Letter pubs.acs.org/JPCL
Unprecedented External Electric Field Effects on S‑Nitrosothiols: Possible Mechanism of Biological Regulation? Qadir K. Timerghazin* and Marat R. Talipov Department of Chemistry, Marquette University, P.O. Box 1881, Milwaukee, Wisconsin 53201-1881, United States S Supporting Information *
ABSTRACT: Reactions of S-nitrosothiols (RSNOs), ubiquitous carriers of nitric oxide NO and its physiological activity, are tightly regulated in biological systems, but the mechanisms of this regulation are not well understood. Here, we computationally demonstrate that RSNO properties can be dramatically altered by biologically accessible external electric fields (EEFs) by modulation of the two minor antagonistic resonance structures of RSNOs, which have opposite formal charge distributions and bonding patterns. As these resonance contributions relate to the two competing modes of RSNO reactivity with nucleophiles, via N- or S-atom directed nucleophilic attack, EEFs are predicted to be efficient in controlling biologically important RSNO reactions with thiols. For instance, EEF catalysis might be one of the mechanisms behind the high selectivity of protein trans-S-nitrosation reactions, or putative nitroxyl HNO formation via RSNO S-thiolation reactions. SECTION: Biophysical Chemistry and Biomolecules
S
protein S-nitrosation or denitrosation via interactions with other proteins and peptides.3,10,11 However, the origins of the high specificity of these processes are not understood, as extensive statistical analyses failed to reveal specific structural motifs that discriminate Cys residues with respect to Snitrosation.12,13 The alternative S-thiolation pathway (Scheme 1) involves a nucleophilic attack of the thiol at the S atom of the −SNO group and yields disulfide and nitroxyl, HNO,14 an elusive but highly active biologically nitrogen species with great pharmacological potential;15,16 emergent experimental evidence supports this possibility.7,17,18 As S-thiolation pathway is expected to have higher reaction barriers and be less thermodynamically favorable,7,19 it is unlikely to occur unless specifically catalyzed. However, there is no data on possible enzymatic mechanisms of this hypothetical, but potentially very important, pathway of endogenous HNO production. What mechanisms, then, could be used by proteins to selectively catalyze RSNO reactions? Recently, we demonstrated that the dichotomy of RSNO reactivity with nucleophiles can be rationalized from the complex20,21 and unusual electronic structure of the −SNO group.22,23 On the basis of detailed natural bond orbital (NBO) and natural resonance theory (NRT) analyses,24 we proposed an elegant representation of the −SNO group in terms of the three resonance structures (Scheme 2a).22 This resonance representation includes, besides the standard RSNO Lewis structure S with a single S−N bond, a zwitterionic structure D with a double SN bond, and a no-bond ionic structure I. A
-Nitrosothiols (RSNOs) have been recognized as an integral part of the chemical biology of nitric oxide and related species.1−4 These thiolusually cysteinederivatives have been implicated as one the main pools of NO storage and transport in biological systems. Besides small peptides, such as the ubiquitous glutathione, certain cysteine residues in proteins can undergo selective S-nitrosation. This tightly regulated reversible post-translational protein modification is emerging as one of the major pathways how NO exerts its physiological role, while disregulated protein S-nitrosation is a hallmark of many pathological processes.5,6 Moreover, recent experiments and computational studies suggest that the smallest, hydrogensubstituted RSNO−HSNO can also be formed in biological milieu and may play a crucial role in connecting the biological chemistries of the two potent vasodilating agents, NO and hydrogen sulfide H2S.7−9 Despite the immense biological importance of RSNOs, their chemistry and the possible enzymatic mechanisms that control their reactivity in vivo are poorly understood. For instance, biological control of RSNO reactions with thiols that have two possible pathways (Scheme 1) is not clear. Trans-S-nitrosation pathway involves a nucleophilic attack of a thiol or thiolate at the RSNO nitrogen atom and is one of the main mechanisms of Scheme 1. Two Pathways of RSNO Reactions with Thiols
Received: February 16, 2013 Accepted: March 13, 2013 Published: March 13, 2013 © 2013 American Chemical Society
1034
dx.doi.org/10.1021/jz400354m | J. Phys. Chem. Lett. 2013, 4, 1034−1038
The Journal of Physical Chemistry Letters
Letter
D- or I-effects, depending on the field direction orientation (Figure 1a).
Scheme 2. (a) RSNO Representation in Terms of the Three Resonance Structures with a Single S and Double D S−N Bonds, and a No-Bond Ionic Structure I; (b) The Antagonistic Structures D and I Drive the RSNO Reactions with Nucleophiles at the S and N Atoms of the −SNO Group, Respectively
surprising coexistence of the two antagonistic resonance structures D and I with the opposite formal charge distributions and bonding patterns is possible because the underlying orbital interactions occur in mutually orthogonal π- and σ-orbital manifolds.22,25 The antagonistic nature of the −SNO group elegantly reconciles the observed double-bond traits of the S−N bond26−28 (structure D) with its unusual elongation and weakness29 (structure I), as well as the duality of RSNO reactivity with nucleophiles (Scheme 1b). Indeed, D favors a nucleophilic attack at the electrophilic S+ atom implied by that structure, whereas I describes the NO group as a nitrosonium ion NO+ and thus favors a nucleophilic attack at the N atom. As these two structures imply opposite formal charges on the −SNO group atoms, coordination of charged or neutral Lewis acids and bases can effectively modulate RSNO properties, leading either to “D-effect” (shorter, stronger S−N bond, and electrophilic S atom) or to “I-effect” (longer, weaker S−N bond, and electrophilic NO+ moiety).30 First examples of Lewis-acid catalysis of RSNO reactions via D- or I-effects include acid-catalyzed RSNO hydrolysis,23 as well as RSNO reactions with alkenes and alkynes.31 Acid catalysis of Sthiolation reactions recently reported by Liebeskind et al.17 also agrees with the D-effect achieved via N-protonation of the −SNO group. We have recently shown30 that complexation with charged and polar residues provide proteins with a plethora of possibilities to modulate the reactivity of biological RSNOs through D- and I-effects. However, specific interactions with charged/polar residues is not the only way to control the RSNO chemistry in proteins. For instance, catalytic effects of external electric fields (EEFs) created by the protein environment have recently attracted considerable attention.32−38 Although intense enough EEFs (∼0.1 au, 1 au = 51.4 V Å−1) can break strong covalent bonds and lead to reactions such as CO2 decomposition,39 biologically accessible EEFs (∼0.01 au)32,40 do not significantly affect the properties of closed-shell, ground-state covalently bound molecules. On the other hand, these “physiological” EEFs have been shown to significantly alter open-shell iron−porphyrin complexes,33−37 charge-transfer complexes and excited states,38 as well as transition states (TSs) of various reactions.32,37,40 Although RSNOs are relatively stable closed-shell molecules, their antagonistic nature suggests that the −SNO group may be as prone to the EEF modulation as fleeting TSs or reaction intermediates. Depending on the field direction, one could expect that moderate EEFs oriented along the S−N or S−O vectors could lead to sizable
Figure 1. EEF effects on the S−N bond length in RSNOs: (a) qualitatively predicted from the resonance considerations and (b) calculated at the CCSD(T)/aug-cc-pV(Q+d)Z level for trans-HSNO and model molecules with single, double, and π-conjugated S−N bonds. The results of DFT calculations at the PBE0/def2-TZVPPD level are also shown for HSNO. Following the convention used in the electronic structure codes, the positive direction of EEF is from the negative to the positive charge; the EEF vector is oriented along the S−N bonds.
Here, we test this hypothesis by computationally examining the effects of biologically relevant EEFs on the HSNO geometry in comparison with model molecules with single (H2N−SH), double (HONS), and π-conjugated (HNS O) S−N bonds at the CCSD(T)/aug-cc-pV(Q+d)Z level41,42 (Figure 2b). With the EEF vector oriented along the S−N bonds and varied within the ±0.0125 au range, these three model molecules demonstrated moderate polarization resulting in nearly linear changes in the S−N bond lengths up to ±0.02 Å relative to the unperturbed molecules. On the other hand, the HS−NO bond demonstrated an order of magnitude larger changes in length, in agreement with the resonance picture (Figure 2a). At one end of the range, FZ = +0.0125 au, the S−N bond is shortened by ∼0.1 Å relative to zero-field geometry, consistent with a strong D-effect, whereas the same EEF oriented in the opposite direction, FZ = −0.0125 au, leads to even stronger I-effect, with >0.2 Å S−N bond elongation. The S−N bond length dependence on the EEF strength calculated with density functional theory (DFT) at the PBE0/ def2-TZVPPD level43,44 is in excellent agreement with CCSD(T)/aug-cc-pV(Q+d)Z results across the entire range of the field strengths studied (Figure 1b). The DFT calculations suggest that further increase of the EEFs’ strength in the negative direction leads to eventual S−N bond dissociation into HS− and NO+ ions. Thus, I, which is a minor component 1035
dx.doi.org/10.1021/jz400354m | J. Phys. Chem. Lett. 2013, 4, 1034−1038
The Journal of Physical Chemistry Letters
Letter
Figure 3. Correlations between the ΔAS values and S−N bond length changes induced by EEFs in trans-HSNO and cis-MeSNO are nearly linear, except for the strongest fields when the RSNO molecule starts to lose its antagonistic character.
initio coupled-cluster results (Figures 1b and 2b), instills further confidence in using DFT calculations with PBE0 functional followed by NRT analysis to characterize the modulation of the −SNO group electronic structure by the environmental factors. Although comparing the D and I resonance weights relative to those in the unperturbed RSNO molecule provides an adequate measure of the D- or I-effects of an external perturbation, a single metric to describe the −SNO group would be more practical. To this end, we introduced a ΔAS parameter30 calculated as a difference in the NRT weights of the two antagonistic structures (ASs) normalized relative to the respective unperturbed RSNO molecule:
Figure 2. (a) EEF effects on the resonance weights obtained with NRT from PBE0/def2-TZVPPD calculations of trans-HSNO optimized in EEF (solid lines) and kept in its zero-field geometry (dashed lines). (b) The dependence of the dipole moment projection on the S−N vector calculated with CCSD(T)/aug-cc-pV(Q+d)Z and PBE0/ def2-TZVPPD for relaxed HSNO geometries.
within a formal resonance description of an unperturbed RSNO molecule, becomes the physical reality in a sufficiently strong EEF. Similar results have been obtained with the EEF oriented along the S−O vector of HSNO (Figure S1 in the Supporting Information), as well as for a CysNO model, cis-MeSNO (Figure S2). The resonance weights calculated with NRT24,45,46 from the PBE0/def2-TZVPPD density matrix show the evolution of the HSNO electronic structure as the EFF strength and direction change along the S−N vector (Figure 2a). The D and I contributions change synchronously and linearly, while the contribution of the conventional structure S remains nearly constant (65−70%), and decreases only in highly negative EEFs. Importantly, the resonance weights calculated for the fixed unperturbed HSNO geometry demonstrate very similar dependence (Figure 3a, dashed lines). Interestingly, rather low contributions from I (∼15%), even in combination with similar or larger D contribution, already correspond to significant elongations of the S−N bond (∼0.1 Å). Thus, a question may arise if the resonance weights calculated with NRT provide a quantitatively correct picture of the RSNO electronic structure. We compared NRT-calculated resonance structure weights evolution with the changes of the dipole moment projection to the S−N vector, μZ, calculated with both coupled-cluster and DFT methods (Figure 2b). According to NRT, the contribution of D is larger than that of I for FZ > −0.004 au, while I dominates in the more negative fields. Practically at the same FZ ∼ 0.004 au when D and I compensate each other and switch in order, μZ reaches zero and reverses its sign, in accord with the NRT description. This remarkable agreement between the NRT resonance weights and a physical observablea dipole moment projection provides qualitative and quantitative validation of the NRT results. This, and the comparison the DFT and high-level ab
ΔAS = (D% − I %) − (D% − I %)0
so the positive and negative ΔAS values correspond to D- and Ieffects, respectively. Conveniently, the ΔAS parameter is less dependent on the level of theory as well as a particular RSNO molecule: e.g., the ΔAS dependences on FZ are virtually identical for HSNO and MeSNO (Figure S3 in Supporting Information). ΔAS values correlate very well with various properties of the −SNO group, e.g., the S−N bond length (Figure 3) or its homolytic dissociation energy, BDE.30 In the case of recently reported RSNO-charged residue model complexes, the calculated ΔAS values vary from −10 to +20, which corresponds to significant variation of BDE(S−N), from 28 to 40 kcal/ mol.30 Further, we found that |ΔAS| = 8−14 due to RSNO complexation with neutral Lewis acids such as BF3 can translate into 11−13 kcal/mol barrier reduction in the RSNO-alkyne/ alkene reactions.31 As the ΔAS values obtained here span an even larger range (−20 to +15, Figure 3), we can expect that EEFs could prove efficient in catalyzing or inhibiting biologically relevant RSNO reactions such as trans-S-nitrosation and S-thiolation (Scheme 1). Calculated activation barriers of RSNO-thiol reactions are highly dependent on the inclusion of the polarizable environment effects and explicit consideration of water molecules;19 Sthiolation barriers appear to be especially sensitive to the number of water molecules included. A detailed mechanistic analysis of these two reactions, including the effects of environment, water molecules, charged residues, and EEFs is beyond the scope of this Letter, and will be presented elsewhere. However, the preliminary data that we show here 1036
dx.doi.org/10.1021/jz400354m | J. Phys. Chem. Lett. 2013, 4, 1034−1038
The Journal of Physical Chemistry Letters
Letter
species. For instance, the puzzling selectivity of S-nitrosation of cysteine residues in proteins might be in part determined by local EEFs exerted by nearby charged residues, α-helix dipoles,47 or membrane potentials. Although the possible role of RSNO S-thiolation reactions in production of endogenous HNO remains an open question, our preliminary results suggest that enzymatic catalysis of S-thiolation via EEFs and/or specific interactions with charged and polar residues is not outside the realm of possibility. Further detailed studies of the possible catalytic mechanisms that exploit the unusual antagonistic nature of the −SNO group are clearly warranted, as they may not only reveal new important facets of RSNO biochemistry, but also help in designing new RSNO reactions of bioanalytical and pharmacological interest.31
(Figure 4) illustrate well the potential effectiveness of EEFs in controlling RSNO reactions with nucleophiles.
■
ASSOCIATED CONTENT
S Supporting Information *
Computational details, Figures S1−S3, additional references, Cartesian coordinates and energies of all calculated structures. This material is available free of charge via the Internet at http://pubs.acs.org.
■
AUTHOR INFORMATION
Corresponding Author
*E-mail:
[email protected]. Notes
The authors declare no competing financial interest.
■
Figure 4. (a) TS structures for model trans-S-nitrosation and Sthiolation assisted by four water molecules, optimized at the PBE0/ def2-SV(P)+d level in a dielectric environment; solvating water molecules not directly involved in a proton transfer are shown semitransparent. (b) Estimated effects of EEFs directed along the S−O axis on the reaction barriers, evaluated as single-point electron energy differences between the TS and the reagent cluster, ΔEe‡, with PBE0/ def2-TZVPPD for zero-field geometries.
ACKNOWLEDGMENTS Q.K.T. thanks Christine Morales (University of Wisconsin-Eau Claire) for fruitful suggestions. This work has been funded by Marquette University; calculations were performed on the highperformance computing cluster Père at Marquette University funded by NSF awards OCI-0923037 and CBET-0521602.
■
TS calculations of water-assisted MeSNO + MeSH model reactions (Figure 4a) suggest moderately high barriers, 21.1 and 22.4 kcal/mol for trans-S-nitrosation and S-thiolation, respectively, estimated from the zero-point corrected energies calculated at the PBE0/def2-TZVPPD//PBE0/def2-SV(P)+d level. However, EEFs directed along the S−O vector of the RSNO molecule can lead to dramatic alterations of these barriers (Figure 4b). The concerted trans-S-nitrosation reaction barrier significantly lowers in EEFs with FZ < 0, and even medium-range EFFs (FZ = −0.005 au) lead to >5 kcal/mol drop in the barrier height. On the other hand, in EEFs with FZ > 0 the trans-S-nitrosation reaction barrier quickly increases, while the S-thiolation barrier quickly drops and becomes lower than the trans-S-nitrosation barrier for FZ > 0.002 au. Moreover, preliminary estimations shown in Figure 4b suggest that in stronger fields, |FZ| ≥ 0.01 au, depending on the field direction, one of the two reaction barriers could reach exceptionally low values (
