups in which geometries were observed that indicated possible π-hole interactions of nitrate anions with electron-rich moieties with a focus on carbonyl groups. This interaction appeared to occur in proteins that are part of N-acetylglucosamine deacetylases in bacteria, tyrosinases in bacteria and kinases in humans. 40% of the selected N-acetylglucosamine deacetylases were found to have equal structures and had relevance towards the overall tertiary structure of the proteins and a possible influence towards the processes occurring in the active sites. A consistency of 83% was found in the protein structures of the tyrosinases, which had clear structural relevance in the tertiary structures and also in connecting two monomers of the protein, but no convincing prove was found for any functional relevance. The kinase group showed that 88% of the structures had similar structures and these were present in a rather flexible loop of the proteins, indicating a structural relevance of the π-hole interactions, but no obvious functional relevance since the loop is not connected to the active site. In addition to the human kinases, a bacterial kinase (3ezh) was found to also have a similar nitrate binding pocket as the other kinases. However, besides having a structural relevance by connecting two monomers, it was found that the two probable π-hole interactions of the nitrate anion with both monomers have a functional relevance in the protein. Since the protein is active as a dimer, the fact that the nitrate anion holds the dimer together and is needed to crystallize the dimer, is proof of the indirect influence on the function of the kinase itself. From this literary study, it can be concluded that π-hole interactions with nitrate anions in enzymes are likely to have a significant influence towards the structural integrity and the enzymatic functions of N-acetylglucosamine deacetylases in bacteria, tyrosinases in bacteria and kinases in humans. In order to confirm or to negate these speculations, experiments have to be performed that can prove whether or not the π-hole interactions in the nitrate binding pocket are structurally and functionally relevant. This could include a crystallization study, in which the concentration of nitrate buffer is varied to observe structural changes and determining the impact of not using nitrate on the function on the enzymes, which could change reaction rate or selectivity of the enzymes.