Molecular Basis of the Recognition Process: Hydrogen-Bonding

De Swarte, J.; De Vos, S.; Langhorst, U.; Steyaert, J.; Loris, R. Eur. J. Biochem. 2001, 268, 3993−4000. [Crossref], [PubMed], [CAS]. (9) . The cont...
0 downloads 0 Views 373KB Size
Download PDF
13590

J. Phys. Chem. B 2006, 110, 13590-13596

Molecular Basis of the Recognition Process: Hydrogen-Bonding Patterns in the Guanine Primary Recognition Site of Ribonuclease T1 Jiande Gu,*,†,‡ Jing Wang,‡ and Jerzy Leszczynski*,‡ Drug Design & DiscoVery Center, State Key Laboratory of Drug Research, Shanghai Institute of Materia Medica, Shanghai Institutes for Biological Sciences, Chinese Academy of Sciences, Shanghai 203201, People’s Republic of China, and Computational Center for Molecular Structure and Interactions, Department of Chemistry, Jackson State UniVersity, Jackson, Mississippi 39217 ReceiVed: March 4, 2006

Investigation of the intrinsic H-bonding pattern of the guanine complex with a sizable segment (from Asn43 to Glu46) of the primary recognition site (PRS) in RNase T1 at the B3LYP/6-311G(d,p) level of theory enables the electronic density characteristics of the H-bonding patterns of the guanine-PRS complexes to be identified. The perfect H-bonding pattern in the guanine recognition site is achieved through the guanine complex interactions with the large segment of the PRS. Two significant short H-bonds, O1‚‚‚HN1 and O2‚‚‚HN2, have been identified. The similar short H-bond distances found in the anionic GC- base pair and in this study suggest that the short hydrogen-bond distances may be characteristic of the multiple H-bonded anionic nucleobases. The H-bonding energy distribution, the geometric analysis of the H-bonding pattern, and the electron structure characteristics of the H-bonds in the guanine PRS of RNase T1 all suggest that the O1‚‚‚HN1 and O2‚‚‚HN2 side-chain H-bonds dominate the binding at the guanine recognition site of RNase T1. Also, the geometry evidence, the electron structure characteristics, and the properties of the bond critical points of the H-bonds reveal that the side-chain H-bonding and the main-chain H-bonding are mutually intensifying. Thus the positive cooperativity between Asn43 to Tyr45 and Glu46 is proposed.

Introduction Ribonuclease T1 (RNase T1) is a well-characterized enzyme that catalyzes the hydrolysis of single-stranded RNA at guanylyl residues.1,2 Experimental studies of protein-guanine interactions in RNase T1 revealed its pronounced specificity for the nucleobase guanine. The guanine binding site, referred to as the primary recognition site (PRS), demonstrates the greatest affinity in substrate binding.3-7 Multiple H-bonding and aromatic stacking have been found to be the driving force in the molecular recognition process in RNase T1-nucleotides complexes. The crystal structures of RNase T1 complexes with nucleotides or substrate analogues indicate that the guanine moiety is involved in extensive hydrogen bonding with three main-chain amide N-H donors and three oxygen acceptors from the binding site of RNase T1.8-11 While the N-H donor from Asn43 interacts with the N7 acceptor of guanine, the amide N-H donors from Asn44 and Asn45 are bonded to the O6 atom of guanine, forming a bifurcated H-bond. The anionic side-chain carboxylate (O1 and O2) of Glu46 and the main-chain amide CdO of Asn98 are H-bonded to the HN1, HN2, and H′N2 of guanine, respectively. In addition, guanine is located between the phenolic groups of Tyr42 and Tyr45. One of the remarkable aspects of the guanine-bonded site of RNase T1 is that the residues which are H-bonded to the guanine moiety form a continuous fragment (from Asn43 through Glu46), with the exception of Asn98. Guanine-binding site-directed mutagenesis studies of residues 42 through 46 in the PRS have revealed that substitutions at all * Corresponding authors. E-mail: J.G.: [email protected]; J.L.: [email protected]. † Chinese Academy of Sciences. ‡ Jackson State University.

of these locations were either neutral or deleterious to the enzyme function.3,5,6,12,14 All of these findings suggest that the PRS may have achieved its optimal structure for guanine recognition. The free energy change through mutation of Tyr45 (Tyr45Ala) of the guanine-RNase T1 complex suggests that the stacking interaction between the phenolic side chain of Tyr45 and the aromatic ring of guanine is less important compared with the multiple H-bonding between the PRS and the guanine moiety. The influence of stacking between Tyr42 and guanine is also expected to be less important than the H-bonding in the guanine-PRS system since a recent comprehensive study of the H-bonding and stacking energy of guanine and cytosine concluded that the stacking energy for the GC pair is considerably lower than the corresponding H-bonding energy.15 Therefore, the detailed characterization of the H-bonding patterns in the guanine-PRS complexes is a key for understanding the primary role of the PRS in the molecular recognition. It is also crucial for the development of the designed nucleobase recognition sites, which are relevant to biological systems. However, due to the influence of the surroundings, the characteristics of the H-bonds of the PRS-guanine interactions from the crystallographic data deposited in the RCSB Protein Data Bank vary considerably. The H-bond distance between the O2 of Glu46 and the N2 of guanine varies from 2.65 (PDB li0v9) to 2.98 Å (PDB li0x16). Therefore, knowledge of the intrinsic properties of the H-bonding of such systems may be secured only through theoretical investigations. Theoretical studies have been performed at different levels to explore the molecular recognition between nucleobases and amino acids.8,17-19 Interaction energies of 21 pairs of H-bonded amino acid side chains with nucleic acid bases were calculated at the LMP2/6-31G(d,p)//HF/6-31G(d,p) level of theory to

10.1021/jp061360x CCC: $33.50 © 2006 American Chemical Society Published on Web 06/20/2006

Molecular Basis of the Recognition Process

J. Phys. Chem. B, Vol. 110, No. 27, 2006 13591

Figure 2. Optimized structure of the guanine-PRS complex model (G-PRS). The side chains of Asn43, Asn44, and Tyr45 and the carbonyl group of Glu46 are simplified by hydrogen atoms. The carbonyl group is added to N43 to mimic the main-chain amide of Asn43. H-Bond distances are in angstroms. Figure 1. Structure of the primary recognition site (PRS) of the RNase T1 2′-GMP complex. Color representation: gray, carbon; blue, nitrogen; white, hydrogen; red, oxygen; yellow, phosphorus. The guanine complex with a continuous segment (from Asn43 to Glu46) of the PRS in tube-frame represents the model adopted in this research.

explore the intrinsic energies of discrete sequence-specific interactions.17 The intermolecular interactions between positively charged residues and adenine were studied at the MP2 level of theory with different basis sets.18 Recently, a density functional theory (DFT) approach has been applied to investigate the individual interactions between guanine and the residues of the PRS of RNase T1.19 However, the studies of interactions between the PRS and guanine based on the crystallographic data are not able to reveal the intrinsic properties of the H-bonding pattern of guanine-PRS interactions because, as mentioned above, the parameters of the H-bonds differ for different crystal structure studies. To acquire the description of the intrinsic characteristics, the structure of the guanine-PRS complex has to be fully optimized and investigated at a reliable level of theory. In this paper, we report investigations of the intrinsic H-bonding pattern of the guanine complex with a sizable segment (from Asn43 to Glu46) of the PRS in RNase T1 (Figure 1). The structure of the guanine-PRS complexes have been fully optimized by the DFT approach with a relatively large (triple-ζ) basis set. We constructed various molecular models that differ in complexity in order to reveal the details of influence of molecular parameters and the formed H-bonds on the structure and properties of the complex. Our attention is focused on the electronic density characteristics of the Hbonding patterns of the guanine-PRS complexes. Specifically, the H-bonding pattern of the main-chain H-bonds and the anionic side-chain H-bonds are studied by a combined approach that utilizes electron density deformation analysis and the atoms in molecule (AIM) theory.20,21 These electron density characteristics provide deeper insight into the cooperative effects in the multiple H-bonding interactions. Also, the total interaction energies of the guanine-PRS complexes are appropriately decomposed into the individual components according to the Kitaura-Morokuma scheme22,23 to characterize the contributions of the electrostatics, exchange repulsion, polarization, and other contributions of the intermolecular interactions. These physical properties are of vital importance for the prediction of molecular recognition and for the development of the newly designed nucleobase recognition sites.

Method of Calculation. The local minima of the guaninePRS complexes have been fully optimized by analytic gradient techniques. The density functional theory (DFT) with Becke’s three-parameter (B3)24 exchange functional along with the LeeYang-Parr (LYP) nonlocal correlation functional (B3LYP)25,26 was applied in this study. The standard valence triple-ζ basis set augmented with d-type and p-type polarization functions, 6-311G(d,p),27 was used for all of the elements. It is well-known that the geometries and frequencies of the molecules calculated at the B3LYP/6-311G(d,p) level agree well with experiment.28 Our previous studies of the hydrogen-bonded systems involving NA bases have shown that the B3LYP approach predicts reliable interaction energies and is compatible with the MP2/6-31(d,p) method.29,30 The Gaussian 98 package of programs31 was used in the geometry optimization and the energy calculations. To analyze the H-bonding pattern in the guanine binding site, the atoms-in-molecules (AIM) theory of Bader20,21 was applied, and the analysis was carried out by the AIM2000 program (version 2.0).32 The AIM analysis is based on the density obtained at the B3LYP/6-311G(d,p) level. The Morokuma energy decomposition analysis was carried out by the GAMESS program33 employing the 6-311G(d,p) basis set. Results and Discussion Geometry. The guanine-PRS complex model (G-PRS) consists of the peptide main chain from Asn43 to Glu46, the side chain of Glu46 with an anionic carboxylate group, and the nucleic acid base guanine. The side chains of Asn43, Asn44, and Tyr45 and the carbonyl group of Glu46 are terminated by hydrogen atoms. To complete the main-chain amide of Asn43, the carbonyl group of Tyr42 is added. The structure of the fully optimized G-PRS is depicted in Figure 2 along with the geometric parameters of the H-bonds. A comparison with crystallographic parameters of these H-bonds enables the reliability of the model selection to be justified. Table 1 compares the H-bond distances predicted from the theoretical model with the experimental measurements. It is remarkable that the theoretical H-bond parameters in the gas phase are close to those of the crystallographic data of the structure in PDB li0v9 with the exception of R(N44-O6), which is about 0.24 Å too long, and R(N43-N7), which is about 0.17 Å too short. Considering that the influence of the environmental surroundings in the crystal structure is significantly different from that in the gas phase, the general

13592 J. Phys. Chem. B, Vol. 110, No. 27, 2006

Gu et al.

TABLE 1: Comparison of Theoretical H-Bond Distances with Crystallographic Dataa

R(O2-N2) R(O1-N1) R(N45-O6) R(N44-O6) R(N43-N7) R(O1-N46) a

GPRSb

PDB 1i0v9

PDB 1i0x16

PDB 1bvi33

PDB 1loy8

PDB 1rnt10

2.689 2.737 2.958 2.904 3.030 2.862

2.65 2.74 2.86 2.66 3.20 2.82

2.98 2.95 2.85 2.70 2.89 2.82

2.91 2.85 2.87 2.83 2.99 2.81

2.78 2.85 2.75 2.62 3.04 2.75

2.96 2.73 2.82 2.78 3.33 2.79

H-Bond distances are given in angstroms. b Calculated values.

consistency between the theoretical and experimental results justifies the simplifications in the model selection. Two remarkable short H-bonds have been predicted. The bond distances of O1‚‚‚HN1 and O2‚‚‚HN2 have been found to be 1.680 and 1.636 Å, respectively. The anionic side-chain carboxylate of Glu46 seems to tightly bond to guanine. The amide N-H donors from Asn44 and Asn45 are bonded to the O6 atom of guanine, forming the oxygen-bifurcated H-bond with a N44H‚‚‚O6 bond length of 2.008 Å and a N45H‚‚‚O6 bond length of 1.952 Å. The amide N-H of Asn43 interacts with the proton acceptor N7 of guanine. The N43H‚‚‚N7 H-bond distance amounts to 2.055 Å. In addition to the five intermolecular H-bonds between guanine and the PRS of RNase T1, there is also an intramolecular H-bond found in the Glu46 residue. The theoretically predicted O1‚‚‚HN46 bond distance amounts to 1.970 Å. The O1 atom of Glu46 in the complex forms a bifurcated H-bond. This intramolecular O1‚‚‚HN46 bond is expected to reduce the H-bonding between O1 and HN1, resulting in the longer bond distance of O1‚‚‚HN1 (1.680 Å) compared to that of O2‚ ‚‚HN2 (1.636 Å). The remarkable discovery in the recent investigations35 of the anionic GC- base pair reveals the significant change in the H-bond distance of O2(C)‚‚‚HN2(G) with respect to that of neutral GC (1.645 vs 1.884 Å). The similar short H-bond distances between the anionic carboxylate of Glu46 and guanine suggests that this short hydrogen-bond distance may be an important characteristic of the multiple H-bonded anionic nucleobases. Considering that the H‚‚‚O distance in the water dimer is about 1.94 Å,36 the H‚‚‚O distance in the anionic G-PRS complex is so short that one could consider it to be a possible partial chemical bond rather than a conventional H-bond. The tight H-bonding between the anionic side-chain carboxylate of Glu46 and guanine is expected to influence the mainchain H-bonding. And inversely, different main-chain bonding will affect the characteristics of the side-chain H-bond. To examine how the length of the main-chain bonds (the Asn43, Asn44, and Asn45 components) affects the H-bonding between Glu46 and guanine, three other models (G-PRS1, G-PRS2, and G-PRS3 in Figure 3) were investigated at the same level of theory. In G-PRS1, the Asn43 residue of G-PRS is replaced by a carbonyl group; in G-PRS2, both Asn43 and Asn44 are substituted by a carbonyl group; and finally, residues from Asn43 to Tyr45 are eliminated in G-PRS3. The optimized structures of these three models are depicted in Figure 3. One important trend has been revealed by examining the H-bond distances of the G-PRSn models: the bond distance of O2‚‚‚HN2 increases monotonically as the number of mainchain H-bond decreases. The final H-bond length of O2‚‚‚HN2 is 1.749 Å in G-PRS3, 0.113 Å longer than that in G-PRS. The opposite trend can be deduced for the intramolecular H-bond O1‚‚‚HN46, where this bond distance decreases evenly

Figure 3. Structures of the optimized models. H-Bond distances are in angstroms.

TABLE 2: Predicted H-Bond Distances in Different Modelsa R(O2-N2) R(O1-N1) R(N45-O6) R(N44-O6) R(N43-N7) R(O1-N46)

G-PRS3

G-PRS2

G-PRS1

G-PRS

2.789 2.783

2.720 2.807 2.876

2.700 2.773 2.913 2.828

2.769

2.801

2.867

2.689 2.737 2.958 2.904 3.030 2.862

a

G-PRS is the optimized structure of the guanine-PRS complex modelG-PRS. The side chains of Asn43, Asn44, and Tyr45 and the carbonyl group of Glu46 are simplified by hydrogen atoms. The carbonyl group is added to N43 to mimic the main-chain amide of Asn43. H-Bond distances are given in angstroms.

from G-PRS (1.970 Å) to G-PRS3 (1.811 Å). The increasing bond length of O1‚‚‚HN1 reaches a maximum in G-PRS2 (1.763 Å). The formation of the N45H‚‚‚O6 H-bond weakens the O1‚‚‚HN1 bond. However, the formation of the other two H-bonds, N44H‚‚‚O6 and N43H‚‚‚N7, intensifies the O1‚‚‚ HN1 bonding interaction. On the other hand, the N44H‚‚‚O6 and N43H...N7 H-bonds elongate the N45H‚‚‚O6 bond distance by 0.09 Å. The H-bond variations in the different models are summarized in Table 2. To explore the influence of the formation of side-chain H-bonds on main-chain H-bonds (N43H‚‚‚N7, N44H‚‚‚O6, and N45H‚‚‚O6), the model that lacks the Glu46 moiety and the carbonyl group of Tyr45 (model G-PRS4) has been studied. However, at the B3LYP/6-311G(d, p) level of theory, this model with three H-bonds (Figure 4) does not represent a local minimum on the potential energy surface; rather, it is a saddle point. The long bond distances of N44H‚‚‚O6 and N45H‚‚‚O6 (2.533 and 2.247Å) suggest that without the presence of the side-chain H-bonds O1‚‚‚HN1 and O2‚‚‚HN2, these mainchain H-bonds are very weak. Therefore, the geometric features of H-bonds in guanine recognition sites seem to imply that the side-chain H-bonds O1‚‚‚HN1 and O2‚‚‚HN2 dominate the binding at the PRS and that the side-chain H-bonding and the

Molecular Basis of the Recognition Process

J. Phys. Chem. B, Vol. 110, No. 27, 2006 13593 TABLE 4: NPA Charge Distribution between Guanine and the PRS Moiety in Different Modelsa PRSn guanine a

Figure 4. Structure of the model G-PRS4 with three H-bonds. This structure has been located as the saddle point on the potential energy surface.

TABLE 3: H-Bonding Energies of the G-PRS Models and the Components of the Kitaura-Morokuma Analysisa ∆EDFT b ∆Ec ESc EXc PLc CTc MIXc

G-PRS3

G-PRS2

G-PRS1

G-PRS

-40.17 (-9.64) -37.56 (-8.34) -45.43 (-14.16) 30.06 (12.25) -10.76 (-2.56) -10.48 (-3.80) -0.95 (-0.07)

-49.81 (-10.96) -45.90 (-10.83) -59.59 (-13.51) 42.31 (10.17) -13.32 (-4.19) -14.28 (-3.28) -1.02 (-0.02)

-60.77 (-9.36) -56.73 (-8.17) -73.10 (-10.12) 52.48 (7.59) -17.51 (-3.17) -17.56 (-2.51) -1.04 (0.04)

-70.13 -64.90 -83.22 60.07 -20.68 -20.07 -1.00

a Analyzed at the HF/6-311G(d, p) level with the geometry optimized at the B3LYP/6-311G(d,p) level of theory unless noted otherwise. Values in parentheses are the energy changes due to the formation of an additional H-bond in the model listed in the next right column. Energies are given in kilocalories per mole. b Evaluated at the B3LYP/ 6-311G(d,p) level of theory. c H-Bonding energy (∆E) is decomposed into the electrostatic energy (ES), the exchange energy (EX), the polarization energy (PL), the charge transfer (CT), and the mixed term (MIX).22,23

main-chain H-bonding do strengthen each other. Thus the positive cooperativity between Asn43 to Tyr45 and Glu46 is revealed. Hydrogen Bonding Energies. Without consideration of the relaxation energy of the separated species, the H-bonding energy of the side-chain H-bonds O1‚‚‚HN1 and O2‚‚‚HN2 is evaluated to be 40.2 kcal/mol (Table 3) at the B3LYP/6-311G(d, p) level of theory. The interaction energies of the three other H-bonds are 9.64 kcal/mol for the N45H‚‚‚O6 bond, 11.0 kcal/ mol for the N44H‚‚‚O6 bond, and 9.4 kcal/mol for the N43H‚ ‚‚N7 bond. It should be noted that this distribution does not include the cooperative effects between the Asn43 to Tyr45 segment and the Glu46 residue as suggested on the basis of the geometric changes in different G-PRSn models. The cooperativity between the H-bonds should increase the binding energies of O1‚‚‚HN1 and O2‚‚‚HN2 in the G-PRS model. It is important that the H-bonding energy distribution is consistent with the conclusion from the geometric analysis above that the side-chain H-bonds O1‚‚‚HN1 and O2‚‚‚HN2 dominate the binding at the guanine recognition site of RNase T1. The Kitaura-Morokuma energy decomposition analysis was carried out to reveal the deeper physical background of the H-bonding interactions. Among the energy components, the electrostatic energy (ES) term represents the classical Coulombic force between the charge distributions of the interacting parts. The large negative values of the ES term in G-PRS3 to G-PRS (-45.43 to -83.2 kcal/mol) demonstrate that the electrostatic energy is decisive in the guanine recognition site. It is interesting to note that the increase of the polarization term (PL) is more

G-PRS3

G-PRS2

G-PRS1

G-PRS

-0.879 -0.121

-0.891 -0.109

-0.898 -0.102

-0.898 -0.102

Charge distributions are given in atomic units (au).

significant from G-PRS2 to G-PRS1 (-4.2 kcal/mol) as compared with that from G-PRS3 to G-PRS2 (-2.6 kcal/ mol) and from G-PRS1 to G-PRS (-3.2 kcal/mol). Notice that the main-chain H-bonds in G-PRS1 are the O6-bifurcated H-bonds. The polarization alteration seems to suggest that the bifurcated H-bonds are more polarizable. The natural population atomic (NPA) charges were determined by use of the B3LYP functional and the triple-ζ 6-311G(d, p) basis set with the natural bond order (NBO) analysis of Reed and Weinhold and co-workers.37-40 The charge distribution between guanine and PRS summarized in Table 4 reveals that the charge transfer from PRS to guanine occurs mainly through the O1‚‚‚HN1 and O2‚‚‚HN2 side-chain H-bonds. As the size of the PRS model increases, the negative charge decreases on guanine, and consequently, the rate of the energy increase due to the charge transfer (CT) and the electrostatic interaction (ES) slows down as the number of H-bonds accumulates (as shown in Table 3). It is worthwhile to note that the charge distribution on guanine reaches -0.102 au in both G-PRS1 and G-PRS models. The total charge value on guanine is larger than that found in the anionic GC- base pair (-0.08 au35). When the negative electron affinity of guanine is considered,41 it is reasonable to expect that the PRS modifications that decrease the charge distribution on guanine might improve the guanine recognition specificity in RNase T1. On the other hand, due to the high negative charge on the carbonyl group of Asn98, the formation of the H-bond between the oxygen of the carbonyl group of Asn98 and the H′N2 fragment of guanine might increase the negative charge on guanine and thus reduce the existing H-bonding in the guanine complex with the continuous segment (from Asn43 to Glu46) of the PRS in RNase T1. Electron Structure of Hydrogen Bonding in G-PRS. H-Bonding can be characterized by the change of electron density for the bonded moiety. The electron density around the proton and the proton acceptor decreases while the density between the proton and its acceptor increases as the result of the formation of a H-bond.42 The deformation electron density maps of the side-chain H-bonds O1‚‚‚HN1 and O2‚‚‚HN2 in the guanine binding site have been plotted in Figure 5. The electronic structures of the O1‚‚‚HN1 and O2‚‚‚HN2 H-bonds are clearly noticeable in the deformation density map. The concentration of electron density accumulates between the proton (HN1 and HN2) and the proton acceptor (O1 and O2) with a corresponding electron density deficiency at the positions of a proton and the acceptor lone pair. The electron density accumulation along the shorter H-bond (O2‚‚‚HN2) is larger compared to that of the longer H-bond (O1‚‚‚HN1) in all G-PRSn models. The electron structure of the O2‚‚‚HN2 H-bond in Figure 5 indicates that the electron density between the carboxylate O2 and the amino proton HN2 increases as the result of formation of the main-chain H-bonds. The density peak value in the electron structure of the O2‚‚‚HN2 H-bond is about 0.003 au in G-PRS3. This value increases to 0.004 au in G-PRS1 and 0.005 au in G-PRS. The electron structure of O1‚‚‚HN1 displays a similar trend as shown in Figure 5. The

13594 J. Phys. Chem. B, Vol. 110, No. 27, 2006

Gu et al.

Figure 5. Deformation electron density maps of the side-chain H-bonds O1‚‚‚HN1 and O2‚‚‚HN2 in G-PRSn. ∆F ) F[G-PRSn] - F[G] F[PRSn]. Contours in the deformation density map are shown at (0.001 au. Dashed pink indicates deficiency of density, and solid blue indicates increasing density. The plan is formed by the HN1, HN2, and C (carboxylate) atoms.

Figure 6. Deformation electron density maps of the O6-bifurcated H-bond (N44H‚‚‚O6 and N45H‚‚‚O6) in G-PRSn. ∆F ) F[G-PRSn] - F[G] - F[PRSn]. Contours in the deformation density map are shown at (0.001 au. Dashed pink indicates deficiency of density, and solid blue indicates increasing density. The plan is formed by the O6, HN44, and HN45 atoms.

trend of electron density variation of the H-bonds in different models corresponds to the alteration in the geometry of the H-bonds. The electronic structure of the O6-bifurcated H-bond (N44H‚ ‚‚O6 and N45H‚‚‚O6) in G-PRS1 reveals that the N45H‚‚‚O6 bonding is weaker than that of the N44H‚‚‚O6 fragment (Figure

6). However, this order reverses due to the formation of the N43H‚‚‚N7 bond in G-PRS. The density maximum between O6 and HN44 moves away from the bond path of N44H‚‚‚O6 due to the existence of the neighboring N43H‚‚‚N7 bonding. Introduction of the N43H‚‚‚N7 bond in G-PRS thus reduces the existing N43H‚‚‚N7 bonding.

Molecular Basis of the Recognition Process

J. Phys. Chem. B, Vol. 110, No. 27, 2006 13595 in G-PRS4 (0.0128, 0.0074, and 0.0214 au) illustrate the importance of the existence of H-bonds of anionic side-chain carboxylate of Glu46 and the amino donors of guanine. Therefore, the conclusions derived from the geometric features of H-bonds in the guanine recognition site that the side-chain H-bonds O1‚‚‚HN1 and O2‚‚‚HN2 dominate the binding at the PRS and that the side-chain H-bonding and the main-chain H-bonding are mutually intensifying are also supported both by the electron structure characteristics and by the AIM analysis. Finally, the intramolecular H-bond in Glu46 (O1‚‚‚HN46) is weakened by the formation of the O6-bifurcated H-bonds and the N43H‚‚‚N7 H-bonds, as suggested by the steady decreasing of the corresponding density at BCP6 from G-PRS3 to G-PRS. This effect is expected to contribute to intensifying the O1‚‚‚HN1 intermolecular H-bond. Concluding Remarks

Figure 7. Definition and position of the bond critical points (BCPs) in the guanine recognition site.

TABLE 5: Density at the BCPs in Different Modelsa G-PRS4 BCP1 BCP2 BCP3 BCP4 BCP5 BCP6 a

G-PRS3

G-PRS2

G-PRS1

G-PRS

0.0421 0.0418

0.0505 0.0403 0.0289

0.0529 0.0450 0.0250 0.0286

0.0361

0.0303

0.0265

0.0551 0.0497 0.0232 0.0213 0.0237 0.0256

0.0128 0.0074 0.0214

Densities are given in atomic units (au).

As demonstrated in the electron deformation map, the O6bifurcated H-bonding in G-PRS4 is very weak. There is basically no density increase detected between the N44H and O6 atoms. Therefore, the existence of the O2‚‚‚HN2 and O1‚‚‚HN1 H-bonds in the guanine recognition site is necessary for the effective formation of the main-chain H-bonds, which is consistent with the conclusion based on the geometry analysis. AIM theory has been proven to be a useful and successful tool in the interpretation of charge density toward a wide variety of chemical concepts.43,44 The density at the bond critical point (BCP) is of paramount importance in AIM. It has been used to characterize different types of chemical bonds. In this study the AIM calculations were performed at the B3LYP/6-311G(d, p) level of theory. The positions of the BCPs of the H-bonds in G-PRS have been illustrated in Figure 7, and the charge densities of the BCPs have been summarized in Table 5. The density at the BCP1 increases monotonically as the number of main-chain H-bonds increases. With the exception of G-PRS2, the density at the BCP2 raises its value with the accumulation of the main-chain H-bonds. These changes are also consistent with the alterations in the electronic structure of the corresponding H-bonds. Therefore, the cooperative effects of main-chain H-bonds on the side-chain H-bonds formed by the anionic side-chain carboxylate (O1 and O2) of Glu46 and the amino donors (N1 and N2) of guanine are confirmed. The density decrease at BCP3 and BCP4 with the formation of the N43H‚‚‚N7 bond in G-PRS corresponds to the elongations of the corresponding N45H‚‚‚O6 and N44H‚‚‚O6 bond distances in G-PRS as compared to those in G-PRS1 and G-PRS2. Along with the electron structure alteration of the O6-bifurcated H-bonds in G-PRS1 and G-PRS, one can conclude that the existence of the N43H‚‚‚N7 bond weakens the O6-bifurcated H-bonds. The significant large values at BCP3, BCP4, and BCP5 (0.0232, 0.0213, and 0.0237 au) in G-PRS compared to those

The investigation of the intrinsic H-bonding pattern of the guanine complex with a continuous segment (from Asn43 to Glu46) of the PRS in RNase T1 enables the electronic density characteristics of the H-bonding patterns of the guanine-PRS complexes to be identified. These electron density characteristics provide deeper insight concerning the cooperative effects in the multiple H-bonding interactions in the nucleobase recognition sites. The consistency between the theoretical H-bond parameters in the gas phase and those of the crystallographic data of PDB li0v9 demonstrates that the H-bonding pattern of perfect guanine recognition at the PRS site of RNase T1 is achievable under select crystallization conditions. Two significant short H-bonds, O1‚‚‚HN1 and O2‚‚‚HN2, have been identified. The similar short H-bond distances found in the anionic GC- base pair34 and in this study suggest that short hydrogen-bond distances may be characteristic of the multiple H-bonded anionic nucleobases. The Kitaura-Morokuma energy decomposition analysis reveals that the electrostatic energy is a decisive factor in guanine recognition by RNase T1. The H-bonding energy distribution, the geometric analysis of the H-bonding pattern, and the characteristics of electron structure of the H-bonds in the guanine PRS of RNase T1 all suggest that the side-chain H-bonds O1‚‚‚HN1 and O2‚‚‚HN2 dominate the binding at the guanine recognition site of RNase T1. The geometric evidence, the characteristics of electronic structure, and the properties of the bond critical points of the H-bonds reveal that the side-chain H-bonding and the mainchain H-bonding are mutually reinforced. This mutual reinforcement suggests positive cooperativity between the bonding of Asn43 to the Tyr45 fragment and Glu46. The perfect H-bonding pattern in the guanine recognition site seems to be achieved through the guanine complex with a continuous segment (from Asn43 to Glu46) of the PRS in RNase T1. Further H-bonding interaction between the H′N2 of guanine and the Asn98 residue of RNase T1 might diminish the perfectness of the recognition binding due to negative cooperative effects. It should be interesting in future studies to determine if this perfect H-bonding pattern in the guanine recognition site can be improved through the stacking interaction between guanine and the phenolic groups of Tyr42 and Tyr45. These properties are of vital importance for the prediction of molecular recognition and development of the newly designed nucleobase recognition sites. Acknowledgment. This research project in the People’s Republic of China was supported by the “Knowledge Innovation

13596 J. Phys. Chem. B, Vol. 110, No. 27, 2006 Program”, Chinese Academy of Sciences. In the United States the project was supported by NIH Grant G1 2RR13459-21 and NSF CREST Grant HRD-0318519. References and Notes (1) Takahashi, K.; Moore, S. In The Enzymes, Academic Press Inc.: New York, 1982; Vol. XV, pp 435-468. (2) Irie, M. In Ribonucleases; D’Alessio, G., Riordan, J. F., Eds.; Academic Press: San Diego, CA, 1997; pp 101-130. (3) Chitester, B. J.; Walz, F. G., Jr. Arch. Biochem. Biophys. 2002, 406, 73-77. (4) Walz, F. G., Jr.; Osterman, H. L.; Libertin, C. Arch. Biochem. Biophys. 1979, 195, 95-102. (5) Kumar, K.; Walz, F. G., Jr. Biochemistry 2001, 40, 3748-3757. (6) Loverix, S.; Doumen, J.; Steyaert, J. J. Biol. Chem. 1997, 272, 9635-9639. (7) Walz, F. G., Jr. Biochim. Biophys. Acta 1997, 1350, 183-188. (8) Mignon, P.; Steyaert, J.; Loris, R.; Geerlings, P.; Loverix, S. J. Biol. Chem. 2002, 277, 36770-36774. (9) De Swarte, J.; De Vos, S.; Langhorst, U.; Steyaert, J.; Loris, R. Eur. J. Biochem. 2001, 268, 3993-4000. (10) Arni, R.; Heinemann, U.; Maslowska, M.; Tokuoka, R.; Saenger, W. Acta Crystallogr., Sect. B 1987, 43, 548. (11) Arni, R.; Heinemann, U.; Tokuoka, R.; Saenger, W. J. Biol. Chem. 1988, 263, 15358-15368. (12) Ikehara, M.; Ohtsuka, E.; Tokunaga, T.; Nishikawa, S.; Uesugi, S.; Tanaka, T.; Aoyama, Y.; Kikyodani, S.; Fujimoto, K.; Yanase, K.; Fuchimura, K.; Morioka, H. Proc. Natl. Acad. Sci. U.S.A. 1986, 83, 46954699. (13) Steyaert, J.; Opsomer, C.; Wyns, L.; Stanssens, P. Biochemistry 1991, 30, 494-499. (14) Hubner, B.; Haensler, M.; Hahn, U. Biochemistry 1999, 38, 13711376. (15) Jurecka, P.; Hobza, P. J. Am. Chem. Soc. 2003, 125, 15608-15613. (16) De Swarte, J.; De Vos, S.; Langhorst, U.; Steyaert, J.; Loris, R. PDB code 1i0x, 2001. (17) Cheng, A. C.; Frankel, A. D. J. Am. Chem. Soc. 2004, 126, 434435. (18) Mao, L.; Wang, Y.; Liu, Y.; Hu, X. J. Am. Chem. Soc. 2003, 125, 14216-14217. (19) Mignon, P.; Loverix, S.; Steyaert, J.; Geerlings, P. Int. J. Quantum Chem. 2004, 99, 53-58. (20) Bader, R. F. W. Atoms in Molecules: a Quantum Theory; Clarendon Press: Oxford, U.K., 1990. (21) Bader, R. F. W. Chem. ReV. 1991, 91, 893. (22) Kitaura, K.; Morokuma, K. Int. J. Quantum Chem. 1976, 10, 325. (23) Morokuma, K.; Kitaura, K. Chemical Applications of Atomic and Molecular Electrostatic Potentials; Politer, P., Truhlar, D. G., Eds.; Plenum

Gu et al. Press: New York, 1981. (24) Becke, A. D. J. Chem. Phys. 1993, 98, 5648-5652. (25) Lee, C.; Yang W.; Parr, R. G. Phys. ReV. B 1988, 37, 785-789. (26) Miehlich, B.; Savin, A.; Stoll, H.; Preuss, H. Chem. Phys. Lett. 1989, 157, 200-206. (27) Hehre, W. J.; Radom, L.; Schleyer, P. R.; Pople, J. A. Ab initio Molecular Orbital Theory; Wiley: New York, 1986. (28) Mebel, A. M.; Morokuma, K.; Lin, C. M. J. Chem. Phys. 1995, 103, 7414-7421. (29) Sponer, J.; Leszczynski, J.; Hobza, P. J. Phys. Chem. 1996, 100, 1965-1974. (30) Gu, J.; Leszczynski, J. J. Phys. Chem. A 1999, 103, 577-584. (31) Frisch, M. J.; Trucks, G. W.; Schlegel, H. B.; Gill, P. M. W.; Johnson, B. G.; Robb, M. A.; Cheeseman, J. R.; Keith, T.; Petersson, G. A.; Montgomery, J. A.; Raghavachari, K.; Al-Laham, M. A.; Zakrzewski, V. G.; Ortiz, J. V.; Foresman, J. B.; Cioslowski, J.; Stefanov, B. B.; Nanayakkara, A.; Challacombe, M.; Peng, C. Y.; Ayala, P. Y.; Chen, W.; Wong, M. W.; Andres, J. L.; Replogle, E. S.; Gomperts, R.; Martin, R. L.; Fox, D. J.; Binkley, J. S.; Defrees, D. J.; Baker, J.; Stewart, J. P.; HeadGordon, M.; Gonzalez, C.; Pople, J. A. Gaussian 98, ReVision D.3; Gaussian, Inc., Pittsburgh, PA, 1998. (32) Biegler-Ko¨nig, F.; Scho¨nbohm, J.; Bayles, D. J. Comput. Chem. 2001, 22, 545-559. (AIM2000 homepage: http://www.aim2000.de). (33) Schmidt, M. W.; Baldridge, K. K.; Boatz, J. A.; Elbert, S. T.; Gordon, M. S.; Jensen, J. H.; Koseki, S.; Matsunaga, N.; Nguyen, K. A. J. Comput. Chem. 1993, 14, 1347. (34) Langhorst, U.; Loris, R.; Denisov, V. P.; Doumen, J.; Roose, P.; Maes, D.; Halle, B.; Steyaert, J. Protein Sci. 1999, 8, 722. (35) Richardson, N. A.; Wesolowski, S. S.; Schaefer, H. F. J. Am. Chem. Soc. 2002, 124, 10163-10170. (36) Smith, B. J.; Swanton, D. J.; Pople, J. A.; Schaefer, H. F.; Radom, L. J. Chem. Phys. 1990, 92, 1240. (37) Reed, A. E.; Weinstock, R. B.; Weinhold, F. J. Chem. Phys. 1985, 85, 735. (38) Reed, A. E.; Weinhold, F. J. Chem. Phys. 1985, 85, 1736. (39) Reed, A. E.; Curtiss, L. A.; Weinhold, F. Chem. ReV. 1988, 88, 899. (40) Reed, A. E.; Schleyer, P. R. J. Am. Chem. Soc. 1990, 112, 1434. (41) Wesolowski, S. S.; Leininger, M. L.; Pentchev, P. N.; Schaefer, H. F. J. Am. Chem. Soc. 2001, 123, 4023. (42) Vanquickenborne, L. G. Quantum Chemistry of Hydrogen Bonds. In Intermolecular Forces; Huyskens, P. L., Luck, W. A. P., ZeegersHuyskens, T., Eds.; Springer-Verlag: Berlin and Heidelberg, Germany, 1991; p 41. (43) Koch, U.; Popelier, P. L. A. J. Phys. Chem. 1995, 99, 9747. (44) Popelier, P. L. A. J. Phys. Chem. A 1998, 102, 1873-1878.