- Email: [email protected]
Zinc enzyme modelling with zinc complexes of polar pyrazolylborate ligands
Inorganica Chimica Acta 360 (2007) 1510–1516 www.elsevier.com/locate/ica
Zinc enzyme modelling with zinc complexes of polar pyrazolylborate ligands C. Pe´rez Olmo, K. Bo¨hmerle, H. Vahrenkamp
*
Institut fu¨r Anorganische und Analytische Chemie der Universita¨t Freiburg, Albertstr. 21, D-79104 Freiburg, Germany Received 16 July 2006; accepted 15 August 2006 Available online 26 August 2006
Abstract The Zn–OH2 and Zn–OH complexes of the new tris(pyrazolyl)borate ligands with pyridyl and carboxamido substituents were investigated for their reactivity towards hydrolyzeable substrates. Tp4Py,MeZn–OH inserted CO2 and CS2 in methanol forming the Zn–OCOOMe and Zn–SCSOMe products. In non-aqueous media, both types of complexes with both types of substituents on the Tp ligands effected stoichiometric cleavage of tris(p-nitrophenyl)phosphate and p-nitrophenyl acetate. In solutions containing water and the MOPS buffer, up to eight p-nitrophenyl groups per equivalent of zinc complex could be cleaved from the esters, and the resulting bis(p-nitrophenyl)phosphate was also degraded to mono(p-nitrophenyl)phosphate. This is the first time that pyrazolylborate–zinc complexes have shown catalytic activity in hydrolytic reactions. Ó 2006 Elsevier B.V. All rights reserved. Keywords: Bioinorganic chemistry; CO2 activation; Ester cleavage; Pyrazolylborate ligands; Zinc
1. Introduction The hydrolysis of esters and phosphates, particularly tris(p-nitrophenyl)phosphate, is the most thoroughly studied reaction in the modelling of zinc enzymes with coordination compounds of zinc. The review articles bear witness to this [1–5], the leading researchers in the field have worked on it [6–12], and there is continued activity until today [13–18]. We too have worked in this area [19–24], and we have derived a mechanistic pathway. [25,26]. During these studies, like in other fields of enzyme modelling, it became evident that the more well-defined the model complexes were the lower was their catalytic efficiency. This holds particularly for those cases where fouror five-coordinate Zn–OH2 or Zn–OH complexes could actually be isolated and used as enzyme models. Thereby the best structural models, typically (tripod) Zn–X species reproducing the tetrahedral environment of zinc in the *
Corresponding author. Tel.: +49 761 203 6120; fax: +49 761 203 6001. E-mail address: [email protected] (H. Vahrenkamp).
0020-1693/$ - see front matter Ó 2006 Elsevier B.V. All rights reserved. doi:10.1016/j.ica.2006.08.021
enzymes [27], were often useful only for stoichiometric, but rarely for catalytic hydrolyses. We experienced this with our pyrazolylborate–zinchydroxide complexes, which were the first structurally correct and functional models of the active state of hydrolytic zinc enzymes [28]. They could be used in a stoichiometric fashion for all types of hydrolyses [19,20,24,25]. But their application for catalyses was prevented by product inhibition. Typically, both products of the hydrolysis of a triarylphosphate, i.e. the phenolate and the diarylphosphate, ended up in stable TpZn–phenolate and TpZn–phosphate complexes. We assume that the major reason for this catalytic inability is the non-polar nature of the hitherto employed pyrazolylborates and their zinc complexes. The synthetic procedure for the Tp ligands, i.e. heating a pyrazole and KBH4 up to 200 °C, has limited the substituents allowed on the pyrazoles almost entirely to hydrocarbon groups. As a result the ligands and their complexes have been insoluble in water and water-containing solvents. Consequentially, during the hydrolytic reactions very little water was
C. Pe´rez Olmo et al. / Inorganica Chimica Acta 360 (2007) 1510–1516
available to overcome the product inhibition and to reconstitute the hydrolytically active complexes. The way out of this dilemma had to be the synthesis of water-soluble pyrazolylborates. We, among others (see Ref. [29]), have recently succeeded in this task [30]. In addition, we could convert the new 4-pyridyl- and carboxamido-substituted Tp ligands to Zn–OH2 and Zn–OH complexes [31]. Thereby we had available the desired enzyme models which could be employed in the presence of water. This paper reports our investigations on their stoichiometric and catalytic hydrolyses of esters and organophosphates. The Zn–OH2 and Zn–OH complexes of L1–L3 were used. H
H B N
N
N
N
B N
N
N
N
N
N
N
N
O HN
O
HN N
O
NH
N
N
L2
L1 H B
HN
N
N
N
N
N
N
O HN
O
O
NH
L3
2. Results and discussion The aqua- and hydroxo-complexes of ligands L1–L3 which could be isolated are 1–5 [31]. Of these, the mononuclear hydroxides 2 and 4 are rather labile and cannot be handled in solution for extended periods, for which reason only 2 was used, and for some stoichiometric reactions only. Aqua complexes 1 and 3 and the hydroxide-bridged dinuclear complex 5 persisted in solutions containing water and buffer, and hence were suitable for catalytic reactions. 1
1
½L Zn–OH2 ClO4 L Zn–OH 2 1 2 3 ½L Zn–OH2 ClO4 L Zn–OH 4 3 3 3 ½L Zn–OH–ZnL ClO4 5
1511
that hydroxides 2 and 4 would react accordingly. This was borne out by the hydrolytic reactions of 2. When CO2 was bubbled through a methanolic solution of 2, it could be seen by 1H NMR that all of 2 was consumed after four hours. The isolation of the expected methylcarbonate complex proved difficult, however, being hampered by slow decomposition of the starting material and the product. Thus only a mediocre yield of impure Tp4Py,MeZn–OCOOMe could be obtained by speedy isolation, or careful crystallization produced a very small amount of the desired product in the unusual composition 6. We assume that the chloride in 6 originates from traces of HCl often found in dichloromethane. L1 Zn–OCOOMe L1 ZnCl 6 The composition of 6 was determined by a structure determination and by 1H NMR. Fig. 1 shows the contents of the asymmetric unit in the crystals. 6 is a chain-like coordination polymer with alternating L1Zn–OCOOMe and L1ZnCl units, each of which uses one of its Tp’s pyridyl substituents to coordinate to the next unit in the chain. This way all zinc ions become five-coordinate with a rather distorted trigonal-bipyramidal geometry. The formation of coordination dimers and polymers is very common for zinc complexes of L1–L3 [30,31], and the related complex Tp3Py,MeZn–OCOOMe is a pyridyl bridged dimer [32]. The bond lengths at both zinc ions (see Fig. 1) are normal, including the elongations at the axes of the trigonal bipyramids. Thus the main value of the structure determination of 6 consists in the proof that it contains L1Zn–OCOOMe. The reaction of 2 with CS2 in the presence of methanol proceeded more smoothly and produced the xanthogenate complex 7 in a good yield. The main indicators for the presence of a Zn–SC(S)OMe unit in 7 are the IR band for C@S at 1208 cm1 and the OMe resonance in the 1H NMR spectrum at 3.17 ppm. The NMR spectra of 7 do not point to association by a splitting of the pyridyl resonances. We therefore assume that 7 is monomeric, using the OMe function of the xanthogenate ligand to make zinc five-coordinate, just as was observed in L1Zn–OC(O)Me [30]. L1 Zn–SCðSÞOMe 7 The cleavage of activated esters by 2 could be performed with the p-nitrophenolates PO(ONit)3 and CH3COONit. As expected, the reactions were stoichiometric, and they consumed two equivalents of 2 per equivalent of substrate according to the following equations:
2.1. Stoichiometric reactions
POðONitÞ3 þ 2L1 Zn–OH 2 ! L1 Zn–OPOðONitÞ2 þ L1 Zn–ONit þ H2 O 9 8
After introducing the TpZn–OH complexes as enzyme models [28], we had used them extensively for stoichiometric model reactions [26]. It could be expected, therefore,
CH3 COONit þ 2L1 Zn–OH 2 ! L1 Zn–OCðOÞCH3 þ L1 Zn–ONit þ H2 O 9 10
ð1Þ
ð2Þ
1512
C. Pe´rez Olmo et al. / Inorganica Chimica Acta 360 (2007) 1510–1516
Fig. 1. Contents of the asymmetric unit of compound 6. The linking of the units of 6 in the coordination polymer is indicated by dotted lines. Bond lengths ˚ ): Zn1–O 1.930(3), Zn1–N1 2.070(4), Zn1–N2 2.038(4), Zn1–N3 2.300(4), Zn1–N18 2.192(4), Zn2–Cl 2.285(3), Zn2–N9 2.181(4), Zn2–N10 2.050(4), (A Zn2–N11 2.035(4), Zn2–N12 2.575(4).
The cleavage of the phosphate proceeded about twice as fast as that of the acetate. Following reactions (1) and (2) by 1H NMR indicated that they were quantitative, using the spectroscopic data of their products 8, 9 and 10 which we had prepared previously [30]. The workup of both reactions, however, was cumbersome, and after their separation 8 and 9 from reaction (1) and 9 and 10 from reaction (2) were obtained only in moderate yields. Hoping for amide hydrolysis, complex 2 was also reacted with trifluoroacetamide. However, instead of cleavage deprotonation took place with formation of the amidate complex 11. The identification of 11 rests in its m(CO) IR band at 1679 cm1, which compares well with those of other TpZn–NHC(O)R complexes [33,34] and which rules out the hoped-for formation of the acetate complex L1Zn–OC(O)CF3 for which a m(CO) band around 1710 cm1 would be expected [19,34]. Thus, in contrast to previous propositions by us [19], it must be stated that until now amide cleavage by TpZn–OH complexes could not be achieved yet.
gressively more acidic which in turn resulted in the hydrolytic destruction of the Tp ligands. The substrate hydrolyses by 1 and 3 became considerably faster when water was present. Thus PO(ONit)3 was quantitatively hydrolyzed to HOPO(ONit)2 and HONit by complex 1 after 15 min in DMSO/water (3:1). After this time complex 1 was still intact. This is demonstrated by Fig. 2 which shows only the hydrolysis products and no more PO(ONit)3. The unmarked signals in Fig. 2 belong to complex 1 with its characteristic 1:5 splitting of the pyrazole CH singlet at 6.6 ppm and the two pyridyl doublets near 7.7 and 8.5 ppm [31]. Similar observations were made when reacting 1 with CH3COONit in DMSO/H2O. It took, however, five days to achieve quantitative cleavage to CH3COOH and HONit with one equivalent of 1.
HOP(O)(ONit)2
HONit
L1 Zn–NHCðOÞCF3 11 Stoichiometric hydrolytic cleavages were also possible using the aqua complexes 1 and 3. They were performed in NMR tubes and followed by 1H and 13P NMR. When water-free conditions were applied, just as during the reactions with 2, the cleavages by 1 and 3 were considerably slower. Thus, only two thirds of the PO(ONit)3 were cleaved by 3 in DMSO after six days. NMR revealed (for representative spectra see below) that the cleavage products are free HONit and HOPO(ONit)2 and not their LZn complexes. This implies that the reaction solutions became pro-
9.0 8.8 8.6 8.4 8.2 8.0 7.8 7.6 7.4 7.2 7.0 6.8 6.6 6.4 6.2
[ ppm] Fig. 2. 1H NMR spectrum of the 1:1 reaction mixture of 1 and PO(ONit)3 in DMSO-d6/D2O (3:1) after 15 min. For the unmarked resonances see text.
C. Pe´rez Olmo et al. / Inorganica Chimica Acta 360 (2007) 1510–1516
1513
2.2. Catalytic hydrolyses The cleavages of the esters with water generate p-nitrophenol and the corresponding acids. For the attempts to catalyze these reactions with the zinc complexes therefore water-containing solvents had to be used, and a buffer had to be present in order to retard the hydrolytic selfdestruction of the Tp ligands which is fast in acidic solutions. All reactions were carried out in NMR tubes, those employing complex 1 in DMSO-d6/D2O (3:1), those employing complexes 3 and 5 in CDCl3/CD3OD/D2O (4:10:1). The buffer MOPS (3-morpholinopropanesulfonic acid, pKa = 7.2) was applied in a 5–20-fold excess relative to the zinc complexes. Control experiments showed that under these conditions the uncatalyzed hydrolyses are at least 1–2 orders of magnitude slower than those in the presence of the zinc complexes. All reactions were performed at 310 K and followed by 1H NMR. The cleavages of PO(ONit)3 were also followed by 31P NMR, which in combination with the signal intensities in the 1H NMR spectra ensured the identification of the solution constituents PO(ONit)3 (d = 18.8), [PO2(ONit)2] (d = 11.3) and [PO3ONit]2 (d = 3.9). When PO(ONit)3 was treated with 1 in a 4:1 ratio, all substrates were consumed after 1 h. The only products were [PO2(ONit)2] and HONit, while complex 1 remained in solution. The hydrolysis of CH3COONit under the same conditions proceeded much slower. After 20 days only 50% of the substrates were consumed, being cleaved to acetate and HONit. At the same time, all proton resonances of 1 in the aromatic range were replaced by new ones of the same intensity nearby, and after 20 days this change of 1, which we assume to be a hydrolytic destruction, was complete. Complex 3 needed about 40 h to achieve consumption of three equivalents of PO(ONit)3. After this time, however, it was already evident that the first cleavage product [PO2(ONit)2] is further hydrolyzed to [PO3(ONit)]2. After about 4 days, two thirds of the [PO2(ONit)2] had been converted to [PO3(ONit)]2. But one third of complex 3 had also been destroyed, and the hydrolysis had become considerably slower. Fig. 3 shows the progression of this catalytic process, the most important observation of which is that these TpZn complexes are able to catalyze also the second step of PO(ONit)3 hydrolysis, which is of much higher biological significance than the first one [1,2]. The most efficient catalyst was the dinuclear complex 5. Although we have no proof for it, we assume that the active catalyst is not 5, but one of the mononuclear complexes [L3Zn–OH2]+ or L3Zn–OH (4) which should be in equilibrium with 5 in the presence of MOPS [31]. This equilibration should be favourable, generating only small amounts of the labile mononuclear complexes, of which [L3Zn–OH2]ClO4 could not be isolated [31] and of which 4 was found here to be too labile to be useful for the hydrolytic reactions. It took 40 h with complex 5 to effect the consumption of six equivalents of PO(ONit)3, and after this time 30% of the
Fig. 3. 1H NMR spectra of the reaction mixture containing PO(ONit)3 and 3 (3:1) in CDCl3/CDOD/D2O (4:10:1) in the presence of 5 equivalents of the MOPS buffer.
resulting [PO2(ONit)2] had already been hydrolyzed further to [PO3(ONit)]2. However, after this time the decomposition of 5 was already evident, the reaction became much slower, and after about 3 days there was no more 5 in the solution. Thus a total of 8 ester functions have been hydrolyzed by one equivalent of 5, compared to about 4 for 3 and also 4 for 1. When one counts the number of hydrolytic steps per zinc ion, the numbers for 5 have to be cut in half, and all three zinc complexes are of comparable hydrolytic strength. It would have suggested itself to perform kinetic measurements on the catalytic reactions or to study saturation phenomena. In all cases, however, the imminent decomposition of the catalysts caused an erratic course of the hydrolyses, for which reason there was no solid basis for this kind of studies. 2.3. Conclusions Our new pyrazolylborate ligands have come up to the expectations put into them. They form zinc complexes which are soluble in water-containing media [30]. Among these zinc complexes are those which contain the essential hydrolytic functions Zn–OH2 and Zn–OH [31]. And now
1514
C. Pe´rez Olmo et al. / Inorganica Chimica Acta 360 (2007) 1510–1516
these functions have been put to use both in a stoichiometric and in a catalytic fashion. Although the structures in the solid state of all four functional complexes employed here are untypical for TpZn–OH species (each of them contains five-coordinate zinc and three of them are dinuclear [31]), their hydrolytic activity parallels or exceeds that of the previously used TpZn complexes. We assume that the reason for this is their existence as (possibly solvated) monomers in solution. The characteristic hydrolytic reactions related to the modelling of zinc enzymes, namely heterocumulene insertion and the cleavage of esters and phosphates, have been performed with the new complexes. Only the most difficult hydrolysis, that of amides, could not be achieved. Yet it has not been achieved with any TpZn complex so far, and there are few reports in the literature [3–5] on its realization with other zinc complexes. The major achievements of this work are the catalytic hydrolyses. They are unprecedented in TpZn chemistry, because previously there were no hydrolytically active TpZn complexes which were soluble and useable in water containing solvents. Although the catalytic turnovers are small due to the decomposition of the catalysts, the catalytic activity of the latter seems to be high. Proof for this comes from the fact that complexes 3 and 5 are able to catalyze two steps of the hydrolysis of PO(ONit)3, of which the second one, the cleavage of bis(p-nitrophenyl)phosphate, is known to be more difficult than the first one [3–5]. The major limitation of this work was the lability of the catalytically active complexes in solution. It calls for further refinements in the ligand design, aiming at both a better protection of the hydrolytically sensitive parts of the ligand, i.e. the B–N bonds, and at an even more hydrophilic and hydrogen bond promoting environment of the zinc ion. We leave it to our successors to meet this challenge.
stream, complex 6 could be isolated as follows: A small amount of the precipitate was removed by filtration and the filtrate kept at 5 °C overnight. 10 mg (2%) of 6 precipitated as colourless crystals, m.p. 238 °C, which were identified by a structure determination. The mother liquor was concentrated to 5 ml in a stream of CO2, precipitating another 58 mg (12%) of impure 6. 1 H NMR (CDCl3): d(ppm) = 2.50 [s, 12H, Me(pz)], 2.56 [s, 6H, Me(pz)], 2.82 [s, 3H, OMe], 5.85 [s, 2H, H(pz)], 6.30 [s, 4H, H(pz)], 6.36 [dd, J = 1.7 and 4.7 Hz, 4H, Py], 7.45 [d, J = 6.2 Hz, 6H, Py], 7.53 [dd, J = 1.6 and 4.4 Hz, 4H, Py], 7.74 [dd, J = 1.6 and 4.6 Hz, 4H, Py], 8.59 [d, J = 5.2 Hz, 8H, Py]. IR(KBr): 2559w (BH), 1606s (Py). 3.1.2. Complex 7 371 ll (470 mg, 6.18 mmol) of CS2 was added dropwise with stirring over a period of 30 min into a solution of 352 mg (0.62 mmol) of 2 in 30 ml of methanol/dichloromethane/water (2:2:1) at 5 °C. After 6 h of stirring, all volatiles were removed in vacuo and the residue was taken up in chloroform. After 30 min of stirring the mixture was filtered and the filtrate evaporated to dryness, leaving behind 289 mg (71%) of 7 as a colourless powder, m.p. 218 °C. 1 H NMR (CDCl3): d(ppm) = 2.59 [s, 9H, Me(pz)], 3.17 [s, 3H, OMe], 6.30 [s, 3H, H(pz)], 7.52 [dd, J = 1.6 and 4.4 Hz, 6H, Py], 8.60 [dd, J = 1.6 and 4.4 Hz, 6H, Py]. IR(KBr): 2562m (BH), 1606s (Py). Anal. Calc. for C29H28BN9OS2Zn (658.93): C, 52.86; H, 4.28; N, 19.13; S, 9.73; Zn 9.92. Found: C, 52.32; H, 4.42; N, 20.11; S, 8.96; Zn, 9.48%.
3.1. Stoichiometric reactions
3.1.3. Cleavage of PO(ONit)3 by 2 To a solution of 424 mg (0.75 mmol) of 2 in 50 ml of methanol/dichloromethane (1:1), which was cooled to 5–10 °C in an ice bath, was added dropwise a solution of 172 mg (0.37 mmol) of P(O)(ONit)3 in 10 ml of methanol/dichloromethane (1:1). After 2 h of stirring, the yellow solution was concentrated in vacuo until a solid started to precipitate and then it was kept at 5 °C overnight to yield 83 mg (25%) of 8 as a light yellow powder. The mother liquor was evaporated to dryness and the residue was recrystallized from 15 ml of methanol/dichloromethane (1:2) at 20 °C to yield after 2 days 183 mg (71%) of 9.
3.1.1. Complex 6 223 mg (0.39 mmol) of 2 was dissolved in 50 ml of methanol/dichloromethane/water (2:2:1). The solution was cooled to 5 °C and a slow stream of CO2 was passed through it until all solvent was evaporated. A 1H NMR spectrum of the residue indicated that 2 was consumed and that the main constituent of the residue was a compound whose 1H NMR signals are close to those of Tp4-Py,MeZn–OCOOMe (see below). This compound could not be purified by crystallization. When the reaction mixture was worked up after 4 h, i.e. when half of the solvent had been evaporated in the CO2
3.1.4. Cleavage of CH3COONit by 2 To a solution of 387 mg (0.68 mmol) of 2 in 25 ml of methanol/dichloromethane (2:1), which was cooled to 5–10 °C in an ice bath, was added dropwise a solution of 62 mg (0.34 mmol) of CH3C(O)ONit in 10 ml of methanol. After stirring overnight, the yellow solution was evaporated to dryness in vacuo, suspended in 20 of methanol and then filtered off. The filtrate was concentrated in vacuo to ca. 10 ml and kept at 5 °C overnight to yield 162 mg (69%) of 10 as a crystalline powder. Recrystallization of the residue from methanol/dichloromethane (1:2) at 20 °C yielded after 2 days 114 mg (55%) of 9.
3. Experimental For the preparative work, the general working and measuring procedures were as before [35]. The synthesis of complexes 1–5 is described in Ref. [31].
C. Pe´rez Olmo et al. / Inorganica Chimica Acta 360 (2007) 1510–1516
3.1.5. Complex 11 A solution of 397 mg (0.70 mmol) of 2 in 30 ml of methanol/dichloromethane (2:1) was added dropwise with stirring over a period of 30 min to a solution of 79 mg (0.70 mmol) of trifluoroacetamide in 5 ml of methanol, kept at 5 °C. After 6 h of stirring all volatiles were removed in vacuo, and the residue was crystallized from 15 ml of chloroform by evaporation at room temp. to yield 343 mg (74%) of 11 as a colourless powder, m.p. 205 °C (dec.). 1 H NMR (CDCl3): d(ppm) = 2.58 [s, 9H, Me(pz)], 5.05 [s, b, 1H, NH], 6.34 [s, 3H, H(pz)], 7.39 [d, J = 5.9 Hz, 6H, Py], 8.60 [d, J = 5.8 Hz, 6H, Py]. 19 F NMR (CDCl3): d(ppm) = 76.03. IR(KBr): 3416m (NH), 2559w (BH), 1679s (CO), 1607m (Py). Anal. Calc. for C29H26BF3N10OZn (663.79): C, 50.47; H, 3.95; N, 21.10; Zn, 9.85. Found: C, 49.88; H, 3.93; N, 19.68; Zn, 9.00%. 3.1.6. Stoichiometric cleavages by 1 and 3 These experiments were carried out in NMR tubes. 0.5 ml each of the two reagents from 0.3 M stock solutions in deuterated solvents was mixed and kept at room temp. 1 H NMR spectra were recorded at appropriate intervals. Section 2 names the solvents. Fig. 2 shows a representative spectrum. The resulting zinc complexes were identified by recording the spectra of authentic samples [30], the organic reaction products were all commercially available compounds and could be identified as such. 3.2. Catalytic hydrolyses The procedures for mixing the reagents and identifying the hydrolysis products were as above. Section 2 names the solvents. The buffer was added as such. All reaction mixtures were kept at 310 K, the temperature inside the NMR cavity. Fig. 3 shows a representative sequence of spectra. 3.3. Structure determination The crystal of 6 was taken from the reaction product. The data set was obtained with a Bruker Smart CCD diffractometer at 236 K with Mo Ka radiation and subjected to an empirical absorption correction (program SADABS). The structure was solved with direct methods and refined anisotropically with the SHELX program suite [36]. In addition to one formula unit of 6, two water molecules were found in a highly disordered form in the asymmetric unit and treated as a diffused contribution using the program SQUEEZE [37]. The figure of 6 was produced with SCHAKAL [38]. 6: Formula C56H53B2ClN18O3Zn2 Æ 2H2O, mol. wt. = 1213.97 + 36.04, space group C2/c, Z = 8, a = 18.409(4) ˚ , b = 25.757(6) A ˚ , c = 29.688(6) A ˚ , b = 96.427(4)°, V = A 3 3 ˚ 13988(5) A , Dcalc = 1.15 g/cm , l(Mo Ka) = 0.77 mm1, R1 (obs. refl.) = 0.072, wR2 (all refl.) = 0.219.
1515
4. Supplementary material The crystallographic data of the structure described in this paper have been deposited with the Cambridge Crystallographic Data Centre as supplementary publication no. CCDC-609699. Copies of these data are available free of charge from the following address: The Director, CCDC, 12 Union Road, GB Cambridge CB2 1EZ (Telefax: Int. +1223/336033; e-mail: [email protected]). Acknowledgements This work was supported by the Deutsche Forschungsgemeinschaft and by the Fonds der Chemischen Industrie. We are indebted to Mrs. P. Klose for her contributions in the laboratory. References [1] H. Sigel (Ed.), Zinc and its Role in Biology and Nutrition, Marcel Dekker, New York, 1983. [2] R.H. Prince, in: G. Wilkinson, R.D. Gillard, J.A. McCleverty (Eds.), Comprehensive Coordination Chemistry, vol. 5, Pergamon, Oxford, 1987, pp. 926–1045. [3] M.F. Dunn, Struct. Bonding (Berlin) 23 (1975) 61. [4] R.S. Brown, J. Huguet, N.J. Curtis, in Ref. [1], pp. 55–99. [5] G. Parkin, in: A. Sigel, H. Sigel (Eds.), Metal Ions in Biological Systems, vol. 38, Marcel Dekker, New York, 2001, pp. 411–460. [6] V. Scheller-Krattiger, H. Sigel, Inorg. Chem. 25 (1986) 2628. [7] R.S. Brown, M. Zamkanei, J.L. Cocho, J. Am. Chem. Soc. 106 (1984) 5222. [8] S. Albedyhl, M.T. Averbuch, C. Belle, B. Krebs, J.L. Pierre, E. Saint-Aman, S. Torelli, Eur. J. Inorg. Chem. (2001) 1457. [9] S. Hikichi, M. Tanaka, Y. Moro-oka, N. Kitajima, J. Chem. Soc., Chem. Commun. (1992) 814. [10] R.H. Chapman, R. Breslow, J. Am. Chem. Soc. 117 (1995) 5462. [11] T. Koike, E. Kimura, J. Am. Chem. Soc. 113 (1991) 8935. [12] H. Adams, N.A. Bailey, D.E. Fenton, Q.Y. He, J. Chem. Soc., Dalton Trans. (1996) 2857. [13] T. Itoh, Y. Fujii, T. Tada, Y. Yoshikawa, H. Hisada, Bull. Chem. Soc. Jpn. 69 (1996) 1265. [14] A. Bencini, E. Berni, A. Bianchi, V. Fodi, C. Giorgi, P. Paoletti, B. Valtincoli, Inorg. Chem. 38 (1999) 6323. [15] M.M. Ibrahim, K. Ichikawa, M. Shiro, Inorg. Chim. Acta 353 (2003) 187. [16] B. Bauer-Siebenlist, F. Meyer, E. Farkas, D. Vidovic, J.A. CuestaSeijo, R. Herbst-Irmer, H. Pritzkow, Inorg. Chem. 43 (2004) 4189. [17] D.P. Goldberg, R.C. diTargiani, F. Namuswe, S.C. Chang, L.N. Zakharov, A.L. Rheingold, Inorg. Chem. 44 (2005) 7559. [18] G. Feng, J.C. Mareque Rivas, N.H. Williams, Chem. Commun. (2006) 1845. [19] M. Ruf, H. Vahrenkamp, Chem. Ber. 129 (1996) 1025. [20] K. Weis, M. Rombach, M. Ruf, H. Vahrenkamp, Eur. J. Inorg. Chem. (1998) 263. [21] A. Tro¨sch, H. Vahrenkamp, Inorg. Chem. 40 (2001) 2305. [22] H. Brombacher, H. Vahrenkamp, Inorg. Chem. 43 (2004) 6050. [23] F. Gross, H. Vahrenkamp, Inorg. Chem. 44 (2005) 3321. [24] F. Gross, H. Vahrenkamp, Inorg. Chem. 44 (2005) 4433. [25] M. Rombach, C. Maurer, K. Weis, E. Keller, H. Vahrenkamp, Chem. Eur. J. 5 (1999) 1013. [26] H. Vahrenkamp, Acc. Chem. Res. 32 (1999) 589. [27] D.W. Christianson, Adv. Protein Chem. 42 (1991) 281. [28] R. Alsfasser, M. Ruf, S. Trofimenko, H. Vahrenkamp, Chem. Ber. 126 (1993) 703.
1516
C. Pe´rez Olmo et al. / Inorganica Chimica Acta 360 (2007) 1510–1516
[29] G. Parkin, Chem. Rev. 104 (2004) 699. [30] C. Pe´rez Olmo, K. Bo¨hmerle, G. Steinfeld, H. Vahrenkamp, Eur. J. Inorg. Chem. (2006), early view. [31] M.M. Ibrahim, C. Pe´rez Olmo, T. Tekeste, J. Seebacher, G. He, J.A. Maldonado Calvo, K. Bo¨hmerle, G. Steinfeld, H. Brombacher, H. Vahrenkamp, Inorg. Chem. 45 (2006) 7439. [32] A. Peter, H. Vahrenkamp, Z. Anorg. Allg. Chem. 631 (2005) 2347. [33] H. Brombacher, H. Vahrenkamp, Inorg. Chem. 43 (2004) 6054.
[34] J.A. Maldonado Calvo, H. Vahrenkamp, Inorg. Chim. Acta 359 (2006) 4079. [35] M. Fo¨rster, R. Burth, A.K. Powell, T. Eiche, H. Vahrenkamp, Chem. Ber. 126 (1993) 2643. [36] G.M. Sheldrick, SHELXTL Program Package of the Bruker Smart CCD Diffractometer, version 5.1, Universita¨t Go¨tingen, 2002. [37] A.L. Spek, J. Appl. Cryst. 36 (2003) 7. [38] E. Keller, SCHAKAL for Windows, Universita¨t Freiburg, 2002.
Zinc enzyme modelling with zinc complexes of polar pyrazolylborate ligands C. Pe´rez Olmo, K. Bo¨hmerle, H. Vahrenkamp
*
Institut fu¨r Anorganische und Analytische Chemie der Universita¨t Freiburg, Albertstr. 21, D-79104 Freiburg, Germany Received 16 July 2006; accepted 15 August 2006 Available online 26 August 2006
Abstract The Zn–OH2 and Zn–OH complexes of the new tris(pyrazolyl)borate ligands with pyridyl and carboxamido substituents were investigated for their reactivity towards hydrolyzeable substrates. Tp4Py,MeZn–OH inserted CO2 and CS2 in methanol forming the Zn–OCOOMe and Zn–SCSOMe products. In non-aqueous media, both types of complexes with both types of substituents on the Tp ligands effected stoichiometric cleavage of tris(p-nitrophenyl)phosphate and p-nitrophenyl acetate. In solutions containing water and the MOPS buffer, up to eight p-nitrophenyl groups per equivalent of zinc complex could be cleaved from the esters, and the resulting bis(p-nitrophenyl)phosphate was also degraded to mono(p-nitrophenyl)phosphate. This is the first time that pyrazolylborate–zinc complexes have shown catalytic activity in hydrolytic reactions. Ó 2006 Elsevier B.V. All rights reserved. Keywords: Bioinorganic chemistry; CO2 activation; Ester cleavage; Pyrazolylborate ligands; Zinc
1. Introduction The hydrolysis of esters and phosphates, particularly tris(p-nitrophenyl)phosphate, is the most thoroughly studied reaction in the modelling of zinc enzymes with coordination compounds of zinc. The review articles bear witness to this [1–5], the leading researchers in the field have worked on it [6–12], and there is continued activity until today [13–18]. We too have worked in this area [19–24], and we have derived a mechanistic pathway. [25,26]. During these studies, like in other fields of enzyme modelling, it became evident that the more well-defined the model complexes were the lower was their catalytic efficiency. This holds particularly for those cases where fouror five-coordinate Zn–OH2 or Zn–OH complexes could actually be isolated and used as enzyme models. Thereby the best structural models, typically (tripod) Zn–X species reproducing the tetrahedral environment of zinc in the *
Corresponding author. Tel.: +49 761 203 6120; fax: +49 761 203 6001. E-mail address: [email protected] (H. Vahrenkamp).
0020-1693/$ - see front matter Ó 2006 Elsevier B.V. All rights reserved. doi:10.1016/j.ica.2006.08.021
enzymes [27], were often useful only for stoichiometric, but rarely for catalytic hydrolyses. We experienced this with our pyrazolylborate–zinchydroxide complexes, which were the first structurally correct and functional models of the active state of hydrolytic zinc enzymes [28]. They could be used in a stoichiometric fashion for all types of hydrolyses [19,20,24,25]. But their application for catalyses was prevented by product inhibition. Typically, both products of the hydrolysis of a triarylphosphate, i.e. the phenolate and the diarylphosphate, ended up in stable TpZn–phenolate and TpZn–phosphate complexes. We assume that the major reason for this catalytic inability is the non-polar nature of the hitherto employed pyrazolylborates and their zinc complexes. The synthetic procedure for the Tp ligands, i.e. heating a pyrazole and KBH4 up to 200 °C, has limited the substituents allowed on the pyrazoles almost entirely to hydrocarbon groups. As a result the ligands and their complexes have been insoluble in water and water-containing solvents. Consequentially, during the hydrolytic reactions very little water was
C. Pe´rez Olmo et al. / Inorganica Chimica Acta 360 (2007) 1510–1516
available to overcome the product inhibition and to reconstitute the hydrolytically active complexes. The way out of this dilemma had to be the synthesis of water-soluble pyrazolylborates. We, among others (see Ref. [29]), have recently succeeded in this task [30]. In addition, we could convert the new 4-pyridyl- and carboxamido-substituted Tp ligands to Zn–OH2 and Zn–OH complexes [31]. Thereby we had available the desired enzyme models which could be employed in the presence of water. This paper reports our investigations on their stoichiometric and catalytic hydrolyses of esters and organophosphates. The Zn–OH2 and Zn–OH complexes of L1–L3 were used. H
H B N
N
N
N
B N
N
N
N
N
N
N
N
O HN
O
HN N
O
NH
N
N
L2
L1 H B
HN
N
N
N
N
N
N
O HN
O
O
NH
L3
2. Results and discussion The aqua- and hydroxo-complexes of ligands L1–L3 which could be isolated are 1–5 [31]. Of these, the mononuclear hydroxides 2 and 4 are rather labile and cannot be handled in solution for extended periods, for which reason only 2 was used, and for some stoichiometric reactions only. Aqua complexes 1 and 3 and the hydroxide-bridged dinuclear complex 5 persisted in solutions containing water and buffer, and hence were suitable for catalytic reactions. 1
1
½L Zn–OH2 ClO4 L Zn–OH 2 1 2 3 ½L Zn–OH2 ClO4 L Zn–OH 4 3 3 3 ½L Zn–OH–ZnL ClO4 5
1511
that hydroxides 2 and 4 would react accordingly. This was borne out by the hydrolytic reactions of 2. When CO2 was bubbled through a methanolic solution of 2, it could be seen by 1H NMR that all of 2 was consumed after four hours. The isolation of the expected methylcarbonate complex proved difficult, however, being hampered by slow decomposition of the starting material and the product. Thus only a mediocre yield of impure Tp4Py,MeZn–OCOOMe could be obtained by speedy isolation, or careful crystallization produced a very small amount of the desired product in the unusual composition 6. We assume that the chloride in 6 originates from traces of HCl often found in dichloromethane. L1 Zn–OCOOMe L1 ZnCl 6 The composition of 6 was determined by a structure determination and by 1H NMR. Fig. 1 shows the contents of the asymmetric unit in the crystals. 6 is a chain-like coordination polymer with alternating L1Zn–OCOOMe and L1ZnCl units, each of which uses one of its Tp’s pyridyl substituents to coordinate to the next unit in the chain. This way all zinc ions become five-coordinate with a rather distorted trigonal-bipyramidal geometry. The formation of coordination dimers and polymers is very common for zinc complexes of L1–L3 [30,31], and the related complex Tp3Py,MeZn–OCOOMe is a pyridyl bridged dimer [32]. The bond lengths at both zinc ions (see Fig. 1) are normal, including the elongations at the axes of the trigonal bipyramids. Thus the main value of the structure determination of 6 consists in the proof that it contains L1Zn–OCOOMe. The reaction of 2 with CS2 in the presence of methanol proceeded more smoothly and produced the xanthogenate complex 7 in a good yield. The main indicators for the presence of a Zn–SC(S)OMe unit in 7 are the IR band for C@S at 1208 cm1 and the OMe resonance in the 1H NMR spectrum at 3.17 ppm. The NMR spectra of 7 do not point to association by a splitting of the pyridyl resonances. We therefore assume that 7 is monomeric, using the OMe function of the xanthogenate ligand to make zinc five-coordinate, just as was observed in L1Zn–OC(O)Me [30]. L1 Zn–SCðSÞOMe 7 The cleavage of activated esters by 2 could be performed with the p-nitrophenolates PO(ONit)3 and CH3COONit. As expected, the reactions were stoichiometric, and they consumed two equivalents of 2 per equivalent of substrate according to the following equations:
2.1. Stoichiometric reactions
POðONitÞ3 þ 2L1 Zn–OH 2 ! L1 Zn–OPOðONitÞ2 þ L1 Zn–ONit þ H2 O 9 8
After introducing the TpZn–OH complexes as enzyme models [28], we had used them extensively for stoichiometric model reactions [26]. It could be expected, therefore,
CH3 COONit þ 2L1 Zn–OH 2 ! L1 Zn–OCðOÞCH3 þ L1 Zn–ONit þ H2 O 9 10
ð1Þ
ð2Þ
1512
C. Pe´rez Olmo et al. / Inorganica Chimica Acta 360 (2007) 1510–1516
Fig. 1. Contents of the asymmetric unit of compound 6. The linking of the units of 6 in the coordination polymer is indicated by dotted lines. Bond lengths ˚ ): Zn1–O 1.930(3), Zn1–N1 2.070(4), Zn1–N2 2.038(4), Zn1–N3 2.300(4), Zn1–N18 2.192(4), Zn2–Cl 2.285(3), Zn2–N9 2.181(4), Zn2–N10 2.050(4), (A Zn2–N11 2.035(4), Zn2–N12 2.575(4).
The cleavage of the phosphate proceeded about twice as fast as that of the acetate. Following reactions (1) and (2) by 1H NMR indicated that they were quantitative, using the spectroscopic data of their products 8, 9 and 10 which we had prepared previously [30]. The workup of both reactions, however, was cumbersome, and after their separation 8 and 9 from reaction (1) and 9 and 10 from reaction (2) were obtained only in moderate yields. Hoping for amide hydrolysis, complex 2 was also reacted with trifluoroacetamide. However, instead of cleavage deprotonation took place with formation of the amidate complex 11. The identification of 11 rests in its m(CO) IR band at 1679 cm1, which compares well with those of other TpZn–NHC(O)R complexes [33,34] and which rules out the hoped-for formation of the acetate complex L1Zn–OC(O)CF3 for which a m(CO) band around 1710 cm1 would be expected [19,34]. Thus, in contrast to previous propositions by us [19], it must be stated that until now amide cleavage by TpZn–OH complexes could not be achieved yet.
gressively more acidic which in turn resulted in the hydrolytic destruction of the Tp ligands. The substrate hydrolyses by 1 and 3 became considerably faster when water was present. Thus PO(ONit)3 was quantitatively hydrolyzed to HOPO(ONit)2 and HONit by complex 1 after 15 min in DMSO/water (3:1). After this time complex 1 was still intact. This is demonstrated by Fig. 2 which shows only the hydrolysis products and no more PO(ONit)3. The unmarked signals in Fig. 2 belong to complex 1 with its characteristic 1:5 splitting of the pyrazole CH singlet at 6.6 ppm and the two pyridyl doublets near 7.7 and 8.5 ppm [31]. Similar observations were made when reacting 1 with CH3COONit in DMSO/H2O. It took, however, five days to achieve quantitative cleavage to CH3COOH and HONit with one equivalent of 1.
HOP(O)(ONit)2
HONit
L1 Zn–NHCðOÞCF3 11 Stoichiometric hydrolytic cleavages were also possible using the aqua complexes 1 and 3. They were performed in NMR tubes and followed by 1H and 13P NMR. When water-free conditions were applied, just as during the reactions with 2, the cleavages by 1 and 3 were considerably slower. Thus, only two thirds of the PO(ONit)3 were cleaved by 3 in DMSO after six days. NMR revealed (for representative spectra see below) that the cleavage products are free HONit and HOPO(ONit)2 and not their LZn complexes. This implies that the reaction solutions became pro-
9.0 8.8 8.6 8.4 8.2 8.0 7.8 7.6 7.4 7.2 7.0 6.8 6.6 6.4 6.2
[ ppm] Fig. 2. 1H NMR spectrum of the 1:1 reaction mixture of 1 and PO(ONit)3 in DMSO-d6/D2O (3:1) after 15 min. For the unmarked resonances see text.
C. Pe´rez Olmo et al. / Inorganica Chimica Acta 360 (2007) 1510–1516
1513
2.2. Catalytic hydrolyses The cleavages of the esters with water generate p-nitrophenol and the corresponding acids. For the attempts to catalyze these reactions with the zinc complexes therefore water-containing solvents had to be used, and a buffer had to be present in order to retard the hydrolytic selfdestruction of the Tp ligands which is fast in acidic solutions. All reactions were carried out in NMR tubes, those employing complex 1 in DMSO-d6/D2O (3:1), those employing complexes 3 and 5 in CDCl3/CD3OD/D2O (4:10:1). The buffer MOPS (3-morpholinopropanesulfonic acid, pKa = 7.2) was applied in a 5–20-fold excess relative to the zinc complexes. Control experiments showed that under these conditions the uncatalyzed hydrolyses are at least 1–2 orders of magnitude slower than those in the presence of the zinc complexes. All reactions were performed at 310 K and followed by 1H NMR. The cleavages of PO(ONit)3 were also followed by 31P NMR, which in combination with the signal intensities in the 1H NMR spectra ensured the identification of the solution constituents PO(ONit)3 (d = 18.8), [PO2(ONit)2] (d = 11.3) and [PO3ONit]2 (d = 3.9). When PO(ONit)3 was treated with 1 in a 4:1 ratio, all substrates were consumed after 1 h. The only products were [PO2(ONit)2] and HONit, while complex 1 remained in solution. The hydrolysis of CH3COONit under the same conditions proceeded much slower. After 20 days only 50% of the substrates were consumed, being cleaved to acetate and HONit. At the same time, all proton resonances of 1 in the aromatic range were replaced by new ones of the same intensity nearby, and after 20 days this change of 1, which we assume to be a hydrolytic destruction, was complete. Complex 3 needed about 40 h to achieve consumption of three equivalents of PO(ONit)3. After this time, however, it was already evident that the first cleavage product [PO2(ONit)2] is further hydrolyzed to [PO3(ONit)]2. After about 4 days, two thirds of the [PO2(ONit)2] had been converted to [PO3(ONit)]2. But one third of complex 3 had also been destroyed, and the hydrolysis had become considerably slower. Fig. 3 shows the progression of this catalytic process, the most important observation of which is that these TpZn complexes are able to catalyze also the second step of PO(ONit)3 hydrolysis, which is of much higher biological significance than the first one [1,2]. The most efficient catalyst was the dinuclear complex 5. Although we have no proof for it, we assume that the active catalyst is not 5, but one of the mononuclear complexes [L3Zn–OH2]+ or L3Zn–OH (4) which should be in equilibrium with 5 in the presence of MOPS [31]. This equilibration should be favourable, generating only small amounts of the labile mononuclear complexes, of which [L3Zn–OH2]ClO4 could not be isolated [31] and of which 4 was found here to be too labile to be useful for the hydrolytic reactions. It took 40 h with complex 5 to effect the consumption of six equivalents of PO(ONit)3, and after this time 30% of the
Fig. 3. 1H NMR spectra of the reaction mixture containing PO(ONit)3 and 3 (3:1) in CDCl3/CDOD/D2O (4:10:1) in the presence of 5 equivalents of the MOPS buffer.
resulting [PO2(ONit)2] had already been hydrolyzed further to [PO3(ONit)]2. However, after this time the decomposition of 5 was already evident, the reaction became much slower, and after about 3 days there was no more 5 in the solution. Thus a total of 8 ester functions have been hydrolyzed by one equivalent of 5, compared to about 4 for 3 and also 4 for 1. When one counts the number of hydrolytic steps per zinc ion, the numbers for 5 have to be cut in half, and all three zinc complexes are of comparable hydrolytic strength. It would have suggested itself to perform kinetic measurements on the catalytic reactions or to study saturation phenomena. In all cases, however, the imminent decomposition of the catalysts caused an erratic course of the hydrolyses, for which reason there was no solid basis for this kind of studies. 2.3. Conclusions Our new pyrazolylborate ligands have come up to the expectations put into them. They form zinc complexes which are soluble in water-containing media [30]. Among these zinc complexes are those which contain the essential hydrolytic functions Zn–OH2 and Zn–OH [31]. And now
1514
C. Pe´rez Olmo et al. / Inorganica Chimica Acta 360 (2007) 1510–1516
these functions have been put to use both in a stoichiometric and in a catalytic fashion. Although the structures in the solid state of all four functional complexes employed here are untypical for TpZn–OH species (each of them contains five-coordinate zinc and three of them are dinuclear [31]), their hydrolytic activity parallels or exceeds that of the previously used TpZn complexes. We assume that the reason for this is their existence as (possibly solvated) monomers in solution. The characteristic hydrolytic reactions related to the modelling of zinc enzymes, namely heterocumulene insertion and the cleavage of esters and phosphates, have been performed with the new complexes. Only the most difficult hydrolysis, that of amides, could not be achieved. Yet it has not been achieved with any TpZn complex so far, and there are few reports in the literature [3–5] on its realization with other zinc complexes. The major achievements of this work are the catalytic hydrolyses. They are unprecedented in TpZn chemistry, because previously there were no hydrolytically active TpZn complexes which were soluble and useable in water containing solvents. Although the catalytic turnovers are small due to the decomposition of the catalysts, the catalytic activity of the latter seems to be high. Proof for this comes from the fact that complexes 3 and 5 are able to catalyze two steps of the hydrolysis of PO(ONit)3, of which the second one, the cleavage of bis(p-nitrophenyl)phosphate, is known to be more difficult than the first one [3–5]. The major limitation of this work was the lability of the catalytically active complexes in solution. It calls for further refinements in the ligand design, aiming at both a better protection of the hydrolytically sensitive parts of the ligand, i.e. the B–N bonds, and at an even more hydrophilic and hydrogen bond promoting environment of the zinc ion. We leave it to our successors to meet this challenge.
stream, complex 6 could be isolated as follows: A small amount of the precipitate was removed by filtration and the filtrate kept at 5 °C overnight. 10 mg (2%) of 6 precipitated as colourless crystals, m.p. 238 °C, which were identified by a structure determination. The mother liquor was concentrated to 5 ml in a stream of CO2, precipitating another 58 mg (12%) of impure 6. 1 H NMR (CDCl3): d(ppm) = 2.50 [s, 12H, Me(pz)], 2.56 [s, 6H, Me(pz)], 2.82 [s, 3H, OMe], 5.85 [s, 2H, H(pz)], 6.30 [s, 4H, H(pz)], 6.36 [dd, J = 1.7 and 4.7 Hz, 4H, Py], 7.45 [d, J = 6.2 Hz, 6H, Py], 7.53 [dd, J = 1.6 and 4.4 Hz, 4H, Py], 7.74 [dd, J = 1.6 and 4.6 Hz, 4H, Py], 8.59 [d, J = 5.2 Hz, 8H, Py]. IR(KBr): 2559w (BH), 1606s (Py). 3.1.2. Complex 7 371 ll (470 mg, 6.18 mmol) of CS2 was added dropwise with stirring over a period of 30 min into a solution of 352 mg (0.62 mmol) of 2 in 30 ml of methanol/dichloromethane/water (2:2:1) at 5 °C. After 6 h of stirring, all volatiles were removed in vacuo and the residue was taken up in chloroform. After 30 min of stirring the mixture was filtered and the filtrate evaporated to dryness, leaving behind 289 mg (71%) of 7 as a colourless powder, m.p. 218 °C. 1 H NMR (CDCl3): d(ppm) = 2.59 [s, 9H, Me(pz)], 3.17 [s, 3H, OMe], 6.30 [s, 3H, H(pz)], 7.52 [dd, J = 1.6 and 4.4 Hz, 6H, Py], 8.60 [dd, J = 1.6 and 4.4 Hz, 6H, Py]. IR(KBr): 2562m (BH), 1606s (Py). Anal. Calc. for C29H28BN9OS2Zn (658.93): C, 52.86; H, 4.28; N, 19.13; S, 9.73; Zn 9.92. Found: C, 52.32; H, 4.42; N, 20.11; S, 8.96; Zn, 9.48%.
3.1. Stoichiometric reactions
3.1.3. Cleavage of PO(ONit)3 by 2 To a solution of 424 mg (0.75 mmol) of 2 in 50 ml of methanol/dichloromethane (1:1), which was cooled to 5–10 °C in an ice bath, was added dropwise a solution of 172 mg (0.37 mmol) of P(O)(ONit)3 in 10 ml of methanol/dichloromethane (1:1). After 2 h of stirring, the yellow solution was concentrated in vacuo until a solid started to precipitate and then it was kept at 5 °C overnight to yield 83 mg (25%) of 8 as a light yellow powder. The mother liquor was evaporated to dryness and the residue was recrystallized from 15 ml of methanol/dichloromethane (1:2) at 20 °C to yield after 2 days 183 mg (71%) of 9.
3.1.1. Complex 6 223 mg (0.39 mmol) of 2 was dissolved in 50 ml of methanol/dichloromethane/water (2:2:1). The solution was cooled to 5 °C and a slow stream of CO2 was passed through it until all solvent was evaporated. A 1H NMR spectrum of the residue indicated that 2 was consumed and that the main constituent of the residue was a compound whose 1H NMR signals are close to those of Tp4-Py,MeZn–OCOOMe (see below). This compound could not be purified by crystallization. When the reaction mixture was worked up after 4 h, i.e. when half of the solvent had been evaporated in the CO2
3.1.4. Cleavage of CH3COONit by 2 To a solution of 387 mg (0.68 mmol) of 2 in 25 ml of methanol/dichloromethane (2:1), which was cooled to 5–10 °C in an ice bath, was added dropwise a solution of 62 mg (0.34 mmol) of CH3C(O)ONit in 10 ml of methanol. After stirring overnight, the yellow solution was evaporated to dryness in vacuo, suspended in 20 of methanol and then filtered off. The filtrate was concentrated in vacuo to ca. 10 ml and kept at 5 °C overnight to yield 162 mg (69%) of 10 as a crystalline powder. Recrystallization of the residue from methanol/dichloromethane (1:2) at 20 °C yielded after 2 days 114 mg (55%) of 9.
3. Experimental For the preparative work, the general working and measuring procedures were as before [35]. The synthesis of complexes 1–5 is described in Ref. [31].
C. Pe´rez Olmo et al. / Inorganica Chimica Acta 360 (2007) 1510–1516
3.1.5. Complex 11 A solution of 397 mg (0.70 mmol) of 2 in 30 ml of methanol/dichloromethane (2:1) was added dropwise with stirring over a period of 30 min to a solution of 79 mg (0.70 mmol) of trifluoroacetamide in 5 ml of methanol, kept at 5 °C. After 6 h of stirring all volatiles were removed in vacuo, and the residue was crystallized from 15 ml of chloroform by evaporation at room temp. to yield 343 mg (74%) of 11 as a colourless powder, m.p. 205 °C (dec.). 1 H NMR (CDCl3): d(ppm) = 2.58 [s, 9H, Me(pz)], 5.05 [s, b, 1H, NH], 6.34 [s, 3H, H(pz)], 7.39 [d, J = 5.9 Hz, 6H, Py], 8.60 [d, J = 5.8 Hz, 6H, Py]. 19 F NMR (CDCl3): d(ppm) = 76.03. IR(KBr): 3416m (NH), 2559w (BH), 1679s (CO), 1607m (Py). Anal. Calc. for C29H26BF3N10OZn (663.79): C, 50.47; H, 3.95; N, 21.10; Zn, 9.85. Found: C, 49.88; H, 3.93; N, 19.68; Zn, 9.00%. 3.1.6. Stoichiometric cleavages by 1 and 3 These experiments were carried out in NMR tubes. 0.5 ml each of the two reagents from 0.3 M stock solutions in deuterated solvents was mixed and kept at room temp. 1 H NMR spectra were recorded at appropriate intervals. Section 2 names the solvents. Fig. 2 shows a representative spectrum. The resulting zinc complexes were identified by recording the spectra of authentic samples [30], the organic reaction products were all commercially available compounds and could be identified as such. 3.2. Catalytic hydrolyses The procedures for mixing the reagents and identifying the hydrolysis products were as above. Section 2 names the solvents. The buffer was added as such. All reaction mixtures were kept at 310 K, the temperature inside the NMR cavity. Fig. 3 shows a representative sequence of spectra. 3.3. Structure determination The crystal of 6 was taken from the reaction product. The data set was obtained with a Bruker Smart CCD diffractometer at 236 K with Mo Ka radiation and subjected to an empirical absorption correction (program SADABS). The structure was solved with direct methods and refined anisotropically with the SHELX program suite [36]. In addition to one formula unit of 6, two water molecules were found in a highly disordered form in the asymmetric unit and treated as a diffused contribution using the program SQUEEZE [37]. The figure of 6 was produced with SCHAKAL [38]. 6: Formula C56H53B2ClN18O3Zn2 Æ 2H2O, mol. wt. = 1213.97 + 36.04, space group C2/c, Z = 8, a = 18.409(4) ˚ , b = 25.757(6) A ˚ , c = 29.688(6) A ˚ , b = 96.427(4)°, V = A 3 3 ˚ 13988(5) A , Dcalc = 1.15 g/cm , l(Mo Ka) = 0.77 mm1, R1 (obs. refl.) = 0.072, wR2 (all refl.) = 0.219.
1515
4. Supplementary material The crystallographic data of the structure described in this paper have been deposited with the Cambridge Crystallographic Data Centre as supplementary publication no. CCDC-609699. Copies of these data are available free of charge from the following address: The Director, CCDC, 12 Union Road, GB Cambridge CB2 1EZ (Telefax: Int. +1223/336033; e-mail: [email protected]). Acknowledgements This work was supported by the Deutsche Forschungsgemeinschaft and by the Fonds der Chemischen Industrie. We are indebted to Mrs. P. Klose for her contributions in the laboratory. References [1] H. Sigel (Ed.), Zinc and its Role in Biology and Nutrition, Marcel Dekker, New York, 1983. [2] R.H. Prince, in: G. Wilkinson, R.D. Gillard, J.A. McCleverty (Eds.), Comprehensive Coordination Chemistry, vol. 5, Pergamon, Oxford, 1987, pp. 926–1045. [3] M.F. Dunn, Struct. Bonding (Berlin) 23 (1975) 61. [4] R.S. Brown, J. Huguet, N.J. Curtis, in Ref. [1], pp. 55–99. [5] G. Parkin, in: A. Sigel, H. Sigel (Eds.), Metal Ions in Biological Systems, vol. 38, Marcel Dekker, New York, 2001, pp. 411–460. [6] V. Scheller-Krattiger, H. Sigel, Inorg. Chem. 25 (1986) 2628. [7] R.S. Brown, M. Zamkanei, J.L. Cocho, J. Am. Chem. Soc. 106 (1984) 5222. [8] S. Albedyhl, M.T. Averbuch, C. Belle, B. Krebs, J.L. Pierre, E. Saint-Aman, S. Torelli, Eur. J. Inorg. Chem. (2001) 1457. [9] S. Hikichi, M. Tanaka, Y. Moro-oka, N. Kitajima, J. Chem. Soc., Chem. Commun. (1992) 814. [10] R.H. Chapman, R. Breslow, J. Am. Chem. Soc. 117 (1995) 5462. [11] T. Koike, E. Kimura, J. Am. Chem. Soc. 113 (1991) 8935. [12] H. Adams, N.A. Bailey, D.E. Fenton, Q.Y. He, J. Chem. Soc., Dalton Trans. (1996) 2857. [13] T. Itoh, Y. Fujii, T. Tada, Y. Yoshikawa, H. Hisada, Bull. Chem. Soc. Jpn. 69 (1996) 1265. [14] A. Bencini, E. Berni, A. Bianchi, V. Fodi, C. Giorgi, P. Paoletti, B. Valtincoli, Inorg. Chem. 38 (1999) 6323. [15] M.M. Ibrahim, K. Ichikawa, M. Shiro, Inorg. Chim. Acta 353 (2003) 187. [16] B. Bauer-Siebenlist, F. Meyer, E. Farkas, D. Vidovic, J.A. CuestaSeijo, R. Herbst-Irmer, H. Pritzkow, Inorg. Chem. 43 (2004) 4189. [17] D.P. Goldberg, R.C. diTargiani, F. Namuswe, S.C. Chang, L.N. Zakharov, A.L. Rheingold, Inorg. Chem. 44 (2005) 7559. [18] G. Feng, J.C. Mareque Rivas, N.H. Williams, Chem. Commun. (2006) 1845. [19] M. Ruf, H. Vahrenkamp, Chem. Ber. 129 (1996) 1025. [20] K. Weis, M. Rombach, M. Ruf, H. Vahrenkamp, Eur. J. Inorg. Chem. (1998) 263. [21] A. Tro¨sch, H. Vahrenkamp, Inorg. Chem. 40 (2001) 2305. [22] H. Brombacher, H. Vahrenkamp, Inorg. Chem. 43 (2004) 6050. [23] F. Gross, H. Vahrenkamp, Inorg. Chem. 44 (2005) 3321. [24] F. Gross, H. Vahrenkamp, Inorg. Chem. 44 (2005) 4433. [25] M. Rombach, C. Maurer, K. Weis, E. Keller, H. Vahrenkamp, Chem. Eur. J. 5 (1999) 1013. [26] H. Vahrenkamp, Acc. Chem. Res. 32 (1999) 589. [27] D.W. Christianson, Adv. Protein Chem. 42 (1991) 281. [28] R. Alsfasser, M. Ruf, S. Trofimenko, H. Vahrenkamp, Chem. Ber. 126 (1993) 703.
1516
C. Pe´rez Olmo et al. / Inorganica Chimica Acta 360 (2007) 1510–1516
[29] G. Parkin, Chem. Rev. 104 (2004) 699. [30] C. Pe´rez Olmo, K. Bo¨hmerle, G. Steinfeld, H. Vahrenkamp, Eur. J. Inorg. Chem. (2006), early view. [31] M.M. Ibrahim, C. Pe´rez Olmo, T. Tekeste, J. Seebacher, G. He, J.A. Maldonado Calvo, K. Bo¨hmerle, G. Steinfeld, H. Brombacher, H. Vahrenkamp, Inorg. Chem. 45 (2006) 7439. [32] A. Peter, H. Vahrenkamp, Z. Anorg. Allg. Chem. 631 (2005) 2347. [33] H. Brombacher, H. Vahrenkamp, Inorg. Chem. 43 (2004) 6054.
[34] J.A. Maldonado Calvo, H. Vahrenkamp, Inorg. Chim. Acta 359 (2006) 4079. [35] M. Fo¨rster, R. Burth, A.K. Powell, T. Eiche, H. Vahrenkamp, Chem. Ber. 126 (1993) 2643. [36] G.M. Sheldrick, SHELXTL Program Package of the Bruker Smart CCD Diffractometer, version 5.1, Universita¨t Go¨tingen, 2002. [37] A.L. Spek, J. Appl. Cryst. 36 (2003) 7. [38] E. Keller, SCHAKAL for Windows, Universita¨t Freiburg, 2002.