Metal Clusters in Proteins - ACS Publications - American Chemical

or three (13,21) imidazole units from protein histidine residues coordinate to each ..... limited spectroscopic handles available for the d 1 0 diamag...
0 downloads 0 Views 2MB Size
Chapter 5

Models for Copper Proteins

Downloaded by UNIV OF CALIFORNIA SAN DIEGO on April 12, 2016 | http://pubs.acs.org Publication Date: June 21, 1988 | doi: 10.1021/bk-1988-0372.ch005

Reversible Binding and Activation of Dioxygen and the Reactivity of Peroxo and Hydroperoxo Dicopper(II) Complexes 1

Zoltan Tyeklar, Phalguni Ghosh, Kenneth D. Karlin , Amjad Farooq, Brett I. Cohen, Richard W. Cruse, Yilma Gultneh, Michael S. Haka, Richard R. Jacobson, and Jon Zubieta Department of Chemistry, State University of New York at Albany, Albany, NY 12222 Copper coordination complexes serving as models for copper­ -containing proteins which either reversibly bind or activate dioxygen are described. These include dinuclear copper(I) compounds which react with O2 resulting in the hydroxylation of the arene-containing ligand, thus mimicking the action of tyrosinase. Dicopper(I) complexes which can duplicate the reversible dioxygen binding action of hemocyanin are also described. A related set of hydroperoxo and acylperoxo dicopper(II) complexes can also be generated. Comparisons of the complexes described to related model systems are made and the biological relevance is discussed. Copper compounds have been established to be important and versatile catalysts for dioxygen-mediated reactions in both biological and synthetic systems (1-11). Studies on non-protein, low molecular weight compounds designed to mimic structural and spectral features as well as the chemistry of enzyme active sites have attracted a great deal of attention; these can and have contributed to the understanding of reversible binding and activation of molecular oxygen by coppercontaining proteins (1-9,11,12). In addition, information gained in these investigations may lead to the discovery of synthetic systems capable of effecting mild, selective oxidations of organic substrates with dioxygen, and the improvement of existing catalysts of synthetic chemical or even industrial importance. Our biomimetic investigations have focused on the metalloproteins hemocyanin (He) (11-17) and tyrosinase (11,12,14,16,18,29), which contain two copper ions in their active center. The function of hemocyanin is to bind and transport dioxygen in the hemolymph of molluscs and arthropods. Studies employing E X A F S spectroscopy have shown that in the deoxy form, two (19-21) or three (13,21) imidazole units from protein histidine residues coordinate to each cuprous ion. Upon addition of O2 to give oxy-Hc, considerable changes take place in the coordination sphere giving rise to tetragonally coordinated Cu(II) ions 1

Correspondence should be addressed to this author. 0097-6156/88/0372-0085$06.00/0 • 1988 American Chemical Society

Que; Metal Clusters in Proteins ACS Symposium Series; American Chemical Society: Washington, DC, 1988.

86

METAL CLUSTERS IN PROTEINS

bridged by an exogenous peroxide (O2 ") ligand. A n additional bridging ligand such as an oxygen atom from a tyrosine phenolate, serine hydroxyl, or water (i.e. as H2O or OH") had been suggested; this could account for the presence of a 420 nm optical absorption band in oxy-Hc and the observed diamagnetism resulting from strongly antiferromagnetically coupled ("super-exchange" via the bridging ligands) Cu(II) ions (11,22,41). The recent crystallographic studies on a deoxy-Hc (23-25) indicate that three imidazole groups are ligated to each copper® ion. Along with arnino acid sequence homology studies (23), phenolate and probably other protein-derived groups appear to be ruled out as possible bridging ligands. Downloaded by UNIV OF CALIFORNIA SAN DIEGO on April 12, 2016 | http://pubs.acs.org Publication Date: June 21, 1988 | doi: 10.1021/bk-1988-0372.ch005

2

deoxy-Hc Cu...Cu = 3.6 A

oxy-Hc Cu...Cu = 3.6 A

Tyrosinase is a monooxygenase which catalyzes the incorporation of one oxygen atom from dioxygen into phenols and further oxidizes the catechols formed to 0-quinones (oxidase action). A comparison of spectral (EPR, electronic absorption, CD, and resonance Raman) properties of oxy-tyrosinase and its derivatives with those of oxy-Hc establishes a close similarity of the active site structures in these proteins (26-29). Thus, it seems likely that there is a close relationship between the binding of dioxygen and the ability to "activate" it for reaction and incorporation into organic substrates. Other important copper monooxygenases which are however of lesser relevance to the model studies discussed below include dopamine p-hydroxylase (16,30) and a recently described copper-dependent phenylalanine hydroxylase (31). OH OH O

A Copper Monooxygenase Model Reaction In our studies with dinuclear copper(I) and copper(H) complexes (32), we have employed the dinucleating ligand, X Y L (1), in which two tridentate donor units are connected by a m-xylyl group (32-35). Ligand 1 was fashioned following the approach of Martell (36) and Bulkowski, Osborn and coworkers (89), in which ligands capable of placing two metal ions in close proximity (i.e., dinucleating ligands), were formed by appending multidentate groups to a xylyl moiety. Here, pyridine donors are utilized to mimic the aromatic nitrogen ligands (imidazole from histidine) known to bind to copper in the protein active sites. Dinucleating ligand 1 also contains a tridentate unit for Cu(I), a feature which would be desirable in modeling a three-coordinate deoxy-Hc (reduced) active site (35).

Que; Metal Clusters in Proteins ACS Symposium Series; American Chemical Society: Washington, DC, 1988.

5. TYEKLAR ET AL.

FY

Models for Copper Proteins

FY

Downloaded by UNIV OF CALIFORNIA SAN DIEGO on April 12, 2016 | http://pubs.acs.org Publication Date: June 21, 1988 | doi: 10.1021/bk-1988-0372.ch005

1

FY 2

PY = 2-pyridyl

FY

4

3

Ligand 1 does form a dinuclear copper(I) compound, 2, with roughly planar three-coordinate moieties. Complex 2 reacts quantitatively with O2 in the molar ratio 1:1 in dimethylformamide or ctichloromethane, resulting in the hydroxylation of the m-xylyl connecting unit in 1 and the formation of a phenoxo and hydroxo-bridged dinuclear complex 3 (33). The structure of 3 consists of two crystallographically independent but very similar copper(II) coordination environments (Figure 1). The geometry around each copper is essentially undistorted from square-based pyramidal, and Cu...Cu = 3.1 A. The copper(II) ions can be leached out of complex 3 to produce the free phenol 4, completing the sequence involving the copper-mediated hydroxylation of an arene. The observed oxygen atom insertion into an aromatic C-H bond and the stoichiometry of the reaction 2—»3 is reminiscent of the action of the copper monooxygenase tyrosinase; the reaction thus serves as a model system which may be of use in providing insights into possible mechanisms of copper-mediated dioxygen activation. While other copper complex systems have been reported to exhibit monooxygenase activity (11), the present system enjoys a number of advantages for further study, including the clean, high yield (> 90 %) reaction with O2 and definitive structural characterization of both starting compound and the oxygenated product. We have also found that reaction of hydrogen peroxide with the dinuclear copper(II) complex containing ligand 1 (prepared in situ) also results in the hydroxylation of 1 to give the same phenoxo-bridged dicopper(II) product (37). This observation suggests that a peroxo copper(H) species [Cu(II)2(02 ")] is a common intermediate in pathways developing either from Cu(I)2/02 or Cu(II)2/H202. Evidence for such an intermediate has also been obtained using copper(I) complexes of the ligands X Y L - F (38) and N5PY2 (39). The ligand X Y L - F is the 2-fluoro-substituted derivative of 1 (the 2-position is the site of hydroxylation). The ligand N5PY2 has a five-membered alkyl chain between the two tridentate donor units, instead of the m-xylyl group in 1. The dinuclear 2

Que; Metal Clusters in Proteins ACS Symposium Series; American Chemical Society: Washington, DC, 1988.

METAL CLUSTERS IN PROTEINS

Downloaded by UNIV OF CALIFORNIA SAN DIEGO on April 12, 2016 | http://pubs.acs.org Publication Date: June 21, 1988 | doi: 10.1021/bk-1988-0372.ch005

χ

ο ο

g IL υ ο 3, suggesting that these reactions proceed by electrophilic attack of a Cu(I)2/02 derived species upon the aromatic ring.

5

6

7

Que; Metal Clusters in Proteins ACS Symposium Series; American Chemical Society: Washington, DC, 1988.

Downloaded by UNIV OF CALIFORNIA SAN DIEGO on April 12, 2016 | http://pubs.acs.org Publication Date: June 21, 1988 | doi: 10.1021/bk-1988-0372.ch005

90

METAL CLUSTERS IN PROTEINS

We have recently prepared a number of 5-substituted (e.g., -NO2 or -f-Bu para to the site of hydroxylation) X Y L ligands, and their dicopper(I) complexes also give high yields of the corresponding hydroxylated dicopper(II) products (38). However, it seems that certain changes on the donor groups in 1 ehminate or drastically lower the monooxygenase activity. Thus, when a 6-methyl-2-pyridyl group is utilized instead of 2-pyridyl, the dicopper(I) complex is inert to 62 (38). Also, when one of the four -CH2CH2PY (PY = 2-pyridyl) groups in 1 is replaced by a -CH2PY unit, the yield of the hydroxylation is only ca. 40% and there is no hydroxylation at all when all four ligand arms consist of the -CH2PY unit. Other researchers have also studied systems related to these m-xylyl dicopper compounds. Sorrell and coworkers (48) have described the dinuclear copper© complex of a ligand which is structurally analogous to 1 with 1-pryazolyl groups replacing the 2-pyridyl groups as donors for copper. No ligand hydroxylation occurs upon oxygenation of the pyrazolyl-containing compound and a dinuclear dihydroxo-bridged copper(H) complex is isolated instead (i.e. fourelectron reduction of dioxygen took place). Casella and coworkers (49) have also described dicopper(I) complexes of analogous m-xylyl containing ligands having Schiff base nitrogen donors; their three-coordinate dicopper(I) complexes are completely inert to dioxygen. However, using m-xylyl dinucleating ligands containing two nitrogen donors per Cu(I) ion, Casella and Rigoni (50) observe hydroxylation in certain instances to give a phenoxo and hydroxo-bridged dicopper(II) compound. The course of the reaction depends critically upon the R group of the imidazole ligand donor; hydroxylation occurs with R = Me (monooxygenase stoichiometry Cu:02 = 2:1) but irreversible oxidation takes place withR = H ( C u : 0 = 4:l). 2

Ll R = H L2 R = Me

I

I

R

R

[Cu^Ll)] - - + 1/2 0 2

1

[Cu ^)] * +0 1

2

2

2

— • [

]

(no hydroxylation)

— • [Cu (L2-0-)(HO-)] n

2

2+

At present, we do not completely understand why only some of these very similar m-xylyl dicopper(I) complexes systems described above undergo ligand oxygenation reactions. However, based on the results outlined above, we can speculate on a number of aspects of this 02-activation process. Our studies implicate the presence of a copper-dioxygen (peroxo dicopper(II)) adduct as an intermediate in the oxygenation reaction and more recent kinetic studies (51) further support this conclusion. This adduct then either direcdy or via some further intermediate undergoes an electrophilic attack of the arene. The unique nature of this very fast reaction 2->3, and the observed inability to intercept the active

Que; Metal Clusters in Proteins ACS Symposium Series; American Chemical Society: Washington, DC, 1988.

Downloaded by UNIV OF CALIFORNIA SAN DIEGO on April 12, 2016 | http://pubs.acs.org Publication Date: June 21, 1988 | doi: 10.1021/bk-1988-0372.ch005

5. TYEKLAR ET AL.

91

Models for Copper Proteins

oxygenating species by the addition of external substrates (52) suggests that the active copper/oxy intermediate is formed in a proximity to the xylyl substrate suitable for reaction. Since it seems unlikely that replacing a pyridine by a pyrazole or Schiff base group would change the proximity of a copper/oxy intermediate to the xylyl ring, other explanations are required to account for the different courses of reactions in the these latter systems. The answer perhaps lies in the variation in electronic environment provided by these different donor groups as reflected by the difference in the E\/2 (cyclic voltammetry in CH3CN), which is 0.15 V more negative for 2 than for Sorrell's pyrazolyl complex (48). Thus, peroxo dicopper(H) intermediates formed in these analogous systems will have different characteristics, since dioxygen binding to Cu(I) is a redox process involving at least some degree of electron transfer from copper(I) to dioxygen (11). A related explanation has to do with the stability of the peroxo dicopper(II) intermediate, since it will either attack the substrate or 'decompose'; the kinetics of formation of the intermediate relative to those of the ensuing decomposition reactions will thus be important. Nelson (53) and Sorrell (90) have both described systems that undergo a Cu:02 = 4:1 reaction stoichiometry for dicopper(I) complexes where they propose that 'degradation' of the peroxo dicopper(II) intermediate proceeds by the fast bimolecular two-electron transfer from a second dicopper(I) molecule to the putative peroxo-dicopper(II) intermediate to give an aggregated oxo-copper(II) product. [The latter may form hydroxo-Cu(II) species in the presence of protic solvents]. Cu(I)...Cu(I) 2

+

Cu(II)-(0 -)-Cu(II) 2

0

2

^

2

Cu(II)-(0 -)-Cu(II)

Cu(I)...Cu(I)

2

->

ki/k.i

2 [Cu(II)-0-Cu(II)]

n

k

2

The stability of the intermediate will depend in part on the relative values of k i and k . The more negative Cu(II)/Cu(I) redox potential observed for the pyridyl complex 2 would probably result in a larger k i in the reaction of this compound with 0 , compared to the pyrazolyl analog. With the assumption that comparable k values would be observed in the two systems, the peroxo dicopper(II) intermediate will be more 'stable' (i.e. longer lived) for the case of the pyridyl donors. Further investigations with these and new synthetic model systems will be required to fully deterrnine the nature of 02-activation in these m-xylyl compounds and others. The systems described do provide a nice illustration of the model approach in bioinorganic chemistry. Although carried out by a number of different research groups having undoubtedly somewhat different objectives, studies of the structural and physical properties and reactivity studies of compounds which have been systematically varied with respect to parameters such as coordination number, chelate ring size and ligand donor type allow us to gain insights and draw conclusions concerning the possible mechanism of hydroxylation and to identify those special characteristics apparently present and/or required for 02-activation. 2

2

2

Reactions of Dioxygen with Copper Complexes - Reversible Binding While there has been a great deal of effort and success in the synthesis and

Que; Metal Clusters in Proteins ACS Symposium Series; American Chemical Society: Washington, DC, 1988.

92

METAL CLUSTERS IN PROTEINS

characterization of dioxygen complexes of cobalt (54,55) and iron (54-57), attention to copper has come only more recently. This is in part due to the difficulty in handling of the kinetically labile copper(I) and copper(II) complexes, and the limited spectroscopic handles available for the d diamagnetic copper(I) ion (6,7). However, in trying to mimic the behavior of the biological copper active sites, there has been recent progress in finding ways of stabilizing appropriate copper(I) complexes and their dioxygen adducts (6,7,11,58,59). In a recent review article (11), we have described systems either proposed or established to be dioxygen/copper adducts, including those systems which bind O2 reversibly. Caution must be taken in this field, since oxygenation of a Cu(I) complex which is accompanied by color changes, even if reversible, does not guarantee that a dioxygen adduct was formed. A number of criteria for establishing the existence of Cun/02 adducts should be used (12,54), and often not all of these can be fulfilled for legitimate technical/experimental reasons. Amongst the best characterized systems are those due to Thompson (6062). He has described both a monomelic superoxo copper(II) complex (61) and a dinuclear peroxo dicopper(II) compound (60). CuL(C2H4) [L = hydrotris(3,5dimethyl-l-pyrazolyl)borate anion] reacts with O2 to give CuL(02) which can be isolated as a stable crystalline diethyl ether solvate. The observed EPR silence, normal H NMR, and IR band at 1015 cm* (observed in the presence of 0 2 but not (>2) firmly establish the complex as a superoxide species. The binding is reversible since ethylene can displace the bound O2 ligand to again give CuL(C H4). When the copper(I) complex [Cu(TEEN)(C2H4)]C104 (TEEN = N,N,N',N'-tetraethylethylenecuamine) reacts with O2 in wet methanol, the solid blue product which can be isolated is a peroxo-dicopper(II) complex [ v ( 0 - 0 ) = 825 cm ], based on a variety of analytical data (60,62). Reversible behavior is indicated by the observation that (TEEN)Cu(I)-ethylene or carbonyl complexes can be isolated by treating the CU2/O2 complex with C2H4 or CO, respectively. We have found that the dinuclear phenoxo-bridged copper(I) complex, [Cu2(XYL-0-)] (8), can bind dioxygen reversibly (63,64). The x-ray structure of 8 (Figure 2) shows that it possesses some features which are strikingly similar to those of the proposed sites of either deoxy- and/or oxy-hemocyanin, including the Cu...Cu distance of 3.6-3.7 A, an 'endogenous' phenoxo bridging group, and an empty "pocket" where a second small bridging group (X), such as OH", N3-, CI", B r , and RCO2" (65) is known to coordinate in dicopper(II) complexes, [Cu (XYL-0-)(X)]2+. When an orange dichloromethane solution of 8 is exposed to dioxygen below -50 °C, an intense purple color develops due to the formation of the dioxygen adduct 9. Manometric measurements at -80 °C indicate that 1 mole of dioxygen is taken up per mole of 8 to give the product formulated as 9, [Cu 2(XYL-0-)(02 " )] . In resonance Raman spectra there are two enhanced vibrations, the copperoxygen stretch at 488 c n r (464 c m using 0 2 ) and the O-O stretch at 803 c m (750 c m with 0 2 ) (Figure 3). Oxygenation of 8 with mixed isotope dioxygen, 16Q-180, reveals that the peroxide is asymmetrically bound, since the copperperoxide stretch is split into two components at 465 and 486 c m (66). Recent E X A F S results have indicated a 3.31 A copper-copper separation for 9 (67). This finding rules out the |i-1,1-bridging peroxo geometry for [Cu 2(XYL-0-)(02 ")] (9) since structural data for the doubly bridged complexes [Cu 2(XYL-0-)(X)] show that a | i - l , l - X (X = oxygen atom) bridging geometry is only compatible with

Downloaded by UNIV OF CALIFORNIA SAN DIEGO on April 12, 2016 | http://pubs.acs.org Publication Date: June 21, 1988 | doi: 10.1021/bk-1988-0372.ch005

1 0

l

1

1 8

16

2

16

16

-1

+

II

2

n

2

+

1

-1

-1

18

-1

18

-1

n

2

n

Que; Metal Clusters in Proteins ACS Symposium Series; American Chemical Society: Washington, DC, 1988.

+

2+

Downloaded by UNIV OF CALIFORNIA SAN DIEGO on April 12, 2016 | http://pubs.acs.org Publication Date: June 21, 1988 | doi: 10.1021/bk-1988-0372.ch005

5. TYEKLAR ET AL.

Models for Copper Proteins

Figure 2. ORTEP diagram of the phenoxo-bridged dicopper(I) complex [ C u ^ X Y L - O - ) ] * (8), reference 64.

Que; Metal Clusters in Proteins ACS Symposium Series; American Chemical Society: Washington, DC, 1988.

93

94

METAL CLUSTERS IN PROTEINS

a Cu...Cu distance of < 3.15 A (65,68,69). Consequently, the geometry of the peroxide moiety in [ C u ^ Q C Y L - O - ) ^ ) ] (9) is either non-symmetrical ji-1,2bridging (e.g. axial to one Cu, but equatorial to the other) or terminal in character (i.e. O-0-Cu...Cu).

Downloaded by UNIV OF CALIFORNIA SAN DIEGO on April 12, 2016 | http://pubs.acs.org Publication Date: June 21, 1988 | doi: 10.1021/bk-1988-0372.ch005

2 _

10a 10b

+

L = CO L = PPh

3

The application of a vacuum while rapidly warming solutions of 9 results in the removal of O2 and regeneration of dicopper(I) complex 8 and quasi-reversible cycling between 8 and 9 using such vacuum-purge applications can be followed spectrophotometrically (64). Dioxygen can also be liberated by addition of carbon monoxide or triphenylphosphine to the purple solution of 9; the purple color fades and the bis(carbonyl)dicopper(I) complex 10a, or the bis(triphenylphosphine) adduct 10b ([Cu (XYL-0-)(L) ] , 10, L = CO, PPI13) is formed. Carbon monoxide can be removed from [Cu 2(XYL-0-)(CO)2] (10a) by applying a vacuum to give back [Cu 2(XYL-0-)] (8) and several cycles of oxygenation, dioxygen displacement by CO and decarbonylation can be carried out without a severe amount of decomposition, Figure 4. As previously discussed in another context, Borovik and Sorrell (90) have described a phenoxo-bridged dicopper(I) close analog of 8, but with pyrazole donors instead of pyridine. In their case, oxygenation even at low temperature results in a reaction with a 4Cu:02 stoichiometry, and no dioxygen complex is detected. As described above, they conclude that the kinetics of the ligand-complex oxygenation and ensuing decomposition reactions are unfavorable and that a metastable dioxygen adduct reacts rapidly with more dicopper(I) compound to give irreversible dioxygen reduction. Again, it seems likely that these kinetic differences are governed by the electronic environment around copper, as provided by the ligands, and that pyridine ligands favor a more rapid reaction of copper(I) with dioxygen compared to pyrazole, leading to a spectroscopically observable dioxygen-adduct (64). As alluded to above, we have also recently characterized a related series of dioxygen-copper complexes utilizing ligands like N5PY2, shown above. The generalized ligand type, NnPY2, contains a variable length (n) methylene chain I

+

2

2

I

J

+

+

Que; Metal Clusters in Proteins ACS Symposium Series; American Chemical Society: Washington, DC, 1988.

Downloaded by UNIV OF CALIFORNIA SAN DIEGO on April 12, 2016 | http://pubs.acs.org Publication Date: June 21, 1988 | doi: 10.1021/bk-1988-0372.ch005

5. TYEKLAR ET AL.

Models for Copper Proteins

95

300 3 5 0 4 0 0 450 5 0 0 5 5 0 Frequency (cm" ) 1

n

2

Figure 3. Resonance Raman spectra of the intraligand region of [Cu (XYL-0)(0 ")] (9) prepared with 0 , O , and 0 0 . (Reproduced from reference 66. Copyright 1987 American Chemical Society.) 2

1 6

l s

1 6

1 8

2

WAVELENGTH

(nm)

Figure 4. Absorption spectra showing the "carbonyl cycling" behavior of the dioxygen adduct [Cu 2(XYL-0-)(02 ")] , 9, reference 64. n

2

+

Que; Metal Clusters in Proteins ACS Symposium Series; American Chemical Society: Washington, DC, 1988.

2

+

METAL CLUSTERS IN PROTEINS

96

which connects two N3 tridentate groups, the same one used in the X Y L and X Y L O- dinucleating ligands. Dicopper(I) complexes [Cu2(NnPY2)] , containing two three-coordinate Cu(I) ions, react reversibly at -80 °C in CH2CI2 (Cu:C>2 = 2:1, manometry) to give 62-adducts, [Cu2(NnPY2)(02)] . These are characterized by multiple and strong charge-transfer bands in the visible region, e.g., ^ (£, M " cm- ) 360 (21400), 423 (3600), 520 (1200) nm, for the complex containing the N5PY2 ligand; this spectral pattern is unique in copper/dioxygen coordination chemistry (11) and also closely resembles that observed for oxy-Hc [ A , (e); 345 (20000), 570 (1000), 485 (CD) nm] (40,41). The observation of d-d bands at > 650 nm, and results from x-ray absorption edge studies (41,80) suggest that Cu(II) is present, thus the 02-adducts are best formulated as peroxo-dicopper(IT) complexes. As in the X Y L - O - containing system, carbon monoxide can displace the bound dioxygen (peroxo) ligand forming the bis-adduct, [Cu2(NnPY2)(CO)2r , undoubtedly occuring by forcing the dioxygen binding equilibrium towards the deoxy form, [Cu2(NnPY2)] , with subsequent reaction with CO. Since the ligands NnPY2 have no potential Cu...Cu bridging group, the results indicate that a bridging ligand besides O2 ' itself is not a prerequisite for systems capable of binding CO and O2 reversibly, nor for generating spectral characteristics reminiscent of oxy-Hc. Recent experiments indicate that complexes [Cu2(NnPY2)(02)] are diamagnetic (i.e. EPR silent, 'normal' NMR) suggesting that these dioxygen complexes possess a peroxo-bridging group capable of mediating strong magnetic coupling between Cu(II) ions. 2+

2+

m a x

l

f

Downloaded by UNIV OF CALIFORNIA SAN DIEGO on April 12, 2016 | http://pubs.acs.org Publication Date: June 21, 1988 | doi: 10.1021/bk-1988-0372.ch005

max

+

2+

2

2+

/ < ™ 2 > . \

2 +

^ ( C H

GO GO PY

I PY CO

PY ^ ^ P Y

/(CH ) 2

Cu

^

/ \ PY

)

n

\

2 +

GO GO

PY^I PY CO

PY^

2

(

11

py/ ^ P Y

n X

Cu

11

\ - PY

, / \ ? o FY 2

2

}

Hydroperoxo and Acylperoxo Dicopper(IT) Complexes Transition metal hydroperoxo species are well established as important intermediates in the oxidation of hydrocarbons (8,70,71). As they relate to the active oxygenating reagent in cytochrome P-450 monooxygenase, (porphyrin)MOOR complexes have come under recent scrutiny because of their importance in the process of (porphyrin)M=0 formation via 0 - 0 cleavage processes (72-74). In copper biochemistry, a hydroperoxo copper species has been hypothesized as an important intermediate in the catalytic reaction of the copper monooxygenase, dopamine (J-hydroxylase (75,76). A Cu-OOH moiety has also been proposed to be involved in the disproportionation of superoxide mediated by the copper-zinc superoxide dismutase (77-78). Thus, model Cu -OOR complexes may be of n

Que; Metal Clusters in Proteins ACS Symposium Series; American Chemical Society: Washington, DC, 1988.

5. TYEKLAR ET AL.

97

Models for Copper Proteins

relevance to active site intermediates in these and other copper proteins involved in dioxygen activation. In our studies on reversible binding and activation of 02 by dinuclear copper complexes, we have found that the dioxygen complex [Cu 2(XYL-0-)(02 *)] (9) reacts with acids to form a green hydroperoxo complex [Cu 2(XYL-0)(OOH )] (11) (Figure 5) (79). In fact, this species can be generated via three different routes: a) by protonation of 9 with one equivalent of HBF4*Et20 in CH2CI2 at -80 °C; here, spectrophotometric titration of [Cu 2(XYL-0-)(02 ")] (9) with the acid shows an isosbestic point indicating that only two species are involved in this reaction and thus 9 is straightforwardly converted to 11, b) by decarbonylation of the bis(carbonyl)-dicopper(I) complex [ C u ^ X Y L OH)(CO)2] (12) under reduced pressure at 0 °C, followed by oxygenation at -80 °C (Cu:02 = 2:1). Here, the uncoordinated and protonated phenol group serves as a stoichiometric source of H to protonate a putative dioxygen adduct of the decarbonylated form of 12, and c) by the addition of excess hydrogen peroxide to a dimethylformamide/CH2Cl2 solution of [ C u ( X Y L - 0 - ) ( O H ) ] (3), Figure 5. Attempts to isolate [Cu 2(XYL-0-)(OOH")] (11) as a solid to determine its structure have not been successful. However, the close similarity of the UV-vis spectrum of 11 [ J l = 395 nm, e = 8000 (M-cm)" ] with that of [ C u ^ X Y L - O )(OH)] (3) and a solution E X A F S derived Cu...Cu distance of 3.05 A in 11 (80) suggest that [Cu 2(XYL-0-)(OOH")] possess a p.-1,1 -bridged hydroperoxo moiety. Recently, we have prepared a derivative of 11, an acylperoxo dicopper(II) complex [Cu 2(XYL-0-)(RC03-)] (R = m-QC6H4,13), and its structure has been determined by x-ray diffraction (Figure 6) which shows that indeed the peroxide moiety has a |i-1,1-bridging structure (81). Complex 13 can be prepared by two different pathways, a) by reacting the hydroxo-bridged complex 3 with mchloroperbenzoic acid, or b) by acylation of the peroxo complex 9 with m~ chlorobenzoyl chloride, followed by a metathesis reaction (Figure 5). The reactivity of the dioxygen (peroxo) complex 9 is markedly different from that observed for both the hydroperoxo and acylperoxo complexes 11 and 13. As described above, the addition of triphenylphosphine to 9 results in the quantitative displacement of the dioxygen ligand with the concomitant production of the bis(triphenylphosphine) adduct, [Cu 2(XYL-0-)(PPh ) ] (10b). By contrast, the CU2-OOR complexes 11 and 13 react with one equivalent of triphenylphosphine to give essentially quantitative yields of triphenylphosphine oxide (79,81), Figure 5. These results are in accord with observations on other transition metal peroxide complexes where the oxidation of organic substrates is enhanced by the presence of electrophiles such as H+ or R C O (82-86). In the present case, protonation (or acylation) of the dioxygen-copper complex appears to result in activation via formation of the Cu -OOR species which is capable of transferring an oxygen atom to a substrate while leaving behind a stable hydroxo, 3, (or carboxylato, 14)-copper(n) moiety. Further mechanistic work is required to distinguish between a metal-based reaction or a pathway involving displacement of the coordinated RO2" ligand followed by its direct reaction with a substrate (8788). As suggested above, we can speculate that the biological relevance of the observations described above pertain to the activation of dioxygen in copper monooxygenases via protonation of a peroxo-Cu (derived from Cu(I) and O2) n

2

+

n

-

2+

Downloaded by UNIV OF CALIFORNIA SAN DIEGO on April 12, 2016 | http://pubs.acs.org Publication Date: June 21, 1988 | doi: 10.1021/bk-1988-0372.ch005

!I

2

2+

+

n

2+

2

n

2+

1

m a x

2+

n

2+

n

2+

I

+

3

2

+

n

n

n

Que; Metal Clusters in Proteins ACS Symposium Series; American Chemical Society: Washington, DC, 1988.

+

METAL CLUSTERS IN PROTEINS

Downloaded by UNIV OF CALIFORNIA SAN DIEGO on April 12, 2016 | http://pubs.acs.org Publication Date: June 21, 1988 | doi: 10.1021/bk-1988-0372.ch005

98

•H ι

ο υ Bis?

11 ν c ο S §

> Il

il* «si « ο 2 ο