Cu(II)-catalyzed reactions in ternary [Cu(AA)(AA – H)]+ complexes (AA = Gly, Ala, Val, Leu, Ile, t-Leu, Phe)

P. Wang, G. Ohanessian and C. Wesdemiotis, Eur. J. Mass Spectrom. 15, 325–335 (2009) received: 2 August 2008 n revised: 6 January 2009 n Accepted: 8 January 2009 n Publication: 6 March 2009 325 EuroPEAN JourNAL of MASS SPEctroMEtry Dedicated to Peter Derrick in recognition of his contributions to mass spectrometry Cu(II)-catalyzed reactions in ternary [Cu(AA)(AA – H)]+ complexes (AA = Gly, Ala, Val, Leu, Ile, t-Leu, Phe) Ping Wang,a,c Gilles Ohanessianb and Chrys Wesdemiotisa,* a Department of chemistry, the university of Akron, Akron, oH 44325-3601, uSA. E-mail: wesdemiotis@uakron.edu b Laboratoire des Mécanismes réactionnels, Département de chimie, Ecole Polytechnique, cNrS, f-91128 Palaiseau cedex, france c current address: the Dow chemical company, 2301 N. Brazosport Blvd, B-1219 freeport, tX 77541-3257, uSA The unimolecular chemistry of [Cu(II)AA(AA – H)]+ complexes, composed of an intact and a deprotonated amino acid (AA) ligand, have been probed in the gas phase by tandem and multistage mass spectrometry in an electrospray ionization quadrupole ion trap mass spectrometer. The amino acids examined include Gly, Ala, Val, Leu, Ile, t-Leu and Phe. Upon collisionally-activated dissociation (CAD), the [Cu(II)AA(AA – H)]+ complexes undergo decarboxylation with simultaneous reduction of Cu(II) to Cu(I); during this process, a radical site is created at the a-carbon of the decarboxylated ligand (H2N1 – • C aH – C b H2 – R; R = side chain substituent). The radical site is able to move along the backbone of the decarboxylated amino acid to form two new radicals (HN1• – C aH2 – C b H2–R and H2N1 – C aH2 – • C b H – R). From the complexes of Gly and t-Leu, only C a and N1 radicals can be formed. The whole radical ligand can be lost to form [Cu(I)AA]+ from these three isomeric radicals. Alternatively, further radical induced dissociations can take place along the backbone of the decarboxylated amino acid ligand to yield [Cu(II)AA(AA – 2H – CO2)]+, [Cu(I)AA(• NH2)]+, [Cu(I)AA(HN = C a H2)]+, or [Cu(I)AA(H2N – C aH = C b H – R′]+ (R′ = partial side chain substituent). The sodiated copper complexes, [Cu(II)(AA – H + Na)(AA – H)]+, show the same fragmentation patterns as their non-sodiated counterparts; sodium ion is retained on the intact amino acid ligand and is not involved in the CAD pathways. The amino groups of both AA units, the carbonyl group of the intact amino acid, and the deprotonated hydroxyl oxygen coordinate Cu(II) in squareplanar fashion. Ab initio calculations indicate that the metal ion facilitates hydrogen atom shuttling between the N1, C a and C b atoms of the decarboxylated amino acid ligand. The dissociations of the decarboxylated radical ions unveil important insight about the so far largely unknown intrinsic chemistry of a-amino acid and peptide radicals, which are implicated as intermediates in numerous pathogenic biological processes. Keywords: Cu–amino acid complexes, a-amino acid radicals, H-atom shuttling, radical-induced fragmentation, reductive decarboxylation Introduction copper ion, an essential species in living systems, is involved in several biological processes, including oxidation, dioxygen Dedicated to Professor Peter J. Derrick for his outstanding contributions in mass spectrometry and to the mass spectrometry community. ISSN: 1469-0667 doi: 10.1255/ejms.987 transport and electron transfer.1–6 It plays a significant role in maintaining normal neurological function, immune system and respiration.1,5 Imbalance of copper in living systems can cause various diseases, including cancer.1,5,6 copper ions interact with proteins when they perform biological functions.2,4 understanding of how copper ions work in biological systems © IM Publications LLP 2009 All rights reserved 326 has been gained by studying their binding interactions with amino acids and peptides in the gas phase, using mass spectrometry methods.7–13 the absence of solvent in these experiments allows for the determination of the intrinsic binding modes of copper to proteins and the intrinsic reactivities of the resulting complexes. Metal–ligand interactions are usually metal specific and depend on the oxidation state of the metal. copper(I) has a closed-shell d 10 configuration and very different chemical properties than the open-shell d9 copper(II). copper(I) prefers tetrahedral, 4,7 trigonal three-co-ordinated, 4,7 or linear geometries. 8,14 copper(II) prefers a square-planar geometry15–25 and is one of the most strongly binding divalent metal ions to organic molecules.6 the major structural difference between the +1 and +2 oxidation states is a slight lengthening of the copper–ligand bonds in the reduced forms.5,9,11 the main fragmentation pathway of the complexes of copper(I) and amino acids in the gas phase involves the elimination of H 2o and co (46 u). 13,26 According to theory, the mechanism of lowest critical energy leading to these products proceeds via successive insertions of cu(I) into the c–c and c–oH bonds. on the other hand, many complexes of cu(II) and amino acids / peptides produce radical cations and closed-shell fragment ions that result from radical reactions.15,21–25,27–29 Hu et al. studied the gas-phase co-ordination properties of divalent metal ions with His-containing peptides.21 cu(II) complexes of EfVyVHPV were found to lose co2 from the c-terminal carboxylate group. It was proposed that the hydrogen atom of the c-terminal cooH group transferred to cu(II), thus enabling concomitant co 2 loss to create a radical site at the a-carbon of the c-terminal residue; this radical induced further fragmentation at the side chain. Gatlin et al. 22 and Seymour and tureč ek 23 reported that copper(II) complexes of a deprotonated amino acid and 2,2′-bi pyridine underwent reductive decarboxylation, also forming a radical at the a-carbon which caused further fragmentation at the amino acid side chain; the 2,2′-bipyridine ligand remained intact during these reactions. As will be shown here, similar reactions can take place in ternary cu(II) complexes containing a deprotonated and an intact amino acid (or peptide) ligand. Amino-acid derived radicals with the unpaired electron at the a-carbon atom have been proposed as intermediates in the oxidative damage of amino acids and proteins. 30 Here, we introduce a novel method for forming such a-amino acid radicals, viz. cu(II)-catalyzed decarboxylation of ternary, singly charged complexes [cu(II)AA(AA – H)]+, containing cu(II) ions bound to two amino acid ligands, one in neutral form, AA, and one in deprotonated form, (AA – H)–. the unimolecular chemistry of the copper ion bound radicals produced this way is subsequently probed by collisionally activated dissociation (cAD) and multi-stage mass spectrometry (MS2 and MS3).31 reaction mechanisms are proposed based on the cAD results and corroborated by computational modeling. Cu(II)-Catalyzed Reactions in Ternary [Cu(AA)(AA – H)]+ Complexes Experimental the copper(II) bound dimer complexes [cu(II)AA(AA – H)] + (AA = Gly, Ala, Val, Leu, Ile, t-Leu, and Phe) were formed in the gas phase by electrospray ionization and their consecutive fragmentations were examined by cAD in a quadrupole ion trap mass spectrometer (Bruker Esquire-Lc, Billerica, MA, uSA). the solution sprayed contained 10 –4 M amino acid and 2.5 × 10–5 M cucl2 in a water : methanol (50 : 50 by volume) mixture; ammonium hydroxide was added to obtain final solutions containing 2% NH4oH (by volume). the final solution was introduced into the ion source by a syringe pump at a rate of 100–300 µL h–1. the spraying needle was grounded and the entrance of the sampling capillary was set at –4 kV. Nitrogen was used as the nebulizing gas (1–10 psi) and drying gas (10 L min–1, 150°c). He was used as the buffer gas in the ion trap. for cAD (MS2), the monoisotopic cu(II) bound dimer precursor ions were isolated and excited to fragment with a radio frequency (rf) that was resonant with their frequency of motion in the trap; the kinetic energies of the ions increase during this excitation, causing cAD with the He buffer gas. the excitation time was set at 40 ms and the rf amplitude (Vp–p) at 0.30–0.5 V. the isolation and cAD steps were repeated with specific fragments from the cu(II) bound dimer ions for the acquisition of MS3 spectra. thirty scans per spectrum were collected and the experiments were repeated at least three times. the solvents (HPLc grade) and cucl2 were purchased from Sigma–Aldrich and fisher, respectively. Ala, Leu and Val were acquired from Sigma, Ile from Eastman and Gly, t-Leu, Phe and NH4oH from Aldrich. All chemicals were used without further purification. the computations were carried out with the Gaussian 03 suite of programs. 32 Geometry optimization of the cu(II)amino acid complexes and neutral ligands and vibrational frequency analysis were performed at the B3LyP/6-31G(d) level. final energetics were obtained at the B3LyP/6311+G(2d,2p) level with use of the B3LyP/6-31G(d) geometries. All values given in the text and in figure 3 refer to relative enthalpies at 298 K. Results and discussion Cu(II)-bound dimer with Ala ligands A solution of cucl2 and Ala was injected into the electrospray source. the resulting mass spectrum is shown in figure 1. In this spectrum, the peaks at m/z 90 and m/z 179 originate from protonated alanine and the proton-bound Ala dimer, respectively, and m/z 112 originates from sodiated alanine. the peak at m/z 240 corresponds to [cu(II)Ala(Ala – H)]+, m/z 262 to [cu(II) (Ala – H + Na)(Ala – H)]+ and m/z 196 represents a fragment from the [cu(II)Ala(Ala – H)]+ complex (vide infra). copper-bound complexes dominate the spectrum. cu-containing ions are easily identified, because they appear as doublets originating from incorporation of either 63cu or 65cu (~3 : 1 abundance ratio). the ion at m/z 240, [cu(II)Ala(Ala – H)]+, was selected for cAD, P. Wang, G. Ohanessian and C. Wesdemiotis, Eur. J. Mass Spectrom. 15, 325–335 (2009) 262 196 327 196 - 44 u 240 (a) - 28 u 90 - 15 u 240 - 44 u 152 168 181 220 180 140 179 m/z 196 152 112 - H• (b) 100 m/z 220 160 Figure 1. ESI mass spectrum of a methanol / water solution containing Ala, CuCl2, and ammonium hydroxide (see Experimental). The peak at m/z 240 corresponds to [Cu(II) Ala(Ala – H)]+, m/z 262 to [Cu(II)(Ala – H + Na)(Ala – H)]+, and m/z 196 represents a fragment from the [Cu(II)Ala(Ala – H)]+ complex (see text). The peaks at m/z 90, 112, 152, and 179 correspond to [Ala + H]+, [Ala + Na]+, [Ala + Cu]+ and [2Ala + H]+, respectively. 152 140 Cu NH2 181 200 170 m/z + Figure 2. (a) CAD mass spectrum of [Cu(II)Ala(Ala – H)] (m/z 240) and (b) MS3 mass spectrum of the m/z 196 fragment formed from [Cu(II)Ala(Ala – H)]+. m/z 181, m/z 168 and m/z 152. the MS3 spectrum of m/z 196 is shown in figure 2(b). It shows the same fragmentation pattern as that obtained in the MS2 experiment. which led to the spectrum shown in figure 2(a). the major fragment in this spectrum appears at m/z 196 and arises from the loss of 44 u from the parent ion. there are also peaks at m/z 195, Ala 168 - C2H4 Ala -H H2 N Cu CH2 C H2 Cβ m/z 168 Ala H2 N Cu CH2 C H m/z 195 rH HO O II H2 N α Cu α N H2 O - CO2 I Ala Cu H2 N Cα O CH3 -H Ala Cu C H C H -H rH Ala Cu Ala m/z 152 H N Cu N1 Ala Cu CH3 m/z 195 m/z 196 m/z 240 H N CH3 C H2 - CH3 H N CH2 m/z 181 Scheme 1. Dissociation pathways of [Cu(II)Ala(Ala – H)]+ (m/z 240); the acronym rH indicates a hydrogen atom rearrangement. 328 cu(II) complexes prefer square-planar geometries, as shown in Scheme 1 for [cu(II)Ala(Ala – H)]+ (m/z 240). the intact alanine ligand binds cu(II) with its amino and carbonyl groups, while the deprotonated alanine ligand binds cu(II) with its amino and carboxylate groups. 24 Scheme 1 also provides plausible pathways to the observed fragments. fragmentation begins with reductive decarboxylation; during this process, co2 (44 u) is lost from the deprotonated Ala, as cu(II) is reduced to cu(I) and a radical site is formed at the a-carbon of the deprotonated Ala, viz. ca (m/z 196). two isomeric radicals, N 1 (radical site on the amino nitrogen) and cb (radical site on the b-carbon), can be formed from ca by hydrogen rearrangement. two types of consecutive decompositions can take place from each of these radicals. one reaction is detachment of the entire partially reacted alanine ligand (i.e. H 2 N – • cH – cH 3 , H 2 N – cH 2 – • cH 2 , or HN • – cH 2 – cH 3) from the dimeric complex to form monomeric Ala – cu+ (m/z 152). the other reaction is an elimination via either a- or b-bond scission in the partially reacted ligand. through such processes, a hydrogen radical (from ca, N1 and cb), a cH3 radical (from N1), and a c2H4 molecule (from cb) can be eliminated to form the products at m/z 195, m/z 181 and m/z 168, respectively. these latter products can dissociate further to form Ala – cu+ (m/z 152). the striking similarity of the MS2 and MS3 spectra in figure 2 corroborates that the fragments at m/z ≤ 195 arise sequentially after reductive decarboxylation. figures 3(a) and 3(b) show the calculated energy profiles for several fragmentation pathways of the copper-bound Ala dimer complex [cu(II)Ala(Ala – H)]+. the parent ion [cu(II) Ala(Ala – H)]+ is the lowest energy ion (0 kJ mol–1 in the relative energy scale). formation of the initial reductive decarboxylation product, ca, carrying the radical site at the a–c of the decarboxylated ligand, is endothermic by only 20 kJ mol –1. this step is relatively easy, requiring an activation enthalpy of 119 kJ mol–1, which accounts for the high intensity of m/z 196 in the cAD spectrum of figure 2(a). As shown in figure 3(a), ca can rearrange to cb via direct (one-step) 1,2-hydrogen transfer or stepwise. In the stepwise c a ® c b rearrangement, a b–H atom first moves to cu(I) ion and subsequently to the a-c atom of the reacting ligand. the transition state (tS) of the first step lies at 162 kJ mol–1 and leads to a cu(I)hydride / enamine complex, which is 90 – 20 = 70 kJ mol –1 less stable than c a; the tS of the second step is located at a slightly higher energy, 169 kJ mol –1, and leads to the c b radical, which is 42 – 20 = 22 kJ mol –1 less stable than ca. the one-step 1,2-H rearrangement, ca ® cb, requires 203 kJ mol–1, i.e. 34 kJ mol–1 more energy than the stepwise reaction and, thus, is less competitive. the calculations show that cu(I) ion facilitates nominal 1,2-H shifts by shuttling the H atom between the heavy atoms. According to the potential energy diagram of figure 3(a), cb dissociates spontaneously (via b-bond scission) to Ala – cu+ – •NH2 + c2H4 because it is formed well above its dissociation threshold. on the other hand, formation of Ala – cu + (m/z 152) by further elimination of •NH2 is computed to be much too demanding to be Cu(II)-Catalyzed Reactions in Ternary [Cu(AA)(AA – H)]+ Complexes competitive. A significantly more favorable route to m/z 152 is direct elimination of H2N – •cH – cH3 from intermediate ca (see below). As shown in figure 3(b), the rearrangement ca ® N1 can proceed in a fashion analogous to ca ® cb, via an intermediate cu(I)-hydride / imine complex formed by migration of an N–H hydrogen to cu(I) ion and subsequent transfer of this H atom to the a-c atom of the decarboxylated ligand; these steps have calculated activation enthalpies of 135 kJ mol–1 and 139 kJ mol –1, respectively. the cu(I)-hydride intermediate in this case lies 43 – 20 = 23 kJ mol –1 above c a. Here again, the stepwise ca ® N1 pathway is found to be energetically more favorable than the direct 1,2-H rearrangement. It is noteworthy that the Ala – cu+(H) – NH2 – cH=cH2 complex is less stable than the Ala – cu+(H) – NH = cH – cH3 complex (by 90 – 43 = 47 kJ mol–1), indicating that an imine ligand binds more strongly to cu(I) than an isomeric enamine ligand. this also applies to minimum N1, which is more stable than its ca and cb isomers by 20 kJ mol–1 and 42 kJ mol–1, respectively. these differences plausibly explain why the activation enthalpies for the ca ® N1 isomerization are computed to be smaller than those for the ca ® cb process by ca 30 kJ mol–1. from minimum N1, elimination of •cH3 to form Ala – cu+ – NH=cH2 is expected to be spontaneous, because N1 is formed above its dissociation threshold. the calculations predict that the direct elimination of the decarboxylated ligand (H 2N – •cH – cH 3) from c α requires 158 kJ mol–1. the resulting product, Ala–cu+, lies 178 kJ mol–1 higher in energy than the dimeric precursor ion [cf. figure 3(a)]. Hence, this direct elimination is competitive with the interconversion of ca and cb via H migration to copper, consistent with the comparable relative intensities of the fragments generated by the direct elimination and by c2H4 loss from cβ (at m/z 152 and m/z 168, respectively, cf. figure 2 and Scheme 1). the main fragmentation channel after reductive decarboxylation is elimination of an H• radical (cf. figure 2). this reaction may occur from ca, cb, and/or N1, yielding one of two possible isomers in which the newly formed ligand is either an amine or an imine. As discussed above, cu(I) is more strongly bound to imines than to enamines and this should also apply to the products of H• elimination. Indeed, our calculations show that the complex Ala – cu+ – NH = cH – cH3 is more stable than Ala – cu+ – NH2 – cH = cH2 by 57 kJ mol–1. this makes H• elimination one of the energetically most favorable processes in the fragmentation of the decarboxylated parent dimer (see m/z 195 in figure 3). It is instructive to summarize the comparison of experimental and computational results for the fragmentations of [cu(II)Ala(Ala – H)]+. reductive decarboxylation leads to an intense peak in the cAD mass spectrum, for which computations indicate an activation enthalpy of 119 kJ mol–1. from this intermediate, four fragment ions are observed, viz. the ions at m/z 195, m/z 181, m/z 168 and m/z 152, which are associated with overall activation enthalpies of 127 kJ mol–1, 139 kJ mol–1, 169 kJ mol–1 and 178 kJ mol–1, respectively. the lowest demand P. Wang, G. Ohanessian and C. Wesdemiotis, Eur. J. Mass Spectrom. 15, 325–335 (2009) 329 293 m/z 152 (a) AlaCu + CO2 + C2H4 + 203 m/z 152 AlaCu + NH2 CH3 178 H2N 169 162 + H CO2 119 90 80 AlaCu H H2 N H AlaCu + CO2 42 20 AlaCu H2 N AlaCu H2 N 0 CH3 + CO2 AlaCu H m/z 168 CH2 + CO2 NH2 + CO2 + C2H4 H2 N CH2 H H Cβ m/z 196 Cα m/z 196 O O m/z 240 (b) 167 m/z 195 AlaCu H N + CO2 + H 135 127 H N AlaCu 139 + CO2 + CH3 CH2 106 CH3 m/z 181 H 43 20 AlaCu H2 N + CO2 m/z 196 CH3 H Cα AlaCu H H N CH3 0 + H CO2 H N AlaCu + CO2 CH3 H H N1 m/z 196 Figure 3. Calculated energy profile for the fragmentations of [Cu(II)Ala(Ala – H)]+ (m/z 240) via the charged radical intermediates (a) Ca and Cb and (b) Ca and N1. The numbers in boldface give relative energies in kJ mol–1. Cu(II)-Catalyzed Reactions in Ternary [Cu(AA)(AA – H)]+ Complexes 330 is for elimination of H•, which is consistent with m/z 195 being the most intense ion in the cAD spectrum. out of the three minor cAD fragments, m/z 152 is most intense but has the largest computed activation enthalpy. It is, however, formed from cα via a simple bond breakage and, therefore, should be kinetically favored over the formation of m/z 168 and m/z 181, which both require passing two transition states [cf. figures 3(a) and 3(b), respectively]. overall, the agreement between experimental and computational results is very good, reinforcing the mechanistic scenario detailed above. this may then be extended to the fragmentations of copper-bound dimers of other amino acids. Cu(II)-bound dimer with Gly ligands figure 4(a) depicts the cAD spectrum of [cu(II)Gly(Gly – H)]+ (m/z 212). After reductive decarboxylation (m/z 168), the only major consecutive fragmentation is a 30-u loss to form m/z 138. there is almost no hydrogen radical loss from the decarboxylated product, which shows that the loss of the whole side chain from the partially reacted amino acid (i.e. of •cH2NH2) is a more competitive process. the sodiated product [cu(II) (Gly – H + Na)(Gly – H)] + (m/z 234) was also selected for a tandem MS study; the resulting cAD spectrum is shown in figure 4(b) and is completely analogous to that of figure 4(a) except for the mass shifts due to the exchange of the acidic proton in the intact glycine ligand with a sodium ion. this finding indicates that the originally intact amino acid ligand in [cu(II)AA(AA – H)]+ complexes with unfunctionalized AAs does not participate in the decarboxylation and consecutive radical induced dissociation processes and that the salt bridge formed by H+ / Na+ replacement does not influence the fragmentation pattern of the parent ion, which is dominated by the radical site reactions. Cu(II)-bound dimer with Leu, Ile, or t-Leu ligands the copper–Leu and copper–Ile complexes behave the same way as the copper–Ala complex under the experimental conditions used [figure 5(a) and 5(b)]. Leu and Ile could be differentiated by the fragmentation patterns of their [cu(II)Bpy(Xle – H)]+ complexes (Bpy: bipyridine; Xle: Leu or Ile). 22,23 these differences are also observed here for the corresponding [cu(II)Xle(Xle – H)] + complexes. the ca radicals emerging after reductive decarboxylation undergo loss of 43 u (cH3•cHcH3) if Xle = Leu and loss of 29 u (•cH2cH3) if Xle = Ile. there is no noticeable amount of 15-u loss (•cH3) from Ile, which is probably due do the fact that the reaction pathway releasing the larger radical (•cH2cH3) is kinetically favored over the reaction pathway releasing the smaller, less stable radical (•cH3).33 the dissociations of the t-Leu–copper complex under cAD conditions are shown in figure 5(c). the major difference between the dissociation pathways of the t-Leu and Leu / Ile complexes is the significant amount of • cH 3 loss (–15 u, m/z 265) from the t-Leu complex. Since the b-c atom of t-Leu does not carry a hydrogen atom, a cb radical cannot be formed from the cu(II) complex of t-Leu. As a result, there is no m/z 210 peak in figure 5(c). In contrast, there is such a peak (minor) in both figure 5(a) and 5(b), because the Leu / Ile complexes are able to form a cβ radical, which can decompose to Xle – cu+ – •NH2 (m/z 210) + c5H10 168 (a) - 30 u 212 - 44 u 138 120 160 200 m/z 190 (b) 160 - 30 u 234 - 44 u 150 180 + 210 m/z + Figure 4. CAD mass spectra of (a) [Cu(II)Gly(Gly – H)] (m/z 212) and (b) [Cu(II)(Gly – H + Na)(Gly – H)] (m/z 234). P. Wang, G. Ohanessian and C. Wesdemiotis, Eur. J. Mass Spectrom. 15, 325–335 (2009) 331 280 (a) NH2 324 - 57 u COOH - 44 u 237 - 43 u 223 194 200 250 300 m/z 280 (b) NH2 - 57 u COOH 324 - 29 u 223 - 44 u 251 194 250 200 m/z 300 280 NH2 COOH (c) - 86 u 223 324 - 57 u - 15 u 265 194 200 - 44 u 250 300 m/z Figure 5. CAD mass spectra of (a) [Cu(II)Leu(Leu – H)]+, (b) [Cu(II)Ile(Ile – H)]+ and (c) [Cu(II)t-Leu(t-Leu – H)]+ (all at m/z 324). The dashed arrows point at the m/z 210 fragment (minor), which is only observed from the Leu and Ile complexes. via b-bond scission (cf. Scheme 1). there are two common products for all three complexes. the first arises from loss of the entire decarboxylated ligand (86 u), i.e. of the ligand with the unpaired electron, and generates m/z 194. the second results from elimination of the side chain of the decarboxylated Leu, Ile, or t-Leu ligand, which is cleaved via N1 radical intermediates and creates fragment ions with a NH = cH2 unit attached to cu(I) (at m/z 223). overall, the fragments generated from the ca radical provide the most valuable structure information about the identity of the AA bound to copper(II). the sodiated copper complexes [cu(II)(AA – H + Na)(AA – H)]+ have also been studied and showed identical fragmentation patterns to those of their non-sodiated counterparts (spectra not shown); this further confirms that sodium ion is retained on the intact amino acid ligand and is not involved in cAD reactions and that the latter reactions are radical-induced and, hence, insensitive to the introduction of a coo–Na+ salt bridge. the salt bridge would have influenced the outcome of charge-induced dissociations. Cu(II)-bound dimer with Val or Phe ligands the cAD spectrum of [cu(II)Val(Val – H)]+, figure 6(a), shows the occurrence of reductive decarboxylation and subsequent losses of 15 u (m/z 237, loss of •cH3 radical from ca), 43 u (m/z 209, loss of cH3•cHcH3 radical from N1), 56 u (m/z 196, loss of cH2 = c(cH3)2 from cb), and 72 u which is the loss of the entire decarboxylated Val ligand (m/z 180). thus, the reaction pathways of [cu(II)Val(Val – H)]+ are completely analogous to those of [cu(II)Ala(Ala-H)]+. for [cu(II)Phe(Phe – H)] +, cf. figure 6(b), there is no c a radical product detected (it would have appeared at m/z 271), and the product from the corresponding N 1 radical cation (m/z 257 according to the mechanism of Scheme 1) is minuscule. Here, the dominant fragment, m/z 244, is generated from the cb radical cation. the isomerization ca ® cb involves transfer of a benzylic hydrogen atom and produces a resonance-stabilized benzylic radical,34 therefore, the formation of cb is both kinetically and thermodynamically favored, explaining the high relative abundance of m/z 244 which is Cu(II)-Catalyzed Reactions in Ternary [Cu(AA)(AA – H)]+ Complexes 332 252 NH2 (a) - 72 u - 56 u - 43 u COOH 296 - 15 u 180 237 196 209 180 - 44 u 220 348 244 NH2 392 (b) - 44 u - 104 u COOH m/z 260 - 16 u 228 257 220 300 380 m/z 344 (c) 300 - 104 u - 72 u - 44 u 196 228 175 250 + 325 m/z + Figure 6. CAD mass spectra of (a) [Cu(II)Val(Val – H)] (m/z 296), (b) [Cu(II)Phe(Phe – H)] (m/z 392) and (c) [Cu(II)Phe(Val – H)]+ or [Cu(II) Val(Phe – H)]+ (m/z 344). formed by elimination of a styrene molecule (104 u) from cβ. –72 u in figure 6(c), while the corresponding homodimeric Experiments with the mixed dimer complex cu(II)–Phe / Val complex preferably loses only the side chain of the decarboxywere also carried out and the cAD spectrum is shown in lated ligand (–56 u in figure 6(a)). this change is attributed figure 6(c). After the reductive decarboxylation, the major to differences in the relative dissociation energetics of homofragments are loss of 72 u, which proceeds from a decar- and heterodimeric complexes. boxylated Val ligand; and the loss of styrene (104 u), which Structure/fragmentation pattern proceeds from the cb radical resulting from a deprotonated correlations Phe ligand. the higher relative abundance of the 104-u loss vis à vis the 72-u loss suggests that the major component of the table 1 summarizes the fragments observed in the cAD spectra isomeric mixture is [cu(II)Val(Phe – H)]+, in agreement with the of [cu(II)AA(AA – H)]+ complexes. Loss of co2 is the major frag35 higher acidity of Phe vis à vis Val both in solution as well as mentation pathway for all amino acid ligands examined in the gas phase.36,37 the spectrum of the mixed complex, figure this study. Hydrogen atom elimination after decarboxylation 6(c), shows that styrene loss is the decomposition preferred proceeds most efficiently from the Ala and Val complexes; by the deprotonated Phe ligand (after initial co2 expulsion), H• loss is the only b bond scission possible from ca of Ala (cf. as was the case for the homodimeric complex [figure 6(b)]. Scheme 1) and involves cleavage of a tertiary hydrogen atom on the other hand, the mixed complex portion that contains from ca of Val, explaining the increased yield of this reaction a deprotonated Val, rather loses the decarboxylated ligand, from the Ala and Val complexes. the product from b c–c bond P. Wang, G. Ohanessian and C. Wesdemiotis, Eur. J. Mass Spectrom. 15, 325–335 (2009) 333 Table 1. Fragments generated upon CAD of [63Cu(II)AA(AA – H)]+ complexes.a AA Complex Loss of CO2 Loss of H • b Fragment from C α c (loss) Fragment from C β d (loss) Fragment from N1 e (loss) AA–Cu+ f f 138 m/z Gly 212 168 167 f Ala 240 196 195 f Val 296 252 251 Leu 324 280 Ile 324 t-Leu Phe 168 (c2H4) 181 (cH3 • ) 152 237 (cH3 • ) 196 (i-c 4H8) 209 (i-c 3H7• ) 180 279 237 (i-c 3H7• ) 210 (c5H10) g 223 (i-c 4H9• ) 194 280 279 251 (c2H5 • ) 210 (c5H10) h 223 (s-c 4H9• ) 194 324 280 279 265 (cH3 • ) f 223 (t-c 4H9• ) 194 392 348 347 f 257 (c7H7• ) j 228 244 (c 8H8) i the m/z ratios of minor products are italicized; bfrom c a , c b , or N1; c[cu(I)AA(H2NcH = cHr ¢] + (r′ = partial side chain substituent), formed via b c–c bond scission in c a and loss of the radical given in parenthesis; d[cu(I)AA( • NH2)] +, formed by b c–N bond scission in c b and loss of the alkene given in parenthesis; e [cu(I)AA(HN = cH2] +, formed b c–c bond scission in N1 and loss of the radical given in parenthesis; fnot possible from these complexes; g3-methyl-1-butene. h 2-methyl-1-butene; istyrene; jbenzyl radical a scission in the ca radical has moderate intensity, with the exception of the Phe complex which does not undergo detectably this reaction, probably because it would have released a high energy phenyl radical.38 for the complexes, in which b c–c bond scission from ca takes place, the abundance of the resulting product increases with increasing size of the detached radical: loss of i-c3H7• (from the Leu complex) > loss of c2H5• (from the Ile complex) > loss of cH3• (from t-Leu and Ala complexes).33 the product from b c–N bond scission in the cb radical has low intensity except for Phe ligands (table 1). As reasoned above, the high yield of this reaction from the Phe complex must result from the transfer of a relatively mobile benzylic hydrogen atom in the rearrangement ca ® cb that precedes b c–N bond scission. All dimeric complexes, which can undergo b c–c bond scission from the N1 radical, do so (table 1). the relative abundance of the corresponding product depends on the ligand: loss of c4H9• from the Xle complexes > loss of c3H7• from the Val complex > loss of cH3• from the Ala complex ≈ loss of c7H7• from the Phe complex; based on this trend, the extent of the rearrangement c α ® N 1 decreases in the direction Xle > Val > Ala ≈ Phe ligand. Conclusions the dissociations of [cu(II)AA(AA – H)]+ complexes (AA = Gly, Ala, Val, Leu, Ile, t-Leu and Phe) have been examined by cAD experiments in an electrospray ionization mass spectrometer. In the single-stage mass spectra, cu(II)-cationized complexes of the amino acid dominate, but the protonated and sodiated amino acids are present as well. In the complexes, the amino groups of both amino acids, the carbonyl group of the intact amino acid, and the deprotonated hydroxyl oxygen coordinate cu(II) in square-planar fashion.23 under cAD conditions, the [cu(II)AA(AA – H)]+ complexes undergo decarboxylation with simultaneous reduction of cu(II) to cu(I); during this process, a radical site is created at the a-carbon of the decarboxylated ligand (H2N1 – •caH – cbH2 – r; r = side chain). the radical site is able to move along the backbone of the decarboxylated amino acid to form two new radicals (HN1•฀– caH2 – cbH2 – r and/or H 2N 1 – c aH 2 – •c bH – r). from the complexes of Gly and t-Leu, only ca and N1 radicals can be formed. the whole radical ligand can be lost to form [cu(I)AA]+ from these three isomeric radicals. Alternatively, further radical-induced dissociations can take place along the backbone of the decarboxylated amino acid (table 1). the sodiated copper complexes ([cu(II)(AA – H + Na)(AA – H)]+) show the same fragmentation patterns as their non-sodiated counterparts; sodium ion and the coo–Na+ salt bridge are not involved in the cAD pathways. on the basis of the cAD spectra, dissociation mechanisms are proposed for the [cu(II)AA(AA – H)]+ complexes and substantiated by the calculated potential energy diagrams for fragmentation of the [cu(II)Ala(Ala – H)]+ complex. the calculations provide information about the hydrogen migration pathways. After decarboxylation, a hydrogen from the amino group or the b-carbon migrates to the copper ion, from where it is shuttled to ca to form radicals at the amino nitrogen (N1) or the b-carbon (cb), respectively. Such pathways require lower 334 activation energies than direct 1,2-H atom rearrangements and may also take place with a-peptide radicals generated in vivo where cu(II) ions are available. Migration from the amino group is significantly more favorable because imines are more strongly bound to the cu ion than amines. for copper complexes with aliphatic amino acids, like the ones in the present study, only the deprotonated amino acids are involved in the dissociation reactions. When functionalized side chains are present, or with larger ligands (dipeptides), interligand reactions start occurring, which will be discussed in a separate, upcoming study.31,39 Acknowledgments We thank Dr Michael J. Polce for helpful comments and the National Science foundation for generous financial support. References 1. M.c. Linder and c.A. Goode (Eds), Biochemistry of Copper. Plenum Press, New york, Ny, uSA (1991). 2. t.G. Spiro (Ed.), “copper Proteins”. John Wiley & Sons, New york, Ny, uSA (1981). 3. c. Kies (Ed.), Copper Bioavailability and Metabolism. Plenum Press, New york, Ny, uSA (1989). 4. S.J. Lippard and J.M. Berg, Principles of Bioinorganic 5. 6. 7. 8. 9. 10. 11. Chemistry. university Science Books, Mill Valley, cA, uSA (1994). D.D. Karlin and Z. tyeklar (Eds), Bioinorganic Chemistry of Copper. chapman & Hall, New york, Ny, uSA (1993). J.J.r. frausto da Silva and r.J.P. Williams, The Biological Chemistry of the Elements, 2nd Edn. oxford university Press, oxford, uK (2001). K. Wang and K. Ma, “Binding of copper(II) ions to the polyproline II helices of PEVK modules of the giant elastic protein titin as revealed by ESI-MS, cD, and NMr”, Biopolymers 70, 297–309 (2003). B.A. cerda and c. Wesdemiotis, “Gas phase copper(I) ion affinities of valine, lysine, and arginine based on the dissociation of cu+ -bound heterodimers at varying internal energies,” Int. J. Mass Spectrom. 185–187, 107–116 (1999). doi: 10.1016/S1387-3806(98)14132-5 r.c. Burnier, t.J. carlin, W.D. reents Jr, r.B. cody, r.K. Lengel and B.S. freiser, “coordination chemistry of copper(I) in the gas phase”, J. Am. Chem. Soc. 101, 7127– 7129 (1979). doi: 10.1021/ja00517a085 N.r. Walker, r.r. Wright, P.E. Barran, J.N. Murrell and A.J. Stace, “comparisons in the behavior of stable copper(II), silver(II), and gold(II) complexes in the gas phase: are there implications for condensed-phase chemistry?”, J. Am. Chem. Soc. 123, 4223–4227 (2001). doi: 10.1021/ja003431q t. Marino, N. russo and M. toscano, “Gas-phase metal ion (Li+, Na+, cu+) affinities of glycine and alanine”, J. Cu(II)-Catalyzed Reactions in Ternary [Cu(AA)(AA – H)]+ Complexes Inorg. Biochem. 79, 179–185 (2000). doi: 10.1016/S01620134(99)00242-1 12. S. Hoyau, J.-P. Pélicier, f. rogalewicz, y. Hoppilliard and G. ohanessian, “complexation of glycine by atomic metal cations in the gas phase”, Eur. J. Mass Spectrom. 7, 303–311 (2001). doi: 10.1255/ejms.440 13. y. Hoppilliard, G. ohanessian and S. Bourcier, “fragmentation mechanisms of glycine-cu+ in the gas phase. An experimental and theoretical study”, J. Phys. Chem. A 108, 9687–9696 (2004). doi: 10.1021/jp047905+ 14. A. Luna, B. Amekraz, J.-P. Morizur, J. tortajada, o. Mo and M. yanez, “reactions between guanidine and cu+ in the gas phase. An experimental and theoretical study”, J. Phys. Chem. A 101, 5931–5941 (1997). doi: 10.1021/ jp9710300 15. W.A. tao, D. Zhang, E.N. Nikolaev and r.G. cooks, “copper(II)-assisted enantiomeric analysis of D,L-amino acids using the kinetic method: chiral recognition and quantification in the gas phase”, J. Am. Chem. Soc. 122, 10598–10609 (2000). doi: 10.1021/ja000127o 16. t. Marino, N. russo and M. toscano, “Interaction of cu+ and cu2+ ions with a-alanine. A density functional study” J. Mass Spectrom. 37, 786–791 (2002). doi: 10.1002/jms.331 17. J. Bertran, L. rodriguez-Santiago and M. Sodupe, “the different nature of bonding in cu+ –glycine and cu2+ – glycine”, J. Phys. Chem. B 103, 2310–2317 (1999). doi: 10.1021/jp984534m 18. A. rimola, E. constantino, L. rodriguez-Santiago and M. Sodupe, “Binding properties of cu+/2+ -(glycyl)nglycine complexes (n = 1–3)”, J. Phys. Chem. A 112, 3444-3453 (2008). doi: 10.1021/jp711189s 19. W.A. tao and r.G. cooks, “Parallel reactions for enantiomeric quantification of peptides by mass spectrometry”, Angew. Chem. Int. Ed. 40, 757–760 (2001). doi: 10.1002/1521-3773(20010216)40:4<757::AIDANIE7570>3.0.co;2-H 20. W.A. tao, D. Zhang, f. Wang, P.D. thomas and r.G. cooks, “Kinetic resolution of D,L-amino acids based on gas-phase dissociation of copper(II) complexes”, Anal. Chem. 71, 4427–4429 (1999). doi: 10.1021/ac990530h 21. P. Hu and J.A. Loo, “Gas-phase coordination properties of Zn2+, cu2+, Ni2+ and co2+ with histidine-containing peptides”, J. Am. Chem. Soc. 117, 11314–11319 (1995). doi: 10.1021/ja00150a032 22. c.L. Gatlin, r.D. rao, f. turecek and t. Vaisar, “copper(II) amino acid complexes in the gas phase”, J. Am. Chem. Soc. 117, 3637–3638 (1995). doi: 10.1021/ja00117a043 23. J.L. Seymour and f. tureček, “Structure, energetics and reactivity of ternary complexes of amino acids with cu(II) and 2,2-bypyridine by density functional theory. A combination of radical-induced and spin-remote fragmentations”, J. Mass Spectrom. 37, 533–540 (2002). doi: 10.1002/jms.311 24. f. tureček, “copper–biomolecule complexes in the gas phase. the ternary way”, Mass Spectrom. Rev. 26, 563– 582 (2007). doi: 10.1002/mas.20137 P. Wang, G. Ohanessian and C. Wesdemiotis, Eur. J. Mass Spectrom. 15, 325–335 (2009) 25. D.r. Brown, V. Guantieri, G. Grasso, G. Impellizzeri, G. Pappalardo and E. rizzarelli, “copper(II) complexes of peptide fragments of the prion protein. conformation changes induced by copper(II) and the binding motif in c-terminal protein region”, J. Inorg. Biochem. 98, 133– 143 (2004). doi: 10.1016/j.jinorgbio.2003.09.006 26. H. Lavanant and y. Hoppilliard, “formation and fragmentation of a-amino acids complexed by cu+”, J. Mass Spectrom. 32, 1037–1049 (1997). doi: 10.1002/(SIcI)10969888(199711)32:10<1037::AID-JMS556>3.0.co;2-L 27. E. Bagheri-Majdi, y. Ke, G. orlova, I.K. chu, A.c. Hopkinson and K.W.M. Siu, “copper-mediated peptide radical ions in the gas phase”, J. Phys. Chem. B 108, 11170–11181 (2004). doi: 10.1021/jp049531q 28. y. Ke, J. Zhao, u.H. Verkerk, A.c. Hopkinson and K.W.M. Siu, “Histidine, lysine, and arginine radical cations: isomer control via the choice of auxiliary ligand (L) in the dissociation of [cuII(L)(amino acid)] •2+ complexes”, J. Phys. Chem. B 111, 14318–14328 (2007). doi: 10.1021/ jp0746648 29. I.K. chu, c.f. rodriguez, A.c. Hopkinson, K.W.M. Siu and t.-c. Lau, “formation of molecular radical cations of enkephalin derivatives via collision-induced dissociation of electrospray-generated copper (II) complex ions of amines and peptides”, J. Am. Soc. Mass Spectrom. 12, 1114–1119 (2001). doi: 10.1016/S1044-0305(01)00297-5 30. c.J. Easton, “free-radical reactions in the synthesis of a-amino acids and derivatives”, Chem. Rev. 97, 53–82 (1997). doi: 10.1021/cr9402844 31. P. Wang, “Interactions of peptides with simple Lewis acids and fragmentation mechanisms of adducts studied by tandem mass spectrometry”, Ph.D. Dissertation, the university of Akron, (2005). 32. M.J. frisch, G.W. trucks, H.B. Schlegel, G.E. Scuseria, M.A. robb, J.r. cheeseman, J.A. Montgomery Jr, t. Vreven, K.N. Kudin, J.c. Burant, J.M. Millam, S.S. Iyengar, J. tomasi, V. Barone, B. Mennucci, M. cossi, G. Scalmani, N. rega, G.A. Petersson, H. Nakatsuji, M. Hada, M. Ehara, K. toyota, r. fukuda, J. Hasegawa, M. Ishida, t. Nakajima, y. Honda, o. Kitao, H. Nakai, M. 335 Klene, X. Li, J.E. Knox, H.P. Hratchian, J.B. cross, c. Adamo, J. Jaramillo, r. Gomperts, r.E. Stratmann, o. yazyev, A.J. Austin, r. cammi, c. Pomelli, J.W. ochterski, P.y. Ayala, K. Morokuma, G.A. Voth, P. Salvador, J.J. Dannenberg, V.G. Zakrzewski, S. Dapprich, A.D. Daniels, M. Strain, o. farkas, D.K. Malick, A.D. rabuck, K. raghavachari, J.B. foresman, J.V. ortiz, Q. cui, A.G. Baboul, S. clifford, J. cioslowski, B.B. Stefanov, G. Liu, A. Liashenko, P. Piskorz, I. Komaromi, r.L. Martin, D.J. fox, t. Keith, M.A. Al-Laham, c.y. Peng, A. Nanayakkara, M. challacombe, P.M.W. Gill, B. Johnson, W. chen, M.W. Wong, c. Gonzalez and J.A. Pople, Gaussian 03, rev. B.05. Gaussian, Inc., Pittsburgh, PA, uSA (2003). 33. f.W. McLafferty and f. turecek, Interpretation of Mass Spectra, 4th Edn. university Science Books, Mill Valley, cA, uSA, pp. 53, 60–64 and 169 (1993). 34. P.J. Lindstrom and W.G. Mallard (Eds), NIST Standard Reference Database, Number 69. National Institutes of Standards and technology, Gaithersburg, MD, uSA (2003); (http://webbook.nist.gov). 35. David r. Lide (Ed.), Handbook of Chemistry and Physics, 71st Edn. crc Press, Boca raton, fL, uSA, Section 7, p. 1 (1990). 36. r.A.J. o’Hair, J.H. Bowie and S. Gronert, “Gas phase acidities of the a amino acids”, Int. J. Mass Spectrom. Ion Processes 117, 23–36 (1992). doi: 10.1016/01681176(92)80083-D 37. c.M. Jones, M. Bernier, E. carson, K.E. colyer, r. Metz, A. Pawlow, E.D. Wischow, I. Webb, E.J. Andriole and J.c. Poutsma, “Gas-phase acidities of the 20 protein amino acids”, Int. J. Mass Spectrom. 267, 54–62 (2007). doi: 10.1016/j.ijms.2007.02.018 38. D.f. McMillen and D.M. Golden, “Hydrocarbon bond dissociation energies”, Ann. Rev. Phys. Chem. 33, 493–532 (1992) doi: 10.1146/annurev. pc.33.100182.002425 39. P. Wang and c. Wesdemiotis, “cu (II)-catalyzed reactions in complexes with functionalized amino acids (trp, tyr, His, Arg, Lys) and dipeptide (GG) ligands”, manuscript in preparation.