Binding of coenzyme B induces a major conformational change in the active site of methyl-coenzyme M reductase

Article abstract of DOI:10.1021/ja906367h

Methyl-coenzyme M reductase (MCR) is the key enzyme in methane formation by methanogenic Archaea. It converts the thioether methyl-coenzyme M and the thiol coenzyme B into methane and the heterodisulfide of coenzyme M and coenzyme B. The catalytic mechanism of MCR and the role of its prosthetic group, the nickel hydrocorphin coenzyme F₄₃₀, is still disputed, and no intermediates have been observed so far by fast spectroscopic techniques when the enzyme was incubated with the natural substrates. In the presence of the competitive inhibitor coenzyme M instead of methyl-coenzyme M, addition of coenzyme B to the active Ni(I) state MCR_red1 induces two new species called MCR _red2a and MCR_red2r which have been characterized by pulse EPR spectroscopy. Here we show that the two MCR_red2 signals can also be induced by the S-methyl- and the S-trifluoromethyl analogs of coenzyme B. ¹⁹F-ENDOR data for MCR_red2a and MCR_red2r induced by S-CF₃-coenzyme B show that, upon binding of the coenzyme B analog, the end of the 7-thioheptanoyl chain of coenzyme B moves closer to the nickel center of F₄₃₀ by more than 2 A as compared to its position in both, the Ni(I) MCR_red1 form and the X-ray structure of the inactive Ni(II) MCR_ox1-silent form. The finding that the protein is able to undergo a conformational change upon binding of the second substrate helps to explain the dramatic change in the coordination environment induced in the transition from MCR_red1 to MCR_red2 forms and opens the possibility that nickel coordination geometries other than square planar, tetragonal pyramidal, or elongated octahedral might occur in intermediates of the catalytic cycle.

Full text of DOI:10.1021/ja906367h

Published on Web 12/16/2009
Binding of Coenzyme B Induces a Major Conformational
Change in the Active Site of Methyl-Coenzyme M Reductase
Sieglinde Ebner,† Bernhard Jaun,*,† Meike Goenrich,‡ Rudolf K. Thauer,‡ and
Jeffrey Harmer*,§
Department of Chemistry, Centre for AdVanced Electron Spin Resonance, UniVersity of Oxford,
South Parks Road, Oxford OX1 3QR, United Kingdom, Laboratory of Organic Chemistry, ETH
Zurich, 8093 Zurich, Switzerland, and Department of Biochemistry, Max Planck Institute for
Terrestrial Microbiology, Karl-Von-Frisch Strasse, 35043 Marburg, Germany
Received July 29, 2009; E-mail: jeffrey.harmer@chem.ox.ac.uk; jaun@org.chem.ethz.ch
Abstract: Methyl-coenzyme M reductase (MCR) is the key enzyme in methane formation by methanogenic
Archaea. It converts the thioether methyl-coenzyme M and the thiol coenzyme B into methane and the
heterodisulfide of coenzyme M and coenzyme B. The catalytic mechanism of MCR and the role of its
prosthetic group, the nickel hydrocorphin coenzyme F430, is still disputed, and no intermediates have been
observed so far by fast spectroscopic techniques when the enzyme was incubated with the natural
substrates. In the presence of the competitive inhibitor coenzyme M instead of methyl-coenzyme M, addition
of coenzyme B to the active Ni(I) state MCRred1 induces two new species called MCRred2a and MCRred2r
which have been characterized by pulse EPR spectroscopy. Here we show that the two MCRred2 signals
can also be induced by the S-methyl- and the S-trifluoromethyl analogs of coenzyme B. 19F-ENDOR data
for MCRred2a and MCRred2r induced by S-CF3-coenzyme B show that, upon binding of the coenzyme B
analog, the end of the 7-thioheptanoyl chain of coenzyme B moves closer to the nickel center of F430 by
more than 2 Å as compared to its position in both, the Ni(I) MCRred1 form and the X-ray structure of the
inactive Ni(II) MCRox1-silent form. The finding that the protein is able to undergo a conformational change
upon binding of the second substrate helps to explain the dramatic change in the coordination environment
induced in the transition from MCRred1 to MCRred2 forms and opens the possibility that nickel coordination
geometries other than square planar, tetragonal pyramidal, or elongated octahedral might occur in
intermediates of the catalytic cycle.
Introduction
methyl-coenzyme M has to diffuse to reach F430. The channel
is designed such that the terminal SH-group of the heptanoyl
Methyl-coenzyme M reductase (MCR) catalyzes the reduction
of methyl-coenzyme M (CH3-S-CoM, 2-(methylthio)ethane
sulfonate) with coenzyme B (HS-CoB, 7-thioheptanoyl-thre-
oninephosphate) to methane and CoM-S-S-CoB, which is
the methane forming step in the energy metabolism of all
methanogenic archaea.1,2 MCR harbors the nickel porphyrinoid
F430, which is in the Ni(I) oxidation state when the enzyme is
active.
arm of coenzyme B remains 8 Å from the nickel. CH3-S-CoM
binding increases the affinity for HS-CoB binding (ordered Bi
Bi mechanism), thus ensuring CH3-S-CoM enters the channel
first. The available crystal structures differ in the species that
are axially coordinated to the nickel center of F430. All structures
have the carboxamide oxygen of the R′-glutamine residue 147
(Scheme 1) coordinated axially from the distal face to the nickel
ion. In MCRsilent, a sulfonate oxygen of CoM-S-S-CoB is
CH3-S-CoM + HS-CoB f CH4 +
CoM-S-S-CoB ∆G0′ ) -30 kJ mol-1 (1)
(3) Ermler, U.; Grabarse, W.; Shima, S.; Goubeaud, M.; Thauer, R. K.
Science (Wash.) 1997, 278 (5342), 1457.
(4) Grabarse, W.; Mahlert, F.; Duin, E. C.; Goubeaud, M.; Shima, S.;
Thauer, R. K.; Lamzin, V.; Ermler, U. J. Mol. Biol. 2001, 309 (1),
315.
Several crystal structures of MCR in an inactive state (NiII,
d8, S ) 1) have been determined.3-6 The enzyme has an
interlinked R2ꢀ2γ2 subunit structure and contains two molecules
of the nickel porphyrinoid cofactor F430 (Scheme 1) that are
approximately 50 Å apart. Coenzyme F430, the prosthetic group
of MCR,7-10 is buried deep within the protein and is accessible
from the exterior only via a 50 Å long channel, through which
(5) Grabarse, W.; Mahlert, F.; Shima, S.; Thauer, R. K.; Ermler, U. J.
Mol. Biol. 2000, 303 (2), 329.
(6) Shima, S.; Goubeaud, M.; Vinzenz, D.; Thauer, R. K.; Ermler, U.
J. Biochem. (Tokyo) 1997, 121 (5), 829.
(7) Ellefson, W. L.; Whitman, W. B.; Wolfe, R. S. Proc. Natl. Acad. Sci.
U.S.A. 1982, 79 (12), 3707.
(8) Faerber, G.; Keller, W.; Kratky, C.; Jaun, B.; Pfalz, A.; Spinner, C.;
Kobelt, A.; Eschenmoser, A. HelV. Chim. Acta 1991, 74 (4), 697.
(9) Livingston, D. A.; Pfaltz, A.; Schreiber, J.; Eschenmoser, A.; Ankel-
fuchs, D.; Moll, J.; Jaenchen, R.; Thauer, R. K. HelV. Chim. Acta
1984, 67 (1), 334.
† ETH Zurich.
‡ Max Planck Institute for Terrestrial Microbiology.
§ University of Oxford.
(1) Thauer, R. K. Microbiology 1998, 144 (9), 2377.
(10) Pfaltz, A.; Jaun, B.; Faessler, A.; Eschenmoser, A.; Jaenchen, R.; Gilles,
H. H.; Diekert, G.; Thauer, R. K. HelV. Chim. Acta 1982, 65 (3), 828.
(2) Jaun, B.; Thauer, R. K. Met. Ions Life Sci. 2007, 2, 323.
9
10.1021/ja906367h  2010 American Chemical Society
J. AM. CHEM. SOC. 2010, 132, 567–575 567

A R T I C L E S
Ebner et al.
coordinated to the NiII ion, whereas in MCRox1-silent and
MCRred2-silent, the thiol(ate) sulfur of (H)S-CoM is coordi-
nated from the proximal face. Attempts to obtain a crystal
structure of active MCR have failed until now, probably due
to the low redox potential of the Ni(II)/Ni(I) couple in F430
which is below -600 mV and thus 200 mV below that of
the hydrogen electrode at pH 7.
Scheme 1. Coenzyme F430, the Prosthetic Group of
Methyl-Coenzyme M Reductasea
The catalytic mechanism of the reduction (eq 1) at the nickel
center is disputed,1,4,11,12 but all mechanistic proposals that
appeared in the literature after the X-ray structures explicitly
or tacitly assume that the terminal SH-group of the heptanoyl
arm of coenzyme B remains 8 Å from the nickel as indicated
by the crystal structure of the inactive enzyme containing NiIIF430
(Scheme 1). In essence there are two proposed mechanisms
which mainly differ in the nature of the initial cleavage of the
sulfur-carbon bond of CH3-S-CoM. In mechanism “A”,
proposed by Pelmenschikov et al.11,12 on the basis of density
functional theory (DFT) calculations, the NiI center is assumed
to attack the thioether sulfur of CH3-S-CoM, releasing as
intermediates a methyl radical and the thiolate complex
CoM-S-NiIIF430. Since HS-CoB is assumed fixed at 8 Å from
NiF430, the generated methyl radical from CH3-S-CoM must
move ca. 4 Å to abstract a hydrogen atom from HS-CoB, a
controversial assumption. In mechanism “B”, the NiI center
initially acts as a nucleophile attacking methyl-coenzyme M at
the carbon of the CH3-S group, generating a CH3-NiIIIF430
intermediate and HS-CoM.2,4,13 A CH3-NiIIIF430 species has
been shown to form when CH3Br (or CH3I) is added to active
MCR,14-16 and a Ni-alkyl bond is formed when the irreversible
inhibitor BPS (BrCH2CH2CH2SO3-) reacts with the active
enzyme.17 Theoretical hybrid DFT calculations on a CH3-
NiIIIF430 model complex indicate that it is stable and has the
a The orientation of the axis g3 is shown. Shown are natural substrates
CH3-S-CoM and HS-CoB, and inhibitors HS-CoM (ꢀ position indi-
cated), CH3-S-CoB, and CF3-S-CoB.
(“m” for methyl coenzyme M), and in the presence of HS-CoM
it is denoted MCRred1c (“c” for coenzyme M). By addition of
the natural substrates CH3-S-CoM and HS-CoB to active
MCR (red1a) it has not been possible to trap or observe any
intermediates of reaction eq 1. Substrate analogs are thus used
to explore structural changes upon binding which are important
in understanding the catalytic mechanism. One important set is
MCRred1a + HS-CoM + HS-CoB, where the addition of
HS-CoB to MCRred1a + HS-CoM induces the conversion of
the MCRred1c signal into the MCRred1cc signal (which is very
similar to MCRred1c) and up to 50% into two EPR species called
MCRred2a and MCRred2r after their axial and rhombic EPR
spectra, respectively.27,28 The MCRred1 type signal generated
from MCRred1a in the presence of HS-CoM and HS-CoB
differs slightly from the species generated only in the presence
of HS-CoM and is therefore in the following addressed to as
MCRred1cc (“cc” for coenzyme M and coenzyme B). MCRred2a
is assigned to a NiIII hydride complex, whereas MCRred2r has a
Ni-S coordination from HS-CoM, rhombic principal g-values
which are unusually high, and a very asymmetric spin density
distribution on the four hydropyrrolic nitrogens of the macrocyle
14,18,19
1
2
2
2
2
2
unusual ground-state configuration (dxy) (dxz) (dyz) (dx -y ) .
Active MCR exhibits an axial EPR spectrum characteristic
9
1
of a d, S ) /2, NiI complex with the unpaired electron in the
20-26
2
2
3dx -y orbital
and is designated MCRred1. In the absence
of substrates the species is designated MCRred1a (“a” for
absence), in the presence of CH3-S-CoM it is called MCRred1m
(11) Pelmenschikov, V.; Blomberg, M. R. A.; Siegbahn, P. E. M.; Crabtree,
R. H. J. Am. Chem. Soc. 2002, 124 (15), 4039.
(12) Pelmenschikov, V.; Siegbahn, P. E. M. J. Biol. Inorg. Chem. 2003, 8
(6), 653.
(13) Ermler, U. J. Chem. Soc., Dalton 2005, (21), 3451.
(14) Yang, N.; Reiher, M.; Wang, M.; Harmer, J.; Duin, E. C. J. Am. Chem.
Soc. 2007, 129 (36), 11028.
(15) Dey, M.; Kunz, R. C.; Lyons, D. M.; Ragsdale, S. W. Biochemistry
2007, 46 (42), 11969.
(16) Dey, M.; Telser, J.; Kunz, R. C.; Lees, N. S.; Ragsdale, S. W.;
Hoffman, B. M. J. Am. Chem. Soc. 2007, 129 (36), 11030.
(17) Hinderberger, D.; Piskorski, R. P.; Goenrich, M.; Thauer, R. K.;
Schweiger, A.; Harmer, J.; Jaun, B. Angew. Chem., Int. Ed. 2006, 45
(22), 3602.
F430.
27,29 This loss of near-equivalence of the four nitrogens in
the tetrapyrrole macrocycle together with the pronounced
rhombicity of the MCRred2r signal must be due to a strong
electronic and/or geometric distortion of the coordination sphere
of the nickel center. It has been proposed by us that binding of
HS-CoB results in a conformational change of the protein that
induces or is coupled to the geometric changes observed when
(18) Wondimagegn, T.; Ghosh, A. J. Am. Chem. Soc. 2001, 123 (7), 1543.
(19) Sarangi, R.; Dey, M.; Ragsdale, S. W. Biochemistry 2009, 48 (14),
3146.
(20) Goubeaud, M.; Schreiner, G.; Thauer, R. K. Eur. J. Biochem. 1997,
243 (1/2), 110.
(21) Jaun, B.; Pfaltz, A. J. Chem. Soc., Chem. Commun. 1986, (17), 1327.
(22) Mahlert, F.; Bauer, C.; Jaun, B.; Thauer, R. K.; Duin, E. C. J. Biol.
Inorg. Chem. 2002, 7 (4-5), 500.
(23) Mahlert, F.; Grabarse, W.; Kahnt, J.; Thauer, R. K.; Duin, E. C. J. Biol.
Inorg. Chem. 2002, 7 (1-2), 101.
(27) Harmer, J.; Finazzo, C.; Piskorski, R.; Ebner, S.; Duin, E. C.; Goenrich,
M.; Thauer, R. K.; Reiher, M.; Schweiger, A.; Hinderberger, D.; Jaun,
B. J. Am. Chem. Soc. 2008, 130 (33), 10907.
(24) Rospert, S.; Boecher, R.; Albracht, S. P. J.; Thauer, R. K. FEBS Lett.
1991, 291 (2), 371.
(25) Rospert, S.; Voges, M.; Berkessel, A.; Albracht, S. P. J.; Thauer, R. K.
Eur. J. Biochem. 1992, 210 (1), 101.
(28) Kern, D. I.; Goenrich, M.; Jaun, B.; Thauer, R. K.; Harmer, J.;
Hinderberger, D. J. Biol. Inorg. Chem. 2007, 12 (8), 1097.
(29) Finazzo, C.; Harmer, J.; Jaun, B.; Duin, E. C.; Mahlert, F.; Thauer,
R. K.; Van Doorslaer, S.; Schweiger, A. J. Biol. Inorg. Chem. 2003,
8 (5), 586.
(26) Telser, J.; Fann, Y. C.; Renner, M. W.; Fajer, J.; Wang, S. K.; Zhang,
H.; Scott, R. A.; Hoffman, B. M. J. Am. Chem. Soc. 1997, 119 (4),
733.
9
568 J. AM. CHEM. SOC. VOL. 132, NO. 2, 2010

Active Site of Methyl-Coenzyme M Reductase
A R T I C L E S
27
was stirred for 2 h at 0 °C and 1 h at room temperature. After
adding 2 mL of water, the reaction mixture was stirred at room
temperature for 24 h, then poured on 25 mL of a 10% aqueous
hydrochloride solution, extracted with diethylether, and dried to
give 0.748 g (98%) of a yellow oil. H NMR (CDCl3; 300 MHz)
δ 1.38 (m, 4H), 1.61 (m, 4H), 2.08 (s, 3H), 2.34 (t, 2H), 2,47 (t,
2H) ppm.
going from the axial MCRred1/2a species to MCRred2r
.
Our
principal aim in this work is to determine what structural
changes occur when HS-CoB binds to active MCR.
Interestingly, the MCRred2 signals can also be induced when
the terminal SH group of CoB is replaced by a methylthioether
group (CH3-S-CoB).27 In fact, this combination of substrate
analogs has a very similar overall size as the natural substrates
CH3-S-CoM + HS-CoB (the methyl group of CH3-S-
CoM and thiol group of HS-CoB have been swapped). This
observation prompted us to speculate that a CF3-S- group,
despite being somewhat larger than CH3-S, might still be able
to induce MCRred2. Such a fluorine label at the end of the
heptanoyl tail of coenzyme B would allow geometric changes
to be monitored through measurement of the 19F (I ) 1/2)
hyperfine interactions to determine the position of the CF3 group
of bound CF3-S-CoB relative to the nickel. If the conformation
in the active NiI states is similar to that found in the X-ray
structure of inactive NiIIF430, then the distance between the three
F atoms and the nickel ion will be 6.2-7.7 Å. In this paper we
first show that, like HS-CoB or CH3-S-CoB, CF3-S-CoB
induces the two red2 species, and then determine the positions
of CoB relative to NiF430, for red1cc/red2a/red2r, using measured
19F hyperfine couplings.
1
7-(Methylthio)heptanoyl-N-hydroxysuccinimide Ester30 (4).
3 (0.735 g, 4.2 mmol) and 0.503 g of N-hydroxysuccinimide (4.4
mmol) were dissolved in 25 mL anhydrous dioxane under nitrogen.
After addition of 858 mg of dicyclohexylcarbodiimide (4.2 mmol),
the reaction mixture was stirred for 16 h at room temperature. The
precipitated dicyclohexylurea was removed by filtration and washed
with dioxane (3 × 10 mL) followed by isopropanol (2 × 10 mL).
The combined filtrates were dried to give a yellow oil which was
purified by flash column chromatography (rf ) 0.2; silica gel; dietlyl
ether/ hexane, 2:1, v/v). The combined fractions containing pure 4
were dried to give 0.667 g (62%) of a viscous yellowish oil which
1
solidified at -18 °C. H NMR (CDCl3; 200 MHz) δ 1.35-1.43
(m, 4H), 1.54-1.60 (m, 2H), 1.68-1.72 (m, 2H), 2.05 (s, 3H),
2.45 (t, 2H), 2.57 (t, 2H), 2.80 (s, 4H). 13C NMR (CDCl3; 200
MHz) δ 15.4, 24.3, 25.5 (2C), 28.1, 28.2, 28.7, 30.7, 33.9, 168.6,
169.2 (2C).
(+)-(2S,3R)-N-[7-(Methylthio)heptanoyl]-O-phospho-L-threo-
nine30 (1) [CH3S-CoB (NH4+ Form)]. A solution of 495 mg of
O-phospho-L-threonine (2.48 mmol) and 849 µL of diisopropyl-
ethylamine (4.96 mmol) in 5 mL water was added to a solution of
4 (612 mg; 2.23 mmol) in 40 mL THF. Addition of acetonitrile
(14 mL) resulted in a homogeneous colorless reaction mixture which
was stirred at room temperature under nitrogen for 18 h. The dried
residue was purified by RP-HPLC, then desalted and further purified
by applying to an XAD-2-polystyrene column, dried, dissolved in
approximately 2 mL of water, treated with concentrated NH3(aq)
to pH 7-8 and lyophylized to give 198 mg of pure 1 (ammonium
salt; 0.53 mmol; 23%). 1H NMR (CD3OD; 400 MHz) δ 1.31 (d, J
) 2.89, 2H), 1.35-1.45 (m, 4H), 1,56-1.66 (m, 4H), 2.06 (s, 3H),
2.29-2.33 (m, 2H), 2.48 (t, J ) 7.29, 2H), 4.30 (d, J ) 2.65, 1H),
4.72 (m, 1H). 13C NMR (CD3OD; 400 MHz) δ 15.4, 19.6, 26.9,
30.1, 35.0, 37.1, 61.1, 73.8, 172.2, 176.0. 31P NMR (CD3OD; 300
MHz): δ 1.32 (d, J ) 8.77). ESI-MS: m/z 388.1 (6), 139.1 (6),
193.8 (6), 95.2 (9), 357.1 (15), 177.6 (51), 356.1 (100).
Materials and Methods
General. THF was destilled over K under nitrogen, DMF was
freshly destilled under vacuum at ca. 10 mbar and 40 °C over a
fractionating column (110 cm) packed with glass beads. The reflux
ratio was 10:1, and the middle fraction (30%) of the destillate was
used. Amberlite IR-120 (H+ form, 16-45 mesh) was conditioned
by washing with ethanol until the eluent was colorless, stirring in
20% H2SO4 for 1 h, and washed with water until the eluent was
neutral. XAD-2-polystyrene was purchased from Serva Electro-
phoresis, Germany. All other routine chemicals were obtained from
Fluka, Aldrich, J.T. Baker and Merck and were used without further
purification. Acid-base extraction was performed with concentrated
HCl(aq) and saturated aqueous bicarbonate solution. RP-HPLC was
performed with an Atlantis dC18 5 µm column, 19 mm × 50 mm
with the following gradient program: H2O/acetonitrile (4:1) f 100%
acetonitrile in 70 min. Detection was performed on Alugram SIL
Synthesis of (+)-(2S,3R)-N-[7-(Trifluoromethylthio)heptanoyl]-
O-phospho-L-threonine (5) [CF3S-CoB (NH4+ Form)]. 7-Mercap-
toheptanoic Acid (6). According to a modified procedure of Noll
et al.,32 a solution of 3.5 g of 2 (16.7 mmol) and 6.37 g of thiourea
(83.7 mmol) in 100 mL acetone was refluxed under nitrogen at 70
°C for 17 h. The cooled (room temperature) and dried residue was
added to a degassed (Ar) solution of 6.56 g of KOH (116.9 mmol)
in 140 mL of ethanol, refluxed for another 4 h at 90 °C under
nitrogen, and cooled to room temperature. Acid-base extraction
G/UV254 TLC plates (silica gel 60 with fluorescent indicator UV254
)
in n-butanol/H2O/acetic acid, 2:1:1 (v/v), staining with Mostain-
solution (20 g (NH4)6Mo7O24 ·4H2O in 10% H2SO4, v/v (400 mL)
and 0.4 g CeSO4 in H2SO4 conc. (4 mL)). For desalting of 5 an
Oasis HLB 35 cm3 cartridge was conditioned first with MeOH,
second with 1 N HCl. Then, the product was applied, desalted with
5 loads of water, and eluted with 100% MeOH. For the synthesis
of coenzyme B derivatives, appropriately modified procedures of
Ellerman et al.30 were followed.
1
was performed to give 2.61 g of pure 6 (16.08 mmol; 96%). H
Synthesis of (+)-(2S,3R)-N-[7-(Methylthio)heptanoyl]-O-phos-
pho-L-threonine (1) [CH3-S-CoB (NH4+ Form)]. 7-Bromohep-
tanoic Acid (2).31 Ethyl 7-bromo heptanoate (4.1 mL, 21.04 mmol)
and aqueous hydrobromic acid (48%; 14.4 mL) were stirred at 130
°C (oil bath temperature) for 5 h. The cooled reaction mixture was
diluted with water, extracted with dichloromethane and dried. Then
acid-base extraction was perfomed to give 3.80 g (86%) of a white
solid. 1H NMR (CDCl3, 300 MHz) δ 1.37-1.47 (m, 4H), 1.63-1.68
(m, 2H), 1.84-1.89 (m, 2H), 2.37 (t, 2H). 13C NMR (CDCl3, 300
MHz) δ 24.53, 27.88, 28.57, 32.60, 33.86, 34.00, 179.81.
7-(Methylthio)heptanoic Acid30 (3). 2 (0.9 g, 4.30 mmol) was
dissolved in 12 mL of ethanol under nitrogen. First, 0.43 mL of a
10 M sodium hydroxide solution, then 0.313 g of sodium meth-
anethiolate (4.47 mmol) were added slowly at 0 °C. The suspension
NMR (CDCl3; 200 MHz) δ 1.29-1.43 (m, 4H), 1.59-1.69 (m,
4H), 2.36 (t, 2H), 2.47-2.58 (q, 2H). 13C NMR (CDCl3; 200 MHz)
δ 24.5, 27.9, 28.4, 33.7, 33.8, 180.0.
7-(Trifluoromethylthio)heptanoic Acid (7). According to a
33
procedure of Soloshonok et al. for the synthesis of S-trifluoro-
methyl-containing amino acids, 7 was synthesized by gradually
introducing 90 mL of liquid ammonia (1.6 equiv) to 2 g of 6 (12.33
mmol) at -45 °C, and adding approximately 2 mL of trifluorom-
ethyliodide (1.6 equiv, by cold pipetting). The reaction mixture was
kept (stirring was not possible, since 7 solidified in the flask and
the stirrer could not be moved) under UV irradiation (Hg lamp)
for 4 h at -45 °C. The negative result of the Ellmann-Test34 showed
(32) Noll, K. M.; Donnelly, M. I.; Wolfe, R. S. J. Biol. Chem. 1987, 262
(2), 513.
(30) Ellermann, J.; Kobelt, A.; Pfaltz, A.; Thauer, R. K. FEBS Lett. 1987,
220 (2), 358.
(31) Rao, A. V. R.; Reddy, S. P. Synth. Commun. 1986, 16 (10), 1149.
(33) Soloshonok, V.; Kukhar, V.; Pustovit, Y.; Nazaretian, V. Synlett 1992,
(8), 657.
(34) Ellman, G. L. Arch. Biochem. Biophys. 1959, 82 (1), 70.
9
J. AM. CHEM. SOC. VOL. 132, NO. 2, 2010 569

A R T I C L E S
Ebner et al.
that the thiol had been fully converted. After evaporation of the
excess of ammonia, the residue was dissolved in 5 mL of water,
acidified with 2 M HCl, extracted with diethyl ether, and dried.
Flash column chromatography (silica gel; dichloromethane/ ethyl
acetate/ acetic acid, 98:10:2, v/v) gave 2.30 g of pure 7 (9.97 mmol;
81%). 1H NMR (CDCl3; 300 MHz) δ 1.36-1.44 (m, 4H),
1.61-1.73 (m, 4H), 2.36 (t, 2H), 2.87 (t, 2H), 8.5 (s, b). 13C NMR
(CDCl3; 300 MHz, τ ) 4 s) δ 24.3, 28.0, 28.3, 29.2, 29.6-29.8
(q, J ) 2.3), 33.8, 134.2 (q, J ) 305.7), 179.7.
using 10 mM copper perchlorate as a standard (10 mM CuSO4; 2
M NaClO4; 10 mM HCl). All signal intensities are expressed as
spins per mol F430. These EPR data were used exclusively for
sample quality control, all experimental spectra shown in this paper
are described below.
EPR Spectroscopy. Measurements at X- and W-band were
carried out on a Bruker Elex 680 spectrometer, and at Q-band (35.3
GHz) on a home-built instrument,35 both of which were equipped
with a Helium gas-flow cryostat from Oxford Inc. The field-swept
frozen-solution (15 K) W-band EPR spectra were recorded by
integrating over the echo created with the mw pulse sequence
π/2-τ-π-τ-echo, with mw pulse lengths tπ/2 ) 100 ns, tπ ) 200
ns, and an interpulse delay of τ ) 700 ns. The first derivative of
this spectrum was calculated numerically. The field was calibrated
using the two central lines from a CaO sample containing
manganese ions. The 19F Mims-ENDOR spectra were measured at
15 K with the mw pulse sequence π/2-τ-π/2-Τ-π/2-τ-echo,
with mw pulses of length tπ/2 ) 32 ns, τ ) 200-500 ns, and a shot
repetition time of 5 ms. A radio frequency pulse of length 23 µs
and variable frequency νENDOR was applied during time T. The
7-(Trifluoromethylthio)heptanoyl-N-hydroxysuccinimide Ester30
(8). According to the synthesis of the analog ester 4, 744 mg of 7
(3.23 mmol) and 409 mg of N-hydroxysuccinimide (3.55 mmol)
were dissolved in 20 mL anhydrous dioxane under nitrogen. After
addition of 667 mg of dicyclohexylcarbodiimide (3.23 mmol), the
reaction mixture was stirred for 24 h at room temperature. The
precipitated dicyclohexylurea was removed by filtration and washed
with dioxane (3 × 10 mL) followed by isopropanol (2 × 10 mL).
The combined filtrates were dried and 136 mg of the crude product
was purified by flash column chromatography (silica gel; dietlyl
ether/ hexane, 2:1, v/v). The combined fractions containing pure 8
were dried to give 640 mg of 8 (1.95 mmol; 61%). 1H NMR
(CDCl3; 300 MHz) δ 1.45 (m, 4H), 1.74 (m, 4H), 2.61 (t, 2H),
2.84 (s, 4H), 2.88 (t, 2H).
1
Q-band H Davies-ENDOR spectra were measured at 20 K with
the mw pulse sequence π-T-π/2-τ-π-τ-echo, with mw pulses
of length tπ/2 ) 28 ns and tπ ) 56 ns, and τ ) 160 ns. A radio
frequency pulse of length 8.5 µs and variable frequency νENDOR
was applied during time T. X-band HYSCORE experiments
at 25 K employed the pulse sequence π/2-τ-π/2-t1-π-t2-π/
2-τ-echo with mw pulses of lengths tπ/2 ) tπ ) 16 ns, τ ) 120
ns, and starting times t1,0 ) t2,0 ) 96 ns with increments ∆t ) 16
ns (data matrix 300 × 300). An eight-step phase cycle was used to
remove unwanted echoes. The HYSCORE data were processed with
MATLAB 7.0 (The MathWorks, Inc.). The time traces were
baseline corrected with an exponential, apodized with a Gaussian
window and zero filled. After a two-dimensional Fourier transfor-
mation absolute-value spectra were calculated. Spectra recorded
with different τ values were added to eliminate τ-dependent blind
spots. The EPR and Davies-ENDOR spectra were simulated with
the program EasySpin.36 HYSCORE spectra were simulated with
a program written in-house.37
(+)-(2S,3R)-N-[7-(Trifluoromethylthio)heptanoyl]-O-phospho-
L-threonine30 (5) [CF3S-CoB (NH4+ Form)]. A solution of 427
mg O-phospho-L-threonine (2.15 mmol) and 668 µL diisopropyl-
ethylamine (3.90 mmol) in 5 mL water was added to a solution of
8 (640 mg; 1.95 mmol) in 20 mL THF. Addition of acetonitrile
(10 mL) resulted in a homogeneous colorless reaction mixture which
was stirred at room temperature under nitrogen for 36 h. The dried
residue was purified by applying to an XAD-2-polystyrene column,
dried, and desalted on an Oasis HLB 35 cm3 cartridge, dissolved
in 2 mL of water, treated with concentrated NH3(aq) to pH 7-8
and lyophylized to give 98 mg of pure 5 (ammonium salt; 0.20
1
mmol; 10%). H NMR (CD3OD; 400 MHz) δ 1.32 (d, J ) 6.34,
3H), 1.37-1.48 (m, 4H), 1.62-1.73 (m, 4H), 2.32 (t, J ) 7.51,
2H), 2.94 (t, J ) 7.39, 2H), 4.31 (d, J ) 3.54, 1H), 4.70-4.72 (m,
1H). 13C NMR (CD3OD; 400 MHz) δ 19.6, 26.7, 29.2, 29.7, 30.6,
30.8, 36.9, 60.5, 73.2, 128.4-137.5 (q, J ) 304.4), 175.7, 176.2.
31P NMR (CD3OD; 300 MHz): δ 0.87 (d, b, J ) 8.71). 19F NMR
(CD3OD; 300 MHz): δ -61.7 (s). ESI-MS: m/z 410 (100), 312
(27), 205 (23), 411 (15), 330 (14), 216 (11).
Results
First EPR data are presented which show that the sample
preparation [CF3-S-CoB]-MCR, like [HS-CoB]-MCR,
produces the three components red1c/red2a/red2r. Figure 1
presents W-band echo-detected EPR spectra for the com-
Protein Purification and Sample Preparation. Methanothermo-
bacter marburgensis is the strain deposited under DSM 2133 in
the Deutsche Sammlung von Mikroorganismen and Zellkulturen
(Braunschweig). Methyl-coenzyme M reductase isoenzyme I from
Methanothermobacter marburgensis was purified as previously
described.23 All the buffers used during purification contain 10 mM
HS-CoM. Therefore the obtained enzyme was in the MCRred1c
form. The spin concentration per mole of F430 was approximately
0.8-0.9. The protein concentration was determined by measuring
the absorbance difference of oxidized enzyme (MCRsilent) at 420
nm using ε ) 44 000 M-1 cm-1 for a molecular mass of 280 000
Da. For sample preparation the enzyme was concentrated exten-
sively using an Amicon centrifugation cell with a 100 kDa cutoff
(Millipore, Bedford MA). The two red2 species were induced by
addition of CF3-S-CoB to a concentration of 20 mM. EPR
samples were frozen in liquid nitrogen for shipping and storage.
EPR Spectroscopy (Sample Control). As a control of the
sample quality and concentration, 1-10 diluted samples (0.35 mL)
were analyzed for EPR spectra at 77 K in 0.3 cm (inner diameter)
quartz tubes with 95% N2/5% H2 as gas phase and closed with a
closed-off rubber tube. The samples contained approximately 9.6
mg MCR (35 nmol) in 10 Tris/HCl pH 7.6. CW EPR spectra at
X-band were recorded with a Bruker EMX-6/1 EPR spectrometer
composed of an EMX 1/3 console, an ER 041 X6 bridge with built-
in ER-0410-116 microwave (mw) frequency counter, an ER-070
magnet and an ER-4102st standard universal rectangular cavity.
Spin quantifications were carried out under nonsaturating conditions
plexes (A) MCRred1c (red1a + HS-CoM), (B) MCRred1c
+
HS-CoB, and (C) MCRred1c + CF3-S-CoB. To simulate
the MCRred1c spectrum, a two component model was required
as can be appreciated by inspection of the g3 feature at the
low field end of the spectrum where there are two partially
resolved peaks. The three features marked with a “*” belong
to the species MCRox1.
38 Addition of either HS-CoB (Figure
1B) or CF3-S-CoB (Figure 1C) to MCRred1c results in a
dramatic spectral change that in both cases can be simulated
by three components, a MCRred1cc, virtually unchanged from
MCRred1c, and two MCRred2 species labeled MCRred2r (rhombic
spectrum) and MCRred2a (axial spectrum).27,28 The g-values
and relative intensities of the three species are given in Table
1 (the related preparation with CH3-S-CoB is included for
reference). Clearly the induction of the two MCRred2 species
(35) Gromov, I.; Shane, J.; Forrer, J.; Rakhmatoullin, R.; Rozentzwaig,
Y.; Schweiger, A. J. Magn. Reson. 2001, 149 (2), 196.
(36) Stoll, S.; Schweiger, A. J. Magn. Reson. 2006, 178 (1), 42.
(37) Madi, Z. L.; Van Doorslaer, S.; Schweiger, A. J. Magn. Reson. 2002,
154 (2), 181.
(38) Harmer, J.; Finazzo, C.; Piskorski, R.; Bauer, C.; Jaun, B.; Duin, E. C.;
Goenrich, M.; Thauer, R. K.; Van Doorslaer, S.; Schweiger, A. J. Am.
Chem. Soc. 2005, 127 (50), 17744.
9
570 J. AM. CHEM. SOC. VOL. 132, NO. 2, 2010

Active Site of Methyl-Coenzyme M Reductase
A R T I C L E S
Figure 1. W-band (94.2556 GHz) echo-detected EPR spectra (numerical
1st derivative) measured at 15 K of (A) MCRred1a + HS-CoM (MCRred1c),
(B) MCRred1a + HS-CoM + HS-CoB, and (C) MCRred1a + HS-CoM +
CF3-S-CoB. Experimental, black. Simulations: red, sum; (A) dark green,
red1c (1st); light green, red1c (2nd); (B/C) blue, red2r; magenta, red2a;
green, red1cc. “*” indicates MCRox1 features.
Figure 2. (A) 1H X-band HYSCORE spectrum (9.723 GHz, 304.1 mT, τ
) 120 ns). The marked cross-peaks shifted behind the 1H antidiagonal
originate from a proton hyperfine coupling with a large anisotropy in
MCRred2r. (B) 14N Q-band HYSCORE spectrum (35.3515, 1120.9 mT, τ )
124 ns). The double- (d) and single-quantum (s) cross-peaks, which can be
simulated with a nitrogen hyperfine coupling of ca. 12 MHz, belong to
MCRred2r. The double-quantum cross-peaks positions allow the hyperfine
coupling to be evaluated as indicated. (C) Q-band Davies ENDOR spectrum
(34.6767 GHz, 1193.0 mT) revealing a large proton hyperfine coupling of
|A(1H)| ) 37 MHz from MCRred2a. In A-C, the EPR spectrum and ENDOR/
HYSCORE field position are shown in the inset.
Table 1. Amplitudes (A%) and g-Values for the Three States in
MCRred1/red2 Samples Prepared with HS-CoB, CH3-S-CoB, and
CF3-S-CoB
preparation
state
A%
g1, g2, g3a
MCRred1a + HS-CoM +
HS-CoB
red1cc
red2a
red2r
69
14
17
2.0598, 2.0671, 2.2467
2.0731, 2.0771, 2.2727
2.2885, 2.2339, 2.1771
MCRred1a + HS-CoM +
CH3-S-CoB
red1cc
red2a
red2r
56
27
17
2.0595, 2.0677, 2.2479
2.0743, 2.0777, 2.2730
2.2886, 2.2339, 2.1797
interaction,27 and an asymmetric spin density distribution on
the four hydropyrrolic nitrogens.29 The latter two features
MCRred1a + HS-CoM +
CF3-S-CoB
red1cc
red2a
red2r
68
24
8
2.0599, 2.0673, 2.2490
2.0745, 2.0705, 2.2742
2.2819, 2.2295, 2.1810
were confirmed for the new species [CF3-S-CoB]-MCRred2r
.
Figure 2A shows an X-band HYSCORE spectrum measured
at the field position g1 of MCRred2r. Simulation of the set of
1H X-band HYSCORE (Figure S1, Supporting Information)
and Q-band ENDOR (Figure S2, Supporting Information)
spectra yielded the hyperfine couplings (Hrh) given in Table
2, which are very similar to those of the corresponding
[HS-CoB]-MCRred2r preparation, |A(1H)| ) [29, 26, 5] MHz.
The small hydropyrrolic nitrogen hyperfine interaction previ-
ously found in MCRred2r induced with HS-CoB is also
present with [CF3-S-CoB] as is evident by the Q-band
HYSCORE spectrum given in Figure 2B (Figure S3 displays
the full set of data and simulations, Supporting Information).
The hyperfine couplings given in Table 2 are very similar to
those from the preparation [HS-CoB]-MCRred2r where A(14N)
) [13.5, 16.0, 11.8] MHz. The remaining hydropyrrolic
nitrogen hyperfine couplings are in the range A(14N) ) 20-27
MHz (data not shown). Measurement at the position of g1,2
(≡g⊥) of [CF3-S-CoB]-MCRred2a (Figure 2C) reveals a very
large and anisotropic proton (Hax) hyperfine coupling that is
very similar to Hax of [HS-CoB]-MCRred2a, A(1H) ) [-43,
a Error in principle values: ∆gi ) 0.0005.
is achieved with CF3-S-CoB. In both [HS-CoB]-MCR and
[CF3-S-CoB]-MCR preparations the amount of MCRred2
(and MCRred1cc) induced is dependent upon the temperature
of the sample before it is rapidly frozen, an effect reported
in ref 28.
For each one of the EPR components MCRred1cc, MCRred2a
,
and MCRred2r the g-values listed in Table 1 are nearly the same
regardless of whether HS-CoB, CH3-S-CoB, or CF3-S-CoB
is bound. This indicates that coenzyme B and its two anaolgues
induce very similar structural changes upon binding to the
MCRred1c form.
In [HS-CoB]-MCR preparations, it is known that the
induced MCRred2r species is distinguished by an Ni-S
coordination,39 a very large and anisotropic proton hyperfine
(39) Finazzo, C.; Harmer, J.; Bauer, C.; Jaun, B.; Duin, E. C.; Mahlert, F.;
Goenrich, M.; Thauer, R. K.; Van Doorslaer, S.; Schweiger, A. J. Am.
Chem. Soc. 2003, 125 (17), 4988.
9
J. AM. CHEM. SOC. VOL. 132, NO. 2, 2010 571

A R T I C L E S
Ebner et al.
Table 2. 1H, Pyrrole 14N, and 19F Hyperfine Couplings and Error Estimates (∆) for the MCRred2a, MCRred2r, and MCRred1cc Species Induced
with CF3-S-CoB
nucleus/comment
A1, A2, A3 (∆Α1, ∆Α2, ∆Α3) [MHz]
R, ꢀ, γa (∆R, ∆ꢀ, ∆γ)
aiso [MHz]
T1, T2, T3 [MHz]
MCRred2a
-, 0, 45 (-, 5, 5)
-
320, 11, 49
85, 31, 287
160, 38, 114
1Hax (hydride)
-37.5, -34, -0.5 (1, 2, 1)
∼22-38
-23.6
-
-13.8, -10.3, -24.2
-
-1.0, -0.9, 1.9
-0.7, -0.7, 1.4
-2.3, -1.2, 3.5
14N (×4)g
f
19F1
-0.8, -0.8, 2.0
1.7, 1.7, 3.9
0.1
19F2
19F3
2.4
1.6
-0.8, 0.4, 5.0
MCRred2r
-130, 125, 60 (30, 10, 10)
-, 30, 45
d
1Hrh
28.0, 25.0, 4.0 d (1, 2, 1)
11.5, 17.0, 11.5
∼20-27
19.0d
13.3
-
9.0, 6.0, -15.0d
-1.8, 3.6, -1.8
-
-0.9, -0.8, 1.7
-0.7, -0.7, 1.4
-1.6, -1.4, 3.0
14Nb
14N (×3)g
-
e
19F1
-0.1, -0.1, 2.4
0.8, 0.8, 2.9
260, 21, 89
83, 15, 5
170, 38, 100
0.7
19F2
1.5
0.1
19F3
-1.5, -1.3, 3.1
MCRred1cc
-
-, 7, -
-, 21, -
-, 15, -
14N (×4)g
19F1
∼27
-
0
0
-
-0.24, -0.24, 0.48
-0.24, -0.24, 0.48
-0.49, -0.49, 0.96
-0.24, -0.24, 0.48
-0.27, -0.27, 0.56
-0.49, -0.49, 0.96
19F2
19F3
0
a Euler angles define the passive rotation of the hyperfine or nuclear quadrupole principal axis system into the g-matrix principal axis system, A )
b
R(R,ꢀ,γ)AdiagonalR†(R,ꢀ,γ). Nuclear quadrupole parameters |e2qQ/h| ) 3.25 MHz, η ) 0.6, and [R,ꢀ,γ] ) [0,-60,135], where κ ) |e2qQ/h|/4I(2I - 1))
and asymmetry parameters η ) (Q1 - Q2)/Q3 with Q1 ) -κ(1 - η), Q2 ) -κ(1 + η), and Q3 ) 2κ. c Several values of γ (R) give similar simulations;
no error estimate is given. d Absolute sign of interaction unknown. e Values for the F(S) ) 0.17. f Values for the F(S) ) 0.04. g Range of pyrrole
nitrogen hyperfine coupling from refs 23, 29, and 38.
-42, -5] MHz (Figures S1-S2 display the data set,
Supporting Information).
thiol F(S) ) 17-6%,41 three hydropyrrolic nitrogen as F(N) )
3% and the fourth as F(N) ) 1.5%.
The g-values and hyperfine couplings show conclusively that
the nature of the three EPR components induced by addition of
either the natural substrate H-S-CoB or the analog
CF3-S-CoB results in nickel-centered paramagnetic species
with very similar structures and CoB bindings.
For each species, ENDOR spectra are simulated with three
19F hyperfine couplings,
Ai ) aiiso + Ti(R) (i ) 1,2,3)
(3)
We now turn our attention to locating the CF3 position (of
CF3-S-CoB) for the three species. Fluorine signals from the
[CF3-S-CoB] preparation are observed with W-band Mims
ENDOR as shown in Figure 3; spectra c-j have contributions
from MCRred1cc, b-i from MCRred2a, and MCRred2r contributes
to traces a-e. The three fluorine C19F3-S-CoB hyperfine
couplings for each species (red1cc/red2a/red2r) were modeled
with an isotropic aiso and a dipole contribution with principal
values [T1, T2, T3] MHz, the latter calculated according to the
and fitted to the experimental traces by optimizing the three
aiiso values and the positions of CF3-(S-CoB) relative to NiF430
by changing R and rk in eqs 2A/B. This equates to moving CF3
in the x, y, or z directions and rotation of CF3 around the S-C
bond, dTk ) dTk(dx,dy,dz,dθ). Optimized 19F couplings and the
corresponding fits are shown in Table 2 and Figure 3,
respectively.
The ENDOR spectrum recorded at the highest field position,
Figure 3j, is contributed to by only MCRred1cc and has a hyperfine
coupling at g1,2 (≡ g⊥) of |A1,2| ≈ 0.4 MHz. This coupling can
be accounted for when [CF3-S-CoB] is positioned as in the
NiII crystal structure (i.e., simply by replacing the thiol H with
CF3, see Scheme 2A) giving the three Ni-F distances between
40
point-dipole model,
T )
R (R, ꢀ, γ)T diag(-1, -1, 2)R†(R, ꢀ, γ)
k k k
∑
k
(2A)
(2B)
r ) 6.3-7.6 Å. The point-dipolar model of Eq. 2A then gives
calc
A
) -(0.21-0.46) MHz and Ac3alc ) +(0.43-0.91) MHz42
1,2
with
(see Table 2) since the g3-axis (g|-axis) is known to point
perpendicular to the NiF430 plane. The fluorine signals for
MCRred1cc are thus accurately modeled by only dipolar contribu-
tions (aiiso ) 0) with [CF3-S-CoB] positioned according the
crystal structure of NiII MCR (Scheme 2A).
1
rk3
Tk ) (µo/4πh)(geꢀegnꢀn)Fk
where rk is the distance between the unpaired electron and the
kth nucleus with spin population Fk, and R(R,ꢀ,γ) is the rotation
matrix transforming the kth point-dipole contribution into the
g-matrix principal axis system. In calculations using eq 2A, the
following spin populations were assumed unless stated other-
wise; MCRred1cc F(Ni) ) 88%, each hydropyrrolic nitrogen F(N)
) 3%; MCRred2a F(Ni) ) 84%, thiol F(S) e 4%; each
hydropyrrolic nitrogen F(N) ) 3%; MCRred2r F(Ni) ) 71-82%,
To simulate the MCRred2a signals at field positions b-i the
hyperfine model needed both isotropic and dipolar contributions
as listed in Table 2. This is a significant result as it indicates an
electronic connection between the nickel ion and the F nuclei,
(41) Spin population is based on the 33S hyperfine interaction, for which
the sign of the principal values could not be determined. This results
in the two value of 6 or 17%.
(42) The hyperfine coupling between an electron localised at a point and
a nuclear spin is (-T,-T,2T), where the largest value has its axis along
the electron-nuclear vector.
(40) Po¨ppl, A.; Kevan, L. J. Phys. Chem. 1996, 100 (9), 3387.
9
572 J. AM. CHEM. SOC. VOL. 132, NO. 2, 2010

Active Site of Methyl-Coenzyme M Reductase
A R T I C L E S
Scheme 2. Schematic Representation of the CF3-S-CoB
Positions as Determined by 19F ENDOR Measurementsa
a (A) MCRred1cc with distances r(Ni-F) ) 6.3, 7.5, 7.6 Å, which are
essentially those expected assuming the crystal structure of inactive Ni(II)
enzyme. (B) A DFT model of MCRred2a with H-S-CoM coordinated,
r(Ni-S) ) 3.1 Å and r(Ni-H) ) 1.7 Å.27 Arrows indicate the shortest
and longest Ni-F distances for sulfur spin densities of 0 and 0.04 (for F(S)
) 0.04; r(Ni-F) ) 4.0, 5.0, 5.4 Å and r(S-F) ) 1.5, 2.1, 2.5 Å). (C)
MCRred2r where EPR has established a strong Ni-S coordination, a strongly
coupled proton, and asymmetric pyrrole nitrogen hyperfine couplings; the
structure is a hypothesis consistent with this data.27 The Ni-F distances
indicated by the arrows are for F(S) ) 0.06 and 0.17, and r(Ni-S) ) 2.5
Å (for F(S) ) 0.17; r(Ni-F) ) 5.1, 5.9, 6.8 Å and r(F-S) ) 2.1, 2.9, 3.9
Å).
Ni-F distances. The optimized parameters for MCRred2a
,
obtained from simulations using F(S) ) 0.04, are listed in Table
2. The dipolar parts of the three hyperfine couplings indicate a
shorting of the Ni-F distances from 6.2-7.7 Å to 4.0-5.5 Å
(Scheme 2B), consistent with direct contact between fluorine
atoms and the sulfur of CoM. The result is sensitive to the spin
density distribution on the S of CoM, and F(S) ) 0.04 is an
upper limit and thus the distances are also upper limits. Using
F(S) ) 0 reduces the Ni-F distances to 3.2-4.6 Å.
Figure 3. W-band (94.255 GHz) Mims ENDOR spectra recorded at 15 K
showing the 19F frequency region. Experimental, black. Simulations: total,
red; red1cc, green (b-j); red2a, magenta (b-i); and red2r, blue (a-e). The
B0 observer positions (mT) and τ values (ns) are (A) 2954, τ ) 200, 400,
(B) 2969, τ ) 200, (C) 3019, τ ) 200, 500, (D) 3051, τ ) 200, 500, (E)
3082, τ ) 200, 400, (F) 3096, τ ) 200, 400, (G) 3169, τ ) 200, 600, (H)
3219, τ ) 200, 500, (I) 3242, τ ) 200, and (J) 3251, τ ) 200, 400.
Simulations for MCRred2r are determined from the five spectra
of Figure 3a-e. The spectrum at the lowest field position (a)
showing |A1(19F)| ) 0.9-1.5 MHz is contributed to only by the
MCRred2r species and only those molecular orientations close
to its g1 axis. In this species the spin density on the coordinated
thiol(ate) sulfur is estimated from 33S EPR data in the range
F(S) ) 0.06-0.17.39 Taking the upper limit of F(S) ) 0.17,
the maximum fluorine dipolar coupling according to Eq. 2A
with CF3-S-CoB positioned according the X-ray structure of
the inactive Ni(II) enzyme would be Ac|al ) 1.2 MHz. This value
occurs directly along the electron-nuclear vector, which is
expected to be close to the g3 axis that points closely along the
Ni-S bond.39 The calculated two perpendicular values would
be Ac⊥al ) -0.6 MHz. Thus even in the impossible case where
the g1 axis points near to the Ni-F and S-F vectors, the
calculated coupling is too small to account for the experimental
one, indicating the CF3-S-CoB nickel distance has shortened,
and probably that there is also an isotropic contribution. The
final optimized hyperfine parameters are given in Table 2, and
as for MCRred2a, the presence of the isotropic 19F couplings
indicates an orbital overlap with the thiol sulfur of (H)S-CoM
and a reduction in the Ni-F distances to r(Ni-F) ) 5.1-6.8
Å. Simulation using the lowest thiol(ate) spin density, F(S) )
probably mediated through the coordinated Ni-H-S-CoM
moiety (see Scheme 2B). Clearly the hyperfine couplings are
much larger than those from MCRred1cc as can be appreciated
by comparing trace j (only MCRred1cc) to i (g1,2 of MCRred2a
/
MCRred1cc) which shows a hyperfine coupling of |A1,2(19F)|max
) 4 MHz belonging to MCRred2a. The magnitude of this very
important coupling was confirmed by X-band HYEND (Figure
S4, Supporting Information) and Q-band Mims ENDOR mea-
surements (Figure S5, Supporting Information) and cannot be
purely dipolar for several reasons. First, this would imply
A(19F)max ≈ [-4,-4,8] MHz since the g3 axis and A3 axis are
perpendicular to the F430 macrocycle and CF3-S-CoB is located
near the g3 axis. This would give r(Ni-F) ) 2.7 Å for F(Ni) )
1, improbably short since, in MCRred2a, the “hydride” hydrogen
is located axially at ca. 1.7 Å above the nickel, and H-S-CoM
is known to be close to the nickel as well (via hyperfine
couplings of the ꢀ-protons).27 Second, the field dependence of
the 19F signal is not dramatic and at all ENDOR observer field
positions A(19F)max < 4.5 MHz. Third, such a large anisotropy
would easily be detected with X-band HYSCORE measure-
ments, which is not the case (data not shown). For an isotropic
hyperfine contribution aiso the fluorine orbital must overlap with
a thiol(ate) S orbital from CoM, implying a shortening of the
9
J. AM. CHEM. SOC. VOL. 132, NO. 2, 2010 573

A R T I C L E S
Ebner et al.
0.06, yields considerably shorter Ni-F distances of r(Ni-F)
) 4.5-5.7 Å.
Conclusion
The Ni-19F hyperfine couplings presented above leave no
doubt that in MCRred2a and MCRred2r, the two species induced
upon binding of CF3-S-CoB to MCRred1c, the distance between
the trifluoromethyl group and the nickel center is distinctly
reduced as compared to MCRred1cc (with couplings that are
described very well assuming CF3-S-CoB positioned as in
the NiIIF430 X-ray structure).
Since the EPR and ENDOR spectra of the MCRred2 forms
induced by binding of coenzyme B or its analogs S-methyl
coenzyme B and S-trifluoromethyl coenzyme B are very similar
and show the same three EPR species, it is reasonable to assume
that all three are bound to the enzyme in the same way and
induce the same structural changes in the protein and the
coordination environment of the nickel in F430. The enzyme is
not inactivated by the fluorine labeled inhibitor: CF3-S-CoB
in the presence of coenzyme M does not quench the MCRred1
signal and reversibly converts MCRred1 to MCRred2 signals to
the same extent as CH3-S-CoB. The latter has been shown to
reversibly inhibit MCR, inhibition being competitive with
respect to HS-CoB and noncompetitive with respect to
CH3-S-CoM.30,43
Inspection of the X-ray crystal structure of the inactive Ni(II)
form MCRox1-silent, which also has one molecule of coenzyme
B and one molecule of coenzyme M bound in the active site,
reveals that the polar groups of the threonine posphate end of
coenzyme B are fixed by numerous hydrogen bonds to the amino
acids lining the funnel-shaped entrance near the enzyme surface
whereas the 7-thioheptanoyl chain tightly fits into a long narrow
channel formed by hydrophobic amino acid side chains, which
connects the entrance funnel to the dome-shaped cavity above
the proximal face of the nickel hydrocorphin macrocycle (Figure
4). The 7-thioheptanoyl chain assumes a t,t,g+,g+,g+,t conforma-
tion with a bend that follows a corresponding kink of ca. 120°
in the protein channel. The terminal thiol group is located at
the exit of the channel into the cavity in which coenzyme M is
bound.
An additional trifluoromethyl group bound to the sulfur
of coenzyme B can be modeled into the rigid structure of
MCRox1-silent without steric clashes; it protrudes into the proximal
pocket directly above the thiol group of bound coenzyme M.
In this model, in which the protein structure remains the same
as in MCRox1-silent, the shortest distance between a fluorine atom
and nickel is 6.25 Å. This distance corresponds closely to the
value deduced from the hyperfine data for the MCRred1cc
component of our MCRred2a/r-MCRred1cc sample, namely 6.3 Å.
In contrast to MCRred1cc, the closest distances between 19F
and Ni deduced for the MCRred2a and the MCRred2r species of
3.2-4.0 Å and 4.5-5.1 Å, respectively, are much shorter than
anticipated based on the MCRox1-silent crystal structure. The pos-
sibility that the trifluoromethyl group is transferred from the
sulfur of coenzyme B to the one of coenzyme M inside the
enzyme can be excluded because we have shown earlier that
CF3-S-CoM in the presence of coenzyme B quenches the Ni(I)
signal to give an organic radical.44 Within the rigid framework
of the MCRox1-silent structure, the possibility that coenzyme B
Figure 4. Active site of MCR-I in the inactive Ni(II) form MCRox1-silent
with bound coenzymes B and M. Shown is a cut through the surfaces of
the protein at the interfaces of the R-, R′- and ꢀ-chains, which form the
polar entrance funnel, the long, narrow, hydrophobic channel with its bend,
and the dome-shaped cavity above the Ni-hydrocorphinoid F430. PDB code:
1MRO.3
could slide down the channel into a second binding site can be
excluded because this would lead to a loss of the hydrogen bonds
of the threonine phosphate moiety with polar protein residues,
and the channel is too narrow to accommodate this part of
coenzyme B (Figure 4). Attempts to model a more extended
t,t,t,t,t,t conformation of coenzyme B, which might be able to
reach to within the measured distances, into the MCRox1-silent
structure failed because they lead to prohibitive van der Waals
clashes with the protein in the region of the bend in the channel.
We therefore conclude that, upon the transformation of MCRred1c
into MCRred2a and MCRred2r, induced by binding of coenzyme
B and its analogs, the protein must undergo a significant
conformational change.
It has been shown earlier that the binding of coenzyme B
converts at most 50% of the spins of MCRred1c into MCRred2a
plus MCRred2r and that the two MCRred2 species are in a
temperature dependent equilibrium.28 The fact that at least 50%
of the MCRred1cc signal remains after addition of coenzyme B
has been discussed in the terms of a hypothetical two-stroke
mechanism in which the two symmetry related active sites of
the enzyme are coupled and not synchronously in the same state
during catalysis. The weak but significant hyperfine coupling
between the 19F atoms of CF3-S-CoB and the unpaired
electron on the nickel center of the MCRred1cc component
reported here proves that the coenzyme B analog is also bound
to active sites that are not converted into MCRred2r or MCRred2a
.
Since binding of coenzyme B or its analogs is triggering the
formation of the MCRred2 species, the question arises why active
sites remain in the MCRred1cc state despite the fact that they
have coenzyme B bound. This question has to remain open
because ENDOR experiments do not directly measure pair-pair
relationships and cannot correlate the presence of MCRred1cc and
MCRred2 forms in the same enzyme molecule. However, one
possible explanation would be that the conformational change
in the protein induced upon transformation to MCRred2r or
(43) Goenrich, M.; Duin, E. C.; Mahlert, F.; Thauer, R. K. J. Biol. Inorg.
Chem. 2005, 10, 333.
(44) Goenrich, M.; Mahlert, F.; Duin, E. C.; Bauer, C.; Jaun, B.; Thauer,
R. K. J. Biol. Inorg. Chem. 2004, 9, 691.
9
574 J. AM. CHEM. SOC. VOL. 132, NO. 2, 2010

Active Site of Methyl-Coenzyme M Reductase
A R T I C L E S
MCRred2a prevents the formation of these species in the other
active site. Although the X-ray structures do not directly provide
support for an asymmetric “2-stroke” catalytic cycle, they
interestingly reveal that both active sites are interlinked via the
two R subunits, both of which contribute to the formation of
both active sites.
face of the macrocycle and one of the nitrogens of F430 moved
into an off-equatorial position. The finding that the protein does
undergo a significant conformational change upon formation
of the MCRred2 species helps to rationalize these earlier data
because any movement of one of the tetrapyrrolic nitrogens
away from the equatorial plane requires a conformational change
in the protein chains tightly anchoring the sidechains of F430 in
the MCRox1-silent structure. Together, these studies open up the
possibility that coordination geometries other than square planar,
square pyramidal or elongated octahedral are accessible. The
mechanisms proposed for nonenzymatic C-H activation reac-
tions at transition metal centers typically require intermediates
in which two fragments of the reacting hydrocarbon occupy
two cis-coordination sites on the metal.46-48 Furthermore, there
is increasing evidence that anaerobic methane oxidation by
archaea of the ANME type, that is, activation of a C-H bond
in methane, is catalyzed by enzymes that are close homologues
None of the species investigated here (MCRred1cc, MCRred2a
,
MCRred2r) is a logical candidate for an intermediate in the
catalytic cycle of methane formation because, instead of the
natural substrate methyl coenzyme M, they contain the competi-
tive substrate analog coenzyme M. So far, no intermediate of
the catalytic cycle with the natural substrates has been observed,
the only state detected by EPR during catalysis being the
MCRred1m state with the first substrate, methyl coenzyme M,
bound and the second substrate, coenzyme B, absent.45 How-
ever, the detection of a significant conformational change of
the protein concomitant with the formation of the MCRred2
species reported here is relevant to the mechanistic question.
The X-ray structure of the inactive Ni(II) forms of MCR, which
are very compact and tightly packed around the active site,
certainly biased thinking about possible mechanisms because
it is difficult to envisage a change in the coordination geometry
and conformation of the tetrapyrrolic cofactor without a major
conformational change involving several protein chains. To our
knowledge, the mechanisms proposed so far only postulate
catalytic intermediates in which the coordination geometry of
the nickel center was either square planar, square pyramidal,
or elongated octahedral. We have shown previously27,39 that in
MCRred2r at least one of the tetrapyrrolic nitrogens of F430 has
a much reduced hyperfine coupling compared to the others and
that the nickel is bound to both, the sulfur of coenzyme M and
a hydrogen at <2.25 Å. This lead us to suggest a coordination
geometry with two substrate derived cis-ligands on the same
of MCR and contain either coenzyme F430 or the variant 172-
49,50
methylthio-coenzyme F430.
In view of these observations,
the possibility that methyl-coenzyme M reductase can adjust
its conformation to allow coordination geometries with two cis-
ligands on the proximal face of F430 is of particular interest.
Acknowledgment. We thank the ETH Zu¨rich, the Swiss
National Science Foundation (SNF), the Max Planck Society, and
the EPSRC and BBSRC for financial support.
Supporting Information Available: Figures S1-S5. This
material is available free of charge via the Internet at http://
pubs.acs.org.
JA906367H
(46) Janowicz, A. H.; Bergman, R. G. J. Am. Chem. Soc. 1982, 104, 352.
(47) Perutz, R. N.; Sabo-Etienne, S. Angew. Chem., Int. Ed. 2007, 46, 2578.
(48) Crabtree, R. H. J. Organomet. Chem. 2004, 689, 4083.
(49) Krueger, M.; Meyerdierks, A.; Gloeckner, F. O.; Amann, R.; Widdel,
F.; Kube, M.; Reinhardt, R.; Kahnt, J.; Boecher, R.; Thauer, R. K.;
Shima, S. Nature 2003, 426 (6968), 878.
(45) Krzycki and Prince have observed formation of the MCRred2r signal
in whole cells of Methanosarcina barkeri after cessation of linear
methane formation from acetate: Krzycki, J. A.; Prince, R. C. Biochim.
Biophys. Acta 1990, 1015, 53.
(50) Mayr, S.; Latkoczy, C.; Kruger, M.; Gunther, D.; Shima, S.; Thauer,
R. K.; Widdel, F.; Jaun, B. J. Am. Chem. Soc. 2008, 130 (32), 10758.
9
J. AM. CHEM. SOC. VOL. 132, NO. 2, 2010 575