TMCnet News
Isomerism of Cycloserine and Its Protonated Form [ChemPlusChem](ChemPlusChem Via Acquire Media NewsEdge) A comprehensive ab initio investigation has been performed on the structure and stability of the isomers of cycloserine and its protonated forms in the unsolvated state. Many conformers of cycloserine in the ketonic (K), enolic (E4 and E2), and zwitterionic (Z7 and Z2) forms have been characterized. Enols E2 are only a few kilocalories per mole less stable than the K isomers. Enols E4, as well as Z7 and Z2 zwitterions, are several tens of kilocalories per mole less stable than K. All the above isomeric structures exhibit pronounced isoxazolidine ring puckering, which generates very rich conformeric landscapes. The relative stability of the conformers of K, E2, and E4 responds essentially to a complex balance between the attractive and repulsive electrostatic interactions among their functional groups. The preferred site of protonation of cycloserine in the gas phase has been also investigated computationally and experimentally by IR multiphoton dissociation (IRMPD) spectroscopy. The most basic center of cycloserine is the N(7) nitrogen atom (proton affinity (PA)= 215.3 kcalmol-1). Another important basic site is the O(6) oxygen atom (PA =213.0 kcalmol-1). Their most populated conformers have been identified by IRMPD spectroscopy. Their predominance responds to the electrostatic interactions among the functional groups of the protonated molecule. Keywords : ab initio calculations · antibiotics · conformation analysis · IR spectroscopy · isomers Introduction Cycloserine (4-amino-1,2-oxazolidin-3-one ; K in Figure 1) is a broad-spectrum antibiotic that is produced by a strain of Streptomyces orchidaceous or by synthesis.[1] The R enantiomer (d-cycloserine) is used as a second-line drug in the treatment of pulmonary and extra-pulmonary tuberculosis. In vitro stud- ies demonstrated that d-cycloserine is also an effective, moder- ately specific partial agonist at the glycine site coupled to N- methyl-d-aspartate (NMDA) receptors.[2] Because of its action on NMDA receptors, d-cycloserine is commonly used in a number of clinical treatments of anxiety[3-5] and obsessive- compulsive disorders.[6, 7] The S enantiomer (l-cycloserine) ex- hibits the ability to increase the levels of the inhibitory neuro- transmitter g-aminobutyric acid (GABA) as well as to inhibit the pyridoxal-5'-phosphate (PLP) enzyme GABA-transferase.[8] Since its commercial availability in 1952, there have been very few studies on the physicochemical properties of cycloser- ine.[9, 10] Evaluation of data from spectral measurements and ti- tration at varied temperature and solvent composition show that cycloserine in neutral aqueous solutions exists in the zwit- terion form Z7, whereas in acidic water the protonated form P7 predominates (Figure 1).[11, 12] The ketonic form K of cycloserine is most favored in organic solvents. X-rays structural analysis of monohydrated cycloserine crystals revealed the exclusive pres- ence of the [Z7·H2O] adducts with the isoxazolidine ring puck- ered at C(5).[13] The same analysis on [P7·Cl^] crystals pointed to the isoxazolidine ring as approximately planar.[14-16] Few pieces of information are also available on the structure, conformation, and relative stability of cycloserine in the unsol- vated state, that is, under conditions mimicking the NMDA re- ceptor sites, wherein the drug may undergo extensive desolva- tion and proton transfer to/from the receptor cavity. Interest in this point is owing to a recent computational study showing that the biological activity of some NMDA receptor ligands, in- cluding cycloserine, is determined by their highest occupied molecular orbitals (HOMOs) and lowest unoccupied molecular orbitals (LUMOs), the features of which strongly depend on the specific conformation adopted by the drug.[17] The electronic structure and properties of cycloserine have been recently in- vestigated using X-ray photoelectron spectroscopy (XPES) and DFT calculations.[18] Two conformers of cycloserine have been predicted to predominate in the gas phase, separated in energy by 1.1 kcal mol^1. However, the experimental spectra did not provide any proof of the presence of these conformers. Therefore, we believe it is important to shed more light on the structure and the conformation of unsolvated cycloserine in the ketonic (K), enolic (E4 and E2), and zwitterionic (Z7 and Z2) forms (Figure 1). Another important question to be raised is the preferred site of protonation of cycloserine in the gas phase. Cycloserine is a polyfunctional molecule and the prediction of its protonation site(s) is a challenging problem from the experimental as well as the theoretical standpoint, especially if one considers the possible changes of the conformation of the molecule upon protonation.[19] Therefore, the present investigation on the structure and conformation of unsolvated cycloserine is ex- tended to its protonated isomers (P7, P6, P2, and P1 in Figure 1) as well. The study is mainly based on ab initio calculations. Ex- perimental validation of the computational results is obtained by using variable-wavelength IR multiphoton dissociation (IRMPD) spectroscopy[20-25] of protonated cycloserine generated in the electrospray ionization source of a mass spectrometer (ESI-MS). The three-dimensional (3D) shape of the structures of Figure 1 is determined by the following : 1) The puckering of the isoxazolidine ring, which is described by the sign of the dihedral angle x= N(2)-O(1)-C(5)-C(4) (Figure 2). In particular, a positive x value is associated with the isoxazolidine ring puckering O(1)-exo-C(5)-endo, relative to the position of the N(7) atom (North conformers denot- ed with the N symbol). A negative x value is instead associ- ated with the isoxazolidine ring puckering O(1)-endo-C(5)- exo (South conformers denoted with the S symbol). In the N conformers, the N(7)H2 group is equatorially oriented, whereas it is axially oriented in the S ones. The absolute value of the dihedral angle x gives a measure of the ampli- tude of the corresponding ring puckering. 2) The dihedral angle w = C(5)-O(1)-N(2)-H (Figure 2), which describes the orientation of the N(2)^H hydrogen in the isoxazolidine ring. A 1308 < j w j < 1808 value indicates that the N(2)^H hydrogen is equatorially oriented (denoted with the symbol eq in Figure 2), whereas a 808 < jwj < 1308 value indicates that the N(2)^H hydrogen is axially oriented (denoted with the symbol ax in Figure 2). 3) The torsional angle g defining the orientation of the N(7)H2 lone pair relative to the C(3) atom. The g value is calculated by adding 1808 to the angle y between C(3) and the bisec- tor of the H-N(7)-H angle (Figure 3a). Depending upon the value of the g angle, the symbolism illustrated in Figure 3b is applied. 4) Only for structures E4, E2, and P6, the torsional angle d= H- O(6)-C(3)-C(4). An absolute value of d around 08 indicates that the O(6)^H hydrogen points to the N(7)H2 group (syn), whereas an absolute value of approximately 1808 indicates that the O(6)^H hydrogen points to the N(2) center (anti). 5) Only for the P1 structure, the torsional angle c =C(3)-N(2)- O(1) + -H. A negative c value indicates that the O(1) +^H hy- drogen is oriented up relative to the C(3)-N(2)-O(1) plane, whereas a positive c value indicates that its oriented down relative to the same plane. Results and Discussion Structure and stability of the K conformers Table 1 reports the stable conformers of K, together with the corresponding MP2/6-311 +G**-calculated enthalpy and free- energy gaps at 298 K relative to the global minimum, that is structure 0KNe. Their geometrical parameters are listed in Table S1 of the Supporting Information. As shown in Table 1, the N conformers are invariably more stable than the corre- sponding S forms and, within each family, the t > + g >^g sta- bility order is observed. It should be also noted that the ax structures are always less stable than their eq analogues. The reduced repulsion between the trans-anti-oriented O(1) and N(2) lone pairs in the eq structures may account for their greater stability relative to the corresponding ax conformers (Figure S1). Repulsive effects may develop also between the N(7) lone pair and the C(3)=O(6) p electrons. Among the N- and S-type K structures, the t rotamers are most stable by virtue of the increased distance and, then, reduced repulsion between the N(7) lone pair and the C(3)=O(6) p electrons. This repulsive interaction may be partially compensated in the S- type structures by the attractive N(7)^H···p and N(7)^H···O(1) ones that are thought important in their t conformers.[18] The N > S stability trend points to the repulsive forces as prevailing over any attractive interactions. The MP2/6-311 + G**-calculated relative stability of the cyclo- serine conformers of Table 1 is reported in Figures 4 and 5 to- gether with the activation enthalpies for the corresponding in- terconversions. Figure 4 illustrates the energetics of the torsion of the g angle in the North equatorial (Ne), North axial (Na), South equatorial (Se), and South axial (Sa) conformers. Figure 5 illustrates the energetics of the ring repuckering (con- tinuous lines) and of the inversion of configuration of the N(2) center (broken lines) in cycloserine. As shown in Figure 4, the + g Q t interconversion in the North conformers is essentially barrierless. Small, but appreciable activation enthalpies are in- stead calculated for the same angle g torsion in the South con- formers (up to 1.1 kcal mol^1) and in the ^gQt and ^gQ + g torsions in both the N and S families (up to 1.7 kcal mol^1). Analysis of Figure 5 indicates that the inversion of the N(2) configuration involves very small barriers (broken lines in Figure 5). The activation enthalpy of the KNaQKNe and KSe QKSa processes never exceeds 0.5 kcal mol^1. In contrast, the North- QSouth repuckering involves more pronounced activation bar- riers, ranging from 1.1 kcal mol^1 in +KSaQ+KNe, to 3.0 kcal mol^1 in ^KNaQ^KSe (continuous lines in Figure 5). As shown, no first-order saddle-point structures could be identified on the cycloserine potential-energy surface (PES) involving simul- taneous ring puckering and inversion of the N(2) configura- tion. According to the relative stability of the cycloserine con- formers (Table 1) and taking into account their relatively facile interconversion ( + KNeQ0KNe (Figure 4) and 0KSeQ0KNaQ0KNe (Figure 5)), it is suggested that, at 298 K, over 99 % of the K structures are represented by the eq conformers 0KNe (51 %), + KNe (26 %), and 0KSe (22 %). This conclusion diverges from the results of recent B3LYP/6-311 ++G** calculations that did not take into consideration the N-type structures and predicted a 0KSe/+KSe/^KSe = 77:21:2% distribution for cycloserine at 395 K.[18] These computational results were found inconsistent with the XPES spectra of cycloserine, which have been inter- preted with the formation of a single K conformer, that is 0KSe, at 395 K.[18] To look into the origin of such a discrepancy, we re- peated our calculations at the B3LYP/6-311 ++G** level of theory. The results, reported in the last two columns of Table 1, confirm the MP2/6-311 + G** indication that the 0KSe conformer is less stable than the 0KNe one. On these grounds, a reconsider- ation of the experimental cycloserine XPES in light of the cor- rect conformational distribution of Table 1 is in order. Studies are in progress along this line. Structure and stability of cycloserine enols E2 and E4 The stability of the cycloserine enols E2 and E4, relative to the global minimum 0KNe, are reported in Tables 2 and 3, respec- tively. Their geometrical parameters are listed in Tables S2 and S3. Inspection of Table 2 reveals that, in general, the anti con- formers of both the N and S families are more stable than the syn ones. The t rotamers of the anti families are most stable, whereas the ^g forms are most stable in the syn families. The structure + ENs is not a local minimum on the E2 PES because it undergoes a barrierless C(4)^ N(7) bond rotation to ^ENs.No similar aptitude has been no- ticed for the corresponding + ESs, which corresponds to a local, though high-energy minimum on the E2 PES. These findings can be explained by the attrac- tive N(7)^H···O(6)^H interaction that is most favored in the t ro- tamers 0ENa and 0ESa by the prox- imity of the two N(7)^H hydro- gen atoms with the O(6) center (hydrogen-bond lengths : 2.90 and 3.00 ^ (0ESa) ; 3.01 and 3.43 ^ (0ENa)). When allowed by the syn orientation of the O(6)^H bond, a further stabilizing O(6)^ H···N(7)^H interaction may operate, which accounts for the en- hanced stability of the ^ESs and ^ENs conformers relative to the corresponding 0ESs and 0ENs ones (for instance, hydrogen-bond length : 2.24 (^ENs) versus 2.64 ^ (0ENs)). The 0KNe!0ENe step is the lowest-energy intramolecular path connecting cycloserine K with its E2 tautomers. This process involves a MP2/6-311 + G**-calculated activation barrier of 50.4 kcal mol^1. As shown in Table 3, nine are the stable N-type structures belonging to the E4 family. Among the syn structures, the + Ees structure is not stable since it evolves spontaneously to the ^Ees one. No similar aptitude has been observed for the corre- sponding axial form +Eas. It is also worth noting that, among the anti structures, both the 0Eea and the 0Eaa are not local minima on the E4 PES, since they evolve directly to + Eea and + Eaa, respectively. Because the chirogenic C(4) center of K and E2 is clearly destroyed in the E4 tautomers, their stable S-type structures are the mirror images of the corresponding N-type ones (Figure 6). The relative structural features are given in pa- rentheses in Table 3 and Table S3. It is evident that the racemi- zation of the chirogenic C(4) center in K may take place in the isolated state only through the resonant K (l enantiomer)- QE4QK (d enantiomer) tautomeric equilibrium. The intramo- lecular 0KNe!0Ees step is a possible path for this process, which involves a MP2/6-311 + G**-calculated activation barrier of 79.0 kcal mol^1. Structure and stability of cycloserine zwitterions Z2 and Z7 The formal proton transfer C(4)!H + !N(2) in both S- and N- type cycloserine conformers leads to the same zwitterionic structure Z2 (DH298 = 37.0 kcal mol^1; DG298= 35.5 kcal mol^1, above the global minimum 0KNe). The formal proton transfer N(2)!H + !N(7) in the S-type cycloserine conformers gives rise to a stable Z7 S-type structure (DH298 = 36.0 kcal mol^1; DG298= 36.0 kcal mol^1, above the global minimum 0KNe). Any attempt to generate the corresponding N-type zwitterion leads invaria- bly to the ^ENs enol through a barrierless N(7)!H + !O(6) proton transfer. The geometrical parameters of Z2 and Z7 are listed in Table S4. Structure and stability of protonated cycloserine Tables 4 and S5 report all the stable structures of the P7, P6, P2, and P1 families of protonated cycloserine, together with the corresponding MP2/6-311 + G**-calculated enthalpy and free- energy gaps at 298 K relative to the global minimum, that is, the North conformer of P7 (PN). Only the equatorial conformers of P7 and P6 can be identified as local minima on the relative PES. The corresponding axial conformers are not stable and evolve to the equatorial forms with no appreciable energy bar- rier. The N conformer of P7 (PN) is slightly more stable than the S one ( PS ). This may suggest the operation in PN of an intramo- lecular interaction between N(7)^H + and the O(6) lone pair that is more effective than in the PS conformer (hydrogen- bond length : 2.15 ^ (PN); 2.67 ^ (PS)). In agreement with the crystal structure of the hydrochloride salt of cycloserine,[14-16] the N(7) atom is the most basic site of cycloserine also in the gas phase (proton affinity = 215.3 kcal mol^1). According to MP2/6-311 + G** calculations, the 298 K activation enthalpy and free energy of the PN !PS conformational transition are 3.6 and 4.2 kcal mol^1, respectively. In the P6 family, the proton bonding between the N(7) lone pair and the H +^O(6) moiety is allowed only in the syn con- formers and is stronger in the N conformer than in the S ana- logue (for instance, hydrogen-bond length : 2.22 ^ (^PNes); 2.75 ^ (^PSes)). Such an interaction is so stabilizing that ^PNes and ^PSes act as thermodynamic sinks for structures, like 0PNes, + PNes, and + PSes, not identifiable as local minima on the P6 PES. A similar situation is observed for the ^PS rotamer, which is not a local minima on the P2 PES since it rearranges directly to the most stable conformer + PS. Besides this, the stability of the P2 conformers appears to be not very sensitive to structural features. In contrast, the stability of the P1 conformers depends on the C(3)-N(2)-O(1) + -H dihedral angle c (Table S5 and Figure S2). The PNad> PNeu > PNed> PNau stability order (Table 4) is attributed to an interplay between the attractive interaction between the H +^O(1) moiety and the N(2) lone pair. Indeed, the c values of about + 120 and ^1508 of PNad and PNeu, respectively, indicate that the H +^O(1) bond approximately eclipses the N(2) lone pair and, therefore, the electrostatic interaction is maximized (Table S5 and Figure S2). In the least-stable PNed and PNau con- formers, a repulsive interaction may develop between the re- sidual lone pair of the O(1) moiety and that of N(2), as suggest- ed by the corresponding c values of about + 80 and ^1508. The same arguments apply to the PSau > PSed > PSeu > PSad stabil- ity order. Some hypothetical P1 structures, such as 0PNau, 0PSeu, 0PSau, + PSeu,and + PSau are not listed in Table 4 because they're not identifiable as local minima on the corresponding PES. For instance, the last two structures directly transform to the PS structure by proton transfer. IRMPD spectra of ESI-protonated cycloserine Collision-induced dissociation (CID) of protonated cycloserine (m/z 103), formed by ESI-MS, yields two predominant frag- ments at m/z 75, by formal loss of CO, and at m/z 58, by formal loss of CO and NH3. The same fragmentation pattern is observed in the corresponding IRMPD spectrum recorded in the 900-2000 and 2800-3600 cm^1 frequency ranges (Fig- ure 7a). The curve describes the wavelength dependence of the fragmentation efficiency, defined as R =^log[(I103)/ 03)] , in which I103 is the intensity of the parent m/z 103 ion and I75 and I58 are the intensity of its m/z 75 and m/z 58 fragments, respectively.[26] Three intense IR bands are ob- served in the 2800-3600 cm^1 range : two relatively sharp peaks at 3300 and 3350 cm^1 and a broader one at approxi- mately 3440 cm^1. These bands are accompanied by a barely visible bump below 3200 cm^1. The frequencies of these IR bands fall in the spectral range typical of N^H bond stretching. In the 900-2000 cm^1 range, two intense signals at 1770 and 1460 cm^1 are observed, which are accompanied by weaker absorptions at 1580, 1200, 1020, and 909 cm^1. The intense band at 1770 cm^1 is typical of C=O bond stretching, whereas the others can be attributed to N^H bending. The black curve of Figure S3 describes the fragmentation ef- ficiency of protonated cycloserine in the 2800-3600 cm^1 fre- quency range, defined as R =^log[(I103)/(I75+I103)]. The gray curve refers to the IR frequency dependence of R = ^log- [(I103)/(I58+I103)] . The decrease of the m/z 75 ion intensity be- tween 3428 and 3458 cm^1 (black curve), corresponding to the absorption maximum of the gray curve (see also Figure S4), provides strong support to the hypothesis that protonated cy- closerine (m/z 103) releases CO to give m/z 75 which, in turn, yields m/z 58 by further photon absorptions. According to the DG values of Table 4 and taking into ac- count the moderate activation free energy for the PS !PN con- formational transition at 298 K (3.8 kcal mol^1), most of the pro- tonated cycloserine is formed in the PN (90 %) and PS (9 %) con- formations. The residual 1 % may be assigned to the ^PNes isomer. Such a distribution finds strong support in the experi- mental IRMPD spectrum of ESI-formed protonated cycloserine (Figure 7a). While no appreciable differences could be dis- cerned among the MP2/6-311+ G**-calculated vibrational spectra of the PN, PS,and^PNes in the 900-2000 cm^1 range (left side of Figure 7a-c), their comparison in the 2800- 3600 cm^1 range (right side of Figure 7a-c) confirms the large predominance of PN over PS and the virtual absence of ^PNes. The signal at 3134 cm^1, present in the MP2/6-311 + G**-cal- culated vibrational spectra of PN (Figure 7b), is barely detecta- ble in the IRMPD spectrum of protonated cycloserine (see inset in Figure 7a). Such a situation is by no means unusual. Indeed, the intensity of the experimental IRMPD signals is determined by the probability of depositing enough excess energy into the specific bond(s) involved in the complex fragmentation. This does not depend only on the efficiency of resonant photon absorption, but also on the efficiency of the intramo- lecular redistribution of the excess of the vibrational energy as well as on the dissociation energy barrier.[27] Thus, it is possible that the resonant absorption by a given IR-active vibrational mode in a complex could produce a signal with a relative in- tensity different from the corresponding calculated linear IR absorption intensity. The very weak band below 3200 cm^1, shown in the inset of Figure 7a, is attributed to the stretching of the N(7)^H bond coordinated with the carbonyl oxygen in the equatorial PN structure. The sharp IRMPD peaks at 3350 and 3300 cm^1 are assigned to the free H-N(7)-H asymmetric and symmetric stretchings, respectively. Finally, the N(2)^H stretching in PN is mostly responsible for the 3440 cm^1 band observed in the IRMPD spectrum of protonated cycloserine. Conclusion In conclusion, ESI-formed protonated cycloserine exhibits an IRMPD spectrum that fully agrees with the predominant forma- tion of the N(7) protonated conformers PN, as correctly predict- ed on the grounds of MP2/6-311+ G** calculations. Structure PN, as well as its less stable PS conformer, exhibit pronounced isoxazolidine ring puckering. This observation contrasts with the approximately planar isoxazolidine ring, observed in [P7·Cl^] crystals,[14-16] which can be explained by the combined ion pairing and crystal-packing effects. Experimental Section IRMPD experiments Cycloserine (d enantiomer) was purchased from a commercial source and used without further purification. The protonated cy- closerine has been generated by ESI-MS of calibrated solutions of cycloserine (10^5 m) either in pure CH3OH or CH3CN. The ion was then introduced in a modified Bruker Esquire 6000 quadrupole ion trap and isolated using the standard Bruker Esquire Control (v6.2) software. ESI conditions used were as follows : syringe pump rate : 180 mLh^1; spray voltage: 3500 V; capillary temperature: 250 8C. Mass-selected ions were irradiated using the MS2 step, in which the excitation amplitude was set to zero to avoid any collision-in- duced dissociation (CID) process. Mass spectra were recorded after five accumulations, using the standard mass range (m/z 50-3000) and normal scan-mode resolution (13000 Da s^1), with the accumu- lation time in the range of 200-500 ms in the free-electron laser (FEL) region and 1-2 s in the optical parametric oscillator/amplifier (OPO/OPA) one, depending on the fragmentation efficiency of the process. IR spectroscopy in the 900-2000 cm^1 wavenumber range was performed using the CLIO FEL. The light is produced in 8 ms long pulse trains, the macropulses, of IR laser pulses a few picosec- onds in duration, the micropulses. The macropulses repetition rate is 25 Hz, whereas that of the micropulses is 62.5 MHz. Typical ener- gies reached within one macropulse can be 40-60 mJ : the macro- pulse energy in the present experiments is about 20 mJ.[28] The 2800-3600 cm^1 wavenumber range was explored using an IR OPO/OPA system of LaserVision, pumped by a 10 Hz Nd :YAG laser (Excel Technology Europe GmbH Surelite-II, 650 mJ per pulse, 8 ns pulse duration). The output energy, measured between 3400 and 3600 cm^1, is about 23 mJ per pulse with a spectral bandwidth of approximately 5 cm^1. The loss of energy in the other spectral re- gions is not more than 14 %. The IR-FEL photon energy was in- creased at a rate of about 2.5 cm^1 s^1, whereas that from the OPO/OPA systems was at a rate of about 0.1 cm^1 s^1. Computational details Geometry optimization of the different isomers of cycloserine and all its protonated forms was achieved without any structural con- strain at the MP2/6-311 + G**[29] level of theory using the 1 May 2012 (R1) version of the general atomic and molecular electronic structure system (GAMESS) software[30, 31] running on a six-core (8 Intel-Xeon E5520 2.27 GHz CPU and 24 GB DDR3 RAM each) cluster (48 CPU total) with a Debian GNU/Linux 6.04 operating system. All the conformational degrees of freedom were taken into account, as well as prototropic isomers, which generated tens of structures, including zwitterionic and enolic forms. Harmonic vibrational fre- quencies were determined at this level to characterize the station- ary points as local minima and to estimate their zero-point vibra- tional energy (ZPE) corrections as well as the relevant 298 K enthal- pies and free energies. As far as the positions of the IR bands are concerned, a scaling factor of 0.95 was applied to the calculated frequencies in the 2800-3600 cm^1 range on the basis of their good agreement with the experimental values. Neutral cycloserine in its canonical form was further investigated at the same level of theory to establish the energy barriers separating its conformation- al and prototropic forms. All the optimized transition structures were characterized as true first-order critical points by frequency analysis. The MP2/6-311 +G**-optimized canonical structures of cy- closerine were also used as input geometries and reoptimized at the B3LYP/6-311 ++G**[32-35] level of theory for comparison pur- poses. Acknowledgements This study was supported by the Ministero dell'Istruzione dell'Uni- versit^ e della Ricerca of Italy (PRIN 2010-2011: CUP B81J1200283001). Annito Di Marzio is gratefully acknowledged for technical assistance. We are also grateful to Felice Grandinetti for the sharing his computational resources and competence. [1] F. D. Lowy, J. Clin. Invest. 2003, 111, 1265 - 1273. [2] R. Dall'Olio, O. Gandolfi, R. Gaggi, Behav. Pharmacol. 2000, 11, 631 - 637. [3] G. Kaushal, R. Ramirez, D. Alambo, W. Taupradist, K. Choksi, C. Sirbu, Drug Discov. Ther. 2011, 5, 253 - 260. [4] S. G. Hofmann, A. E. Meuret, J. A. Smith, N. M. Simon, M. H. Pollack, K. Ei- senmenger, M. Shiekh, M. W. Otto, Arch. Gen. Psychiatry 2006, 63, 298 - 304. [5] M. W. Otto, D. F. Olin, N. M. Simon, G. D. Pearlson, S. Basden, S. A. E. Meunier, S. G. Hoffmann, K. Eisenmenger, J. H. Krystal, M. H. Pollack, Biol. Psychiatry 2010, 67, 365 - 370. [6] G. S. Chasson, U. Buhlmann, D. F. Tolin, S. R. Rao, H. E. Reese, T. Rowley, K. S. Welsh, S. Wilhelm, Behav. Res. Ther. 2010, 48, 675 - 679. [7] M. G. Kushner, S. W. Kim, C. Donahue, P. Thuras, D. Adson, M. Kotlyar, J. McCabe, J. Peterson, E. B. Foa, Biol. Psychiatry 2007, 62, 835 - 838. [8] P. Scotto, P. Monaco, V. Scardi, V. Bonavita, J. Neurochem. 1963, 10, 831- 839. [9] F. A. Kuehl, Jr., F. J. Wolf, N. R. Trenner, R. L. Peck, R. P. Buhs, E. Howe, I. Putter, B. D. Hunnewell, R. Ormond, J. E. Lyons, L. Chaiet, K. Folkers, J. Am. Chem. Soc. 1955, 77, 2344- 2345. [10] L. F. Iakhontova, B. P. Bruns, V. D. Kartseva, S. N. Kobzieva, N. A. Perevozskaia, Antibiotiki 1969, 14, 205- 210. [11] J. B. Neilands, Archiv. Biochem. Biophys. 1956, 62, 151 -162. [12] H. H. Lee, N. Takeuchi, H. Senda, A. Kuwae, K. Hanai, Spectrosc. Lett. 1997, 30, 685- 700. [13] H. H. Lee, N. Takeuchi, H. Senda, A. Kuwae, K. Hanai, Spectrosc. Lett. 1998, 31, 1217 -1231. [14] J. W. Turley, R. Pepinsky, Acta Crystallogr. 1956, 9, 948- 51. [15] J. Donohue, Acta Crystallogr. 1957, 10, 383- 384. [16] J. W. Turley, R. Pepinsky, Acta Crystallogr. 1957, 10, 480 -481. [17] J. Yosa, M. Blanco, O. Acevedo, L. R. Lareo, Eur. J. Med. Chem. 2009, 44, 2960 -2966. [18] M. Ahmed, F. Wang, R. G. Acres, K. C. Prince, arXiv.org, e-Print Archive, Physics 2013, 1 -28, arXiv:1305.6384v1 [physics.chem-ph]. [19] D. Kaur, R. Sharma, Struct. Chem. 2012, 23, 905- 919. [20] J. Roithov, Chem. Soc. Rev. 2012, 41, 547-559. [21] N. C. Polfer, Chem. Soc. Rev. 2011, 40, 2211 -2221. [22] J. R. Eyler, Mass Spectrom. Rev. 2009, 28, 448 -467. [23] J. S. Brodbelt, J. J. Wilson, Mass Spectrom. Rev. 2009, 28, 390- 424. [24] L. MacAleese, P. Maitre, Mass Spectrom. Rev. 2007, 26, 583 -605. [25] Y. M. E. Fung, T. Besson, J. Lemaire, P. Maitre, R. A. Zubarev, Angew. Chem. 2009, 121, 8490- 8492; Angew. Chem. Int. Ed. 2009, 48, 8340 - 8342. [26] J. Lemaire, P. Boissel, M. Heninger, G. Mauclaire, G. Bellec, H. Mestdagh, A. Simon, S. L. Caer, J. M. Ortega, F. Glotin, P. Maitre, Phys. Rev. Lett. 2002, 89, 273002. [27] J. Oomens, B. G. Sartakov, G. Meijer, G. von Helden, Int. J. Mass Spectrom. 2006, 254, 1 -19. [28] L. Mac Aleese, A. Simon, T. B. McMahon, J. M. Ortega, D. Scuderi, J. Lemaire, P. Maitre, Int. J. Mass Spectrom. 2006, 249 - 250, 14- 20. [29] J. S. Binkley, J. A. Pople, Int. J. Quantum Chem. 1975, 9, 229- 236. [30] M. W. Schmidt, K. K. Baldridge, J. A. Boatz, S. T. Elbert, M. S. Gordon, J. H. Jensen, S. Koseki, N. Matsunaga, K. A. Nguyen, S. J. Su, T. L. Windus, M. Dupuis, J. A. Montgomery, J. Comput. Chem. 1993, 14, 1347- 1363. [31] M. S. Gordon, M. W. Schmidt in Theory and Applications of Computational Chemistry: The First Forty Years (Eds.: C. E. Dykstra, G. Frenking, K. S. Kim, G. E. Scuseria), Elsevier, Amsterdam, 2005, pp. 1167- 1189. [32] A. D. Becke, Phys. Rev. A 1988, 38, 3098 -3100. [33] A. D. Becke, J. Chem. Phys. 1993, 98, 1372 - 1377. [34] A. D. Becke, J. Chem. Phys. 1996, 104, 1040 -1046. [35] C. Lee, W. Yang, R. G. Parr, Phys. Rev. B 1988, 37, 785 -789. Received: January 3, 2014 Published online on February 24, 2014 Caterina Fraschetti,[a] Antonello Filippi,[a] Stefano Borocci,[b] Vincent Steinmetz,[c] and Maurizio Speranza*[a] [a] Dr. C. Fraschetti, Prof. A. Filippi, Prof. M. Speranza Dipartimento di Chimica e Tecnologie del Farmaco Universit^ "La Sapienza", P.le A. Moro 5, 00185 Roma (Italy) Fax: (+ 39)06-49913602 E-mail : [email protected] [b] Dr. S. Borocci Dipartimento per la Innovazione nei Sistemi Biologici Agroalimentari e Forestali (DIBAF) Universit^ della Tuscia, L.go dell'Universit^, s.n.c. , 01100 Viterbo (Italy) [c] Dr. V. Steinmetz Laboratoire Chimie Physique UMR8000 CNRS, Universit^ Paris Sud 11, Orsay (France) Supporting information for this article is available on the WWW under http://dx.doi.org/10.1002/cplu.201400006. (c) 2014 Blackwell Publishing Ltd. |