Electron paramagnetic resonance spectroscopy as a probe of hydrogen bonding in heme-Thiolate proteins

Article

pubs.acs.org/IC

Cite This: Inorg. Chem. XXXX, XXX, XXX −XXX

Electron Paramagnetic Resonance Spectroscopy as a Probe of

Hydrogen Bonding in Heme-Thiolate Proteins

Matthew R. Dent,^†Michael W. Milbauer,^†Andrew P. Hunt,^‡Michael M. Aristov,^†Ilia A. Guzei,^†

Nicolai Lehnert,^‡and Judith N. Burstyn*^,†

^†Department of Chemistry, University of Wisconsin-Madison, 1101 University Avenue, Madison, Wisconsin 53706, United States

^‡Department of Chemistry, University of Michigan, 930 North University Avenue, Ann Arbor, Michigan 48109, United States

S

* Supporting Information

ABSTRACT: Despite utilizing a common cofactor binding motif,

hemoproteins bearing a cysteine-derived thiolate ligand (heme-

thiolate proteins) are involved in a diverse array of biological

processes ranging from drug metabolism to transcriptional regulation.

Though the origin of heme-thiolate functional divergence is not well

understood, growing evidence suggests that the hydrogen bonding

(H-bonding) environment surrounding the Fe-coordinating thiolate

inﬂuences protein function. Outside of X-ray crystallography, few

methods exist to characterize these critical H-bonding interactions.

Electron paramagnetic resonance (EPR) spectra of heme-thiolate

proteins bearing a six-coordinate, Fe(III) heme exhibit uniquely

narrow low-spin (S = 1/2), rhombic signals, which are sensitive to

changes in the heme-thiolate H-bonding environment. To establish a well-deﬁned relationship between the magnitude of g-

value dispersion in this unique EPR signal and the strength of the heme-thiolate H-bonding environment, we synthesized and

characterized of a series of six-coordinate, aryl-thiolate-ligated Fe(III) porphyrin complexes bearing a tunable intramolecular H-

bond. Spectroscopic investigation of these complexes revealed a direct correlation between H-bond strength and g-value

dispersion in the rhombic EPR signal. Using density functional theory (DFT), we elucidated the electronic origins of the

narrow, rhombic EPR signal in heme-thiolates, which arises from an Fe−S p_π−d_πbonding interaction. Computational analysis

of the intramolecularly H-bonded heme-thiolate models revealed that H-bond donation to the coordinating thiolate reduces

thiolate donor strength and weakens this Fe−S interaction, giving rise to larger g-value dispersion. By deﬁning the relationship

between heme-thiolate electronic structure and rhombic EPR signal, it is possible to compare thiolate donor strengths among

heme-thiolate proteins through analysis of low-spin, Fe(III) EPR spectra. Thus, this study establishes EPR spectroscopy as a

valuable tool for exploring how second coordination sphere eﬀects inﬂuence heme-thiolate protein function.

cytochrome P450 enzymes (a large group of type-1 heme-

thiolate proteins), and a growing body of work implicates this

H-bonding network in maintaining stability of the Fe−S bond

and in modulating reactivity of the heme.²⁻¹⁰In a number of

type-2 heme-thiolate proteins, H-bonding interactions may

facilitate ligand switching at the heme,¹¹⁻¹⁴although structural

characterization of these interactions is limited.

Type-1 heme-thiolate proteins, which include a number of

small-molecule-activating enzymes such as cytochromes P450

(Cyt P450s), chloroperoxidases, and nitric oxide synthases/

reductases, possess a well-characterized H-bonding network

that stabilizes thiolate coordination and modulates heme

reactivity. In the exemplar camphor-hydroxylating P450_cam

from Pseudomonas putida, three amide N−H groups from

three amino acid residues (Leu³⁵⁸, Gly³⁵⁹, and Gln³⁶⁰) act as H-

bond donors to the Cys(S⁻) ligand (Cys³⁵⁷) (Figure 1,

left).¹⁵⁻¹⁸Additionally, the side chain amide of Gln³⁶⁰serves as

1. INTRODUCTION

Hemoproteins bearing an axial, cysteine-derived thiolate ligand

(heme-thiolate proteins) are an important class of metal-

loproteins with diverse biological functions ranging from drug

metabolism to transcriptional regulation. Importantly, proper-

ties of the b-type heme cofactor that give rise to diverse

biological functions in heme-thiolate proteins also impart

distinct spectroscopic characteristics. These properties include

sixth axial ligand identity, iron coordination number and spin

state, solvent exposure, and the propensity to react with

substrates or bind small gaseous molecules. Two classes of

heme-thiolate proteins emerge based on these functional and

spectroscopic features: type-1 heme-thiolates, which act as

small-molecule activators, and type-2 heme-thiolates, which act

primarily as small molecule sensors.¹

We hypothesize that a key contributing factor to functional

divergence in heme-thiolate proteins is the structure of the

hydrogen-bonding (H-bonding) environment surrounding the

coordinating Cys(thiolate) (Cys(S⁻)) ligand. A well-deﬁned

H-bonding pocket envelops the heme-bound Cys(S) in

Received: August 19, 2019

A

DOI: 10.1021/acs.inorgchem.9b02506

Inorg. Chem. XXXX, XXX, XXX−XXX

Inorganic Chemistry

Article

Figure 1. Comparison of heme-thiolate H-bonding environments in archetypical type-1 and type-2 heme-thiolate proteins. Left: Crystal structure

of the thiolate H-bonding pocket in Cyt P450_camfrom P. putida, a type-1 heme-thiolate.¹⁵Residues involved in the thiolate H-bonding pocket are

labeled, and important H-bonds are depicted as black dashes. Donor−acceptor distances range from 3.08 to 3.56 Å. Structure was visualized using

Pymol (v1.3). Right: Proposed H-bonding environment in Rr CooA, a type-2 heme-thiolate.

a

Table 1. Comparison of Experimental g-Values for Low-Spin, Rhombic EPR Signals Observed in Fe(III) Hemoproteins

protein

cytochrome b₅

cytochrome c

hCBS

PxRcoM-1

RrCooA

axial heme ligands

g_z

g_y

g_x

ref

His/His

His/Met

3.03

3.08

2.49

2.48

2.46

2.41

2.56

2.21

2.14

2.31

2.26

2.25

2.26

2.27

1.40

1.35

1.87

1.88

1.89

1.93

1.87

33

34

35

28

30

4

Cys(S⁻)/His

Cys(S⁻)/Pro

Cys(S⁻)/H₂O

Cys(S⁻)/Im

cyt P450_cam

cyt P450_cam+ imidazole (Im)

36

a

Historical assignments of g_max, g_mid, and g_minare given as g_z, g_y, and g_xin this table.

a H-bond donor to the carbonyl O atom of Cys³⁵⁷. Synthetic

model studies highlight the importance of this H-bonding

environment: Varying the number of H-bonds to thiolate

modulates Fe−S covalency and inﬂuences the redox potential

of thiolate-ligated porphyrins.¹⁹⁻²¹In Cyt P450 proteins,

disruptions to this H-bonding pocket give rise to increased

formation of the inactive, thiolate-protonated P420 species via

weakening the Fe−S bond.⁴⁻⁶Thus, an important role of the

thiolate H-bonding network is to maintain the thiolate

coordination that is critical in (1) facilitating O−O bond

cleavage of the hydroperoxo intermediate via the “thiolate-

push” eﬀect and (2) increasing the basicity of the protonated

ferryl species in compound II.^2,3,22−26Hunt and Lehnert

recently elucidated the electronic origin of the thiolate-push

eﬀect in synthetic models of cytochrome P450 nitric oxide

reductase.¹⁰They found that the thiolate ligand exhibits a

strong σ-trans eﬀect, which attenuates Fe−NO and N−O bond

strengths in Fe(III) heme-thiolate nitrosyl adducts through

population of an antibonding Fe−N−O σ* orbital. DFT

calculations demonstrated that the O−O bond in Compound 0

of the Cyt P450 monooxygenase catalytic cycle is also

weakened via this thiolate σ-trans eﬀect. Importantly, this

study demonstrated that thiolate σ-donor strength, a property

that may be tuned via changes in the thiolate H-bonding

environment, directly modulates the strength of the σ-trans

eﬀect.

mediated ligand switch” occurs in which the thiolate ligand is

replaced with another protein-derived ligand upon reduction

from Fe(III) to Fe(II).^11,12,27,28A second change in heme

coordination, in which a small molecule replaces an axial

protein-derived ligand, occurs at the six-coordinate, Fe(II)

heme only after the thiolate ligand is replaced. Binding of a

small gaseous molecule (such as carbon monoxide or nitric

oxide) to heme allosterically modulates protein function, often

leading to enhanced binding to a DNA promoter site.²⁹⁻³¹In

the archetypical type-2 heme-thiolate protein, CooA (CO

oxidation activator), we speculate that a strong, directional H-

bond may exist between the coordinating Cys⁷⁵and the

protonated imidazolium ring of His⁷⁷(Figure 1, right); in

support of this hypothesis, distinct changes in the EPR

spectrum of the low-spin, Fe(III) heme are observed upon

mutation of His⁷⁷to Tyr.¹²No crystal structure of CooA with

heme in the Fe(III) state currently exists, possibly due to the

dynamic nature of this protein in the inactive state.³²There are

limited crystallographic data for type-2 heme-thiolate proteins

in general; therefore, it is diﬃcult to compare and contrast

thiolate H-bonding networks between type-1 and type-2

proteins. This limitation precludes a comprehensive assess-

ment of the role of thiolate H-bonding in the functional

diﬀerentiation between these protein types.

Ferric heme-thiolate proteins exhibit a rhombic, low-spin (S

= 1/2), electron paramagnetic resonance (EPR) signal with

characteristically small shifts from g_e∼ 2.00 (g-shifts). Table 1

summarizes the set of g-values for a number of hemeproteins

with diﬀerent axial ligands. Hemoproteins with axial His/His

(cytochrome b₅) and His/Met ligation (cytochrome c) exhibit

a very anisotropic, rhombic signal with g-values in the range of

g ≈ 3.1−1.35.^33,34In contrast, heme-thiolate proteins exhibit a

rhombic signal with signiﬁcantly less g anisotropy, with g-values

The thiolate H-bonding environment in type-2 heme-

thiolate proteins likely facilitates redox-mediated ligand

switching and may diﬀer signiﬁcantly from the environment

in type-1 heme thiolate proteins. Type-2 heme-thiolate

proteins are often small-molecule sensors involved in signal-

dependent transcriptional regulation.¹The type-2 heme exists

as a low-spin species with two axial ligands; however, a “redox-

B

DOI: 10.1021/acs.inorgchem.9b02506

Inorg. Chem. XXXX, XXX, XXX−XXX

Inorganic Chemistry

Article

in a very narrow range between g ≈ 2.5 and 1.9.^{4,28,30,35,36}

Blumberg and Peisach noted the unique features of heme-

thiolate proteins in their analysis of low-spin, Fe(III)

hemoprotein EPR spectra in the 1970s, and the characteristic

EPR signal facilitated the identiﬁcation of Cys(S⁻) as an axial

heme ligand when Cyt P450s were originally character-

ized.³⁷⁻³⁹This unique rhombic signal continues to be useful

in identifying new heme-thiolate proteins;⁴⁰however, the

underlying features of the heme electronic structure that give

rise to low g anisotropy in heme-thiolate proteins have not

been elucidated.

Herein, we uncover the electronic origins of the character-

istically small g-shifts observed in rhombic EPR spectra of low-

spin, Fe(III) heme-thiolate proteins. Additionally, we establish

a relationship between the magnitude of the g-shift and heme-

thiolate H-bond strength. We characterize a series of six-

coordinate, aryl-thiolate-ligated Fe(III) porphyrin models

bearing a tunable, intramolecular H-bond using computational

and spectroscopic methods. Our ﬁndings demonstrate that

EPR spectroscopy, which is extremely sensitive to changes in

Fe−S bonding, is particularly well-suited to probe second

coordination sphere eﬀects in heme-thiolate proteins. Specif-

ically, we show that EPR spectroscopy reports on the strength

of the thiolate H-bonding environment and therefore

represents an important tool for understanding how this

critical second coordination sphere eﬀect inﬂuences heme-

thiolate protein function.

suspension of 2,2′-disulfanediyldibenzoic acid (5.5 g, 18.0 mmol) and

catalytic dimethylformamide in CH₂Cl₂(75 mL) at room temper-

ature. The reaction mixture was heated to reﬂux and stirred for 4 h

until no solids remained. The reaction mixture was then hot-ﬁltered,

and solvent was removed from the collected ﬁltrate in vacuo. The

resulting amber, crystalline solid was used without further puriﬁcation.

Aniline (0.8 mL, 8.7 mmol, 3 equiv) was added slowly to a suspension

of 2,2′-disulfanediyldibenzoyl chloride (1.00 g, 2.9 mmol) and

NaHCO₃(0.979 g, 11.7 mmol, 4 equiv) in CH₂Cl₂(30 mL), and

the reaction mixture was stirred overnight at room temperature. The

precipitate was collected by vacuum ﬁltration, suspended in HCl

(aqueous; 3.7%, 40 mL) and stirred for 15 min. Solids were collected

by vacuum ﬁltration and washed with Et₂O (40 mL) to aﬀord 2,2′-

1

disulfanediylbis(N-phenylbenzamide) as an oﬀ-white solid. H NMR

(DMSO-d₆) δ 10.57 (s, 2H), 7.81−7.73 (m, 8H), 7.52 (td, J = 7.7, 1.5

Hz, 2H), 7.43−7.36 (m, 6H), 7.17−7.10 (m, 2H). ¹³C NMR (126

MHz, DMSO-d₆) δ 166.17, 139.33, 136.92, 135.17, 131.91, 129.22,

128.95, 126.84, 126.74, 124.46, 120.60. ESI-MS [M + Na]⁺: 479.0864

(calcd), 479.0825 (exptl). Yield: 1.17 g (2.6 mmol), 87%.

2.3.1.1. 2,2′-Disulfanediylbis(N-(p-tolyl)benzamide) (SS-CH₃). p-

Toluidine (0.937 g, 8.7 mmol, 3 equiv) was used as a starting material

following the general procedure above to aﬀord 2,2′-disulfanediylbis-

1

(N-(p-tolyl)benzamide) as an oﬀ-white solid. H NMR (400 MHz,

DMSO-d₆) δ 10.48 (s, 2H), 7.77 (dd, J = 7.7, 1.4 Hz, 2H), 7.74 (dd, J

= 8.1, 1.1 Hz, 2H), 7.65 (d, J = 8.2 Hz, 4H), 7.51 (td, J = 7.7, 1.5 Hz,

2H), 7.39 (td, J = 7.4, 1.2 Hz, 2H), 7.18 (d, J = 8.2 Hz, 4H), 2.30 (s,

6H). ¹³C NMR (126 MHz, DMSO-d₆) δ 165.47, 136.45, 136.34,

134.71, 132.96, 131.34, 129.11, 128.40, 126.31, 126.20, 120.11, 20.52.

¹³C NMR (126 MHz, DMSO-d₆) δ 165.95, 136.93, 136.82, 135.19,

133.43, 131.82, 129.59, 128.88, 126.79, 126.68, 120.59, 21.00. ESI-

MS [M + Na]⁺: 507.1177 (calcd), 507.1152 (exptl). HRMS (+ESI)

m/z: [M + Na]⁺Calculated for C₂₈H₂₄N₂O₂S₂Na: 507.1177. Found:

507.1152. Yield: 1.23 g (2.5 mmol), 85%.

2.3.1.2. 2,2′-Disulfanediylbis(N-(4-chlorophenyl)benzamide) (SS-

Cl). 4-Chloroaniline (1.16 g, 8.7 mmol, 3 equiv) was used as a starting

material following the general procedure above to aﬀord 2,2′-

disulfanediylbis(N-(4-chlorophenyl)benzamide) as an oﬀ-white solid.

¹H NMR (DMSO-d₆) δ 10.71 (s, 2H), 7.86−7.77 (m, 6H), 7.75 (dd,

J = 8.1, 1.1 Hz, 2H), 7.53 (td, J = 7.7, 1.5 Hz, 2H), 7.47−7.43 (m,

4H), 7.40 (td, J = 7.5, 1.1 Hz, 2H). ¹³C NMR (126 MHz, DMSO-d₆)

δ 166.22, 138.29, 136.96, 134.84, 132.07, 129.15, 129.04, 128.07,

126.90, 126.82, 122.12. ESI-MS [M + Na]⁺: 547.0084 (calcd),

547.0078 (exptl). HRMS (+ESI) m/z: [M + Na]⁺Calculated for

C₂₆H₁₈Cl₂N₂O₂S₂Na: 547.0084. Found: 547.0078. Yield: 1.39 g (2.7

mmol), 91%.

2. EXPERIMENTAL SECTION

2.1. General Methods. Unless otherwise speciﬁed, reagents were

used as received from commercial suppliers (Sigma-Aldrich, Acros

Organics, Alfa Aesar, and Cambridge Isotope Laboratories, Inc.). The

preparation and handling of all O₂- and H₂O-sensitive materials was

carried out under inert conditions (N₂or Ar gas) using standard

Schlenk techniques or in an N₂-atmosphere MBraun glovebox

equipped with a circulating puriﬁer (O₂, H₂O < 0.1 ppm). For

procedures involving the synthesis of thiol ligand precursors, solvents

were dried by elution through alumina. For procedures involving

porphyrin complexes, all dry solvents were distilled from CaH₂under

N₂, degassed via ﬁve freeze−pump−thaw cycles, and stored over

appropriately sized (3 or 4 Å) activated molecular sieves in a glovebox

until used, unless otherwise stated.

2.3.1.3. 2,2′-Disulfanediylbis(N-(4-(triﬂuoromethyl)phenyl)-

benzamide) (SS-CF₃). 4-(Triﬂuoromethyl)aniline (1.1 mL, 8.7

mmol, 3 equiv) was used as a starting material following the general

procedure above to aﬀord 2,2′-disulfanediylbis(N-(4-

(triﬂuoromethyl)phenyl)benzamide) as an oﬀ-white solid_.¹H NMR

(DMSO-d₆) δ 10.92 (s, 2H), 7.99 (d, J = 8.6 Hz, 4H), 7.83 (dd, J =

7.7, 1.4 Hz, 2H), 7.78−7.75 (m, 6H), 7.55 (td, J = 7.7, 1.4 Hz, 2H),

7.42 (td, J = 7.5, 1.1 Hz, 2H). ¹³C NMR (126 MHz, DMSO-d₆) δ

166.16, 142.47, 136.56, 134.19, 131.81, 128.77, 126.53, 126.49,

126.08 (q, J = 3.9 Hz), 124.37 (q, J = 271.3 Hz), 123.97 (q, J = 32.0

Hz), 120.04. HRMS (+ESI) m/z: [M + Na]⁺Calculated for

C₂₈H₁₈F₆N₂O₂S₂Na: 615.0612. Found: 615.0598. Yield: 1.61 g (2.7

mmol), 93%.

2.3.1.4. 2,2′-Disulfanediylbis(N-(4-nitrophenyl)benzamide) (SS-

NO₂). The general procedure was modiﬁed by heating the reaction to

reﬂux and stirring for 2 days following the addition of 4-nitroaniline

(1.21 g, 8.7 mmol, 3 equiv). 2,2′-Disulfanediylbis(N-(4-nitrophenyl)-

benzamide) was obtained as an oﬀ-white solid. ¹H NMR (DMSO-d₆)

δ 11.13 (s, 2H), 8.32−8.27 (m, 4H), 8.03 (d, J = 9.3 Hz, 4H), 7.86

(dd, J = 7.6, 1.5 Hz, 2H), 7.81−7.76 (m, 2H), 7.57 (td, J = 7.8, 1.5

Hz, 2H), 7.47−7.41 (m, 2H). ¹³C NMR (126 MHz, DMSO-d₆) δ

166.31, 145.04, 142.74, 136.66, 133.90, 132.03, 128.92, 126.62,

126.60, 124.93, 119.81. HRMS (+ESI) m/z: [M + Na]⁺Calculated

for C₂₆H₁₈N₄O₆S₂Na: 569.0565. Found: 569.0561. Yield: 1.25 g (2.3

mmol), 77%.

2.2. Physical Measurements. 2.2.1. Electronic Absorption

Spectroscopy. An Analytik Jena Specord S600 spectrometer was

used to record electronic absorption spectra of solutions in anaerobic

screw-cap quartz cuvettes, which were prepared in a glovebox.

2.2.2. NMR Spectroscopy. Proton and ﬂuorine NMR spectra were

recorded on a Bruker Avance III 500 MHz instrument or on Varian

NMRS 500 or 700 MHz spectrophotometers at room temperature

(20−22 °C). All spectra were referenced to internal solvent peaks

(e.g., CD₂Cl₂: 5.32 ppm).

2.2.3. EPR Spectroscopy. Electron paramagnetic resonance (EPR)

spectra were recorded on a Bruker X-band EMX or ELEXSYS E500

spectrophotometer equipped with an Oxford Instruments liquid

helium cryostat. EPR spectra were obtained on frozen solution

samples (0.5−2 mM) in a 1:1 mixture of CH₂Cl₂and toluene, using

1−20 mW microwave power and 100 kHz ﬁeld modulation with the

amplitude set to 3 or 8.3 G. The g-values for each EPR spectrum were

extracted from simulations performed using EasySpin (v5.2.23).⁴¹

2.2.4. Elemental Analysis. Elemental analyses (C, H, and N) were

performed at Atlantic Microlab.

2.2.5. Mass Spectrometry. High-resolution mass spectrometry

(HRMS) experiments were carried out using a Thermo Q Exactive

Plus electrospray ionization-quadrupole-ion trap mass spectrometer.

2.3. Syntheses. 2.3.1. General Procedure for the Synthesis of

2,2′-Disulfanediylbis(N-phenylbenzamide) (SS-H) and Derivatives.

Oxalyl chloride (4 mL, 47 mmol, 2.6 equiv) was added to a

C

DOI: 10.1021/acs.inorgchem.9b02506

Inorg. Chem. XXXX, XXX, XXX−XXX

Inorganic Chemistry

Article

2.3.1.5. 2-Mercapto-N-phenylbenzamide (HS-H, Method A). 2,2′-

Disulfanediylbis(N-phenylbenzamide) (456 mg, 1 mmol) and

triphenylphosphine (1.18 g, 4.5 mmol) were suspended in a 50/50

mixture of CH₃CN/H₂O and stirred at reﬂux for 1 h. The suspension

was then added to a separatory funnel and mixed with NaCl

(aqueous) brine (15 mL) and hexanes (15 mL). The middle CH₃CN

layer was collected and dried with MgSO₄, and solvent was removed

in vacuo. The crude solid was puriﬁed by silica column

chromatography (100% CH₂Cl₂) to aﬀord 2-mercapto-N-phenyl-

C₁₃H₉N₂O₃S: 273.0339. Found: 273.0640. Yield: 343 mg (1.3 mmol),

62%.

2.3.2. General Procedure for the Synthesis of Five-Coordinate,

Fe(III) Thiolate-Ligated Porphyrin Complexes. All iron-containing

tetraphenylporphyrinate (TPP²⁻) complexes were prepared using the

same general procedure recently described for the synthesis of high-

spin, ﬁve-coordinate [Fe(TPP)(S-R)] complexes, by mixing the

desired thiol ligand precursor with [Fe(TPP)(OCH₃)].¹⁰All

complexes were recrystallized by dissolution in a minimal amount

of toluene, followed by layering of the solutions with hexanes and/or

methanol.

1

benzamide as a white solid. H NMR (500 MHz, CDCl₃) δ 7.77 (s,

1H), 7.61 (d, J = 8.0 Hz, 2H), 7.58 (dd, J = 7.7, 1.5 Hz, 1H), 7.39−

7.34 (m, 3H), 7.30 (td, J = 7.6, 1.5 Hz, 1H), 7.22−7.14 (m, 2H), 4.56

(s, 1H). ¹³C NMR (126 MHz, CDCl₃) δ 166.59, 137.61, 133.55,

132.98, 131.39, 131.01, 129.14, 128.03, 125.50, 124.87, 120.25.

HRMS (+ESI) m/z: [M + H]⁺Calculated for C₁₃H₁₂NOS: 228.0489.

Found: 228.0489. Yield: 395 mg (1.7 mmol), 86%.

2.3.1.6. 2-Mercapto-N-(p-tolyl)benzamide (HS-CH₃). 2,2′-

Disulfanediylbis(N-(p-tolyl)benzamide) (485 mg, 1 mmol) was

used as a starting material following Method A above to aﬀord 2-

mercapto-N-(p-tolyl)benzamide as a white solid. ¹H NMR (CDCl₃) δ

7.65 (s, 1H), 7.58 (d, J = 7.6 Hz, 1H), 7.50 (d, J = 7.9 Hz, 2H), 7.38

(dd, J = 7.8, 1.3 Hz, 1H), 7.31 (td, J = 7.6, 1.5 Hz, 1H), 7.22 (dd, J =

7.6, 1.3 Hz, 1H), 7.19 (t, J = 7.1 Hz, 2H), 4.62 (s, 1H), 2.35 (s, 3H).).

¹³C NMR (126 MHz, chloroform-d) δ 166.69, 135.15, 134.64,

133.66, 133.04, 131.36, 130.97, 129.67, 128.14, 125.50, 120.49, 21.03.

HRMS (+ESI) m/z: [M + H]⁺Calculated for C₁₄H₁₄NOS: 242.0645.

Found: 242.0645. Yield: 482 mg (2.0 mmol), 99%.

2.3.2.1. [Fe(TPP)(S-CH₃)]. Electronic absorption (CH₂Cl₂): 380

1

(shoulder), 407, 515, 576, 719 nm. H NMR (CD₂Cl₂): 71.4 (br, β-

pyrrole H), 64.1 and 56.4 (s, meta-SPh’s), 12.7 (br-s, meta-Ph TPP),

9.6 (br-s, ortho-Ph TPP), 6.9 (br-s, para-Ph TPP), 4.1 (s, SPh-NHPh-

para-CH₃), −85.2 (s, para-SPh), −92.6 (br, ortho-SPh) ppm. Other

SPh-NHPh-pCH₃signals observed: 8.3 (br), 4.8 (br) and −109.9

(br) ppm. Anal. Calcd for C₅₈H₄₀FeN₅OS: C, 76.48; H, 4.43; N, 7.69.

Found: C, 76.35; H, 4.47; N, 7.61. Yield: 47.9 mg, 78.8% (starting

[Fe(TPP)(OCH₃)] mass was 46.66 mg). EPR (1:1 CH₂Cl₂/toluene):

g_x= 1.95, g_y= 1.88, g_z= 2.04; E/D = 0.040.

2.3.2.2. [Fe(TPP)(S-H)]. Electronic absorption (CH₂Cl₂): 380

1

(shoulder), 407, 515, 576, 717 nm. H NMR (CD₂Cl₂): 71.6 (br,

β-pyrrole H), 63.7 and 56.5 (s, meta-SPh’s), 12.8 (br-s, meta-Ph

TPP), 9.8 (br-s, ortho-Ph TPP), 6.9 (br-s, para-Ph TPP), 6.5 (s, SPh-

NHPh-para-H), −85.2 (s, para-SPh), −92.8 (br, ortho-SPh) ppm.

Other SPh-NHPh signals observed: 8.3 (br), 4.9 (br) and −111.3

(br) ppm. Anal. Calcd for C₅₇H₃FeN₅OS: C, 76.34; H, 4.27; N, 7.81.

Found: C, 75.66; H, 4.29; N, 7.62. Yield: 117.6 mg, 77.8% (starting

[Fe(TPP)(OCH₃)] mass was 101.4 mg). EPR (1:1 CH₂Cl₂/toluene):

g_x= 1.95, g_y= 1.88, g_z= 2.085; E/D = 0.034.

2.3.1.7. 2-Mercapto-N-(4-chlorophenyl)benzamide (HS-Cl). 2,2′-

Disulfanediylbis(N-(4-chlorophenyl)benzamide) (525 mg, 1 mmol)

was used as a starting material following Method A above to aﬀord 2-

1

mercapto-N-(4-chlorophenyl)benzamide as a white solid. H NMR

2.3.2.3. [Fe(TPP)(S-Cl)]. Electronic absorption (CH₂Cl₂): 380

(500 MHz, CDCl₃) δ 7.79 (s, 1H), 7.59−7.53 (m, 3H), 7.37 (dd, J =

7.9, 1.3 Hz, 1H), 7.35−7.29 (m, 3H), 7.21 (td, J = 7.5, 1.3 Hz, 1H),

4.48 (s, 1H). ¹³C NMR (126 MHz, CDCl₃) δ 166.50, 136.19, 133.20,

133.01, 131.55, 131.22, 129.90, 129.16, 128.05, 125.59, 121.46.

HRMS (+ESI) m/z: [M + H]⁺Calculated for C₁₃H₁₁ClNOS:

262.0099. Found: 262.0097. Yield: 384 mg (1.5 mmol), 73%.

2.3.1.8. 2-Mercapto-N-(4-(triﬂuoromethyl)phenyl)benzamide

(HS-CF₃, Method B). 2,2′-Disulfanediylbis(N-(4-(triﬂuoromethyl)-

phenyl)benzamide) (593 mg, 1 mmol) and NaBH₄(378 mg, 10

mmol) were placed in an oven-dried ﬂask and evacuated and reﬁlled

with N₂three times. Methanol (15 mL) was added slowly to prevent

vigorous bubbling, and the resulting reaction mixture was stirred for 1

h at ambient temperature. Methanol was then removed in vacuo, and

the resulting solid was redissolved in ethyl acetate (20 mL). This

solution was stirred with HCl (aqueous; 7%, 20 mL) for 1 h. The

compound of interest was extracted from this mixture into ethyl

acetate (2 × 10 mL). Combined organic layers were dried over

MgSO₄and solvent was removed in vacuo. The crude solid was

puriﬁed by silica column chromatography (100% CH₂Cl₂) to aﬀord 2-

mercapto-N-(4-(triﬂuoromethyl)phenyl)benzamide as a white solid.

¹H NMR (400 MHz, CDCl₃) δ 7.88 (s, 1H), 7.76 (d, J = 8.4 Hz, 2H),

1

(shoulder), 408, 515, 578, 720 nm. H NMR (CD₂Cl₂): 71.5 (br,

β-pyrrole H), 62.5 and 55.6 (s, meta-SPh’s), 12.8 (br-s, meta-Ph TPP),

9.6 (br-s, ortho-Ph TPP), 6.8 (br-s, para-Ph TPP), −84.5 (s, para-

SPh), −91.9 (br, ortho-SPh) ppm. Other SPh-NHPh-pCl signals

observed: 8.3 (br), 4.8 (br), and −111.6 (br) ppm. Anal. Calcd for

C₅₇H₃₇ClFeN₅OS: C, 73.51; H, 4.00; N, 7.52; S, 3.44. Found: C,

72.21; H, 4.07; N 7.42, S, 3.60. Yield: 32.8 mg, 41.7% (starting

[Fe(TPP)(OCH₃)] mass was 59.1 mg). EPR (1:1 CH₂Cl₂/toluene):

g_x= 1.95, g_y= 1.88, g_z= 2.06; E/D = 0.034.

2.3.2.4. [Fe(TPP)(S-CF₃)]. Electronic absorption (CH₂Cl₂): 380

1

(shoulder), 408, 515, 578, 721 nm. H NMR (CD₂Cl₂): 68.4 (br, β-

pyrrole H), 61.2 and 54.8 (s, meta-SPh’s), 12.8 (br-s, meta-Ph TPP),

9.9 (br-s, ortho-Ph TPP), 6.9 (br-s, para-Ph TPP), −83.2 (s, para-

SPh), −90.4 (br, ortho-SPh) ppm. Other SPh-NHPh-pCF₃signals

observed: 8.0 (br), 5.0 (br) and −110.5 (br) ppm. ¹⁹F-NMR

(CD₂Cl₂): −60.3 ppm. Anal. Calcd for C₅₈H₃₇F₃FeN₅OS: C, 72.20;

H, 3.87; N, 7.26. Found: C, 71.66; H, 3.86; N, 7.26. Yield: 34.8 mg,

46.7% (starting [Fe(TPP)(OCH₃)] mass was 54.0 mg). EPR (1:1

CH₂Cl₂/toluene): g_x= 1.93, g_y= 1.87, g_z= 2.06; E/D = 0.037.

2.3.2.5. [Fe(TPP)(S-NO₂)]. Electronic absorption (CH₂Cl₂): 336

(shoulder), 380 (shoulder), 408, 516, 578, 722 nm. ¹H NMR

(CD₂Cl₂): 70.5 (br, β-pyrrole H), 69.9 and 54.3 (s, meta-SPh’s), 12.9

(br-s, meta-Ph TPP), 10.5 (br-s, ortho-Ph TPP), 6.9 (br-s, para-Ph

TPP), −83.7 (s, para-SPh), −90.2 (br, ortho-SPh) ppm. Other SPh-

NHPh-pNO₂signals observed: 8.2 (br), 4.9 (br) and −112.4 (br)

ppm. Anal. Calcd for C₅₇H₃₇FeN₆O₃S: C, 72.69; H, 3.96; N, 8.92.

Found: C, 72.35; H, 3.75; N, 8.71. Yield: 87.5 mg, 64.0% (starting

[Fe(TPP)(OCH₃)] mass was 101.5 mg). EPR (1:1 CH₂Cl₂/toluene):

g_x= 1.94, g_y= 1.95, g_z= 2.00, E/D = 0.048. Crystal data for

C₅₇H₃₇FeN₆O₃S (M = 941.83 g/mol): monoclinic, space group P2₁/c

(no. 14), a = 14.124(4) Å, b = 12.805(4) Å, c = 24.419(8) Å, β =

90.916(13)°, V = 4416(2) Å³, Z = 4, T = 100.0 K, μ(Mo Kα) = 0.445

mm⁻¹, Dcalc = 1.417 g/cm³, 96775 reﬂections measured (2.884° ≤ 2θ

≤ 53.52°), 9382 unique (R_int= 0.0684, R_sigma= 0.0337) which were

used in all calculations. The ﬁnal R₁was 0.0360 (I > 2σ(I)) and wR₂

was 0.0887 (all data). Detailed methods for X-ray crystallography and

additional crystallographic data can be found in the Supporting

Information.

7.63 (dd, J = 9.0, 7.4 Hz, 3H), 7.41 (dd, J = 7.8, 1.4 Hz, 1H), 7.35 (td,

J = 7.6, 1.5 Hz, 1H), 7.28−7.23 (m, 2H), 4.44 (s, 1H). ¹³C NMR

(126 MHz, chloroform-d) δ 166.79, 140.82, 133.22, 133.10, 131.80,

131.56, 128.25, 126.72 (q, J = 32.9 Hz), 126.52, 125.80, 124.16 (q, J =

271.7 Hz), 119.93. ¹⁹F NMR (471 MHz, chloroform-d) δ −62.19.

HRMS (+ESI) m/z: [M + H]⁺Calculated for C₁₄H₁₁F₃NOS:

298.0508. Found: 298.0504. Yield: 164 mg (0.55 mmol), 55%.

2.3.1.9. 2-Mercapto-N-(4-nitrophenyl)benzamide (HS-NO₂). 2,2′-

Disulfanediylbis(N-(4-nitrophenyl)benzamide) (547 mg, 1 mmol)

was used as a starting material following Method B above to aﬀord 2-

mercapto-N-(4-nitrophenyl)benzamide as a pale yellow microcrystal-

1

line solid. H NMR (400 MHz, CDCl₃) δ 8.28−8.22 (m, 2H), 8.18

(s, 1H), 7.85−7.80 (m, 2H), 7.63 (dd, J = 7.7, 1.5 Hz, 1H), 7.41 (dd,

J = 7.9, 1.4 Hz, 1H), 7.36 (td, J = 7.6, 1.5 Hz, 1H), 7.28−7.22 (m,

1H), 4.35 (s, 1H). ¹³C NMR (126 MHz, chloroform-d) δ 166.63,

144.09, 143.55, 133.37, 132.71, 132.08, 131.91, 128.36, 125.99,

125.33, 119.64. HRMS (−ESI) m/z: [M − H]⁻Calculated for

D

DOI: 10.1021/acs.inorgchem.9b02506

Inorg. Chem. XXXX, XXX, XXX−XXX

Inorganic Chemistry

Article

Table 2. Experimental and DFT-Computed Metal-Ligand Bond Distances for Low-Spin, Fe(III) Hemoprotein Models

experimental

computed

a

protein

L₁

L₂

d_Fe−Npor(Å)

d_Fe−L1(Å)

d_Fe−L2(Å)

d_Fe−Npor(Å)

d_Fe−L1(Å)

d_Fe−L2(Å)

cytochrome b₅

cytochrome c

hCBS/PxRcoM

His(N)

Met(S)

Cys(S⁻)

1.988

2.005

2.010

2.075

2.037

2.210

2.003

2.351

2.281

2.008

2.002

2.011

1.997

2.009

2.104

1.996

2.338

2.186

a

Average Fe−N bond distance for pyrrole N atoms.

Figure 2. Electronic structure analysis of metal−ligand bonding in computational models of low-spin, Fe(III) porphyrins with diﬀerent axial

ligands. Left: The MO diagram depicts the relative energies of Fe 3d-based MOs (spin up, single-electron orbitals). Orbital energies are relative to

2

the energy of the 3d_{x −y}orbital for each model. Two ligand ﬁeld parameters, rhombic (V) and axial (Δ) splitting, are deﬁned in terms of the relative

energies of Fe 3d_xy-, 3d_xz-, and 3d_yz-based MOs. Right: Geometry-optimized structures of each porphyrin model shown with an overlay of the Fe

3d_yz-based, spin-down MO. Orbital plots were generated with an isosurface value of 0.05, and structures were visualized using Pymol (v1.3).

2.3.3. Preparation of Six-Coordinate, Fe(III) Thiolate-Ligated

Porphyrin Complexes for EPR Spectroscopy. Six-coordinate (6-c)

low-spin, Fe(III) porphyrin-thiolate complexes were prepared with

either 1-methylimidazole (1-MeIm) or water as a second axial ligand.

All complexes were prepared in EPR tubes in a glovebox. Porphyrin-

thiolate complexes bound to 1-MeIm were prepared by mixing a

solution of [Fe(TPP)(S-R)] (500 μM in 50:50 CH₂Cl₂/toluene) with

3 molar equiv of 1-MeIm. Porphyrin-thiolate complexes axially ligated

by a water molecule were prepared by dissolving each 5-c compound

in solvent (50:50 CH₂Cl₂/toluene) that had been dried only by

elution through alumina and without further drying. The concen-

tration of water in the solvent (≥2 mM, as determined by

potentiometric titration) was such that it was in excess of the 5-c

porphyrin species in each of these samples (500 μM).

the energy diﬀerence between Fe 3d_xzand 3d_yzorbitals (i.e., E_xz= V/2

= −E_yz). The axial distortion (Δ/ξ), scaled to ξ, approximates the

energy diﬀerence between Fe 3d_xyand the barycenter of the 3d_xzand

3d_yzorbitals (i.e., E_xy= Δ).

All ligand ﬁeld parameters were computed using the program

DLSD5, originally written by McGarvey,⁴²as modiﬁed by Telser et

al.⁴⁴which systematically assigns the sign and magnitude of each g-

value assuming maximal distortion along the z-axis (i.e., |Δ| ≥ 2|V|/3).

For most low-spin, Fe(III) porphyrin complexes, the axis of maximal

distortion lies normal to the heme plane and thus aligns with the

molecular z-axis.³⁹In this case, computed values for V/ξ and Δ/ξ

were used directly. However, for thiolate-ligated porphyrins, the axis

of maximal distortion lies within the heme plane, leading to an

interchange of the x- and z-axes.⁴⁵To provide meaningful information

about electronic structure in terms of the molecular axes (in which z is

normal to the heme plane), computed values for V/ξ and Δ/ξ in

thiolate-ligated porphyrins were converted using eqs 1 and 2.

2.4. Relationship between Experimentally Determined g-

Values and Ligand Field Parameters. Ligand ﬁeld parameters

were estimated from experimentally determined g-values for low-spin,

Fe(III) porphyrin models using the tetragonally distorted, strong ﬁeld

d⁵model described by McGarvey.⁴²In the strong ﬁeld approximation,

the wave function for the ground state Kramers doublet (²T_2gin O_h

symmetry) is modeled in terms of three low-lying, metal-based d-

(1)

V′ = 0.5V

− Δ

(2)

Δ′ = −0.5 − 0.75V

2

orbitals, d_xy, d_xz, and d_yz. The two other metal-based d-orbitals, d_zand

2.5. Computational Methods. 2.5.1. Density Functional

Theory Calculations. For 6-c, low-spin Fe(III) porphyrin models,

geometry optimizations, single-point calculations, and time-dependent

density functional theory (TD-DFT) computations were carried out

using ORCA v4.0,⁴⁶and frequency calculations were carried out using

Gaussian09.⁴⁷Geometry optimizations utilized Becke’s functional for

exchange along with Perdew’s functional for correlation (BP86),^48,49

and all atoms were described using Alrich’s polarized triple ζ-valence

(TZVP) basis set along with the def2/J auxiliary basis.^50,51Frequency

2

d_{x −y}, are much higher in energy and therefore are not considered in

this model. Wave function coeﬃcients for each orbital (A, B, and C

for d_xy, d_yz, and d_xz, respectively), as well as the orbital reduction

factor k, are computed directly from experimentally determined values

of g_x, g_y, and g_z. Two additional ligand ﬁeld parameters, rhombic and

axial distortions, are then computed from these wave function

coeﬃcients. The rhombic distortion (V/ξ), scaled to the single

electron spin−orbit constant for Fe³⁺(ξ = 464 cm⁻¹),⁴³approximates

E

DOI: 10.1021/acs.inorgchem.9b02506

Inorg. Chem. XXXX, XXX, XXX−XXX

Inorganic Chemistry

Article

a

Table 3. Computational Analysis of Low-Spin, Fe(III) Hemoprotein Models

spin population

protein

axial heme ligands

ΔE₁(cm⁻¹

)

ΔE₂(cm⁻¹

)

Fe 3d_yz

Fe 3d_xz

Fe 3d_xy

S 3p

N/A

cytochrome b₅

cytochrome c

hCBS

His/His

His/Met

His/Cys(S⁻)

3785

2072

5865

4118

2571

6379

0.4666

0.6091

0.7787

0.445

0.2863

0.01453

0.0316

0.04704

0.03369

0.1565

a

̈

Values for ΔE₁and ΔE₂were computed using TD-DFT, and Loewdin spin populations were determined from single-point calculations.

calculations utilized Becke’s three-parameter hybrid functional for

exchange and the Lee−Yang−Parr functional for correlation

(B3LYP), with the 6-31G(d) basis set.⁵²Spin-unrestricted single-

point calculations were performed using the B3LYP functional along

with the polarized split valence potential (SVP) basis set and def2/J

auxiliary basis for all atoms except Fe and S, which were modeled

using the TZVPP basis set, and Fe-coordinated N atoms, which were

modeled using the TZVP basis set.^51,53Energy diﬀerences between

ground and excited states were calculated using TD-DFT with the

camB3LYP functional along with the polarized SVP basis set and

def2/J auxiliary basis for all atoms except Fe and S, which were

modeled using the TZVPP basis set, and Fe-coordinated N atoms,

which were modeled using the TZVP basis set.

Initial coordinates for archetype hemoprotein models were

obtained from protein crystal structures for cytochrome b₅(PBD

1CYO),⁵⁴cytochrome c (PBD 1CCR),⁵⁵and human cystathionine β-

synthase (PBD 1JBQ).⁵⁶Histidine ligands were truncated at the β-

methylene C atom, which was replaced with a methyl group, the

methionine ligand was truncated at the γ-methylene C atom and

replaced with a methyl group, and the cysteine ligand was truncated at

the β-methylene C atom, which was replaced with a methyl group.

Porphine (P²⁻) was used to model all porphyrins by replacing

porphyrin substituents with H atoms. Initial coordinates for synthetic

models were derived from the crystal structure of [Fe(TPP)(S-

NO₂)]. To generate starting coordinates for each 6-c compound, the

N-phenyl para-substituent was modeled using idealized geometries,

and a second axial ligand was added based on an appropriate crystal

structure of a six-coordinate heme or porphyrin model complex.⁵⁶⁻⁵⁸

To ensure proper positioning of the Fe atom in each model, the Fe

atom was shifted to achieve coplanarity with the coordinating

porphyrin N atoms in starting coordinates. All structural modiﬁca-

tions were performed using the molecular builder in WebMO

(v19.0.009e).⁵⁹Geometry-optimized coordinates are compiled in

Tables S9 −S22.

lowest energy Fe 3d-based MO was primarily d_xyin character.

For the His/Met model, the lowest energy orbital was also

predominantly Fe 3d_xyin character; however, there was also a

signiﬁcant contribution from a sulfur 3p orbital. This ﬁlled−

ﬁlled π-type interaction results in no net Fe−S bonding

character. The next lowest energy orbitals observed are the

nearly degenerate Fe 3d_xz- and 3d_yz-based MOs, which both

exhibit π-antibonding character with respect to the porphyrin

p_zorbitals. The second-highest energy Fe 3d-based MO is

2

largely d_zin character. This MO is primarily σ-antibonding

with respect to the two axial heme ligands, is slightly σ-

antibonding with respect to the porphyrin N atoms, and is

slightly higher in energy in the His/His model than in the His/

Met model. The highest energy Fe 3d-based MO in both

2

models is largely d_{x −y}in character and strongly σ-antibonding

with respect to the porphyrin N atoms.

In the thiolate-bound His/Cys(S⁻) model, the orbital

contributions to axial iron-ligand bonding diﬀer from those

of the His/His and His/Met models. Speciﬁcally, one of the

thiolate S 3p orbitals has the appropriate symmetry to engage

in a p_π−d_πbonding interaction with the Fe 3d_yzorbital (Figure

S5). This bonding interaction raises the energy of the Fe 3d_yz-

based MO and lifts the near-degeneracy of d_xzand d_yz(Figure

2). Additionally, the energy of Fe 3d_xz-based MO is lowered

slightly, relative to those of the His/His and His/Met models.

Axial thiolate ligation to low-spin, Fe(III) heme increases the

diﬀerences in energy between the ground state and two low-

lying excited states. Using TD-DFT, we computed the

diﬀerences in energy between the ground state and ﬁrst two

excited states in low-spin, Fe(III) hemoprotein models.

Consistent with greater energy separation between Fe 3d_xz-

and 3d_yz-based MOs, the energy of the (d_xy)²(d_xz)²(d_yz)¹→

(d_xy)²(d_xz)¹(d_yz)²transition (ΔE₁) increased substantially for

the His/Cys(S⁻) model (ΔE₁= 5865 cm⁻¹) relative to His/

His (ΔE₁= 3785 cm⁻¹) and His/Met (ΔE₁= 2072 cm⁻¹)

models (Table 3). The energy of the (d_xy)²(d_xz)²(d_yz)¹→

(d_xy)¹(d_xz)²(d_yz)²transition (ΔE₂) also increased for the heme-

thiolate model, consistent with a stronger overall axial ligand

ﬁeld.

3. RESULTS

3.1. Electronic Origins of the Unique Rhombic EPR

Signal Associated with Low-Spin, Fe(III) Heme-Thiolate

Proteins. We utilized density functional theory (DFT)

computations to analyze metal−ligand bonding in three low-

spin, Fe(III) hemoprotein models with His/His (cytochrome

b₅), His/Met (cytochrome c), and His/Cys(S⁻) (hCBS,

PxRcoM) axial ligands. In all cases, the computed structures

were consistent with those observed experimentally (Table 2).

Axial metal−ligand bond distances were slightly (up to 0.1 Å)

shorter than those observed experimentally, which is

unsurprising given that the computational models do not

account for secondary coordination sphere eﬀects of the

protein heme pocket. Using geometry-optimized model

structures, we computed single-point energies and ground-

state wave functions using DFT. We compared metal−ligand

bonding in each of these hemoprotein models by examining

the energies and orbital compositions of the predominantly Fe

3d-based frontier molecular orbitals (MOs) (Figure 2).

Our computational results establish a connection between

experimentally observed g-values and heme-thiolate electronic

structure. In EPR spectroscopy, the shift from g_e∼ 2.00 arises

from spin−orbit coupling of low-lying excited states into the

paramagnetic ground state. The extent of this spin−orbit

coupling cannot be directly calculated using DFT; however,

spin−orbit coupling is inversely proportional to two

parameters that can be estimated using computations: (1)

metal−ligand covalency and (2) ground-state/excited-state

energy diﬀerences. Our single-point DFT computations predict

that thiolate ligation increases metal−ligand covalency, as

indicated by the signiﬁcant contribution of the sulfur 3p orbital

to the unpaired spin population in the His/Cys(S⁻) model

(Table 3). This increased metal−ligand covalency likely

diminishes spin−orbit coupling and contributes to the small

The orbital contributions to iron-ligand bonding were

similar to one another in the His/His and His/Met models

(Figures S3 and S4). In the case of the His/His model, the

F

DOI: 10.1021/acs.inorgchem.9b02506

Inorg. Chem. XXXX, XXX, XXX−XXX

Inorganic Chemistry

Article

Table 4. Ligand Field Parameters Computed from Experimental g-Values for a Number of Hemoproteins with Varying Axial

Ligands

a

b

c

d

e

f

protein

cytochrome b₅

cytochrome c

hCBS

PxRcoM-1

RrCooA

axial heme ligands

g_x

g_y

g_z

V/ξ

|V| (cm⁻¹

)

Δ/ ξ

|Δ| (cm⁻¹

)

k

V/Δ

His/His

His/Met

1.40

1.35

−2.21

−2.14

−2.26

−2.31

−2.25

−2.26

−2.27

−3.03

−3.08

1.88

1.87

1.89

−1.63

−1.49

−4.32

−4.39

−4.51

−5.98

−4.06

750

685

−3.21

−3.39

−5.03

−4.26

−5.24

−6.01

−5.46

1477

1559

2314

1960

2410

2765

2512

1.01

0.99

1.05

1.10

1.05

1.22

1.15

0.51

0.44

0.86

1.03

0.86

0.99

0.74

Cys(S⁻)/His

Cys(S⁻)/Pro

Cys(S⁻)/H₂O

Cys(S⁻)/Im

−2.48

−2.49

−2.46

−2.41

−2.56

1987

2019

2075

2751

1868

cyt P450_cam

cyt P450_cam+ imidazole (Im)

1.93

1.87

a

b

c

V/ξ = rhombic splitting, V′/ξ for thiolate-ligated species, ξ = 460 cm⁻¹|V′| for thiolate-ligated species Δ/ξ= axial splitting, Δ′/ξ for thiolate-

d

e

f

ligated species |Δ′| for thiolate-ligated species k = orbital reduction factor V/Δ = rhombicity, V′/Δ′ for thiolate-ligated species

g-shifts observed in EPR spectra of heme-thiolate proteins. We

qualitatively estimated ground-state/excited-state energy diﬀer-

ences in low-spin, Fe(III) porphyrins by comparing the DFT-

calculated ground state energies of Fe 3d-based MOs:

Thiolate-ligated porphyrins exhibit large diﬀerences in energy

between d_xy-, d_xz-, and d_yz-based MOs and therefore experience

diminished spin−orbit coupling. TD-DFT computations are

consistent with this trend: Thiolate ligation gives rise to an

increase in ground state-excited state energy diﬀerences (ΔE₁

and ΔE₂) and thereby attenuates spin−orbit coupling in low-

spin, Fe(III) heme centers bound by an axial Cys(S⁻) ligand.

TD-DFT calculations, which are unable to fully account for the

correlation energy contributions associated with these

electronic transitions, provide limited quantitative information;

however, we argue that ΔE₁and ΔE₂values provide a

comparative measure of spin−orbit coupling, since these values

agree qualitatively with single-point calculations.

accounting for the increased axial splitting observed in thiolate-

ligated porphyrins. Our single-point DFT computations

predicted that Fe−S p_π−d_πbonding destabilizes the Fe 3d_yz-

based MO relative to the Fe 3d_xz-based MO (Figure 2), an

observation that is consistent with a thiolate-dependent

increase in rhombic splitting. Furthermore, TD-DFT compu-

tations qualitatively recapitulate the thiolate-dependent in-

crease in ground-state/excited-state energy diﬀerences that are

observed experimentally, although computed ΔE₁and ΔE₂

values overestimate axial and rhombic splitting parameters by

several thousand wavenumbers (Table 3).

The g-shifts observed in EPR spectra of low-spin, Fe(III)

porphyrins, as well as the resulting ligand ﬁeld parameters, are

sensitive to the identity of the second axial ligand trans to

thiolate. Speciﬁcally, weaker axial ligands give rise to larger

values for |V′| and |Δ′| (e.g., |V′| = 2751 cm⁻¹and |Δ′| = 2765

cm⁻¹for cyt P450_camwith water trans to thiolate), while

stronger axial ligands give rise to smaller values for |V′| and |Δ′|

(e.g., |V′| = 1687 cm⁻¹and |Δ′| = 2314 cm⁻¹for hCBS with

His-based imidazole ring trans to thiolate, Table 4). This trend

can be explained in terms of a trans inﬂuence; a weaker trans

ligand allows for a stronger Fe−S bond and greater thiolate

character, giving rise to larger values for |V′| and |Δ′|, as well as

smaller g-shifts in the rhombic EPR signal. Interestingly, even

heme-thiolate proteins bearing identical second axial ligands

exhibit diﬀerent ligand ﬁeld parameters. For example, hCBS,

PxRcoM-1, and cyt P450_cam+ imidazole exhibit axial Cys(S⁻)/

imidazole ligation; however, rhombic and axial splittings diﬀer

between these proteins. It is likely that protein-dependent

variations to the second coordination sphere (including

changes to the thiolate H-bonding environment) give rise to

these variations in porphyrin electronic structure.

Our computational data predict that protonation of the

coordinating Cys(S⁻) in a low-spin, Fe(III) heme leads to

diminished spin−orbit coupling and should give rise to a

rhombic EPR signal with larger g-shifts. We modeled this case

by protonating the previously optimized His/Cys(S⁻) model

and reoptimizing the geometry using DFT. We then compared

ground-state bonding and electronic transitions between the

His/Cys(S⁻) and His/Cys(SH) models in a manner analogous

to that described above for other hemoprotein models.

Protonation of the coordinating S atom weakens the donor

strength of the coordinating cysteine in two ways. First,

protonation neutralizes the negative charge and eliminates a

strong Coulombic attraction between the metal and ligand

(d_Fe−S= 2.330 Å for His/Cys(SH) model; d_Fe−S= 2.186 Å for

His/Cys(S⁻) model). Second, as was the case for the His/Met

model, there is no Fe−S π-bonding interaction in the His/

Cys(SH) model, and the Fe 3d_yz-based MO is stabilized upon

Using a tetragonally distorted, strong ﬁeld d⁵model, we

estimated ligand ﬁeld parameters for a number of hemopro-

teins with varying axial ligands. This model, introduced by

Griﬃth in the late 1950s and reﬁned through the 1990s,

approximates energy diﬀerences between the three low-lying,

metal-based d_xy, d_xz, and d_yzorbitals from experimentally

determined g-values.^{36,39,42,45,60,61}In low-spin, Fe(III) por-

phyrins, these orbital energies primarily correspond to Fe 3d-

based MOs. The rhombic distortion, V/ξ, approximates the

energy diﬀerence between d_xzand d_yz, while the axial

distortion, Δ/ξ, approximates the energy diﬀerence between

d_xyand the barycenter of the d_xzand d_yz(Figure 2). Both

rhombic and axial distortions are scaled by the spin−orbit

coupling parameter, ξ, which has a value of 464 cm⁻¹for

Fe³⁺.⁴³Table 4 summarizes g-value assignments and ligand

ﬁeld parameters for hemoproteins bearing a variety of axial

ligands, computed using the program developed by McGarvey

and Telser.^42,44By McGarvey’s convention, the tetragonal

(axial) distortion stabilizes d_xyrelative to d_xz/d_yzwhen Δ/ξ <

0, and the rhombic distortion stabilizes d_xzrelative to d_yzwhen

V/ξ < 0.

Ligand ﬁeld parameters, estimated from experimental g-

values, validate the computational prediction that thiolate

ligation increases the orbital energy separation between Fe

3d_xy-, 3d_xz-, and 3d_yz-based MOs. Magnitudes for rhombic (V,

V′ in thiolate-ligated porphyrins) and axial (Δ, Δ′ in thiolate-

ligated porphyrins) splittings increased signiﬁcantly for

hemoproteins bearing an axial Cys(S⁻) ligand (|V′| = 1987−

2751 cm⁻¹and |Δ′| = 1960−2765 cm⁻¹) compared to those

without a thiolate ligand (|V| = 685−750 cm⁻¹and |Δ| =

1477−1559 cm⁻¹, Table 4). Cys(S⁻) is a charged, strong ﬁeld

ligand that destabilizes d_xzand d_yzrelative to d_xy, likely

G

DOI: 10.1021/acs.inorgchem.9b02506

Inorg. Chem. XXXX, XXX, XXX−XXX

Inorganic Chemistry

Article

Figure 3. Eﬀects of thiolate protonation on porphyrin electronic structure. Left: Comparison of energy levels between the ground state and two

lowest-lying excited states (with corresponding electron conﬁgurations shown) in low-spin, Fe(III), d⁵porphyrin models, computed using TD-

DFT. Right: Geometry-optimized structures of His/Cys(S⁻) and His/Cys(SH) models with an overlay of the Fe 3d_yz-based, spin-down MO.

Orbital plots were generated with an isosurface value of 0.05, and structures were visualized using Pymol (v1.3).

a

Scheme 1. Synthesis of 2-Mercapto-N-phenylbenzamides Derivatives from 2,2′-Disulfanediyldibenzoyl Chloride

a

R = CH₃, H, Cl, CF₃, and NO₂. Disulﬁdes are abbreviated as “SS-R” and aryl thiols are abbreviated as “HS-R”.

protonation (Figure S6). Consequently, values for ΔE₁and

ΔE₂are reduced by ∼2000 cm⁻¹(Figure 3). A stabilized Fe

3d_yz-based MO and smaller values for ΔE₁and ΔE₂reﬂect

greater spin−orbit coupling, which should give rise to larger

magnitude g-shifts. H-bonding represents an intermediate case

between a fully deprotonated thiolate and a protonated thiol;

therefore, we hypothesize that the dispersion of g-values in the

rhombic, low-spin EPR signal observed for a six-coordinate,

Fe(III) heme-thiolate protein should reﬂect the strength of the

thiolate H-bonding network.

3.2. Synthesis of Intramolecularly Hydrogen-Bonded

Aryl Thiolate Ligands. To test our hypothesis that EPR

spectroscopy may be sensitive to changes in the H-bonding

environment in heme-thiolate proteins, we utilized a series of

aryl thiolate ligands in which the strength of an intramolecular

hydrogen bond (H-bond) to the Fe-coordinating thiolate may

be systematically tuned. Our chosen ligand set takes inspiration

from the substituted 2-acylaminobenzenethiolate ligands

reported by Ueyama et al.,⁶²whose syntheses we have been

unable to successfully reproduce. The newly utilized ligand

series contains a 2-mercaptobenzamide core, in which an

intramolecular N−H···S H-bond exists between the amide and

coordinating aryl thiolate S atom. This intramolecular H-bond

results in the formation of a six-membered ring, an improve-

ment over the originally employed ligand in which a ﬁve-

membered ring was formed. Expansion of the ring by one atom

allows the D−H···A angle to further approach 180°, the ideal

angle for this H-bonding interaction. To systematically tune

the strength of the intramolecular H-bond, we varied the

electronics of the amide by changing the identity of an N-

phenyl para substituent. We chose this para substitution

strategy for three reasons. First, the eﬀects of para substituent

electron-donating/-withdrawing properties on the pK_aof a

heteroatom proton are well-characterized and concisely

encompassed in the Hammett parameter, σ.⁶³We anticipate

para substituents with a more positive σ value will lower the

pK_aof the amide proton and result in a stronger H-bond and

vice versa. Second, by making substitutions at the amide

phenyl ring, we avoid substantially inﬂuencing the thiolate

donor strength directly, and instead isolate the eﬀects of

changes in H-bond strength. Finally, synthesis of these

compounds relies on a facile two-step syntheses from an acyl

chloride and commercially available, para-substituted aniline

derivatives.^64,65

We successfully synthesized ﬁve 2-mercapto-N-phenyl-

benzamide derivatives (HS-R, R = CH₃, H, Cl, CF₃, NO₂),

which serve as intramolecularly H-bonded thiolate ligand

precursors, using a two-step synthetic method (Scheme 1). We

H

DOI: 10.1021/acs.inorgchem.9b02506

Inorg. Chem. XXXX, XXX, XXX−XXX

Inorganic Chemistry

Article

a

Table 5. Comparison of the NMR Chemical Shifts (ppm) of the 5-c Aryl Thiolate Porphyrin Complexes, [Fe(TPP)(S-R)]

R

β-pyrrole

m-Ph, TPP

o-Ph, TPP

p-Ph, TPP

m-Ph, SPh’s

o-Ph, SPh

p-Ph, SPh

o/m-PhL−H

p-PhL−H

CH₃

H

Cl

CF₃

NO₂

72.5

71.6

71.5

68.4

70.5

12.8

12.9

9.6

9.8

9.6

9.9

10.5

6.9

6.8

6.9

6.7

64.3, 56.6

63.7, 56.4

62.5, 55.6

61.2, 54.8

59.9, 54.3

−93.0

−92.8

−91.9

−90.4

−90.2

−85.6

−85.3

−84.5

−83.2

−82.7

−110.7, 8.4, 4.8

−111.3, 8.3, 4.9

−111.6, 8.3, 4.8

−110.5, 8.0, 5.1

−112.4, 8.2, 4.9

4.1

6.5

a

R refers to the identity of the N-phenyl para substituent, recorded in CD₂Cl₂at room temperature (between 20 and 22 °C).

reacted 2,2′-disulfanediyldibenzoyl chloride with ﬁve diﬀerent

commercially available, para-substituted anilines under basic

conditions to generate 2,2′-disulfanediylbis(N-phenylbenza-

mide) derivatives with varying N-phenyl para substitution (SS-

R, R = CH₃, H, Cl, CF₃, NO₂, Figures S7 −S17). Strategically,

we chose to carry out the amide formation reaction using these

disulﬁdes in order to eliminate side-reactions between the thiol

moiety and acyl chloride intermediate. Reduction of the 2,2′-

disulfanediylbis(N-phenylbenzamide) disulﬁde bond using

either triphenylphosphine or sodium borohydride, followed

by protonation of the charged thiolate, resulted in formation of

Scheme 2. Labeling Scheme for the Observed Signals in the

¹H NMR Spectra Reported in Table 5

1

our desired thiol ligand precursors (Figures S18 −S28). In H

NMR spectra of the thiols, we observed a direct correlation

between the chemical shift observed for the amide proton and

electron-withdrawing strength of the N-phenyl para substitu-

ent, as represented by the two extremes: δ_N−H(HS-CH₃) =

7.65 ppm and δ_N−H(HS-NO₂) = 8.18 ppm. No such

correlation was observed for protons in the aryl thiol ring.

Taken together, these observations serve as preliminary

evidence of a localized, systematic change in the electronic

properties of the amide proton. The nature of this systematic

change in electronics is enumerated below.

3.3. Preparation of Five-Coordinate, Fe(III) Thiolate-

Ligated Porphyrins. We prepared a series of ﬁve complexes

of the form [Fe(TPP)(S-R)] using the intramolecularly H-

bonded thiol ligand precursors, HS-R (R = CH₃, H, Cl, CF₃,

NO₂). We assembled each complex using the procedure

recently reported by Hunt and Lehnert for the synthesis of 5-c,

Fe(III) tetraphenylporphyrinate complexes bearing an axial

aryl thiolate ligand.¹⁰This method involves mixing [Fe(TPP)-

(OCH₃)] with a slight excess of thiol ligand precursor in dry

CH₂Cl₂. The strongly basic methoxide ligand deprotonates the

aryl thiol, resulting in coordination of the aryl thiolate and

formation of 1 equiv of methanol. We used electronic

absorption spectroscopy to monitor the reaction, which

typically progresses to completion within 1 min. Characteristic

spectroscopic features of each [Fe(TPP)(S-R)] complex

include a split Soret peak with a shoulder at 380 nm and

maximum at 407−408 nm, in addition to other peaks at 515,

576−578, and 717−722 nm (Figures S30, S33, S36, S39, and

S43).

complexes bound to aryl thiolate ligands.¹⁰Importantly,

peaks for β-pyrrole protons were observed between 68.4 and

72.5 ppm, a range that is characteristic of paramagnetic shifts

due to the presence of a high-spin, Fe(III) porphyrin species

(Figures S32, S35, S38, S41, and S45). We observed a

correlation between the identity of the N-phenyl para

substituent and the magnitude of the chemical shift observed

for protons in the ortho, meta, and para positions of the

coordinating aryl-thiolate: as the electron-withdrawing charac-

ter of the N-phenyl para substituent increases, giving rise to a

stronger intramolecular H-bond, the magnitude of the

paramagnetic shift decreases. Considering that a stronger H-

bond gives rise to a weaker Fe−S bond (vide infra),

attenuation of the interaction between the paramagnetic,

high-spin (S = 5/2) Fe(III) center and the ligand-derived

protons may cause the observed trends in paramagnetic shifts;

however, in-depth analysis of the observed paramagnetic shifts

would be required to fully substantiate this hypothesis.

The crystal structure of [Fe(TPP)(S-NO₂)] revealed an

intramolecular H-bonding interaction between the ligand

amide and Fe-coordinated thiolate (Figure 4). Geometric

parameters for [Fe(TPP)(S-NO₂)] (Fe−S bond distance, Fe−

S = 2.3078(8) Å, average Fe−N bond distance, Fe−N_av

=

2.064(5) Å, and Fe atom displacement from the mean plane of

the coordinating N atoms, Fe atom displacement = 0.4181(8)

Å) are comparable to those reported for [Fe(TPP)(SPh)] by

Byrn and Strouse (Fe−S = 2.315(2) Å, Fe−N_av= 2.063 Å, Fe

atom displacement = 0.470(1) Å)⁶⁶and are therefore

consistent with aryl-thiolate coordination. The conformation

of the amide relative to the coordinating S atom is consistent

with the presence of an intramolecular interaction in which the

amide acts as an H-bond donor to the coordinating S atom. A

donor−acceptor distance of 3.0087(19) Å and a D−H···A

angle of 149(2)° are indicative of a weak H-bonding

interaction. The contribution of this intramolecular H-bond

to the attenuation of thiolate ligand strength is supported by

spectroscopic and computational results detailed below.

Each recrystallized [Fe(TPP)(S-R)] compound was charac-

1

terized using EPR and H NMR spectroscopies and exhibited

distinct features indicative of a high-spin, Fe(III) porphyrin

complex. Eﬀective g-values were observed at g = 7.1, 4.8, and

1.9 in the EPR spectrum of each compound, suggesting the

presence of a high-spin (S = 5/2), rhombic paramagnetic

species (Figures S31, S34, S37, S40, and S44). An additional

sharp signal around g = 6.0 was observed in all EPR spectra,

derived from a very small amount of [Fe(TPP)Cl] impurity

1

(Figure S29). Assignment of H NMR chemical shifts are

summarized in Table 5 and Scheme 2. These chemical shift

assignments draw on those made for other [Fe(TPP)]⁺

I

DOI: 10.1021/acs.inorgchem.9b02506

Inorg. Chem. XXXX, XXX, XXX−XXX

Inorganic Chemistry

Article

Figure 4. Molecular structure of [Fe(TPP)(S-NO2)] shown with 50% probability ellipsoids and selected atom IDs. The amide H atom is

displayed, and all other H atoms are omitted for clarity. The yellow dots denote the intramolecular H-bond between the amide proton and

coordinating S atom.

3.4. Assembly and EPR Spectroscopic Character-

ization of Six-Coordinate, Fe(III) Thiolate-Ligated

Porphyrins. By mixing each of the ﬁve-coordinate [Fe(TPP)-

(S-R)] complexes with a neutral donor ligand, we successfully

assembled six-coordinate model complexes that exhibit EPR

spectra characteristic of low-spin, Fe(III) thiolate-ligated

porphyrins. We utilized two biologically relevant neutral

donor ligands to mimic heme-thiolate protein environments:

1-methylimidazole (1-MeIm), which gives rise to axial ligation

similar to the Cys/His environment observed in Fe(III) hCBS

and RcoM, and water, which gives rise to axial ligation similar

to the Cys/H₂O environment observed in Fe(III), substrate-

free cytochromes P450. Upon addition of 1-MeIm (liquid) to a

solution of each ﬁve-coordinate complex (in 50:50 CH₂Cl₂/

toluene), we observe an EPR signal characteristic of a low-spin,

Fe(III) thiolate-ligated porphyrin complex (g_z= 2.479−2.491,

g_y= 2.280−2.283, and g_x= 1.868−1.877), which we attribute

to the six-coordinate [Fe(TPP)(S-R)(1-MeIm)] complexes

(Figure S46). We also observe a very minor high-spin (S = 5/

2) signal (less than 1% total signal intensity) with g_⊥= 5.9 and

g_∥= 2.0 for each complex, which we attribute to a small

[Fe(TPP)Cl] contaminant (Figure S46). We were unable to

observe formation of [Fe(TPP)(S-R)(OH₂)] complexes when

adding exogenous water to a solution of ﬁve-coordinate

complex due to the immiscibility between water and the 50:50

CH₂Cl₂/toluene solvent. To overcome this limitation, we

prepared EPR samples using solvent that was not rigorously

dried by distillation over CaH₂. Potentiometric titration using

an automated Karl Fischer titrator revealed that this “wet”

50:50 CH₂Cl₂/toluene mixture contained a concentration of

water greater than 2 mM. Upon dissolution of each ﬁve-

coordinate porphyrin model complex to a concentration of

∼500 μM in wet solvent, we observe an EPR signal

characteristic of a low-spin, Fe(III) thiolate-ligated porphyrin

complex which is distinct from that observed in the presence of

1-MeIm. This new signal, which we attribute to a six-

coordinate [Fe(TPP)(S-R)(OH₂)] complex (g_z= 2.416−

2.429, g_y= 2.267−2.272, and g_x= 1.912−1.917), is less

anisotropic than the signal observed for the 1-MeIm-bound

complexes (Figure S46). The minor [Fe(TPP)Cl] contami-

nant is also observed for the water-bound complexes. We

observed a similar rhombic, low-spin signal upon addition of

dry methanol to 5-c [Fe(TPP)(S-R)] complexes (data not

shown). A representative overlay of low-spin, Fe(III) EPR

spectra for both six-coordinate [Fe(TPP)(S-CH₃)(L′)]

complexes is displayed in Figure 5.

Figure 5. Overlay of low-spin, Fe(III) EPR signals observed for six-

coordinate [Fe(TPP)(S-CH₃)(L′)] complexes where L′ = 1-MeIm

(black) or H₂O (red). Samples were dissolved in 50:50 CH₂Cl₂/

toluene, and spectra were recorded at 10 K. Signal intensities are

normalized to the most intense feature at g_y.

A positive correlation exists between g-value dispersion and

intramolecular H-bond strength for six-coordinate [Fe(TPP)-

(S-R)(L′)] complexes. Figure 6 compares the low-spin, Fe(III)

EPR signals observed for six-coordinate complexes with either

1-MeIm or water bound as the sixth axial ligand trans to the

aryl thiolate. A clear trend is apparent in both sets of

complexes: the g-value dispersion of the low-spin, rhombic

EPR signal increases as the electron-withdrawing character of

the amide N-phenyl para substituent increases. As described

above, the electron-donating/-withdrawing character of this

substituent tunes the acidity of the amide proton and thereby

modulates the strength of the N−H···S hydrogen bond without

signiﬁcantly altering the electronics of the thiophenolate ring

J

DOI: 10.1021/acs.inorgchem.9b02506

Inorg. Chem. XXXX, XXX, XXX−XXX

Inorganic Chemistry

Article

computed negative vibrational frequencies. Representative

geometry-optimized structures for both [Fe(P)(S-NO₂)(L′)]

models are displayed in Figure 7.

Computed geometric parameters and vibrational frequencies

for [Fe(P)(S-R)(L′)] models support our hypothesis that the

electron-donating or -withdrawing character of each N-phenyl

para substituent directly correlates with the strength of the

intramolecular N−H···S hydrogen bond. Increasing the

electron-withdrawing strength of the para substituent resulted

in a marked decrease in calculated amide N−H vibrational

frequency, as well as a slight lengthening of the N−H bond

(Table 7). These observations suggest that changing the

identity of the para substituent modulates the amide N−H

bond strength, and by extension, the acidity of the amide

proton. For example, in [Fe(P)(S-R)(1-MeIm)] models, the

electron-donating methyl substituent gave rise to a stronger N−

H bond (ν_N−H= 3258 cm⁻¹) and a less acidic amide proton, as

compared to the strongly electron-withdrawing nitro sub-

stituent, which gave rise to a signiﬁcantly weaker N−H bond

(ν_N−H= 3176 cm⁻¹) and a more acidic amide proton. Amide

proton acidity, approximated by N−H bond strength, directly

inﬂuences the strength of the intramolecular N−H···S

hydrogen bond: As the amide N−H bond is weakened, the

H-bond donor−acceptor distance decreases, as does the

distance between the amide H atom and thiolate S atom

(Table 7). For example, the [Fe(P)(S-NO₂)(1-MeIm)] model,

which exhibited the weakest amide N−H bond, possessed the

strongest H-bond, represented by the shortest donor−acceptor

distance (3.050 Å). The same trends are observed in to a

similar degree in [Fe(P)(S-R)(H₂O)] models.

Computations predict that strengthening the intramolecular

N−H···S hydrogen bond results in a weakening of the Fe−S

bond. We observed a small but consistent increase in Fe−S

bond distance as the electron-withdrawing character of the

para substituent (and H-bond strength) increased, ranging

from 2.234 Å in [Fe(P)(S-CH₃)(1-MeIm)] to 2.245 Å in

[Fe(P)(S-NO₂)(1-MeIm)] (Table 7). We also observed this

trend in the water-bound series, where the Fe−S bond lengths

ranged from 2.192 Å in [Fe(P)(S-CH₃)(H₂O)] to 2.198 Å in

[Fe(P)(S-NO₂)(H₂O)]. Weakening of the Fe−S bond results

in a concomitant strengthening of the bond between iron and

the sixth axial ligand: The Fe−N(1-MeIm) bond shortens from

2.076 Å in [Fe(P)(S-CH₃)(1-MeIm)] to 2.070 Å in [Fe(P)(S-

NO2)(1-MeIm)], while the Fe−O bond shortens from 2.215

Å in [Fe(P)(S-CH₃)(H₂O)] to 2.207 Å in [Fe(P)(S-

NO₂)(H₂O)]. Using DFT, Paulat et al. predicted a similar

direct correlation between intramolecular H-bond strength and

Fe−S bond distance for a series of porphyrin-thiolate model

compounds with zero, one, or two intramolecular H-bonds.⁶⁷

The electronic origins of the observed changes in bonding of

axial porphyrin ligands will be discussed below.

Figure 6. Comparison of rhombic, low-spin spectra observed for six-

coordinate [Fe(TPP)(S-R)(L′)] complexes where L′ = 1-MeIm (left)

or H₂O (right). Samples were prepared in 50:50 CH₂Cl₂/toluene, and

spectra were recorded at 10 K. Signal intensities are normalized to the

most intense feature at g_y. The vertical lines, centered on g_zand g_xfor

[Fe(TPP)(S-Cl)(L′)] (red), help depict changes in g-value dispersion

as the electron-withdrawing character of the amide para substituent

(R) increases.

directly. We therefore conclude that the observed trend is

indicative of a direct correlation between intramolecular N−

H···S hydrogen bond strength and low-spin, Fe(III) EPR signal

g-value dispersion (Table 6). The connection of these trends

to the porphyrin electronic structure are discussed in detail

below.

Table 6. EPR Parameters for [Fe(TPP)(S-R)(L′)]

a

Complexes

L′ = 1-MeIm

L′ = H₂O

R

g_z

g_y

g_x

g_z

g_y

g_x

CH₃

H

Cl

CF₃

NO₂

2.479

2.478

2.481

2.484

2.491

2.280

2.279

2.280

2.281

2.283

1.877

1.875

1.874

1.868

2.416

2.417

2.419

2.421

2.429

2.267

2.268

2.270

2.272

1.917

1.919

1.916

1.912

a

g_z, g_y, and g_xare given the historical assignments of g_max, g_mid, and g_min

in this table to facilitate comparison of g-values to those of heme-

thiolate proteins reported in the literature.

3.5. Computationally-Observed Changes in Thiolate

H-Bonding. We generated geometry-optimized models of six-

coordinate [Fe(TPP)(S-R)(L′)] (L′ = 1-MeIm, H₂O)

complexes using DFT calculations (BP86, TZVP). To generate

initial coordinates for each model, we used the crystal structure

of ﬁve-coordinate [Fe(TPP)(S-NO₂)]. We modiﬁed these

atomic coordinates to reﬂect the appropriate para substitution

and added a second axial ligand, either 1-MeIm or H₂O.

Additionally, we employed porphinate (P²⁻) as a simpliﬁed

porphyrin structure by replacing each of the tetraphenylpor-

phyrin meso phenyl rings with a hydrogen atom. Frequency

calculations ensured that each geometry-optimized model

represents an energy minimum, as reﬂected by the lack of any

3.6. Insight into H-Bond-Dependent Changes in

Model Porphyrin-Thiolate Electronic Structure. Rhombic

and axial splittings decreased as intramolecular H-bond

strength increased in our 6-c porphyrin-thiolate models.

Using the tetragonally distorted, strong ﬁeld d⁵model

described above, we estimated ligand ﬁeld parameters for

these compounds from experimentally observed g-values.

(Table 8). For both sets of [Fe(TPP)(S-R)(L′)] compounds,

we observe a small decrease in rhombic and axial splittings as

H-bond strength increases (for L′ = 1-MeIm, Δ|V′| = 115 cm⁻¹,

Δ|Δ′| = 82 cm⁻¹; for L′ = H₂O, Δ|V’| = 77 cm⁻¹, Δ|Δ′| = 61

cm⁻¹). This observation demonstrates that the intramolecular

K

DOI: 10.1021/acs.inorgchem.9b02506

Inorg. Chem. XXXX, XXX, XXX−XXX

Inorganic Chemistry

Article

Figure 7. DFT-optimized structures of [Fe(P)(S-NO₂)(L′)] complexes with L′ = 1-MeIm (left) and L′ = H₂O (right). Structures were visualized

using Pymol (v1.3).

Table 7. Geometric Parameters and Amide N−H Vibrational Frequencies for [Fe(TPP)(S-R)(L′)] Complexes Computed

Using DFT

L′ = 1-MeIm

L′ = H₂O

ν_N−H

d_N−H

(Å)

d_D−A

(Å)

d_S−H(amide)

(Å)

d_Fe−S

(Å)

d_Fe−N(ax)

(Å)

ν_N−H

d_N−H

(Å)

d_D−A

(Å)

d_S−H(amide)

(Å)

d_Fe−S

(Å)

d_Fe−O

(Å)

R

(cm⁻¹

)

(cm⁻¹

)

CH₃

H

Cl

CF₃

NO₂

3258

3252

3230

3214

3176

1.032

1.034

1.036

3.070

3.068

3.061

3.059

3.050

2.162

2.159

2.145

2.142

2.124

2.234

2.237

2.239

2.243

2.245

2.076

2.073

2.071

2.070

3306

3301

3279

3266

3234

1.029

1.030

1.031

1.032

1.034

3.089

3.087

3.079

3.075

3.067

2.198

2.193

2.178

2.171

2.155

2.192

2.194

2.195

2.197

2.198

2.215

2.212

2.211

2.207

Table 8. Ligand Field Parameters Computed from Experimental g-Values for [Fe(TPP)(S-R)(L′)] Compounds

a

b

c

d

L′

R

g_x

g_y

g_z

V′/ξ

|V| (cm⁻¹

)

Δ′/ξ

|Δ| (cm⁻¹

)

k

V′/Δ′

CH₃

H

Cl

CF₃

NO₂

CH₃

H

Cl

CF₃

NO₂

−2.479

−2.478

−2.481

−2.484

−2.491

−2.416

−2.417

−2.419

−2.421

−2.429

−2.280

−2.279

−2.280

−2.281

−2.283

−2.267

−2.268

−2.270

1.877

1.875

1.874

1.868

1.917

1.919

1.916

1.912

−4.380

−4.377

−4.339

−4.320

−4.213

−5.486

−5.564

−5.449

−5.459

−5.314

2015

2013

1996

1987

1938

2524

2559

2507

2511

2444

−4.665

−4.668

−4.637

−4.631

−4.535

−5.354

−5.436

−5.340

−5.335

−5.257

2146

2147

2133

2130

2086

2463

2501

2456

2454

2418

1.068

1.065

1.063

1.065

1.056

1.147

1.165

1.148

1.155

1.147

0.939

0.938

0.936

0.933

0.929

1.025

1.024

1.020

1.023

1.011

1-MeIm

H₂O

−2.272

b

a

c

d

V′/ξ = rhombic splitting, ξ = 460 cm⁻¹. Δ′/ξ = axial splitting. k = orbital reduction factor. V′/Δ′ = rhombicity.

H-bond attenuates the thiolate ligand strength, ultimately

giving rise to smaller g-shifts in the rhombic, low-spin EPR

signal. The trans inﬂuence observed in heme-thiolate proteins

was also observed in our porphyrin-thiolate models: Rhombic

and axial splittings were larger for the water-bound complexes

than for 1-MeIm-bound complexes.

As was the case for hemoprotein models with variable axial

ligands, we can rationalize the g-shift variations observed in our

H-bonding models in terms of Fe−S bonding interactions. H-

bond donation to the coordinating thiolate (much like

protonation of the thiolate) alters Fe−S bonding in two

ways. First, H-bonding eﬀectively reduces the negative charge

on sulfur and thereby reduces the ligand donor strength.

Second, the H-bond diminishes the Fe−S π-bonding

interaction, as evidenced by a decrease in sulfur 3p character

observed in the Fe-based 3d_yzMOs of H-bonding models

(Figure S47). Increasing the strength of the H-bond enhances

these two eﬀects and leads to a weakening of the Fe−S bond,

as evidenced by the correlation between H-bond strength and

Fe−S bond distance observed in computational models of our

complexes (Table 7). Weakening of the Fe−S bond causes a

reduction in the π*-antibonding character and subsequent

stabilization of the singly occupied Fe 3d_yz-based MO.

Stabilization of d_yzbrings the lowest-lying Kramers doublets

closer together in energy, giving rise to enhanced spin−orbit

coupling and larger magnitude g-shifts in the presence of a

stronger H-bond. TD-DFT-computed transition energies for

[Fe(P)(S-R)(H₂O)] models provide additional evidence that

H-bonding enhances spin−orbit coupling (Table 9). Values for

both ΔE₁and ΔE₂fell between those computed for the His/

L

DOI: 10.1021/acs.inorgchem.9b02506

Inorg. Chem. XXXX, XXX, XXX−XXX

Inorganic Chemistry

Article

Table 9. Electronic Excitation Energies for

[Fe(P)(S-R)(H₂O)] Computed Using TD-DFT

R

E₁(cm⁻¹

)

E₂(cm⁻¹

)

CH₃

H

Cl

CF₃

NO₂

5442

5428

5388

5374

5337

5812

5791

5756

5724

5687

Cys(S⁻) and His/Cys(SH) models, consistent with the fact

that H-bonding represents an intermediate protonation state.

We observed a decrease in values for ΔE₁and ΔE₂of

approximately 100 cm⁻¹when comparing values between

[Fe(P)(S-CH₃)(H₂O)], which contains the weakest H-bond,

and [Fe(P)(S-NO₂)(H₂O)], which contains the strongest H-

bond. As was the case with the hemoprotein models above, we

caution against interpreting TD-DFT data in a quantitative

fashion (ΔE₁and ΔE₂largely overestimate V′ and Δ′);

however, the trends observed in TD-DFT computations

support the hypothesis that H-bond donation to thiolate

leads to enhanced spin−orbit coupling.

4. DISCUSSION

We have uncovered the electronic origins of the unique

rhombic EPR signal observed in low-spin, Fe(III) heme-

thiolate proteins. Our DFT computations revealed a π-bonding

interaction between Fe and S that is unique to porphyrins

bearing an axial thiolate ligand. This p_π−d_πinteraction

destabilizes the singly occupied, Fe 3d_yz-based MO and gives

rise to larger energy diﬀerences between the ground state and

two lowest-lying excited states. These large ground-/excited-

state energy diﬀerences, predicted using TD-DFT, reﬂect

diminished spin−orbit coupling between the ground-state

Kramers doublet and two low-lying excited states. Diminished

spin−orbit coupling ultimately gives rise to the reduced

magnitude of g-shifts observed for heme-thiolate proteins.

Ligand ﬁeld parameters for low-spin, Fe(III) porphyrins,

estimated from experimentally observed g-values, corroborate

our computational results: We observed signiﬁcantly larger

rhombic and axial splittings in hemoproteins bearing an axial

thiolate ligand compared to hemoproteins with no axial

thiolate.

We visualized the unique features of the heme-thiolate

electronic structure by plotting rhombic and axial splittings for

a number of low-spin, Fe(III) hemoproteins bearing diﬀerent

axial ligands (Figure 8A). The resulting correlation diagram

resembles that originally developed by Blumberg and Peisach

in the 1970s;⁶⁸however, our new correlation diagram utilizes

McGarvey’s systematic approach for identifying the proper g-

tensor axis system for each species.⁴²As is the case in the

Blumberg−Peisach diagram, porphyrin species in Figure 8A

cluster based on the identity of the two axial heme ligands.

Species bearing neutral donors (i.e., His/Met- and His/His-

ligated porphyrins) exhibit the lowest magnitude values for

rhombic and axial splittings. His/His(N⁻)-bound porphyrins

exhibit slightly larger splittings than neutral His/His-bound

species, and hydroxide/phenoxide-bound porphyrins exhibited

large rhombic splitting. As detailed above, thiolate-bound

porphyrins exhibit the largest values for axial splitting and

moderate-to-large rhombic splitting. Within the broad group of

thiolate-ligated porphyrins, two distinct clusters emerge based

Figure 8. (A) Rhombic versus axial splitting correlation diagram for

low-spin, Fe(III) hemoproteins and porphyrin complexes computed

from experimentally derived g-values. EPR parameters utilized to

create the correlation diagram were obtained in this study or from

literature values.^{4,28,30,33−36,69}Porphyrin species are grouped accord-

ing to the identity of the two axial heme ligands. (B) Relationship

between rhombic splitting, axial splitting, and H-bond strength in

[Fe(P)(S-R)(L′)] complexes. For both sets of complexes, the

magnitude of rhombic and axial splittings decreases as H-bond

strength increases.

on the identity of the second axial ligand, reﬂecting the neutral

donor trans inﬂuence.

Through synthesis and characterization of porphyrin-thiolate

model compounds bearing a tunable, intramolecular H-bond,

we have established a direct correlation between the magnitude

of g-shifts observed in the rhombic EPR signal of low-spin,

Fe(III) heme-thiolates and thiolate H-bond strength. To

model H-bonding in heme-thiolates, we utilized a thiolate

ligand with an intramolecular N−H···S H-bond between an

amide and aryl thiolate S atom, and we systematically tuned

the strength of this H-bond by varying the electronics of the

amide. Spectroscopic investigation of low-spin, Fe(III)

porphyrin complexes bearing our model thiolates (in addition

to a second axial ligand, water, or 1-methylimidazole) revealed

a direct correlation between H-bond strength and g-shift in the

rhombic EPR signal. Axial and rhombic splittings, computed

from experimental g-values, also correlate with H-bond

strength. Interestingly, changes in thiolate H-bond strength

give rise to the same slope on the correlation diagram in

compounds with water or 1-methlyimidazole bound as the

trans ligand (Figure 8B). This observation demonstrates that

for heme-thiolates with a single H-bond and a ﬁxed geometry

the thiolate H-bond donor strength inﬂuences heme electronic

structure in a similar manner irrespective of sixth axial ligand

identity. By elucidating this relationship between H-bond

strength and heme-thiolate electronic structure, we provide a

M

DOI: 10.1021/acs.inorgchem.9b02506

Inorg. Chem. XXXX, XXX, XXX−XXX

Inorganic Chemistry

Article

new means to assess the H-bonding environment of heme-

thiolate proteins using the correlation diagram.

The intramolecularly H-bonded porphyrin complexes

characterized in this study oﬀer a new way to model the

thiolate H-bonding network and its role in tuning the reactivity

of heme-thiolate proteins. Hunt and Lehnert recently shed

light on the electronic nature of the “thiolate push” eﬀect,

which gives rise to O−O bond cleavage in Cyt P450s, in terms

of an admixture of Fe−S σ-bonding and Fe−O−O σ*-

antibonding in a doubly protonated model of the Compound 0

ferric peroxo species.¹⁰This study demonstrated that

modulating the thiolate donor strength inﬂuences this critical

bonding interaction, and the authors hypothesize that H-

bonding provides a means to attenuate thiolate donor strength.

Our current study lends support to this hypothesis. Through

computational and spectroscopic analysis of a model system in

which we systematically tune the strength of a heme-thiolate

H-bonding interaction, we have identiﬁed a clear connection

between H-bonding and thiolate ligand strength. On the basis

of changes in the magnitude of g-shifts in low-spin, rhombic

EPR spectra, we demonstrate that increasing the strength of H-

bonding to thiolate leads to a decrease in thiolate donor

strength. The functional implications of these changes in H-

bonding in our model system have yet to be explored.

In conclusion, we developed a synthetic model of H-bonding

in heme-thiolate proteins in which the strength of an

intramolecular H-bond may be ﬁnely tuned. We utilized this

model system to demonstrate that H-bond donation to thiolate

inﬂuences the Fe−S bonding interaction in heme-thiolates

using a combination of computational methods and EPR

spectroscopy. We deﬁned a direct correlation between g-shifts

in low-spin, rhombic EPR spectra of Fe(III) heme-thiolates

and the strength of H-bond donation to thiolate. This

relationship may be utilized to better understand how H-

bonding controls function in the diverse family of proteins

bearing an axial thiolate ligand.

We employed DFT computations to better understand this

correlation in terms of heme-thiolate electronic structure.

From DFT, we observe that H-bonding attenuates the Fe−S π-

bonding interaction and thereby stabilizes the Fe 3d_yz-based

MO. An important consequence of this stabilization is to

decrease ground-/excited-state energy diﬀerences, and TD-

DFT predicts a systematic decrease in these energy diﬀerences

as H-bond strength increases in [Fe(P)(S-R)(H₂O)] com-

pounds. These computational observations suggest that H-

bonding gives rise to greater spin−orbit coupling through

reduction in ground-/excited-state energy diﬀerences, as well

as through decreased Fe−S covalency. Spin−orbit coupling

directly contributes to the magnitude of the EPR g-shift; thus,

an increase in heme-thiolate H-bond strength leads to an

increase in the magnitude of g-shifts.

Given the well-deﬁned relationship between heme-thiolate

electronic structure and rhombic EPR signal, it is possible to

compare thiolate donor strengths among heme-thiolate

proteins through analysis of low-spin, Fe(III) EPR spectra.

As an example, we compare the EPR signals (and resulting

ligand ﬁeld parameters) observed in imidazole-bound Fe(III)

Cyt P450_cam, a type-1 heme-thiolate protein, with that of

PxRcoM-1, a type-2 heme-thiolate protein (Table 4). These

hemoproteins bear axial ligands that are nearly identical.

However, rhombic splitting is slightly larger in PxRcoM-1 (|V′|

= 2019 cm⁻¹) than in Cyt P450_cam(|V′| = 1868 cm⁻¹), while

axial splitting is signiﬁcantly larger in Cyt P450_cam(|Δ′| = 2512

cm⁻¹) than in PxRcoM-1 (|Δ′| = 1960 cm⁻¹). Slightly larger

rhombic splitting in PxRcoM-1 suggests that the heme in this

protein possesses slightly greater Fe−S p_π−d_πbonding

character than in Cyt P450_cam. In contrast, the signiﬁcantly

smaller axial splitting in PxRcoM-1 suggests that the thiolate in

this protein is a weaker donor than the thiolate in Cyt P450_cam

.

ASSOCIATED CONTENT

* Supporting Information

The Supporting Information is available free of charge on the

ACS Publications website at DOI: 10.1021/acs.inorg-

chem.9b02506.

These diﬀerences in heme-thiolate electronic structure are

consistent with each protein’s function. Cyt P450_camis a

monooxygenase enzyme that requires a strong thiolate donor

ligand to facilitate O−O bond cleavage and increase the

basicity of the protonated ferryl species in compound

II.^{2,3,10,22−26}In PxRcoM-1, the thiolate ligand is replaced by

a protein-derived methionine ligand upon heme reduction,^27,28

which may explain why the heme-coordinating thiolate is a

■

S

Detailed methods for X-ray crystallography and addi-

tional crystallographic data, molecular orbital diagrams

for hemoprotein models, spectroscopic data for synthe-

sized compounds, and atomic coordinates for geometry-

optimized structures (PDF)

weaker donor than that in Cyt P450_cam

.

It is likely that second coordination sphere interactions, such

as H-bonding, impart diﬀerential thiolate donor strengths in

heme-thiolate proteins. The crystal structure of Cyt P450_cam

reveals a well-deﬁned thiolate H-bonding network that

includes three backbone amide H-bond donors.¹⁵Long

donor−acceptor distances (3.08−3.56 Å) and nonlinear D−

H−A angles (88−126°) demonstrate that H-bonding to

thiolate is relatively weak in this protein. Our EPR analysis,

which points to a relatively strong thiolate donor in Cyt

P450_cam, is consistent with a weak thiolate H-bonding network.

While no structural data exist for PxRcoM-1, it is likely that the

second coordination sphere diﬀers signiﬁcantly from that of

Cyt P450_cam. Analysis of the PxRcoM-1 rhombic EPR signal

suggests that the overall thiolate donor strength is weaker than

that of Cyt P450_cam. We speculate that a strong H-bond may

exist between the coordinating thiolate and a protein-derived

H-bond donor in PxRcoM-1. This strong H-bond would

facilitate redox-mediated ligand switching by weakening the

Fe−S bond.

Accession Codes

CCDC 1939954 contains the supplementary crystallographic

data for this paper. These data can be obtained free of charge

via www.ccdc.cam.ac.uk/data_request/cif, or by emailing

data_request@ccdc.cam.ac.uk, or by contacting The Cam-

bridge Crystallographic Data Centre, 12 Union Road,

Cambridge CB2 1EZ, UK; fax: +44 1223 336033.

AUTHOR INFORMATION

Corresponding Author

*Phone: 608-262-0328. Fax: 608-262-6143. E-mail: burstyn@

chem.wisc.edu.

■

ORCID

Matthew R. Dent: 0000-0002-7704-4867

Andrew P. Hunt: 0000-0003-0565-9050

Ilia A. Guzei: 0000-0003-1976-7386

N

DOI: 10.1021/acs.inorgchem.9b02506

Inorg. Chem. XXXX, XXX, XXX−XXX

Inorganic Chemistry

Article

(9) McQuarters, A. B.; Speelman, A. L.; Chen, L.; Elmore, B. O.;

Fan, W.; Feng, C.; Lehnert, N. Exploring second coordination sphere

effects in nitric oxide synthase. J. Biol. Inorg. Chem. 2016, 21, 997.

(10) Hunt, A. P.; Lehnert, N. The thiolate trans eﬀect in heme

{FeNO}⁶complexes and beyond: Insight into the nature of the push

eﬀect. Inorg. Chem. 2019.5811317

(11) Aono, S.; Ohkubo, K.; Matsuo, T.; Nakajima, H. Redox-

controlled ligand exchange of the heme in the CO-sensing

transcriptional activator CooA. J. Biol. Chem. 1998, 273, 25757.

(12) Shelver, D.; Thorsteinsson, M. V.; Kerby, R. L.; Chung, S.-Y.;

Roberts, G. P.; Reynolds, M. F.; Parks, R. B.; Burstyn, J. N.

Identification of two important heme site residues (Cysteine 75 and

Histidine 77) in CooA, the CO-sensing transcription factor of

Rhodospirillum rubrum. Biochemistry 1999, 38, 2669.

Nicolai Lehnert: 0000-0002-5221-5498

Author Contributions

M.R.D. carried out computational analyses and wrote the

manuscript. M.W.M. synthesized and characterized ligands and

ligand precursors in the lab of J.N.B. A.P.H. developed the

methodology for the synthesis of 5-c aryl-thiolate complexes in

the lab of N.L. A.P.H., M.W.M., and M.R.D. synthesized and

characterized porphyrin complexes. M.M.A. and I.A.G. carried

out single crystal X-ray diﬀraction experiments.

Notes

The authors declare no competing ﬁnancial interest.

(13) Dhawan, I. K.; Shelver, D.; Thorsteinsson, M. V.; Roberts, G.

P.; Johnson, M. K. Probing the heme axial ligation in the CO-sensing

CooA protein with magnetic circular dichroism spectroscopy.

Biochemistry 1999, 38, 12805.

(14) Vogel, K. M.; Spiro, T. G.; Shelver, D.; Thorsteinsson, M. V.;

Roberts, G. P. Resonance Raman evidence for a novel charge relay

activation mechanism of the CO-dependent heme protein tran-

scription factor CooA. Biochemistry 1999, 38, 2679.

(15) Poulos, T. L.; Finzel, B. C.; Howard, A. J. High-resolution

crystal structure of cytochrome P450cam. J. Mol. Biol. 1987, 195, 687.

(16) Cupp-Vickery, J. R.; Poulos, T. L. Structure of cytochrome

P450 eryF: An enzyme involved in erythromycin biosynthesis. Nat.

Struct. Mol. Biol. 1995, 2, 144.

(17) Ravichandran, K. G.; Boddupalli, S. S.; Hasermann, C. A.;

Peterson, J. A.; Deisenhofer, J. Crystal structure of hemoprotein

domain of P450BM-3, a prototype for microsomal P450’s. Science

(Washington, DC, U. S.) 1993, 261, 731.

(18) Hasemann, C. A.; Kurumbail, R. G.; Boddupalli, S. S.; Peterson,

J. A.; Deisenhofer, J. Structure and function of cytochromes P450: a

comparative analysis of three crystal structures. Structure 1995, 3, 41.

(19) Das, P. K.; Chatterjee, S.; Samanta, S.; Dey, A. EPR, Resonance

Raman, and DFT calculations on thiolate- and imidazole-bound

iron(III) porphyrin complexes: Role of the axial ligand in tuning the

electronic structure. Inorg. Chem. 2012, 51, 10704.

(20) Dey, A.; Jiang, Y.; Ortiz de Montellano, P.; Hodgson, K. O.;

Hedman, B.; Solomon, E. I. S K-edge XAS and DFT Calculations on

cytochrome P450: Covalent and ionic contributions to the cysteine-

Fe bond and their contribution to reactivity. J. Am. Chem. Soc. 2009,

131, 7869.

ACKNOWLEDGMENTS

■

We gratefully acknowledge Prof. Thomas Brunold (University

of Wisconsin, Madison, Chemistry) for helpful discussions

regarding the analysis of computational results. We also thank

Prof. Joshua Telser (Roosevelt University) for helpful

comments and for providing his version of Prof. Bruce R.

McGarvey’s original ligand-ﬁeld theory program, DLSD5.

Finally, we thank Prof. Sam Pazicni (University of Wisconsin,

Madison, Chemistry), whose dissertation laid the foundation

for this work. This research was supported by the National

Science Foundation (CHE-1213739 to J.N.B., CHE-1464696

to N.L.). M.R.D. was supported by the NIH/NIGMS under

award number T32 GM008293. Instrumentation in Paul

Bender Chemical Instrumentation Facility at UW-Madison

were supported as follows: Thermo Q Exactive Plus mass

spectrometer by NIH 1S10 OD020022-1; Bruker Avance III

500 MHz NMR spectrophotometer by a generous gift from

Paul J. and Margaret M. Bender; Bruker EleXsys E500 EPR

spectrophotometer by NSF CHE-0741901.

REFERENCES

■

(1) Smith, A. T.; Pazicni, S.; Marvin, K. A.; Stevens, D. J.; Paulsen,

K. M.; Burstyn, J. N. Functional divergence of heme-thiolate proteins:

A classification based on spectroscopic attributes. Chem. Rev. 2015,

115, 2532.

(2) Sono, M.; Roach, M. P.; Coulter, E. D.; Dawson, J. H. Heme-

containing oxygenases. Chem. Rev. 1996, 96, 2841.

(21) Dey, A.; Okamura, T.-a.; Ueyama, N.; Hedman, B.; Hodgson,

K. O.; Solomon, E. I. Sulfur K-Edge XAS and DFT calculations on

P450 model complexes: Effects of hydrogen bonding on electronic

structure and redox potentials. J. Am. Chem. Soc. 2005, 127, 12046.

(22) Groves, J. T. Key elements of the chemistry of cytochrome P-

450: The oxygen rebound mechanism. J. Chem. Educ. 1985, 62, 928.

(23) Dawson, J. H.; Sono, M. Cytochrome P-450 and chloroper-

oxidase: Thiolate-ligated heme enzymes. Spectroscopic determination

of their active-site structures and mechanistic implications of thiolate

ligation. Chem. Rev. (Washington, DC, U. S.) 1987, 87, 1255.

(24) Behan, R. K.; Green, M. T. On the status of ferryl protonation.

J. Inorg. Biochem. 2006, 100, 448.

(25) Davydov, R.; Im, S.; Shanmugam, M.; Gunderson, W. A.; Pearl,

N. M.; Hoffman, B. M.; Waskell, L. Role of the proximal cysteine

hydrogen bonding interaction in cytochrome P450 2B4 studied by

cryoreduction, electron paramagnetic resonance, and electron−

nuclear double resonance spectroscopy. Biochemistry 2016, 55, 869.

(26) Yosca, T. H.; Ledray, A. P.; Ngo, J.; Green, M. T. A new look at

the role of thiolate ligation in cytochrome P450. JBIC, J. Biol. Inorg.

Chem. 2017, 22, 209.

(3) Yosca, T. H.; Rittle, J.; Krest, C. M.; Onderko, E. L.; Silakov, A.;

Calixto, J. C.; Behan, R. K.; Green, M. T. Iron(IV)hydroxide pK(a)

and the role of thiolate ligation in C-H bond activation by cytochrome

P450. Science (Washington, DC, U. S.) 2013, 342, 825.

(4) Yoshioka, S.; Tosha, T.; Takahashi, S.; Ishimori, K.; Hori, H.;

Morishima, I. Roles of the proximal hydrogen bonding network in

cytochrome P450cam-catalyzed oxygenation. J. Am. Chem. Soc. 2002,

124, 14571.

(5) Galinato, M. G. I.; Spolitak, T.; Ballou, D. P.; Lehnert, N.

Elucidating the role of the proximal cysteine hydrogen-bonding

network in ferric cytochrome P450cam and corresponding mutants

using magnetic circular dichroism spectroscopy. Biochemistry 2011,

50, 1053.

(6) Lehnert, N. Elucidating second coordination sphere effects in

heme proteins using low-temperature magnetic circular dichroism

spectroscopy. J. Inorg. Biochem. 2012, 110, 83.

(7) Mak, P. J.; Yang, Y.; Im, S.; Waskell, L. A.; Kincaid, J. R.

Experimental documentation of the structural consequences of

hydrogen-bonding interactions to the proximal cysteine of a

cytochrome P450. Angew. Chem., Int. Ed. 2012, 51, 10403.

(8) Krest, C. M.; Silakov, A.; Rittle, J.; Yosca, T. H.; Onderko, E. L.;

Calixto, J. C.; Green, M. T. Significantly shorter Fe−S bond in

cytochrome P450-I is consistent with greater reactivity relative to

chloroperoxidase. Nat. Chem. 2015, 7, 696.

(27) Kerby, R. L.; Youn, H.; Roberts, G. P. RcoM: A new single-

component transcriptional regulator of CO metabolism in bacteria. J.

Bacteriol. 2008, 190, 3336.

(28) Bowman, H. E.; Dent, M. R.; Burstyn, J. N. Met104 is the CO-

replaceable ligand at Fe(II) heme in the CO-sensing transcription

factor BxRcoM-1. JBIC, J. Biol. Inorg. Chem. 2016, 21, 559.

O

DOI: 10.1021/acs.inorgchem.9b02506

Inorg. Chem. XXXX, XXX, XXX−XXX

Inorganic Chemistry

Article

(29) Roberts, G. P.; Youn, H.; Kerby, R. L. CO-sensing mechanisms.

Microbiol. Mol. Biol. Rev. 2004, 68, 453.

(30) Roberts, G. P.; Kerby, R. L.; Youn, H.; Conrad, M. CooA, a

paradigm for gas sensing regulatory proteins. J. Inorg. Biochem. 2005,

99, 280.

(31) Kerby, R. L.; Roberts, G. P. Burkholderia xenovorans RcoMBx-1,

a transcriptional regulatorsystem for sensing low and persistent levels

of carbon monoxide. J. Bacteriol. 2012, 194, 5803.

J. E.; Cross, J. B.; Bakken, V.; Adamo, C.; Jaramillo, J.; Gomperts, R.;

Stratmann, R. E.; Yazyev, O.; Austin, A. J.; Cammi, R.; Pomelli, C.;

Ochterski, J. W.; Martin, R. L.; Morokuma, K.; Zakrzewski, V. G.;

Voth, G. A.; Salvador, P.; Dannenberg, J. J.; Dapprich, S.; Daniels, A.

D.; Farkas, O.; Foresman, J. B.; Ortiz, J. V.; Cioslowski, J.; Fox, D. J.

Gaussian 09, revision E.01; Gaussian, Inc.: Wallingford, CT, 2009.

(48) Becke, A. D. Density-functional exchange-energy approxima-

tion with correct asymptotic behavior. Phys. Rev. A: At., Mol., Opt.

Phys. 1988, 38, 3098.

(49) Perdew, J. P. Density-functional approximation for the

correlation energy of the inhomogeneous electron gas. Phys. Rev. B:

Condens. Matter Mater. Phys. 1986, 33, 8822.

(50) Schafer, A.; Horn, H.; Ahlrichs, R. Fully optimized contracted

gaussian basis sets for atoms Li to Kr. J. Chem. Phys. 1992, 97, 2571.

(51) Weigend, F. Accurate Coulomb-fitting basis sets for H to Rn.

Phys. Chem. Chem. Phys. 2006, 8, 1057.

(52) Lee, C. T.; Yang, W. T.; Parr, R. G. Development of the Colle-

Salvetti correlation energy formula into a functional of the electron

density. Phys. Rev. B: Condens. Matter Mater. Phys. 1988, 37, 785.

(53) Weigend, F.; Ahlrichs, R. Balanced basis sets of split valence,

triple zeta valence and quadruple zeta valence quality for H to Rn:

Design and assessment of accuracy. Phys. Chem. Chem. Phys. 2005, 7,

3297.

(54) Durley, R. C.; Mathews, F. S. Refinement and structural

analysis of bovine cytochrome b5 at 1.5 A resolution. Acta Crystallogr.,

Sect. D: Biol. Crystallogr. 1996, 52, 65.

(32) Hines, J. P.; Dent, M. R.; Stevens, D. J.; Burstyn, J. N. Site-

directed spin label electron paramagnetic resonance spectroscopy as a

probe of conformational dynamics in the Fe(III) “locked-off” state of

the CO-sensing transcription factor CooA. Protein Sci. 2018, 27, 1670.

(33) Deng, B.; Parthasarathy, S.; Wang, W.; Gibney, B. R.; Battaile,

K. P.; Lovell, S.; Benson, D. R.; Zhu, H. Study of the individual

cytochrome b5 and cytochrome b5 reductase domains of Ncb5or

reveals a unique heme pocket and a possible role of the CS domain. J.

Biol. Chem. 2010, 285, 30181.

(34) Silkstone, G. G.; Cooper, C. E.; Svistunenko, D.; Wilson, M. T.

EPR and optical spectroscopic studies of Met80X mutants of yeast

ferricytochrome c. Models for intermediates in the akaline transition.

J. Am. Chem. Soc. 2005, 127, 92.

(35) Smith, A. T.; Su, Y.; Stevens, D. J.; Majtan, T.; Kraus, J. P.;

Burstyn, J. N. Effect of the disease-causing R266K mutation on the

heme and PLP environments of the human enzyme cystathionine ß-

synthase. Biochemistry 2012, 51, 6360.

(36) Bohan, T. L. Analysis of low-spin ESR spectra of ferric heme

proteins: A reexamination. J. Magn. Reson. 1977, 26, 109.

(37) Peisach, J.; Blumberg, W. E. Electron paramagnetic resonance

study of the high- and low-spin forms of cytochrome P-450 in liver

and in liver microsomes from a methylcholanthrene-treated rabbit.

Proc. Natl. Acad. Sci. U. S. A. 1970, 67, 172.

(38) Blumberg, W. E.; Peisach, J. Probes of Structure and Function of

Macromolecules and Membranes; Academic Press: New York, 1971;

Vol. 2.

(39) Palmer, G. In The Porphyrins; Dolphin, D., Ed. Academic Press,

Inc.: New York, NY, 1979; Vol. IV, p 313.

(40) Barr, I.; Smith, A. T.; Senturia, R.; Chen, Y.; Scheidemantle, B.

D.; Burstyn, J. N.; Guo, F. DiGeorge critical region 8 (DGCR8) is a

double-cysteine-ligated heme protein. J. Biol. Chem. 2011, 286, 16716.

(41) Stoll, S.; Schweiger, A. EasySpin, a comprehensive software

package for spectral simulation and analysis in EPR. J. Magn. Reson.

2006, 178, 42.

(42) McGarvey, B. R. Survey of ligand field parameters of strong

field d⁵complexes obtained from the g matrix. Coord. Chem. Rev.

1998, 170, 75.

(55) Ochi, H.; Hata, Y.; Tanaka, N.; Kakudo, M.; Sakurai, T.;

Aihara, S.; Morita, Y.; Huber, R. Structure of rice ferricytochrome c at

2·0 Å resolution. J. Mol. Biol. 1983, 166, 407.

(56) Meier, M.; Janosik, M.; Kery, V.; Kraus, J. P.; Burkhard, P.

Structure of human cystathionine β-synthase: a unique pyridoxal 5′-

phosphate-dependent heme protein. EMBO J. 2001, 20, 3910.

(57) Tang, S. C.; Koch, S.; Papaefthymiou, G. C.; Foner, S.; Frankel,

R. B.; Ibers, J. A.; Holm, R. H. Axial ligation modes in iron(III)

porphyrins. Models for the oxidized reaction states of cytochrome P-

450 enzymes and the molecular structure of iron(III) protoporphyrin

IX dimethyl ester p-nitrobenzenethiolate. J. Am. Chem. Soc. 1976, 98,

2414.

(58) Verras, A.; Alian, A.; de Montellano, P. R. Cytochrome P450

active site plasticity: attenuation of imidazole binding in cytochrome

P450(cam) by an L244A mutation. Protein Eng., Des. Sel. 2006, 19,

491.

(59) Schmidt, J. R.; Polik, W. F.; WebMO LLC: Holland, MI, USA,

2013; Vol. 2019.

(60) Griffith, J. S. Binding in hæmoglobin azide as determined by

electron resonance: Theory of electron resonance in ferrihæmoglobin

azide. Nature 1957, 180, 30.

(43) Bendix, J.; Brorson, M.; Schaffer, C. E. Accurate empirical spin-

orbit coupling parameters. zeta. nd for gaseous ndq transition metal

ions. The parametrical multiplet term model. Inorg. Chem. 1993, 32,

2838.

(44) Jiang, Y.; Telser, J.; Goldberg, D. P. Evidence for the formation

of a mononuclear ferric−hydroperoxo complex via the reaction of

dioxygen with an (N4S(thiolate))iron(ii) complex. Chem. Commun.

2009, 6828.

(45) Taylor, C. P. S. The EPR of low spin heme complexes Relation

of the t2g hole model to the directional properties of the g tensor, and

a new method for calculating the ligand field parameters. Biochim.

Biophys. Acta, Protein Struct. 1977, 491, 137.

(61) Palmer, G. The electron paramagnetic resonance of metal-

loproteins. Biochem. Soc. Trans. 1985, 13, 548.

(62) Ueyama, N.; Nishikawa, N.; Yamada, Y.; Okamura, T.-a.;

Nakamura, A. Cytochrome P-450 model (Porphinato)(thiolato)iron-

(III) complexes with single and double NH···S hydrogen bonds at the

thiolate site. J. Am. Chem. Soc. 1996, 118, 12826.

(63) McDaniel, D. H.; Brown, H. C. An extended table of hammett

substitutent constants based on the ionization of substituted benzoic

acids. J. Org. Chem. 1958, 23, 420.

(64) Shinkai, H.; Maeda, K.; Yamasaki, T.; Okamoto, H.; Uchida, I.

Bis(2-(Acylamino)phenyl) disulfides, 2-(Acylamino)benzenethiols,

and S-(2-(Acylamino)phenyl) alkanethioates as novel inhibitors of

cholesteryl ester transfer protein. J. Med. Chem. 2000, 43, 3566.

(65) Sarma, B. K.; Mugesh, G. Redox regulation of protein tyrosine

phosphatase 1B (PTP1B): A biomimetic study on the unexpected

formation of a sulfenyl amide intermediate. J. Am. Chem. Soc. 2007,

129, 8872.

(66) Byrn, M. P.; Strouse, C. E. Porphyrin sponges. Inversion

disorder and inversion twinning in lattice clathrates based on five-

coordinate metallotetraarylporphyrin complexes. J. Am. Chem. Soc.

1991, 113, 2501.

(46) Neese, F. Software update: The ORCA program system,

version 4.0. Wiley Interdiscip. Rev. Comp. Mol. Sci. 2018, 8, No. e1327.

(47) Frisch, M. J.; Trucks, G. W.; Schlegel, H. B.; Scuseria, G. E.;

Robb, M. A.; Cheeseman, J. R.; Scalmani, G.; Barone, V.; Mennucci,

B.; Petersson, G. A.; Nakatsuji, H.; Caricato, M.; Li, X.; Hratchian, H.

P.; Izmaylov, A. F.; Bloino, J.; Zheng, G.; Sonnenberg, J. L.; Hada, M.;

Ehara, M.; Toyota, K.; Fukuda, R.; Hasegawa, J.; Ishida, M.;

Nakajima, T.; Honda, Y.; Kitao, O.; Nakai, H.; Vreven, T.;

Montgomery, J. A., Jr.; Peralta, J. E.; Ogliaro, F.; Bearpark, M.;

Heyd, J. J.; Brothers, E.; Kudin, K. N.; Staroverov, V. N.; Kobayashi,

R.; Normand, J.; Raghavachari, K.; Rendell, A.; Burant, J. C.; Iyengar,

S. S.; Tomasi, J.; Cossi, M.; Rega, N.; Millam, J. M.; Klene, M.; Knox,

P

DOI: 10.1021/acs.inorgchem.9b02506

Inorg. Chem. XXXX, XXX, XXX−XXX

Inorganic Chemistry

Article

(67) Paulat, F.; Lehnert, N. Electronic structure of ferric heme

nitrosyl complexes with thiolate coordination. Inorg. Chem. 2007, 46,

1547.

(68) Blumberg, W. E.; Peisach, J. Low-spin compounds of

hemoproteins. Adv. Chem. Ser. 1971, 100, 271.

(69) Gadsby, P. M. A.; Thomson, A. J. Assignment of the axial

ligands of ferric ion in low-spin hemoproteins by near-infrared

magnetic circular dichroism and electronic resonance spectroscopy. J.

Am. Chem. Soc. 1990, 112, 5003.

Q

DOI: 10.1021/acs.inorgchem.9b02506

Inorg. Chem. XXXX, XXX, XXX−XXX

Article Doi

Electron paramagnetic resonance spectroscopy as a probe of hydrogen bonding in heme-Thiolate proteins

DOI: 10.1021/acs.inorgchem.9b02506

Source and publish data:

Authors:

Article abstract of DOI:10.1021/acs.inorgchem.9b02506

Full text of DOI:10.1021/acs.inorgchem.9b02506

Products guided by the article

R&D Labs maybe for 20054-42-6

Relevant to this article

Hot Product