3224 Journal of the American Chemical Society / 102:9 / April 23, 1980

possible. One such interpretation is that the deviation of the slow phase from (pseudo-)first-order behavior is due to the presence of two first-order decays with similar (but not identical) rate constants. However, the excellent second-order fits place severe restrictions on the relative values of such rate constants; and, for this reason, the second-order interpretation is preferred.

the reactions (Scott, R. A., unpublished results). (29) Yu, C.-A.; Yu, L. Biochem. Biophys. Res. Commun. 1976, 70, 1115. (30) Gilmour, M. V.; Wilson, 0. F.: Lemberg, R. Biochim. Biophys. Acta 1967,

143, 487. (31) However, we emphasize at this time that one such reasonable alternative

is a simple bimolecular mechanism involving electron transfer from Ru(NH3)eZ+ to a3+.

(32) The intermolecular redox process could also involve a reduced copper site as the primar electron donor to a33+ (az+as3+(red Cu) + a2+a33+ - (ki~d s+a33Y(ox CU) + 82+a3'+). Additional experiments Will be required terpretations Of the second-order dependence of the slow phase are also

(33) Wherland, S.; Gray, H. 6. "Biological Aspects of Inorganic Chemistry", Addison, A. W., Cullen, W.. James, 8. R.. Dolphin, D., Eds.; Wiley: New York, 1977; p 289.

(34) Holwerda, R. A.; Wherland, S.; Gray, H. B. Annu. Rev. Biophys. Bioeng. 1976, 5, 363.

Models for the Active Site of Oxygen-Binding Hemoproteins. Dioxygen Binding Properties and the Structures of (2-Methylimidazole)-meso-tetra( a,a,a,a-o- pivalamidophenyl)porphyrinatoiron(II)-Ethanol and Its Dioxygen Adduct

Geoffrey B. Jameson,Ia Frank S. M~l inaro , '~ James A. Ibers,*la James P. Collman,*lb John I. Brauman,Ib Eric Rose,lb and Kenneth S. Suslicklb,c Contribution from the Departments of Chemistry, Northwestern Unicersity, Euanston, Illinois 60201, Stanford University. Stanford, California 94305, and Unicersity of Illinois, Urbana, Illinois 61801. Receiced August I , 1979

Abstract: When crystals of (2-methylimidazole)-~eso-tetra(oc,cu,cu,-o-pivalamidophenyl)porphyrinatoiron(~l)-ethanol, Fe(TpivPP)(2-Melm).EtOH, are exposed to dioxygen, the crystals of the resultant dioxygen adduct are still suitable for dif- fraction studies. The direct, precise determination of the stereochemical changes accompanying oxygenation of an iron(1l)- (porphyrinato)(base) complex has been carried out using conventional X-ray diffraction methods. The structures have been refined by full-matrix, least-squares methods, using 4176 and 5 183 reflections for the deoxy and oxy complexes, respectively, to A indices on F 2 of 0.1 62 and 0.1 20. For the portion of data where F02 > 3u(F02) the respective indices on F are 0.086 and 0.083. Crystal data for the deoxy compound follow: space group C:,,-C2/c, 2 = 4, molecular symmetry c '2 , a = 18.871 ( 1 I ) A, b = 19.425 (13) 8, c = 18.434 ( 1 1 ) A, @ = 91.48 (3)O, V = 6755.0 A', The oxy complex is nearly isomorphic with 2 = 4 in space group C2/c with a cell of dimensions a = 18.864 (5) A, b = 19.451 (5) A, c = 18.287 (5) A, 0 = 91.45 (2)O, V = 6707.0 A3. Some selected parameters for the coordination spheres, with those in square brackets pertaining to the dioxygen adduct, follow: Fe-Nporph = 2.068 ( 5 ) , 2.075 (5) t1.997 (4), 1.995 (4)] A; Fe-Nl, = 2.095 (6) [2.107 (4)] A. The iron atom is displaced 0.399 tO.0861 8, from the least-squares plane of the porphinato nitrogen atoms toward the imidazole ligand. The F e - 0 separation is 1.898 (7) 8. The average 0-0 separation is 1.22 (2) A and the F e - 0 - 0 angle is 129 (1)'. In the presence of ethanol the deoxy complex binds dioxygen reversibly, noncooperatively, and with lower affinity than when the sample is de- solvated-in the latter case dioxygen uptake has been found to be cooperative. The structure and properties of these possible models for T-state deoxy- and oxyhemoglobin are correlated and then compared with the I -methylimidazole analogue. The sterically active 2-methyl substituent appears to perturb the F e - 0 bond but not the Fe-31, bond.

Introduction The structural changes which occur a t the heme center upon

the binding of dioxygen, carbon monoxide, nitric oxide, and alkyl isocyanides to hemoglobins, normal and abnormal, are of great importance to an understanding of structure-function relationships. In particular, discussions on the mechanism of cooperativity in the binding of small molecules to hemoglobin have been strongly influenced by structural data and physical measurements obtained from simple model systems;2-8 this influence is manifested in the evolving concepts of P e r ~ t z . ~ . ~ ~ In the absence of precise data from protein crystal structures, an absence that arises not from faults in the experiment but is intrinsic in the nature of the problem, model complexes have been invaluable as a means of establishing the general geom- etry of small molecule-hemoglobin complexes"-I5 and of understanding some general structure-function relation- ships.

Complexes of Fe(TPP)( 1 -Melm)I6 with small molecules ( 0 2 , CO, and NO) can be taken as models for the corre-

0002-7863/80/ 1502-3224$01 .OO/O

sponding myoglobin adducts or for hemoglobin in the R (or high ligand affinity) states as the binding properties of the base, 1 -methylimidazole, closely resemble those of the biological axial base, histidine. The recent studies of abnormal hemo- globins, with either unusually high or low ligand affinities,I7 are of considerable use in understanding the behavior of normal hemoglobins. It is often difficult to find appropriate model analogues. But by adding a sterically active substituent to the axial base, such as exists in 2-methylimida~ole,~~~I~-~~ or by creating steric strain in a covalent chain linking an imidazole base to the porphyrin skeleton,*O affinities for small molecules are reduced. These systems at least mimic if not model low-

close structural correspondence between such model com- pounds and the metal sites of oxygen-binding hemoproteins has been established using techniques such as MO~sbauer ,**-~~ extended X-ray absorption fine structure (EXAFS),6 reso- nance Raman spectroscopy,2s%26 and infrared spectros-

affinity hemoglobins or T-state h e m ~ g l o b i n . ~ ~ ~ * ~ ~ ~ - ~ ~ A v ery

copy.2',28

0 I980 American Chemical Society

Ibers, Collman, et al. / Fe( TpiuPP)(2-Melm)-EtOH

Table 1. Crystal Data and Data Collection Procedures for Fe(TpivPP)(2-MeIm).EtOH (I ) and Fe(02)(TpivPP)(2-Melm)-EtOH (11)

3225

I I 1

formula formula wt space group a b C

P V Z

temp crystal shape crystal vol radiation linear absorption coeff transmission factors detector aperture

take-off angle scan speed A-' sin 8 limits

Pobsdr Pcalcd

background counts

scan range data collected P unique data unique data with Fo2 >

FeC7oN 1 0 0 5 H 7 6 1193.30 amu

18.871(12) A 19.425( 13) A 18.434(11) 8, 91.48(3)' 6755.0 A' 4

C4h - c2/c

l . l9(2) , 1.17 g c m V 3 21 "C needle, 0.21 X 0.29 X 0.71 mm 0.042 mm3 graphite-monochromated Mo Ka, X ( K a , ) = 0.7093 8, 2.73 cm-'

5.1 m m wide, 5.3 mm high

2.8' 2.0' in 28 per min 0.0554-0.528 1

initially I O s, then 20 s a t each end of

0.8' below Kal, 0.8' above K a 2 fh. k . I 0.04 4176

0.92-0.96'

32 cm from crystal

4.5 < 28(Mo Ka) < 44.0'

scan with rescan optionb

FeC7oN 1007H76 1225.30 amu

18.864(5) 8, 19.451(5) A 18.287(5) A 91.45(2)' 6707.7 8,' 4 -, 1.21 gem-' 21 'C needle, 0.27 X 0.29 X 0.75 mm 0.046 mm3 Ni-filtered, Cu K a , A(Ka1) = 1.540 562 A 22.6 cm-l

5.2 mm wide, 6.0 mm high

4.2' 2.0° in 28 per min

cqh-c2/c

0.44-0.62

32 cm from crystal

0.0198-0.5691 3.5 < 28(Cu Ka) < 122.5'

similar to I

0.9' below Kal, 0.9' above Ka2 h h , k , I 0.05 5183

3086 2263 3 d F o 2 )

' N o absorption correction applied for I . Lenhert, P. G . J . Appl . Crystaltogr. 1975,8, 568-570.

Using the compounds Fe(TpivPP)(2-MeIm)-EtOH (I) and Fe(O*)(TpivPP)(2-MeIm)-EtOH ( I I ) , we may probe quan- titatively the structural changes which could occur upon oxy- genation of T-state hemoglobin to give T-state oxyhemoglobin. Furthermore, this pair of structures provides a unique oppor- tunity to relate changes in geometry to ligand binding prop- erties in a situation where the crystal packing influences remain very similar. W e have already communicated the salient fea- tures of these s t r u c t ~ r e s . ~ ~ Here we provide a more complete description and discussion of our results. Experimental Section

Preparation of Compounds. The deoxy compound, Fe(TpivPP)- (2-Melm)eEtOH ( I ) , was prepared as dcscribed in ref 18. Crystals suitable for single-crystal X-ray diffractometry were obtained by recrystallization from refluxing ethanol, and were sealed under dry dinitrogen gas in thin-walled, quartz capillaries. The dioxygen adduct, Fe(O2)(TpivPP)(2-Melm).EtOH ( I I ) , was prepared by exposing crystals of I for 3 days at 20 'C to 1-2 atm of dioxygen gas which was saturated with ethanol vapor. Crystals were then rapidly sealed in ca pi I la r i es.

Crystallographic Data for 1. Precession and Weissenberg photo- graphs displayed symmetry and systematic absences consistent with the monoclinic space groups C$,-C2/c or C:-Cc. Diffraction data were collected on a Picker FACS-I automatic diffractometer with graph- ite-monochromated Mo Ka radiation. Lattice parameters were ob- tained as previously described29 by hand centering 14 reflections i n the range 0.2169 < A-' sin 8 < 0.2581-' using Mo Kal radiation ( A = 0.7093 A). Crystal data and details of data collection are sum- marized in Table I .

Crystallographic Data for 11. Formation of the dioxygen adduct was not accompanied by loss of crystallinity. With respect to I , no changes in symmetry or systematic absences and only small changes in unit cell constants were observable, although there were many significant changes in intensities of non-h0l reflections. Identical behavior was observed for all crystals of I I which were tested (more than five). Except where stated explicitly, experimental conditions similar to those for 1 prevailed. Data were collected in this instance with Ni- prefiltered Cu Ka radiation. Crystal mosaicities (peak width at half

u8)'- ,

\ C J A 3 )

C(E5)

Figure 1. Atom labeling scheme for Fe(TpivPP)(2-Melm).EtOH.

peak height for w scans) ranged from 0.20 to 0.30'; some reflections had a minor shoulder. Lattice parameters were derived from the set- ting angles of 16 reflections in the range 0.1952 < X-I sin 6' < 0.29 12

Solution and Refinement of Structure 1. The usual procedures, computer programs, atomic scattering factors, and anomalous dis- persion terms were e m p l ~ y e d . ' ~ . ~ ~ In the initial development and re- finement of the structure, the Northwestern University CDC6600 computer was used. In the final cycles of refinement the Lawrence Berkeley Laboratory CDC7600 computer was accessed by remote telephone connection.

Space group C2/c was assumed; consequently twofold symmetry is imposed on the molecular species. Subsequent successful refinement of the structure in this space group, negligible differences in the in- tensities of 544 Friedel pairs, and the failure of refinements in the alternative space group Cc to yield a better defined model justified the initial assumption. Since the solution and refinement of both this

A-' (CU K a l , A = 1.540 562 A).

FE 0

N O ) 0.098451281

C I I ) 0.115341361

Cl2l 0.19066140)

Cl31 0.219431361

C14) 0,16182137)

Cl5) 0.169251371

Nl21 0.043461271

Cl6) 0 . 1 1 3 5 8 ( 3 7 )

C17l 0.122411391

Cl8l 0.057601401

C19) 0.00749137)

C 110) -0.06589 137)

Nldll 6.229191361

CfA7l 0.24857159)

O t A 1 ) 0.309821471

CIA81 0.19345164)

C l A 9 ) 0.12199175)

CIA10) 0.20737l751

C I A l l ) 0.1992l151

Nf01) -0.11332151)

C 187) -0.1 I312 I80 I

0181) -0.11415l80)

CIB8) -0.10917153)

ClB91 -0.17984(59)

CI B I O ) -0.0798110)

CfBIIl -0.05914179)

CIA11 0.243621391

CIA21 0.284681491

c ( a 3 i 0.35398156)

C I A4 I 0.38 154 I 5 6 I

CIA51 0.342261501

CIA61 0.27242 144)

CIEI) -0.09457136)

Cl821 -0.09746142)

C f 8 3 ) -0.12192l51)

Cl841 -0.14257(46)

Cl05I -0.14079144)

-

Ct86) -0.11641142)

0.1173411781

0.137641271

0.13754(351

0.1421Ol391

0.145371381

0.142491341

0.14290134l

0. I 3 P I4 I261

0,141 00 ! 3 4 )

0.141191381

0. I 3 7 2 3 I 3 8 1

0.13613132)

0.133511341

0.26792141 I

0~331371661

0.34738 149)

0.38589161)

0.361041771

0.425361841

0.4303(10)

0.2470914l1

0.304871811

0.32381 154)

0.367601501

0.396331691

0.34676174)

0.4202hf73)

0.14603f44)

0.08800(451

0.09234161)

0.152741751

0.21100(59)

0.20947150)

0.12697143)

0.06311 ( 4 5 )

0.05664155)

0.11303l651

0.17535152)

0.18324148t

3226

Table 11. Final Atomic Parameters for Fe(TpivPP)(2-Melm).EtOH Journal of the American Chemical Society / 102:9 / April 23, 1980

A B A T O M x Y 2 61 I 822 833 812 813 LIP3 *..*......... O ~ ~ ~ ~ ~ ~ I . . . . . . . . . . . . . . . . . . . ~ ~ ~ ~ ~ ~ O ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ b * ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ W ~ ~ ~ ~ ~ O ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ o w ~ ~ ~ ~ ~ O ~ ~ ~

.............. t.............~.~.*.*.*~~~~~***~~**...~..***.*..*.*.~~~b..***~~~*....~o*~*~~.***b~..**+*..~**~o~+o~~~~*~~.~n*wo.o~o~ Estimated standard deviations in the least significant figure(s) are given in parentheses in this and all subsequent tables. The form of

the anisotropic thermal ellipsoid is: exp[-(bl I h 2 + b22k2 + b33P + 2b12hk + 2 b l 3 h l + 2b23kI) l . The quantities given in the table are the thermal coefficients X 1 04.

structure and I I were nonroutine more detail than usual is sup- plied.

From a sharpened, origin-removed Patterson synthesis coordinates for the iron atom and the two crystallographically independent por- phyrinato nitrogen atoms were deduced. Positions for remaining nonhydrogcn atoms were obtained from Fourier syntheses interspersed with structure factor calculations and cycles of least-squares refine- mcnt. Initial refinements were on F and utilized only those 2268 re- flections having FO2 > 3r7(FO2). Disorder of the 2-Melm ligand about the crystallographic twofold axis leads to near superposition of atomic positions. Consequently, in rigid group refinements of this moiety, its geometry was constrained to that shown in Figure 1. which alsodefines the atom labeling scheme for the remainder of the molecule. Initially the phenyl rings were constrained to Dbh symmetry (C-C = 1.395 A). All atoms in rigid groups were allowed individual isotropic thermal parameters. Values for R and R, of 0.156 and 0.190, respectively,

were obtained after two cjcles of least-squares refinemcnt of a model comprising 30 atoms (each with an isotropic thermal parameter) and three rigid groups (two phenyl and one 2-Melm). However, the i l l - defined solvate species, ethanol, which was disordered about the twofold axis continued to refine unreasonably and hence in subsequent calculations its geometry was constrained in a rigid group (C-C = 1.531 A, C - 0 = 1.465 A. C-C -0 = 109.5O).

Atoms not constrained in rigid groups were now assigned an an- isotropic model for their thermal motion. Two cycles of least-squares refinement yielded values for R and R, on F of 0.1 15 and 0.136, re- spectively. A search for hydrogen atoms was then conducted. The very high apparent thermal motion of the methyl carbon atoms of the pi- valamide groups precludcd an entirely unambiguous location of the methyl hydrogen atoms. Nonethcless, all hydrogen atoms, with the exceptions of those on the solvate species and the pivalamide nitrogen atoms, were included at their calculated idealized positions as a fixed

1 /4

0.20811 1271

0. I3540 I371

0.12936140)

0.197071391

0.24601138)

0.32209 1381

0.35221127)

0.371 06 1361

0.44892(37)

0.476101371

0.416301361

0.42180134)

0,369271431

0.36099159)

0.36541 I691

0.349051591

0.3474 I141

0.28433(811

0.4142112)

0.49828 I41 I

0.51568l601

0.58225(62)

0.46839 (591

0.45544183)

0.3981fl1901

0.4928 110)

0.353131361

0.35921 1481

0.38525 (591

0.401551621

0.402631521

0,375321431

0,497271381

0.53027(451

0.59906154)

0.63653145)

0.60584142)

0.5354614))

29.71151) 36.70155) 24.70(501 0

30.7 120)

32,01271

37.5 I 3 0 1

31 0 1126)

31 a 9 (26)

32 e 5 (26)

30.7 I21 I

30 -6 I261

38 a5 129)

38.3 1291

35.1 1271

36.91271

39.3 128)

44.3 (42)

60.0 I381

72.31531

59.7159)

106.4 I781

327.124)

110.9(501

139.5 I851

308. I 1 31

53.21401

49.81471

219.(14)

I20 .o 1841

34.3 129)

48.5 1381

39.31401

45.1 1451

36.0 I36 I

35.6133)

35.2(27)

47.8 I35 I

58.0 1431

48.2 136)

50.5 136)

46.5 133)

39.2 I231

38.6 I301

45 -6 133 I

47.9133)

36.4 1281

32.0 (27 I

36.6 I2 31

35.7 128 1

46.9 I331

48.3 I341

28.8 I26 I

31.01271

37.u1271

60.2151)

95.5 148)

51.9(421

82.5 1751

163,110)

129. I I 1 I

32.5 (271

81.7(711

85.3150)

46.5 139)

106.1 (71)

81.0 I681

82.61681

44.7 132)

45.1 1361

76.0156)

88.4 I67 I

76.21521

51.21381

40.51311

47.61361

66.2149)

79.3153)

66.6 I44 I

47.9136)

22.8120)

27,61261

35.5 131 I

32.4 (291

31.41271

3 3 . 1 1281

21.51191

27.0 I26 1

25.5(26)

27.9127)

28.3 (25)

24.1 I241

80.3(41)

81.2155)

265. I 1 1 )

64.7 149)

393.1241

135.7 I 9 8 I

152.1141

37.3131)

26.7 I40 1

9 3 . 4 1571

61.8146)

178. ( 1 0 1

120.2(911

228.115)

25,5126)

58,8142)

70.2153)

70.4153)

63.1f461

45.6 1 3 5 )

29,21271

39.7 (36)

40.0141)

29.1133)

24.1 (301

27,0130)

-1.71161

-0. I(21 I

2.71241

-4.8 (221

- I , 0 1 2 1 )

0.2120~

-3.61161

-0.21211

-3.5124)

-2.1125)

-2.0 I 20 )

-3.51211

-3.91K1

-7.7(421

-22.3 I35 I

0.01431

23.6151)

47.3(70)

86.114)

4.5 1321

12.91671

-12 -6 158)

0.4132)

2.61471

9.3 I761

-31 , I (63)

-0.7126)

5.4130)

15.4 1 3 6 )

- 3 . 3 I46 I

- I 0.8 (351

-2.91291

0.2125)

-0.8 177)

-7.8(35)

-4.1 1401

6.0132)

-0. I I781

- 1 -86 1371

-0.8116)

3.91211

7.11241

2.5(23)

0.1 1221

-3.5 (22 I

-1.3116)

-4,6121 I

-0.8 I231

-4 .1 (231

-2.8(21)

2.9(211

-1.2126)

15.2 139)

-4 .6151)

3 1 . 4 1 4 1 )

55.5 192)

44.7 171)

ie..(14)

15.3(291

1.11471

-3.31701

-1.71351

-8.9 I55 I

82.6(931

-108,8192)

-2.71211

-5.4 132)

-8.1136)

-28.1 138)

-14.6 (32)

-3.3 126)

0.2121)

2.9128)

4.21321

9.3126)

8.61261

3.2 124 I

0

2.0116)

-0.8(211

3.01241

1.1124)

2.01221

-2. I(21l

-1.6(161

-2.7(201

1.5 (221

-2.61231

-1.0 I201

-0.1 ( 1 9 )

-16.21281

1 0 . 4 1 4 3 )

67.6152)

0.3(41)

81 . (11 )

92.1 186)

-55.110)

- 1 1.8 (251

5.7148)

1.8145)

4.5(371

76.4170)

54.2(66)

46.3 180)

1.9(231

13.01301

20.6l411

8.01501

-6 - 5 1 4 0 )

-6.5 1 3 0 )

4.7(26)

6.2(28)

18.6 1 3 5 )

15.6117)

-1.0129)

6.41271

Ibers, Collman, et al. Fe(TpicPP)(2-Melm).EtOH 3227

Table Ill. Rigid Group Parameters for Fe(TpivPP)(2-Melm).EtOH 2 2

ATOM X Y z 8.A ATOM X Y z B l A .................................................................................................................................. I N [ ) ) O.nI409157) fl.01099I30) 0.23843161) 4.941261 EOII) -0 ,0564119) -0 .2272115) 0.3OOI115l 17 .911) )

1 N l 2 ) 0.00658170) -O. l0260(30l 0.24701171) 6 . 4 7 1 2 4 ) ECIII - 0 . 0 3 0 5 i 1 4 ) -0 .28209195) 0 .25331151 7.97 I871

IC(]) -0 .01589146) -0 .04256134) 0 .27040148) 6 .17139) ECl2l -0 .0036 l l81 -0.2507116) 0 . 1 8 2 3 1 1 3 l 8 .27 I96 )

I C t Z I 0 .05492IhO) -0.08768147) 0 .19585162) 6 .49138) E O 1 2 1 0 -0.4709117l 1/4 22 .3 I I I )

I C I 3 ) 0.05733147) -0 .01996148) 0 .19292151) 6 .69143) E C l 3 ) -0 .0411117) -0 .4 lOI117) 0 .2280135) 2 2 . 3 l 1 1 )

ICIMEI -Q.O7166lLOl - 0 . 0 3 3 9 3 ( 6 5 ) 0 .32926161) 6 .65142) E C l 4 ) 0 -0.34451171 1 / 4 22.31 1 1 ) .................................................................................................................................. R I G 1 0 GROUP P A R A M E T E R S

A B

C C C GROUP x Y 2 DELTA EPSILON E T A .................................................................................................................................. I M O l o . o n r 2 1 4 2 ) -0.031261311

E T 0 H - l -0.0564 I 1 8 ) -0 .22721151 0 .3001115) 0 . 0 1 1 ( 4 3 ) 2 .932 124) -1 .155125)

ETOH-2 0 - 0 . 3 4 4 5 1171 I /4 -1 .5708 -2 .681184) 0

-0 .8024144) 0 .23166145) 0 . 1 5 1 4 1781 -2 .9887176)

.................................................................................................................................. a X C , Yc, and Z C are the fractional coordinates of the origin of the rigid group. The rigid group orientation angles delta, epsilon, and

eta(radians) have been defined previously: La Placa, S. J. ; Ibers, J . A. Acta Crystallogr., 1965, 18, 51 1-519.

contribution to F,. A C-H separation of 0.95 8, and, for methyl hy- drogen atoms, a C-C-H bond angle of 109.5' were assumed. The thermal parameter for a hqdrogen atom was fixed at 1 .O A* greater than the equivalent isotropic thermal parameter of its attached carbon atom.

When, after further cycles of least-squares refinement, now on F 2 and using all data including Fo2 < 0, a region of electron density along and near the twofold axis persisted, a closer investigation of the solvate species was undertaken. As ill-resolved solvate species also plagued the structure analysis of I I , parallel calculations were performed on both structures in the hope that the problem could be unraveled in a synergistic manner. As deduced from Fourier maps, the best model for I and also for 11, judged both on chemical grounds and by subse- quent successful least-squares refinement. involved the disorder of one ethanol molecule per asymmetric unit over two crystallographi- cally independent sites. At one site (occupancy a) the ethanol oxygen atom was hydrogen bonded to the noncoordinating nitrogen atom of the imidazole ligand. The other site (occupancy I - a ) was also close to the twofold axis but was considerably further from the imidazole ligand. There is additional disorder created by the twofold axis.

Group constraints on the phenyl rings were relaxed and this final model, described by 373 variable parameters, was refined to satis- factory convergence. The final values for R and R , on F* are 0.140 and 0.204, respectively, and the standard error in an observation of unit weight is 1.62 e?. For the portion of data having F02 > 3u(FO2) the conventional values of R and R,. on F a r e 0.086 and 0.098. The final value for the ethanol occupancy parameter cy is 0.68 (2), but because of correlation the final thermal parameters and the distri- bution of the ethanol molecule between the two solvate sites should not be taken too seriously. There wJas some dependence of the mini- mized function on lF,I and A-' sin 8, but this may be attributed to the inadequacies in this or any other model in coping with the very high thermal motion and/or unresolved disorder u hich afflicts many parts of the structure. The final difference Fourier synthesis WBS mostly flat and featureless, with the highest peaks being near the iron atom (height 1.01 (9) e k3); the height of the next peak is 0.58 e A-3. I n earlier syntheses peaks associated with the minor solvate site (final refined occupancy 32%) had been in the range 0.82-0.58 e Tables I I and I l l contain final nonhydrogen atomic parameters. Hydrogen atom parameters are given in Table IV.3' Table V lists the values of 10IFOI vs. 10(F,1.3' A negative entry indicates that Fo2 5 0.

Solution and Refinement of Structure 11. Starting from the coor- dinates for the meso-tetraphenylporphyrinato component of I , the structure of I I was developed in a similar manner with only those 3205 reflections having Fo? > 3u(Fo2) being utilized initially. As i n I , the methyl carbon atoms of each pivalamide group were observed as three distinct, though diffuse, regions of electron densitq; disorder similar to that observed for the Fe(Oz)(TpivPP)( I -Melm) structure'* was not apparent. Unfortunately, the terminal oxygen atom for I I occu- pied, unequally, two crystallographically independent sites, and hence

suffered from disorder. In order to establish a meaningful overall occupancy for the dioxygen molecule and a meaningful apportionment of the terminal atom between its two sites the following refinement was conducted. Variable occupancies, 0 and y, were assigned to the bonded oxygen atom and one terminal oxygen site, respectively; the occupancy of the other was constrained to be 0 - y. Isotropic thermal parameters described the thermal motion of the dioxygen ligand; these for the two terminal oxygen positions were constrained to be equal. In other respects the refinement paralleled the isotropic refinement of I described above. Values for 0 and y of 0.93 (3) and 0.54 (3) were obtained after two cycles of refinement. At this stage the values for R and R , on F were 0.149 and 0.196.

The thermal parameters of nongroup atoms were now allowed to vary anisotropically. After three cycles of least-squares refinement, which lowered values for R and R, on F to 0.107 and 0.138, a search for hydrogen atoms was made. All hydrogen atoms, with the exception of thoseon the solvate species, were inclilded as a fixed contribution to F,. After the solvate species was charac erized, as described above for I , refinements on all data, including Fa2 < 0, were commenced. Because of correlation between occupancy and (anisotropic) thermal parameters for the dioxygen molecule which occurred in the previous three cycles and also because a better model for the solvate species was now being employed, more meaningful constraints were applied to the dioxygen moiety and the relative occupancy of the two terminal sites was reassessed. The occupancy parameter 0 was fixed at 1 .O; otherwise the refinement was as previously described. A value for y of 0.602 ( 1 7 ) was obtained. This value of y was now held constant and thermal parameters for the dioxygen molecule were allowed to vary anisotropically. Group constraints on the phenyl rings were relaxed and this final model, specified by 396 variable parameters, was refined to satisfactory convergence with final values for R and R,. on F2 of 0.1 I9 and 0.214 and with a standard error in an observation of unit weight of I .92 e2. For the portion of data having Fa2 > 3u(Fo2) the values of R and R, on F are 0.083 and 0.104. The final value for a, the occupancy of the hydrogen-bonded solvate species, is 0.45 (2), but interpretation of the final values for the thermal parameters and oc- cupancies of the solvate species must again be made with caution. Furthermore, there was also a dependence of the value of the mini- mized function on IFo\ and X- l sin 8. Except for a peak near the iron atom (height 0.95 (7) e the final difference Fourier map is flat and featureless with the second highest peak having a height of 0.46 e k3. Tables V I and VI1 list the final nonhydrogen parameters for I I . Hydrogen atom parameters are given in Table V111.3' Table IX lists thevalues of 10)F,I vs. 101F,1.3' A negative entry indicates Fo2 I O .

Solid State Oxygen Equilibria. The simple manometric adsorption apparatus and techniques used to determine dioxygen adsorption isotherms for these solid samples are essentially the same as those previously d e s ~ r i b e d . ' ~ . ~ ~ I n studies using an ethanol-saturated at- mosphere, the sample volume contained a subcompartment filled with

3228 Journal of the American Chemical Society / 102:9 / April 23, 1980

0

0

0.0163(16)

0.0470(16)

0.096381 19)

0.11322125)

0.18946127)

0.21757(26)

0.16013125)

0.16954 1241

0.04257 I I8 1

0. I I437 (24 )

0.12297127)

0.059631291

0.00825125)

-0.06501(25)

0.23246 (24 )

0.25309(36)

0.31 4 35 129)

0.1971908)

0.12657(56)

0,19769 189)

0.21584(90)

-0.10949(32)

-0. I I182(44)

-0.12209(511

-0.10848(361

-0.17929 145)

-0.07283 I84 I

-0.06602(66)

*.24348(26)

0.28405(32)

0.35400(371

0.38200137)

0.34321 (35)

0.27412(28)

-0.09339(26)

-0.09692(29)

-0,12158 (33)

-0.14268 (321

-0.1 3904 (32)

-0.11428(28)

n. 1340651591 0.23165133)

0.2707110)

0.2711(16)

0.14003(19)

0.14049124)

0.14580 (28)

0.14806128)

0.14405(24)

0.14268(23)

0.13703( 18)

0.13940(23)

0.13786127)

0.13537127)

0.13466123)

0.13626123)

0.26735127)

0.33119136)

0.34582130)

0.385813 133)

0.36899 (671

6,4125 ( 1 1 )

0 . 4 4 0 4 0 (76)

0.25089128)

0.31229145)

0.32091(35)

0.37162(34)

0.39718(61)

0.35266156)

0.42919(52)

0.14530128)

0.08651132)

0.0892R(44)

0.15047(53)

0.20957140)

0.20732(32)

0.17941(28)

0.06521129)

0.05929136)

0.11487(41)

0.17994136)

0.18689131)

I / 4

1 /4

0.2021(14)

0.2718(281

0.20775(19)

0.13434 (25)

0.12696(291

0.194901311

0.24538126)

0.32072(26)

0.35096 119)

0.37032 125)

0 .C4787 I26 1

0.47614(27)

0,41562124)

0.42355(25)

0.3690102)

0.35956(37)

0.36293(47)

0.35024 (42)

0.3682 I I O )

0.28046(82)

0.4011 (14)

0.50221126)

0.52300(40)

0.59012(411

0.47262(411

0.45563(66)

0.40568(77)

0 .SO228 (83)

0.351 38 (27)

0.35659(36)

0.38146(43)

0.40391(46)

0 . 4 0 1 4 1 ( 4 0 )

0.37437(32)

0.49990(26)

0.53231(31)

0.60339 (36)

0.64022 0 3 )

0.60869(31)

0.53824 (28)

40.17142) 30.131381 0

70.7 (34)

119. ( 1 3 )

127.(23)

31 .o ( 1 4 )

31.8(17)

40.8121)

48.3123)

36.8(181

35.6(171

32-21 13)

36.2 ( 17)

32.4 ( 17)

33.4(18)

29.4115)

33.2116)

93.0(30)

69.8 132)

221.5(61)

67.5132)

348. (17)

122.0180)

395. (24)

35.0118)

41.5 (27)

81 .SI331

65.9 (31 )

160.1 (75)

153.7(83)

221. ( IO)

39.4(191

68.1 (30)

84. 7 ( 39)

85.9141)

84.2 138)

56.1 (26)

30.1(16)

45.8 (23)

49.1 127)

38.5 (23)

35.8 (22)

31.8(19)

0

-6.1(701

-42. ( I I )

-0.7110)

0.8(13)

-3 .0 (16 )

-1.2114)

0.2(131

1.1(131

-0.34 (98 i

-0.2t 121

-1.2(16)

-4 -8 ( 16)

-0.1 (13)

- I .2(13)

-4.6 ( 1 4 )

-7.2 (201

-13.7118)

-2.8120)

42.9 (45)

180. ( 1 I )

35 -9 ( 6 6 1

4.6(19)

0 -9 (28)

5.303)

2.2(20)

-7.0 (361

28.8(55)

-36.3 ( 4 5 1

1.7115)

7.0118)

11.3 (24)

- 1 .7 (28)

-3.2 I221

-2.6 I171

-2.7 ( I51

-2.5( 16)

-2.2 (20 1

1.9(211

2.0120)

1-71 16)

0

0

48.21791

-17. I141

0.31111

0.6(141

2.5(17)

-2.5 (171

-1.7(141

O.O(I4)

1.1IlOJ

-0.3(131

1.6(16)

-0 .5(16)

Oal(13)

0.1(13)

-15.31191

2.5(22)

48.1 ( 3 1 )

-4.3 (22)

121.2(88)

115.6 (96)

-123.1111

-3.5(15)

1 a7 (261

-3.9125)

7.7(231

90.5 (57)

59.31531

44 .* (55 )

4.0116)

7.2121 1

19.4 (29)

14.0 ( 3 4 )

-4.6 ( 26 )

-4.4119)

3.3115)

10.51181

20.3 122)

13.SI2Zl

-1.21191

3.6(171

ethanol. This contributed a constant and measurable background of -60 Torr at 25 'C. The electric manometer, a Datametrics Inc. Barocell Model 1 174-570, was calibrated against a new MKS In- struments Inc. Barotron, Model 310BHS-1000, over the full range of the instrument (0-IO00 Torr), and found to agree to within 0.1% and to be reproducible to within 0.5% of the observed pressure for measurements made under ethanol. The errors in the determination of moles of 0 2 adsorbed were generally less than 3%; in the absence of ethanol errors were less than 0.5%.

By extrapolation to infinite 0 2 pressure the stoichiometry of 0 2 binding was established to be 1 mol of dioxygen per mol of iron to within 1 % in all cases. Complete reversibility was observed with each sample; data were collected at random over the range of 0 2 pressures used. In the absence of ethanol, even after 50 cycles between 0 2 and vacuum, no discernible change had occurred in the binding of dioxygen

by the sample ( to within experimental error, <OS% of moles of 0 2 adsorbed). For Fe(TpivPP)(2-Melm).EtOH under ethanol vapor, oxidation occurred more rapidly, although it was still less than 5% after 25 cycles. In all cases dioxygen binding studies have been made on at least two separately prepared samples using both adsorption and de- sorption techniques. The desolvated Fe(TpivPP)(Z-Melm) could be prepared in situ from Fe(TpivPP)(2-Melm).EtOH by simple evac- uation without disassembly or recalibration. In this way the dramatic differences in their 0 2 binding properties were confirmed.

In studies of Fe(TpivPP)(2-Melm).EtOH solvate analyses were performed before and after oxygenation studies; the Et0H:Fe ratios (0.9 ( ] ) : I ) were determined by vapor phase chromatography of pyr- idine solutions. Elemental analyses, magnetic susceptibilities (under Ar and 0 2 ) , and visible spectra were in excellent agreement with the presumed compositions for I and l l . l 8

Ibers, Collman. et al. 1 Fe( TpiuPP)(Z-MeIm).EtOH

Table VII. Rigid Group Parameters for Fe(O~)(TpivPP)(Z-Melm).EtOH

3229

2 2 ATOM X I 2 0.A ATOM X I 2 0.1 .................... . ~ . . ~ ~ . ~ . . ~ ~ ~ ~ ~ 4 + ~ ~ ~ * ~ ~ ~ ~ ~ . ~ . ~ ~ ~ ~ ~ . ~ ~ . . ~ . ~ . * . . ~ . . . ~ ~ ~ ~ . . . ~ ~ . ~ . . . . ~ ~ . . ~ . . . . ~ ~ . ~ . . . ~ ~ . . . . . . . ~ . . ~ . . . . . . ~ . ~ . . . . . .

IN111 0.01083143) 0.02659121) 0.24130146) 4.93116) E O 1 1 ) -0,0527113) -0.2072112l 0.3051113) 4.11179)

IN12) 0.00346152) -0.08695121) 0.24822153) 6.08114) EC1I) -0.0350117l -0.26363193) 0.2558114) 10.0112l

IC111 -0.01920132) -0.02736(24) 0.27262133) 6.08125) ECtZ) 0.0155122) -0.2374123) 0.1971121) 11.1112)

IC(.?) 0.051991451 -0.07130132) 0.197061471 6.59(25) €012) 0 -0.4758117) 1 / 4 29.9 1 1 1 )

IC13) 0.05431135) -0.00363133) 0.19512138) 6.50125) EC13) -0.0358126) -0.6151117) 0,2194132) 29.91111

IC1MEI -0.07522143) -0.01962147) 0.33188144) 1.00128) EC141 0 -0.34961171 114 29.9 I1 1 I

o...................... .............1....~.~~~~~..~.~~~~*.~~~~~..~.~~....~~~......~~.....~.....~..~............~........~.........

RIG10 GROUP PARAMETERS

Figure 2. Stereodiagram of Fe(TpivPP)(Z-Melm). Ellipsoids are drawn at the 30% probability level. Hydrogen atoms are omitted for clarity.

Figure 3. Stereodiagram of Fe(TpivPP)(2-Melm) looking down the twofold axis. Ellipsoids are drawn a t the 30% probability level. Hydrogen atoms are omitted for clarity.

Results and Discussion General Description of the Structures. The deoxy complex,

Fe(TpivPP)(2-MeIm).EtOH ( I ) , is illustrated in Figures 2 and 3. In order that small differences may be better appreciated, perspectives and orientation are identical for the corresponding diagrams, Figures 4 and 5, of the dioxygen complex Fe(O1)- (TpivPP)(2-MeIm).EtOH (11). Similarly, Figures 6 and 7 show the arrangements of 1 and I1 in the unit cells.

The twofold symmetry, which is imposed on molecular species, leads to disorder about the crystallographic twofold axis of the 2-methylimidazole ligand, the ethanol of solvation, and, for 11, the angularly bound dioxygen ligand. The ethanol solvate species suffers from additional disorder: in one of its two crystallographically independent sites it is weakly hy- drogen bonded to the 2-Melm ligand. The N-H-0 separa- tions are 2.88 (3) 8, for I and 2.78 ( 2 ) A for 11. Near super- position of atomic positions for the 2-Melm and the ethanol

moieties necessitated the application of rigid group constraints. For the 2-MeIm ligand an idealized geometry was used (see Figure 1 ) . Small systematic errors in parameters associated with this group may, therefore, exist, but differences in pa- rameters between I and I1 are meaningful. Because Fourier maps of the solvate species were i l l resolved and interpretation is not unambiguous, intermolecular contacts involving these atoms may well be in considerably greater error than that in- dicated by least-squares refinement. The dioxygen ligand in I 1 shares a generally similar angular geometry and disorder with its 1-MeIm analogue. The occupancy ratio for the two crystallographically independent terminal oxygen atom sites is 0.60 (2):0.40. As previously observed' the Fe-0-0 planes approximately bisect the N,-Fe-N, right angles. The 0-0 separations and F e - 0 - 0 bond angles are uncorrected for the effects of thermal motion and/or a small off-axis displacement of the coordinated oxygen atom which could be inferred from

3230 Journal of the American Chemical Society 1 102:9 / April 23, 1980

Figure 4. Stereodiagram of Fe(02)(TpivPP)(Z-Melm) Ellipsoids are drawn a t the 30% probabrlit) level Disorder of the terminal oxygen atom but not of the 2-Melm IS shown Hydrogen atoms are omitted for clarit)

Figure 5. Stereodiagram of Fe(02)(TpivPP)(2-Melm) looking down the twofold axis. Ellipsoids are drawn at the 30% probability level. Disorder of the terminal oxygen atom but not of th’e 2-Melm is shown. Hydrogen atoms are omitted for clarity.

V I

Figure 6. Stereodiagram of the unit cell for Fe(TpivPP)(2-Melm).EtOH. The two crystallographically independent ethanol moieties are shown. Hydrogen atoms are omitted for claritb, The view is approximately down the c axis; the b axis i s vertical.

V 1

Figure 7. Stereodiagram of the unit cell for Fe(O>)(TpivPP)(2-Melm).EtOH. The two crystallographically independent ethanol moieties are shown; disorder of the terminal oxjgen atom but not of the 2-Melm is shown. Hydrogen atoms are omitted for clarity. The view is approximately down the c axis; the b axis is vertical.

the flattened shape of its thermal ellipsoid (root mean square com onents of displacement are 0.277 (8), 0.329 (9), and 0.362 (8) 1). Nonetheless, we have defined more precisely than be- fore a lower bound for the 0-0 separation (1.21 (2) A) and an upper bound for the F e - 0 - 0 angle ( 1 29 (1)’). Compre- hensive compilations of bond distances and angles are available in Tables X and XI. The isolation of a nondisordered, iron- dioxygen complex obviously remains a high priority.

In common with all other “picket-fence’’ porphyrin struc- t u r e ~ . ~ ~ ~ ~ ~ , ~ ~ atoms of the pivalamide “pickets” a re afflicted

with very high thermal motion and/or irresolvable disorder. However, the immediate coordination sphere and the por- phyrinato skeleton are reasonably precisely defined, although for the dioxygen complex the scatter of formally chemically equivalent parameters about their mean is a little larger than that expected from the estimated standard deviation of an individual parameter (see Figure 8 and Tables X and XI). Inspection of Figures 2-7 shows that oxygenation of I induces large changes only in the immediate coordination sphere; changes in crystal packing and the porphyrin periphery are

Ibers, Collman, et al. / Fe( TpiuPP)(Z-MeIm).EtOH 3231

Table X. Bond Distances ( A ) . for 1 and I I

atoms distance atoms distance

Fe-N(I ) 2.068(5) 1.997(4) Fe-N(2) 2.075(5) 1.995(4) Fe-IN ( 1 ) 2.095(6) 2.107(4) Fe-O( I ) 1.898(7) O( 1)-0(2A) 1.205( 16) O( 1)-O(2B) 1.232(22) N(1 ) -C( I ) 1.384(8) I .387(6) N( l ) -C(4 ) 1.373(8) 1.373(6) N(2)-C(6) 1.361(8) 1.392(6) N(2)-C(9) 1.378(8) 1.363(6)

C(6)-C(7) 1.441(9) 1.424(6) C(8) -C(9) 1.434(9) 1.453(6)

C(4)-C(S) 1.406(9) 1.385(6) C U - C ( 6 ) I .403(9) I .399(6) C ( 1 )-C( 10) 1.393(9) 1.381 (6) C(9) -C( 10) 1.392(9) 1.394(6) C(S)-C(A1) 1.503(9) 1.491 (6) C( IO)-C(B1 ) 1.511(9) 1.5 14(6) C ( A 1 )-C(A2) 1.371 ( I O ) 1.378(7) C(BI)-C(B2) 1.383( 10) 1.385(7) C (A2) -C( A3) 1.384( 1 1 ) 1.386(8) C ( 82 ) -C( B3) 1.380( 1 I ) 1.396(8) C(A3)-C(A4) 1.344( 13) 1.361 ( I O ) C(B3)-C( B4) 1.350( 12) I .340(8) C(A4)-C(A5) I .35S( 13) 1.363( 10) C(B4)-C(B5) 1.337( 12) 1.393(9) C(A5)-C(A6) I .399( I O ) 1.383(7) C(B5)-C(B6) 1.396( I O ) 1.388(7) C(A6)-C(AI) I .404( I O ) 1.398(7) C(B6)-C(BI) 1.369( IO) I .382(7) C(A6)-N(A 1 ) 1.401 ( I O ) 1.410(7) C( B6)-N( BI ) 1.419( I O ) 1.412(7) N ( A 1 )-C(A7) 1.296(1 I ) 1.3 l4(7) N (BI )-C( B7) 1.167( 14) 1.254(8) C(A7)-O(AI) 1.198(11) I . 190(7) C ( B 7 ) - 0 ( B I ) 1.282( 12) 1.279(8)

C(I ) -C(2) 1.43 l (9) 1.452(7) C(3)-C(4) 1.432(9) 1.443(7)

c w c ( 3 ) I .350(9) 1.339(7) C(7)-C(8) 1.336(9) l .315(7)

C (A 7) - C ( A 8 ) I .497( 13) 1.505(9) C(B7)-C(B8) I . % ) I ( 14) 1.479(9) C(A8)-C( A9) 1.432( 16) 1.419(11) C(B8)-C(B9) 1.460( 13) I .452( I O ) C(A8)-C(A I O ) 1.448( 14) 1.377( 12) C(B8)-C(Bl0) 1.478( 16) I .459( 13) C(A8)-C(AI I ) 1.480( 19) I .448( 15) C(BS)-C(BlI ) I .456( 14) I .472( 1 1 ) IN(2)-EO( I ) 2.88(3) 2.78(2)

a First entry is for Fe(TpivPP)(2-Melm).EtOH; second entry for Fe(Ol)(TpivPP)(2-Melm). EtOH

generally small. Intra- and intermolecular structural changes upon oxygenation will be discussed after a closer examination of the stereochemistries of 1 and 11.

The Deoxy Complex. Figure 9 summarizes some of the metrical details of the coordination spheres of I and I1 and two closely related compounds, Fe(TPP)(2-MeIm) and Fe(O2)- (TpivPP)( I -MeIm). Differences between the Fe-N,, Ct-N,, or Fe-Ct separations for I and Fe(TPP)(2-Melni) are of marginal significance and these first two distances are typical of other five-coordinate, high-spin metalloporphyrins, such as Mn(TPP)( 1-MeIm).35 However, the Fe-NI, separation and the doming parameter-the difference between the Ct-Fe separation and the displacement of the iron atom from the least-squares plane of the 24-atom porphyrinato skeleton-are very different for I and Fe(TPP)(2-MeIm). The doming pa- rameter, which is rarely larger than 0.05 A,34 and which for I is 0.026 A, is 0.15 A for Fe(TPP)(2-MeIm).34 By way of comparison, the Ct-Mn separation for Mn(TPP)( I -Melm) is 0.5 12 A, yet the doming parameter35 is only 0.045 A. Con- struction of a space-filling model of the “picket-fence’’ por- phyrin indicates that intramolecular steric interaction among the “pickets” should not prevent doming for “picket-fence’’ porphyrin complexes. Since TPP and TpivPP are very similar electronically the longer (by 0.066 A) separation observed for Fe(TPP)(2-MeIm) could be attributed to the more nearly eclipsing conformation which the imidazole plane adopts with respect to the Fe-N, bonds (see Figure 9).

The Oxy Complex. The Fe-N, bond lengths and the Ct-N, radius are typical of other low-spin iron(1I) systems (see Table XII). However, the bulky 2-methyl substituent on the 2-Melm ligand results in lengthened axial bonds relative to the sterically less demanding I-MeIm ligand. The sum of the F e - 0 and Fe-NI, separations is 4.005 8, for I1 but only 3.813 8, for Fe(Oz)(TpivPP)( I -Melm). Of this 0.1 92-8, difference, most (0.150 A) arises from a lengthening of the F e - 0 bond to a value similar to that observed for Schiff-base cobalt-dioxygen compounds. This lengthened Fe -0 separation and also the greater sum may be correlated with the decreased dioxygen affinities in solution which have been observed for systems with hindered imidazole^.^^'*-'^ Note also that atoms 0, Fe, and

\

Figure 8. Averaged bond parameters for the porph) rinato skeleton for I and 11. Parameters for I are to the left of the dotted line; those for I I are to the right. The general labeling scheme for porphjrinato species is also shown.

NI,, are collinear for Fe(Oz)(TpivPP)( 1-Melm); for I 1 the angle is 172.9 (2)’.

For 11, the compromise between maximum bonding and minimum destabilizing nonbonding contacts results in the iron atom remaining 0.086 A out of the plane of the porphyrinato nitrogen atoms toward the 2-Melm ligand. This is in marked contrast to that observed for Fe(02)(TpivPP)( 1-Melm) where the displacement is 0.03 8, out of the plane toward the dioxygen ligand and for other related compounds whose stereochemis- tries are summarized in Table XI I . The porphyrinato-dioxygen contacts for I1 and Fe(Oz)(TpivPP)( 1 -Melm) are not signif- icantly different, dispelling the slight reservations expressed about the accuracy of the Fe -0 separation for Fe(O2)- (TpivPP) ( 1 -Me1 m) . I

Structural Changes Occurring upon Oxygenation. A. In- tramolecular. The expected contractions in the average Fe-N, separation (0.076 A) and in the Ct.-N, radius (0.039 A), at- tributable to the transition from the five-coordinate, high-spin. deoxy form to the six-coordinate, low-spin, oxy form are ob- served, as well as a general contraction of the porphyrinato skeleton. The average change i n the separation of dyadically

3232

Table XI. Bond Angles (deg)“ for I and I I

Journal of the American Chemical Society / 102:9 / April 23, 1980

atoms anale atoms

h ( 1 ) -Fe-N( I ) ’ N ( I )- Fe- N (2) N ( 1 )- Fe-N (2)’ N ( 1) -Fe- IN(I ) O( I)-Fe-N( 1) Fe-O( I ) -0(2A) Fe-IN( ] ) - IC( I ) Fe -N( I ) -C( I ) Fe-Y( I)-C(4) C ( I)-N( I ) -C(4) N ( I )-C( I )-C(2) N ( I)-C(4)-C(3) N ( I)-C( I)-C(l0)’ C(2)-C( I ) -C( I O ) ’ C(3)-C(4)-C(S) Y ( 1 ) -C( 4) -C( 5 )

C(4)-C(3) -C(2)

C(4) -C(S)-C(A 1 )

C ( 5 ) -C (A 1 ) -C (A2) C ( 5 ) - C ( A I ) C(A6)

C ( l)-C(2)-C(3)

C(4)-C(5)-C(6)

C(6)-C(5)-C(A I )

C(AZ)-C(A I)-C(A6) C(A I )-C(AZ)-C(A3) C ( A 2) -C( A 3) -C( A4) C(A3)-C(A4)-C(A5) C(A4)-C(A5)-C(A6) C(AS)-C(A6)-C(AI) C(A5)-C(A6) N ( A I ) C ( A 1 ) C ( A 6 ) - h ( A I ) C(A6)- \ (A I )-C(A7) N(A I ) -C(A7)-O(AI) h ( A 1 )-C(A7)-C(A8) C(A8)-C(A7)-O(A 1 ) C(A7)-C(A8)-C(A9) C(A7)-C(A8)-C( 4 I O ) C(A7)-C(A8)-C(AI I ) C( 49)-C(A8)-C(A I O )

158.0(3) 87.9(2) 8 7.8 (2) 91.9(3)

132.1 (8) 126.3(4) 127.0(4) 105.8(6) l09.3(6) 110.3(6) 124.5(6) 126.2(7) 124.8(7) 124.8(6) I07.8(6) I 06.7( 6) I25.8(6) 1 l6.6(6) 1 17.6(6) 12 I .2(8) I19.8(7) 1 I9.0(8) 120.3(9) l20.9( I O ) 120. I ( I 0) 12 I .O( I O ) 118.5(9) 123.6(9) I17.9(7) 127.9(9) 120.9( 12) 119.5(10) 1 19.5( 12) I14.5(1 I ) I l0.8(10) 104.9( 13) 110.6(15)

173.2(2) 90.4( 1) 89.4(1) 86.4(2) 86.7(1)

I29.0( 12) 135.0(6) 127.4(3) 127.2(3) 1 05.4( 4) 1 I0.0(4) 110.2(4) 125.4(4) 124.6(4) I23.9(5) 125.9(5) 106.6(4) 1 07.8 (4) 124.5(4) 118.0(4) 117.5(4) 120.7(5) I2 1.2(5) I18.1(5) I20.9(6) I19.5(7) 121.6(6) I l9.0(7) 120.9(6)

117.9(5) 128.7(5) 120.6(7) 118.3(6) 120.9(7) 118.0(7) 110.5(7) 1 06.7 (9)

1 2 1 . I (6)

109.2( 12) 105.3( 14) l03.6( 12)

c ( A I O ) - C ( A ~ ) - C ( A I ’ I ) I I O . ~ ( I ~ ) 108.2(13) IN( 2)- H IN (2)-EO( 1 ) 166.8 166.0

anele -

N(2)-Fe-N(2) O( I)-Fe-IN( 1 )

N (2)-Fe-I N (2) O( I )- Fe-N (2) Fe-O( I)-O(2B) Fe- IN ( 1 )-IC( 3) F e - h (2)-C(6) Fe-N(2)-C(9) C(6)-N(2)-C(9) K(2)-C(6)-C(7) N(2)-C(9)-C(X) N(2)-C(6)-C(5) C ( 7)-C (6) -C( 5)

N (2) -C( 9)-C( 1 0) C(6)-C(7)-C(8) C(9)-C(8)-C(7) C(9)-C(lO)-C( I ) ’ C(9) -C( 10)-C( BI ) C( I)’-C( IO)-C(BI) C( IO)-C(B I)-C(B2) C ( 10)-C(BI )-C(B6) C(BZ)-C(B 1 )-C(B6)

C(S)-C(9)-C( I O )

C(B l ) -C( B2)-C(B3) C(B2)-C(B3) -C(B4) C(B3)-C( B4)-C(B5) C(B4)-C(B5)-C(B6) C ( B5)-C( B6)-C(B I ) C( B5)-C( B6)-N(B I ) C(B I )-C(B6)-h’(BI ) C(B6)-N(B I )-C(B7) N ( BI )-C( B7)-0( BI) N ( B 1)-C(B7)-C( B8) C(B8)-C(B7)-O(B 1 ) C( B7)-C(B8)-C(B9) C(B7)-C(B8)-C(B I O ) C(B7)-C(B8)-C(BI 1) C(B9)-C( B8)-C(B I O ) C(B9)-C(B8)-C(BI I ) C(B 10)-C(B8)-C(B 1 1 )

157.5(3) 170.4(3)b

103.7( 3)

126.3(7) 126.8(4) 125.5(4) 106.2(5) l10.0(6) 109.2(6) 125.1(6) 124.9(7) 125.6(7) 1 2S.2(6) 106.9(6) 107.7(6) 126.2(6) l16.8(6) I 17.0(6) 120.0(7) l21.7(7) 1 l8.3(7) I20.6(8) 120.0(9) 120.8(8) I 20.3 (9)

124.3(9) I l5.6(7) I35.0( I O ) 122.6( 13) 128.4(11) 109.0( 1 3) 110.3(10) l08.2( I O ) 115.6(11) l08.9( 12) 1 1 1.4(1 1)

120.1(8)

l02.0( 12)

I

176.7(2) 172.9(2)

93.5(2) 88.3(1)

128.5(18) 123.3(5) 127.0(3) 127.8(3) 105.1(4) I09.8(4) 109.8(4) 124.9(4) 325.4(4) 124.4(4) 125.7(4) 108. I (4) 107.2(4) 123.9(4) 117.9(4) 118.1(4) 119.7(5) I20.6(5)

1 19.6(6) 120.8(6) 120.4(6) 119.6(6) 119.9(5) l23.0(6) 117.0(5) 134. I (6) 120.6(7) 123.5(7) 1 15.7(8) 110.3(7) I10.7(7) 1 13.3(8) I 10.4( I O ) 109.9(8)

119.7(5)

104.1 ( I O ) -

a First entrv is for Fe(ToivPP)~2-MeImbEtOH: second entrv for Fe(Oz)(TpivPP)(2-Melm).EtOH. Angle quoted is that which the vector ~ , / \

Fe-lK( I ) makes wi th the twofold axis.

related porphyrinato atoms for I and I1 is 0.048 8, with ex- tremes of 0.002 (atom C ( l 0 ) ) and 0.082 8, (atom N(2)). Atoms in phenyl group A move in while those in B move out slightly upon oxygenation (see Table XIII).3’

While the orientation of the imidazole plane with respect to Fe-N, vectors remains essentially constant upon oxygen- ation (Figure 9), its movement toward the porphyrinato plane “pushes” against the porphyrinato nitrogen atom, N( 1 ) ; for I the four porphyrinato nitrogen atoms lie within 0.005 (5) 8, of the least-squares plane but for I 1 this increases to 0.030 (4) 8,. The porphyrinato skeleton is rather buckled-the mean displacements from the least-squares plane of the 24-atom skeleton are 0.056 A for 1 and a slightly greater 0.065 A for 11. The displacements are illustrated in Figure 10. The relative orientation of the pyrrole rings follows no systematic pattern and probably reflects the nonsymmetrical orientation of the imidazole ligand which neither bisects an Np-Fe-N, right angle nor eclipses an Fe-N, vector (see Figure 9). Selected least-squares planes and the dihedral angles between them are presented i n Tables XIV and XV.31

Upon oxygenation of the complex, the iron atom moves 0.3 16 8, toward but not into the plane of the porphyrinato ni- trogen atoms, while preserving the Fe-NI, separation. Non- bonding porphyrinato-imidazole contacts show decreases of

up to 0.27 8, (Table XVI),31 indicative of the strong bonding in low-spin, six-coordinate iron-porphyrinato complexes. The 2-Melm ligand adjusts for this motion by increased tipping: the angles Fe-IN(I)-IC( 1) and Fe-IN( I)-IC(3) increase their asymmetry upon oxygenation (see Table XI) while the angle of tilt between the Fe-N, vector and the normal to the porphyrinato plane decreases from 9.6 to 7 . I O (see also next section).

B. Intermolecular. Since crystallinity is maintained upon oxygenation large changes in intermolecular contacts are not expected nor are they observed. Changes in unit-cell dimen- sions are small and the small shrinkage in the c axis of 0.8% may be correlated with the general shrinkage of the porphy- rinato skeleton. The iron atom and atoms belonging to the 2-MeIm and solvate moieties move substantially in a direction parallel to the b axis upon oxygenation. In addition, it is clear from Fourier maps, as well as from the final refined parame- ters, that the ill-resolved ethanol solvate species hydrogen bonded to the 2-Melm ligand changes its orientation. But the remaining atoms show an average change in position of only 0.1 I 8, with a maximum of 0.44 8, for one of the pivalamide methyl carbon atoms, C(A1

Table XVII summarizes intermolecular contacts for 1 and 11 and also indicates the changes in contacts which occur when

Ibers. Collman, et al. / Fe( TpiuPP)(Z-MeIm)-EtOH

Fe-0

3233 0

?;698171 N, Ct 2 044 2 033 1994 I

N N N N N N v

2 :-%51 0399 1

Fe-N, Fe C t Fe-N (I 2 16115) \ 2 095m 1 2107141

. $3 . 6 Figure 9. Structural changes upon oxygenation of an iron(ll)(porphyrinato)(base) complex.

Table XII. Bond Parameters (A) for Some Six-Coordinate, Low- Suin. Iron(l1) Poruhvrinato Compounds

comDd Fe-N, Fe-L Fe-NI, Fe-Ct"

Fe(Qz)(TpivPP)- l.98( I ) 1.75(2) 2.07(2) +0.03

Fe(Oz)(TpivPP)- 1.996(4) 1.898(7) 2.107(4) -0.086

Fe(CO)(TPP)(py)d 2.02(3) 1.77(2) 2.10(1) 0.02 Fe(NO)(TPP)- 2.008(9) 1.743(4) 2.1 80(4) 0.07

Fe(TPP)( I-Melm)$' 1.997(4) 2.016(5) 0

( 1 -Melm)

(2-Melm)

( I -MeIm)e

0 A negative value denotes a displacement toward the imidazole ligand, b Reference 12. This work. Reference 13. Reference 14. .f Hoard, J . L., personal communication.

an oxygenated molecule is surrounded by deoxygenated mol- ecules. Table XVII13' gives selected intermolecular contacts involving hydrogen atoms. For both I and I1 there are few contacts less than 3.5 A. However, several of the shortest contacts involve atoms with very high apparent thermal motion and, hence, such separations cannot be regarded as realistic. This, together with the ill-resolved solvate species, cripples any attempt to perform meaningful, unbiased packing energy calculations. No dramatically close contacts arise when a molecule of 11 is surrounded by molecules of I , or vice versa. I n other words, no single contact or group of contacts can be said clearly to influence the binding of dioxygen to I , Fe(T- pivPP)(2-Melm).EtOH. Of the atoms with relatively modest thermal motion only contacts between phenyl-B and the 2- MeIm ligand show some rearrangement. The 2-Melm ligand in I is tilted such that the O(A1)-IC(Me) contact is relieved (see also Figures 4 and 5). Upon oxygenation the 2-MeIm li- gand moves away from this oxygen atom and the angle of tilt decreases from 9.6 to 7.1 '. Moreover, the contraction of the porphyrinato core pulls in phenyl A and further reduces this contact.

Solid State Dioxygen Binding. The presence or absence of ethanol in the solid state dramatically affects dioxygen binding. The reversible binding of dixoygen to powdered samples of Fe(TpivPP)( 1 -MeIm) , Fe(TpivPP)( 1,2-Me*Im), Fe(T- pivPP)(2-MeIm), and Fe(TpivPP)(2-MeIm)-EtOH is shown in Figure 1 1 and collated in Table XIX. The binding of diox- ygen under EtOH vapor to I , Fe(TpivPP)(2-Melm).EtOH, shows simple, noncooperative behavior over the range of 10-60% oxygenation with Pll2 = 620 (30) Torr. This was surprising since Fe(TpivPP)(2-MeIm) without ethanol had already been found to bind dioxygen in a cooperative manner18 (see Figure 11) . This is not an experimental artifact, having been confirmed on several separately prepared samples, using both adsorption and desorption measurements. I n addition, after being oxygenated under ethanol, a sample of Fe(T- pivPP)(Z-MeIm).EtOH can be converted to Fe(TpivPP)(2-

261 209

'\

'131 108

0 h75121

0 03

2 07121

-510 ;484

037(7)--040(7) -037(5) -0856) -239 '

I \ -20 I

Figure IO. Displacements of atoms (A X I 03) from the least squares plane of the 24-atom porphyrinato core. The displacements for I are quoted above those for 11; the atom labeling system is indicated. Only those dis- placements with estimated standard deviations i n parentheses were used to define the least-squares plane.

MeIm) by evacuation, and the 0 2 binding of the desolvated solid regains its high affinity and cooperativity (with all measurements taken in the same apparatus without any dis- assembly or recalibration).

Having been able to gather X-ray diffraction data on both Fe(TpivPP)(2-MeIm).EtOH and Fe(Oz)( TpivPP)(2- MeIm).EtOH, we had hoped that the structures of these complexes and, especially, an analysis of the intermolecular contacts might have revealed the origin of the cooperativity observed in solid-state oxygenation. Unfortunately, since the solvated solids do not bind dioxygen cooperatively, we can only use these structures in a general way to speculate on the mechanism of cooperativity in this system. For I the motion of the 2-MeIm ligand toward the porphyrinato skeleton upon oxygenation is 0.30 A and there is also a small contraction in the porphyrinato skeleton. Presumably, as molecules oxy- genate, the change in molecular dimensions induces strain in the crystallite. Eventually this strain must be sufficient to in- duce a conformational change in the solid, which enhances the dioxygen affinity of the remaining deoxy sites. Why such a process occurs in some solids and not in others remains un- known. Indeed there is no a priori reason why cooperative and noncooperative rather than anticooperative behavior should be observed in the present system. Spin-crossover systems show similar cooperative behavior in the solid state as the ground spin states change with temperature. The nature of these spin transitions depends variously on crystal size, counterion and solvate choice, and defect content. Of particular interest here, the high-spin to low-spin transition of several Fe( 11) complexes is dramatically influenced by the presence of solvate molecules. For example, bis(thiocyanato)bis(4,7-dimethyl- 1 , l O-phen-

3234

Table XVII. Intermolecular Contacts, Less Than 4.0 A, for I (Deoxy) and I I (Oxy) and for I I Surrounded bq Molecules of I

Journal of the American Chemical Society / 102:9 / April 23, 1980

deoxy surrounded' Fe(O?)(TpivPP)- atoms sym op" deoxyh in deoxy by oxy in oxy cell oxy i n d oxy ( 1 me 1111) 1'

3.48 3.58 3.63 3.96 3.93 3.12 3.42 3.60

C(2)-O( B I ) C ( 3)-O( B I ) C(6)-C( B3) C(7) -O(AI) C(7) -C( B3) C(7)-C( B2)

C(8)-C(B2) C(8)-O(A 1 )

C(A7)-E(OI) O(A I )-IC(2) O(AI) -EO( I ) O(A I)-IC( ME) O(A 1 )-IC(3) O(A I ) -EC( 2)

C(A1 I)-C(A3) C(AI I)-C(A4) N (B 1 )-EC( 2) O( B I ) -EC(2) O( B I ) -EC(3) O( BI)-EC( I )

C ( A lO)-C(B4)

O( B I )-EC(4) C(B7) -EC(2) C(B9)-C( 8 4 ) C ( B9)-C( B5)

C(B1 I)-C(BI I ) C(A3)-EC(3)

C ( B9)-C(B3)

C(A3)-EC(4) C ( A 3 ) - E 0 ( 2 ) C(A3)-EC(3) C(A4)-EO( I ) C(A4)- EC(2) C(A4)-EC( I ) C(A4)-EC(4) C(A4)- EC( 3) C(A4)-EC(3) C(AS)-EO( I ) c ( A 5) - EC ( 2) C(A5)-EC( I ) C(A6)-EO( I ) C(B3)- IC( I ) C(B3)-IY(2)

C(B3) IY(2) C ( B3)-IC( 2) C(B3)-1N( I )

C(B3)-IC( I )

C( B3)-IN( I ) C(B3)- IC(3) C ( B 3 ) - I C ( M E) C ( B3)-IC(2) C ( B3)-IC(3) C ( B4)- I W (2 ) C ( B4)- I N (2 ) C ( B4)- IC( 2) C(B4) - IC( I ) C(B4)-IC( I ) C( B4)-EC( 2) C(B4)--IC(2) C ( B4)- IC( 3) C ( B4)- I C( M E) C(B4)-EO( I ) C(B5) EC(2) C( B5) - lh ( 2)

C(SS)-EO( I )

C(B6) EC(2) IN (2)-EO( I ) Ih (2) -EC( 2) IN(Z)-EC( I ) I C ( I ) EO(1)

C(B5)-FC( I )

C(SS)-EO( I )

55405,45505 55405.45505 55606, 55606 55607,55607 55606, 55606 55606, 55606 55607, 55607 55606. 55606

4, 44504 8, 54508 4,44504 4,44504 8. 54508 8, -

55405, - 55607, - 55607, -

3,55403 3,55403 3, 55403 3. 55403 3, 55403 3, 55403

45607,45607 45607,45607 45607.45607

4, 44504 4,44504 4,44504 8, - 4,44504 8. 54508 4,44504 4,44504 4,44504

4,44504 8, 54508 4,44504 4.44504

55606, 55606 55606, 55606

3, 55403 3.55403 3, 55403

55606, 55606 3. 55403 3. 55403

55606, 55606 55606, 55606 - ,55606

55606, 55606 3,55403

55606, 55606 3, 55603

55606.55606 3, 55403 3, 55403

- , 55606 - , 55403

3. 55403 3, 55403

55606, 55606 3. 55403 3, 55403

55606, - 3, 55403 1 I I 1

56606, -

8. -

3.86 3.88 3.88 3.63 3.95

3.81 3.97

3.08 3.17 3.29 3.77

-

3.88 4.00 3.93 3.09 3.43 3.58 3.74 3.81 3.72 3.80 3.96 3.82 3.55 3.96 3.96

3.29 3.43 3.56 3.7 1 3.85

2.98 3.54 3.70 3.76 3.50 3.63 3.69 3.70 3.79 3.80 3.8 1 3.82 3.87 3.99

3.3 1 3.44 3.50 3.66 3.67 3.83 3.89 3.93

3.26 3.93 3.97

3.64 2.88 3.12 3.56 3.71

3.67, 3.86 3.65, 3.87 3.93, 3.85 3.62. 3.62 - , 3.88 - , 3.96 3.78, 3.77 - 3.93 3.98, 3.95 3.28. 3.01 3.04, 3.07 3.44. 3.25 __ . 3.72

3.88, - 3.99, - 3.91, - - _ -, 3.81

3.60, 3.14 3.38, 3.33 3.67. 3.52 3.74, 3.69 - . 3.71

3.75. 3.73 3.88. 3.82 3.97. 3.96 3.76, - 3.72, 3.47 3.92, 3.92 3.98. 3.88 3.97, - 3.55. 3.28 3.49, 3.41 3.61, 3.50 3.69. 3.63 - , 3.75

3.99. - 3.13. 2.95 3.43, 3.5 I 3.62, 3.66 3.92. 3.73 3.54, 3.44 3.57, 3.55 3.7 I , 3.62 3.60, 3.62 3.70, 3.74 3.89, 3.75 3.91. 3.77 3.85. 3.79 3.96, 3.85 3.92, 3.89

~~ , 3.96 3.36, 3.26 3.43. 3.39 3.55, 3.43 3.75. 3.62 3.82, 3.65 3.98, 3.79 3.89. 3.88 _- , 3.88 - , 4.00 3.93, 3.97 3.56, 3.17 - , 3.89 3.77. 3.85

~~ ,3 .91

3.96. 3.55 .- _-

3.67 3.43 3.65 3.48 3.90 3.64 3.50 3.95 4.00 3.75 3.53 3.97 3.89 3.2 I 2.90 2.93 3.40 4.00 3.78

3.65

3.62 3.70

3.33 3.39 3.50 3.73 3.81 3.94 3.88

3.87 3.48

3.67 3.94 3.94 3.88 2.78 3.08 3.52 3.61

3.24-3.64 solvate

3.54-3.75 solvate

3.42-3.71 solvate

3.55 imid

3.46 itnid

Ibers, Collman, et al. / Fe( TpiuPP)(2-MeIm).EtOH

Table XVlI (Continued)

3235

I I 7 ‘ I

deoxy surroundedC Fe(Ol)(TpivPP)- atoms sym opO deoxyb i n deoxy by oxy in oxy cell oxy in oxy ( I -Melm)‘

IC(2)-EC(2) 1 3.36 3.30 IC(21-EO( I ) 1 3.96 3.87

~~~ ~~~~~~~~

The symmetry operation applied to the second atom is in ORTEP format (Johnson, C. K. ORNL-5 138, 1976). The first three digits specify u n i t cell translations parallel to a, b, and c, respectively, e.g., 546 denotes no translation along a , a translation of - I along b, and + I along c. The last two digits specify the symmetry operation in the order: 01 (x, y , z ) ; 02 (-x, J, I/* - z ) ; 03, ( x . -J, ‘/z + z ) : 04, (I/* + x . ’/z + J, z ) : 05, ( ‘ / I + x , I/* - y , ‘12 + z ) ; 06, (-x, - y , - z ) ; 07 (I,$ - x, I/* - y , - z ) : 08, (l/l - x , I/* + J, ‘/z - z ) . Intermolecular contacts for I . Inter- molecular contacts for a molecule of I surrounded by molecules of I I . The uni t cell for I 1 is used. See footnote 36. Intermolecular contacts for 11 . Analogous contacts for Fe(Ol)(TpivPP)(l-MeIm).

4- / i

Log P ( tor r1 02

Figure 1 1 . Uptake of dioxygen by. ( a ) solid Fe(TpivPP)(I-Melm), (b) Fc(TpivPP)(Z-Melm), (c) Fe(TpivPP)( 1,2-Mezlm), and (d) Fe(T- pi v P P) ( 2- M c I m ) a E t 0 H ,

anthroline)iron( 11) and also its a-picoline solvate display a sharp, cooperative transition, whereas the pyridine solvate remains high spin over the entire range of 300 to 80 K.57a,b Other examples involving solvates of benzene, chloroform, and water have been r e p ~ r t e d . j ~ ~ . ~ - ~ Structural studies indicate the expected reduction of iron-ligand bond distances upon tran- sition from high to low spin, and this is thought t o b e the origin of the observed c ~ o p e r a t i v i t y . ~ ~ ~ Although a full explanation of these phenomena is not yet in hand, nucleation and domain growth are central in the evolving theoriess8 proposed for the crossover systems. In general terms, these same concepts may be applicable to solid state 0 2 binding. In this analogy, our complexes could be viewed as related systems in which 0 2 pressure, rather than temperature, is the variable influencing the spin state.

The general trend in dioxygen affinities, seen in Figure 1 1, can be rationalized in terms of increasing steric restraint to motion of the axial base upon oxygenation. The change from the unhindered I-Melm to the hindered 2-MeIm ligand pro- duces a tenfold decrease in dioxygen affinity. Another tenfold decrease occurs on replacing 2-Melm with 1,2-MezIm as the axial base. This is the result of the buttressing interaction of the I-methyl against the 2-methyl substituent which increases steric contact between the 2-methyl substituent and the por- phyrinato skeleton. The even lower affinity observed for I cannot be attributed to reduced access of dioxygen to the binding site-this is a kinetic, not thermodynamic, phenom- enon. The diminished affinity of I for dioxygen presumably reflects an increased resistance to motion toward the porphy- rinato skeleton by the 2-Melm moiety together with its hy- drogen-bonded ethanol. This is consistent with the apparently decreased occupancy of the hydrogen-bonding site i n the oxy

Table XIX. Cooperativity in 0 2 Binding

Pl l2 , Torr ( 2 5 “C)

affinity affinity Hill high low

system “R” “T“ coefficient

Fe(TpivPP)( I - 0.5 (noncooperative) I .o Melm)a

Melm)a

Mezlm)‘

Melm).EtOHb

U H 7.4c

Fe(TpivPP)(2- 0.7 14 2.6

Fe(TpivPP)( 1,2- 14 I12 3 .O

Fe(TpivPP)( 1- (noncooperative) 620 1 .o HbA, stripped, 0.3 9 2 .5

Reference 18. This work. Tyuma, 1 . ; Imai, K.: Shimizu, K. Biochemisfry 1973, 12, 1491-1498. Imai, K.; Yonetani, T.: Ikeda- Saito, M. J . Mol . B i d 1977, 109, 83-97.

form, 11. Inspection of Table XVlI shows the substantial changes in nonbonded contacts made by the hydrogen-bonded ethanol molecule.

Structural Comparison with Hb and Mb. One intent of studies of synthetic analogues of metalloproteins is to generate information which can serve as a base line for understanding structure and function relationships i n proteins. One should not assume that the behavior of model compounds represents that of an active site free of external influence by the protein. In fact, external influences, such as solvation and crystal packing effects, also exist in model compounds. For example, the differences in the structures of Fe(TPP)(Z-Melm).EtOH and Fe(TpivPP)(2-Melm)-EtOH, discussed earlier, although small and beyond detection in structural analyses of hemo- proteins, are nonetheless significant. Thus, strictly speaking, neither is truly representative of the “free molecule”, undis- torted by crystal packing influences. However, it is possible to change the conditions responsible for these effects and thus gain some knowledge of the isolated site behavior. Certain caveats must be given in making comparisons between these model systems and the hemoproteins: TpivPP is not proto- porphyrin IX; intermolecular solid-state interactions are not intersubunit conformational changes; the hindrance created by the 2-MeIm ligand is not a protein-generated restraint. Nonetheless, the deoxy structure, 1, with the sterically de- manding 2-Melm ligand, remains a plausible model for T-state deoxyhemoglobin. The oxy structure, 11, should not be com- pared with R-state oxyhemoglobin: the 2-Melm ligand of I I can never be unrestrained (as are relaxed oxyhemoproteins or Fe(O*)(TpivPP)( I -Melm)) . However, complex 11 may be a simple analogue of T-state oxyhemoglobin. The Fe-N Im sep- aration is essentially the same for complexes I and 11, and also for Fe(Or)(TpivPP)( I-MeIm),” consistent with the Moss- bauer results of Huynh et on isolated hemoglobin chains and T-state deoxyhemoglobin and with the recent resonance Raman results of Kincaid et al.25c on R- and T-state deoxy-

3236

Table XX. Stereochemistry of Hemoproteins and Their Model Analogues“

Journal of the American Chemical Society / 102.9 / April 23, 1980

complex M-Ct M-*P, M-N, M - N I ~ M - 0 0-0 LMO0,deg Fe(TPP)(2-Melm).EtOHh 0.42 0.55 2.086(4) 2.161(5) Fe(TpivPP)(2-Melm)-EtOH c 0.399 0.43 2.072(5) 2.095(6)

H b e CY 0.60(10) 2.1(1) 2.0(3) P 0.63(10) 2 .1(1) 2.2(3)

Mbd 0.42 0.55 2.06 2.1

erythrocruorinf 0.23 0.17 2.02 2.2

M b(02) 0.24 0.24 2.07(7) 2.03( 7) 2.02( 7) 1.21(10) I I 1 Fe(02)(TpivPP)( I -Melm)g ,h -0.02 -0.03 l.98(1) 2.07(2) 1.75(2) >1.16(4) <131(2)

oxycobaltomyoglobinj 0.25 1.99(8) I .97(8) 1.89(8) 1.26(8) 131 oxyerythrocruorink 0.38 0.30 2.04( I O ) 2.1(2) 1.8(2) 1.25(20) 170(30) Fe(Oz)(TpivPP)(2-Melm).EtOHc,h 0.086 0.1 19 1.996(4) 2.107(4) 1.898(7) >1.22(2) <129(2)

All distances in A. Estimated standard deviations, when available, are in parentheses. NI,,, coordinated nitrogen atom of the imidazole ligand; P,, center of mean porphyrinato plane. See Figures 8 and 9 for definition of N, and Ct. Reference 48; resolution 2.0 A. e References 49 and 50; resolution 2.5 A. ,f References 5 1-53; resolution 1.4 A. g Reference 12. Fe-0-0 angle and 0-0 separations affected by disorder. Reference 45 and Phillips, S. E. V.. personal communication; resolution 1.6 A. j Reference 46: resolution I .5 A.

Reference 34. This work.

References 47 and 53: resolution I .4 A.

hemoglobin. Significantly, it is the Fe-0 bond length which is lengthened in 11. This is a result of considerable interest since some discussions on c ~ o p e r a t i v i t y ~ . ~ ~ have stressed the sig- nificance of assumed changes in the Fe-NI,, and Fee-Ct sep- arations upon oxygenation. It has been proposed that i t is the Fe-Ni,,, bond which is under “tension” in T-state oxyhemo- g l ~ b i n . ~ Furthermore, in the very recent studies by Budge et a1.38339 on ligand binding to iron( 11) “capped” porphyrins40 in solution, a porphyrin with a larger “cap” possesses a lower affinity for dioxygen. Since the axial base was kept constant, the porphyrin rather than the axial base would appear to be controlling dioxygen affinity in this case. In the solid-state structure of the free-base ‘‘H2Cap” porphyrin, the porphyrin skeleton is highly buckled,41 a situation likely to occur for iron(]]) derivatives of this “capped” porphyrin.

In contrast, long Fe-Ni, and short Fe-NNo separations are observed for Fe(NO)(TPP)( I-MeIm).I4 This, and also the observation that in T-state nitrosylhemoglobin there is ap- parently rupture of the Fe-NI, bond,42-44 reflects the stronger affinity of an Fe(porphyrinato)(imidazole) complex for N O compared with 0 2 and the different trans-labilizing properties of these ligands.

Recently, preliminary reports of the structures of oxymyo- globin, oxycobaltomyoglobin, and oxyerythrocruorin have a ~ p e a r e d . ~ ~ - ~ ’ Relevant bond parameters for these and other hemoprotein^^^-^^ and their synthetic analogues are included in Table X X . For o x y m y ~ g l o b i n ~ ~ the F e - 0 - 0 unit is, reas- suringly, bent as observed for the “picket-fence” porphyrin complexes. But for o x y e r y t h r o ~ r u o r i n , ~ ~ the very large error (30’) associated with the apparently near-linear Fe-0-0 bond angle of 1 70’ should preclude any definitive conclusion. For o x y c o b a l t ~ m y o g l o b i n , ~ ~ hydrogen bonding between the coordinated oxygen atom and the distal histidine is postulated; the C o - 0 - 0 geometry is bent and similar to that for Mb(O2) and cobalt-Schiff base dioxygen complexes.54

Unexpectedly large displacements of the iron atom from the porphyrinato plane toward the coordinated histidine residue are observed for both Mb(02)45 and o x y e r y t h r o ~ r u o r i n ~ ~ (see Table XX). For these proteins this may result from the eclipsed orientation of the imidazole plane relative to the Fe-N, bonds, which does not occur for Fe(O2)(TpivPP)(I-MeIm) or 11. Upon oxygenation of M b the iron atom moves -0.3 8, toward the porphyrin plane, whereas in erythrocruorin the iron atom appears to move even further out of the plane (by -0.1 A). For oxycobaltomyoglobin the cobalt atom is 0.25 8, from the plane of the porphyrinato nitrogen atoms, whereas for the model compound^^,^^ for deoxycobaltomyoglobin, Co(porphyrin)- (1-Melm), the displacement is 0.1 3-0 16 Whether one perceives significant differences in these metrical comparisons

of proteins and models is very much a function of the faith one puts in those standard deviations estimated on occasion in the protein studies.

In the solid state and in the hemoproteins, oxygen affinity should be sensitive to the orientation of the axial base. A lower affinity is likely when the axial base is in the eclipsing ( 4 = 0’) rather than the bisecting conformation (4 = 45’). The latter conformation probably occurs for model compounds in solution where the steric constraints of crystal packing or of a protein envelope are absent. It has been generally assumed that the structures of myoglobin in single crystals and in solution a re identical, a t least in the immediate environment of the heme. One could test this assumption by comparing the dioxygen affinities of Mb in solution and in single crystals. It is possible that a more nearly eclipsing conformation for the histidine l i - gand i n low-affinity T-state hemoglobin relative to high-af- finity R-state hemoglobin may be a significant factor in the cooperative binding of small molecules by hemoglobin.

Acknowledgments. This work was supported at North- western University by the National Institutes of Health (HL-13157) and at Stanford University by the National In- stitutes of Health (GM-17880) and the National Science Foundation (CHE-75-17018). We thank the Hertz Founda- tion (K.S.S.) and the North Atlantic Treaty Organization (E.R.) for fellowship support.

Supplementary Material Aiailable: Idealized hydrogen positions (Tables IV and VI I I ) , observed and calculated structure amplitudes (Tables V and IX). metrical changes on oxygenation (Table X I I I ) , least-squares planes (Table X I V ) . dihedral angles between planes (Table XV) , nonbonded contacts (Table XVI) , and intermolecular hydrogen contacts (Table X V I I I ) (44 pages). Ordering information is given on a n y current masthead page.

References and Notes (1) (a) Northwestern University. (b) Stanford University. (c) University of Iili-

nois. (2) (a) Ibers. J. A,; Lauher, J. W.; Little, R. G. Acta Crystallogr., Sect. B 1974,

30, 268-272. (b) Little, R. G.; Ibers, J. A. J. Am. Chem. SOC. 1974, 96, 4452-4463.

(3) Basolo, F.; Hoffman, B. M.; Ibers. J. A. Acc. Chem. Res. 1975, 8, 384-

(4)

(5)

392. (a) Collman, J. P. Acc. Chem. Res. 1977, 70, 265-272. (b) Collman. J. P.; Brauman, J. I.; Doxsee. K. M.; Halbert. T. R.; Suslick. K. S. Proc. Natl. Acad. Sci. U.S.A. 1978, 75, 564-568. (a) Collman. J. P.; Halbert, T. R.; Suslick. K:S. In “02 Binding and Activation by Metal Centers”, Spiro, T. G., Ed.; Wiley: New York, 1980. (b) Jones, R. D.; Summerville, D. A,; Basolo. F. Chem. Rev. 1979, 79, 139-178. @)Reed, C. A. In “Metal Ions in Biological Systems”, Sigel, H., Ed.; Marcel Dekker: New York, 1978: pp 278-310. (d) Traylor, T. G. In “Bioorganic Chemistry’, van Tamelen, E. E., Ed.; Academic Press: New York, 1978; Chapter 4. (e) Buchler, J. W. Angew. Chem., Int. Ed. Engl. 1978, 17,407-423. (f) James, B. In “The Porphyrins”, Dolphin, D., Ed.; Academic Press: New York, 1978; Chapter 5. (9) Vaska, L. Acc. Chem. Res. 1976, 9, 175-183. (h) Erskine, R. W.; Field, B. 0. Struct. Bonding(Berlin) 1976, 28, 3-50.

Ibers, Collman. et al. / Fe(TpiL;PP)(Z-MeIm)-EtOH (6) Eisenberger, P.; Shulman, R. G.; Kincaid. B. M.; Brown, G. S.; Ogawa, S.

(7) Warshel, A. Proc. Natl. Acad. Sci. U.S.A. 1977, 74, 1789-1793. (8) Gelin, B. R.; Karplus. M. Proc. Nati. Acad. Sci..U.S.A. 1977, 74, 801-

(9) (a) Perutz, M. F. Nature (London) 1970, 228, 726-739. (b) /bid. 1972, 237,

110) (a) Perutz. M. F. Br. Med. Bull. 1976. 32. 195-208. (b) Sci. Am. 1978,

Nature (London) 1978, 274, 30-34.

805.

495-499.

3231

. . . . . . 239(6), 92-125.

(11) Collman, J. P.; Gagne, R. R.; Reed, C. A.; Robinson, W. T.; Rodley, G. A.

(12) Jameson, G. B.; Rodley, G. A.; Robinson, W. T.; Gagne, R. R.; Reed, C. A,; Proc. Nati. Acad. Sci. U.S.A. 1974, 77, 1326-1330.

COlh" . J. P. lnoro. Chem. 1978. 77. 850-857, (13) Peng, S.-M.: Ibers,-J. A. J. Am. Chem: SOC. 1976, 98, 8032-8036. (14) Scheidt, W. R . ; Piciulo, P. L. J. Am. Chem. SOC. 1976, 98, 1913-1919. (15) Jameson, G. B.; Ibers, J. A. Inorg. Chem. 1979, 78, 1200-1208. (16) Abbreviations: TPP. meso-tetraphenylporphyrinato; TpivPP, meso-te-

tra(cu,cu,cu.cu-o-pivalamidepheny1)porphyrinato; 1-Melm, 1-methylimidazole; 2-Melm, 2-methylimidazole; 1 ,2-Me21m. 1,2-dimethylimidazole; Hb, he- moglobin A; Mb, sperm whale myoglobin; EtOH, ethanol.

(17) See, for example, Bunn, H. F.; Forget, B. G.; Ranney, H. M. "Human He- moglobins"; w. B. Saunders: Philadelphia, 1977.

(18) Collman, J. P.: Brauman, J. I.; Rose, E.; Suslick, K. S. Proc. Nati. Acad. Sci. U.S.A. 1978, 75, 1052-1055.

(19) Rougee, M.; Brault, D. Biochemistry 1975, 14, 4100-4106. (20) (a) Geibel, J.; Cannon, J.; Campbell, D.; Traylor, T. G. J. Am. Chem. SOC.

1978, 700, 3575-3985. (b) White, D. K.; Cannon, J. B.; Traylor, T. G. bid. 1979, 707. 2443-2454.

(21) Although it is implicitly assumedz0 that the conformation of the porphyrin skeleton is unaffected by the nature of the covalent linkage between im- idazole and porphyrin, we would expect that at least some of the steric strain so induced In the Fe-NI, bond would be dissipated by conformational changes in the porphyrin skeleton.

(22) Marchant, L.; Sharrock, M.; Hoffman, B. M.; Munck, E. Proc. Nati. Acad. Sci. U.S.A. 1972, 69, 2396-2399.

(23) Spartalian, K.; Lang, G.; Collman, J. P.; Gagne, R. R.; Reed, C. A. J. Chem. Phys. 1975, 63, 5375-5382,

(24) Lang, G.; Marshall, W. Proc. Pbys. SOC., London 1966, 87, 3-34. (25) (a) Burke, J. M.; Kincaid, J. R.; Peters, S.; Gagne, R. R.; Collman, J. P.; Spiro,

T. G. J. Am. Chem. SOC. 1978, 700, 6083-6088. (b) Spiro, T. G.; Burke, J. M. bid. 1976, 98, 5482-5489. (c) Kincaid, J.; Stein, P.; Spiro, T. G. Proc. Nati. Acad. Sci. U.S.A. 1979, 76, 549-552. Correction: ibid. 1979, 76, 4156.

(26) Brunner, H. Naturwissenschaften 1974, 61, 129. (27) Maxwell, J. C.; Caughey. W. S. Biocbem. Biopbys. Res. Commun. 1974,

(28) Jones, R. D.; Budge, J. R.; Ellis, P. E. Jr.; Linard, J. E.; Summerville, D. A,;

(29) Jameson, G. B.; Molinaro, F. S.; Ibers, J. A,; Collman, J. P.; Brauman, J.

(30) For example, Waters, J. M.; Ibers, J. A . inorg. Cbem. 1977, 16, 3273-

(31) See paragraph at end of paper regarding supplementary material. (32) Collman, J. P.; Brauman, J. I.; Doxsee, K. M.; Halbert, T. R . ; Hayes, S. E.;

(33) Jameson, G. 8.; Robinson, W. T.; Collman, J. P.; Sorrell, T. N. inorg. Chem.

(34) Hoard, J. L. In."Porphyrins and Metalloporphyrins", Smith, K. M., Ed.; El-

60, 1309-1314.

Basolo, F . J. Organomet. Cbem. 1979, 787, 151-158.

I.; Rose, E.; Suslick, K . S. J. Am. Chem. Soc. 1978, 700, 6769-6770.

3277.

Suslick, K. S. J. Am. Cbem. SOC. 1978, 100, 2761-2766.

1978, 77, 858-864.

sevier: Amsterdam, 1975; pp 317-380.

(35) Kirner, J. F.; Reed, C. A.; Scheidt. W. R. J. Am. Chem. SOC. 1977, 99, 2557-2563.

(36) Changes in unit cell dimensions are sufficiently small that the simple dif- ference in fractional coordinations between I and II suffices for the calcu- lation of atom movements.

(37) Huynh, B. H.; Papaefthymiou, G. C.; Groves, J. L.; Wu. C. S. Biocbem. Biopbys. Res. Commun. 1974, 60, 1295-1301.

(38) Budge, J. R.; Ellis, Jr., P. E.; Jones, R. D.; Linard. J. E.; Szymanski, T.; Basolo, F.; Baldwin, J. E.; Dyer, R. L. J. Am. Cbem. SOC. 1979, 701, 4762- 4763.

(39) Budge, J. R.; Ellis, Jr., P. E.; Jones, R. D.; Linard, J. E.; Basolo. F.; Baldwin, J. E.; Dyer, R. L. J. Am. Cbem. SOC. 1979, 701, 4760-4762.

(40) (a) Almog, J.; Baldwin, J. E.; Dyer, R. L.; Peters, M. J. Am. Cbem. Soc. 1975, 97, 226-227. (b) Almog, J.; Baldwin, J. E.; Huff, J. Ibid. 1975, 97, 227- 228.

(41) Jameson, G. B.; Ibers, J. A. J. Am. Chem. Soc., in press. (42) Perutz, M. F.; Kilmartin, J. V.; Nagai, K.; Szabo, A.; Simon, S. R . Bio-

chemistry 1976, 75, 378-387. (43) Maxwell, J. C.; Caughey, W. S. Biochemistry 1976, 75, 388-396. (44) It has been suggested (Chevion, M.; Salhany. J. M.; Peisach, J.; Castillo,

C. L.; Blumberg, W. E. Isr. J. Chem. 1977, 75,311-317) that theESR results of Maxwell and C a ~ g h e y ~ ~ were on denatured nitrosylhemoglobin. The former propose that the proximal histidine remains coordinated in both R- and T-state nitrosylhemoglobin, although the proximal histidine proton at position N(1) of the chains disappears in the transition from T- to R-state hemoglobin with a consequent change in the NNO superhyperfine struc- ture.

(45) Phillips, S. E. V. Nature (London) 1978, 273, 247-248. (46) Petsko, G. A.; Rose, D.; Tsernoghon, D.; Ikeda-Saito, M.; Yonetani, T. J.

Bioi. Cbem., in press. (47) Weber, E.; Steigemann, W.; Jones, T. A.; Huber, R. J. Mol. Biol. 1978, 120,

327-336. (48) Takano, T. J. Mol. Biol. 1977, 170, 569-584. (49) Fermi, G. J. Mol. Bioi. 1975, 97, 237-256. (50) Ten Eyck. L. F.; Arnone. A. J. Mol. Bioi. 1976, 700, 3-1 1. (51) Huber, R . ; Epp, 0.; Formanek, H. J. Mol. Bioi. 1970, 52, 349-354. (52) Huber, R.; Epp, 0.; Steigemann, W.; Formanek, H. Eur. J. Biochem. 1971,

(53) Steigemann, W.; Weber, E. J. Mol. Biol,, in press. (54) For example, Rodley, G. A,; Robinson, W. T. Nature (London) 1972, 235,

438-439. Gall, R . S.; Schaefer, W. P. Inorg. Cbem. 1976, 75, 2758-

(55) Scheidt, W. R. J. Am. Chem. SOC. 1974, 96, 90-94. (56) In the report of the structure of meso-porphyrinato IX cobalt(l1) myoglobin

(Padlan, E. A.; Eaton, W. A.; Yonetani, T. J. Bioi. Cbem. 1975, 250, 7069-7073) no mention was made of the out-of-plane displacement of the cobalt atom.

(57) (a) Goodwin, H. A. Coord. Cbem. Rev. 1976, 78, 293-325. (b) Konig, E.; Ritter, G.; Madeja, K.; Bohmer. W. H. Ber. Bunsenges. Pbys. Chem. 1973, 77, 390-398. (c) Sylva, R . N.; Goodwin, H. A. Aust. J. Chem. 1968, 27, 1081-1084. (d) Konig, E.; Ritter. G.; Goodwin, H. A. Cbem. Pbys. 1974, 5, 211-223. (e) Sylva. R. N.; Goodwin, H. A. Aust. J. Cbem. 1967, 20, 479-496. (f) Dosser, R . J.; Eilbeck, W. J.; Underhill, A. E.; Edwards, P. R.; Johnson, C. E. J. Chem. SOC. A 1969,810-816.

(58) (a) Sorai, M.; Seki, S. J. Phys. Chem. Solids 1974, 35, 555-570. (b) Gutlich, P.; Koppen, H.; Link, R.: Steinhause, H. G. J. Cbem. Phys. 1979, 70, 3977-3983. (c) Rao, C. N. R.; Rao, K. J. "Phase Transitions in Solids"; McGraw-Hill: New York, 1978. (d) Haddad, M. S.; Federer, W. D.; Lynch, M. W.: Hendrickson, D. N. J. Am. Chem. Soc. 1980, 702, 1468-1470.

79, 42-50.

2763.