The Semireduced Mechanism for Nitric Oxide ... - ACS Publications

Subscriber access provided by READING UNIV

Article

The Semireduced Mechanism for Nitric Oxide Reduction by NonHeme Diiron Complexes: Modeling Flavodiiron NO Reductases Corey J. White, Amy L. Speelman, Claudia Kupper, Serhiy Demeshko, Franc Meyer, James P. Shanahan, E. Ercan Alp, Michael Hu, Jiyong Zhao, and Nicolai Lehnert J. Am. Chem. Soc., Just Accepted Manuscript • DOI: 10.1021/jacs.7b11464 • Publication Date (Web): 19 Jan 2018 Downloaded from http://pubs.acs.org on January 19, 2018

Just Accepted “Just Accepted” manuscripts have been peer-reviewed and accepted for publication. They are posted online prior to technical editing, formatting for publication and author proofing. The American Chemical Society provides “Just Accepted” as a free service to the research community to expedite the dissemination of scientific material as soon as possible after acceptance. “Just Accepted” manuscripts appear in full in PDF format accompanied by an HTML abstract. “Just Accepted” manuscripts have been fully peer reviewed, but should not be considered the official version of record. They are accessible to all readers and citable by the Digital Object Identifier (DOI®). “Just Accepted” is an optional service offered to authors. Therefore, the “Just Accepted” Web site may not include all articles that will be published in the journal. After a manuscript is technically edited and formatted, it will be removed from the “Just Accepted” Web site and published as an ASAP article. Note that technical editing may introduce minor changes to the manuscript text and/or graphics which could affect content, and all legal disclaimers and ethical guidelines that apply to the journal pertain. ACS cannot be held responsible for errors or consequences arising from the use of information contained in these “Just Accepted” manuscripts.

Journal of the American Chemical Society is published by the American Chemical Society. 1155 Sixteenth Street N.W., Washington, DC 20036 Published by American Chemical Society. Copyright © American Chemical Society. However, no copyright claim is made to original U.S. Government works, or works produced by employees of any Commonwealth realm Crown government in the course of their duties.

Page 1 of 48 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Journal of the American Chemical Society

The Semireduced Mechanism for Nitric Oxide Reduction by Non-Heme Diiron Complexes: Modeling Flavodiiron NO Reductases Corey J. White†, Amy L. Speelman†, Claudia Kupper§, Serhiy Demeshko§, Franc Meyer§, James P. Shanahan†, E. Ercan Alp‡, Michael Hu‡, Jiyong Zhao‡, Nicolai Lehnert*,† †

Department of Chemistry, The University of Michigan, Ann Arbor, Michigan 48109-1055,

United States §

Universität Göttingen, Institut für Anorganische Chemie, Tammannstrasse 4, 37077 Göttingen,

Germany ‡

Advanced Photon Source, Argonne National Laboratory, Argonne, Illinois 60439, United States

ACS Paragon Plus Environment

1

Journal of the American Chemical Society 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 2 of 48

ABSTRACT: Flavodiiron nitric oxide reductases (FNORs) are a subclass of flavodiiron proteins (FDPs) capable of preferential binding and subsequent reduction of NO to N2O. FNORs are found in certain pathogenic bacteria, equipping them with resistance to nitrosative stress, generated as a part of the immune defense in humans, and allowing them to proliferate. Here, we report the spectroscopic characterization and detailed reactivity studies of the diiron dinitrosyl model complex [Fe2(BPMP)(OPr)(NO)2](OTf)2 for the FNOR active site that is capable of reducing NO to N2O [Zheng et al., J. Am. Chem. Soc. 2013, 135, 4902-4905]. Using UV-Vis spectroscopy, cyclic voltammetry, and spectro-electrochemistry, we show that one reductive equivalent is in fact sufficient for the quantitative generation of N2O, following a semireduced reaction mechanism. This reaction is very efficient and produces N2O with a first order rate constant k > 102 s-1. Further isotope labelling studies confirm an intramolecular N-N coupling mechanism, consistent with the rapid timescale of the reduction and a very low barrier for N-N bond formation. Accordingly, the reaction proceeds at -80o C, allowing for the direct observation of the mixed-valent product of the reaction. At higher temperatures, the initial reduction product is unstable and decays, ultimately generating the diferrous complex [Fe2(BPMP)(OPr)2](OTf), and an unidentified ferric product. These results combined offer deep insight into the mechanism of NO reduction by the relevant model complex [Fe2(BPMP)(OPr)(NO)2]2+, and provide direct evidence that the semireduced mechanism would constitute a highly efficient pathway to accomplish NO reduction to N2O in FNORs and in synthetic catalysts.


2

Page 3 of 48 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


INTRODUCTION: Nitric oxide (NO) is an endogenously generated molecule that plays a dual role in biology. In mammals, NO is a signaling molecule at low (nM) concentrations, acting as an anti-inflammatory agent, and inhibiting platelet aggregation.1,2 On the other hand, macrophages express iNOS in response to bacterial infection, which in turn generates high (μM) concentrations of NO as an immune defense agent,3 toxic to most invasive bacterial pathogens. However, some pathogens have evolved flavodiiron nitric oxide reductases (FNORs) as a response to mitigate this nitrosative stress. These enzymes can detoxify NO via the two electron reduction to N2O, resulting in pathogen proliferation.4 FNORs are therefore of significant interest in bacterial pathogenesis, and evaluating the mechanism of these enzymes contributes to the greater goal of finding new ways to fight bacterial infection. FNORs belong to the family of flavodiiron proteins (FDPs), found in certain bacteria, archaea, and protozoa.5-7 These FDPs act primarily as dioxygen scavengers, protecting these microbes from oxidative stress by reducing of O2 to H2O. In addition to this O2 reductase activity, some FDPs, termed FNORs, exhibit preferential nitric oxide reductase activity, protecting microbes against nitrosative stress by reducing two NO by one electron each to yield the less toxic N2O. For example, FNORs have been found to be expressed in E. Coli specifically as a response to NO exposure, implicating them as part of a defense strategy against NO.4 More recently, FNOR activity has been observed for some members of the FprA family of flavodiiron proteins, found in Desulfovibrio vulgaris8 and Desulfovibrio gigas.9 Kinetic studies comparing NO and O2 reductase activities in Morella thermoacetica,10 an anaerobic bacterium that contains genes for an FprA protein and a high molecular weight ruredoxin, further reveal a 6-fold higher catalytic efficiency for NO vs. O2 reductase activity in the corresponding enzyme.11


3


Page 4 of 48

Crystallographic studies on M. thermoacetica12 and Thermotoga maritima13 FDPs reveal an active site consisting of a non-heme diiron core, with two central iron atoms, each coordinated by two histidine and one carboxylate ligands. In addition, the two iron centers are bridged by water/hydroxide and a carboxylate ligand. The D. gigas14 FDP was found to have a similar coordination sphere, but with one histidine replaced by a water molecule. This, however, was found to be functionally irrelevant by mutagenesis studies.15 In either case, each iron is fivecoordinate with an open 6th coordination site for substrate binding. Additionally, a flavin (FMN) cofactor rests across the subunit interface, within 4-6 Å of the active site (Figure 1) and is within range to facilitate rapid electron transfer. The diiron core of FNORs shows a quite typical coordination sphere compared to other non-heme diiron enzymes, for example ribonucleotide reductase (RNR). However, RNRs have minimal NO reductase activity,16,17 and it is therefore unclear how FNORs are capable of achieving efficient NO reduction to N2O. Several pathways for NO activation and reduction have been proposed for FNORs (Scheme 1). Recent enzyme studies on the T. maritima FDP by Kurtz and coworkers13,18,19 have shown that a diferrous dinitrosyl is formed as the critical intermediate preceding N2O generation. In this case, NO sequentially binds the active site prior to the rate determining (N2O producing) step, implicating the diferrous dinitrosyl complex, [{FeNO}7]2, as the catalytically competent intermediate. Here, we use the Enemark-Feltham notation, {MNO}n, to classify metal-nitrosyl (MNO) complexes.20 This notation is useful in counting electrons of transition metal NO complexes by treating the metal-NO unit as a single entity. In our case, the superscript ‘7’ in {FeNO}7 corresponds to the sum of iron(II)-d and NO(*) electrons. Non-heme high-spin (hs) {FeNO}7 complexes have a total spin S = 3/2, originating from the antiferromagnetic coupling of a high spin FeIII to a bound triplet NO ligand as determined previously.21,22,23 Additionally, the hs-{FeNO}7


4

Page 5 of 48 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


units in the [{FeNO}7]2 intermediate of FNORs are antiferromagnetically coupled, yielding a complex with a total spin of S=0.17-19 From the [{FeNO}7]2 intermediate formed first, the two NO units may react in a direct NN coupling pathway, where no external reductant is necessary, to facilitate N-N bond formation, generating N2O and the diferric form of the diiron active site. The latter is subsequently re-reduced by FMNH2 to the diferrous active state of the enzyme. This pathway is supported by stopped-flow UV/Vis, and rapid freeze quench Mössbauer and EPR studies on the T. maritima FDP.13,18,19 Importantly, the deflavinated FDP is capable of carrying out a turnover of 2 NO to 0.7 equivalents of N2O, suggesting that the FMN cofactor has no catalytic role, but rather acts as a means of rereducing the active site. However, the T. maritima FDP is thought to serve primarily as an O2 reductase in vivo,24 and as such, its NO reductase activity is rather low. Nevertheless, this finding is quite striking, as mono- and dinuclear non-heme {FeNO}7 model complexes have been found to be stable in solution25,26 due to the rather covalent nature of the Fe-NO bond.27 Accordingly, these complexes have no propensity to form N2O without the assistance of an external reductant. It is therefore unclear whether the direct NO reduction pathway accurately reflects the mechanism of NO reduction in designated FNORs under turnover. For this reason, we have employed model complexes to further identify chemically facile pathways for efficient NO reduction by diiron active sites. Since the FMN cofactor of FDPs distinguishes FNORs from other non-heme diiron enzymes with low NO reductase activity, like RNR and methane monooxygenase (MMO), it is possible that the FMN cofactor must provide reducing equivalents to the {FeNO}7 units to activate them for N-N coupling. In this superreduced mechanism proposed previously,7 FMNH2 delivers two reducing equivalents to the [{FeNO}7]2 intermediate, yielding the [{FeNO}8]2 dimer, which


5


Page 6 of 48

is proposed to be activated for N-N coupling. After N2O release, the diferrous active site would be regenerated. The most likely mechanisms for FNORs are summarized in Scheme 1. A number of FNOR model complexes have been reported in the literature, beginning with Lippard’s [Fe2(Et-HPTB)(O2CPh)(NO)2](BF4)2 complex.26,28 This complex is capable of substoichiometric N2O production upon photocleavage of one of the Fe-NO bonds. It is proposed that this generates a bridging mononitrosyl species, which is then electrophillically attacked by the previously photocleaved NO. Interestingly, recent studies on a mononuclear {FeNO}7 complex, [Fe(NO)(N3PyS)]BF4,29 demonstrate that chemical reduction to the {FeNO}8 state leads to substoichiometric N2O formation via bimolecular coupling in a superreduced mechanism. Along the same lines, the diiron mononitrosyl complex, [Fe2(N-Et-HPTB)(NO)(DMF)3](BF4)3,30 is able to generate quantitative (0.5 equivalents) amounts of N2O upon reduction via a superreduced mechanism. We

have

previously

reported

the

diferrous

dinitrosyl

complex,

[Fe2(BPMP)(OPr)(NO)2](BPh4)2,25 which is stable in solution and was structurally characterized. Each iron center has a pseudo-octahedral coordination environment. The BPMP ligand induces a slight asymmetry in the complex, where one {FeNO}7 unit has the tertiary amine as the trans ligand to NO, while the other one has a pyridine bound trans to NO. This complex exhibits antiferromagnetic coupling between the two S=3/2 {FeNO}7 units, as in FDPs16,17,31 and other diiron model complexes26 with short Fe-Fe bond distances and several bridging ligands. Importantly, upon addition of two equivalents of reductant (cobaltocene, CoCp2), this complex shows fast and quantitative N2O generation, which constitutes evidence that this complex produces N2O via the superreduced pathway. These results demonstrate that reduction is a very potent way to activate otherwise stable {FeNO}7 complexes to become more reactive.


6

Page 7 of 48 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


In this work, we sought to gain deeper mechanistic insight into the N-N coupling mechanism of [Fe2(BPMP)(OPr)(NO)2]2+, using spectroscopic and kinetic studies. First, we devised a new synthesis for this complex (as the triflate salt), which produces the compound in greater purity and mitigates solubility issues that were encountered with the previously used tetraphenylborate counterion. Importantly, [Fe2(BPMP)(OPr)(NO)2]2+ shows rapid and quantitative N2O production with the addition of just one equivalent of reductant, which we refer to as the “semireduced” pathway. Using spectro-electrochemistry (SEC) and stopped-flow measurements, we demonstrate that N2O production is much faster than previously assumed based on IR gas headspace N2O detection. N-N coupling additionally occurs at -80o C as evidenced by gas-headspace IR, solution IR, and UV-Vis spectroscopy, yielding a mixed-valent FeIIFeIII species that then decays at room temperature. Cyclic voltammetry, Mössbauer spectroscopy, and crystallographic studies on the room temperature product indicate that over time, this product disproportionates to form the diferrous complex [Fe2(BPMP)(OPr)2]+ and a ferric product. In this way,

we

have

obtained

detailed

insight

into

the

NO

reduction

mechanism

of

[Fe2(BPMP)(OPr)(NO)2]2+ with direct relevance for FNORs and synthetic catalysts for NO remediation.

EXPERIMENTAL SECTION: General Considerations. Air sensitive materials were prepared under a dinitrogen atmosphere in either an MBraun glovebox, or using standard Schlenk techniques. Unless specified otherwise, low temperature experiments were conducted in an Inert Technologies glovebox modified with a coldwell. All solvents were dried using standard techniques, freeze-pump-thawed to remove dioxygen, and stored over molecular sieves. All reagents were purchased from commercial


7


Page 8 of 48

sources and were used as received, unless noted below. Fe(OTf)2 was recrystallized from ~10 mL of acetonitrile with ~70 mL of diethyl ether and dried under high vacuum to give [Fe(OTf)2](CH3CN)2.32 Cobaltocene (CoCp2) was sublimed under static vacuum and stored at 35o C in the glovebox freezer.

Nitric oxide was passed over an Ascarite II column and

subsequently through a -80o C cold trap to remove higher NxOy impurities. Nitric oxide-15N18O and 15NO were used without any further purification. [Fe(BF4)2]·2.5(CH3CN). Under an inert atmosphere, 103 mg (1.8 mmol) of Fe powder and 399 mg (3.4 mmol) of NOBF4 were combined in a Schlenk flask and dissolved in 12 mL of acetonitrile. The reaction mixture was then stirred on the Schlenk line under static vacuum, while periodically refilling and evacuating the headspace with argon, always leaving the flask under a static vacuum. After 6 hours, the solvent was removed under high vacuum and the flask was brought into the glovebox. The precipitate was then re-dissolved in 5 mL of acetonitrile and filtered through a 0.2 μm PTFE syringe filter to remove excess iron powder. The filtrate was precipitated with 50 mL of diethylether and stored at -35o C overnight. The product was then filtered off, yielding an offwhite precipitate, which was dried under high vacuum overnight. Yield: 480 mg (1.4 mmol, 85%) of the dried product was collected as a white powder. Characterization: Elemental anal. Calcd for C5/2H7/2B2F8FeN5/2: C, 17.93; H, 2.26; N, 10.46; found (%): C, 18.16; H, 2.59; N, 10.40. IR (KBr) ν(CH3CN) 2317, 2289 cm-1. 2,6-bis[[bis(2-pyridylmethyl)amino]methyl]-4-methylphenol (H[BPMP]). H[BPMP] was generated via a modified literature procedure.33,34 In a dry 3-neck flask under a dinitrogen atmosphere, equipped with a dropping funnel and a condenser, 1mL (5.6 mmol) of di-2picolylamine and 1 mL (7.17 mmol) of triethylamine were combined in 12 mL anhydrous tetrahydrofuran. 608 mg (2.9 mmol) of 2,6-bis(chloromethyl)-4-methylphenol were dissolved in


8

Page 9 of 48 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


12 mL anhydrous tetrahydrofuran, loaded in the dropping funnel, and added slowly to the 0o C chilled reaction mixture. The reaction was then heated to reflux for 5 hrs before cooling the solution back to RT and filtering off the triethylammonium chloride. The filtrate was concentrated under reduced pressure to a yellow oil, which was then taken up in minimal dichloromethane and precipitated with hexanes. The resulting cloudy solution was immediately filtered through a celite plug and the filtrate was allowed to precipitate at -35o C overnight. The product was filtered off, giving 1.06 g (2.0mmol, 74 % yield) of a pale-yellow powder which was used without further purification. Characterization: 1H NMR (400 MHz, Chloroform-d) δ 10.75 (s, 1H), 8.50 (dd, 4H), 7.58 (td, 4H), 7.47 (d, 4H), 7.10 (t, 4H), 6.98 (s, 2H), 3.85 (s, 8H), 3.76 (s, 4H), 2.22 (s, 3H). [Fe2(BPMP)(OPr)](OTf)2 (1). Under an inert atmosphere, 235 mg (0.45 mmol) of H[BPMP] and 35 mg (0.50 mmol) of KOMe were combined in a Schlenk flask and dissolved in 3 mL of MeOH. 370 mg (0.88 mmol) of [Fe(OTf)2](CH3CN)2 dissolved in 1 mL of MeOH were added and the solution was stirred for 5 min before slowly adding 43 mg (0.45 mmol) NaOPr dissolved in 1 mL of methanol, giving a yellow-orange solution. The solution was stirred for 1hr before precipitating it overnight at -35o C with 70 mL of diethylether. The resulting yellow precipitate was filtered through a frit and washed with dichloromethane until only the white KOTf/NaOTf salt remained on the frit. The filtrate was then precipitated with 30 mL of hexanes, and the yellow powder was separated with a frit, giving 392 mg (0.39 mmol, 88% yield) of the final product. Characterization: Elemental anal. calcd for C38H38F6Fe2N6O9S2: C, 45.08; H, 3.78; N, 8.30; found (%): C, 44.83; H, 3.78; N, 8.23. Mössbauer (80K, powder) 𝛿 = 1.19 mm/s, EQ = 2.89 mm/s. UV-Vis (CH2Cl2) 400 nm. [Fe2(BPMP)(OPr)(NO)2](OTf)2 (2a). Under an inert atmosphere, 335 mg (0.331 mmol) of (1) was dissolved in 3 mL CH2Cl2, and the solution was stirred in a Schlenk flask under a headspace


9


Page 10 of 48

of NO gas for 15 min. 24 mL of hexanes were syringed in to precipitate the product, and the flask was stored at -35o C overnight. The solution was filtered through a frit, making sure to stop the vacuum while there was still a small layer of solvent remaining to prevent NO loss. 291 mg (0.271 mmol, 82% yield) of the product, as a brown powder, were collected after the residual solvent had evaporated. Characterization: IR (KBr) ν(NO) 1768 cm-1. Mössbauer (80K, powder) 𝛿 = 0.70 mm/s, EQ = 1.72 mm/s. UV-Vis (CH2Cl2) 400 nm. [Fe2(BPMP)(OPr)(15NO)2](OTf)2 (2b). A 50 mL Schlenk flask was equipped with a T-joint and connected to the NO tank and the Schlenk line, with an additional T-joint with a 3-way stopcock in-between the Schlenk flask and the Schlenk line leading to a balloon to monitor the pressure of the system (see Figure S1). The balloon was purged in 30 quick cycles of evacuation and refilling with Ar, and was then held under a positive pressure while the Schlenk flask was purged five times with 30 minute evacuation cycles. The balloon was then evacuated and the three way stopcock was turned to be open to the balloon and the NO tank. At this point, 1 dissolved in minimal CH2Cl2 was syringed into the Schlenk flask and the NO tank was opened until the balloon was under a positive pressure. The flask was briefly manually shaken under 15NO pressure and then taken into the glovebox to stir for an additional hour before precipitating the product with ~20 mL of hexanes at -35o C overnight. The reaction mixture was filtered through a frit, taking care to stop the vacuum while there was still a small layer of solvent above the precipitate. Characterization: IR (KBr) ν(NO) 1734 cm-1. [Fe2(BPMP)(OPr)(15N18O)2](OTf)2

(2c). The same procedure was followed as for 2b.

Characterization: IR (KBr) ν(NO) 1695 cm-1. [Fe2(BPMP)(OPr)(NO)2](BF4)2 (3).

Under an inert atmosphere, 84.5 mg (0.16 mmol) of

H[BPMP] and 12.7 mg (0.18 mmol) of KOMe were combined in 2 mL of MeOH. 110.0 mg (0.31


10

Page 11 of 48 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


mmol) of [Fe(BF4)2]·2.5(CH3CN) dissolved in 1 mL of methanol were then added to the solution, followed by the slow addition of 16.6 mg (0.17 mmol) NaOPr in 1 mL of MeOH, yielding a yellow-orange solution. This solution was then stirred under a headspace of NO gas for 15 min before precipitating the crude product with 24 mL of diethyl ether at -35o C overnight. The resulting brown powder was filtered off with a frit and the precipitate was washed with CH2Cl2, being careful to apply minimal vacuum to avoid NO loss. The filtrate was precipitated with hexanes at -35o C overnight. The powder was again filtered off with a frit, yielding 71.0 mg (0.02 mmol, 46.8% yield) of the final product. Characterization: IR (ATR) ν(NO) 1760 cm-1. [Fe2(BPMP)(OPr)2](OTf) (4). Under an inert atmosphere, 233 mg (0.45 mmol) of H[BPMP] and 40.1 mg (0.57 mmol) of KOMe were combined in a Schlenk flask and dissolved in 3 mL of MeOH. 375 mg (0.89 mmol) of [Fe(OTf)2](CH3CN)2 dissolved in 1 mL of MeOH were added and the solution was stirred for 5 min before slowly adding 89.1 mg (0.93 mmol) NaOPr dissolved in 1 mL of MeOH, giving an orange solution. The solution was stirred for 1 hr before precipitating it overnight at -35o C with 70 mL of diethyl ether. The resulting solution was filtered through a frit and the precipitate was washed with dichloromethane until only the white KOTf/NaOTf salt remained. The filtrate was then precipitated with 30 mL of hexanes, and the resulting orange powder was filtered off, giving 248 mg (0.26 mmol, 59% yield) of product. Characterization: Elemental anal. calcd for C40H44F3Fe2N6O8S: C, 51.24; H, 4.73; N, 8.96; found (%): C, 51.01; H, 4.71; N, 9.10. Mössbauer (80 K, powder) 𝛿 = 1.18 mm/s, EQ = 2.69 mm/s. UV-Vis (λmax) 444 nm. 1H NMR (400 MHz, Acetonitrile-d3, broad singlets) δ 185.10, 176.19, 142.00, 79.95, 73.56, 62.59, 46.78, 42.33, 41.46, 35.91, 31.61, 27.13, 24.42, 12.85, 11.70, 10.80, 5.45, 3.42, 2.14, 1.95, 1.29, 1.14, 0.89.


11


Page 12 of 48

[Zn2(BPMP)(OPr)2](OTf)(CH2Cl2) (5). This complex was synthesized in an analogous manner as 4, using Zn(OTf)2. Characterization: Elemental anal. calcd for C41H47Cl2F3N6O8SZn2: C, 47.23; H, 4.54; N, 8.06; found (%): C, 47.45; H, 4.06; N, 8.36. 1H NMR (400 MHz, Acetonitrile-d3, broad singlets) δ 8.80, 8.56, 8.23, 7.94, 7.50, 7.44, 7.31, 7.02, 6.66, 6.47, 6.42, 5.45, 4.51, 4.03, 3.91, 3l49, 3.42, 3.21, 2.28, 2.02, 1.29, 0.95, 0.89. [Zn2(BPMP)(OAc)2](OTf). Due to disorder in the terminal methyl group of propionate in 5, the crystal structure of this complex could not be completely refined. The acetate analog of 5 was therefore prepared, following the procedure for 4 and using NaOAc in place of NaOPr. Crystals suitable for X-ray diffraction were obtained by vapor diffusion of diethyl ether into a -30o C solution of [Zn2(BPMP)(OAc)2](OTf) dissolved in MeOH (Figure S25). Preparation of 57Fe Samples: 57Fe-labelled complexes of 1 and 2a were synthesized in the same manner as the unlabeled complexes from 57

57

Fe(OTf)2, generated in situ from the metathesis of

FeCl2 with Ag(OTf) as previously described.23

Physical Measurements: UV-Vis/Immersion Probe: Spectra were taken using an Analytik Jena Specord S-600 UV-Vis spectrometer. Dip probe experiments used the same spectrometer, with a Helma low-temperature immersion probe. IR (ATR/KBr/Gas): Spectra of solids were taken on PerkinElmer BX and GX, and on a Bruker Alpha-E FTIR spectrometer. Solution samples were measured in a thin-layer solution cell equipped with CaF2 windows. Gas samples were measured using a Pike HT gas cell (10 cm) with CaF2 windows on the same instruments. 1

H NMR: Measurements were conducted on a Varian MR 400 instrument at room temperature and

were referenced against tetramethylsilane.


12

Page 13 of 48 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


Cyclic Voltammetry: Cyclic voltammograms were obtained on a CH Instruments CHI600E electrochemical workstation with a glassy-carbon working electrode, platinum counter electrode, and silver wire psuedoreference electrode. All potentials were referenced to a ferrocene standard, taken before each experiment under the same experimental conditions. Data were collected on ~5 mM samples in CH2Cl2 with 0.1 M tetrabutylammonium triflate as the supporting electrolyte. IR Spectroelectrochemistry: IR spectroelectrochemical experiments were conducted in a LabOmak UF-SEC thin layer cell, with Pt mesh working and counter electrodes, and an Ag wire pseudoreference electrode. EPR: X-Band electron paramagnetic resonance spectra were obtained on a Bruker X-band EMX spectrometer equipped with an Oxford Instruments liquid helium cryostat. Spectra were recorded on ~2 mM frozen solutions using 20 mW microwave power and 100 kHz field modulation at a 1 G amplitude. NRVS: Nuclear resonance vibrational spectroscopy (NRVS) data were obtained as described previously.35 Spectra were taken at beamline 3ID at the Advanced Photon Source (APS) at Argonne National Laboratory. Powder samples were loaded into copper sample cells with a 4x7x1 mm sample compartment. Spectra were taken from 0 to 80 or to 120 meV (depending on the compound) in 0.25 meV steps, and represent the sum of 6 to 12 scans. X-Ray Crystallography: Orange blocks of 4 were grown from layering a solution of the room temperature reduction product in tetrahydrofuran with hexanes at 22o C. A crystal of dimensions 0.20 x 0.04 x 0.03 mm was mounted on a Rigaku AFC10K Saturn 944+ CCD-based X-ray diffractometer equipped with a low temperature device and Micromax-007HF Cu-target microfocus rotating anode ( = 1.54187 Å) operated at 1.2 kW power (40 kV, 30 mA). The X-ray intensities were measured at 85(1) K with the detector placed at a distance of 42.00 mm from the


13


Page 14 of 48

crystal. A total of 2028 images were collected with an oscillation width of 1.0 in  The exposure times were 1 sec. for the low angle images, 5 sec. for high angle. Rigaku d*trek images were exported to CrysAlisPro for processing and corrected for absorption. The integration of the data yielded a total of 36133 reflections to a maximum 2 value of 138.70 of which 8598 were independent and 8072 were greater than 2(I). The final cell constants (Table S2) were based on the xyz centroids of 16330 reflections above 10(I). Analysis of the data showed negligible decay during data collection. The structure was solved and refined with the Bruker SHELXTL (version 2016/6) software package, using the space group P1bar with Z = 2 for the formula C41H51N6O9F3SFe. All non-hydrogen atoms were refined anisotropically with the hydrogen atoms placed in idealized positions. Full matrix least-squares refinement based on F2 converged at R1 = 0.0522 and wR2 = 0.1458 [based on I > 2sigma(I)], R1 = 0.0544 and wR2 = 0.1494 for all data. The SQUEEZE subroutine of the PLATON program suite was used to address a disordered THF solvate molecule present in the structure. Additional details are presented in the Supporting Information. Note that the crystal structure was deposited with the CCDC, deposition no. 1572848. Yellow needles of [Zn2(BPMP)(OAc)2](OTf) were grown from a methanol/diethyl ether solution of the compound at -30o C. A crystal of dimensions 0.15 x 0.02 x 0.02 mm was mounted as described above. The integration of the data yielded a total of 31369 reflections to a maximum 2 value of 139.21 of which 7555 were independent and 7142 were greater than 2(I). The final cell constants (Table S8) were based on the xyz centroids of 17240 reflections above 10(I). Analysis of the data showed negligible decay during data collection; the data were processed with CrystalClear 2.0 and corrected for absorption. The structure was solved and refined with the Bruker SHELXTL (version 2014/6) software package, using the space group P1bar with Z = 2 for the formula C38H39N6O8F3SZn2 + [solvent]. All non-hydrogen atoms were refined anisotropically


14

Page 15 of 48 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


with the hydrogen atoms placed in idealized positions. Full matrix least-squares refinement based on F2 converged at R1 = 0.0363 and wR2 = 0.0959 [based on I > 2sigma(I)], R1 = 0.0380 and wR2 = 0.0985 for all data. The SQUEEZE subroutine of the PLATON program suite was used to address the disordered solvent present in the structure. Additional details are given in the Supporting Information. Note that the crystal structure was deposited with the CCDC, deposition no. 1816883. Stopped-Flow IR: Stopped flow IR measurements were conducted at room temperature with a VERTEX 70 FTIR spectrometer with a liquid nitrogen cooled MCT detector at a 4cm-1 resolution with OPUS software in Rapid Scan mode. 5-10mM sample solutions were prepared in the glovebox and loaded into luer-lock plastic syringes. The syringes were connected to a TgK Scientific SF-16 syringe drive operated by KinetaDrive software, which was used to mix the solutions in a 100μm pathlength cell. The tubing was initially washed with degassed CH2Cl2, which was also used as a solvent blank before loading solutions of (2a) and/or CoCp2. Mössbauer Spectroscopy: Mössbauer spectra were recorded using an alternating constant WissEl Mössbauer spectrometer, consisting of an MR 360 Drive Unit, an MV-1000 velocity transducer, and an LND 45431 proportional counter mounted on an LINOS precision bench. The system was operated in a horizontal transmission geometry with source, absorber, and detector in a linear arrangement. The temperature was controlled and maintained using a Janis SHI closed-cycle helium cryostat. Data acquisition was performed using a 512 channel analyzer. Isomer shifts were referenced to α-iron metal foil at ambient temperatures. Simulation of experimental data was performed using the Mfit program.36 SQUID: Magnetic susceptibility measurements were conducted on a Quantum-Design MPMS XL5 SQUID magnetometer, equipped with a 5 T magnet. Powder samples were loaded into a gelatin


15


Page 16 of 48

capsule, fixed in place using a perfluorinated oil, and placed in a plastic straw sample holder. The raw data were corrected for the diamagnetic contribution of the gelatin capsule as well as the diamagnetic contribution of the complex using Pascal constants. The data were fit using the julX program.37 Elemental Analysis: Elemental analyses were conducted by Atlantic Microlabs (Norcross, GA). Chemical Reduction/N2O tests: N2O tests were performed as previously described.25 Briefly, ~5 μmol of (2a) were dissolved in 2 mL of CH2Cl2 in a 25 mL round bottom flask with a septa. 1eq of CoCp2 dissolved in 0.5 mL of CH2Cl2 was added and the reaction mixture was then stirred for various times before transferring the gas headspace into the evacuated gas-IR cell for exactly 20 s via a cannula. The N2O band was integrated from 2150 to 2275 cm-1 after the subtraction of a CH2Cl2 blank taken under the exact same conditions, and the integration was compared to a standard curve (Figure S16). RESULTS: Synthesis. Our initially reported synthesis of [Fe2(BPMP)(OPr)(NO)2](BPh4)2 suffers from several drawbacks. In the original procedure, the precursor complex [Fe2(BPMP)(OPr)](BPh4)2 is obtained via a salt metathesis step of the initially formed triflate complex with NaBPh4. The following reaction with NO then gives [Fe2(BPMP)(OPr)(NO)2](BPh4)2. While some batches of this complex yielded N2O quantitatively, more typical yields were in the 60-80% range. Elemental analyses of the precursor complex showed that this was largely due to varying amounts of salt impurities in the bulk material, due to the salt metathesis step.

Additionally, the

tetraphenylborate salts of the complexes exhibit low solubility in organic solvents and are insoluble in water, likely due to the tetraphenylborate counterion. In order to mitigate both of these issues, we then worked out the synthesis of the analogous triflate complex [Fe2(BPMP)(OPr)](OTf)2 (1),


16

Page 17 of 48 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


which can be obtained in consistent (high) purity, as evidenced by elemental analysis, and Mössbauer, IR and UV-Vis spectroscopy. Note that the spectroscopic data obtained for 1 are fully consistent with those of the tetraphenylborate analog. Accordingly, the nitrosylated complex [Fe2(BPMP)(OPr)(NO)2](OTf)2 (2a) consistently gives quantitative N2O yields, and is completely soluble in CH2Cl2, CH3CN, and MeOH. Spectroscopic

Characterization

of

1

and

2a.

The

diferrous

precursor

[Fe2(BPMP)(OPr)](OTf)2 (1) is a bright yellow complex, which displays a moderately strong absorption band at 400 nm (ε = 2000 M-1cm-1; see Figure 2) in its UV-Vis spectrum, assigned to an iron(II)-to-pyridine metal-to-ligand charge transfer (MLCT) band.38 Mössbauer experiments performed on powder samples of 1 show a single quadrupole doublet, indicative of electronically identical iron sites in 1. Figure S8 shows the experimental data along with a fit, delivering an isomer shift of 1.19 mm/s and a quadrupole splitting Eq of 2.89 mm/s, which are both indicative of high-spin iron(II). The spin state of the iron(II) centers is further confirmed by magnetic susceptibility data, shown in Figure S9. The data are fit assuming two high-spin Fe(II) centers (S = 2 each), using the Spin Hamiltonian, H = -2J•SA•SB

(1)

where SA and SB are the spin operators for centers A and B, and J is the exchange coupling constant. A fit of the data in Figure S9 to equation (1) gives a coupling constant of J = -3.6 cm-1, indicating weak antiferromagnetic coupling between the two Fe(II) ions. Upon nitrosylation of 1 to form [Fe2(BPMP)(OPr)(NO)2](OTf)2 (2a), there is an immediate color change of the solution to dark brown. The UV-Vis absorption spectrum of 2a shows a broad band at 400 nm (ε = 2000 M-1cm-1; see Figure 2). As mentioned in the Introduction, non-heme high-spin {FeNO}7 centers are best described as Fe(III)-NO, where a high-spin Fe(III) center (S


17


Page 18 of 48

= 5/2) is coordinated to a triplet NO ligand (S = 1), and the spins are antiferromagnetically coupled (total spin: S = 3/2). These complexes typically exhibit relatively weak NO to Fe(III) ligand-tometal charge transfer (LMCT) transitions in the 400 – 500 nm range,21,27,39 and we assign the 400 nm feature of 2a to a corresponding LMCT transition. SQUID data for 2a are shown in Figure S9. These data reveal two {FeNO}7 centers with S = 3/2 spin states that are antiferromagnetically coupled, confirming that the {FeNO}7 units are in the high-spin state. A fit of the data to equation (1) delivers an exchange coupling constant J of -7.0 cm-1. Antiferromagnetic coupling is weaker than in Lippard’s [{FeNO}7]2 complex [Fe2(Et-HPTB)(O2CPh)(NO)2](BF4)2 (J = -23 cm-1; EtHPTB = N,N,N’N’-tetrakis(N-ethyl-2-benzimidazolyl-methyl)-1,3-diaminopropane).26 Mössbauer experiments performed on powder samples of 2a show a single quadrupole doublet, indicative of similar iron sites in this complex, despite the slight asymmetry in the ligand coordination.25 Figure S8 shows the experimental data along with a fit, which delivers an isomer shift of 0.70 mm/s and a quadrupole splitting of 1.72 mm/s for 2a, consistent with similar reported high-spin non-heme ferrous {FeNO}7 complexes in the literature.19,26,40 The decrease in the isomer shift upon nitrosylation in 2a compared to 1 is consistent with the idea that binding of NO to a high-spin iron(II) leads to an oxidation of the iron center, in agreement with the Fe III-NO electronic structure description of these types of complexes (see above).21,22,41 In turn, the fact that the isomer shift is larger compared to analogous Fe(III) complexes with the BPMP ligand, which are typically around 0.5 mm/s,34 reflects the large amount of electron density that is donated from the NO ligand back to the Fe(III) center (by forming two strong -donor bonds), as documented previously.23,42 The FT-IR spectrum of 2a shows the N-O stretch as a strong band at 1768 cm-1, which shifts to 1734 cm-1 upon 15NO labeling (Figure 3). The frequency of the N-O stretch is consistent


18

Page 19 of 48 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


with other mono- and dinuclear non-heme high-spin {FeNO}7 model complexes,25,26,34 which generally exhibit this feature around 1700 – 1800 cm-1. In comparison, the NO adduct of RNR shows the N-O stretch at 1742 cm-1,17,31 and, importantly, the N-O stretching frequency of 2a compares very well to that of the dinitrosyl adduct of Tm FDP, reported at 1749 cm-1.13 In order to measure the Fe-NO stretch of 2a, nuclear resonance vibrational spectroscopy (NRVS) of the 57Fe analogue of this complex was applied. As shown in Figure S7, NRVS exhibits an isotope-sensitive band at 486 cm-1 that downshifts to 466 cm-1 upon 15N18O labelling. We assign this feature to the Fe-NO stretch of 2a. In comparison, the Fe-NO stretching frequencies of the mono- and dinitrosyl adducts of the deflavinated T. maritima FDP are reported at lower energies of 451 cm-1 and 459 cm-1, respectively, indicating that the electronic environment of the Fe centers in the enzyme is more electron-rich compared to our model complex.13,43 We had previously established a correlation between the N-O and the Fe-NO stretching frequencies for mononuclear high-spin {FeNO}7 complexes with the TMPA (tris(2-pyridylmethyl)amine) and BMPA-Pr (N-propanoate-N,N-bis(2-pyridylmethyl)amine) coligands.23 These serve as model complexes for the nonheme FeB site of bacterial NO reductase, with N4 and N3O coordination spheres, respectively. As shown in Figure S10, complex 2a lies right on the correlation line that was previously established for the mononuclear compounds. This is actually not surprising, considering the very similar coordination spheres of the iron centers in 2a and the mononuclear {FeNO}7 complexes used to establish the correlation. However, what this shows is that the individual {FeNO}7 units in 2a have very similar electronic structures as the mononuclear compounds, and that the {FeNO}7 units in 2a are not strongly electronically coupled. Finally, cyclic voltammetry of 2a displays an irreversible reductive wave at -1.15 V vs. Fc/Fc+ (Figure 4) consistent with reduction of the {FeNO}7 complex and subsequent fast transformation of NO to N2O. A new wave forms at -0.45


19


Page 20 of 48

V (vs. Fc/Fc+) after reduction, which we attribute to the reduction product. These data are essentially identical to the electrochemical data reported previously for the analog [Fe2(BPMP)(OPr)(NO)2](BPh4)2. Comparison of the Semi- and Superreduced Mechanisms. For our initially reported complex [Fe2(BPMP)(OPr)(NO)2](BPh4)2, we observed quantitative N2O generation after the addition of two equivalents of reductant (CoCp2).25 We initially interpreted this as evidence that the complex uses a superreduced mechanism for N2O formation. In any case, this is clear evidence that reduction of otherwise stable hs-{FeNO}7 complexes serves as a potent means to activate them and make them more reactive. Our results for the mononuclear complex [Fe(TMG3tren)(NO)]2+ demonstrate that one-electron reduction leads to a decrease in the covalency of the Fe-NO bond, and increases negative charge and radical character on the triplet NO ligand, which further supports these ideas.42,44 The [Fe2(BPMP)(OPr)(NO)2]2+ dimer with the triflate counter ion, 2a, studied here also generates N2O quantitatively upon addition of two equivalents of cobaltocene, as determined by IR gas headspace analysis, on the same timescale as our previously reported complex. This result is therefore consistent with our initial report, but this also poses the question of how the complex would perform if only one equivalent of reductant is added. To our surprise, not only does 2a produce quantitative amounts of N2O in the presence of only one equivalent of CoCp2, the rate of N2O production (as determined by IR gas headspace analysis) is identical to the case where two equivalents of reductant are used. This is illustrated in Figure S17. First of all, this result demonstrates that 2a actually uses a “semireduced” (one-electron reduced) mechanism for N2O generation. The second equivalent of reductant is not needed for N2O formation, and does not have an effect on the overall reaction rate. This further implies that the one-electron reduced


20

Page 21 of 48 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


{FeNO}7/{FeNO}8 dimer is in fact activated for N-N coupling, releasing N2O and forming a mixed-valent reaction product: [{FeNO}7]2 + e  {FeNO}7/{FeNO}8  “Fe(II)-O-Fe(III)” + N2O

(2)

This poses the question of how the second electron is consumed when two equivalents of reductant are added to 2a. In order to investigate this, we performed a redox titration where 2a was added stepwise to a solution of 110 M of CoCp2. The data in Figure S19 show that consumption of up to 0.55 equivalents of 2a (60 M) leads to the appearance of a new absorption band at 435nm, with an isosbestic point at 382 nm. Addition of further equivalents of 2a simply shows the absorption spectrum of 2a appear on top of the spectrum of the reduced reaction product. Therefore, complex 2a is in fact able to consume a total of two equivalents of reductant, as reported in our initial communication.25 However, since only one equivalent is needed for N2O generation according to equation (2), we speculated that the second equivalent might be consumed by reducing the initially formed, mixed-valent reaction product. Indeed, the cyclic voltammogram of the reaction product that results from the one-electron reduction of 2a shows a major feature at 0.983V vs. Fc/Fc+, which is slightly more positive than the reduction potential of 2a. Hence, the mixed-valent reaction product formed after N2O release is in fact responsible for the consumption of the second equivalent of CoCp2, but this second equivalent of CoCp2 is not needed for N2O generation. The detailed characterization of the reaction product of 2a is provided below. Since [Fe2(BPMP)(OPr)(NO)2](BPh4)2 suffers from poor solubility and inconsistent yields (see

above),

an

analogous

complex

with

a

weakly

coordinating

counter

ion,

[Fe2(BPMP)(OPr)(NO)2](BF4)2 (3), was synthesized in order to investigate a potential counter ion effect on the N2O generation pathway. The reduction products of 3 were similarly characterized by IR gas-headspace analysis. Importantly, N2O was again generated quantitatively with the


21


Page 22 of 48

addition of just one equivalent of reductant (Figure S18), confirming that the nature of the counter ion does not influence the reaction of [Fe2(BPMP)(OPr)(NO)2]2+ with reductant. Hence, we conclude that [Fe2(BPMP)(OPr)(NO)2](BPh4)2 also proceeds via the same semireduced mechanism as 2a and 3, with the second reductive equivalent being consumed by the mixed-valent product generated after N2O release: “Fe(II)-O-Fe(III)” + e  “Fe(II)/Fe(II)”

(3)

Correspondingly, formation of a diferrous reaction product after addition of two equivalents of CoCp2 to a solution of [Fe2(BPMP)(OPr)(NO)2](BPh4)2 had already been confirmed in our initial report.25 Kinetics of N2O Generation. In order to evaluate the kinetics of N2O generation and to identify putative intermediates (especially the reduced {FeNO}8 species), we monitored the reduction of 2a by IR spectroelectrochemistry (IR-SEC). Upon applying a potential of -1.3 V (versus Ag wire) to a solution of 2a, the N-O stretching band at 1765 cm-1 decreases in intensity, with a concomitant increase of the product band of N2O at 2222 cm-1 (the N-N stretch) and a shift of the bridging carboxylate band at 1541 cm-1 to higher energy, towards the pyridine band at 1605 cm-1 (Figure 5). Normalization of the intensities of the NO and N2O bands and plotting these intensities as a function of time (see Figure 5) shows no delay between the consumption of 2a and the formation of N2O, indicating that this reaction is very fast and that no intermediate {FeNO}8 species can be detected on the IR-SEC timescale. This result is surprising, considering that our previous experiments had shown that the reaction of 2a with CoCp2 is instantaneous, but that N2O generation takes about 1 minute, at least according to the IR gas-headspace detection method (see Figure S17). In retrospect, the time delay observed by gas headspace analysis just monitors the


22

Page 23 of 48 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


diffusion of N2O from solution into the gas headspace, and is not suitable to determine rates for very fast reactions. Moving forward, we then decided to probe the reaction of 2a with one equivalent of CoCp2 by stopped-flow IR spectroscopy. Upon mixing of 2a with one equivalent of CoCp2 at room temperature, the NO band has completely disappeared and the N2O band is completely formed in the first scan (Figure 6), revealing that the reduction proceeds remarkably fast, within the 156 ms dead time of the stopped-flow IR instrument. Based on this result, the first order rate constant kobs for N-N coupling and N2O release is estimated to be >102 s-1, assuming a >95% conversion occurred in the 156 ms time frame (this further assumes that the reduction of 2a is instantaneous, and that the following generation of N2O is rate limiting). In contrast, a similar non-heme diiron dinitrosyl complex with a cis NO orientation, [Fe2(Et-HPTB)(O2CPh)(NO)2](BF4)2, photoreduces NO to N2O with a first order rate constant of 0.2 min-1.26 Kinetic studies on the T. maratima FDP,18,19 which is primarily an O2 reductase in vivo, have shown an accumulation of the [{FeNO}7]2 intermediate, preceding a decay of this species to a diferric product and N2O over the course of ~120 s (enzyme concentration: 70M). Due to the rapid reaction of 2a with CoCp2 and the quick N2O release, we were not able to measure the concentration dependence of the rate constant for N2O generation from 2a. Such experiments would have allowed us to determine whether N2O formation is an intramolecular reaction or not. Alternatively, an isotope scrambling experiment can be used to obtain insight into this issue. We therefore reacted a 1:1 mixture of the natural abundance isotopes complex 2a and the 15NO-labeled complex 2b with one equivalent of CoCp2 at -80o C. In this regard, please note that 2a undergoes quite rapid NO exchange at room temperature (see Figure S20), so this experiment has to be performed at low temperature to avoid exchange of NO/15NO between 2a


23


Page 24 of 48

and 2b in the reaction mixture. As shown in Figure 7, this experiment leads to the clean generation of solely 14,14N2O and 15,15N2O and no mixed-isotope product, consistent with an intramolecular N-N coupling reaction. This result also agrees with the rapid kinetics of N2O generation observed by stopped-flow IR, which is facilitated by the fast and efficient coupling of two neighboring triplet NO ligands in the coplanar [{FeNO}7]2 core of 2a. In order to potentially trap intermediates of the NO reduction reaction of 2a, we additionally monitored the reaction of this complex with one equivalent CoCp2 at -80o C via a UVVis dip probe setup. However, the reaction still proceeds at this low temperature (see Figure 8), again indicating that the semireduced pathway has a very low activation barrier, in agreement with the stopped-flow IR results. Interestingly, the reduction product of 2a at -80o C exhibits an absorption band at 437 nm, which is different from the product formed at room temperature (RT), which shows a shoulder in the UV-Vis spectrum at 425 nm. However, upon warming up of the product solution from -80o C to room temperature, the 437 nm species transforms into the 425 nm form (see Figure 8). One possible explanation for this behavior is that we indeed trap an intermediate at low temperature. However, solution IR spectroscopy of the -80o C solution reveals the complete decay of the NO band with the complete formation of N2O (Figure 9). Hence, the 437 nm species forms after N-N coupling and N2O release, and therefore, corresponds to the (mixed-valent) product of the reaction. At room temperature, a somewhat different product complex forms, as indicated by the different UV-Vis properties. This is further analyzed in the next section. The generation of N2O at -80o C combined with the rapid kinetics of the reaction suggest that there is a remarkably low activation barrier for N-N bond formation in the semireduced pathway.


24

Page 25 of 48 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


Characterization of the Reduction Product. Since the two-electron reduction of NO to N2O by 2a only requires the addition of one equivalent of reductant (according to the semireduced mechanism), we anticipated the formation of a mixed-valent, oxo-bridged FeII-O2-FeIII species, originating from the remaining O2 after N2O release, as the reaction product. Whereas 1 and 2a are all EPR silent (Figure S15), the putative mixed-valent product is a non-integer spin compound, and hence, should be EPR active. As shown in Figure 10, this is indeed the case: the reaction product formed at -80o C shows an intense EPR signal with g-values of 1.96, 1.92, and 1.81, indicative of the formation of a mixed-valent Fe(II)/Fe(III) species with S = 1/2, where the spins of the hs FeII (S = 2) and the hs FeIII (S = 5/2) center are antiferromagnetically coupled. A fit of this spectrum in SpinCount, using two interacting spin centers with spin Hamiltonian (1), yields a J-coupling constant of -8 cm-1 (Figure S15). This is quite common for mixed-valent Fe(II)/Fe(III) dimers with bridging ligands, and antiferromagnetic coupling has been observed for several similar dinuclear model complexes,45,46 including [Fe2(BPMP)(OPr)2]2+,34,47 and also for the mixed-valent form of T. maritima FDP. In contrast, the product generated at room temperature (or warmed up to room temperature) is EPR silent as shown in Figure 10. While it is unclear why this species is EPR silent, this is in agreement with the UV-Vis data, which indicate that distinct species are generated at -80o C and at room temperature (see above). The fact that the room temperature product is EPR-silent suggests that a further oligomerization of the mixed-valent dimer formed at low temperature occurs, for example by formation of a tetramer (a dimer of dimers), which would then be EPR silent. Notably, it has also been reported in the literature that in some cases, similar mixed-valent species are EPR silent (at least under standard X-band EPR conditions).48,49 Furthermore, since phenols can act as non-innocent ligands, it is theoretically possible that the low temperature product corresponds to a diferrous complex with a bridging phenoxy radical,


25


Page 26 of 48

which then relaxes to a mixed-valent core with a bridging phenolate at room temperature. In order to test whether the oxidation of the phenol group is viable, the structurally analogous complex [Zn2(BPMP)(OPr)2](OTf) (5) was synthesized, which contains non-redox active Zn2+ centers. We also crystallized the analogous, acetate-bridged complex [Zn2(BPMP)(OAc)2](OTf), which exhibits a Zn-Zn distance of 3.384 Å, similar to the 3.360 Å Fe-Fe distance in the diferrous complex 4. Cyclic voltammetry of complex 5 reveals two semi-reversible waves at 0.625 and 0.850 V vs. Fc/Fc+ (Figure S14), both of which are significantly more positive (~0.4 and 0.6 V, respectively) than the FeIIFeII/FeIIFeIII couple in 1. Hence, formation of a bridging phenoxy radical is improbable. From these data, combined with the low temperature EPR data, we conclude that the product obtained at low temperature is in fact a mixed-valent type complex. We then attempted to crystallize the RT reaction product to obtain further insight into the exact nature of this species. Crystals were attained by vapor diffusion of hexanes into a solution of the RT product in tetrahydrofuran solution. However, X-ray crystallographic analysis of the crystals obtained in this way over four days identified them as the diferrous bis-propionate bridged complex [Fe2(BPMP)(OPr)2](OTf) (4; see Figure S24). This implies that the mixed-valent product is ultimately unstable at room temperature, and slowly disproportionates into the diferrous complex 4 and an unidentified ferric species. Indeed, cyclic voltammograms taken of the RT reduction product right after its formation show three different reversible waves; two small signals with E1/2 at 0.683V and -0.030V, and one large signal at -0.983V (all vs. Fc/Fc+, see Figure 11). The couples at 0.683V and -0.030V belong to the diferrous bis-propionate complex 4, as evident from the literature,34,50 and further confirmed by the CV of our independently synthesized complex 4 (see Figure S13). The main signal in the CV at -0.983V (vs. Fc/Fc+) must then belong to the initially formed room temperature product. The presence of only small amounts of 4 in the initial


26

Page 27 of 48 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


CV in Figure 11 further indicates that the disproportionation of the room temperature product is slow. This is further evidenced by the Mössbauer spectrum of the room temperature product, which contains the expected 49.6% ferric species with an isomer shift of 0.66 mm/s and quadrupole splitting of 1.34 mm/s, and 50.4% ferrous species with an isomer shift of 1.13 mm/s and quadrupole splitting of 2.93 mm/s (Figure S21). Importantly, the ferrous signal is unique from that of 4, so we attribute this Mössbauer spectrum to the room temperature product prior to disproportionation (species [X] in Scheme 2). The redox potential of the product species (-0.983V vs. Fc/Fc+) supports the conclusions drawn from our UV-Vis titration of CoCp2 with 2a (Figure S19), i.e. that 2a can consume two equivalents of CoCp2 in total by first reducing 2a by one electron, and then reducing the mixed-valent reaction product after N2O release by one additional electron (see above). This further implies that our previously reported

complex

[Fe2(BPMP)(OPr)(NO)2](BPh4)2 also follows the semireduced pathway for N2O generation. With these combined results in hand, a reaction scheme for the semireduction of 2a can be constructed, as shown in Scheme 2. Complex 2a can be reduced at -80o C and at RT with one equivalent of CoCp2, leading to intramolecular N-N coupling of the NO units to generate N2O. At -80o C, a mixed-valent species is formed as shown by EPR spectroscopy. Based on the stoichiometry of the reaction, we propose that this is the mixed-valent oxo-bridged dimer. This species then rearranges upon warming of the reaction mixture to RT, forming an EPR silent species. Given that the species trapped at -80o C is formed after N2O release (Figure 9), and that the mixed-valent diiron core cannot oxidize the ligand, we suspect that this may be due to changes in the coordination of the BPMP or bridging propionate ligand, or formation of a tetramer or oligomer when warming the reaction mixture from low to room temperature. Based on Mössbauer spectroscopy and CV, this is a distinct species (not random decomposition), and likely corresponds


27


Page 28 of 48

to another mixed-valent complex. Crystallization attempts of this RT product, as well as CV studies, imply that over time, this species disproportionates, forming the diferrous complex 4 and a ferric product. DISCUSSION: In

this

paper,

we

report

the

synthesis

of

the

FNOR

model

complex

[Fe2(BPMP)(OPr)(NO)2](OTf)2 (2a), using the dinucleating BPMP ligand scaffold. Compared to the complex [Fe2(BPMP)(OPr)(NO)2](BPh4)2 communicated previously,25 the triflate salt 2a is easier to prepare and can be obtained in high purity in a straight-forward and reliable synthesis. Independent of the counter ion, the complexes [Fe2(BPMP)(OPr)(NO)2]2+, which directly model the key dinitrosyl intermediate of FNORs, are stable dimers with two coplanar {FeNO}7 units, which is the proposed coordination mode of NO in the enzymes. This is based on crystallographic data on several FNORs, which consistently show two five-coordinate iron centers with adjacent coordination sites that are open for substrate binding.8,13,14 We have previously communicated the crystal structure of [Fe2(BPMP)(OPr)(NO)2](BPh4)2, and we reported that this complex can mediate quantitative N2O generation in the presence of two equivalents of reductant (CoCp2), but no further mechanistic details were provided.25 In this study, we use 2a to demonstrate that [Fe2(BPMP)(OPr)(NO)2]2+ does not utilize the superreduced mechanism for N2O generation, but instead, a new “semireduced” pathway is used where one-electron reduction of the dimer to the mixed-valent {FeNO}7/{FeNO}8 intermediate is sufficient to activate the complex for N-N coupling and N2O formation. This result is significant, as it suggests that superreduction of a [{FeNO}7]2 dimer is not necessary for NO reduction to N2O, and if FNORs should utilize a reductive activation approach to mediate N2O generation, then the Flavin cofactor would likely


28

Page 29 of 48 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


operate as a one-electron donor. Our results therefore provide a new mechanistic alternative for FNORs that has previously not been realized. A complication in the previous study that masked the true mechanism of N2O generation by [Fe2(BPMP)(OPr)(NO)2]2+ is the fact that this complex will consume two equivalents of reductant. However, our studies now demonstrate that in this case, the first equivalent is used for N2O generation, yielding an Fe(II)/Fe(III) mixed-valent product, which has a slightly more positive redox potential compared to the dinitrosyl complex, and hence, consumes the second equivalent of reductant to generate a diferrous product. Our data therefore indicate that [Fe2(BPMP)(OPr)(NO)2]2+ does not superreduce, even in the presence of two equivalents of CoCp2. Kinetic and isotope labelling studies on 2a further demonstrate that N-N coupling and N2O release occur extremely rapidly (kobs > 102 s-1) and that the reaction proceeds intramolecularly. The rate of the reaction is much faster than initially estimated, based on IR gas headspace analysis. The rate determined with this method is therefore an artifact that simply reflects the diffusion of the N2O product from solution into the gas phase. Furthermore, low temperature studies show that N2O is produced even at -80o C, consistent with an extremely low activation barrier for the reaction of

The Semireduced Mechanism for Nitric Oxide ... - ACS Publications

Recommend Documents