78

Biochimica et Biophysica Acta, 4 2 8 ( 1 9 7 6 ) 7 8 - - 9 0 © Elsevier S c i e n t i f i c P u b l i s h i n g C o m p a n y , A m s t e r d a m - - P r i n t e d in T h e N e t h e r l a n d s

BBA 27840 PROTON NMR STUDY OF COORDINATED IMIDAZOLES IN LOW-SPIN FERRIC HEME COMPLEXES ASSIGNMENT OF SINGLE PROTON HISTIDINE RESONANCE IN HEMOPROTEINS

GERD N. LA MAR *, JAMES S. FRYE and JAMES D. SATTERLEE Department of Chemistry, University of California, Davis, Calif. 95616 (U.S.A.)

(Received July 25th, 1975) (Revised manuscript received November 10th, 1975) Summary The proton signals for the coordinated axial imidazoles in a series of low-spin ferric bis-imidazole complexes with natural porphyrin derivatives have been located and assigned. The methyl signals of several methyl-substituted imidazoles have also been resolved for the mixed ligand complexes of imidazole and cyanide ion. The imidazole spectra for the bis complexes are essentially the same as those reported earlier for synthetic porphyrins, with the hyperfine shifts exhibiting comparable contributions from the dipolar and contact interactions. The contact contribution reflects spin transfer into a vacant imidazole lr orbital. The spectra of both the mono- and bis-imidazole complex concur in predicting that only the 2-H and 5-CH2 signals of an axial histidine are likely to resonate clearly outside the diamagnetic 0 to --10 ppm from TMS region in hemoproteins. However, both the 2-H and 4-H imidazole peaks are found to be too broad to detect in a hemoprotein. Hence, it is suggested that the pair of non-heme, single-proton resonances in low-spin met-myoglobin cyanides arise from the non-equivalent methylene protons at the 5-position of the histidyl imidazole. Both the resonance positions and relative linewidths in the model compounds are consistent with the data for this pair of protons in myoglobins. The possible interpretations of the average downfield bias of these signals as well as the magnitude of their spacing, are discussed in terms of the conformation of the proximal histidine relative to the heme group.

* Author to w h o m coEcespondence s h o u l d be addressed; F e l l o w o f the Alfred P. Sloan F o u n d a t i o n . Abbreviations used: PPDME, protoporphyrin dimethyl ester; DPDME, deuteroporphyrin dimethyl ester; MPDME, mesoporphyrin dimethyl ester; A2DPDME , 2,4-diaeetyldeuteroporphyrin dimethyl ester; TPP, m e s o - t e i x a p h e n y l p o r p h y r i n ; OEP, o c t a e t h y l p o r p h y r i n ; Ira, imidazole; 1-CH3-1m, 1methyl imidazole; 5-CH3-Im, 5-methyl inidazole; hist-ME, histidine m e t h y l ester.

79 Introduction

Extensive proton NMR investigations of a variety of low-spin ferric hemoproteins have firmly established the utility of the hyperfine-shifted resonances as sensitive probes for characterizing structure-function relationships in these macromolecules [1,2]. These hyperfine shifts [3] arise from a combination of the dipolar and contact contribution, i.e. AH = (~)con ~ ~)d~

(~)h,

(1)

The through-space dipolar interaction yields, for axial system, the relation [ 3] : (--~-)

=--(×11 - - X±)(3 c o s : 0

- - 1 ) r -3 ,

(2)

dip

where ×11, Y~ are the principal components of the susceptibility tensor, 0 is the angle between the unique axis and the metal-proton vector, and r is the length of this vector; (3cos:0--1)r -3 is the "geometric factor" which determines the relative dipolar shifts for all nuclei within a complex. The contact shift, which reflects the delocalized unpaired spin distribution is given by

iTH/27r)3kT '

(3)

where A / h is the Fermi contact coupling constant in Hz, g is the spectroscopic splitting factor, ~ is the Bohr magneton, S is the total spin quantum number, ~/H is the proton magnetogyric ratio, k is the Boltzmann constant and T is the temperature. Both of these paramagnetic interactions [3] are relatively short-ranged, and therefore cause substantial hyperfine shifts only for protons near the active site, (i.e., the protoporphyrin ligand), permitting their resolution [4--18] in the presence of the dominant proton band in the 0 to --10 ppm from TMS region due to the diamagnetic polypeptide chain. Analysis of these hyperfine shifts has provided a wealth of structural information on the met-cyano form of the oxygen-binding proteins myoglobin [4--10], MbCN, and hemoglobin [11,12], HbCN, as well as on the oxidized form of the electron transport proteins, the cytochromes [13--18]. These shifts have been employed in many ways, ranging from simple analytic applications of monitoring individual proton peaks subjected to various perturbations [7,8], to detailed analysis of dipolar shift in terms of geometry of the active site [9], or interpretation of contact shifts in terms of the effect of the protein on the electronic structure of the ferric ion [10]. The proton NMR spectra for the heme protons of a variety of model porphyrin complexes have been reported [19--24] whose analysis has served as the basis for interpreting the hyperfine shifts in the hemoproteins. It is now established that the heine contact shifts in both model compounds and proteins are dominated by transfered spin density in the highest filled fr(e) molecular orbital

80 [10,22,24]. The differences in the distribution of this unpaired spin about the prophyrin periphery in model compounds and proteins has been interpreted [10] in terms of the particular electronic structure forced on the iron in the protein environment by the axial bonding of the histidyl imidazole. To date, however, little attention has been paid to hyperfine shifted resonances originating from the axial histidine ligand, although valuable structural information on the axial interaction could, in principle, be obtained from them. The success in developing hyperfine shifted resonances as probes for the environment at the heme periphery [4--18] is due to the relative certainty with which peaks can be assigned either by comparison to model compounds [19-24] or by reconstituting the proteins with altered porphyrins [5]. The prominent features in such spectra have generally been the peaks with the area for three protons, assigned to the heme peripheral methyl groups. Recently, deuteration studies of protoporphyrin have not only confirmed this, but provided [8] specific methyl assignments for sperm whale MbCN. Previously, the emphasis has been on the assignment of the hyperfine-shifted resonances to heme positions. However, systematic studies of hemoproteins have revealed the existance of at least two non-exchangable single proton peaks with substantial hyperfine shifts which do not appear to originate from the heme group. The non-heme one- and two-proton resonances have been observed in MbCN's [4--10], cytochromes [13--18], and single strand HbCN's [11]. These peaks are particularly prominent in the b2 and bs-type cytochromes [17,18], since they exhibit even larger downfield shifts than the heme methyl groups. In view of the fact that the histidyl imidazole comprises the only other proton-containing ligand coordinated to the iron in the myoglobins [25], these peaks were assigned to imidazole protons [4--6,9,16]. In the absence of any data on model compounds, the pair of downfield peaks in the region --10 to --16 ppm from TMS were attributed to 2,4-protons for the histidyl imidazole, A, on the basis that these positions should exhibit the largest hyperfine shifts [4,5]. H

I

NH2\ /CH2~5 ~_ (a)CH

CO2H

IP 2,k,_ H A

Due to the highly tentative nature of these assignments, no attempts have been made to interpret changes in these shifts in terms of the environment at the axial site. In order to make assignments of the histidyl imidazole resonances and thereby develop a probe for characterizing the protein environment near the proximal histidine in a manner similar to that served by the heme methyls for the heme periphery [4--18], we have initiated an investigation of the coordinated imidazole peaks for low-spin bis-imidazole complexes of ferric porphyrins. Our interest focused on two points: one was to characterize the nature of the ironimidazole bonding, and the other was to establish which functional groups on an imida~ole ligand resonate outside the diamagnetic region (0 to --10 ppm from TMS} and thus give rise to resolvable single-proton peaks in hemoproteins. Our initial study showed [26] that for the synthetic porphyrins [27,31],

81

meso-tetraphenylporphyrin, TPP, and octaethylporphyrin, OEP, the imidazole hyperfine shifts reflected both dipolar and contact contributions [3], with the latter arising from spin transfer into the lowest vacant imidazole 7r molecular oribital. Several resonances were located well outside the diamagnetic region. In this report we extend this study to natural porphyrin complexes [28], as well as assess the effect on shifts of having a cyanide (MbCN) versus another imidazole, (cytochromes b2, bs), trans to an imidazole. A variety of imidazoles and porphyrins will be used to establish the relationship between the shifts of the synthetic and natural porphyrin complexes. The porphyrins structure is depicted in B, with the designation R = vinyl, PPDME; R = ethyl, MPDME, R = proton, DPDME, and R = acetyl, A2DPDME.

R CH3 C H 3 ~ R CH3~CH3 CH2 (~t-i2 I CH2 ~H2 I CO2CH 3 CO2CH 3 In order to maintain a consistent numbering scheme for the coordinated imidazole, we will use that depicted in C.

HSH /

H

Fe

C

Experimental The ferric porphyrin complexes of protoporphyrin IX dimethyl ester, PPFeC1, mesoporphyrin IX dimethyl ester, MPDMEFeC1, and deuteroporphyrin IX dimethyl ester, DPFeCI, were prepared from commercially available hemin (Sigma) by standard methods [28], and chromatographed on dry alumina as chloroform solutions. Their purity was established as t>99% by visible spectroscopy and their proton NMR traces as low-spin bis cyano complexes. The diacetyldeuteroporphyrin IX dimethyl ester complex [28], A2DPDMEFeC1, was a gift from W.S. Caughey. The heterocyclic bases, imidazole, Im, 1-methyl imidazole, 1-CH3-Im, and 5methyl imidazole, 5-CH3-Im, were obtained commercially and purifed as described elsewhere [26]. Histidine dimethyl ester hydrochloride was obtained commercially and converted to the free histidine methyl ester, histME, by neutralization with sodium methoxide and recrystallization from methanol.

82 The bis R-imidazole complexes were prepared by adding either a stoichiometric amount (2 equivalents) or a slight excess (2.2 equivalents) of R-imidazole. As the R-imidazole was added stepwise to a chloroform solution ~ 0 . 0 2 M in the ferric porphyrin chloride complex, the high spin peaks were progressively replaced by a set of new peaks which integration of the imidazole versus porphyrin peaks established to be the bis complexes. The equilibrium constants are generally large enough that the bis-adduct is the dominant species at a 2 : 1 Rimidazole : complex ratio (Satterlee, J.D. and LaMar, G.N., manuscript in preparation, and ref. 29). For the case of 1-CH3-Im or histME, an excess of the axial ligand was used. Proton spectra were recorded at 99,5 MHz on a JEOL PS-100 spectrometer operating in the Fourier Trnasform mode. 200 to 2000 transients were collected using bandwidths of 3 to 5 KHz, 8K data points, and ~ 2 0 ps 90 ° pulses. The probe temperature was varied by a VT-3D controller, and calibrated with a thermocouple within the probe both prior and after data collection. Chemical shifts are reported in ppm at 25°C, referenced against internal TMS. Results

The proton NMR trace illustrating the coordinated imidazole peaks in PPDMEFe(Im)2 ÷ in CDC13 is presented in Fig. 1. The observed shifts, relative to TMS, for the coordinated imidazoles of a variety of bis complexes of natural porphyrin derivatives are listed in Table I. The diamagnetic shifts for a related ruthenium(II) porphyrin complex [30] are also included. The hyperfine shifts are determined by the differences between the shift in the low-spin ferric complex and the shift in the diamagnetic Ru(II) complex. Table II lists the

6,7-0CH3

TMS

1,3,5,8-CH3 ilia

zz z |

F -20

F i g . 1. P r o t o n

2,4_~+_CH

I

I -15

®.o +H

Ii

I

I

-10

-5 CHEMICAL SHIFT, in PPM from TMS +

nmr trace of PPDMEFe(Im)

I

,.,,+.,,;

I

I

I

O

5

10

~,

2 Cl- in CDCI 3 at --4°C.

83 TABLE

I

OSBERVED

SHIFTS

P

FOR COORDINATED

M

L,L'

IMIDAZOLES

IN PMLL' a

Imidazole positions 1

2

--14.8 --14.6 --14.9 --13.3 --14.7

+ + + b +

PPDME DPDME MPDME A 2 DPDME TPP d

Fe(III) Fe(III) Fe(III) Fe(III) Fe(III)

Im Im Im Im Im

DPDME MPDME TPP d

Fe(III) Fe(III) Fe(III)

1-CH3-Im 1-CH3-Im 1-CH3-Im

MPDME TPP d

Fe(III) Fe(III)

5-CH 3-Im 5-CH3-Im

--15.8 ~--11

MPDME

Ru(II)

f

~--11 (--1.6) e _

a b c d e f

(--18.4) e (--18.1) e (--18.8) e

4 6.6 6.9 6.3 8.6

+ 6.4 + 5.3 + 7.7 +11.5 +12.3 0.9

5

--

8.9

--

8.3

--

8.7

--

8.5

--

9.0

--

8.5

--

7.5

--

7.0 -- 8.5

--10.2 ~_ --

7.5 c

b

7.8

--

~-- 7.3

7.5

~-- 7.3

b

(--16.3) e

--11

(--15.2) e

--

0.5

--

4.4

(+0.3)

S h i f t s i n p p m , i n CDC13 a t 2 5 ° C ; T M S a s r e f e r e n c e , w i t h u p f i e l d s h i f t s p o s i t i v e . Not resolved. ~ indicates shifts are approximate, due to extrapolation from low temperature. Data for TPP complexes taken from ref. 26. Methyl shifts given in parentheses. Data taken from ref. 30: the shifts are taken from various imidazoles with the appropriate stituents.

e

sub-

typical hyperfine shifts for the various positions, as well as their separation into contact and dipolar contributions by comparison to the analogous complexes of the synthetic porphyrins [26]. Fig. 2. schematically illustrates the methyl region for the mixed ligand complex, MPDMEFe(CN)(1-CH3-Im). The shifts relative to TMS for the imidazole methyl peaks as well as the heme methyls in the mixed-ligand complexes are given in Table III. The proton trace for the complex of histidine methyl ester with the synthetic porphyrin, tetraphenylporphyrin, TPPFe(histMe)2 ÷, is presented in Fig. 3. TABLE

II

TYPICAL CONTACT IMIDAZOLES a

Shift

AND DIPOLAR

CONTRIBUTIONS

TO HYPERFINE

SHIFT

OF COORDINATED

Imidazole position 1

H

(~H/H)hf - - 2 (AH/H)dipb - - 1 2 (~H/H)conC + 1 0

2

4

5

CH 3

H

H

H

CH 3

--17 -- 7 --10

10 --18 +28

--10 --18 + 8

-- 4 --12 + 8

--16 -- 7 -- 9

a Shifts in ppm, in CDCI 3 at 25°C; (AH/H)h f defined as the observed shift relative to TMS minus the s h i f t i n t h e d i a m a g n e t i c R u ( I I ) complex (see Table I) b Obtained as described in ref. 26. c D e t e r m i n e d b y E q n . 1.

84

.if b±

:~

M PDMEFe(I-CI~-Im);

bb

MPDMEFe(I-CH3-1m)(CN) & MPDMEFe(I-CHz-lm)2 ,

M PDMEFe(1-CH3"Im)(CN) e. MPDMEFe(CN)~

1)i)i i

-20

-22

il

-18 -16 SHIFT, in PPM lrom TM$

MPOMEF,(CN)2"

-;4

-/2

Fig. 2. P r o t o n n m r t r a c e s of the m i x e d - l i g a n d c o m p l e x , M P D M E F e ( 1 - C H 3 - I m ) ( C N ) in CDCI 3 a t - - 2 5 ° C . (A) P o s i t i o n of r e s o n a n c e s in M P D M E F e ( 1 - C H 3 - I M ) 2 +. (B) S p e c t r u m in t h e p r e s e n c e o f b o t h C N - and 1C H 3 - I m ~ w i t h t h e l a t t e r in excess. (C) S p e c t r u m in t h e p r e s e n c e o f b o t h C N - a n d 1 - C H 3 - I m , w i t h the f o r m e r in excess. (D) P o s i t i o n o f r e s o n a n c e s in M P D M E F e ( C N - ) ~ . T h e s u p e r s c r i p t -* i n d i c a t e s t h a t this is the 1 - C H 3 - I m m e t h y l signal.

TABLE III OBSERVED SHIFTS FOR LOW-SPIN P F e ( L L ' ) n COMPLEXES a P

L,L'

n

Ring m e t h y l s b ( A v e )

2,4-H

X-CH3 in X-CH3Im

DPDME

Im, Im CN-, CN-

+1 --I

--12.9, --13.8, --14.1, --16.2(--14.2) --11.4, --13.4, --14.6, --16.8(--14.1)

+17.5, +18.2 +17.4, +17.5

---

MPDME MPDME MPDME

1-CH3-Im, 1-CH3-Im 1-CH3-1m, C N CN-, C N -

+1 0

--14.0, --15.8, --15.9, --18.4(--16.0) --14.2, --14.4(2), --16.5 (--14.9)

---

--18.1 --14.9

--i

--13.9(2), --14.3, --15.0

MPDME MPDME MPDME

5-CH3-Im, 5-CH3-Im 5-CH3-Im, CNCN-, CN-

+1 0

--12.2, --13.4, --13.5, --15.1(--13.6) --14.0(2), --14.7, --15.7 (--14.6)

--

--16.3

--

--12

--i

--13.9(2), --14.3, --15.0

(--14.3)

(--14.3)

a Shifts in p p m , in CDCI 3 a t 25 ° C, r e f e r e n c e d a g a i n s t TMS; u p f i e l d shifts are. positive. b D i a m a g n e t i c r e f e r e n c e in F e ( I I ) c o m p l e x is - - 3 . 6 p p m f o r ring m e t h y l s , - - 1 0 . 6 p p m f o r 2,4-H, - - 1 . 6 p p m for 1-CH 3 in I - C H 3 - 1 m , a n d +0.3 p p m f o r 5-CH 3 in 5 - C H 3 - I m .

85 Ug#Sc

~lstME

P~r~-H

i

5-C-~LH

5-CH2

•

-20

I

J

-110 " J''

0 SHICT, i .

Fig. 3. P r o t o n histME.

i

PPM

[

I

10 Uo

/

2

l

=

bIMR t r a c e o f T P P F e ( h i s t M E ) 2 + in CDC13, a t - - 2 5 ° C , in t h e p r e s e n c e o f a n e x c e s s o f

Discussion

Spectra of Bis-Imidazole Complexes The proton NMR trace for PPDMEFe(Im)2 +in Fig. 1 clearly illustrates that, in addition to all the heine resonances, four peaks, each with areas of 2 protons per complex, can be located and assigned to the coordinated imidazole. The 2 : 3 relative areas of an imidazole protofi to heme methyl signal establishes the 2 : 1 stoichiometry of these complexes. The data in Table II demonstrates that these coordinated Im shifts of PPDMEFe(Im)2* have essentially the same shifts as for the previously characterized synthetic porphyrin complex [26], TPPFe(Ira): +. Thus assignment of all Im peaks can be made simply by comparison to the completely assigned spectra of the synthetic porphyrin [ 26], although our present assignments are independently confirmed by the use of the 1-CH3 and 5-CH3 substituents. Comparison of the coordinated Im shifts for the four natural porphyrin derivatives listed in Table I also reveals that they are relatively insensitive to the nature of the 2,4-substituents. The porphyrin methyl shifts in these complexes have been shown to vary from MPDMEFe(Im)2÷, for which all four methyl exhibit about the same shift (Frye, J.S. and LaMar, G.N., manuscript in preparation) to A2DPFe(Im)2÷, for which the heme methyl shift pattern reflects as much asymmetry [10,32] as in MbCN. It appears therefore that the shifts for the axial imidazole are insensitive to the distribution of the unpaired spin within the porphyrin. This probably occurs because rotation of the imidazole about the N-Fe bond averages any differences in spin transfer via dxz and dyz

[zo]. Since the coordinated imidazole hyperfine shifts for the natural porphyrins are essentially the same as for the synthetic porphyrins, and esr studies have

86 shown [33] that they exhibit very similar magnetic anisotropies, it appears reasonable to assume that the analysis of the shifts in the latter complexes reported earlier is directly applicable to our present system. Hence we conclude that the observed hyperfine shifts for a given proton or methyl group is composed of contact and dipolar contributions as found for the analogous TPP complex [26]; these values are given in Table II. This also indicates that the contact contribution reflects Fe-Im ~ bonding with direct transfer of unpaired spin density into the lowest vacant ~ orbital of imidazole. The great similarity of imidazole shifts in natural and synthetic porphyrin complexes is not unexpected, since we have shown elsewhere [22] that the porphyrin shifts for identical functional groups are also nearly the same in the two systems [20]. The shift data in Table I also confirm our earlier conclusion [26] that the only non-exchangable imidazole resonances which are likely to appear outside the 0 to --10 ppm from TMS range in low-spin hemoproteins are 2-H and 5-CH2 of the histidyl imidazole, although other histidine peaks may also be observable near --10 ppm {vide infra). Although the bis-imidazole complexes can be expected to serve as valid models for the hemoproteins containing two coordinated histidyl imidazoles [16--18] (cytochromes c3, b2, bs), it must first be determined whether they can also serve the same purpose for the more extensively studied HbCN and MbCN systems [4--14].

Spectra of Mixed Imidazole/Cyanide Complexes The most c o m m o n sixth ligand employed to induce the desired low-spin configuration for met-Mb and -Hb is the cyanide ion [25]. This suggests that particularly appropriate models would be the mono-imidazole, mono-cyano complexes. Although we could not isolate the desired complexes, we succeeded in generating the mixed-ligand species in situ, as illustrated by the schematic traces for MPDMEFe(CN)-(1-CH3-Im) given in Fig. 2. The two middle traces, which display the heme as well as the imidazole methyl signals for all species, differ in the relative amounts of CN- and 1-CH3-Im. It is clear from inspection of Fig. 2 that, in addition to the easily characterized bis(1-CH3-Im) and bis(CN) peaks, only one additional set of five equally intense methyl peaks is observed, consistent withe the expectation of the mixed ligand complex. One of these five signals is much broader than the other four due to ligand exchange contribution to the linewidth [34], and can therefore be assigned to the 1-CH3 peak of the single coordinated 1-CH3-Im. These observed shifts, as well as those of the 5-CH3-Im/CN- system, are listed in Table III. Since all solutions contained a mixture of at least two species, it was n o t possible to identify any of the aromatic imidazole protons in the mixed-ligand complexes. The shift data for DPDMEFe(Im)2 + and DPDMEFe(CN)2-, also included in Table III, show that the porphyrin shifts are quite insensitive to the nature of the axial ligand. Since both the downfield heme-methyl and upfield heme 2,4proton shifts are insensitive, this requires that both the contact and dipolar contributions to the heme shifts are very similar in the t w o complexes [10,22, 23]. Comparison of the t w o bis complexes with the mixed-ligand complex for both 1-CH3-Im and 5-CH3-Im complexes of MPDME, as listed in Table III, re-

87 veals that, while the methyl peaks (particularly the average shift) are rather insensitive to axial ligand, (change ~ 1 ppm), the coordinated imidazole methyl peaks are shifted upfield by ~ 3 ppm in the mixed-ligand relative to the bisimidazole complexes. Since the magnetic anisotropy is very similar in the bis(Im), bis(CN), and (CN)(Im) complexes, this change in imidazole methyl shifts must reflect primarily changes in the contact shift or transferred spin density. The inability to resolve the other imidazole peak in the mixed ligand complex precludes a more detailed interpretation of the shift changes. The effect of replacing one imidazole by a cyanide can be characterized by a sizable upfield bias for both 5-CH3 and 1-CH3. However, the spectra of the bis(Im) and mixed ligand complex concur in that the 5-CH3 peak resonates clearly out of the 0 to --10 ppm region of the spectrum. The fact that this 5-CH3 peak appears in the same region of the spectrum allows us to use either the bis(Im) or mixed ligand complexes as models for discussing possible peak assignments to histidines in HbCN, MbCN as well as cytochromes.

Histidine Resonances in Hemoproteins The bis-imidazole spectra indicate that only 2-H and 5-CH3 can be expected to yield resonances clearly outside the 0 to --10 ppm range. The similarity of the 5-CH3 shift suggests that this conclusion is valid for the mixed ligand complexes. The 2-H peak, however, is 300--400 Hz wide [26], or about a factor of 10 broader than the heme methyl signals, due to its close proximity to the iron and is therefore extremely unlikely [16] to be resolved in a paramagnetic protein. The 4-H peak, which in our model c o m p o u n d s resonates at the border of the 0 to --10 ppm range, is also likely to be t o o broad ( ~ 2 0 0 Hz) [26] to be observed in proteins. Most useful resonances in low-spin hemoproteins have linewidths [1,2,4--18] < 1 0 0 Hz. The 5-CH3 signal, whose counterpart in histidine is the 5-CH2 group (See A), exhibits linewidths [26] ( ~ 2 5 Hz) comparable to slightly greater than heme methyls (15--20 Hz) and can therefore be expected to be clearly observable in protein spectra. Two non-exchangable single-proton resonances, one in the range --11 to --12 ppm from TMS the other --14 to --17 ppm, have been located in most MbCN spectra [4--14]. The unliklihood that these t w o peaks arise from heme protons is emphasized by the fact that no heme resonances beside methyl groups are found below --10 ppm from TMS in A2DPFeIm2 ÷, even though the methyl shifts closely resemble those in MbCN [1,4--7]. The vinyl groups can also be eliminated since these t w o peaks persist in MbCN reconstituted with deuteroporphyrin [ 5]. Earlier tentative assignments to 2,4-H of the proximal histidyl imidazole [4--6,9], (Hist F8), can now be discarded since 2-H should resonate upfield and both 2- and 4-H of the imidazole are predicted to be too broad to detect in the myoglobin spectra. Hence these protein single-proton peaks cannot arise from an imidazole ring proton. The most likely condidates for the origin of these signals is the 5-CH2 group of Hist F-8. The 5-CH3 peak in both of our model compounds resonates in the same region as these unidentified peaks, and has the appropriate linewidth. Furthermore, the two protons in the 5-CH2 group of Hist F-8 would be nonequivalent, and hence yield two singie-proton resonances. This expectation is clearly confirmed in the low field proton trace of the histidine methylester

88 model compound, TPPFe(histME)2 +, illustrated in Fig. 3. HistME complexes with natural porphyrin derivatives gave very similar 5-CH: shifts, but the heme methyls prevent resolution of both 5-CH2 signals at any one temperature. The pair of lines of the same width centered at --15.8 ppm arise from the diastereotopic 5-methylene protons. The peak at --9.5 ppm is due to the a-CH (see A); its down field shift is consistent with a dominant dipolar interaction. The range over which these single proton resonance are found in MbCN's {similar resonances are found in single chain HbCN's [11] and cytochromes) [13--18], i.e. --11 to --17 ppm, can be rationalized by t w o distinct mechanisms. The large splitting of the pair of lines must arise from the angular dependence of contact shifts of alkyl protons attached to aromatic rings containing 7r spin density [35,36]. The 5-methyl or methylene proton complex coupling constant, as defined in Eqn. 3, can be related to the lr spin density, p, at C-5 by the well-known relation, [35,36] A) = B2 (cos 2 h CH2,CH3

(4)

where B2 is a constant ~ 1 4 0 MHz, and ~ is the angle between the C-C-H plane and the C-5 Pz axis. In the freely rotating methyl group, as in 5-CH3-Im, (cos2¢) = 2/3. However, in the proteins where the methylene group is rigid, cos2~ can range from 0 to 1.0, so that the individual protons in the methylene group can experience contact shifts ranging from zero to 50% larger than for the freely rotating methyl group, depending on their orientation relative to the imidazole ring. Since the analysis of the 5-CH3 hyperfine shift yielded [26] a dipolar contribution of --7 ppm and a contact contribution of --9 ppm, restricted values of ¢ could cause contact shifts to range from 0 to --14 ppm, and hence total shifts to range from --7 to --20 ppm from TMS. (It should be noted that the dipolar shift also depends slightly on the rotational position of a methylene proton, but this effect would be negligible compared to the effect on the contact shift). We suggest therefore that the t w o downfield, non-exchangable single-proton resonances in MbCN's arise from the 5-methylene protons of Hist F-8. Once this assignment is verified, the neglect of the angular part of the dipolar shift in conjunction with Eqn. 4 may permit the determination of the conformation of the histidine backbone relative to the imidazole 7r plane just from the splitting of the pair of lines, (i.e. this splitting is given by B2 [cos2~ -- cos2(120 ° -- ~)] P, in analogy to Eqn. 4). This would suggest that changes in this splitting between the t w o single-proton 5-CH2 peaks may provide a probe for changes in the conformation of the histidine backbone with either genetic origin or external perturbations on the protein. The average position of a pair of 5-CH2 lines from the proximal histidine in MbCN, (or any other hemoprotein), can also be affected by the angular part, b u t in a way that would be predictable upon analysis of the splitting, as discussed above. Another factor that may well be operative concerns the orbital ground state of the iron in the heme group [10]. Shulman, Glarum and Karplus have shown [10] that splitting the orbital degeneracy of the 2E state causes the lone spin on the iron to be placed in a d orbital which interacts differently with

89 the four pyrroles in the heme group. This effect has been suggested to give rise to the asymmetric spin distribution in the heme group in hemoprotein. To form a strong Im-Fe ~r bond, only one d(Tr) orbital can interact with the imidazole ~ orbital. In the imidazole of model c o m p o u n d s [20--22], the free rotation a b o u t the Fe-N bond causes the Fe-Im lr bond to occur equally probably with the d orbital containing the paired spin as with the one with the lone spin. In the proteins, the imidazole has a fixed orientation [25] relative to the d orbital containing the unpaired spin, so that its contact shift can range from zero (imidazole oriented to interact only with the doubly-occupied d orbital) to twice that for the freely rotating imidazole (imidazole oriented to interact only with the spin-containing d orbital). This latter effect could be the cause of the larger downfield shifts for single proton resonances in c y t o c h r o m e c [13--16] (--12 to --25 ppm from TMS), as well as in cytochromes b: [18] and bs [17] (--12 to --31 ppm from TMS). The exchangable l-H, located in the region --14 to --17 ppm from TMS in the model c o m p o u n d s is consistent with the assignments of the Hist F-8 exchangable single proton resonance in most of the hemoproteins [8,9]. The suggestion by Wfithrich and coworkers [17,18] that the much larger downfield shifts of the single-proton resonance in c y t o c h r o m e b2 and b5 relative to c y t o c h r o m e c and MbCN are a characteristic of hemes containing two coordinated histidyl imidazoles is partially supported by our present study, in that bis imidazole complexes exhibited further downfield 5-CH3 shifts than mixed ligand complexes. Thus the single-proton peak at --31 ppm, two-proton peak at --27, and one of the single proton peaks in the region --15 to --19 ppm from TMS could arise from the two 5-CH: groups, with a pair of the lines accidentaUy degenerate. However, the downfield bias in these b-type cytochromes is much larger than observed in model c o m p o u n d s upon replacement of one imidazole. Hence other factors, such as those discussed above, must be operative. In summary, we suggest that the pair of nonexchangable downfield singleproton resonances of met-myoglobin cyanides arise from the nonequivalent pair of 5-CH: protons of the proximal histidine. It is suggested that the splitting between this pair of resonances may provide information on the conformation of the histidine backbond relative to the imidazole ring, while the average shift may provide information on the rotational position of the imidazole plane relative to the porphyrin skelton. Thus confirmation of these assignments may provide a sensitive probe for the environment of the proximal histidine in low-spin hemoproteins.

Acknowledgements The authors are indebted to useful discussion with Professor I. Morishima of K y o t o University, and to the National Institute of Health (Grant HL-16087) for support of this work.

References 1 Wfithrich, K. (1970) Structure and Bonding 8, 5 3 - - 1 2 1 2 Phillips, W.D. (1973) NMR of Paramagnetic Molecules (La Mar, G.N.0 Horrocks, W.D. and H o l m ,

90 R.H., eds.), pP 421--478, Academic Press, New York 3 Jesson, J.P. NMR of Paramagnetic molecules (La Mar, G.N., Horrocks, W.D. and Holm, R.H., eds.), pp 421--478, Academic Press, New York 4 Wiltrich, K., Shulman, R.G. and Peisach, J. (1968) Proc. Natl. Acad. Sci. U.S. 60, 373--380 5 Shulman, R.G., Wiithrich, K., Yamane, T., Antonini, E. and Brunori, M. (1969) Proc. Natl. Acad. Sci. U.S. 63, 623--628 6 Wtithrich, K., Shulman, R.G., Yamane, T., Wyluda, B.J., Hugli, T.E. and Gurd, F.R.N. (1970) J. Biol. Chem. 245, 1947--1953 7 Shulman, R.G., Peisach, J, and Wyluda, B.J. (1970) J. Mol. Biol. 4 8 , 5 1 7 - - 5 2 3 8 Mayer, A., Ogawa, S., Shulman, R.G., Yamane, T., Cavalerio, J.A.S., Roeha Gonsalves, A.M. d'A., Kenner, G.W. and Smith, K.M. (1974) J. Mol. Biol. 8 6 , 7 4 9 - - 7 5 6 9 Sheard, B., Yamane, T. and Shulman, R.G. (1970) J. Mol. Biol. 53, 35--48 10 Shulman, R.G., Glarum, S.H. and Karplus, M. (1971) J. Mol. Biol. 57, 93--115 11 Ogawa, S., Shulman, R.G., and Yamane, T. (1972) J. Mol. Biol. 70, 291--300 12 Shulman, R.G., Ogawa, S., Mayer, A., and Castillo, C.L. (1973) Ann. N.Y. Acad. 8ci. 222, 9--19 and references therein 13 Wiithrieh, K. (1969) Proc. Natl. Aead. Sci. U.S. 63, 1071--1078 14 McDonald, C.C. and Phillips, W.D. (1973) Biochemistry 12, 3170--3186 15 Smith, G.M. and Kamen, M.D. (1974) Proe. Natl. Acad. Sei. U.S. 71, 4303--4306 16 McDonald, C.C., Phillips, W.D. and LeGall, J. (1974) Biochemistry 13, 1952--1959 17 Keller, R.M. and Wiithrieh, K. (1972) Bioehim. Biophys. Aeta 285, 326--336 18 Keller, R., Groudinsky and Wiithrich, K. (1973) Biochim. Biophys. Aeta 3 2 8 , 2 3 3 - - 2 3 8 19 Wiithrich, K., Shulman, R.G., Wyluda, B.J. and Caughey, W.S. (1969) Proc. Natl. Acad. Sci. U.S. 62, 636--643 20 Kurland, R.J., Little R.G., Davis, D.G. and Ho, C. (1971) Biochemistry 10, 2237--2246 21 Momenteau, M. (1973 Biochhn. Biophys. Ac'ta 304, 814--827 22 La Mar, G.N. and Walker, F.A. (1973) J. Amer. Chem. Soc. 95, 1782--1790 23 Frye, J.S. and La Mar, G.N. (1975) J. Amer. Chem. Soc. 97, 3 5 6 1 - - 3 5 6 2 24 Hill, H.A.O. and Morallee, K.G. (1972) J. Amer. Chem. Soc. 94, 731--738 25 Antonini, E., and Brunori, M., (1971) " H e m o g l o b i n and Myoglobin in t he i r Reactions with Ligands", North-Holland, Amsterdam 26 Satterlee, J.D. and La Mar, G.N. (1976) J. Am. Chem. Soc. 27 Adler, A. et al, (1967) J. Org. Chem. 32, 476 28 Caughey, W.S., Alben, J.O., Fujimoto, W.Y. and York, J.L. (1966) J. Org. Chem. 31, 2631--2640 29 Abbott, E.H. and Rafson, P.A. (1974) J. Amer. Chem. Soc. 96, 7378--7379 30 Chen, C.-C. (1972), Ph.D. Thesis, Yale University 31 Wittock, H.W., Hannauer, R. (1969) J. Or~. Chem 33, 2169--2171 32 O'Keefee, Jr., D.H. (1974) Ph.D. Thesis, Arizona State University 33 Peisach, J., Blumberg, W.E. and Adler, A. (1973) Ann. N.Y. Acad. Sci. 206° 310--326 34 La Mar, G . N , and Walker, F.A. (1972), J. Amer. Chem. Soc. 94, 8607--8608 35 La Mar, G.N., (1973) NMR in Paramagnetic Molecules (La Mar, G.N., Horrocks, W.D. and Holm, R.H., eds.), pp. 85--126 Academic Press, New York 36 Stone, E.W. and Maki, A.H. (1962) J. Chem Phys. 37, 1 326--1333