Intermolecular Interactions of Nucleoside Antibiotic Tunicamycin with On-Target MraYCB-TUN and Off-Target DPAGT1-TUN in the Active Sites Delineated by Quantum Mechanics/Molecular Mechanics Calculations.

Tunicamycin (TUN) is a nucleoside antibiotic with a complex structure comprising uracil, tunicamine sugar, N-acetylglucosamine (GlcNAc), and fatty acyl tail moieties. TUN, known as a canonical inhibitor, blocks vital functions of certain transmembrane protein families, for example, the insect enzyme dolichyl phosphate α-N-acetylglucosaminylphosphotransferase (DPAGT1) of Spodoptera frugiperda and the bacterial enzyme phospho-N-acetylmuramoylpentapeptide translocase (MraYCB) of Clostridium bolteae. Accurate description of protein-drug interactions has an immense impact on structure-based drug design, while the main challenge is to create proper topology and parameter entries for TUN in modeling protein-TUN interactions given the structural complexity. Starting from DPAGT1-TUN and MraYCB-TUN crystal structures, we first sketched these structural complexes on the basis of the CHARMM36 force field and optimized each of them using quantum mechanics/molecular mechanics (QM/MM) calculations. By continuing calculations on the active site (QM region) of each optimized structure, we specified the characteristics of intermolecular interactions contributing to the binding of TUN to each active site by quantum theory of atoms in molecules (QTAIM) and natural bond orbital (NBO) analyses at the M06-2X/6-31G** level. The results outlined that TUN insertion into each active site requires multiple weak, moderate, and strong hydrogen bonds accompanying charge-dipole, dipole-dipole, and hydrophobic interactions among different TUN moieties and adjacent residues. The water-mediated interactions also play central roles in situating the uracil and tunicamine moieties of TUN within the DPAGT1 active site as well as in preserving the uracil-binding pocket in the MraYCB active site. The TUN binds more strongly to DPAGT1 than to MraYCB. The information garnered here is valuable particularly for better understanding mode of action at the molecular level, as it is conducive to developing next generations of nucleoside antibiotics.

Introduction

The formation of linkage between an oligosaccharide and the amide nitrogen of an asparagine residue is a complex, multistep, and highly regulated reaction occurring widely as one of the most common post-translational modifications of eukaryotic proteins in a process called N-linked glycosylation.[1,2] This attachment plays a vital role in the structure, function, stabilization, and folding of proteins.[3,4] Dolichyl phosphate α-N-acetylglucosaminylphosphotransferase (DPAGT1) from Spodoptera frugiperda is a human transmembrane enzyme that catalyzes the first and critical step of the synthesis reaction of N-linked glycosylation in the endoplasmic reticulum (ER) membrane by transferring an N-acetyl-d-glucosamine-1-phosphoryl unit (GlcNAc-1-P) from uridine diphosphate-N-acetylglucosamine (UDP-GlcNAc) to dolichyl phosphate (DoI-P).[5,6] DPAGT1 is also known as GlcNAc-1-P transferase (GPT)[1] and belongs to the polyisoprenyl phosphate N-acetylaminosugar-1-phosphoryl transferases (PNPTs) superfamily.[7] Similar to other members of this superfamily, the human N-linked glycosylation pathway of DPAGT1 in the ER membrane can be inhibited and blocked by the natural-product nucleoside antibiotics, such as tunicamycin (TUN).[8−10]

TUN is a nucleoside-analog inhibitor that acts as a competitor for the natural substrate UDP-GlcNAc due to its high binding affinity to DPAGT1.[11−13] Structurally, TUN consists of a uracil base, a fatty acyl tail, and two glycosidically linked sugars called tunicamine and N-acetylglucosamine (GlcNAc) moieties (Figure ).[14] Tunicamine moiety is an unusual 11-carbon aminodeoxydialdose that is bound to uracil and GlcNAc moieties in an N-glycosidic bond and in an O-glycosidic bond, respectively. The amino group of tunicamine is attached to the carbonyl group of fatty acyl tail in an amide linkage.[9,14,15] In addition to DPAGT1, N-acetylmuramoyl pentapeptide translocase (MraYCB) from Clostridium bolteae is well-known as a promising drug target for TUN.[16,17] MraYCB is a bacterial transmembrane enzyme that catalyzes the peptidoglycan biosynthesis of bacterial cell wall by transferring phospho-N-acetylmuramoyl pentapeptide (P-MurNAc-pp) from the hydrophilic precursor uridine diphosphate-MurNAc-pentapeptide (UM5A) to the carrier lipid undecaprenyl phosphate (C55–P) with the Mg2+ as a cofactor.[18,19] Because TUN is also a competitive inhibitor for the natural substrate UM5A, it can block the bacterial peptidoglycan biosynthetic pathway of MraYCB.[20−22]

Figure 1

Chemical structure of TUN, indicating that it comprises the uracil base, tunicamine sugar, GlcNAc, and fatty acyl tail moieties. Tunicamine sugar moiety is attached to the uracil, GlcNAc, and fatty acyl tail through an N-glycosidic bond, an O-glycosidic bond, and an amide linkage, respectively.

It is clear that noncovalent intermolecular interactions play fundamental roles in the potent inhibitor binding to the active site of its targeted enzyme in order to form a tightly bound enzyme–inhibitor complex. These interactions are classified into three categories comprising electrostatic, van der Waals (vdW), and hydrogen-bonding (H-bonding) interactions.[23−25] Deeper insights into the active sites of DPAGT1–TUN and MraYCB–TUN complexes can be gained by determining the nature and strength of intermolecular interactions between TUN and its adjacent residues inside each of these active sites. From a theoretical standpoint, comprehensive description of these interactions requires the use of applicable computational methods. Nowadays, parallel progress in computational softwares and development in computer architectures,[26] the combined quantum mechanics/molecular mechanics calculations have been deemed as a powerful computational chemistry method to study the protein–ligand interactions.[27−29] On the other hand, the M06 family of density functional theory (DFT)[30] comprises suitable functionals to compute noncovalent intermolecular interactions, such as hybrid meta-GGA density functional (M06-2X).[30−32] Bader’s quantum theory of atoms in molecules (QTAIM)[33−35] and natural bond orbital (NBO) analyses[36,37] are two reliable theoretical approaches to gain further insights into the nature and strength of the intermolecular interactions, especially the H-bonding interactions, in the protein structures.[24,25,38−40]

Our purposes in this study can be summarized in five main points. The first is to model DPAGT1–TUN and MraYCB–TUN structures on the basis of the standard CHARMM36 force field[41] protocol. The second is to optimize these two structures by quantum mechanics/molecular mechanics (QM/MM) calculations. The third is to specify the characteristics of H-bonding interactions of TUN with its neighboring residues inside each active site utilizing the QTAIM and NBO analyses in the DFT treatment. The fourth is to compare the intermolecular interaction strengths of the different TUN–residue pairs of these two active sites. The fifth and last is to introduce the more stable active site. On the basis of the analysis, the quantitative ligand–target relationship should enable understanding of the influence of binding factors on biological effectiveness, and with the established relationship, compounds can be constructed with improved and optimal biological profiles but free of unwanted side effects in the future.

Computational Details

Setup of MD Simulations of DPAGT1–TUN and MraYCB–TUN Complexes

The atomic coordinates of the DPAGT1–TUN crystal structure obtained at a 3.40 Å resolution are available in the Protein Data Bank (PDB) with PDB code 5O5E (Figure a).[1] The X-ray crystal structure of MraYCB in complex with TUN determined at the atomic resolution of 2.60 Å was released in the PDB under accession code 5JNQ (Figure b).[17] The atomic coordinates of these low-resolution structures are needed for optimization by MD simulations because each of them has some residues from which only their backbones were detected into its electron density map. In order to run MD simulations, each protein–TUN system was prepared in several steps. First, except the crystallographic water molecules and TUN, other small molecules were removed from each structure. As already mentioned above, TUN has a complex structure consisting of the uracil base, tunicamine sugar, GlcNAc, and fatty acyl tail moieties. This structural complexity is a major challenge for constructing the suitable topology and parameter entries for TUN.

Figure 2

Three-dimensional structures of (a) DPAGT1–TUN and (b) MraYCB–TUN.

Second, by following the standard CHARMM protocol, the proper topology and parameter entries of tunicamycin were created on the basis of the existing topology and parameter information available within the CHARMM36 force field.[41] Topology and parameter entries built for TUN are given in the Supporting Information (Tables S1–S6). Third, the atomic positions of all hydrogen and non-hydrogen atoms missing from DPAGT1 and MraYCB structures were reconstructed by the CHARMM36 protein topology file[41] and the VMD software.[42] In both protein structures, the charge states of all Asp and Glu residues were negative and all Arg and Lys residues were positive, whereas the other residues were neutral. Except for His302 in DPAGT1 and His290 in MraYCB, which were modeled as Hsd302 and Hsd290, all other His residues were modeled as Hsd or Hse on the basis of the local electrostatic environment. The N-terminal and C-terminal of the first and the last residues either in DPAGT1 or in MraYCB were protonated and deprotonated, respectively. Fourth, to sketch protein–TUN structural complexes, TUN was incorporated into each protein active site using the protein–TUN topology file, which was generated by adding the TUN topology entry to the protein topology file, and the VMD.

Fifth, DPAGT1–TUN and MraYCB–TUN complexes were separately immersed in a rectangular box containing TIP3P water molecules[43] extending at least 12 Å from each solute. Then, the total charge of each protein–inhibitor–water system became zero by adding a few sodium and chloride ions to the surface of its protein. Sixth, MD simulations were done on each of these explicitly solvated proteins, first for 2 ns in the isobaric–isothermal (NPT) ensemble and then for 15 ns in the canonical (NVT) ensemble, under periodic boundary conditions by utilizing the NAMD software.[44,45] In both ensembles, the temperature and the pressure were kept constant at 310 K and 1 atm, respectively. The Langevin dynamics[46] was employed to control the temperature with a collision frequency of 1.0 ps–1. The time integration step was 1 fs, and the bonds involving hydrogen atoms were maintained by using the SHAKE algorithm.[47] The energies were reported every 1 ps, and the atomic coordinates saved to the trajectory file (dcd file) were recorded every 5 ps. The particle mesh Ewald (PME) method[48] was employed to calculate the long-range electrostatic interactions, whereas the short-range vdW interactions were evaluated by Lennard-Jones potential. Both types of nonbonded interactions were gradually truncated by utilizing a 12 Å cutoff with a switching function applied beyond 10 Å.

Setup of QM/MM Calculations on the Simulated Structures of DPAGT1–TUN and MraYCB–TUN

According to the four moieties of TUN, the tunicamycin-binding pocket (active site) in DPAGT1 is defined as four regions, including the uracil-, tunicamine-, fatty acyl tail-, and GlcNAc-binding pockets. The DPAGT1–TUN crystal structure[1] shows that the uracil-binding pocket is lined by Asp45, Leu46, Asn47, Gly189, Ile190, Asn191, Gly192, Glu194, Asn242, Phe249, Glu376, water603 (Wt603), and water604 (Wt604). The tunicamine-binding pocket consists of Gln44, Glu56, Asp115, Asn119, Lys125, Asn185, Asp252, Arg301, water601 (Wt601), water606 (Wt606), and water609 (Wt609). The fatty acyl tail-binding pocket is considered as a hydrophobic binding site because it is lined with hydrophobic residues Trp122, Leu126, Leu175, Val178, Phe179, Ile186, Phe286, and Leu293. The GlcNAc-binding pocket contains Ala188, Cys299, Pro300, His302, Arg303, and Ile304. Figure S1 of the Supporting Information displays residues and water molecules constituting these four binding pockets.

It should note that the TUN fatty acyl tail was unresolved in the electron density map of the MraYCB–TUN crystal structure due likely to its high flexibility and the natural variation of the acyl chain length.[17] Since the TUN in this structure lacks the fatty acyl tail, the tunicamycin-binding pocket in the MraYCB includes the uracil-, tunicamine-, and GlcNAc-binding pockets. As shown in Figure S2 of the Supporting Information, Asp175, Leu177, Asp178, Gly179, Cys181, Asn221, Ala225, Phe228, Met229, Glu300, water501 (Wt501), water502 (Wt502), and water503 (Wt503) are located at the uracil-binding pocket. The tunicamine sugar moiety is surrounded by Asp93, Lys97, Lys111, Asn172, Gly176, Gly230, and Asp231. The GlcNAc moiety is enclosed by Phe173, Glu264, Val268, Ala287, Pro288, His290, and His291.

To continue the calculations at the QM level on each tunicamycin-binding pocket, we used two different alternative procedures. The first alternative employed a hybrid QM/MM approach. To set up the QM/MM calculations, the active site (tunicamycin-binding pocket) of the final frame of each structure equilibrated in the NVT ensemble was modeled as a QM region, while the rest of the protein residues along with TIP3P water molecules were selected as the MM region. The QM and MM regions in each simulated structure were modeled by the QwikMD plugin[49] in VMD. QM region I (QM model I) contains all of the atoms of tunicamycin inhibitor, Gln44, Asp45, Leu46, Asn47, Glu56, Asp115, Asn119, Trp122, Lys125, Asn185, Asn191, Gly192, Glu194, Asn242, Phe249, Asp252, Arg301, His302, Arg303, Ile304, Glu376, Wt601, Wt603, Wt604, Wt606, and Wt609 present in the tunicamycin-binding pocket of the DPAGT1–TUN structure (Figure ). QM region II (QM model II) includes all of the atoms of tunicamycin inhibitor, Lys97, Lys111, Asn172, Phe173, Asp175, Gly176, Leu177, Asp178, Cys181, Asn221, Phe228, Asp231, Phe228, Met229, His290, His291, Wt501, Wt502, and Wt503 present in the tunicamycin-binding pocket of the MraYCB–TUN structure (Figure ). Accordingly, QM models I and II consist of 464 and 333 QM atoms, respectively. All QM atoms of each model were optimized at the M06-2X/6-31G** level, whereas the MM regions were described by the CHARMM36 force field.[41] During the QM/MM calculations, the positions of the non-hydrogen atoms of each QM model were kept frozen. The QM/MM calculations were performed with the integration time step of 0.5 fs by using the NAMD/ORCA interface.[50−52] The electrostatic interactions between the QM region and the partial charges of MM atoms surrounding all QM atoms were treated by an electrostatic embedding scheme.[53,54] The QM/MM interface was modeled by hydrogen-linked atoms, which were saturated in the QM region, with the charge shift (CS) treatment.[55,56]

Figure 3

Tunicamycin-binding pocket in the DPAGT1–TUN complex structure, considered as the QM model I. This model includes all of the atoms of tunicamycin inhibitor, residues, and water molecules present in the uracil, tunicamine, fatty acyl tail, and GlcNAc binding pockets, tallying up to a total number of 464 QM atoms.

Figure 4

Tunicamycin-binding pocket in the MraYCB–TUN complex structure, considered as the QM model II. This model includes all of the atoms of tunicamycin inhibitor, residues, and water molecules present in the uracil, tunicamine, and GlcNAc binding pockets, accounting for a total number of 333 QM atoms.

The second alternative is to continue the calculations on the active site (tunicamycin-binding pocket) of the final frame of each structure equilibrated in the NVT ensemble at the DFT level by means of the GAMESS electronic structure package.[57] To implement DFT computations, two structural models were constructed on the basis of the TUN and its encompassed residues within the active sites of equilibrated DPAGT1–TUN (structural model I, Figure S1) and equilibrated MraYCB–TUN (structural model II, Figure S2). In both models, the N-terminal and the C-terminal of each free residue were closed by functional groups of -CH3 and -OCH3, respectively. Because the correct description of the characteristics of a hydrogen bond (H-bond) requires determining accurately hydrogen atom coordinates and the classical MD simulation is unable to specify such properties, the positions of all hydrogens in each structural model were optimized at the M06-2X/6-31G** level. The partial geometry optimizations were done by holding fixed the non-hydrogen atoms and by including the solvent effects using the polarizable continuum model (PCM) approach.[58−60]

Used Quantum Chemical Methods to Characterize Intermolecular Interactions in the Tunicamycin-Binding Pockets

QTAIM analysis is one of the most widely used theoretical tools to characterize the physical nature of intramolecular and intermolecular interactions in terms of the topological properties of the electron density, ρ(r), the bond path (BP), and the bond critical point (BCP).[33−35] From the viewpoint of Bader’s theory[34,61] and in accordance with the Cremer–Kraka formulation,[62,63] the topological parameters extracted from the analysis of electron density at BCP, ρBCP(rcp), including its Laplacian, ∇2ρBCP(rcp), local potential energy density, VBCP, local kinetic energy density, GBCP, total electronic energy density, HBCP, present valuable information concerning the nature and strength of the detected interaction at this BCP. Generally, if ρBCP is larger than 0.20 au and its ∇2ρBCP is a large negative value, the interested interaction is defined as a shared (covalent) bond, whereas a closed-shell (such as vdW, ionic, H-bonding, H–H-bonding, etc.) interaction has ρBCP < 0.10 au and a low positive value of ∇2ρBCP.[26,64,65] A closed-shell interaction is considered as a H-bond when ρBCP on the bond path between the hydrogen atom and proton acceptor (H···B BP) is in the range of 0.002–0.040 au and its ∇2ρBCP lies within 0.020–0.150 au[66]

According to Espinosa’s relationship,[67−69] there is a correlation between VBCP and the H-bonding interaction energy (EHB) asEHB is an appropriate energetic criterion to estimate the H-bonding interaction strength. As a convention, the energy extent of strong H-bonding interactions varies between 62.76 and 167.36 kJ/mol (15 and 40 kcal/mol). The energy range of moderate (normal) H-bonds locates between 16.74 and 62.76 kJ/mol (4 and 15 kcal/mol) and the energy magnitude of weak H-bonds is in the range of 4.18–16.74 kJ/mol (1–4 kcal/mol).[70−75] QTAIM analysis was carried out by employing the AIM 2000 program package.[76]

From a NBO theory point of view,[36,37,77] the H-bonding interaction is the result of a charge transfer (CT) from the lone pair orbital of an electron donor (proton acceptor), nB, into the valence antibonding orbital of an electron acceptor (proton donor), σ*A–H. The energy of CT interaction, nB → σA–H*, called the second-order stabilization energy, E(2), is another suitable criterion for evaluating the H-bonding interaction strength and is calculated by the second-order perturbation theory as follows:where ⟨nB|F|σA–H*⟩ and ε(σA–H*) – ε(nB) are the Fock matrix element and energy difference between the donor and the acceptor orbitals, respectively.

To compare the intermolecular interaction strengths of TUN with the interacting residues in each active site, the interaction energy of each TUN–residue pair was calculated and corrected for the basis set superposition error (BSSE), applying the counterpoise (CP) correction method by the equation given as follows:[78,79]where EAB is the single point energy (SPE) of each pair, while EA and EB are SPEs of the isolated residue and tunicamycin, respectively. The evaluations of interaction energies as well as QTAIM and NBO analyses were implemented on each QM model as well as on each optimized structural model at the respective level. In the Supporting Information (Tables S7–S10), the atomic coordinates of two optimized QM models and two optimized structural models, which were used for QTAIM and NBO analyses, can be found.

Results and Discussion

In this study, since MD simulations were combined with QM/MM calculations, our findings are reported and discussed in two parts. First, the results extracted from the MD simulations are presented and then the QM/MM outcomes are described in detail.

MD Simulations of DPAGT1–TUN and MraYCB–TUN Complexes

DPAGT1–TUN and MraYCB–TUN complexes are not static proteins in the explicit solvent, and their atomic positions repeatedly undergo dynamical rearrangements in the water boxes over the simulation times. This is continued until each of these simulated proteins achieves its equilibrated structure in the explicit solvent under periodic boundary conditions. It is clear that the final atomic coordinates in each equilibrated structure are deviated from the initial atomic positions in the reference structure (i.e., the crystal structure). The root-mean-square deviation, RMSD,[42] is thus applied as a numeric value to evaluate the deviation amount of each equilibrated protein from its reference structure. Generally, the equilibrated protein has an average RMSD, av.RMSD, between 0.5 and 3 Å and its standard deviation of RMSD, sd.RMSD, is less than 1 Å.[80−82]Figure displays the plot of RMSD values of these two simulated proteins against the simulation times. Our results indicate that although the av.RMSD (1.96 Å) of equilibrated DPAGT1–TUN is somewhat smaller than that (2.05 Å) of equilibrated MraYCB–TUN, they have the same equilibrium time (15 ns). Hence, despite very low resolution (3.40 Å) of the initial atomic coordinates in the DPAGT1–TUN reference structure, this protein, especially its active site, has experienced less dynamical rearrangements during the MD simulations than MraYCB–TUN.

Figure 5

Plots of RMSD values of DPAGT1–TUN and MraYCB–TUN versus the simulation times. (RMSD values were calculated for all atoms of each protein–TUN without its hydrogen atoms.)

The nonbonded (total) energy between each protein and TUN is classically defined as the sum of the electrostatic energy and the van der Waals energy between them.[41] The graphical outputs of electrostatic, van der Waals, and nonbonded energies of DPAGT1–TUN and MraYCB–TUN evaluated over the simulation times are depicted in Figures S3 and S4 of the Supporting Information. MD energy analyses show that the electrostatic interactions are mainly responsible for the TUN binding stability to each protein active site because they have larger contributions to the nonbonded energies than the van der Waals interactions. Besides, the TUN binding strength to DPAGT1 is approximately twice that to MraYCB (Supporting Information Table S11).

QTAIM and NBO Analyses on QM Models I and II

In the following sections, the QTAIM analysis is utilized to characterize the diverse conventional and unconventional H-bonds of TUN with the neighboring residues and water molecules inside the tunicamycin-binding pockets of DPAGT1–TUN and MraYCB–TUN. The second-order perturbation theory in the framework of NBO analysis is applied to evaluate the strength of intermolecular orbitals of the partner atoms in each identified H-bond. Although H-bonds are the most important interactions providing the TUN binding to each active site, only some of its residues are able to form H-bonds with TUN. In addition to the H-bonding interactions, it is apparent that the electrostatic and van der Waals interactions also play significant roles in stabilizing the tunicamycin-binding pockets. Accordingly, the CP correction method is used to estimate the total intermolecular interaction energy of per TUN–residue/Wt pair in each model.

Tunicamycin-Binding Pocket in DPAGT1–TUN (QM Model I)

Table presents the geometrical and topological parameters as well as the modulus of the H-bond energies, |EHB|, pertaining to the H-bonding interactions detected in the QM model I. The outcomes of NBO analysis of the donor–acceptor orbital partners in these H-bonds are tabulated in Table . The modulus of interaction energies, |Einteraction|, BSSE energies, and dipole moments of different fragment pairs, which can be TUN–residue/Wt pairs, Wt–residue pairs, or a Wt–Wt pair, of this model are collected in Table . It is worth noting that all of the TUN atoms, except its hydrogen atoms, are numbered according to their positions in the DPAGT1–TUN crystal structure,[1] while the labeling of the hydrogen atoms is based on their positions in the created topology file (see the Supporting Information). In the following, the results of these three tables are discussed in detail in five separate sections.

Table 1

Structural and Topological Parameters of ρBCP(r) Corresponding to the H-Bonds in the TUN–Residue/Wt Pairs and Wt–Residue Pairs of QM Model I Computed at the M06-2X/6-31G** Levela

proton donor	proton acceptor	bond path	d	∠	ρBCP	∇2ρBCP	HBCP	|EHB|
Asp45	TUN	Hα···O15	2.40	139.36	0.0115	0.0381	–0.0079	11.07
Leu46	TUN	H···O15	1.84	150.98	0.0306	0.1130	–0.0246	33.26
Trp122	TUN	Hε1···O8	2.59	146.43	0.0067	0.0246	–0.0036	5.91
Lys125	TUN	Hζ1···O9	2.10	158.86	0.0175	0.0496	–0.0148	18.27
Wt601	TUN	H1···O9	1.82	151.01	0.0316	0.1152	–0.0275	36.72
Asn191	TUN	Hα···O14	2.48	137.69	0.0094	0.0329	–0.0060	8.84
Asn191	TUN	Hδ22···O14	1.87	159.65	0.0283	0.1002	–0.0231	31.16
Wt603	TUN	H2···O14	1.89	165.10	0.0306	0.0992	–0.0284	35.70
Arg301	TUN	Hη11···O1	2.62	135.45	0.0071	0.0242	–0.0038	5.98
Arg301	TUN	Hη21···O1	2.04	164.34	0.0197	0.0649	–0.0164	21.49
Arg301	TUN	Hη11···O2	2.26	160.40	0.0140	0.0408	–0.0105	13.69
His302	TUN	Hα···O5	2.46	129.77	0.0101	0.0339	–0.0064	11.45
His302	TUN	Hδ1···O5	1.77	156.32	0.0357	0.1241	–0.0135	38.15
Arg303	TUN	H···O5	1.74	167.27	0.0426	0.1456	–0.0370	48.30
Arg303	TUN	Hε···O6	2.05	165.75	0.0188	0.0578	–0.0156	19.96
Arg303	TUN	Hη22···O7	1.94	155.13	0.0240	0.0795	–0.0194	25.67
Ile304	TUN	Hγ21···O4	2.50	129.24	0.0098	0.0323	–0.0061	9.78
Ile304	TUN	Hγ21···O5	2.60	156.99	0.0092	0.0300	–0.0051	7.77
Asn191	Wt603	Hα···O	2.72	129.47	0.0062	0.0219	–0.0030	5.00
Gly192	Wt603	H···O	2.41	157.04	0.0112	0.0342	–0.0078	10.55
Asn242	Wt603	Hδ21···O	2.07	155.33	0.0179	0.0556	–0.0149	19.13
Asn47	Wt604	Hβ1···O	2.42	139.96	0.0129	0.0364	–0.0089	11.77
Asn47	Wt604	Hδ22···O	1.95	155.16	0.0252	0.0800	–0.0212	29.95
Asn185	Wt609	Hδ22···O	2.66	115.69	0.0064	0.0239	–0.0034	5.89
Wt606	Wt601	H1···O	2.13	154.31	0.0175	0.0493	–0.0146	20.74
TUN	Asp45	H31···Oδ1	1.68	159.90	0.0428	0.1436	–0.0340	45.46
TUN	Wt604	H31···O	1.99	168.50	0.0265	0.0796	–0.0239	29.63
TUN	Wt603	H351···O	2.73	126.53	0.0061	0.0231	–0.0028	5.00
TUN	Gln44	H12···O	1.65	166.59	0.0452	0.1674	–0.0370	50.73
TUN	Glu56	H131···O	1.60	176.01	0.0588	0.1801	–0.0557	65.35
TUN	Wt609	H···O	2.69	124.99	0.0059	0.0241	–0.0030	5.16
TUN	Wt606	H101···Oε1	2.52	150.36	0.0087	0.0312	–0.0054	8.15
TUN	Trp122	H151···Nε1	2.53	135.00	0.0093	0.0324	–0.0050	7.96
TUN	Asn185	H141···Oδ1	2.28	139.25	0.0133	0.0435	–0.0108	13.51
TUN	Asn185	HN1···Oδ1	1.76	163.84	0.0355	0.1268	–0.0281	38.41
TUN	Asp252	H281···Oδ1	2.40	128.48	0.0124	0.0403	–0.0085	13.07
TUN	Asp252	H91···Oδ2	1.81	153.13	0.0341	0.1130	–0.0272	37.63
TUN	Arg301	H61··· O	2.61	132.81	0.0066	0.0285	–0.0039	6.51
TUN	Arg301	H81···O	2.48	136.63	0.0103	0.0327	–0.0068	9.48
TUN	Arg303	H61···N	2.69	154.77	0.0066	0.0218	–0.0035	5.46
Wt603	Glu194	H1···Oε1	1.86	164.73	0.0285	0.0925	–0.0230	30.24
Wt603	Asn242	H1···Oδ1	2.32	141.95	0.0119	0.0356	–0.0090	12.41
Wt604	Asp45	H2···Oδ1	1.80	168.43	0.0344	0.0577	–0.0269	35.85
Wt604	Glu376	H1···Oε2	1.53	163.70	0.0764	0.1459	–0.0763	89.00
Wt606	Asp115	H2··· O	2.19	169.33	0.0141	0.0455	–0.0111	14.72
Wt606	Asn119	H1···Oδ1	1.87	158.30	0.0281	0.0986	–0.0232	31.08

The H-bond length (d) is in angstroms (Å), and the H-bond angle (∠) is in degrees (deg). All ρBCP, ∇2ρBCP, and HBCP parameters are in atomic units (au). |EHB| is in kJ/mol.

Table 2

NBO Results of Partner Orbitals Participated in Donor–Acceptor Interactions in the TUN–Residue/Wt Pairs and Wt–Residue Pairs of QM Model I Assessed at the M06-2X/6-31G** Level

electron donor	electron acceptor	charge transfer	E(2) (kJ/mol)	qnB→σA–H*(e)
TUN	Asp45	nO15 → σCα–Hα*	6.17	0.0032
TUN	Leu46	nO15 → σN–H*	47.76	0.0135
TUN	Trp122	nO8 → σNε1–Hε1*	2.76	0.0012
TUN	Lys125	nO9 → σNζ–Hζ1*	21.28	0.0071
TUN	Wt601	nO9 → σO–H1*	46.36	0.0165
TUN	Asn191	nO14 → σCα–Hα*	5.09	0.0015
TUN	Asn191	nO14 → σNδ2–Hδ22*	38.94	0.0109
TUN	Wt603	nO14 → σO–H2*	44.08	0.0189
TUN	Arg301	nO1 → σNη1–Hη11*	2.11	0.0008
TUN	Arg301	nO1 → σNη2–Hη21*	35.76	0.0151
TUN	Arg301	nO2 → σNη1–Hη11*	7.95	0.0033
TUN	His302	nO5 → σCα–Hα*	6.36	0.0021
TUN	His302	nO5 → σNδ1–Hδ1*	70.42	0.0259
TUN	Arg303	nO5 → σN–H*	71.95	0.0330
TUN	Arg303	nO6 → σNε–Hε*	35.88	0.0115
TUN	Arg303	nO7 → σNη2–Hη22*	39.85	0.0119
TUN	Ile304	nO4 → σCγ2–Hγ21*	1.14	0.0004
TUN	Ile304	nO5 → σCγ2–Hγ21*	4.54	0.0020
Wt603	Asn191	nO → σCα–Hα*	4.10	0.0017
Wt603	Gly192	nO → σN–H*	7.85	0.0029
Wt603	Asn242	nO → σNδ2–Hδ21*	13.12	0.0043
Wt604	Asn47	nO → σCβ–Hβ1*	4.27	0.0013
Wt604	Asn47	nO → σNδ2–Hδ22*	57.36	0.0212
Wt609	Asn185	nO → σNδ2–Hδ22*	3.12	0.0012
Wt601	Wt606	nO → σO–H1*	31.87	0.0107
Gln44	TUN	nO → σO12–H12*	85.02	0.0248
Asp45	TUN	nOδ1 → σN3–H31*	84.96	0.0272
Wt604	TUN	nO → σN3–H31*	61.55	0.0234
Wt603	TUN	nO → σC35–H351*	3.56	0.0014
Glu56	TUN	nOε1 → σO13–H131*	91.99	0.0421
Wt609	TUN	nO → σO–H*	3.64	0.0015
Wt606	TUN	nO → σO10–H101*	10.51	0.0043
Trp122	TUN	nNε1 → σC15–H151*	1.63	0.0009
Asn185	TUN	nOδ1 → σN1–HN1*	40.90	0.0119
Asn185	TUN	nOδ1 → σC14–H141*	8.62	0.0024
Asp252	TUN	nOδ1 → σC28–H281*	2.68	0.0013
Asp252	TUN	nOδ2 → σO9–H91*	73.91	0.0291
Arg301	TUN	nO → σC8–H81*	3.81	0.0012
Arg303	TUN	nN → σO6–H61*	4.30	0.0023
Glu194	Wt603	nOε2 → σO–H1*	41.89	0.0231
Asn242	Wt603	nOδ1 → σO–H1*	12.62	0.0060
Asp45	Wt604	nOδ1 → σO–H2*	45.04	0.0197
Glu376	Wt604	nOε2 → σO–H1*	102.03	0.0473
Asp115	Wt606	nO → σO–H2*	10.48	0.0030
Asn119	Wt606	nOδ1 → σO–H1*	21.32	0.0059

Table 3

Modulus of the Calculated BSSE-Corrected Interaction Energies, BSSE Energies, and Dipole Moments of the Different Fragment Pairs, Which Can Be TUN–Residue/Wt Pairs or Wt–Residue Pairs in the QM Model I Evaluated at the M06-2X/6-31G** Level

fragment pair	|Einteraction| (kJ/mol)	BSSE energy (kJ/mol)	dipole (debye)
TUN–Gln44	29.36	11.34	12.75
TUN–Asp45	63.15	20.19	58.64
TUN–Leu46	10.60	9.08	13.23
TUN–Glu56	150.10	16.98	30.50
TUN–Trp122	31.23	11.76	12.46
TUN–Lys125	65.87	7.62	35.16
TUN–Leu126	4.18	2.31	9.27
TUN–Leu175	3.66	1.46	10.31
TUN–Val178	2.93	1.02	8.40
TUN–Phe179	1.44	1.28	9.76
TUN–Asn185	29.73	12.05	12.85
TUN–Ile186	3.56	7.85	11.55
TUN–Ala188	6.65	6.78	9.40
TUN–Gly189	6.42	6.35	9.72
TUN–Ile190	9.09	7.50	9.25
TUN–Asn191	17.49	10.34	14.99
TUN–Gly192	3.11	1.91	9.50
TUN–Phe249	9.48	8.53	9.36
TUN–Asp252	183.25	18.73	14.56
TUN–Phe286	6.72	3.59	10.60
TUN–Leu293	6.21	4.12	9.30
TUN–Cys299	0.74	0.29	7.45
TUN–Pro300	5.52	0.74	7.94
TUN–Arg301	80.09	5.57	35.80
TUN–His302	76.42	16.15	19.31
TUN–Arg303	158.43	23.85	19.79
TUN–Ile304	23.28	11.19	14.68
TUN–Wt601	14.80	9.61	9.68
TUN–Wt603	31.40	11.25	10.52
TUN–Wt604	29.86	9.60	10.77
TUN–Wt606	23.07	11.41	12.82
TUN–Wt609	12.98	9.10	10.82
Wt601–Wt606	19.55	7.70	3.04
Wt603–Asn191	3.93	7.39	5.18
Wt603–Gly192	4.45	9.55	1.71
Wt603–Glu194	53.86	10.93	10.92
Wt603–Asn242	36.32	9.75	3.29
Wt604–Asp45	69.15	11.42	12.70
Wt604–Asn47	19.96	8.07	6.43
Wt604–Glu376	70.68	21.05	12.65
Wt606–Asp115	25.60	7.49	14.30
Wt606–Asn119	12.71	9.67	5.24
Wt609–Asn185	10.58	8.04	4.15

Uracil-Binding Pocket

As shown in Figure S1, the TUN uracil base is enclosed by Asp45, Leu46, Asn47, Gly189, Ile190, Asn191, Gly192, Glu194, Asn242, Phe249, Glu376, Wt603, and Wt604. The QTAIM analysis reveals that the uracil base amide nitrogen and its carbonyl oxygens are joined to the backbones and the side chains of Asp45, Leu46, and Asn191 as well as to Wt603 hydrogen and Wt604 oxygen by seven BPs (Figure a,b). The ρBCP, ∇2ρBCP, and |EHB| values of these BCPs demonstrate the formation of weak unconventional Cα–Hα···O15 and moderate conventional N3–H31···Oδ1 H-bonds in the TUN–Asp45 pair, weak unconventional Cα–Hα···O14 and moderate conventional Nδ2–Hδ22···O14 H-bonds in the TUN–Asn191 pair, and three moderate conventional H-bonds of N–H···O15, O–H2···O14, and N3–H31···O in the TUN–Leu46, TUN–Wt603, and TUN–Wt604 pairs, respectively (Table ). Likewise, C35–H351···O with |EHB| of 5.00 kJ/mol exists between the uracil base and Wt603. Among these eight H-bonds, N3–H31···Oδ1 is the strongest interaction because its |EHB| value (45.46 kJ/mol) is the highest compared with the other H-bonds of this moiety.

Figure 6

Uracil-binding pocket in the DPAGT1–TUN, stabilized by the H-bonding interactions among the uracil base: (a) Asn191, Gly192, Glu194, Asn242, and Wt603; (b) Asp45, Leu46, Asn47, Glu376, and Wt604.

The NBO analysis indicates that the nO14 lone pair of the uracil base overlaps synchronously with the antibonding orbitals of σCα–Hα* and σNδ2–Hδ22* in Asn191 as well as with the antibonding σO–H2* in Wt603. Its nO15 donates a charge of 0.0135 e to σN–H* of Leu46 and a charge of 0.0032 e to σCα–Hα* of Asp45. Moreover, its σN3–H31* accepts separately the qCT values of 0.0272 and 0.0234 e from the lone pair orbitals of nOδ1 in Asp45 and nO in Wt604, respectively. Finally, there is an interaction between the σC35–H351* of the uracil base and the nO of Wt603. Consequently, in agreement with the QTAIM findings, TUN is involved in eight CT interactions with Asp45, Leu46, Asn191, Wt603, and Wt604. Of these, the largest E(2) (84.96 kJ/mol) belongs to the nOδ1 → σN3–H31* interaction in the TUN–Asp45 pair (Table ). In line with the |EHB| prediction of N3–H31···Oδ1, it is hence the strongest donor–acceptor interaction of the uracil base.

In addition to TUN, the Wt603 and Wt604 oxygens and hydrogens are connected to the Gly192 backbone amide nitrogen, Asn242 side chain amide group, Asn47 side chain amide nitrogen, and side chain carboxylate oxygens of Asp45, Glu194, and Glu376 through seven BPs (Figure a,b). These two water molecules thus form seven conventional H-bonds with these cited residues in the strength ranging from weak to strong interactions (Table ). There are also two weak unconventional H-bonds of Cα–Hα···O and Cβ–Hβ1···O in Wt603–Asn191 and Wt604–Asn47 pairs, respectively (Figure a,b). Because |EHB| (30.24 kJ/mol) of O–H1···Oε1 in the Wt603–Glu194 pair is approximately equal to that (29.95 kJ/mol) of Nδ2–Hδ22···O in the Wt604–Asn47 pair, the strengths of these two H-bonds are almost the same. It is worth pointing out that O–H1···Oε2 with a length of 1.53 Å, an angle of 163.70°, and a predominantly covalent character[83] in the Wt604–Glu376 pair is the strongest H-bond in the QM model I because it has the largest amounts of ρBCP (0.0764 au), ∇2ρBCP (0.1459 au), and |EHB| (89.00 kJ/mol) compared to the other H-bonds of this model. As seen in Table , the highest values of E(2) (102.03 kJ/mol) and qCT (0.0473 e) are assigned to the nOε2 → σO–H1* interaction that is responsible for the creation of O–H1···Oε2. Therefore, the strongest donor–acceptor interaction of this model is also found in the Wt604–Glu376 pair.

Due to the negative charge nature of the side chains Asp and Glu, the electrostatic interactions of charge–dipole and dipole–dipole types are expected to take place in negatively charged pairs of TUN–Asp45, Wt603–Glu194, Wt604–Asp45, and Wt604–Glu376. Because of the uncharged and polar character of the Asn side chain amide group and water molecules, the significance of dipole–dipole interactions is conspicuous in the polar pairs of TUN–Asn191, TUN–Wt603, TUN–Wt604, Wt603–Asn191, Wt603–Asn242, and Wt604–Asn47. The dipole moments on the uracil functional groups and Wt603 oxygen induce the dipoles, called the induced dipole moments, in the nonpolar side chains of residues Leu46, Gly189, Ile190, Gly192, and Phe249. Therefore, dipole–induced dipole interactions emerge in TUN–Leu46, TUN–Gly189, TUN–Ile190, TUN–Gly192, TUN–Phe249, and Wt603–Gly192 pairs. Because the |Einteraction| and dipole moment values in these four negatively charged pairs are the largest relative to those in the cited polar pairs and those in TUN–nonpolar residue pairs (Table ), the former pairs have the strongest intermolecular interactions compared to the latter two types of pairs.

Since Asp45, Wt604, and uracil base are involved in H-bonding interactions together, hydrogen-bonded (H-bonded) networks of N3–H31···O–H2···Oδ1···H31–N3 are formed among them (Figure b). Similarly, Asn191, Wt603, and uracil base form H-bonded networks of Cα–Hα···O–H2···O14···Hα–Cα with each other (Figure a). These networks lead to an increase in the strengths of uracil–Wt604–Asp45 and uracil–Wt603–Asn191 interactions. Although there is no interaction between TUN and Glu376, it is associated with the uracil–Wt604 interaction by O–H1···Oε2. Moderate–strong N3–H31···O–H1···Oε2 H-bonded networks thus appear in uracil–Wt604–Glu376. Likewise, moderate–moderate N3–H31···O···Hδ22–Nδ2 and moderate–weak N3–H31···O···Hβ1–Cβ H-bonded networks exist in uracil–Wt604–Asn47 (Figure b). Furthermore, the strengths of uracil–Wt603 interactions are enhanced by the appearance of weak–weak C35–H351···O–H1···Oδ1 and moderate–moderate Nδ2–Hδ21···O–H2···O14 H-bonded networks in Uracil–Wt603–Asn242, weak-moderate H-bonded networks of C35–H351···O–H1···Oε1 and N–H···O–H2···O14 in uracil–Wt603–Glu194 and uracil–Wt603–Gly192, respectively (Figure a). Even though Asn47, Gly192, Glu194, Asn242, and Glu376 have no H-bonding and CT interactions with TUN, they accompany uracil–Wt603 and uracil–Wt604 interactions by the formation of H-bonds with water molecules. Hence, they indirectly assist in the stability of this pocket. It is worth stressing that Wt603 and Wt604 play essential roles in maintaining and stabilizing the uracil-binding pocket of DPAGT1–TUN via dipole–dipole interactions concomitant with H-bonded networks, as discussed above (Figure a,b).

Tunicamine-Binding Pocket

Gln44, Glu56, Asp115, Asn119, Lys125, Asn185, Asp252, Arg301, Wt601, Wt606, and Wt609 are located around the tunicamine sugar moiety (Figure S1). The QTAIM analysis reveals that the ribosyl hydroxyls are joined to the Gln44 backbone carbonyl oxygen, Glu56 side chain carboxylate oxygen (Oε1), and Wt609 oxygen by three BPs (Figure a). The values of ρBCP (0.0452 au), ∇2ρBCP (0.1674 au), and |EHB| (50.73 kJ/mol) of H12···O BCP confirm that the O12–H12···O H-bond in the TUN–Gln44 pair is a moderate interaction, while O–H···O with |EHB| of 5.16 kJ/mol in the TUN–Wt609 pair is a weak H-bond. In contrast, on the basis of the topological parameters on H131···Oε1 BCP in the TUN–Glu56 pair, namely, ρBCP = 0.0588 au, ∇2ρBCP = 0.1801 au, and HBCP = −0.0557 au, O13–H131···Oε1 is a strong H-bond with a length of 1.60 Å, an angle of 176.01°, and a basically covalent nature. The appearance of this H-bond is the consequence of a CT (0.0421 e) from the nOε1 of Glu56 into σO13–H131* of the ribose sugar. It is evident from the results in Tables and 2 that the largest values of |EHB| (65.35 kJ/mol) and E(2) (91.99 kJ/mol) are attributed to the O13–H131···Oε1 H-bond and nOε1 → σO13–H131* interaction, respectively. Therefore, H-bonding and donor–acceptor interactions between the ribosyl moiety and Glu56 are the strongest interactions of TUN in this tunicamycin-binding pocket. Moreover, large |Einteraction| (150.10 kJ/mol) and dipole moment (30.50 debye) of this negatively charged pair signify that charge–dipole and dipole–dipole interactions also have significant contributions to its intermolecular interactions.

Figure 7

Tunicamine-binding pocket in DPAGT1–TUN, maintained by the H-bonding interactions among tunicamine sugar: (a) Gln44, Glu56, Asn185, and Wt609; (b) Lys125, Asp252, and Arg301; (c) Asp115, Asn119, Wt601, and Wt606.

The Asn185 side chain carbonyl oxygen forms a moderate N1–HN1···Oδ1 H-bond with tunicamine and a weak C14–H141···Oδ1 H-bond with the fatty acyl tail, while its amide nitrogen is involved in a weak Nδ2–Hδ22···O H-bond with Wt609 (Figure a). It is thus expected that weak–weak O–H···O···Hδ22–Nδ2 H-bonded networks are formed among the ribosyl moiety, Wt609, and Asn185. Because |Einteraction| (29.73 kJ/mol) in the TUN–Asn185 pair is only slightly different from that (29.36 kJ/mol) in the TUN–Gln44 pair, the strengths of intermolecular interactions in these two polar pairs are equivalent. The tunicamine O1 and O2 nuclei play the proton-acceptor (H-acceptor) roles in the formation of three H-bonds of Nη1–Hη11···O1, Nη2–Hη21···O1, and Nη1–Hη11···O2 with Arg301 (Figure b), and its O10 nucleus acts as a proton-donor (H-donor) in the interaction of O10–H101···O with Wt606 (Figure c). Of these, the Nη2–Hη21···O1 with |EHB| of 23.72 kJ/mol is a moderate interaction, but the other three are treated as weak H-bonds (Table ). In contrast, the tunicamine hydroxyl oxygen (O9) acts concurrently as both H-donor and H-acceptor in the formation of three moderate H-bonds of O9–H91···Oδ2, Nζ–Hζ1···O9, and O–H1···O9 with Asp252, Lys125, and Wt601, respectively (Figure b,c). It is worth noting that TUN in the QM model I has the strongest intermolecular interactions with Asp252 because the TUN–Asp252 pair has the largest |Einteraction| (183.25 kJ/mol) compared to the other TUN–residue pairs of this model (Table ). It is reasonable to suggest that Asp252 is one of the most important residues in DPAGT1–TUN, which plays a critical role in the TUN binding to its pocket mainly via extremely strong charge–dipole and dipole–dipole interactions. Because of the positive charge nature of the side chains Lys and Arg, charge–dipole and dipole–dipole interactions have important contributions to the intermolecular interactions of TUN–Lys125 and TUN–Arg301 pairs. Because |Einteraction| (80.09 kJ/mol) in the TUN–Arg301 pair is larger than that (65.87 kJ/mol) in the TUN–Lys125 pair, intermolecular interactions in the former pair are stronger than those in the latter pair.

In addition to TUN, Wt606 interacts weakly with Asp115 through O–H2···O and moderately with Asn119 and Wt601 via O–H1···Oδ1 and O–H1···O, respectively (Table ). The strengths of tunicamine–Wt interactions are thus increased as the result of the formation of H-bonded networks: weak–weak O10–H101···O–H2···O in tunicamine–Wt606–Asp115, weak–moderate O10–H101···O–H1···Oδ1 in tunicamine–Wt606–Asn119, and weak–moderate–moderate O10–H101···O–H1···O–H1···O9 in tunicamine–Wt606–Wt601 (Figure c). Therefore, Asp115 and Asn119 are the other two important residues of DPAGT1–TUN that indirectly affect the tunicamine-binding pocket stabilization by forming cooperative H-bonded networks in tunicamine–Wt interactions. In summary, these five water molecules are intercalated between the key residues and TUN and interact with both of them due to their donor–acceptor duality (Figure a,b and Figure a,c). During charge transfer interactions, Wt601, Wt603, Wt604, Wt606, and Wt609 exchange the charges of 0.0272, 0.0583, 0.1129, 0.0239, and 0.0027 e, respectively, with the uracil and tunicamine moieties as well as with their interacting residues. Accordingly, the water-mediated interactions behave as multibranched wires which exchange electrons between the donor–acceptor orbitals of partner atoms along the H-bonded networks. Therefore, water molecules play critical roles in conserving and stabilizing the situations of the uracil and tunicamine moieties within the tunicamycin-binding pocket via the specified H-bonded networks and dipole–dipole interactions.

Fatty Acyl Tail-Binding Pocket

The fatty acyl tail of TUN is considered as an amphipathic[84] moiety because it contains both a hydrophilic polar headgroup (carbonyl group, e.g., C13=O8) and a hydrophobic chain (monounsaturated aliphatic tail). For this reason, its hydrophobic portion is inserted into the interior of the DPAGT1 active site adjacent to the hydrophobic (nonpolar) residues Trp122 (the aromatic ring of its side chain indole ring), Leu126, Leu175, Val178, Phe179, Ile186, Phe286, and Leu293. Its hydrophilic portion is instead near the pyrrole ring of the Trp122 indole ring (Figure S1). Among these residues, only the Trp122 pyrrole ring is able to participate in H-bonding (Nε1–Hε1···O8 and C15–H151···Nε1) and CT (nO8 → σNε1–Hε1*; nNε1 → σC15–H151*) interactions with the fatty acyl tail (Figure ). Moreover, this hydrophilic group, due to its polar nature, induces a dipole moment in the pyrrole ring, resulting in the addition of a dipole–induced dipole interaction contribution to intermolecular interactions between them. The Trp122 aromatic ring accompanying other cited nonpolar residues interacts with the aliphatic tail via van der Waals interactions of the hydrophobic type. As a consequence, Trp122 is an indispensable residue in the hydrophobic region of the DPAGT1–TUN active site that plays an essential role in situating the fatty acyl tail into its binding pocket through the hydrophobic, electrostatic, and H-bonding interactions. As expected, the fatty acyl tail has the strongest intermolecular interactions with Trp122 because the |Einteraction| (31.23 kJ/mol) and the dipole moment (12.46 debye) in the TUN–Trp122 pair are the largest relative to those in the other TUN–nonpolar residue pairs of this hydrophobic pocket (Table ). Since the hydrophobic interactions are the major contributor to intermolecular interactions among the aliphatic tail and side chains of residues Leu126, Leu175, Val178, Phe179, Ile186, Phe286, and Leu293, these TUN–nonpolar residue pairs have very small |Einteraction| values. Because |Einteraction| (6.72 kJ/mol) in the TUN–Phe286 pair is nearly equal to that (6.21 kJ/mol) in the TUN–Leu293 pair, their intermolecular interaction strengths are the same.

Figure 8

Trp122, Leu126, Leu175, Val178, Phe179, Ile186, Phe286, and Leu293 providing fatty acyl tail binding to the DPAGT1–TUN active site via hydrophobic, dipole–induced dipole, and H-bonding interactions.

GlcNAc-Binding Pocket

As depicted in Figure S1, the GlcNAc moiety is surrounded by Ala188, Cys299, Pro300, His302, Arg303, and Ile304. The QTAIM analysis recognizes H-acceptor roles of the GlcNAc carbonyl oxygen and its hydroxyl oxygens in the formation of three moderate H-bonds of N–H···O5, Nε–Hε···O6, and Nη2–Hη22···O7 with Arg303, two weak H-bonds of Cγ2–Hγ21···O4 and Cγ2–Hγ21···O5 with Ile304, and one weak Cα–Hα···O5 and one moderate Nδ1–Hδ1···O5 H-bonds with His302. GlcNAc also interacts weakly with Arg303 through the O6–H61···N H-bond (Figure and Table ). Besides, two weak H-bonds of O6–H61···O and C8–H81···O exist between GlcNAc and Arg301 (Figure b). Among these ten H-bonds, the largest |EHB| (48.30 kJ/mol) corresponds to N–H···O5 in the TUN–Arg303 pair; it is thus the strongest H-bond of GlcNAc in this model. In accordance with the |EHB| prediction for this H-bond, the NBO analysis also reveals that the nO5 → σN–H* with an E(2) of 71.95 kJ/mol and a qCT of 0.0330 e is the strongest donor–acceptor interaction of this moiety (Table ).

Figure 9

His302, Arg303, and Ile304 interacting with GlcNAc through charge–dipole, dipole–dipole, dipole–induced dipole, and H-bonding interactions.

Because the TUN–Arg303 pair is a positively charged pair with a dipole moment of 19.79 debye, the Arg303 side chain guanidine group and its backbone interact strongly with the GlcNAc carbonyl and hydroxyl oxygens via charge–dipole and dipole–dipole interactions. Very large |Einteraction| (158.43 kJ/mol) of this pair demonstrates the essential role of Arg303 in placing GlcNAc into its binding pocket through extremely strong intermolecular interactions. Due to the uncharged and polar character of the His302 side chain imidazole ring, dipole–dipole interactions also exist between this residue and GlcNAc. |Einteraction| of the polar TUN–His302 pair is 76.42 kJ/mol; this pair has thus relatively strong intermolecular interactions. As a result, Arg303 and His302 are two crucial residues in the GlcNAc-binding pocket that ensure the GlcNAc binding to this pocket mainly via charge–dipole, dipole–dipole, unconventional and conventional H-bonding interactions. Even though the dipole moments on the GlcNAc hydroxyl groups create the induced dipoles in the nonpolar side chains of Ala188, Pro300, and Ile304, resulting in the occurrence of dipole–induced dipole interactions in each of these TUN–nonpolar residue pairs, the TUN–Ile304 pair with |Einteraction| of 23.28 kJ/mol has the strongest intermolecular interactions relative to the TUN–Ala188 and the TUN–Pro300 pairs (Table ). Finally, GlcNAc has the weakest intermolecular interactions with Cys299 because the lowest |Einteraction| (0.74 kJ/mol) is assessed in the TUN–Cys299 pair.

Comparison of the Stabilities of the Uracil, Tunicamine, Fatty Acyl Tail, and GlcNAc Binding Pockets in the QM Model I

The QTAIM outcomes exhibit that the total estimated |EHB| values at the BCPs detected (444.02 kJ/mol) in the uracil-binding pocket are greater than those (400.59 kJ/mol) in the tunicamine-binding pocket and are about 2.5 times those (182.53 kJ/mol) in the GlcNAc-binding pocket as well as being more than 32 times larger than those (13.87 kJ/mol) in the fatty acyl tail-binding pocket. Our NBO results present that the sums of the evaluated E(2) amounts relevant to the CT interactions identified in the uracil-, tunicamine-, fatty acyl tail-, and GlcNAc-binding pockets are 580.39, 497.52, 4.39, and 238.25 kJ/mol, respectively. The whole calculated |Einteraction| values related to the interacting fragment pairs (438.95 kJ/mol) in the uracil-binding pocket are approximately twice those (271.04 kJ/mol) in the GlcNAc-binding pocket and are more than 7 times greater than those (59.93 kJ/mol) in the fatty acyl tail-binding pocket, but these energies are smaller than those (657.69 kJ/mol) in the tunicamine-binding pocket. The |EHB|, E(2), and |Einteraction| amounts thus demonstrate that although the entire H-bonding and CT interactions arising from the uracil-binding pocket are much stronger than those deriving from each of the other three binding pockets, the whole intermolecular interactions appertaining to the total fragment pairs existing in the tunicamine-binding pocket are the strongest relative to those inside the other three pockets of this active site. We hence conclude that the tunicamine-binding pocket is the most stable pocket in the QM model I and the electrostatic interactions, particularly charge–dipole and dipole–dipole interactions in TUN–Glu56 and TUN–Asp252 pairs, have considerable contributions to the total intermolecular interactions of this pocket.

Tunicamycin-Binding Pocket in MraYCB–TUN (QM Model II)

The outcomes extracted from QTAIM and NBO analyses pertaining to the H-bonding and CT interactions identified in the QM model II are presented in Tables and 5, respectively. |Einteraction|, the BSSE energies, and dipole moments of different fragment pairs, which can be TUN–residue/Wt pairs or Wt–residue pairs, of this model are tabulated in Table . In this model, the TUN atomic numbering scheme is according to the labeling of its atoms in the MraYCB–TUN crystal structure[17] and the topology file built for TUN (see the Supporting Information). In the following, the results of these three tables are discussed in detail in four separate sections.

Table 4

Structural and Topological Parameters of ρBCP(r) Corresponding to the H-Bonds in the TUN–Residue/Wt Pairs and Wt–Residue Pairs of QM Model II Computed at the M06-2X/6-31G** Levela

proton donor	proton acceptor	bond path	d	∠	ρBCP	∇2ρBCP	HBCP	|EHB|
Lys97	TUN	Hε2···O12	2.74	120.45	0.0060	0.0226	–0.0027	5.32
Lys111	TUN	Hζ1···O10	2.31	124.25	0.0113	0.0360	–0.0085	11.79
Gly176	TUN	Hα2···O18	2.73	154.36	0.0055	0.0195	–0.0026	4.39
Gly176	TUN	Hα1···O39	2.59	123.80	0.0089	0.0304	–0.0050	7.53
Asp178	TUN	Hα···O28	2.51	150.24	0.0087	0.0281	–0.0055	7.84
Asn221	TUN	Hδ21···O28	2.60	152.71	0.0056	0.0215	–0.0032	4.78
Wt502	TUN	H2···O28	1.58	158.01	0.0558	0.1931	–0.0571	72.16
Wt503	TUN	H1···O31	2.46	128.82	0.0073	0.0264	–0.0040	7.20
Pro288	TUN	Hδ2···O43	2.37	155.50	0.0121	0.0385	–0.0087	11.94
His290	TUN	Hε1···O39	2.28	126.94	0.0131	0.0445	–0.0096	13.26
His291	TUN	Hε2···O41	1.76	164.24	0.0351	0.1224	–0.0276	39.04
Asp175	Wt501	Hα···O	2.58	146.71	0.0080	0.0275	–0.0045	7.82
Cys181	Wt501	H···O	1.91	173.83	0.0272	0.0860	–0.0220	30.19
Cys181	Wt501	Hβ2···O	2.66	131.53	0.0075	0.0264	–0.0039	6.28
Met229	Wt502	H···O	2.01	153.74	0.0256	0.0736	–0.0227	28.53
TUN	Asn172	H71···Oδ1	2.11	127.02	0.0175	0.0654	–0.0141	20.06
TUN	Phe173	H391···O	2.15	152.59	0.0139	0.0461	–0.0118	15.50
TUN	Asp175	H161···O	2.39	145.17	0.0076	0.0295	–0.0055	8.38
TUN	Asp175	H261···Oδ1	2.38	132.88	0.0115	0.0356	–0.0079	10.91
TUN	Wt502	H261···O	2.52	126.77	0.0095	0.0338	–0.0060	8.64
TUN	Asp178	H291···N	2.66	129.71	0.0070	0.0258	–0.0036	5.94
TUN	Asp178	H291···Oδ1	1.61	168.94	0.0516	0.1734	–0.0451	59.11
TUN	Asp231	H101···Oδ2	1.81	164.03	0.0350	0.1051	–0.0285	37.05
Wt501	Asp175	H1···Oδ1	1.55	179.80	0.0602	0.1901	–0.0624	76.17
Wt501	Leu177	H2···O	1.66	157.74	0.0430	0.1848	–0.0366	52.38
Wt502	Asp175	H1···Oδ1	2.02	155.22	0.0232	0.0651	–0.0203	24.58

The H-bond length (d) is in angstroms (Å), and the H-bond angle (∠) is in degrees (deg). All ρBCP, ∇2ρBCP, and HBCP parameters are in atomic units (au). |EHB| is in kJ/mol.

Table 5

NBO Results of Partner Orbitals That Participated in Donor–Acceptor Interactions in the TUN–Residue/Wt Pairs and Wt–Residue Pairs of QM Model II Assessed at the M06-2X/6-31G** Level

electron donor	electron acceptor	charge transfer	E(2) (kJ/mol)	qnB→σA–H*(e)
TUN	Lys111	nO10 → σNζ–Hζ1*	9.56	0.0033
TUN	Gly176	nO18 → σCα–Hα2*	3.26	0.0011
TUN	Gly176	nO39 → σCα–Hα1*	3.56	0.0015
TUN	Asp178	nO28 → σCα–Hα*	5.56	0.0016
TUN	Asn221	nO28 → σNδ2–Hδ21*	4.02	0.0012
TUN	Wt502	nO28 → σO–H2*	104.81	0.0402
TUN	Wt503	nO31 → σO–H1*	5.48	0.0005
TUN	Pro288	nO43 → σCδ−Hδ2*	11.63	0.0049
TUN	His290	nO39 → σCε1–Hε1*	5.84	0.0017
TUN	His291	nO41 → σNε2–Hε2*	80.58	0.0316
Wt501	Asp175	nO → σCα–Hα*	5.02	0.0024
Wt501	Cys181	nO → σN–H*	71.55	0.0206
Wt501	Cys181	nO → σCβ–Hβ2*	2.47	0.0008
Wt502	Met229	nO → σN–H*	50.63	0.0172
Asn172	TUN	nOδ1 → σN7–H71*	13.52	0.0038
Phe173	TUN	nO → σO39–H391*	13.43	0.0036
Asp175	TUN	nO → σO16–H161*	8.66	0.0025
Asp175	TUN	nOδ1 → σC26–H261*	6.59	0.0037
Wt502	TUN	nO → σC26–H261*	1.50	0.0005
Asp178	TUN	nN → σN29–H291*	6.38	0.0024
Asp178	TUN	nOδ1 → σN29–H291*	81.20	0.0314
Asp231	TUN	nOδ2 → σO10–H101*	60.60	0.0256
Asp175	Wt501	nOδ1 → σO–H1*	89.18	0.0241
Leu177	Wt501	nO → σO–H2*	86.53	0.0225
Asp175	Wt502	nOδ1 → σO–H1*	20.61	0.0085

Table 6

Modulus of Calculated BSSE-Corrected Interaction Energies, BSSE Energies, and Dipole Moments of the Different Fragment Pairs, Which Can Be TUN–Residue/Wt Pairs or Wt–Residue Pairs in the QM Model II Evaluated at the M06-2X/6-31G** Level

fragment pair	|Einteraction| (kJ/mol)	BSSE energy (kJ/mol)	dipole (debye)
TUN–Asp93	31.85	0.38	32.38
TUN–Lys97	41.49	5.02	24.79
TUN–Lys111	45.46	5.08	26.77
TUN–Asn172	29.55	13.51	12.57
TUN–Phe173	21.83	8.77	8.82
TUN–Asp175	100.94	23.11	20.83
TUN–Gly176	19.51	17.12	6.33
TUN–Asp178	77.83	15.17	42.55
TUN–Gly179	3.86	1.36	5.43
TUN–Asn221	3.16	2.97	7.45
TUN–Ala225	1.05	0.22	7.62
TUN–Phe228	13.58	7.03	5.44
TUN–Gly230	4.08	0.93	7.17
TUN–Asp231	169.41	19.21	19.69
TUN–Glu264	2.54	0.03	34.85
TUN–Val268	8.89	2.10	7.66
TUN–Ala287	9.85	4.09	6.62
TUN–Pro288	11.82	6.92	10.03
TUN–His290	5.85	5.27	8.39
TUN–His291	92.17	7.64	16.07
TUN–Glu300	4.86	0.05	43.40
TUN–Wt502	23.44	8.74	7.42
TUN–Wt503	6.53	2.28	3.27
Wt501–Asp175	78.43	14.97	11.22
Wt501–Leu177	8.54	8.37	4.35
Wt501–Cys181	38.64	5.63	6.27
Wt502–Asp175	67.38	11.84	14.51
Wt502–Met229	28.40	6.45	5.45

The uracil-binding pocket in MraYCB–TUN consists of Asp175, Leu177, Asp178, Gly179, Cys181, Asn221, Ala225, Phe228, Met229, Glu300, Wt501, Wt502, and Wt503 (Figure S2). The NBO analysis indicates that σN29–H291* of the uracil base accepts the charges of 0.0314 and 0.0024 e from the lone pair orbitals of nOδ1 and nN in Asp178, respectively. Furthermore, its nO28 donates a large qCT (0.0402 e) to σO–H2* in Wt502 and small qCT values to σCα–Hα* in Asp178 and σNδ2–Hδ21* in Asn221. The nO28 → σO–H2* interaction with an E(2) of 104.81 kJ/mol is responsible for the formation of the strong O–H2···O28 H-bond with a length of 1.58 Å, an angle of 158.01°, and an |EHB| of 72.16 kJ/mol in the TUN–Wt502 pair. The moderate H-bond of N29–H291···Oδ1 with an |EHB| of 59.11 kJ/mol between the uracil base and Asp178 arises from the nOδ1 → σN29–H291* interaction with an E(2) of 81.20 kJ/mol (Figure ). From Tables and 5, it is clear that these four interactions are the strongest H-bonding and donor–acceptor interactions of TUN in the QM model II because they have the highest values of ρBCP, ∇2ρBCP, |EHB|, and E(2) compared to the other interactions of TUN. The results of the nO28 → σNδ2–Hδ21*, nN → σN29–H291*, and nO28 → σCα–Hα* interactions are the occurrence of three weak H-bonds of Nδ2–Hδ21···O28, N29–H291···N, and Cα–Hα···O28 in the TUN–Asn221 and TUN–Asp178 pairs (Figure ).

Figure 10

TUN uracil base in the tunicamycin-binding pocket of MraYCB–TUN involved in the H-bonding interactions directly with Asp175, Asp178, Asn221, Wt502, and Wt503 as well as indirectly with Met229, Leu177, Cys181, and Wt501.

As displayed in Figure , H-bonds of the N–H···O type exist in the Wt501–Cys181 and Wt502–Met229 pairs and Asp175 forms O–H1···Oδ1 H-bonds with both Wt501 and Wt502. Wt501 also interacts with Leu177 via O–H2···O H-bond. Of these, O–H1···Oδ1 with an |EHB| of 76.17 kJ/mol, a length of 1.55 Å, and an angle of 179.80° in the Wt501–Asp175 pair is a strong H-bond, but the other four are treated as moderate H-bonds (Table ). Likewise, weak H-bonds of the C–H···O kind are observed in the TUN–Wt502, TUN–Asp175, Wt501–Asp175, and Wt501–Cys181 pairs and O–H1···O31 H-bond with |EHB| of 7.20 kJ/mol is found in the TUN–Wt503 pair. Accordingly, these interactions give rise to H-bonded networks of weak–moderate C26–H261···O···H–N and moderate–strong N–H···O–H2···O28 in TUN–Wt502–Met229, weak–moderate–strong-moderate C26–H261···O–H1···Oδ1···H1–O–H2···O in TUN–Wt502–Asp175–Wt501–Leu177, and weak–moderate–strong–moderate C26–H261···O–H1···Oδ1···H1–O···H–N and moderate–strong–moderate–strong N–H···O–H1···Oδ1···H1–O–H2···O28 in TUN–Wt502–Asp175–Wt501–Cy181 (Figure ). Therefore, Leu177, Cys181, Met229, and Wt501 indirectly affect the uracil conservation inside its binding pocket by forming H-bonds either with Wt502 or with Asp175. Similar to DPAGT1–TUN, the uracil-binding pocket in MraYCB–TUN is stabilized by the water-mediated interactions. In addition to H-bonds, the uracil base interacts with Asp175, Asp178, Gly179, Asn221, Ala225, Phe228, Glu300, Wt502, and Wt503 via charge–dipole, dipole–dipole, and dipole–induced dipole interactions. Of these, the highest |Einteraction| (100.94 kJ/mol) is assigned to the TUN–Asp175 pair (Table ). Consequently, Asp175 is one of the crucial residues in this pocket that enhances the uracil–Wt502 interaction strength by forming H-bonded networks with both Wt501 and Wt502.

Figure S2 displays the locations of residues Asp93, Lys97, Lys111, Asn172, Gly176, Gly230, and Asp231 around the tunicamine sugar moiety. The QTAIM analysis detects four BPs beginning from the tunicamine hydroxyls and terminating at the backbone and the side chains of Lys97, Lys111, Asp175, and Asp231 (Figures and 11). Its amide nitrogen is connected to Asn172 side chain by H71···Oδ1 BCP, and Hα2···O18 BCP is observed between Gly176 and the ribose sugar. Tunicamine thus participates in four conventional H-bonds of O–H···O and N–H···O types and two unconventional H-bonds of C–H···O kind with the aforementioned residues (Figure and Table ). Similarly, the NBO analysis confirms the appearance of the interactions of nO10 → σNζ–Hζ1*, nOδ1 → σN7–H71*, nO → σO16–H161*, nO18 → σCα–Hα2*, and nOδ2 → σO10–H101* in the TUN–Lys111, TUN–Asn172, TUN–Asp175, TUN–Gly176, and TUN–Asp231 pairs, respectively (Table ). In this binding pocket, the greatest amounts of |EHB| (37.05 kJ/mol) and E(2) (60.60 kJ/mol) belong to the O10–H101···Oδ2 and the nOδ2 → σO10–H101* interactions, respectively, in the TUN–Asp231 pair (Tables and 5). Furthermore, this pair has the largest |Einteraction| (169.41 kJ/mol) compared to the other TUN–residue pairs of this model (Table ). As a result, Asp231 is an indispensable residue in the MraYCB–TUN active site that provides the tunicamine binding to its pocket significantly via extremely strong charge–dipole and dipole–dipole interactions. Generally, because of the presence of negatively charged, positively charged, and polar residues in this pocket (Figure S2), charge–dipole and dipole–dipole interactions have the considerable contributions to the intermolecular interactions of the tunicamine moiety with residues Asp93, Lys97, Lys111, Asn172, and Asp231. For this reason, the electrostatic interactions play essential roles in placing the tunicamine moiety into its binding pocket.

Figure 11

Tunicamine moiety in the tunicamycin-binding pocket of MraYCB–TUN forming unconventional and conventional H-bonds with Lys97, Lys111, Asn172, Gly176, and Asp231.

As depicted in Figure S2, Phe173, Glu264, Val268, Ala287, Pro288, His290, and His291 are arranged around the GlcNAc moiety. The GlcNAc hydroxyl oxygen (O39) synchronously accepts the Hα1 of Gly176 and Hε1 of His290 and donates its hydrogen (H391) to Phe173 carbonyl oxygen. GlcNAc also forms Hδ2···O43 and Hε2···O41 BPs with Pro288 and His291, respectively (Figures and 12). The topological parameters of these BCPs are indicative of the presence of a moderate Nε2–Hε2···O41 H-bond between TUN and His291, a weak O39–H391···O H-bond between TUN and Phe173, and three weak H-bonds of C–H···O kind in TUN–Gly176, TUN–Pro288, and TUN–His290 pairs (Table ). As can be seen in Table , the nO39 of GlcNAc is an electron donor for both the σCα–Hα1* of Gly176 and the σCε1–Hε1* of His290, but its σO39–H391* is an electron acceptor for the nO of Phe173. Its nO43 transfers the charge of 0.0049 e into σCδ−Hδ2* of Pro288. Since these CT interactions have very low qCT and E(2) values, attractive interactions between the interacting local orbitals in each cited TUN–residue pair are extremely weak. Additionally, a CT interaction occurs between the nO41 of GlcNAc and the σNε2–Hε2* of His291 with a qCT of 0.0316 e and an E(2) of 80.58 kJ/mol. Indeed, Nε2–Hε2···O41 and nO41 → σNε2–Hε2* in the TUN–His291 pair are the strongest interactions of the GlcNAc moiety in the QM model II (Tables and 5). In line with these findings, the greatest estimated |Einteraction| (92.17 kJ/mol) in the GlcNAc-binding pocket is also assigned to this pair (Table ). Hence, His291 is another important residue in this model that provides the GlcNAc binding to its pocket mainly through the H-bonding and dipole–dipole interactions. Besides, Phe173, Glu264, Val268, Ala287, Pro288, and His290 help preserve and stabilize this pocket through weak conventional and unconventional H-bonds accompanying weak electrostatic interactions of dipole–induced dipole, charge–dipole, and dipole–dipole kinds.

Figure 12

Residues Phe173, Pro288, His290, and His291 providing GlcNAc binding to the tunicamycin-binding pocket of MraYCB–TUN via unconventional and conventional H-bonds.

Comparison of the Stabilities of the Uracil-, Tunicamine-, and GlcNAc-Binding Pockets in the QM Model II

Our results express that the whole assessed |EHB| values at the BCPs found (402.53 kJ/mol) in the uracil-binding pocket are more than 4.5 times larger than both those (86.99 kJ/mol) in the tunicamine-binding pocket and (87.27 kJ/mol) in the GlcNAc-binding pocket. The total estimated E(2) amounts appertaining to the CT interactions characterized in the uracil-, tunicamine-, and GlcNAc-binding pockets are 541.53, 95.60, and 115.04 kJ/mol, respectively. The total evaluated |Einteraction| values relevant to the interacting fragment pairs (456.64 kJ/mol) in the uracil-binding pocket are the highest relative to both those (341.35 kJ/mol) in the tunicamine-binding pocket and those (152.95 kJ/mol) in the GlcNAc-binding pocket. The |EHB|, E(2), and |Einteraction| values hence prove that the uracil-binding pocket is the most stable pocket in the QM model II. This pocket is maintained by the electrostatic, H-bonding, and water-mediated interactions.

A Comparison of the Stabilities of the TUN Binding to DPAGT1 with MraYCB

QTAIM and NBO analyses characterized various types of H-bonding and CT interactions affecting the preservation and the stabilization of the TUN binding to the active site of either DPAGT1 or MraYCB. It is worth noting that the whole H-bonds detected within both active sites by the QTAIM analysis are in accordance with the predicted H-bonds for these two active sites by the experimental findings.[1,2,17] In addition to the intermolecular interactions identified in the uracil-binding pocket, the electrostatic interaction of the π···π stacking type occurs between the π electron clouds of the uracil moiety and the aromatic ring of either Phe249 in DPAGT1[1] or Phe228 in MraYCB.[17] This interaction that plays an important role in stacking the TUN uracil base inside the pertinent active site is also observed between the uracil moiety of muraymycin D2 (MD2) and the aromatic ring of Phe262 in the MraYAA–MD2 complex.[25,85] Finally, in agreement with the prediction of our MD simulations, a comparison of the entireties of the |EHB|, E(2), and |Einteraction| amounts related to the TUN–residue/Wt pairs in each binding pocket of DPAGT1–TUN with those in the corresponding biding pocket of MraYCB–TUN indicates that the active site of DPAGT1–TUN is much more stable than that of MraYCB–TUN.

Comparison of QTAIM and NBO Results Obtained from QM/MM Calculations with Those Derived from QM/PCM Calculations

The results extracted from QTAIM/NBO analyses and the interaction energies pertaining to structural models I and II are tabulated in Tables S12–S17 of the Supporting Information. A comparison of QTAIM/NBO results related to QM models I and II with those relevant to the structural models I and II reveals that although the tunicamycin-binding pockets were optimized by two different alternative procedures, the types of H-bonding and CT interactions identified by QTAIM/NBO analyses from QM/MM calculations are almost identical to those from QM/PCM calculations and there was very little difference between the strength of each interaction in the pertinent QM model from the corresponding interaction in that structural model. Consequently, the H-bonding and CT interactions detected by QTAIM/NBO analyses from both alternative ways are virtually equivalent topologically and geometrically.

Conclusions

The DPAGT1–TUN and MraYCB–TUN complex structures embedded in the explicit solvent were separately optimized in the NVT ensemble under periodic boundary conditions by QM/MM calculations. The characteristics of intermolecular interactions of tunicamycin with the active site (the QM model) of each optimized structure were determined by QTAIM and NBO analyses at the M06-2X/6-31G** level by including the solvent effects using the PCM approach. Our results found multiple H-bonds and CT interactions in the strength ranging from weak to strong among the different TUN moieties, the neighboring residues, and the water molecules within each active site. In the uracil-binding pockets of both active sites, the uracil base participates in H-bonds and CT interactions with Asp45 (N3–H31···Oδ1 and nOδ1 → σN3–H31*) and Asn191 (Nδ2–Hδ22···O14 and nO14 → σNδ2–Hδ22*) in DPAGT1 and Asp178 (N29–H291···Oδ1 and nOδ1 → σN29–H291*) and Asn221 (Nδ2–Hδ21···O28 and nO28 → σNδ2–Hδ21*) in MraYCB. Moreover, these binding pockets are stabilized by cooperative H-bonded networks forming amid the uracil base, Asp45, Asn47, Asn191, Gly192, Glu194, Asn242, Glu376, Wt603, and Wt604 in DPAGT1 as well as among the uracil base, Asp175, Leu177, Cys181, Met229, Wt501, and Wt502 in MraYCB. Besides, the uracil moiety interacts with Phe249 in DPAGT1 and Phe228 in MraYCB through π···π stacking interactions.

In the tunicamine-binding pockets of both active sites, the tunicamine moiety is involved in H-bonds and CT interactions with Asp252 (O9–H91···Oδ2 and nOδ2 → σO9–H91*), Asn185 (N1–HN1···Oδ1 and nOδ1 → σN1–HN1*), and Lys125 (Nζ–Hζ1···O9 and nO9 → σNζ–Hζ1*) in DPAGT1 and Asp231 (O10–H101···Oδ2 and nOδ2 → σO10–H101*), Asn172 (N7–H71···Oδ1 and nOδ1 → σN7–H71*), and Lys111 (Nζ–Hζ1···O10 and nO10 → σNζ–Hζ1*) in MraYCB. The TUN–Asp252 and TUN–Asp231 pairs are the most stable TUN–residue pairs within the active sites of DPAGT1 and MraYCB that play significant roles in conserving the tunicamine-binding pocket of that active site predominantly through considerably strong charge–dipole and dipole–dipole interactions. In DPAGT1–TUN, the tunicamine-binding pocket is sustained by the water-mediated interactions appearing among tunicamine, Asp115, Asn119, Asn185, Wt601, Wt606, and Wt609. The tunicamine moiety also forms additional H-bonds with Gln44, Glu56, and Arg301, neither of which is seen in the MraYCB–TUN.

In the GlcNAc-binding pockets of both active sites, the GlcNAc moiety participates in H-bonds and CT interactions with His302 (Nδ1–Hδ1···O5 and nO5 → σNδ1–Hδ1*) in DPAGT1 and His291 (Nε2–Hε2···O41 and nO41 → σNε2–Hε2*) in MraYCB. Although four H-bonds exist between GlcNAc and Arg303 in DPAGT1, there are no H-bonds between this moiety and the corresponding residue (Arg282) in MraYCB. The fatty acyl tail-binding pocket in DPAGT1–TUN is retained by H-bonds, hydrophobic and dipole–induced dipole interactions arising from the nonpolar side chains of residues Trp122, Leu126, Leu175, Val178, Phe179, Ile186, Phe286, and Leu293 with the TUN fatty acyl tail.

The |EHB|, E(2), and |Einteraction| values conclude that the tunicamine-binding pocket and the uracil-binding pocket are the most stable pockets in the active sites of DPAGT1–TUN and MraYCB–TUN, respectively. Finally, both the MD simulations and the QM calculations affirmed that the tunicamycin binding to DPAGT1 is much stronger than that to MraYCB in support of in vivo and in vitro biochemical observations.

To this end, our analysis should have provided intrinsic physical and chemical characteristics for the naturally occurring nucleoside antibiotic tunicamycin in its own right, while interacting with its native binding partner MraYCB or DPAGT1. This information is critical in terms of better understanding its mode of action at molecular level, which, in the meantime, lays a solid ground for developing new generations of nucleoside antibiotics in a manner of being more effective and selective than tunicamycin but free of any unwanted side effects for deployment in clinical settings.