Journal of the American Chemical Society
I68
1 102.1 /
January 2, 1980
Spectroscopic Studies of Stellacyanin, Plastocyanin, and Azurin. Electronic Structure of the Blue Copper Sites Edward I. Solomon,*la Jeffrey W. Hare, David M. Dooley, John H. Dawson, Philip J. Stephens,Iband Harry B. Gray* Contribution No. 5992 f r o m the Arthur Amos Koyes Laboratory of Chemical Physics, Calvornia Institute of Technology. Pasadena, California 91 125. Receiced March 19, 1979
Abstract: Low-temperature absorption and room temperature circular dichroism and magnetic circular dichroism spectral studies of the blue copper proteins Rhus c.rrnic$era stellacyanin, bean plastocyanin, and Pseudonronas aeruginosa azurin have been made. Low-energy bands attributable to d-d transitions in a flattened tetrahedral ( D ? d ) copper(l1) center are observed in the near-infrared C D spectra at 5000,9 150, and I I 200 cn7-I in plastocyanin, 5 2 5 0 , 8 100. and 10 500 cm-) in stellacyanin, and 5800 and I O 200 em-' in azurin. Bands in the near-infrared M C D spectra are at IO 600, 8800. and I O 500 em-' in plastocyanin, stellacyanin, and azurin. respectively. The low energies of these bands rule out tetragonal six- or five-coordinate as well as square-planar four-coordinate structures, and trigonal four or five coordination is inconsistent with the blue copper E P R parameters. The band positions accord well with ligand field calculations based on a tetrahedral structure that is distorted approximately 6' toward a square plane (?B>(d.,2-12) ground state), the predicted transitions being >B2 'E, 'Bl 2 B ~and , *Ai, in order of increasing energy. EPR parameters for the three blue proteins calculated using this ligand field model are in reasonable agreement with the observed values. Ligand field stabilization energy contributions to the reduction potentials have been calculated; azurin, plastocyanin, and stellacyanin are destabilized bq 250, 345, and 340 niV. respectivelq, relative to the aquo Cu(ll) ion. Based on a comparison of absorption and CD intensities, the characteristic bands at about 13 000, 16 000, and 22 000 cm-' in the blue proteins are assigned to the ligand to metal charge transfer transitions 7rS(Cys) d.,2-,.2, gS(Cys) d.,2-).2, and a S ( H i s ) d,,?-,,z, respectively, in a flattened tetrahedral C u ( l l ) unit.
-
-
4
-
-
-
The once puzzling physical properties of blue copper proteins* are beginning to be understood. Evidence that progress has been made comes from recent absorption, circular dichroism (CD), and magnetic circular dichroism (MCD) spectroscopic experiments on R h u s rernicifera ~tellacyanin,~ Pseudomonas aeruginosa azurin,? bean p l a ~ t o c y a n i ntree ,~ and fungal l a c c a ~ eand , ~ ceruloplasmin,' as in these cases the results have been interpreted successfully in terms of the electronic energy levels derived for a tetragonally ( D 2 d ) distorted tetrahedral blue copper site.3 These and other6 spectroscopic studies have suggested the probable ligand composition of such a distorted tetrahedral site. In the specific case of PO lar plastocyanin, an X-ray crystal structure analysis (2.7- resolution) has established' that its blue copper is bound to two nitrogen and two sulfur donors in a distorted tetrahedral geometry. The copper ion is coordinated by a cysteine thiolate, a methionine thioether group, and two histidine imidazole nitrogens. A 3-A resolution electron density map also has been determined for P . aeruginosa azurin.IOAlthough phasing was difficult and the resulting map noisy, the copper ion could be located and the probable ligands identified via use of the amino acid sequence. Once again, the ligating amino acids are cysteine, two histidines, and a methionine, and the copper-site geometry appears to be close to tetrahedral. Recent X-ray absorption spectroscopic measurements have shown' I that the Cu-S(thiolate) bond distance is relatively short (2.10 f 0.02 A),which is consistent with a near-tetrahedral geometry. Thus all the evidence now available suggests that the structures of the blue copper sites in plastocyanin and azurin are remarkably similar; however, at least one ligand at the copper center in stellacyanin must be different, as this protein does not contain methionine.' * We have now completed a comprehensive spectroscopic investigation of stellacyanin, bean plastocyanin, and P.aeruginosa azurin. The measurements that have been made include variable temperature visible and near-infrared absorption and room temperature near-infrared and visible C D and M C D spectra for the three proteins. An analysis of site symmetry effects on Cu( 1 I ) EPR parameters has also been performed. We report in this paper a detailed formulation of the groundstate electronic structures as well as complete d-d and charge
K
0002-7863/80/ 1502-0 168$0 I .00/0
transfer transition assignments for the blue copper sites in stellacyanin, plastocyanin, and azurin.
Experimental Section French bean (Phusrolus culgaris) p l a ~ t o c y a n i n ,Pseudonionns '~ (iwuginosu azurin.I4 and Rhus cernicifera stellacyanini5were purified by standard methods. Protein concentrations were determined from absorption spectral measurements i n the 600-nm Nearinfrared absorption spectra were obtained on protein films to minimize interference bq water absorption and permit a wide variation of tcinpcrature. The protein solutions werc first dialyzed against deionized. distilled water and then concentrated from roughly I O to 0.5 nil. by membrane ultrafiltration (Amicon). The sample was then furthcr conccntrated by placing a drop of protein solution on a Plexi g l disk ~ in a metal desiccator over Drierite. After 3 or 4 drops had becn successively concentrated ( 1 0 to 15 for thc thick films), the film \+;is prepared by transferring the disk to a desiccator containing a wturated potassium acetate solution. This controlled humidity allowed thc film to form slowly, thereby preventing most of the cracking caused b) rapid removal of excess water. The drying process was halted by transfer to a desiccator charged uith a saturated sodium hydrogen phosphate solution. The humidity of this desiccator allowed slow dissolution of the sample. Films that had been overdried were redissolved by placing them in the sodium phosphate humidifier. Whereas stcllacyanin and azurin appeared unaffected by complete drying and redissolving. plastocyanin never completely redissolved. The undissolved plastocyanin was white, indicating partial denaturation upon complete removal of hater. The plastoqanin samples used in the p r c x n t stud), therefore. uere never allowed todr) past the elastic film form. A lightly greased silicone gasket was placed on the plastic disk and ;I second disk placed over the gasket to prevent further drying upon cv;icuation in a cryocooler u n i t . The sample was masked with aluminuin foil to prevent light leakage and to improve thermal contact \4ith thc cold slation. The sample w a s then slowly cooled to minimum tcmperaturc. I f cracks developed. the unit was warmed to near room tcmpcraturc, slices of black masking tape were applied to cover large cracks. and the unit mas then recooled. Plexiglas ('/2 in. diameter b) in. thick) was chosen over quartz despite poorer optical quality (wcak vibrational overtones at 1250 nm and longer wavelengths) bccause the thermal contraction more closely matched that of the 11 rot e i n fi I ni . T h c near-infrared C D and M C D spectra were run in D20 or deutcratcd phosphate buffer in order to extend the spectral range to 2000 'Q 1980 American Chemical Society
Solomon, Gray, et al.
/
169
Stellacyanin, Plastocyanin, and Azurin
"---
0
1 STELLACYRNIN CD
D
t
d
I' STELLRCYRNIN MCO
15
1 0
1
+ 36000
33000
30000
27000
24000
21000
18090
15000
I2000
WRVENUMBERS
WAVE NUMBE R S
Figure 1. Top frame: the near-infrared and visible absorption spectra of stellacyanin films at 270 (- - - -) and 35 K (-); lower curves, left-hand scale; upper curves (thick film), right-hand scale. Bottom frame: Gaussian resolution of the 35 K near-infrared and visible absorption spectrum of a stellacyanin film; the symbols (+) represent the experimental absorption spectrum.
nm. Although the infrared overtones of water do not give rise to a nicasurable C D or M C D spectrum, they decrease instrumental sensitivity. Equilibration of azurin and plastocyanin samples with D 2 0 was carried out by concentrating the protein solutions in an Amicon ultrafiltration cell (UM-2 membrane) and then adding approximately a tenfold excess of D2O and repeating the procedure four to five times before concentrating the solution to i t s final value. Stellacyanin was equilibrated with deuterated 0.5 M phosphate buffer by twice evaporating a solution to a film and then equilibrating with D20 overnight; a film was then prepared a third time and dissolved in a minimum volume of buffer. Sample concentrations were 1.06, 3.54, and 3.56 mM for stellacyanin, plastocyanin, and azurin, respectively. Solutions for the visible C D and M C D spectra were prepared by diluting the solutions prepared for near-infrared studies with deuterated phosphate buffer of DzO until absorbance less than 1.0 was recorded for the 600-nm band. The final concentrations were 0.206,0.210, and 0.197 niM for stellacyanin, azurin, and plastocyanin, respectively. The absorption spectra were obtained on a Cary 171 spectrometer. Thc spectra were run at constant slit widths of 0.8 mm for X 750 nm. Wavelength variations in lamp output, monochromator transmission, and detector response, which are normally compensated by the slit feedback mechanism, were corrected by a fecdback system on the source voltage supply that was installed by Dr. G . R. Rossman. Higher resolution measurements (0.I -mm slit width in the near-infrared) revealed no fine structure at low temperatures. The samples were cooled by a Cryogenics Technology, Inc., Model 20 cryocooler equipped with quartz windows and a temperature controller that allowed adjustment of the temperature to & I K over most or the range used. The visible C D and M C D spectra w'ere recorded on a Cary 61 spectropolarimeter. The near-infrared C D and M C D spectra were rccorded on an instrument built at the University of Southern California and described elsewhere.Ih A field of 40 kG was supplied by ii Varian superconducting electromagnet. Instrument base lines were determined using deuterated phosphate buffer or D2O as a blank. Spectra a r e reported in terms of molar extinction coefficient (absorption) or differential molar extinction coefficient A€ ( C D and
Figure 2. Top frame: the visible CD spectrum of 0.206 mM stellacyanin in pD 6 deuterated phosphate buffer at 290 K. Bottom frame: the visible MCD spectrum of 0.206 mM stellacyanin in pD 6 deuterated phosphate buffer at 290 K; curve A, left-hand scale; curve B, right-hand scale.
MCD). In the case of MCD,
A6
is normalized to a field of +10
kG.
Results The 270 K absorption spectrum of a stellacyanin film from 28 000 to 7000 cm-l is consistent with solution data reported by Peisach et al." A new band at approximately 8800 cm-' is partially resolved when a thick film is cooled to 35 K (Figure I ) . Gaussian-resolved bands from the 35 K spectrum of stellacyanin are also shown in Figure 1. The position, intensity (c), half-width a t half-maximum (hwhm), and oscillator strength v) of each of the resolved Gaussian peaks as a function of temperature are given in Table I. Our visible C D curve for stellacyanin (Figure 2) accords well with the spectrum reported by Falk and Reinhammar.I8 The M C D curve (Figure 2), which has not been reported previously, shows bands at 16 500 and 22 000 cm-I. The intense bands beginning at 33 000 cm-' are probably attributable to tryptophan residues, by analogy to the interpretation of the M C D curves of azurin in this energy region.I9 The near-infrared C D spectrum of stellacyanin is shown in Figure 3. Bands are observed at 10 500,8 100, and in the region of 5000 cm-l. The whole of the latter band cannot be observed because of D20 vibrational absorption. The M C D curve (Figure 3) in the near-infrared shows positive and negative features at -10 800 and 8800 cm-l, respectively. Correlation of the near-infrared (Figure 3) and visible (Figure 2) M C D spectra reveals an additional negative feature in the 13 000-cm-l region. Electronic absorption spectra of French bean plastocyanin films are presented in Figure 4. The 270 K spectrum is similar to that reported for Chenopodium album plastocyanin.20 The major absorption band has a maximum at 16 700 cm-' and is assigned the solution c of 4500.*' Peaks are also observed at I 2 900 and 21 800 cm-l. The weak, high-energy band shows considerable asymmetry in the thick film spectrum (B), and a shoulder is resolved (-24 000 cm-I) a t 35 K. As the film is cooled, all the bands narrow and increase in intensity. The decreasing breadth of the 12 900-cm-' peak reveals residual intensity a t roughly 10 000 cm-I. A band a t this low energy has not previously been observed. Figure 4 also shows the
I70
Journal of the American Chemical Society
/
102:I
/
January 2, 1980
PLRSTOCYFlNlN FILN 0
0
I
,
-
3
'c
fd
/
\ e
A
i STELLACYANIN
MCO
4
13000
12000
11002
10000
9000
8000
7000
6000
500C
WRVENUMBERS
Figure 3. Top frame: the near-infrared C D spectrum of 1.06 mM stellacyanin in a pD 6 deuterated phosphate buffer a t 290 K; curve A, left-hand scale; curve B, right-hand scale. Bottom frame: the near-infrared MCD spectrum of I .06 mM stellacyanin in pD 6 deuterated phosphate buffer at 290 K ; curve A , left-hand scale: curve B, right-hand scale.
Table 1. Gaussian Analysis of Stellacyanin Absorption Spectrau 200 K
270 K
22 570 1196 0.012 1 I74
120 K Band 1 22 520 I265 0.012 1124
22 380 I457 0.014 Ill0
22 210 1512 0.01 2 942
17 750 967 0.015 1769
Band 2 17 630 990 0.017 1927
17 620 1 I34 0.016 1658
I 7 490 1213 0.016 1542
I6 430 876 0.037 4879
Band 3 I6 350 889 0.034 437 I
16 310 1012 0.035 3995
16 220 1090 0.033 3549
u=
12510 hwhm = 867 / = 0.0049 t = 657
12 590 846 0.0039 535
Band 4 12 600 78 1 0.0030 445
12 680 863 0.003 1 417
12 680 933 0.0027 341
V = I0910
1 1 050
1 I090
1 1 110
hwhm = 999 / = 0.0053 t = 610
I180 0.0064 627
I266 0.0066
1375 0.0067 565
T=35K
Y = 22 550 hwhm = 1 I85 f = 0.012 t = 1177
V = 17840 hwhni = 941.5 f = 0.01 3 E = 1631
V = I6 460 hwhm = 883 f = 0.039 t = 5046
60 K
Band 5 1 1 060 1219 0.0065 615
600
V and hwhm in cm-I.
Gaussian resolution of the 35 K spectrum of plastocyanin. The position, intensity, hwhm, and oscillator strength of each of the Gaussian band components in plastocyanin as a function of temperature are set out i n Table 11. The visible C D and MCD curves for plastocyanin are shown in Figure 5. The plastocyanin C D spectrum has the same general shape as the stellacyanin C D curve. The main difference in the absorption and C D spectra of stellacyanin and plastocyanin occurs i n the activity around 22 500 cm-I. I n
' I
I
I
I
I 8 0
30000
24000
18000
12000
6000
WAVE NU M B E R S
Figure 4. Top frame: the near-infrared and visible absorption spectra of plastoclanin films at 270 (-) and 35 K ( - - - ) : lower curves, left-hand scale: upper curves (thick film). right-hand scale. Bottom frame: Gaussian resolution of the 35 K near-infrured and vihible absorption spcctrum of a plastocbanin film: the sbnibols (t)represent the experimental absorption spectrum.
plastocyanin, two absorption peaks are observed, at 2 I 800 and 24 000 cm-', and the C D shows a negative feature a t 21 200 cm-' and a positive band at 24 000 cm-I. In this region stellacyanin exhibits only one absorption band and one C D feature, both at 22 500 cm-l. The MCD spectrum of plastocyanin shows a band at 14 000 cm-l, very weak activity at -22 500 cm-l, and stronger activity a t 36 000 cm-I. No M C D activity attributable to tryptophan is observed in the 34 000-cm-' region, consistent with the amino acid analysis of Milne and Wells." Low-energy C D and M C D curves for plastocyanin are displayed in Figure 6. Near infrared C D bands are clearly evident at 5000,9150, and 1 1 200 cm-I. The near-infrared M C D spectrum shows a minimum a t 10 600 cm-l. The absorption spectra of Pseudomonas aeruginosa azurin films are shown in Figure 7. The 270 K film spectrum matches the solution spectrum of Tang et al.'? Narrowing of the 12 400-cm-I band at 35 K reveals a weak absorption at -10 000 cm-' in the thick film spectrum (B). The resolved Gaussian bands from the 35 K spectrum of azurin are also shown in Figure 7 . The Gaussian position, hwhm, and oscillator strength of each component as a function of temperature are listed in Table 111. Our C D spectrum of azurin differs from that reported previously" only in that larger negative A€ values are observed in the region below 15 000 cm-l (Figure 8). The visible M C D spectrum of azurin has a negative band at 14 400 cm-I, a weak derivative-like signal centered at 24 300 cm-', and strong activitj associated with tryptophan at -34 000 cm-l (Figure 8). Bands are resolved at 5800 and -12 500 cm-l in the near-infrared C D spectrum (Figure 9). N o other low-energy bands are prominent, although there is an indication of very weak CD activity at about 10 000 cm-I. It should also be noted
Solomon, Gray, et al.
/
171
Stellacyanin, Plastocyanin, and Azurin 5 0
c
2 5
Ai
0 25
00
w
a
0
1
IU
-2 5
10
!
0 50
PLRSTOCYRNIN MCD
/
- 0 25
-0 5 0
PLASTOCYANIN M C D
t
05
00
u
-05
36000
33000
30300
27000
24000
2i000
18000
15000
12000
10
WAVENUMBERS
Figure 5. Top frame: the visible C D spectrum of 0.197 mM plastocyanin in pD 6 deuterated phosphate buffer at 290 K. Bottom frame: the visible MCD spectrum of 0.197 mM plastocyanin in deuterated phosphate buffer at 290 K.
that measurements on a reduced azurin sample revealed no C D bands in the near-infrared region. The near-infrared M C D spectrum of azurin exhibits a negative band at -10 500 cm-I (Figure 9). Discussion Theoretical Considerations. The lowest absorption, CD, and M C D features ( I 1 1 500 cm-l) are attributable3 to d-d transitions in the blue copper centers of the three proteins under study. The positions of these d-d bands ensure that the effective ligand field at the copper site is very weak, which clearly rules out structures based on tetragonal six- and five-coordinate as well as square-planar four-coordinate geometries. Absorption bands at 13 400- 14 300 and 17 100- 17 500 cm-I attributable to d-d transitions have been reported for several tetragonal CuNh’+ species.23The square-pyramidal Cu(NH3)s2+ ion has d-d electronic transitions a t 1 1 400 and 15 000 cm-l .23d The spectra commonly observed24 for square-planar copper(I1)amino acid complexes show d-d absorption in the ranges I 5 400-1 7 700 and 17 500-20 600 cm-’ .25 Bands a t 13 000 and 21 000 cm-I have been observed in the d-d spectrum of a square-planar Cu(l1) complex with an NzSz donor set.z6 Trigonal five- and four-coordinate Cu(I1) structures are inconsistent with the known ground-state (EPR) properties of blue copper (2.0 < g l < gll).’ Typical trigonal bipyramidal copper(I1) complexes show a “reversed” (gli = 2.0 < gl) axial EPR s p e c t r ~ m . ~ This ’ . ~ ~is the expected ordering for an axially compressed trigonal bipyramid (d,2 ground state for a oneelectron Even the Cu(NH3)2(NCS)3- ion, which contains longer axial bond lengths (to ammonia) than equatorial (to isothiocyanate), shows this “reversed” axial spectrum (the higher ligand field strength of ammonia compared to isothiocyanate presumably causes the effective field to be compressed). Expansion along the trigonal axis will eventually result in a ground-state change from 2AI(d,z) to ?E’(dx2-,2, d , , ), assuming D3h symmetry. Using the distorted trigonal bipyramidal energy matrices of Becker et al.,31the energy assignments of Slade et a]..” and the equatorial to axial bond length ratio 1.04 reported by Bernal et values of -1425 and -920 cm-’ are calculated for the radial parameters3’ Db
14000
11 500
9000
6500
too0
WAVENUMBERS
Figure 6. Top frame: the near-infrared C D spectrum of plastocyanin in DlO at 290 K; curve A, left-hand scale; curve B, right-hand scale. Bottom frame: the near-infrared MCD spectrum of plastocyanin i n D20 solution at 290 K .
Table 11. Gaussian Analysis of Plastocyanin Absorption Spectraa T=35K
60 K
120 K
200K
270K
Band I
V = 23 340 f = 0.0024 21 310 hwhm = 1418 1874 0.0060 f = 0.0035 370 E = 288
Band 2 22 010 2648 0.01 1 495
21 910 2757 0.012 518
22 200 2757 0.019 81 1
U = I7 870 hwhm = 985 f = 0.015 E = 1719
18 000 92 I 0.014 I736
Band 3 18 020 926 0.01 I 1314
17 880 I010 0.012 1402
18 100 1 I62 0.012 I I63
16 620 hwhm = 1003 980 0.046 f = 0.044 ~=5111 5407
Band 4 16 540 1063 0.049 5319
16 440 1 I48 0.043 4354
16 500
1301 0.049 4364
hwhm = 959 f = 0.01 1 c = 1354
13 390 962 0.012 1481
Band 5 13 300 966 0.01 1 1371
13 360 1027 0.01 I 2134
13 330 1108 0.012 1289
V = 1 1 990 hwhm = 910 f = 0.0096 c = 1222
I2 080 843 0.0084 1 I47
Band 6 1 1 930 910 0.0089 1136
1 1 870 I043 0.0 I O 1 I48
11 810 1125 0.01 1 1 I62
V = 9970 f = 0.001 3
I O 150 0.0013
Band 7 I O 000 0.0013
9820 0.0014
9530 0.0020
V = 21 390
V = I6 490
V = I3 350
a
V and hwhm in cm-’
172
/ 102:i / January 2 , 1980
Journal ofthe American Chemical Society
k-- '
I
I
I
I
5t
f
I
I
AZURIN C D
\
IO
,
1
1
,,
,
I
,
I
AZURIN MCO
05
$-
W
L
I
0
E
-&\\,
-_
0
*,':\
15000
\r
12500
10000
5000
7500
WAVENUMBERS
Figure 9. Top frame: the near-infrared C D sptctrum of azurin i n D20 solution at 290 K: curve A, left-hand scale; curve B, right-hand scale. Bottom frame: the near-infrared MCD spectrum of azurin in D20 solution ;it 290 K .
Table 111. Gaussian Analysis of Azurin Absorption Spectrau 7=35K
"-
60 K
120 K
200 K
270 K
20060 2090 0.0026 I45
Band I 20400 2984 0.0026 101
21 410 3680 0.0059 I85
20790 4144 0.0071 198
17600 1038 0.007 778
Band 2 17540 1099 0.0072 755
17590 0.0064 640
17 650 1252 0.0055 504
15860 1174 0.042 4167
Band 3 15820 I206 0.042 402 1
15810 1261 0.042 3826
15850 1334 0.044 3798
12920 1592 0.0083 605
Band 4 12800 1380 0.0067 566
12810 1500 0.0084 647
12830 1591 0.0094 686
0
:
I
Y = 21 450
RZURIN c3
hwhm = 2099 f = 0.0034 E = 185
V = I7 820
\
0
St
t
hwhm = 1078 f = 0.006 t = 641
V = I5 870 hwhm = 1210 f = 0.045 6 = 4278
V = 12 770 hwhm = 1468 f = 0.008 t = 632
5 a n d hwhm in cm-I.
N
36000
1 I56
33000
30000
27000
24000
21000
18000
15000
12000
WRVENLJMBERS Figure 8. Top frame: thc visible C D spectrum of 0.21 m M azurin in pD 6 deuterated phosphate buffer a t 290 K . Bottom frame: the visible MCD
\pcctruni of 0.21 mM azurin in pD 6 deuterated phosphate buffer at 290 K: curve A . left-hand scale: curve B, right-hand scale.
and Dq, respectively, for the CuCls3- ion. By varying the axial bond length, the 2E' term becomes the ground state at -2.8 A. Very long axial distances would be required for the electronic transitions to approach those observed in the blue copper
proteins. Although it is possible for the protein to constrain ligation around copper( 11) to such a coordination (essentially trigonal planar geometry), the predicted g values for such a configuration are different from those observed for the blue copper sites. Spin-orbit coupling splits the *E' state into r7 r s representations of the D3h double group.3' For a d9 system I'17 becomes the ground state, and there is no excited state of the same symmetry. The calculated g values for this state are g = 2 ( 2 k 1) and gL = 0, where k is the orbital reduction factor. Thus neither trigonal five-coordinate geometry will yield g values comparable to those observed in the proteins.
+
+
Solomon, Gray, et al. / Stellacyanin, Plastocyanin, and Azurin 5
173 z A
f
Figure 11. Reference coordinate system for the flattened tetrahedral (DZd) ligand field model.
1000
0
-1000
DC Figure 10. Plot ofgll and g, for Dq = 500 cm-I, k = 0.6, and Do/DT = 4 WitIiDpvaryingfrom +1000to-i000cm-'.
The compressed trigonal four-coordinate geometry will also have a 2Al(d,2) ground state for copper(I1). Thus, the "inverted" axial EPR spectrum with g1i N 2.0 and g l = 2 is again expected. This ordering is observed for copper(I1) doped into zinc oxide (811 = 0.74 and gl = 1.53).34 (The magnitude of the g values is a t variance with the theoretical prediction because the very small trigonal distortion in the zinc oxide site allows configurational interaction between the ground state and a spin-orbit component of 2E(t2), which has the same double group representation, thereby invalidating the perturbative approximation used33 to determine the g values.) Expansion along the trigonal axis allows 2E(t2) to become the ground state. For the large axial distortions necessary to account for the observed protein electronic absorption bands, the g values, using the perturbative method, are calculated to be gll = 2 ( k l - 2X0(2 k ) / 3 A 2 ) and g l = - 2 k X o / A l , where A , = 2Al 2E(t2), A2 = 2E(e) 2E(t2), and configurational mixing of the 2E(e) excited state with the ground state in the ligand field is ignored. For the general case where no assumption is made concerning the relative sizes of the ligand field and spin-orbit Hamiltonians, the energy matrices are diagonalized after applying both operators. Since either distortion produces a r4 (C3 double group repre~entation~j) ground state, only the r4 energy matrix need be diagonalized to predict g values. Using Ballhausen's trigonal basis set ((t2d,tZb] = E, t2O = A ] , [ea,eb] = E in C3L)and trigonal ligand field operator,33 the energy matrix shown in eq 1 is obtained.
+
+
-
-
from the spin-orbit Hamiltonian mixing t2Oa with (-t2b)fl. As the g value is defined by the equation g i b H , = ( \k@a1 p, I \kg5a)- ( \kSspI p, I \ k g s p , where pz is the z component of the Zeeman operator, the value is positive when b > a (tzOa is the ground state) and negative when a > b ((-t2b)P is the ground state). The EPR experiment, of course, measures only the magnitude of g. The values of gil and g l do not cross a t 2 (the limit of zero axial field) as the orbital reduction factor was also included in the spin-orbit operator, X(kl)s, yielding an isotropic value of 2.00(k). It is apparent from inspection of Figure 10 that the g values of blue copper (gl 2.25, g l N 2.05)2 cannot be accommodated by a trigonally distorted s t r u c t ~ r e . ~ ' ~ ~ ~ A tetragonally flattened tetrahedral Cu(I1) geometry is consistent with the EPR parameters of blue copper, as in this case we have 8kXo 2kXo 811 = 2 --and gl = 2 - (3)
-
-
A2
V = y2O [3z2- r2]
+ y4O [35z4 - 30z2r2+ 3r4] + [ x 4- 6 x 2 y 2+ y4] ~4~
(-ebb)
- D a - '/jDr - 4 0 9
-
--JZ I
I
--v L x
+ (X/2) - E
L
- d / ? D a+ ( 5 J ? / 3 ) D r
+a2
+ +
r;
+
( - 7 / 3 ) D ~ 6Dq - E
(4)
I =0
x
X
+
+ (5$2/3)Dr +
+ 6 D r - 4Dq - E
This matrix is similar to that of Liehr,36 who diagonalizes the trigonal basis set in the Td double group before adding the trigonal field and spin-orbit coupling. The ground-state wave function is therefore \ k l ~ ~=g a(-tzbp) ~ b(t2Oa) [(-eb&, where the coefficients a, b, and c are determined by diagonalization of the energy matrix. The g values can be written i n terms of these coefficients as follows: gl = 2[(-1 - k ) a 2 b 2 - Z f i k a c ] (2) g l = 2[b2- d k a b 2kbc] where k is the orbital reduction factor. A plot of gl and gl for a Dq value of 500 cm-I, an orbital reduction factor of 0.6, a Da to D r ratio of 4, and variation of D a from +lo00 to -1000 cm-' is shown in Figure 10. The unusual shape of gl arises
+
-400
2
2Ua
A3
where A2 = 2B1 2B2 and A3 = 2E 2B2.33For facile correlation with the square-planar limit, we have chosen a coordinate system (Figure 11) in which the tetrahedral t2 orbital set is ~ ( X Z ) b, z ) , ( x 2- y 2 ) ) .In this system a tetragonal distortion toward the square-planar configuration would be expected to yield a 2B2(dx2-y2) ground state for Cu(I1). The extent of this distortion is given by the angle 6,as defined in Figure 1 1. The ligand field potential should, therefore, retain this angle as a variable, along with the usual radial integrals. The D2d potential has the following form:
(1)
I
+
where yn"' = [ 4 ~ z e / ( 2 n I)a"+']Z;Y,"'(r,, O,, 4,) and Yn"'(r,, O i , 4') is a spherical harmonic. Evaluation of the spherical harmonics for ligands a t (a, /3,0),(a, (3, w ) , ( a , 180 - p, 7r/2), and ( a , 180 - 6, 3 ~ / 2yields ) V = ze(3 cos2 p
- I)]3z2 - r2]/a3
+ ze(35 cos4 p - 30 cos2 0 + 3 )
[35z4- 30z2r2
+ 3r4]16a5+ ze(35 sin4 p) [x4- 6 x 2 y 2+ y 4 ] / 1 6 a 5
(5) The values of the one-electron energies may be obtained by direct integration using this form of the potential. Our procedure, however, employed an operator equivalent form, which saves a great deal of computation. The operator equivalent
I74
Journal of the American Chemical Society
/ 102:I / January 2, 1980
cthylanimonium)tetrachlorocuprate(II) as the square planar limit (the structure of the former compound shows that the copper experiences significant axial interactions with neigh6000 boring ions; such interactions are absent i n the latter complex).j4 The important point to make here is that a reasonable cstiniate of the distortion in a D2d CuC14*- complex may be obtained from an analysis of its d-d spectrum. W e have not included spin-orbit coupling in the calculations, as such effects do not change the predicted fl values in the region of interest ( p L 60'). 0 Systematic changes in the absorption and EPR spectra and rcdox potentials of a series of pyrrole-2-carboxaldehyde Schiff base copper( 11) complexes45effected by decreasing the diheh dral angle 2w between the chelate ring planes by varying the 'r & bulkiness of the ligand side groups are also consistent with the * ligand field model. An earlier series of nitrogen-ligated cop0 a W per( 11) complexes of D2 symmetry was also characterized in z W terms of the dihedral angle 2w by Marakami et al.,46 who showcd the relation to the distortion angle p to be cos = sin CY sin w , where 2 a is the N-Cu-N angle within one of the chelate rings. The EPR spectra of a limited series of CuS4 cornplexes are also predictable in terms of the above ligand I n view of all the electronic spectral and EPR results - 12000 discussed in this and the previous section, we eye with suspicion to the effect that a distortion from square-planar geometry is not a prerequisite for explaining the observed physical properties of blue copper centers. Blue Copper Ligand Fields. For p values near the tetrahedral limit the order of excited states is 2E < 2 B < ~ *AI. Calculations on plastocyanin show that only p's near 60' can fit the three -18000 transitions and at the same time give reasonable energies for I the square-planar limit. Taking AW('E2 - 2B2) = 5000 cm-' and AW('Bl - Bz) = 91 50 cm-I, the procedure is to calculate Figure 12. Thc n n g u l n r depcndcnccs of t h e bl (d,2-,?). e ( d r z . d>..), bl AW('A1 - ?B?) for various values of p near 60". For /3 = 62', (d,) ), a n d a1 (d;2) hole energies a s ;I function of /3 for il d9 ion (07 = 7 6 5 cti1-l. 01 = 444 cm-1). Ds = 407 cm-I, Dt = 430 cm-I, L L W ( ~ A-I *B2) is calculated to be 9938 cm-' (AM'(?Al, - IBlg), the D4h limit, is calculated to be 14 066 cm-I); for /? = 58'. Ds = 1337 cm-l, Dt = 505 cm-I, A W ( 2 A ~- 2Bz) = 12 991 cm-I, D4h limit = 37 135 cxprcssion for Dzd symmetry is as follows: cm-I; for p = 60', Ds = 691 cm-', Dr = 465 cm-I, A W ( 2 A I V = :e(u?(3 cos? p - I)[3iz2- f 2 ] ( r 2 ) / a 3 - 'Bz) = 1 I 412 cm-I. D J / ~limit = 21 200 cm-I. Since the observed 'AI :eo4 (35 cos4 0 - 30 cos? p 3) 'B2 transition falls between 1 1 200 (CD) and 1 1 810 cm-' (absorption), fl = 60" is to be preferred. FurX [35ir4- 30i2i,? 251,? - 6 i 2 3(i2)'] (rJ)/16a5 thermore, the 13 = 58 and 62' calculations give unreasonably 35zea4 sin4 p (r4)/32a5 (6) high and low n 3 h limits, respectively, for a Cu"N2S2 complex. The I 1 8 IO-cm-' absorption band is several times more intense where cy. = 2/2IIcy4 = -2/63, and ( r n ) is the radial integral than the 91 50-cm-' band (Table V), consistent with DZd sel r ' 1 + 2 R ' (,?, , ) d r .The d electron hole energies are calculated lection rules: ?B, 'AI is electric dipole allowed but *Bz to be ?BI is forbidden. W(al = dr:) = -12(3 cos? p - 1)Ds - 36Dt Complete d-d band assignments and ligand field parameters for the three blue proteins are summarized in Table V. For all W(bl = d.,!) = 12(3 cosz - 1)Ds '/?(35 sin4 p - 6)Dt three blue copper centers the p values fall in the narrow range W(e = d,,, d?.,) = -6(3 cosz p - 1)Ds 26Dt 60-6 1 ', providing additional evidence of the close similarities in coordination structure. The ligand field parameters are not W(b2 = d,?-,>) = 12(3 COS' 3 - 1)Ds as well determined for azurin and stellacyanin as they are for -I/? (35 sin4 p 6)Dt ( 7 ) plastocyanin, as in the former two cases there are larger where 6 = 35 cos4 p - 30cos2 p 3, Ds = ~ e ( r ~ ) / 2 1and a , ~ uncertainties in the positions of the d-d transitions. The EPR spectrum of plastocyanin has been fit by an axial D t = ze (r4))/2I u 5 .These energy levels are plotted as a function spin Hamiltonian with the values A0 = -828 cm-I, gil = 2.226, of p (the distortion angle) in Figure 12. and g , = 2.053.'" Using eq 3, values of kli and k L are comModel Systems. We have tested the 6-parametrized D2d puted to be 0.3 1 and 0.16. respectively. These values are much ligand field model on a series of tetrachlorocuprates of varying ( 3 = 64-90') structures. The procedure involved variation of lower than those calculated for square-planar copper-peptide complexes, where k 11 falls in the range 0.64-0.55 and k I bethe radial parameters and the distortion angle p to obtain a best tween 0.71 and 0.45.?5 Using the orbital reduction factors fit with the observed d-d energies and the square-planar limit. calculated for plastocyanin and the observed energy splittings In all four flattened tetrahedral CuC14'- complexes the calfor azurin and stellacyanin, g = 2.20, g i = 2.05, and gl = culated p is within 2' of the experimental value (Table I V ) . Our calculations compare favorably to similar ~ o r k but ~ ~ 2.18, - ~ gl ~ = 2 . O j A y arc predicted for azurin and stellacyanin, respectively. Observed values are g,'= 2.260, gi = 2.052 for the derived radial parameters are somewhat better because a azurin5" and g; = 2.287, g,. = 2.077, g, = 2.025 for stellalarger set of chlorocuprates has been considered and cyanin.' Unusually low hyperfine coupling constants are P t ( N H3)4CuCI4 has been replaced by bis(N-methylphen-
+
-
+
+
+
+ +
-
+
+
+
+
-
Solonqon, Gray, et al.
/
175
Stellacyanin, Plastocyanin, and Azurin
Table IV. Structural and Electronic SDectroscoDic Data and Calculated Ligand Field Parameters for Tetrachlorocuprates ~~~
~
R(Cu-CI) (av), A
P(av), deg
cs+
2.23
6jh
Me3BzNf
2 27
66d
Me2H2N+
2 23
68 e
PhjAs'
2 26
73n
counterion
C(ob5d). cm-1
ti(calcd), cm-'
P(calcd), deg
5 175' 7 900 9 050 5 920"
5 193 7 610 9 070 5 514 1746 9 251 7 937 8 925 I O 407 9 266 10 929 1 1 957
63
415
345
64
415
345
68
400
350
73
445
320
9 200 5 750/ 9 200 I O 300 8 930g 12 200
Ds, cm-l
D I , cm-'
Each set of Ds and Dt values gives an acceptable fit to the D4h CuC14'- spectrum; experimental band positions for the N-methylphenethylammonium salt of C U C I ~ (0 ~ -= 90°, R(Cu-CI) = 2.27A) are 1 I 500,13 600,and 16 100 cm-' (R.L. Harlow, W. J. Wells 111, G . W. Watt, and S. H. Simonsen, Inorg. Chem., 13,2106 (1974));the D4h spectrum is fit exactly by Ds = 431 and Dt = 334 cm-l. J. A. McGinnety, J . Am. Chem. Soc., 94,8406 ( I 972).i' J. Ferguson. J . Chem. Phys., 40,3406(1964). C . Furlani, E.Cervone, F. Calzona, and B. Baldanza, Theor. Chim. Acta, 7,375 (1967). R.D. Willettand M. L. Larsen, Inorg. Chim. Acta, 5, 175 (1971).fR. D. Willett, J. A. Haugen, J. Lesback, ] in Cu~C16'-; R. D. Willett and C. Chow, Acra Crystallogr., and J. Morrey, Inorg. Chem., 13,2510 ( I 974).R Analysis is for the [ C U C I ~fragment Sect. B, 30,207 ( 1974).
Table V. d-d Bands and Ligand Field Parameters for Plastocyanin, Azurin, and Stellacyanin L F parameters
-absorption (270 K ) u, cm-1 E
C D (295 K)
A€
u , cm-1
- M C D (295 K ) u , cni-1 A€
Y
assignment
Plastocyanin
0= 60' Ds = 691 cm-l Dt = 465 cm-' DJh limit = 21 330 cm-l
-
-100
a
9350 I I 810 ( I 1 400lh
200 1 I62
5000 9150 I I 200
0.35 -0.10 1.1
10 600
-0.15
0.0035 0.0005 0.0009
0.54 -0.20
10 500
-0.18
0.0054 0.0016
8800 10 800
-0.08 +0.08
0.0045 0.0035 0.0042
Azurin
@ = 61' Ds = 700 cm-] Dt = 480 cm-l Dqh limit = 21 600 cm-l
a - I O 000" (I2 410)h
5800 I O 200
-I00 -I00 565
5250 8100 I O 500
Stellacyanin"
0 = 60' Ds = 765 cm-] Dt = 444 cm-l Dj), limit = 22 800 cm-'
-I00 -125
a 880OU 1 1 110 ( 1 1 500)h
0.45 -0.35 2.4
-
" Not resolved at 270 K; band positions and E'S are estimated from the 35 K absorption spectrum. Calculated position of 'B' 'AI. Parameters are for the fit to 5250-and 8800-cm-' transition energies; an equally acceptable fit to these energies is obtained for p = 61 O ( D s = 588, Dt = 430,D4h limit = 18 416 cm-I), with 'Bz ? A I at I O 843 cm-'. Alternatively, a fit to 5250 and 8100 cm-' for p = 61' (Ds= 620,Dr = 395,D4h limit = 18 849 cm-I) gives 2B2 'Al at 10 526 cm-I.
--
another spectroscopic characteristic of the blue site. Using the observed g values and the theoretical expressions derived by Bates et a1.,52values of A 11 and A l for stellacyanin, azurin, and plastocyanin were obtained: plastocyanin, A l(calcd) = 0.0115, All(obsd)20 = 0.0063, Al(calcd) = 0.011, AL(obsd)20 < 0.0017 cm-l; azurin, All(calcd) = 0.01 12, A ~ l ( o b s d = ) ~0.006, ~ Al(calcd) = 0.0016, AL(obsd)50 0 cm-I; stellacyanin, A;(calcd) = 0.0083, A , ( o b ~ d ) ~ = ' 0.0035, A,(calcd) = 0.0013, A , ( o b ~ d ) = ~ ~0.0029, A,,(calcd) = 0.0053, A , ( ~ b s d ) ~ ' = 0.0057 cm-I. The effect of differing radial integral values for nitrogen and sulfur ligands on the energy levels and g values predicted by the tetragonal field was also investigated, using a Czl geometry such that the twofold rotation axis remained along the tetragonal z axis. The energies of the electron hole in this field are identical with those given above for the D 2 d field if the average of the nitrogen and sulfur radial parameters is substituted for Ds and Dt. In addition, the d,, and d,,, levels are split by 18 sin2 p ( D s N - Dss) I O sin2 p(7 cos2 /3 - 1 ) ( D t N - D t S ) and d,2 mixes with the ground state dX2-)2. Using the values of p, Ds, and Dt obtained for stellacyanin, and adjusting
-
+
the ratio D t Y / D t s to be 1.55, the splitting of the d,,, d,, pair is calculated to be 3540 cm-I, consistent with the splitting calculated previously49 from the experimental g values of ~tellacyanin.~' Reduction Potentials. The reduction potentials of stellacyanin, azurin, and plastocyanin are 184, 330, and 347 mV, r e ~ p e c t i v e l y Perhaps .~~ the most important point to make is that the ligand field model with /3 60' (Figure 12) accounts in part for the fact that the blue protein potentials are more positive than that of the tetragonal copper(I1) aquo ion (153 mV), owing to the very low theoretical ligand field stabilization energies (LFSE) associated with the tetrahedral site.8d The LFSE for stellacyanin is only -6152 cm-l (-4144 cm-l attributable to the tetrahedral field and -2008 cm-' to the tetragonal distortion). For plastocyanin it is -61 13 cm-l and for azurin it is -6885 cm-l. The LFSE difference between plastocyanin and aqueous copper(I1) is 345 mV (Cu(1I) aquo is -8900 cm-1),8dThus, a value of 498 mV is predicted for the reduction potential of plastocyanin. Azurin is destabilized by 250 m V relative to hexaaquocopper(II), which yields a potential of 403 mV. Stellacyanin is destabilized by 340 mV with
-
I76
Journal of the American Chemical Society
/ 102:I /
January 2, 1980
Table VI. Charge-Transfer Spectral Data for Plastocyanin, Azurin, and Stellscyanin
protein plastocyanin
azurin
stellacyanin
:I
-
absorption nm
-
(I
CD (295 K) nm
A€
Y
13 330 16 500 18 100 22 2006 23 3406
752 606 552 450 428
1289 4364 1 I63 300 -I00
12 800 16 500 19 000 21 200 24 000
78 1 606 526 472 417
3.78 4.08 0.4 -1.32 1.26
0.0029 0.000 93 0.000 34 0.0044 0.013
AS US US* AN
12 830 15 850 17 650 20 790
179 63 1 567 48 1
686 3798 504 198
12 500 16 100 I 9 000 21 400
800 621 526 467
-5.9 6.5 1.2 -1.8
0.0086 0.00 17 0.0024 0.009 I
AS US US* AN
12 680 16 220 17 490 22 210
789 617 676 450
341 3549 1542 942
12 800 16 500 I 9 000 22 400
78 1 606 5 26 446
-5.0 3.6 0.75 -7.35
0.015 0.00 I O 0.000 49 0.0078
AS US
u , cm-1
u, cm-'
c
[' Positions and E'S for the proteins were taken from Gaussian analyses a t 270 K except where noted. Gaussian analysis at 35 K. See text.
assignment
KN
-
--
+
-
+
+
-
+
+
dX2-,2 dX2-)2 dx2-,2 d,2-,,,2 dX2-y2 dX2-?2 d,
