Modelling the impact of geometric parameters on the redox potential of blue copper proteins

JOURNAL OF

Inorganic Biochemistry Journal of Inorganic Biochemistry 100 (2006) 250–259 www.elsevier.com/locate/jinorgbio

Modelling the impact of geometric parameters on the redox potential of blue copper proteins Michelle K. Taylor, Davina E. Stevenson, Leonard E.A. Berlouis, Alan R. Kennedy, John Reglinski * Department of Pure and Applied Chemistry, Strathclyde University, 295 Cathedral Street, Glasgow G1 1XL, UK Received 17 June 2005; received in revised form 11 November 2005; accepted 14 November 2005 Available online 4 January 2006

Abstract The synthesis and structure of a homologous series of cationic N2S2 copper(I) Schiff base complexes constructed using o-tert-butylthiobenzaldehyde and a series of terminal diamines (ethane, propane, butane) are reported. The complexes differ only in the length of the methylene chain between the imine groups. This simple modification forces the copper centre to shift geometry from a planar (1,2-diaminoethane) to a more distorted tetrahedral motif (1,4-diaminobutane). The redox potentials of the three cations were measured using cyclic voltammetry in donor (acetonitrile) and non-donor solvents (dichloromethane). The S–Cu–N angles for each complex are correlated against the respective redox potential allowing an analysis of the geometric impact on the redox potential in soft copper centres. The redox potential is observed to increase as the metal centre moves from a planar towards a tetrahedral motif. Comparing this data with the reported structures of the blue copper proteins (rusticyanin and plastocyanin) allows an assessment of the contribution of the geometry of the metal binding site to the operating potential of these proteins to be made.
2005 Elsevier Inc. All rights reserved. Keywords: Copper Schiff base complexes; Blue copper proteins; Redox potential; Geometry; Models

1. Introduction Of the 15 metals employed by biological systems, three; iron, copper and molybdenum, are favoured for use in oxidation–reduction processes [1–6]. Consequently, a number of metal specific binding motifs within metallo-proteins and metallo-enzymes have been developed to assist and support these metals during redox driven processes. Copper is somewhat unique as it is mainly found complexed directly to highly conserved donor amino acids within the proteins polypeptide chain [3,4]. Subsequently, for copper two common coordination environments are found; namely the two nitrogen (histidine) two sulfur (cysteine, methionine) donor set of the type I blue copper proteins (amicyanin, plastocyanin and rusticyanin) [3,4,7] and the three histidine donor set found for types II and III oxidases *

Corresponding author. Tel.: +44 141 548 2349; fax: +44 141 552 0876. E-mail address: [email protected] (J. Reglinski).

0162-0134/$ - see front matter
2005 Elsevier Inc. All rights reserved. doi:10.1016/j.jinorgbio.2005.11.009

[6]. The polypeptide chains to which the copper cations are directly bound are potentially highly flexible species and need to be folded into the desired conformation during protein synthesis and held in place during electron transfer. Thus, although the copper binding environment may be static once folded, the geometries of the metal binding sites may not be as highly conserved (Table 1) between proteins from related families [4]. The redox potentials of metallo-proteins are modulated in a number of ways. Foremost is the influence of the donor atoms themselves [2–4]. In unison with the donor atoms all metallo-proteins would also seem to make some use of the surrounding polypeptide chain to control the dielectric strength of the metal environment [1,2]. Furthermore, by attaching the redox centre directly to the polypeptide chain it is possible to distort the geometry and symmetry of the metal centre and thus its operating potential. This facet of redox modulation is accepted for the iron sulfur clusters [1], but is as yet still a topic of some debate

M.K. Taylor et al. / Journal of Inorganic Biochemistry 100 (2006) 250–259

251

Table 1 The structural parameters and redox potentials for three structural characterised members of the blue copper protein family [4,7,10–16] in their reduced and oxidised forms (brackets) ˚ Bond distances A Bond angles () Redox potential (mV) Cu–N Cu–N Cu–Scys Cu–Smet N–Cu–Scys N–Cu–Smet Amicyanin [7,8] Plastocyanin

Rusticyanin [15,16]

Cyanobacterium Synechoccus [9] Dryopteris crassirhizoma [10,11] Poplar plastocyanin [12,13]

– 2.38 2.10 2.39 1.95

(2.04) (2.01) (2.06) (2.06) (1.89)

1.91 2.09 1.95 2.12 2.22

(1.95) (1.97) (1.99) (1.91) (2.04)

2.09 2.18 2.21 2.16 2.25

(2.10) (2.14) (2.23) (2.07) (2.26)

2.90 2.80 2.91 2.87 2.75

(2.90) (2.93) (2.94) (2.82) (2.89)

155 140 130 136 127

(137) (131) (128) (132) (128)

– 100 108 106 105

(100) (99) (108) (103) (106)

260 370 387 389 680

Unlike plastocyanin and rusticyanin, amicyanin moves from a trigonal planar to a tetrahedral motif on oxidation. In amicyanin, histidine 95 rotates to facilitate the necessary bond breaking/making event. Consequently, entries are missing for the corresponding Cu–Nhis95 bond length and angle (Nhis95– Cu–Smet) in the reduced form.

for the blue copper proteins [4]. For type I copper centres it is accepted that the factor which has the greatest effect on the operating potential of the protein are the metal donor distances namely the metal–nitrogen and metal–sulfur bond distances [3,7–17]. The more subtle contribution of the geometric parameters (i.e., the bond angles, geometric flexibility) and the effect of the dielectric strength of the metal binding pocket have been more difficult to assess principally because it has been difficult to decouple the effect of the many variables active in these species during the analysis of copper proteins. Three members of the blue copper protein family have been crystallised in their reduced and oxidised forms (Table 1). It is evident from this small data set that there is some variation in bond distances between proteins and specifically between plastocyanin and rusticyanin in their oxidised and reduced forms. These differences all contribute to the operating potential of this family of proteins. However, considering the similarities between plastocyanin and rusticyanin, the redox potentials for three plastocyanins (
370 mV) and rusticyanin (680 mV) are remarkably different. Except for the Cu–Smet distances, the bond distances (Table 1) for both of these proteins generally shorten as the oxidation state of the metal rises. However, it is notable that with these changes the bond angles remain remarkably unchanged. Large changes in geometry need to be resisted for the protein to maintain rapid electron transfer rates [4,17]. What is not apparent from the data in Table 1 are the hydrophobic/hydrophilic nature of the three binding sites and the dielectric strength of the metal environments. However, neither of these factors will change dramatically during oxidation. Since the major determinant of the redox potential of these proteins is believed to be derived from the Cu–Smet distances [3,4], it is difficult to identify the modulating influences of the geometry and environment of the copper. The answer to this problem lies in decoupling all of the factors in action in these species through the preparation and electrochemical analysis of model compounds that seek to mirror the low valent N2S2 copper binding site. In common with a small number of groups we have become interested in the structural effects that occur in symmetric Schiff base complexes as a function of the length

of the methylene chain (–(CH2)n–) separating the imine groups (Fig. 1) [18–21]. Although the structural motifs of some metals (nickel, zinc) change markedly [20] as the alkyl chain increases (n = 2–4), the N2O2 copper Schiff base complexes (Fig. 1a) [18,21] are observed to gently shift from a planar configuration toward a highly distorted tetrahedral orientation. A similar effect was observed by Bereman et al. [19] in their study of the structural and spectroscopic behaviour of copper(II) N2S2 Schiff base complexes (Fig. 1b). They clearly show that this class of complex can be used as an effective model of the metal binding site in blue copper proteins. Furthermore, the manner in which the geometry at the metal centre changes while undergoing only minimal changes to the local bond distance and dielectric effects, makes compounds of this type an ideal system for the interrogation of the effect of binding site geometry on redox potential. Since we already had an interest in salicylidene based complexes [20,21] we opted to build our copper complexes on this scaffold. As such we report here the synthesis and structure of a short series of copper N2S2 Schiff base cations as models of the type I metal binding site in blue copper proteins. These species are subjected to analysis by X-ray crystallography and electrochemical methods thus generating a view of the alteration in oxidation potential as a function of the geometry at the copper centre (viz. the
N

N

N

X

a

N Cu

M

S S

X

b

S

S

Fig. 1. Two Schiff base scaffolds [19–21], which have been used to model the geometric shifts (planar ! tetrahedral: X = O) in copper complexes. Compounds where M = Cu(I) and X = StBut (a) are the subject of this study. = –(CH2)n– n = 2, 3, 4. The cyclopentenedithiocarboxylate ([Cu(Cdten)] framework (b) supports Cu(II) [19].

252

M.K. Taylor et al. / Journal of Inorganic Biochemistry 100 (2006) 250–259

2. Experimental section All experiments were carried out using standard apparatus and commercially available chemicals. Solvents were used as supplied, apart from the acetonitrile and dichloromethane. Prior to their use in the electrochemical studies, acetonitrile and dichloromethane were dried and re-distilled from calcium hydride. NMR analysis was carried out on a JEOL EX 270 instrument operating at 270 MHz for 1H and 67.5 MHz for 13C. The identification of the various resonances follow accepted practice; s: singlet, m: multiplet, d: doublet, bs: broad singlet, and br m: broad multiplet. Infra-red spectra were recorded as nujol mulls using a Nicolet Avatar 360 FT-IR spectrometer. Again the nature of the bands follow accepted practice; s: strong, m: medium, w: weak. Mass spectra were recorded on a ThermoFinnigan LCQ mass spectrometer by direct injection. Reflectance visible spectra (400–800 nm) were recorded on a Photonics CCD array UV–visible spectrophotometer. Micro-analysis was conducted in-house using a Perkin–Elmer 2400 Series II CHNS/O Analyzer. 2.1. Preparation of N,N 0 -bis-(o-tert-butylthiobenzylidene)1,2-diaminoethane (SSalen) [22] o-(tert-Butylthio)benzaldehyde (1.00 g, 5.15 mmol) and 1,2-diaminoethane (0.155 g, 2.58 mmol) were refluxed in ethanol (30 ml) for 6 h whereupon the solution turned dark yellow. The solvent was removed and the oil taken up in hexane (10 ml). The solution was filtered and placed in the freezer (15 C) whereupon pale yellow crystals formed. These were collected, washed with a minimum amount of cold hexane and allowed to air dry (86%, m.p. 74–76 C). Anal. Calc. for C24H32N2S2: C, 69.85; H, 7.82; N, 6.79. Found: C, 69.63; H, 7.90; N, 6.95%. IR (cm1, Nujol mull): 3057, s, (C–H), 1640 m (C@N) 760 (o-disubstituted benzene ring). dH (270 MHz; solvent CDCl3) 9.10 (s, 2H, –CH@N), 8.1 (m, 2H, arom), 7.5 (m, 2H, arom), 7.4 (m, 4H, arom), 4.0 (s, 4H, –CH2–N@), 1.2 (s, 18H, Me). dC (270 MHz; solvent CDCl3) 163 (C@N), 141, 140, 134, 131, 130, 128 (arom), 62 (N–CH2–), 48 (S–C(CH3)3), 31 (S–C(CH3)3). 2.2. Preparation of N,N 0 -bis-(o-tert-butylthiobenzylidene)1,4-diaminobutane (SSalbu) o-(tert-Butylthio)benzaldehyde (1.00 g, 5.15 mmol) and 1,4-diaminobutane (0.23 g, 2.58 mmol) were refluxed in ethanol (30 ml) for 6 h whereupon the solution turned dark yellow. The solvent was removed and the yellow oil taken up in hexane (10 ml) and re-evaporated to drive off residual ethanol. Hexane (1 ml) was added to the oil and the mixture placed in the freezer (15 C) whereupon pale yellow crystals formed. These were collected quickly, washed with a minimum amount of cold hexane and allowed to air dry (80%, m.p. 60–61 C). Anal. Calc. for C26H36N2S2: C, 70.86; H, 8.23; N, 6.36. Found: C, 70.66; H, 8.30; N,

6.53%. IR (cm1, Nujol mull): 3050 s (C–H), 1635 (C@N), 760 m (o-disubstituted benzene ring). dH (270 MHz; solvent CDCl3) 9.0 (s, 2H, –CH@N), 8.0 (d, 2H, arom), 7.5 (d, 2H, arom), 7.3 (m, 4H, arom), 3.6 (bs, 4H, –N–CH2–), 1.7 (bs, 4H, N–CH2–CH2–) 1.1 (s, 18H, Me). dC (270 MHz; solvent CDCl3) 162 (C@N), 141, 140, 133, 130, 129, 127 (arom), 52 (N–CH2–), 48 (N–CH2– CH2–), 31 (S–C(CH3)3), 29 (S–C(CH3)3). 2.3. Preparation of [CuI(SSalen)]BF4 Tetrakisacetonitrile copper(I) tetrafluoroborate (0.37 g, 1.17 mmol) and N,N 0 -bis-(o-tert-butylthiobenzylidene)1,2-diaminoethane (0.50 g, 1.21 mmol) were each dissolved in ethanol (20 ml). The solution containing the ligand was added dropwise to the copper solution whereupon the solution turned green. The solution was refluxed for 2 h and changed from green to dark brown. The solution was allowed to cool, filtered and the solvent removed. The solids were taken up in chloroform and filtered through charcoal and celite to give a deep orange solution. Vapour diffusion of the orange solution with diethyl-ether produced copious amounts of orange crystals suitable for Xray analysis (40%). Anal. Calc. for C24H32BCuF4N2S2: C, 50.88; H, 5.69; N, 4.95. Found: C, 50.56; H, 5.36; N, 4.93%. dH (270 MHz; solvent CDCl3) 8.8 (s, 2H, –CH@N), 7.5–8.0 (m, 8H, arom), 4.4 (s, 4H, –N–CH2–), 1.2 (s, 18H, Me). 2.4. Preparation of [CuI(SSalpr)]BF4 o-(tert-Butylthio)benzaldehyde (1.00 g, 5.15 mmol) and 1,3-diaminopropane (0.20 g, 2.58 mmol) were refluxed in ethanol (30 ml) for 6 h whereupon the solution turned dark yellow. The solvent was removed and the oil taken up in hexane (10 ml) and re-evaporated to drive off residual ethanol. The impure mixture of N,N 0 -bis-(o-tert-butylthiobenzylidene)-1,3-diaminopropane (SSalpr) refused to crystallise and was used without further purification [23]. Tetrakisacetonitrile copper(I) tetrafluoroborate (0.26 g, 0.82 mmol) and N,N 0 -(bis-o-tert-butylthiobenzylidene)1,3-diaminopropane (0.36 g, 0.82 mmol) were each dissolved in ethanol (20 ml). The solution containing the ligand was added dropwise to the copper solution whereupon the solution turned green. The solution was refluxed for 1h changing from green to dark brown. The solution was hot filtered and allowed to cool overnight producing an orange powder. The powder was taken up in chloroform, filtered through charcoal, celite and crystallised by vapour diffusion with diethyl-ether. Crystals suitable for X-ray analysis were grown by vapour diffusion from methanol: diethyl-ether at 15 C (43%). Anal. Calc. for C25H34BCuF4N2S2: C, 52.04; H, 5.94; N, 4.85. Found: C, 52.09; H, 5.75; N, 4.80%. dH (270 MHz; solvent CDCl3) 8.60 (s, 2H, –CH@N), 7.7 (m, 6H, arom), 7.5 (m, 2H, arom), 4.2 (s, 4H, –N–CH2–), 2.3 (s, 2H, N–CH2–CH2–) 1.1 (s, 18H, Me).

M.K. Taylor et al. / Journal of Inorganic Biochemistry 100 (2006) 250–259

2.5. Preparation of [CuI(SSalbu)]BF4

253

2.7. Preparation of dibromo-(N-o-tert-butylthiobenzylidene)diaminoethane copper(II) [CuII(enSSal)Br2]

Tetrakisacetonitrile copper(I) tetrafluoroborate (0.26 g, 0.82 mmol) and N,N 0 -bis-(o-tert-butylthiobenzylidene)1,4-diaminobutane (0.36 g, 0.82 mmol) were each dissolved in ethanol (20 ml). The solution containing the ligand was added dropwise to the copper solution whereupon the solution turned green. The solution was refluxed for 1 h changing from green to dark brown. The solution was hot filtered and allowed to cool overnight producing an orange powder. The powder was taken up in chloroform, filtered through charcoal and celite and crystallised by vapour diffusion with diethyl-ether (45%). Anal. Calc. for C26H36BCuF4N2S2: C, 45.52; H, 5.23; N, 3.93. Found: C, 45.95; H, 5.15; N, 4.03%. dH (270 MHz; solvent CDCl3) 8.4 (s, 2H, –CH@N), 7.5 (br m, 8H, arom), 3.9 (bs, 4H, –N–CH2–), 2.0 (bs, 4H, N–CH2–CH2–) 1.0 (s, 18H, Me).

[CuII(SSalen)Br2] prepared above was redissolved in the minimum amount of dichloromethane (
5 ml) filtered through celite and re-crystallised with diethyl-ether via vapour diffusion. The resulting product was found to be insoluble in dichloromethane. Crystal suitable for X-ray analysis were obtained from the resulting product. Anal. Calc. for C13H20Br2N2SCu: C, 33.96; H, 4.38; N, 6.09. Found: C, 30.18; H, 4.45; N, 6.83%. IR (cm1, KBr): 3350 s (NH), 3210 s (NH), 1651 s, 1566 m (C@N), 763 m. k(max) reflectance, 691 nm. MS (LC direct) m/z 237 (metal free ligand); 181 (-tBu); 138 (-en). This ‘‘species’’ could not be adequately purified by crystallization and although crystals suitable for X-ray analysis were obtained, the microanalysis of the bulk material consistently failed to meet the accepted standard. This finding is likely indicative of the presence of further hydrolysis products.

2.6. Preparation of [CuII(SSalen)Br2] 2.8. Preparation of [Zn(SSalen)Br2] Copper(II) bromide (0.05 g, 0.24 mmol) dissolved in ethanol (15 ml) and N,N 0 -bis-(o-tert-butylthiobenzylidine)-1,2-diaminoethane (0.10 g, 0.24 mmol) dissolved in ethanol (10 ml) were mixed together. The resulting green solution was stirred for 20 min after which time the solvent was removed leaving a green solid which was washed with diethyl-ether (25%). Anal. Calc. for C24H32N2S2Br2Cu: C, 45.32; H, 5.07; N, 4.40. Found: C, 45.01; H, 5.37; N, 4.67%. IR (cm1, KBr): 1696 s, 1651 s, 1582 s (C@N), 763 m. k(max) reflectance, 682 nm. MS (LC direct) m/z 475, 477 ([CuII(SSalen)]); 419 (-tBu); 363, (-tBu).

Zinc bromide (0.59 g, 2.62 mmol) and N,N 0 -bis-(o-tertbutylthiobenzylidene)-1,2-diaminoethane (1.0 g, 2.43 mmol) were dissolved in ethanol (40 ml). The solution was refluxed for 6 h, hot filtered and stored at 15 C for 12 h. The resulting precipitate was collected and recrystallised from hexane producing a white crystalline mass from which crystals suitable for X-ray analysis could be found (40%). Anal. Calc. for C24H32BCuF4N2S2: C, 44.17; H, 5.06; N, 4.39; S, 10.06. Found: C, 44.49; H, 5.02; N, 4.28; S, 9.75%. dH (270 MHz; solvent (CD3)2CO) 8.8 (s, 2H, –CH@N), 7.5–8.0 (m, 8H, arom), 4.4 (s, 4H, –N–CH2–), 1.2 (s, 18H, Me).

Table 2 Experimental details of the crystal structure determination of [Cu(SSalen)]BF4, [CuSSalpr)]BF4, [Cu(SSalbu)]BF4, [Zn(SSalen)Br2], [Cu(enSSal)Br2] and [Cu(en)OH2)2]SO4 Empirical formula Formula wt Crystal system Space group ˚) a (A ˚) b (A ˚) c (A a () b () c () ˚ 3) V (A Z l (Mo Ka) (mm1) Measured reflections Unique reflections Observed reflections R1 wR (all data) Number of parameters

[Cu(SSalen)]BF4

[CuSSalpr)]BF4

[Cu(SSalbu)]BF4

[Zn(SSalen)Br2]

[Cu(enSSal))Br2]

[Cu(en)OH2)2]SO4

C24H32BCuF4N2S2 563.00 Triclinic P 1 11.085(1) 18.478(2) 26.672(3) 100.454(8) 94.473(9) 94.599(8) 5331.0(10) 8 1.019 20,918 12,008 9238 0.0717 0.0871 1360

C25H34BCuF4N2S2 577.01 Orthorhombic Pna21 13.8003(3) 14.4820(4) 13.7185(2) 90.00 90.00 90.00 2741.72(1) 4 0.993 56,428 6257 5373 0.0436 0.0990 331

C26H36BCuF4N2S2 710.40 Monoclinic P21/c 14.4300(6) 15.3350(5) 15.9080(6) 90.00 113.013(2) 90.00 3240.0(3) 4 1.094 14,477 7404 4029 0.062 0.133 388

C24H32Br2ZnN2S2 633.79 Monoclinic C2/c 23.1700(11) 7.2740(3) 16.0890(9) 90.00 97.705(2) 90.00 2687.1(2) 4 4.057 4884 2623 1961 0.0431 0.0934 161

C13H20Br2CuN2S 459.73 Monoclinic P21/n 10.6718(5) 14.0199(9) 11.2804(7) 90.00 96.467(3) 90.00 1677.01(17) 4 6.181 35,369 3889 2163 0.0522 0.0749 175

C2H12CuN2O6S1 255.74 Monoclinic C2/c 7.1450(8) 11.6660(15) 9.7240(13) 90.00 105.993(7) 90.00 779.16(17) 4 3.068 921 921 808 0.0407 0.1008 77

A detailed description of [Cu(enSSal)Br2] and [Cu(en)OH2)2]SO4 is not given here but can be obtained from the site given in the supporting information.

254

M.K. Taylor et al. / Journal of Inorganic Biochemistry 100 (2006) 250–259

2.9. Crystal structure determination Crystals were coated in mineral oil and mounted on glass fibres. Data were collected at 123 K on a Nonius Kappa CCD diffractometer using graphite monochromated Mo Ka radiation. Full-matrix least-squares refinement was based on F2, with all non-hydrogen atoms anisotropic. While hydrogen atoms were mostly observed in the difference maps, they were placed in calculated positions riding on the parent atoms. The structure solution and refinement used the SHELX programs [24] and the graphical interface WinGX [25]. A summary of the crystallographic parameters is given in Table 2. 2.10. Electrochemistry Cyclic voltammetry was carried out using PC driven EG&G model 263a potentiostat with PC programme Powersuite. The voltammograms were acquired using a three electrode cell, which consisted of a platinum disc working electrode (0.00785 cm2), a platinum gauze auxiliary electrode and a solid Ag/AgCl reference electrode. Solutions of the three copper complexes (4 · 103 M) were prepared in the appropriate re-distilled solvent (acetonitrile or dichloromethane) using tetrabutylammonium tetrafluoroborate as background electrolyte (0.1 M). The potentials were scanned between 0.2 and 1.4 V at a scan rate of 50 mV s1. Voltammograms were also collected at a range of scan rates from 20 to 200 mV s1, to examine peak separations and peak current ratios to gain information on reversibility and diffusion control. Potentials are reported vs ferrocene, voltammograms of which were collected under the same conditions. 3. Results and discussion The three ligands N,N 0 -bis-(o-tert-butylthiobenzylidene)1,2-diaminoethane (SSalen), N,N 0 -bis-(o-tert-butylthiobenzylidene)-1,3-diaminopropane (SSalpr), and N,N 0 -bis-(otert-butyl-thiobenzylidene)-1,4-diaminobutane (SSalbu), were

prepared by the condensation of o-tert-butylthiobenzaldehyde with the appropriate diamine in ethanol [22,23]. SSalen (m.p. 74 C) and SSalbu (m.p. 60 C) were easily isolable as off-white solids. However, SSalpr resisted attempts to isolate it as a solid and was used directly as an impure oil. Each of these ligands were converted to their respective copper(I) complex using tetrafluoroborate as the counter ion by treating them with [Cu(NCMe)4] [BF4] in ethanol. To try and limit the lattice effects care was taken to crystallise each of the complexes from identical solvent mixtures (CHCl3: diethyl-ether; vapour diffusion at room temperature). All three systems produced orange crystals, which diffracted. However, with [Cu(SSalpr)]+ problems arose during refinement and only a partial solution could be obtained. In this instance we resorted to the use of methanol: diethyl-ether to obtain crystals from which we could obtain a data set. Comparison of the partial structure obtained in chloroform with that obtained from methanol indicated that the structures were essentially identical. Although the structure of [Cu(SSalen)]+ has been reported previously as the perchlorate salt [26], we required the tetrafluoroborate salt for the electrochemical study and to maintain continuity within the small series of complexes under study. As can be seen (Table 3), although there are only minor differences between the Cu–N and Cu–S bond distances, there is a significant (
5–10) geometric shift as a consequence of the different counter anions. By generating the homologous [BF4] salt, we are of the opinion that we have minimised the lattice effects to such an extent that a detailed structural comparison of [Cu(SSalen)]+, [Cu(SSalpr)]+ and [Cu(SSalbu)]+ can be made. 4. Structure descriptions We are particularly interested in the geometry adopted by the copper(I) centre in response to the increased amount of freedom these three ligands allow. To maintain continuity within the structural analysis we have focused on a select number of factors namely the M–N and M–S bond distances, the two trans S–Cu–N bond angles as a general description of the geometry at the metal centre and the tor-

Table 3 The N–Cu–S bond angles and N–Cu, S–Cu bond lengths and torsion (CN  NC) angles for the series of copper(I) bis-(o-tert-butylthiobenzylidene) complexes and copper(II) cyclopentenedithiocarboxylate [19] ˚ N–Cu–S Bond angles Bond length A Torsion angles E (V) E (V) N1-Cu–S2

N2–Cu–S1

N1–Cu

N2–Cu

S1–Cu

S2–Cu

CN–NC

MeCN (37.5)

DCM (9.1)

[Cu(SSalen)] [ClO4] [27] [Cu(SSalen)] [BF4] [Cu(SSalpr)] [BF4] [Cu(SSalbu)] [BF4]

130.8 135.2 124.8 119.4

132.6 142.6 141.1 128.9

1.99 2.00 1.95 1.98

2.01 2.03 2.05 2.00

2.24 2.26 2.29 2.27

2.25 2.27 2.29 2.27

28 30 12 24

0.10 0.24 0.34

0.14 0.30 0.45

[Cu(Cdten)] [19] [Cu(Cdtpr)] [19] [Cu(Cdtbu)] [19]

166.2 143.6 138.4

166.7 143.6 138.4

1.97 1.95 1.96

1.93 1.95 1.96

2.24 2.23 2.21

2.26 2.23 2.21

23 23 16

The oxidation potentials of bis-(o-tert-butylthiobenzylidene) species in a range of solvents were measured (dielectric constants in brackets). The bond lengths and angles are only given to 2 decimal places for direct comparison to those obtained from the relevant blue copper proteins shown in Table 1. All potentials are reported vs ferrocene, voltammograms of which were collected under the same conditions.

M.K. Taylor et al. / Journal of Inorganic Biochemistry 100 (2006) 250–259

N(1) N(2) Cu(1)

S(2)

S(1)

Fig. 2. The X-ray crystal structure of {[N,N 0 -bis-(o-tert-butylthiobenzylidene)-1,2-diaminoethane] copper(I)} cation [Cu(SSalen)]+. The geometry at the metal centre is intermediate between tetrahedral and square planar. The distortion in the ethane bridge holds the tert-butyl groups at positions above and below the N2S2 ‘‘plane’’ and cants the aromatic rings such that they lie at 74 to one another. The complex cation is not completely symmetric. While showing roughly similar Cu–S and Cu–N bond distances the two N–Cu–S angles differ by 10.

sion angle generated at the imines (NC  CN). These data for the three complexes are shown in Table 3. As can be seen (Fig. 2) [Cu(SSalen)]+ generates the complex which is the most heavily distorted (
255

angles differ by 10 the tert butyl groups lie at different angles (
C15

C4 C14 C13

N1

C11

C16 C6

N2

C5

S1

N2

Cu1

Cu1

S2 C20

C10

C19 C5 S1

C24

C10

C9

C22

N1 S2

C26 C22

C1

Fig. 3. The X-ray crystal structure of {[N,N 0 -bis-(o-tert-butylthiobenzylidene)-1,3-diaminopropane] copper(I)} cation [Cu(SSalpr)]+. The geometry at the metal centre is intermediate between tetrahedral and square planar. The chelate which incorporates the diaminopropane unit adopts a distorted chair configuration (NC  CN falls to 12) which assists the increase in the angle between the aromatic rings such that they lie at 107 to one another. Again the complex cation is not completely symmetric.

Fig. 4. The X-ray crystal structure of {[N,N 0 -bis-(o-tert-butylthiobenzylidene)-1,4-diaminobutane] copper(I)} cation [Cu(SSalbu)]+. Although, the seven membered chelate re-introduces significant torsional strain at the diimine region (NC  CN 24), the influence of the larger internal chelate (CuS2N2C4) ring does not increase the angle between the aromatic rings (112) significantly. Similar to [Cu(SSalen)]+ and [Cu(SSalpr)]+ the complex cation is not completely symmetric.

256

M.K. Taylor et al. / Journal of Inorganic Biochemistry 100 (2006) 250–259

distances vary by only a small amount (
1%, Table 3) it becomes possible to use these three species in an electrochemical study investigating the effect of the geometry on the redox potential of the copper center. The information obtained from such a study has some relevance to the effect of the metal binding site on the operating potential of blue copper proteins. 5. Electrochemical studies Subjecting the three copper(I)N2S2 complexes to analysis by cyclic voltammetry in acetonitrile (0.2–1.4 V) using tetrabutylammonium tetrafluoroborate as background electrolyte revealed that each of the complexes are capable of undergoing a quasi-reversible one electron oxidation in donor solvents (Table 3 and Fig. 5). The data also indicates

that the oxidation potential rises as the length of the alkly chain between the imine groups increases (Fig. 5). Thus, since the copper-donor bond distances remain the same within the synthesised series of complexes, the change in the oxidation potential can be ascribed to the change in geometry effected (Table 3) by the controlled extension of the methylene chain. Consequently, we observe an increase in the oxidation potential as the motif moves from the planar ([Cu(SSalen)]+) to a more distorted tetrahedral [Cu(SSalbu)]+ environment. This is consistent with an increase in the dominance of the tetrahedral motif preferred by copper(I) over the planar motif supported by copper(II). Studies by Nation et al. [27] on a related series of copper(I) N2S2 complexes show that the final product of the oxidation reactions can adopt a five co-ordinate geometry with the donor solvent occupying the fifth site. Although the even-

Fc A

B

C

3.0 2.0

Current / uA

1.0 0.0 0.2

0.4

0.6

0.8

1

1.2

1.4

1.6

-1.0 -2.0 -3.0 -4.0

Potential / V

0.50 0.45

Potential/ (V)

0.40 0.35 0.30 0.25 0.20 0.15 0.10 2

3

4

Number of Carbons in backbone Fig. 5. Top – The cyclic voltammograms of [Cu(SSalen)]+ (A), [Cu(SSalpr)]+ (B) and [Cu(SSalen)]+ (C) in dichloromethane (4 · 103 M) using tetrabutylammonium tetrafluoroborate as background electrolyte (0.1 M). The potentials were scanned between 0.2 and 1.4 V at scan rates 20, 50, 80, 100, 150 and 200 mV s1, in order to gain information on the reversibility of the redox process. However, for clarity only the 50 mV s1 scan is shown. The potentials are reported vs ferrocene (Fc), voltammograms of which were collected under the same conditions – Bottom. The plot of carbon backbone length versus oxidation potential for the three CuN2S2 complexes. The graph indicates that as the alkyl chain between the imine groups becomes larger and more flexible the redox potential rises. Furthermore, the redox potential increases as one moves to a solvent of lower dielectric strength (Table 3).

M.K. Taylor et al. / Journal of Inorganic Biochemistry 100 (2006) 250–259 L Cu

S N

257 L L

S S N

Cu

II

N assoc +L

S

S Cu N

N 2

L

Cu

S dissoc +L

Cu II

N 1

L

L

e-chem S

I

N

4b

2b

L

L

Cu

N

N

S

S

hydrolysis

L Cu

N

N

hydrolysis N

N

N 3

S

4

5

Fig. 6. The degradative pathway of bis-(o-benzylthiobenzylidene)-diamine complexes of copper(II). Figures in bold identify the individual species discussed in the text. It is unlikely that 2b and 3 form in the electrochemistry experiments as these species cannot be expected to demonstrate reversible behaviour.

tual product of the reaction is likely to differ from those being generated at the electrode in the electrochemical experiments reported here, we decided to repeat the experiment in a non-co-ordinating solvent (dichloromethane) in an attempt to observe any additional effects generated by the supporting medium. As expected the oxidation potential of each complex increased, respectively by approximately 0.05–0.10 V1 (Table 3) in the lower dielectric constant solvent. Crucially, the shift in oxidation potential with increasing diimine chain length is maintained.

Br2

C4

N1 C9

1

All potentials are reported vs ferrocene, voltammograms of which were collected under the same conditions. 2 5 was crystallographically characterized. Details of its structural determination have been deposited at the Cambridge crystallographic data center (ccdc 240046).

Cu1 S1 Br1

6. Oxidation products Nation et al. [26–29] have previously had some success in preparing the copper(II) complexes (2) and solvated adducts (2b) of their o-phenylbenzylidene complexes (Fig. 6). These studies encouraged us to re-attempt the isolation of the copper(II) o-tert-butylthiobenzylidene species (Fig. 6), which we believed to be forming in the electrochemical studies discussed above (Fig. 5). However, despite numerous attempts we were unsuccessful in this endeavour. Consistent with the earlier report of Nation et al. [26], the use of poorly coordinating anions such as the tetrafluoroborate and sulfate gave rise to the isolation of hydrolysed copper diamine complexes (5, Table 2).2 In the presence of coordinating anions such as bromide we could isolate products consistent with compound 3, which hydrolysed during recrystallisation to form 4a (Fig. 7). However, since it was problematic deciding whether the sulfurs were coordinated to the copper (2) or not (3) we decided to investigate the behaviour of the ligand further with an alternative divalent species i.e., zinc bromide. In contrast to the copper compounds it was a simple task to obtain a definitive structure of the zinc complex. This confirms that

N2

C1

C10 N2 S2

C8

S1

Br2 C7

C1

Zn1 N1

Br1

Fig. 7. The X-ray crystal structure of dibromo-(N-o-tert-butylthiobenzylidene-diaminoethane) copper(II) (above) and dibromo-(N,N 0 -bis-(o-tertbutylthiobenzylidene-1,2-diaminoethane)zinc (below).

the coordination sphere of the metal is made up of bromine and the SSalen ligand, which coordinates didentate through the two nitrogens (Fig. 7) and, as predicted by Nation et al. in their analysis of the zinc chloride adduct [26], the sulfurs remain uncoordinated. Although a degradative pathway is available in solution to the copper(II) adducts, it is unlikely to be occurring at the electrode in the electrochemical experiments as this would give rise to irreversible electrochemistry.

M.K. Taylor et al. / Journal of Inorganic Biochemistry 100 (2006) 250–259

Although we have been unable to structurally characterise the desired series of copper(II) adducts of the tert-butylthiobenzylidene based species, Berman et al. (Fig. 1b) have previously reported a relevant series of copper(II) cyclopentenedithiocarboxylate adducts [19]. Not only do these compounds maintain the required N2S2 environment (Fig. 1b) in the copper(II) form, they employ exactly the same method of enhancing flexibility (viz. the use of ethane, propane and butane chains) between peripheral rings. It is believed that these cyclopentenedithiocarboxylate complexes are similar to the products believed to be forming in the electrochemistry study (2) above. The most notable difference between these two sets of compounds however, are the presence of six and seven membered rings in the peripheral rings comprising the two S–Cu–N moieties. 7. Blue copper protein modelling The wider influence of the N2S2 coordination environment of blue copper proteins is a subject of some interest [4]. The redox potential of the metal centre can in principal be modulated by a number of factors including bond distances, coordination geometry and the dielectrics of the protein pocket. Although it is generally agreed that the bond distances, especially the Cu–Smethionine distance, are a major determinant of the value of the oxidation potential, the subtle role of the geometry of the metal centre as a finetuning element has thus far been difficult to assess due to an inability to deconvolute all the variables. Since the Cu–S distances are essentially identical (Table 3) within our suite of complexes our data does not shed any light on the influence of the copper-thioether bond distance on redox potential. However, as a result they can be used to explore the influence of geometry on redox potential as they maintain their bond distances (Table 3) while allowing the geometry to change. Thus, correlating the geometric parameters principally the two
145

140

135

Angle/˚

258

130

125

120

115 0.00

0.10

0.20

0.30

0.40

0.50

potential / V

Fig. 8. Plot of the N–Cu–S bond angles (Table 3) against the respective redox potentials for the three complexes studied in donor (j acetonitrile) and non-donor (N dichloromethane) solvents. The plot suggests that the two N–Cu–S moieties change independently of one another and that approximately a 16 mV/ change can be expected as the copper center distorts from a tetrahedral to a ‘‘planar’’ geometry. This assumes that the metrical parameters obtained from X-ray analysis are maintained in solution.

Our attempts to isolate the analogous copper(II) species for our tert-butylthiobenzylidene species were unsuccessful and consequently to obtain some direct comparison with the higher oxidation species we had to resort to the re-analysis of the homologous series of complexes (Table 3) reported by Bereman et al. [19] Although the N–Cu–S ring sizes differ between these two classes of compound, the bond lengths and plane angles around the metal centre are instructive (Table 3). We find only a minor compression of the respective M–N and M–S bond lengths between the two sets of complexes presumably due to the increased charge on the metal centre. The importance of the dielectric constant of the metal environment in blue copper proteins is also believed to contribute a small but significant amount to the redox potential of the system. Thus, the importance of the local environment around the binding site can be inferred from the study of the redox potential of these species in solvents with differing dielectric strengths. We have been able to shift the redox potential by values greater than 0.1 V simply by modulating the local environment of the CuN2S2 motif. The magnitude of the changes observed for the redox potentials as a consequence of dielectric strength (Table 3 and Fig. 5) are commensurate with the shifts obtainable as a result of geometric changes suggesting that these two factors may have a small but equal weighting in the protein. It is accepted that the Cu–methionine distance has a strong influence on the operating potential of blue copper proteins. The importance of the other three factors (the geometry of the metal centre, the dielectric strength of

M.K. Taylor et al. / Journal of Inorganic Biochemistry 100 (2006) 250–259

the protein ‘‘pocket’’ and the hydrophobic nature of the surrounding polypeptide matrix) is still open to debate. The effect of the surrounding polypeptide matrix can only be decided by isolating and characterizing further members of the blue copper protein family. However, this study would suggest that the environment in which the copper is held can have a profound, if small, effect on the redox potential of the protein. 8. Supporting information Details of the X-ray crystal structure determinations may be obtained from the Director, CCDC, 12 Union Road, Cambridge, CB2 1EZ, UK (fax +44 1223 336033; e-mail [email protected] or www:http://ccdc.cam. ac.uk) on request quoting the depository numbers ccdc 240041, 240042, 240043, 240044, 240046 and 281738. Acknowledgements MT would like to thank Strathclyde University for financial assistance. DES gratefully acknowledges the Wellcome Trust and the Nuffield Foundation for vacation scholarships. JR gratefully acknowledges the support of the Cunningham Trust. References [1] E. Capozzi, S. Ciurli, C. Luchinat, Struct. Bond. 90 (1998) 127. [2] S.K. Chapman, S. Daff, A.W. Munro, Struct. Bond. 88 (1997) 39. [3] A. Messerschmidt, Struct. Bond. 90 (1998) 37. [4] E.I. Solomon, R.K. Szilagyi, S. DeBeer-George, L. Basumallick, Chem. Rev. 104 (2004) 419. [5] R. Hille, J. Biol. Inorg. Chem. 2 (1997) 804. [6] W. Kaim, B. Schwederski (Eds.), Bioinorganic Chemistry: Inorganic Elements in the Chemistry of Life, Wiley, England, 1994.

259

[7] Z. Zhu, L.M. Cunane, Z. Chen, R.C. Durley, F.S. Mathews, V.L. Davidson, Biochemistry 37 (1998) 17128. [8] R.C. Durley, L. Chen, L.W. Lim, F.S. Mathews, V.L. Davidson, Protein Sci. 2 (1993) 739. [9] T. Inoue, H. Sugawara, S. Hamanaka, H. Tsukui, E. Suzuki, T. Kohzuma, Y. Kai, Biochemistry 38 (1999) 6063. [10] T. Inoue, M. Gotowda, H. Sugawara, T. Kohzuma, F. Yoshizaki, Y. Sugimura, Y. Kai, Biochemistry 38 (1999) 13853. [11] T. Kohzuma, T. Inoue, F. Yoshizaki, Y. Sasakawa, K. Onodera, S. Nagatomo, T. Kitagawa, S. Uzawa, Y. Isobe, Y. Sugimura, M. Gotowda, Y. Kai, J. Biol. Chem. 274 (1999) 11817. [12] J.M. Guss, H.D. Bartunik, H.C. Freeman, Acta Crystallogr. B 48 (1992) 790. [13] J.M. Guss, P.R. Harrowell, M. Murata, V.A. Norris, H.C. Freeman, J. Mol. Biol. 192 (1986) 361. [14] F.N. Buchi, A.M. Bond, R. Codd, L.N. Huq, H.C. Freeman, Inorg. Chem. 31 (1992) 5007. [15] R.L. Walter, S.E. Ealick, A.M. Friedman, R.C. Blake, P. Proctor, M. Shoham, J. Mol. Biol. 263 (1996) 730. [16] D. Zhao, M. Shoham, Biophys. J. 74 (1998) 233. [17] D.B. Rorabacher, Chem. Rev. 104 (2004) 651. [18] L.C. Nathan, J.E. Koehne, J.M. Gilmore, K.A. Hannibal, W.E. Dewhirst, T.D. Mai, Polyhedron 22 (2003) 887. [19] R.D. Bereman, G.D. Shields, J. Bordner, J.R. Dorfman, Inorg. Chem. 20 (1981) 2165. [20] J. Reglinski, S. Morris, D.E. Stevenson, Polyhedron 21 (2002) 2167. [21] J. Reglinski, S. Morris, D.E. Stevenson, Polyhedron 21 (2002) 2175. [22] M. Yamamura, M. Tadokoro, K. Tanaka, R. Kuroda, Bull. Chem. Soc. Jpn. 66 (1993) 1984. [23] K. Nakajima, Y. Ando, H. Mano, M. Kojima, Inorg. Chim. Acta. 274 (1998) 184. [24] G.M. Sheldrick, SHELX97 – Programs for Crystal Structure Analysis (Release 97-2). Inst. fur Anorg. Chemie, der Univ., Tammanstrasse 4, D-3400 Gottingen, Germany, 1998. [25] L.J. Farrugia, J. Appl. Crystallogr. 32 (1999) 837. [26] D.A. Nation, J.H. Reibenspies, M.R. Taylor, K.P. Wainwright, Inorg. Chim. Acta 258 (1997) 161. [27] D.A. Nation, M.R. Taylor, K.P. Wainwright, Dalton Trans. (1992) 1557. [28] D.A. Nation, M.R. Taylor, K.P. Wainwright, Dalton Trans. (1992) 2411. [29] D.A. Nation, M.R. Taylor, K.P. Wainwright, Dalton Trans. (1996) 3001.