Spectroscopic Studies of the EutT Adenosyltransferase from Salmonella enterica: Mechanism of Four-Coordinate Co(II)Cbl Formation

Article abstract of DOI:10.1021/jacs.5b11708

EutT from Salmonella enterica is a member of a class of enzymes termed ATP:Co(I)rrinoid adenosyltransferases (ACATs), implicated in the biosynthesis of adenosylcobalamin (AdoCbl). In the presence of cosubstrate ATP, ACATs raise the Co(II)/Co(I) reduction potential of their cob(II)alamin [Co(II)Cbl] substrate by >250 mV via the formation of a unique four-coordinate (4c) Co(II)Cbl species, thereby facilitating the formation of a "supernucleophilic" cob(I)alamin intermediate required for the formation of the AdoCbl product. Previous kinetic studies of EutT revealed the importance of a HX₁₁CCX₂C(83) motif for catalytic activity and have led to the proposal that residues in this motif serve as the binding site for a divalent transition metal cofactor [e.g., Fe(II) or Zn(II)]. This motif is absent in other ACAT families, suggesting that EutT employs a distinct mechanism for AdoCbl formation. To assess how metal ion binding to the HX₁₁CCX₂C(83) motif affects the relative yield of 4c Co(II)Cbl generated in the EutT active site, we have characterized several enzyme variants by using electronic absorption, magnetic circular dichroism, and electron paramagnetic resonance spectroscopies. Our results indicate that Fe(II) or Zn(II) binding to the HX₁₁CCX₂C(83) motif of EutT is required for promoting the formation of 4c Co(II)Cbl. Intriguingly, our spectroscopic data also reveal the presence of an equilibrium between five-coordinate "base-on" and "base-off" Co(II)Cbl species bound to the EutT active site at low ATP concentrations, which shifts in favor of "base-off" Co(II)Cbl in the presence of excess ATP, suggesting that the base-off species serves as a precursor to 4c Co(II)Cbl.

Full text of DOI:10.1021/jacs.5b11708

Article
pubs.acs.org/JACS
Spectroscopic Studies of the EutT Adenosyltransferase from
Salmonella enterica: Mechanism of Four-Coordinate Co(II)Cbl
Formation
†
‡
‡
,†
Ivan G. Pallares, Theodore C. Moore, Jorge C. Escalante-Semerena, and Thomas C. Brunold*
†
Department of Chemistry, University of WisconsinMadison, Madison, Wisconsin 53706, United States
‡
Department of Microbiology, University of Georgia, Athens, Georgia 30602, United States
*
S Supporting Information
ABSTRACT: EutT from Salmonella enterica is a member of a class
of enzymes termed ATP:Co(I)rrinoid adenosyltransferases
(
(
ACATs), implicated in the biosynthesis of adenosylcobalamin
AdoCbl). In the presence of cosubstrate ATP, ACATs raise the
Co(II)/Co(I) reduction potential of their cob(II)alamin [Co(II)-
Cbl] substrate by >250 mV via the formation of a unique four-
coordinate (4c) Co(II)Cbl species, thereby facilitating the formation
of a “supernucleophilic” cob(I)alamin intermediate required for the
formation of the AdoCbl product. Previous kinetic studies of EutT
revealed the importance of a HX CCX C(83) motif for catalytic
1
1
2
activity and have led to the proposal that residues in this motif serve
as the binding site for a divalent transition metal cofactor [e.g.,
Fe(II) or Zn(II)]. This motif is absent in other ACAT families,
suggesting that EutT employs a distinct mechanism for AdoCbl formation. To assess how metal ion binding to the
HX CCX C(83) motif affects the relative yield of 4c Co(II)Cbl generated in the EutT active site, we have characterized several
11
2
enzyme variants by using electronic absorption, magnetic circular dichroism, and electron paramagnetic resonance
spectroscopies. Our results indicate that Fe(II) or Zn(II) binding to the HX CCX C(83) motif of EutT is required for
11
2
promoting the formation of 4c Co(II)Cbl. Intriguingly, our spectroscopic data also reveal the presence of an equilibrium between
five-coordinate “base-on” and “base-off” Co(II)Cbl species bound to the EutT active site at low ATP concentrations, which shifts
in favor of “base-off” Co(II)Cbl in the presence of excess ATP, suggesting that the base-off species serves as a precursor to 4c
Co(II)Cbl.
INTRODUCTION
dependent enzymes, as these adenosyltransferases generate
AdoCbl (or coenzyme B ) from ATP and Co(II)Cbl
substrates.
■
1
2
EutT is the most recently identified ATP:Co(I)rrinoid
adenosyltransferase (ACAT) enzyme involved in maintaining
3
,9
1
,2
In Salmonella enterica, the eutT gene is part of a 16-gene
the supply of adenosylcobalamin (AdoCbl) in Salmonella
3
operon, whose expression is triggered by the presence of
enterica. AdoCbl and related corrinoid species are charac-
1
0,11
ethanolamine and adenosylcobalamin (AdoCbl),
leading to
terized by the presence of a redox-active Co(III) ion that is
ligated equatorially by the four nitrogen atoms of a tetrapyrrole
macrocycle, termed the corrin ring, and by the 5,6-
dimethylbenzimidazole (DMB) base at the end of an
intramolecular nucleotide loop in the “lower” axial (Coα)
position (Figure 1). Various ligands can occupy the “upper”
axial (Coβ) position, with biologically relevant forms
containing an adenosyl (AdoCbl), methyl (MeCbl), gluta-
thionyl (GSCbl), or nitrosyl (NOCbl) moiety. In the case of
AdoCbl, homolytic cleavage of the Co−C(Ado) bond to
generate an adenosyl radical and a five-coordinate Co(II)Cbl
species with a vacant Coβ position is the first step in reactions
the formation of proteinaceous polyhedron-shaped bodies.
These so-called eut metabolosomes are postulated to
encapsulate the metabolic machinery employed by the cell for
12,13
the catabolism of ethanolamine.
The ability of S. enterica to
use ethanolamine as a source of nutrition was suggested to
provide this bacterium with a growth advantage over other
1
4
intestinal flora under inflammatory conditions, as ethanol-
11,15
4
−6
amine is readily available in the mammalian gut.
In this
metabolic pathway, the AdoCbl-dependent ethanolamine
ammonia lyase (EAL) enzyme (the product of the eutB and
eutC genes) catalyzes the conversion of ethanolamine to
16
7
acetaldehyde and ammonia. EutT has been postulated to
catalyzed by AdoCbl-dependent eliminases/mutases. Loss or
inactivation of AdoCbl during turnover leads to deactivation o8f
these enzymes, requiring replacement of their cofactor.
ACATs are thus implicated in the reactivation of AdoCbl-
Received: November 8, 2015
©
XXXX American Chemical Society
A
DOI: 10.1021/jacs.5b11708
J. Am. Chem. Soc. XXXX, XXX, XXX−XXX

Journal of the American Chemical Society
Article
solution is displaced by a noncoordinating phenylalanine
residue (F112 in LrPduO and F91 in SeCobA). Subsequent
electron paramagnetic resonance (EPR) and magnetic circular
dichroism (MCD) characterization of Co(II)Cbl in the
presence of SeCobA and LrPduO identified the roles specific
amino acid residues play with regard to the formation of 4c
3
8,39
Co(II)Cbl.
These studies were facilitated by the fact that a
−1
sharp feature centered at ∼12 000 cm appears in the MCD
spectra of Co(II)Cbl bound to SeCobA/MgATP and LrPduO/
MgATP, whose intensity correlates directly with the relative
amount of 4c Co(II)Cbl generated in these enzymes. The
appearance of this feature was shown to reflect a sizable
2
z
stabilization of the singly occupied, redox-active Co 3d -based
molecular orbital in 4c Co(II)Cbl, which was predicted to lead
to an increase in the Co(II)/Co(I) reduction potential by ≥250
24,36,40
mV.
We have recently obtained spectroscopic evidence that also
supports the formation of a 4c Co(II)Cbl species in the active
41
Figure 1. Chemical structure of adenosylcobalamin (AdoCbl), the
final product of the reaction catalyzed by ATP:Co(I)rrinoid
adenosyltransferases (ACATs).
site of EutT in the presence of excesss MgATP. Moreover,
our data indicated that EutT is unable to generate 4c
+
Co(II)Cbi , in contrast to the SeCobA and PduO-type
ACATs. Consistent with this observation, bioassay and kinetic
studies revealed that EutT fails to adenosylate this Co(II)Cbl
supply the levels of AdoCbl required for EAL to perform this
41
2
,17
precursor. Because EutT has thus far eluded characterization
by X-ray crystallography, the approach by which it binds its
Co(II)Cbl substrate and generates 4c Co(II)Cbl remains
unknown. Biochemical characterizations of EutT provided
evidence that this enzyme exists as a dimer in solution and
requires a divalent transition metal cofactor for catalytic activity,
function.
Notably, the presence of eutT genes was observed
18
to correlate with higher incidence of gastroenteritis.
In addition to EutT, two nonhomologous ACATs are present
3
in S. enterica. The SeCobA ACAT is continuously produced to
maintain basal concentrations of AdoCbl and is involved in the
de novo synthesis of AdoCbl and in the scavenging of
27,42
19,20
in striking contrast to the other known families of ACATs.
incomplete corrinoids.
The gene encoding for SePduO is
The activity of EutT was found to be highest in the presence of
Fe(II) ions under anaerobic conditions, followed by Zn(II) and
Co(II) ions, with only the latter two also supporting catalytic
part of the pdu operon that allows S. enterica to utilize 1,2-
propanediol as a source of carbon and energy. PduO-type
enzymes are homologous to the human adenosyltransferase
27
(
hATR), which has been shown to deliver its AdoCbl product
activity in the presence of oxygen. Inspection of the primary
sequence of EutT led to the proposal that that HX11CCXXC-
(83) motif may be involved in divalent metal ion binding, as
similar motifs are found in Bacillus subtilis CbiX, a 4Fe/4S-
containing enzyme, as well as some Zn-finger proteins.
directly to the B -dependent methylmalonyl CoA-mutase
(
1
2
2
1,22
MMCM) enzyme.
Despite the lack of sequence and
structural homologies among the known families of ACATs, a
common Co−C(Ado) bond formation mechanism has
previously been proposed for these enzymes involving the
reduction of a Co(II)rrinoid precursor in their active sites to
generate a “supernucleophilic” Co(I) species capable of
WT
Elemental analyses of wild-type EutT (EutT ) and variants
with a single alanine substitution at either the H67, C79, C80,
WT
or C83 position confirmed that EutT incorporates Fe(II)
3
,23,24
attacking the 5′-carbon of ATP.
A particularly challenging
and Zn(II) ions in an ∼2:1 protein-to-metal ratio, with the
C80A
C83A
step in this mechanism involves the reduction of Co(II)rrinoids
to generate the key Co(I) intermediate, as the reduction
potentials of Co(II)rrinoids free in solution are well below
those of available reducing agents in the cell (e.g., E°(NHE) =
EutT
and EutT
variants exhibiting a dramatically
reduced metal ion affinity. Futhermore, the level of metal
WT
incorporation into EutT and the alanine variants was found
41
to correlate directly with enzymatic activity.
−
[
610 mV for Co(II)Cbl and −490 mV for cob(II)inamide
To obtain further insight into the roles played by the divalent
metal cofactor of EutT with respect to the formation of 4c
Co(II)Cbl, we have employed MCD and EPR spectroscopies
+
Co(II)Cbi ], a Co(II)Cbl precursor that lacks the nucleotide
25
loop and DMB moiety). However, Co(II)Cbl species bound
to the active sites of EutT and the PduO-type ACAT from
Lactobacillus reuteri (LrPduO) are readily reduced by free
dihydroflavins to generate Co(I)Cbl, (e.g., E°(NHE) = −228
WT
to probe the interaction of Co(II)Cbl with EutT containing
WT
WT
either Fe(II) [EutT /Fe] or Zn(II) [EutT /Zn] as well as
variants possessing a single alanine substitution at the H67,
H75, C79, C80, or C83 position. These techniques offer a
uniquely sensitive probe of changes in the coordination
environment of the cobalt ion as Co(II)Cbl binds to the
EutT active site. While MCD spectroscopy allows for the
identification of ligand field (LF) and charge transfer (CT)
transitions, whose energies are substantially different for 4c and
five-coordinate (5c) Co(II)Cbl, EPR spectroscopy provides a
complementary probe of the 5c “base-on” (DMB bound) and
“base-off” (water bound) fractions of Co(II)Cbl generated in
the EutT active site. Collectively, these techniques have enabled
us to determine how the relative populations of 4c and 5c
26,27
mV for FMN at pH 7.5)
or by reduced flavoproteins in the
case of SeCobA (E°(NHE) = −440 mV for the semiquinone/
reduced flavin couple of FldA, the purported physiological
2
8−30
31,32
partner of SeCobA).
graphic,
Extensive biochemical,
crystallo-
3
2−35
36−38
and spectroscopic studies
of SeCobA and
different PduO-type ACATs have revealed that a structurally
unique four-coordinate (4c) Co(II)Cbl species is generated by
these enzymes in the presence of ATP. The most recent X-ray
crystal structures reported for SeCobA and LrPduO complexed
with Co(II)Cbl and MgATP have provided evidence that the
DMB moiety that occupies the Coα position of Co(II)Cbl in
B
DOI: 10.1021/jacs.5b11708
J. Am. Chem. Soc. XXXX, XXX, XXX−XXX

Journal of the American Chemical Society
Article
Co(II)Cbl species generated in the EutTWT active site are
affected by divalent metal ion incorporation and amino acid
substitutions in the HX CCX C(83) motif. The results
obtained in this study provide unprecedented insight into the
mechanism employed by EutT to generate AdoCbl.
was measured with a Varian EIP model 625A CW frequency counter.
All spectra were collected using a modulation amplitude of 10 G and a
modulation frequency of 100 kHz. EPR spectral simulations were
1
1
2
44
performed using Dr. Mark Nilges’s SIMPOW program.
RESULTS
■
METHODS
Effect of Metal Substitution on 4c Co(II)Cbl For-
mation. To determine how the identity of the divalent metal
ion cofactor affects the binding of Co(II)Cbl to EutT, we
collected low-temperature Abs and MCD data of Co(II)Cbl in
the absence and presence of enzyme and cosubstrate ATP. In
the Abs spectrum of Co(II)Cbl obtained in the absence of
EutT (Figure 2C), an intense feature is observed at 21 000
■
(
Cofactors and Chemicals. The chloride salt of aquacobalamin
[H OCbl]Cl) and potassium formate (HCOOK) were purchased
2
from Sigma and used as obtained. Co(II)Cbl was generated by
addition of a small amount (∼40 μL) of saturated HCOOK solution
+
to a degassed sample (0.5 mL) of aqueous H OCbl (∼5 mM) in a
2
sealed vial. The progress of the reduction was monitored
spectrophotometrically. The magnesium salt of ATP was used
throughout all experiments.
−1
cm (the so-called α-band) that has previously been attributed
Protein Preparation and Purification. EutT production and
27
purification was performed as described elsewhere. Briefly, EutT
from Salmonella enterica sv. Typhimurium LT2 (EutT) was overex-
pressed from a pTEV6 vector in Escherichia coli BL21, which generated
a fusion protein with cleavable N-terminal hexahistidine (H ) and
6
maltose binding protein (MBP) tags. EutT variants were generated
using the QuikChange II site-directed mutagenesis kit (Stratagene).
All proteins were purified on a HisTrap nickel-affinity column (GE
Healthcare). The H -MBP tag was cleaved using recombinant tobacco-
6
43
etch protease (rTEV). The tag was separated from the protein using
a HisTrap nickel-affinity column and an amylose column (New
England Biolabs).
To prepare samples of metalated EutT for spectroscopic studies, the
protein was concentrated to a least 25 mg/mL (Amicon Ultra, 10 000
MWCO). EutT was placed into dialysis units (D-tube mini, Novagen)
and moved into an anoxic chamber. The EutT solution was then
dialyzed three times for 30 min against 1.0 L of degassed buffer (2-[4-
(
2-hydroxyethyl)piperazin-1-yl]ethanesulfonic acid (HEPES, 50 mM,
pH 7) containing NaCl (300 mM), tris(2-carboxyethyl)phosphine
TCEP, 0.25 mM), ethylenediaminetetraacetic acid (EDTA, 2 mM),
(
and low (10 μM) or high (1 mM) concentrations of MgATP). To
generate EutT/Fe, ApoEutT samples were dialyzed against chelex-
treated buffer containing FeSO (1 mM) and low or high MgATP (3
4
2
7
×
1L, 30 min each). Since recombinant EutT copurifies with zinc,
EutT/Zn was instead prepared by dialysis against chelex-treated buffer
devoid of EDTA, but containing low or high concentrations of
MgATP (3 × 1L, 30 min each). Metalated EutT samples were gently
resuspended with 100% glycerol to a final concentration of 10 mg/mL
and 60% glycerol. Samples were stored in airtight serum vials
(
Wheaton) and flash-frozen in liquid nitrogen until use.
Sample Preparation. Purified EutT (∼300−600 μM) in HEPES
buffer (50 mM, pH 7) containing NaCl (300 mM) and TCEP (0.25
mM) was incubated with Co(II)Cbl under anoxic conditions. To
generate EutT samples in the presence of excess MgATP, ∼ 10 μL of a
concentrated MgATP solution were added to yield a final MgATP
concentration of >5 mM. Further details of sample compositions are
provided in Table S1. Protein solutions were injected into the
appropriate sample cells in an oxygen-free chamber. Following
removal from the anoxic chamber, the samples were frozen and
stored in liquid nitrogen. For the MCD samples, room temperature
absorption (Abs) spectra were collected under an N atmosphere
2
before they were frozen. No changes in the Abs spectrum of
Co(II)Cbl due to oxidation or sample degradation were observed after
freezing the samples.
Spectroscopy. MCD and low-temperature Abs spectra were
collected on a Jasco J-715 spectropolarimeter in conjunction with an
Oxford Instruments SM-4000 8T magnetocryostat. All MCD spectra
presented herein were obtained by taking the difference between
spectra collected with the magnetic field oriented parallel and
antiparallel to the light propagation axis to remove contributions
from the natural CD and glass strain. X-band EPR spectra were
obtained by using a Bruker ESP 300E spectrometer in conjunction
with an Oxford ESR 900 continuous-flow liquid helium cryostat and
an Oxford ITC4 temperature controller. The microwave frequency
Figure 2. Abs spectra collected at 4.5 K (black traces) and 7 T VT-
MCD spectra of (A) Co(II)Cbl in the presence of EutT /Fe and a
WT
large molar excess (2 mM) of MgATP, (B) Co(II)Cbl in the presence
WT
of EutT /Fe and substoichiometric (8 μL) MgATP, and (C) free
Co(II)Cbl. The dashed vertical line position of the α-band in the Abs
spectra.
C
DOI: 10.1021/jacs.5b11708
J. Am. Chem. Soc. XXXX, XXX, XXX−XXX

Journal of the American Chemical Society
Article
Table 1. Positions of the δ-Band (MCD) and the α-Band (Abs) of 4c Co(II)Cbl Species Generated in the Active Site of EutTWT
a
and Several Variants
Co(II)Cbl
1
Δν(δ)/cm−1
ν(α)/cm−1
Δν(α)/cm−1
metal present
Zn(II)
SeEutT substitution
ν(δ)/cm−
4c-Cbl yield/%
none (WT), >10× MgATP
12 277
12 217
12 225
12 255
12 232
12 195
12 195
n/a
22
21 645
21 552
21 551
21 575
21 505
21 164
21 030
21 142
21 053
615
522
521
545
475
134
0
>95
81
none (WT)
none (WT), no TCEP
none (WT)
C79A
−38
−30
0
71
Fe(II)
>95
46
−23
−60
−60
n/a
n/a
H75A
26
H67A
12
C80A
112
n/d
n/d
C83A
n/a
23
aShifts in the δ-band are shown relative to the position of this feature in the MCD spectrum of Co(II)Cbl in the presence of EutTWT/Fe and excess
MgATP. Also shown are the relative yields of 4c Co(II)Cbl species estimated from the δ-band intensities. Shifts in the α-band are shown relative to
the position of this feature in the Abs spectrum of Co(II)Cbl in the absence of ACATs (λmax = 475 nm).
to the lowest-energy corrin π → π* transition on the basis of its
sample of Co(II)Cbl in the presence of EutTWT/Fe and a
7
,45
high intensity and TDDFT calculations.
The prominent
substoichiometric (∼10 μM) amount of MgATP to enhance
27
shoulder on the low energy side of the α-band was assigned to
protein stability exhibits unique features in the 14 000−
7
,45
−1
metal-to-ligand charge transfer (MLCT) transitions,
while
22 000 cm region (Figure 2B). Most notably, the positive
−1
−1
the intense, sharp feature peaking at 32 000 cm was attributed
features at ∼16 000 and ∼21 000 cm in the MCD spectrum
to another corrin π → π* transition. Upon the addition of wild-
of free Co(II)Cbl (Figure 2C) red-shift by ∼500 and ∼400
WT
−1
WT
type EutT metalated with Fe(II) ions (EutT /Fe) and
cm , respectively, in the presence of EutT /Fe (Figure 2B),
which could potentially reflect the formation of base-off
Co(II)Cbl (vide infra). The corresponding Abs spectrum
reveals that the α-band does not actually red-shift in the
substochiometric amounts (<10 μM) of MgATP to a solution
of Co(II)Cbl, the intensity of this near-UV Abs band decreases
by ∼30% (Figure S7), consistent with substrate Co(II)Cbl
binding to the enzyme active site (vide infra). In the presence
WT
presence of EutT , suggesting that the MCD feature at 21 000
WT
−1
of EutT /Fe and an ∼10-fold molar excess of MgATP
cm arises from one of the many CT transitions of Co(II)Cbl
WT
(
EutT /Fe + MgATP), a similar decrease in the intensity of
whose energies are expected to be particularly sensitive to
changes in the local environment of Co(II)Cbl. Further
changes in the relative intensities and bandwidths of features
that occur in this region of the MCD spectrum provide
additional evidence that a significant fraction of Co(II)Cbl
−1
this Abs feature is observed, along with an ∼550 cm blue-shift
of the α-band (Figure 2A). In previous studies of Co(II)Cbl in
the presence of SeCobA or LrPduO, a sizable blue-shift of the
α-band was shown to reflect the formation of a 4c Co(II)Cbl
species in the enzyme active site.
the α-band of Co(II)Cbl caused by the addition of EutT /Fe
36,37
WT
Intriguingly, the shift of
binds to the EutT /Fe active site in the absence of
WT
cosubstrate MgATP.
−1
+
MgATP is ∼150 cm larger than that induced by the
addition of SeCobA or LrPduO and MgATP.
The low-energy region of the MCD spectrum of Co(II)Cbl
36−38
WT
in the presence of EutT /Fe and a large (>10-fold) molar
The MCD spectrum of Co(II)Cbl in the absence of EutT is
characterized by a series of derivative-shaped features that give
rise to the appearance of four oppositely signed bands centered
excess of MgATP is characterized by an intense, positively
−1
signed sharp feature at ∼12 300 cm (δ-band) and a broad
−1
band at ∼13 600 cm (β-band) (Figure 2A). These features,
−1
at ∼16 000 cm , as well as additional, positively signed bands
which have previously been assigned to the electronic origin
−1
7
2
2
2
around 19 000 cm (Figure 2C). The two lowest-energy
features in this spectrum have previously been assigned to LF
transitions primarily involving electronic excitations from the
doubly occupied Co 3d - and 3d -based molecular orbitals
and a vibrational sideband of the Co 3d ‑ → 3dz LF
x z
transition, are characteristic of 4c Co(II)rrinoids generated in
the active sites of ACATs via elimination of any axial ligand
36,38−41
interactions.
The relative MCD intensities of the δ-band
xz
yz
46
2
(
MOs) to the singly occupied Co 3d -based MO. Additional
associated with 4c Co(II)Cbl and features originating from 5c
Co(II)Cbl indicate that >95% of the Co(II)Cbl substrate was
converted to a 4c species in this sample (Table 1). Consistent
with this high 5c → 4c conversion yield, additional
spectroscopic features from 4c Co(II)Cbl are clearly observed
z
−1
features between 16 000 and 22 000 cm have generally been
attributed to metal-to-ligand charge transfer (MLCT) or
additional LF transitions, with specific assignments made
difficult by the presence of multiple transitions with similar
energies in this region.
−1
at ∼20 000 cm , termed the λ- and σ-bands (Figure S2), which
are usually masked by more intense contributions from the
remaining 5c Co(II)Cbl species. The 5c → 4c Co(II)Cbl
The MCD spectrum of Co(II)Cbl in the presence of
WT
EutT /Fe exposed to oxygen is nearly superimposable on that
WT
of Co(II)Cbl free in solution (Figure S1, bottom). Likewise,
only minor spectral differences are observed between the
corresponding Abs spectra (Figure S1, top). These results
conversion yield achieved by EutT /Fe in the presence of
excess MgATP largely exceeds the conversion yields reported
for SeCobA and LrPduO under similar conditions (∼8% and
WT
38,39
suggest that Co(II)Cbl does not bind to EutT /Fe under
40%, respectively).
aerobic conditions, in support of our proposal that the Fe(II)
ion of EutT /Fe is susceptible to air oxidation and that the
The Abs and MCD spectra of Co(II)Cbl in the presence of
WT
WT
WT
EutT /Fe and its metal-substituted EutT /Zn derivative are
nearly indistinguishable from each other (as well as the Abs and
Fe(III) ion formed in this process is released from the
2
7
WT
protein. In contrast, the MCD spectrum of an anaerobic
MCD spectra of Co(II)Cbl in the presence of EutT /Zn
D
DOI: 10.1021/jacs.5b11708
J. Am. Chem. Soc. XXXX, XXX, XXX−XXX

Journal of the American Chemical Society
Article
4
1
published previously ) under both minimal and excess
MgATP conditions (Figure S2), indicating that the Fe(II)-
WT
and Zn(II)-bound forms of EutT
possess very similar
Co(II)Cbl binding sites. The 5c → 4c Co(II)Cbl conversion
WT
yield in the presence of EutT /Zn and a 3-fold excess of
MgATP is ∼81%, which is somewhat lower than that obtained
WT
with EutT /Fe under similar conditions (Table 1). However,
in the presence of ∼5 mM MgATP (a >10-fold molar excess of
MgATP over EutT), the 5c → 4c Co(II)Cbl conversion of
WT
WT
EutT /Zn becomes analogous to that of EutT /Fe (Table
1
WT
). Even though EutT /Zn is stable in air (in contrast to
WT
EutT /Fe), a small (∼10%) decrease in the 5c → 4c
Co(II)Cbl conversion yield is noted in the absence of TCEP,
possibly due to partial oxidation of the Cys residues (Table 1).
Because the identity of the divalent metal ion that is present
WT
in EutT in vivo is currently unknown, these findings clearly
WT
warrant further spectroscopic studies of both EutT /Fe and
WT
WT
EutT /Zn. Given that EutT
is active with Fe(II) and
Zn(II), it is obvious that the divalent metal ion primarily plays a
structural role. However, circular dichroism (CD) studies of
EutT in the absence and presence of divalent metal ions did not
reveal any significant conformational differences between the
two enzyme forms, suggesting that the metal binding site
27
encompasses a relatively small portion of the protein fold.
Effect of Cosubstrate MgATP Concentration on the 5c
→
4c Co(II)Cbl Conversion Yield. Because the single
2
z
unpaired electron of Co(II)Cbl resides in the Co 3d -based
MO, electron paramagnetic resonance (EPR) spectroscopy
provides an extremely sensitive probe of perturbations to the
axial ligand of Co(II)Cbl by the protein environment. At low
MgATP concentration (∼10 μM), a broad positive feature
around 2620 G can be clearly discerned in the EPR spectrum of
WT
Co(II)Cbl in the presence of EutT /Fe (Figure 3B). This
feature is notably absent in the EPR spectrum of base-on
Co(II)Cbl (Figure 3A), which lacks any visible features below
Figure 3. Experimental X-band EPR spectra collected at 20 K of (A)
free Co(II)Cbl, (B) Co(II)Cbl in the presence of EutT /Fe and
2800 G; rather, it is characteristic of a base-off species in which
WT
the DMB ligand has been replaced by a solvent-derived water
molecule (Figure 3E). The presence of partially resolved
hyperfine structure on the 2620 G feature suggests that the
enzyme active site imposes conformational constraints on the
Co(II)Cbl substrate even in the absence of MgATP. While
both base-off and base-on Co(II)Cbl contribute to the region
of the EPR spectrum above 2800 G, the presence of an intense
derivative-shaped feature centered at 2950 G and peak
splittings near ∼3350 G due to N (I = 1) superhyperfine
coupling are a clear indication of DMB coordination to the
Co(II) ion. From the relative intensities of these features, we
estimate that ∼60% of the Co(II)Cbl substrate is present in the
base-off conformation, while the remaining ∼40% is in the
base-on conformation. The presence of sharp features in the
minimal (<10 μM) MgATP, and (C) Co(II)Cbl in the presence of
WT
EutT /Fe with 3-fold molar excess of MgATP, and (D) a >10-fold
+
excess of MgATP. The spectrum of Co(II)Cbi (E), a base-off
Co(II)Cbl analogue is shown for comparison. EPR spectra were
collected using a 9.36 GHz microwave source, 2 mW microwave
power, 5 G modulation amplitude, 100 kHz modulation frequency,
and a 328 ms time constant. Features from base-off Co(II)Cbl are
highlighted with vertical dashed lines. Features from 4c Co(II)Cbl
species at low magnetic fields are noted.
1
4
MgATP, the features from 4c Co(II)Cbl gain intensity at the
expense of those associated with 5c base-on and base-off
Co(II)Cbl, indicating an increase in the 5c → 4c Co(II)Cbl
conversion yield (Figure 3D). Thus, excess MgATP appears to
facilitate the formation of 4c Co(II)Cbl by promoting DMB
dissociation to generate 5c base-off Co(II)Cbl and/or by
stabilizing the 4c Co(II)Cbl species formed in the active site.
3000 G region of the EPR spectrum suggests that the DMB
moiety of base-on Co(II)Cbl is conformationally constrained
WT
and thus bound to the active site EutT /Fe, as these features
are absent in the EPR spectrum of free Co(II)Cbl and cannot
be attributed to the presence of base-off Co(II)Cbl (vide infra).
Upon the addition of at least a 3-fold molar excess (∼1 mM)
of MgATP, a series of weak, derivative-shaped features appear
in the 1900−2450 G region of the EPR spectrum of Co(II)Cbl
While additional features below 1900 G should be observed in
WT
the EPR spectrum of 4c Co(II)Cbl in the presence of EutT
/
Fe, these features are masked by intense signals from a ferric
minority species in the sample (characteristic derivatives-
shaped feature at ∼1600 G, Figure S3).
WT
WT
in the presence of EutT /Fe (Figure 3C). These broad
As in the case of EutT /Fe, the EPR spectrum of
WT
features are characteristic of square planar, low-spin Co(II)
species, consistent with the formation of 4c Co(II)Cbl in the
Co(II)Cbl in the presence of EutT /Zn and substoiochio-
metric (∼10 μM) MgATP exhibits features consistent with the
presence of both base-on and base-off Co(II)Cbl (Figures 4
and 5B). Although this spectrum is qualitatively similar to that
WT
active site of EutT /Fe, as observed previously for other
3
6,37,47
ACATs.
In the presence of a > 10-fold molar excess of
E
DOI: 10.1021/jacs.5b11708
J. Am. Chem. Soc. XXXX, XXX, XXX−XXX

Journal of the American Chemical Society
Article
Figure 4. Low-field region of X-band EPR spectra collected at 20 K of
Figure 5. High-field region of X-band EPR spectra collected at 20 K of
WT
WT
(
A) free Co(II)Cbl, (B) Co(II)Cbl in the presence of EutT /Zn and
(A) free Co(II)Cbl, (B) Co(II)Cbl in the presence of EutT /Zn and
WT
WT
<
10 μM MgATP, and (C) Co(II)Cbl in the presence of EutT /Zn
<10 μM MgATP, and (C) in the presence of EutT /Zn and a
and a substoichiometric amount of MgATP, (D) an equimolar amount
of MgATP, and (E) a 10-fold molar excess of MgATP. The prominent
feature from base-off Co(II)Cbl is highlighted by a vertical line, and
the feature arising from ferric impurities is marked by an asterisk. EPR
spectra were collected using a 9.36 GHz microwave source, 2 mW
microwave power, 5 G modulation amplitude, 100 kHz modulation
frequency, and a 328 ms time constant. Spectra were simulated using
the parameters provided in Table 2.
substoichiometric amount of MgATP, (D) an equimolar amount of
MgATP, and (E) a 10-fold molar excess of MgATP. The prominent
feature from base-off Co(II)Cbl is highlighted by a vertical line. See
Figure 4 caption for instrument settings. Spectra were simulated using
the parameters provided in Table 2.
features from base-on or 4c Co(II)Cbl are discernible in these
spectra. A fit of these spectra yields spin-Hamiltonian
parameters essentially identical to those obtained for free
obtained in the presence of EutTWT/Fe under similar
conditions (Figure S3), ruling out any strong magnetic
interactions between the Fe(II) and Co(II) ions in Co(II)-
Co(II)Cbi (Table 2), with the exception of a 15−20 MHz
+
smaller A (Co) hyperfine coupling constant. Since A (Co)
z
z
correlates with the covalency of the Co−O(H ) bond (i.e., it
2
WT
Cbl-bound EutT /Fe and thus indicating that the metal ions
decreases as covalency increases), this observation indicates
are separated by at least 10 Å, it is better resolved and thus
provides more compelling evidence for the presence of a
unique base-off Co(II)Cbl species and of perturbed base-on
that the water molecule coordinated to the base-off Co(II)Cbl
species in EutT /Zn forms a more covalent bond with the
WT
Co(II) ion than does the axially bound water molecule in
WT
+
Co(II)Cbl. Intriguingly, in the presence of EutT /Zn and
Co(II)Cbi . This finding suggests that the Co−O(H ) bond is
2
WT
half-molar to equimolar amounts of MgATP, the EPR spectra
lack any features from base-on Co(II)Cbl (Figures 4 and 5C,
D). Instead, these spectra are very similar to that obtained for
slightly compressed in the EutT /Zn active site, thus implying
that the α-face of the corrin ring does not become completely
exposed to solvent upon DMB dissociation.
+
Co(II)Cbi free in solution (Figure 3E), indicating that under
In the presence of a 10-fold molar excess of MgATP, the
WT
WT
these conditions the Co(II)Cbl substrate is bound to EutT
/
EPR spectrum of Co(II)Cbl with EutT /Zn exhibits sharp
Zn in a base-off conformation. Note that no spectroscopic
features between 800 and 1900 G (Figures 4 and 5E), as
F
DOI: 10.1021/jacs.5b11708
J. Am. Chem. Soc. XXXX, XXX, XXX−XXX

Journal of the American Chemical Society
Article
Table 2. Spin-Hamiltonian Parameters for Co(II)Cbl in the Absence and Presence of EutTWT and MgATPa
59
14
g-values
A( Co)/MHz
A( N)/MHz
A(iso)
%
cont
gz
gy
gx
Az
Ay
Ax
40
(
(
A) Co(II)Cbl
base-on
base-on
base-off
base-off
base-off
base-on
base-off
100
2.004
2.004
1.991
1.991
1.993
2.004
1.991
1.800
1.994
1.900
2.060
2.232
2.232
2.315
2.315
2.313
2.232
2.315
2.553
2.314
2.700
2.670
2.269
2.269
2.459
2.459
2.459
2.269
2.459
3.610
2.423
2.720
2.730
303
303
387
387
389
303
387
760
402
770
805
30
30
49
49
B) Co(II)Cbl/EutTWT(Zn)/minimal MgATP
38
40
227
227
224
40
b
62
222
222
217
30
n/a
(
(
(
C) Co(II)Cbl/EutTWT(Zn)/0.5-fold MgATP
100
100
19
n/a
b
WT
b
D) Co(II)Cbl/EutT (Zn)/equimolar MgATP
n/a
WT
E) Co(II)Cbl/EutT (Zn)/>10-fold excess MgATP
49
b
23
222
625
210
755
590
227
1362
226
595
635
n/a
b
4
c
58
n/a
(
(
(
F) Co(II)Cbi+
base-off
4c
100
n/a
b
b
b
G) Co(II)Cbl/LrPduO/MgATP
n/a
n/a
n/a
+
b
b
H) Co(II)Cbi /SeCobA/MgATP
4c
n/a
a
b
Parameters for 4c Co(II)rrinoids generated by SeCobA and LrPduO from refs 36 and 37 are shown for comparison (G and H). Not applicable.
41
reported previously. These features are identical to those
WT
displayed by 4c Co(II)Cbl bound to EutT /Fe + MgATP, but
WT
the minuscule amount of Fe(III) in our EutT /Zn samples
allows for the observation of additional peaks at lower fields. A
fit of this spectrum yields g-shifts and A(Co) values (Table 2E)
that are significantly larger than those reported for the 4c
Co(II)rrinoids generated in the active site of the other ACATs
(Table 2G and H). The increased rhombicity of the g-tensor
(
resulting in g > g > g ) is noteworthy, as the magnitudes of
x
y
z
the g-shifts are sensitive to the conformation of 4c Co(II)-
rrinoids. Particularly intriguing are the very large values
obtained for g and A (Co) (3.61 and 1362 MHz, respectively).
x
x
These values suggest that the conformation of the 4c Co(II)Cbl
species in EutT is substantially different from that adopted in
33,34
the other ACATs.
Consistent with the findings summarized above, the MCD
WT
spectrum of Co(II)Cbl in the presence of EutT /Zn and an
approximately equimolar amount of MgATP is dominated by a
base-off Co(II)Cbl species that is generated in the enzyme
active site (Figure 6C). Intriguingly, this spectrum is not
+
superimposable on that of Co(II)Cbi , as the features near
−
1
1
6 000 and 21 000 cm are red-shifted by ∼500 and ∼300
−1
cm , respectively, in the spectrum of the enzyme-bound base-
off Co(II)Cbl species (Figure 6E). These shifts are very similar
WT
to those observed for Co(II)Cbl in the presence of EutT /Zn
and substoichiometric MgATP (Figure 6D). Based on our EPR
results, these changes can be attributed to the formation of a
slightly perturbed base-off Co(II)Cbl species in the active site
WT
of EutT /Zn. Addition of a 10-fold molar excess of MgATP
WT
to a sample of Co(II)Cbl in the presence of EutT /Zn and
equimolar MgATP leads to the appearance of an MCD
spectrum that is dominated by contributions from a 4c
Co(II)Cbl species (Figure 6B) and nearly indistinguishable
WT
from that obtained with EutT /Zn purified in the presence of
Figure 6. MCD spectra of Co(II)Cbl obtained in the presence of
EutT /Zn and various MgATP concentrations. (A) ∼5 mM MgATP
WT
a large (∼10-fold) molar excess MgATP (Figure 6, A). This
observation suggests that the base-off Co(II)Cbl generated in
is a possible intermediate in the formation of 4c
Co(II)Cbl. This unique base-off Co(II)Cbl species is present in
(∼10-fold molar excess), (C) equimolar amounts of MgATP and
EutT, and (D) substoichiometric (<10 μM) MgATP. Trace (B) was
obtained after the addition of MgATP to the sample used in (C) to
increase the concentration to ∼8 mM (>10-fold excess). The MCD
WT
EutT
WT
the active site of EutT /Zn even under low MgATP
+
spectrum of Co(II)Cbi (E), a base-off Co(II)Cbl analogue, is shown
conditions (i.e., >30-fold Co(II)Cbl:MgATP molar excess),
indicating that MgATP does not have to be present in the
active site before the enzyme can bind Co(II)Cbl and displace
its DMB moiety.
for comparison. Primary spectroscopic features of this species are
highlighted by dashed vertical lines.
Effect of Amino Acid Substitutions on 4c Co(II)Cbl
Formation. The activity of EutT and selected variants has
previously been shown to correlate with the degree of Fe(II) or
Zn(II) incorporation into the enzyme. The C79, C80, C83, and
H67 residues that are part of the conserved HX CCX C(83)
motif of EutT were found to be particularly important for
WT
11
2
WT
G
DOI: 10.1021/jacs.5b11708
J. Am. Chem. Soc. XXXX, XXX, XXX−XXX

Journal of the American Chemical Society
Article
the binding of divalent transition metal ions, as variants
containing alanine substitutions at these positions displayed
significantly lower metalation levels than the wild-type
However, as judged on the basis of the relative intensities of the
MCD features associated with the 4c fraction of Co(II)Cbl, the
5c → 4c Co(II)Cbl conversion yields achieved by these latter
enzyme. Because the three-dimensional structure of EutTWT
has yet to be determined, the nature of the divalent metal ion
binding site and the mechanism by which metal incorporation
into the HX CCX C(83) motif modulates the catalytic activity
2
7
WT
variants are lower than that observed for EutT /Fe. Overall,
the 5c → 4c Co(II)Cbl conversion yields correlate well with
the degree of Fe(II) incorporation into these variants as
determined previously by elemental analysis, which revealed
11
2
WT
C80A
C83A
of EutT are currently unknown. In order to address this
issue, we studied the effects of alanine substitutions within this
motif on the binding of Co(II)Cbl to EutT/Fe and the yield of
that EutT
and EutT
bind negligible amounts of Fe(II),
C79A
H67A
while EutT
and EutT
incorporate ∼50% less Fe(II)
than the wild-type enzyme. However, from these data alone it
remained unclear if the alanine substitutions of residues within
the HX CCX C(83) motif cause a decrease in the 5c → 4c
4
c Co(II)Cbl species.
The MCD spectra of Co(II)Cbl in the presence of EutT/Fe
11
2
variants possessing a single alanine substitution at either the
H67, H75, C79, C80, or C83 position are show in Figure 7 (the
Co(II)Cbl conversion yield by suppressing Co(II)Cbl binding
to the enzyme active site or by impeding the removal of the
DMB moiety from the Co(II) ion.
The MCD spectra of Co(II)Cbl in the presence of EutTC80A/
C83A
Fe and EutT /Fe are virtually superimposable on that of free
Co(II)Cbl (Figure 7), lacking any features attributable to base-
off Co(II)Cbl. This observation indicates that both the
C80A
C83A
EutT
/Fe and the EutT
/Fe variants do not bind
Co(II)Cbl. In contrast, the MCD spectrum of Co(II)Cbl in
C79A
the presence of EutT /Fe contains qualitatively similar,
though considerably smaller contributions from 4c Co(II)Cbl
WT
as the spectrum obtained with EutT /Fe. Despite the modest
fraction of 4c Co(II)Cbl generated by this variant (about 50%
WT
of EutT /Fe), the characteristic features from 4c Co(II)Cbl
can be readily distinguished from those associated with 5c
Co(II)Cbl species. Removal of the 4c Co(II)Cbl contributions
to this MCD spectrum (via subtraction of the properly scaled
WT
spectrum of 4c Co(II)Cbl in the presence of EutT /Fe)
results in a spectrum that is essentially superimposable on that
of free Co(II)Cbl (Figure S8), indicating that the remaining
Co(II)Cbl fraction is not bound to the enzyme. Thus, the 50%
decrease in the 5c → 4c Co(II)Cbl conversion yield in
response to the C79A substitution likely stems from the ∼50%
decrease in Fe(II) incorporation, as the Fe(II)-bound fraction
of the protein appears capable of generating a similar 4c
WT
Co(II)Cbl species as EutT /Fe. These findings indicate that
C79 is not critical for maintaining the structure of the active site
to the same degree as are the C80 and C83 residues, in
particular with regards to the binding of Co(II)Cbl and the
removal of the DMB moiety. In support of this conclusion,
C79A
WT
EutT /Fe has been found to be nearly as active as EutT
/
Fe. It is therefore not surprising that residue C79 is not
conserved among the known EutT ACATs.
A detailed analysis of the MCD spectra of Co(II)Cbl in the
H67A
H75A
presence of EutT
/Fe and EutT
/Fe (Figures S9 and
S10) reveals that these variants produce a significant fraction of
base-off Co(II)Cbl, which is responsible for the features at
Figure 7. MCD spectra of Co(II)Cbl obtained in the presence of
EutT /Fe and selected variants (with >5 mM MgATP unless
−1
WT
∼16 000 and ∼20 600 cm . Moreover, the presence of base-off
otherwise noted). The primary spectroscopic features due to 4c
Co(II)Cbl in these variants is evidenced by the apparent red-
WT
−1
Co(II)Cbl species in the presence of EutT /Fe (A) are highlighted
shift of the feature observed at ∼16 300 cm (Figure 7, feature
a). The absence of a feature at 21 000 cm− shows that no base-
1
by dashed vertical lines. Additional spectroscopic changes in reference
to Co(II)Cbl in the absence of protein (H) are also highlighted.
on Co(II)Cbl remains unbound in solution, while the low
−1
intensity of the δ band at ∼12 000 cm indicates that only a
small fraction of 4c Co(II)Cbl is formed in the active sites of
these variants. EPR data obtained for samples prepared under
the same conditions confirm the presence of base-off and 4c
Co(II)Cbl species, but also contain sizable contributions from
base-on Co(II)Cbl (Figure S4). Because these spectra are very
H75A variant was included in this study because we initially
assumed that residue His75 was part of the conserved divalent
metal ion binding motif and the imidazole side chain of this
residue could potentially coordinate to the divalent metal ion).
C80A
While the spectra of Co(II)Cbl in the presence of EutT /Fe
C83A
WT
and EutT /Fe only contain features from 5c Co(II)Cbl,
similar to that obtained with EutT /Fe (except for differences
C79A
H67A
H75A
those obtained with EutT /Fe, EutT
/Fe, and EutT
/
in relative intensities due to variations in the different
Fe additionally contain contributions from 4c Co(II)Cbl.
Co(II)Cbl populations) and the corresponding MCD spectra
H
DOI: 10.1021/jacs.5b11708
J. Am. Chem. Soc. XXXX, XXX, XXX−XXX

Journal of the American Chemical Society
Article
indicate that all species are enzyme-bound, we conclude that
the Co(II)Cbl binding motif of the enzyme is not disrupted by
the H75A and H67A substitutions. However, since these
substitutions cause an increase in the relative populations of
enzyme-bound base-off and base-on Co(II)Cbl species at the
expense of the 4c Co(II)Cbl fraction, residues H75 and H67
appear to play a role in the mechanism of 4c Co(II)Cbl
formation by promoting axial ligand dissociation from 5c
Co(II)Cbl either directly or indirectly by enhancing the degree
of Fe(II) and Zn(II) incorporation into EutT.
characterized to date (Table 2). While these results provide
compelling evidence that the 4c Co(II)Cbl species generated in
WT
the EutT
active site adopts a unique conformation, it is
difficult to interpret the corresponding g- and A-values in terms
of specific structural perturbations because a proper description
of the electronic structure of this species requires the use of
wave function-based multireference techniques and a non-
perturbative approach for the calculation of spin−orbit
40
coupling effects. The available crystal structures for LrPduO
and SeCobA in the presence of Co(II)Cbl and MgATP show
that the corrin ring of the 4c Co(II)Cbl species generated in
these enzymes adopts a planar conformation in which the four
DISCUSSION
■
33,34
The EutTWT Active Site Promotes DMB Dissociation to
Generate Base-Off Co(II)Cbl. As shown in this study, the
combined use of MCD and EPR spectroscopies is well suited to
characterize the base-off, base-on, and 4c Co(II)Cbl species
nitrogen atoms have become nearly symmetry-equivalent.
A computational analysis of 4c Co(II)Cbl indicated that in this
species the unpaired spin density is localized in an MO that also
2
has sizable contributions from orbitals other than the Co 3dz
orbital. In the case of 4c Co(II)Cbl generated in the EutT
WT
WT
generated in the EutT active site and to determine their
active site, the large difference between the g and g values
relative populations. Our EPR data indicate that in the presence
x
y
WT
indicates that the Co 3d - and Co 3d -based MOs have
of EutT and substoichiometric MgATP, as much as 60% of
xz
yz
substantially different energies, consistent with a nonplanar
conformation of the corrin ring. A perturbed conformation of
the corrin ring could also explain the largely red-shifted α-band
observed in the Abs spectrum of this species, as the corrin π*-
based HOMO that serves as the donor orbital for this transition
Co(II)Cbl is present in the base-off conformation. Never-
WT
theless, we have recently shown that EutT /Zn binds
+
Co(II)Cbi very weakly and is unable to adenoslyate this
41
naturally occurring base-off Co(II)Cbl analogue. Additionally,
we observed small but significant differences between the MCD
+
WT
contains small contributions from the Co 3d and Co 3dyz
spectra of free Co(II)Cbi and the EutT /Zn-bound base-off
Co(II)Cbl species. These findings led us to propose that
specific interactions between certain amino acid residues and
the nucleotide loop are required to produce base-off Co(II)Cbl
xz
36
orbitals.
Another intriguing difference between the 4c Co(II)Cbl
WT
species formed in the EutT active site and those observed in
41
the presence of other ACATs is the large difference in intensity
in the enzyme active site. These interactions could introduce
changes to the conformation of the corrin ring, especially in the
region of C(17) that serves as the anchor of the nucleotide loop
−1
of the δ-band at ∼12 000 cm in the corresponding MCD
spectra (Figure 2A). Even in the presence of a >50-fold molar
excess of MgATP, conditions under which complete conversion
to 4c Co(II)Cbl should occur, the MCD intensity of the δ-band
(
Figure 1), as well as alter the environment around the H O
2
ligand in base-off Co(II)Cbl, both of which could contribute to
the small MCD and EPR spectral differences that exist between
free and enzyme-bound base-off Co(II)Cbl. A more con-
strained conformation of the corrin ring of the base-off
remains less than 0.22 mM− cm for all EutT samples studied.
For the 4c Co(II)Cbl species formed in the active sites of
SeCobA and LrPduO, the intensity of this feature was
1
−1
WT
−1
−1
Co(II)Cbl species generated in the EutT active site would
previously estimated to be 0.505 mM cm (by accounting
for the relative populations of 4c and 5c Co(II)Cbl). The lower
MCD intensity of the δ-band in the spectrum of 4c Co(II)Cbl
also explain the observation of resolved hyperfine structure in
the g region of the corresponding EPR spectrum (Figure 4).
x
41
Similarly, the slight differences between the MCD and EPR
generated in the active site of EutT species remains puzzling,
WT
spectra of the base-on Co(II)Cbl species bound to EutT and
but likely reflects the uniquely perturbed conformation of the
4c Co(II)Cbl substrate bound to this enzyme.
free Co(II)Cbl could arise from enzyme-induced structural
perturbations to the corrin ring prior to DMB dissociation. A
constrained corrin ring conformation may be a critical
prerequisite for the removal of the DMB ligand from the
Co(II) ion, as otherwise the Co(II)Cbl substrate could simply
reorient in response to external forces acting on the nucleotide
loop.
Comparison of the Co(II) → Co(I)Cbl Conversion
Mechanisms of Known ACATs. In all ACATs studied to
date, the redox tuning of the Co(II)Cbl substrate required for
the formation of the essential Co(I)Cbl intermediate is
achieved by inducing a complete dissociation of the axially
bound DMB ligand to generate a square planar 4c Co(II)-
rrinoid species that can be readily reduced by intracellular
Nature of the 4c Co(II)Cbl Species Generated in the
EutT Active Site. The 4c Co(II)Cbl species generated in the
active sites of other ACATs such as LrPduO and SeCobA
exhibit large (∼600 to ∼800 MHz) cobalt hyperfine coupling
parameters, A(Co), consistent with a high degree of localization
of the unpaired spin density on the Co(II) ion in the absence of
axial ligands. The unusually large g-shifts from the free-electron
value of 2.0023 displayed by these species (g ≈ g ≫ g ∼ 2.0)
26
reducing agents (Figure 8). In the case of LrPduO, X-ray
crystallographic studies revealed that the side chains of residues
F112, F187, and V186 form a hydrophobic “wall” upon binding
of Co(II)Cbl and MgATP, serving to block solvent access to
the Coα face of the corrin ring, and a salt bridge interaction
between D35 and R128 near the corrin ring binding site was
found to be critical for promoting the dissociation of the DMB
x
y
z
35,38
have been shown to reflect an increase in the extent of spin−
moiety.
A similar mechanism was proposed for SeCobA,
orbit mixing between the ground state and LF excited states
where residues F91 and W93 were found to be important for
generating 4c Co(II)rrinoids and for properly positioning the
2
z
due to a large stabilization of the Co 3d -based MO in this
3
6,37
conformation.
c Co(II)Cbl species generated in the active site of EutT
displays a rhombic g-tensor (g > g > g ) with the largest g
To this end, it is interesting to note that the
Co(I) intermediate that is subsequently formed during
WT
33,39
4
turnover.
However, the differences in the specific
architectures of these active sites have an effect on the relative
x
y
z
spread and A(Co) values observed for any Co(II)rrinoid
populations of 4c Co(II)rrinoids that are being formed.
I
DOI: 10.1021/jacs.5b11708
J. Am. Chem. Soc. XXXX, XXX, XXX−XXX

Journal of the American Chemical Society
Article
removal of the DMB moiety, the formation of a base-off
intermediate should facilitate the formation of 4c Co(II)Cbl,
which would explain the high 5c → 4c yield observed for
WT
WT
EutT (>95% for EutT /Fe + MgATP). In this proposed
mechanism, specific residues in the enzyme active site serve to
lock the corrin ring in place, giving rise to the unique
spectroscopic features observed experimentally, with additional
protein motifs sequestering the DMB moiety away from the
+
Co(II) ion (Figure 8B). In the case of Co(II)Cbi , the absence
of the nucleotide loop and terminal DMB moiety prevent the
active site from adopting the proper conformation needed for
the formation of a 4c Co(II)Cbi+ species in response to
MgATP binding near the Coβ face of the corrin ring. Given
+
that Co(II)Cbi has been found to bind to EutT much less
Figure 8. Proposed mechanisms for the formation of 4c Co(II)Cbl
employed by the different ACAT families. In the case of the PduO-
and CobA-type ACATs (A), spectroscopic and kinetic studies
tightly than Co(II)Cbl, it is plausible that the former can
respond to MgATP binding to the active site simply by
repositioning itself without loss of the axial ligand. In the case
of enzyme-bound 5c base-off Co(II)Cbl, MgATP binding to
the active site in a position suitable to react with the
subsequently formed Co(I)Cbl intermediate could trigger the
dissociation of an axially bound water molecule on the Coβ face
of the corrin ring (Figure 8B), thereby permitting the enzyme
to control the timing of Co(I)Cbl formation and suppress the
likelihood of undesired side-reactions by this “supernucleo-
phile”. While the specific amino acid residues used to stabilize
enzyme-bound 5c base-off and 4c Co(II)Cbl species remain
unknown in the absence of a crystal structure of EutT, it is
tempting to speculate that the divalent metal site and
HX CCX C(83) motifs are involved in the removal of the
3
,37−39
revealed that the substrate-free enzyme (i) must first bind ATP (ii)
before it can bind Co(II)Cbl. Crystal structures of the tertiary
33−35
complex
(iii) provided evidence that in both enzymes a conserved
phenylalanine residue blocks the position originally occupied by the
DMB moiety in free Co(II)Cbl. In these structures the DMB moiety is
unresolved, which argues against the presence of a specific nucleotide
binding pocket as expected on the basis of the fact that these enzymes
+
can adenosylate both Co(II)Cbl and Co(II)Cbi . In the case of EutT,
the results obtained in the present study indicate that Co(II)Cbl can
bind to the enzyme active site in the absence of ATP (i) to form a
base-off Co(II)Cbl species in which a water molecule is presumably
bound in the upper axial position (ii). In this binary complex, the
DMB moiety is sequestered in a pocket lined by residues of the
HX CCX C(83) motif whose architecture changes in response to
11
2
11
2
Fe(II) or Zn(II) ion binding. The subsequent binding of ATP to the
enzyme active site triggers the dissociation of the axially bound water
molecule to form 4c Co(II)Cbl (iii). Note that our data do not rule
out the possibility that in the case of EutT the order of substrate
binding is unimportant.
DMB moiety, and thus are important for 4c Co(II)Cbl
formation. Studies of Co(II)-substituted EutT, aimed at
elucidating the structure of this metal site and evaluating its
role in the catalytic cycle of EutT, will be presented elsewhere.
ASSOCIATED CONTENT
■
Although SeCobA and LrPduO achieve very similar relative
*
S
Supporting Information
+
yields of 4c Co(II)Cbi species (∼50%), LrPduO is able to
The Supporting Information is available free of charge on the
ACS Publications website at DOI: 10.1021/jacs.5b11708.
generate ∼40% of 4c Co(II)Cbl while in the case of SeCobA
36,37,39
this yield is decreased to ∼8%.
The higher 5c → 4c
Details about compositions of all samples used in this
study, additional low-temperature Abs, MCD, and EPR
+
conversion yields for Co(II)Cbi compared to Co(II)Cbl have
generally been attributed to the weaker axial bond in
WT
spectra of Co(II)Cbl in the presence of EutT /Zn,
+
Co(II)Cbi and the fact that a higher level of coordination
among different active site residues is required to remove the
DMB moiety from the Co(II) ion of Co(II)Cbl. From this
WT
EutT /Fe, and selected variants under various con-
38
ditions, and Gaussian deconvolutions of the MCD
spectra of Co(II)Cbl obtained in the presence of several
EutT/Fe variants (PDF)
WT
perspective, it is rather unusual that EutT + can only
adenoslylate Co(II)Cbl, and not Co(II)Cbi . Although
+
previous studies revealed that Co(II)Cbi does bind to the
EutT active site, albeit weakly, the axially bound water ligand
is retained in this process.
WT
AUTHOR INFORMATION
■
41
Corresponding Author
brunold@chem.wisc.edu
The results obtained in the present study indicate that
binding of Co(II)Cbl to the EutTWT active site and the
formation of a large population of a 5c base-off species occur
even in the absence of MgATP. This is in stark contrast to the
ordered binding scheme observed for SeCobA and LrPduO,
where ATP must bind first for the enzyme to adopt a
conformation suitable for binding of Co(II)rrinoids (Figure
However, 4c Co(II)Cbl is generated only in the
presence of an excess of cosusbtrate MgATP, with higher
concentrations resulting in higher yields. These observations
suggest a mechanism for EutT in which the active site forms
a unique base-off Co(II)Cbl species prior to MgATP
incorporation that serves as a precursor to 4c Co(II)Cbl
Figure 8B). As removal of the H O ligand is more facile than
*
Notes
The authors declare no competing financial interest.
ACKNOWLEDGMENTS
■
This work was supported in part by the National Science
Foundation Grants MCB-238530 (to T.C.B.) and CHE-
0840494. I.G.P. was supported in part by Grant
5T32GM08349 via the Biotechnology Training Grant at UW-
Madison. T.C.M was supported in part by the National
Institutes of Health Grant T32 GM008505 (Chemistry and
Biology Interface Training Grant) and by the National
Institutes of Health Grant R37-GM40313 to J.C.E.-S.
3
,33,34
8A).
WT
(
2
J
DOI: 10.1021/jacs.5b11708
J. Am. Chem. Soc. XXXX, XXX, XXX−XXX

Journal of the American Chemical Society
Article
(
33) Moore, T. C.; Newmister, S. a.; Rayment, I.; Escalante-
REFERENCES
■
Semerena, J. C. Biochemistry 2012, 51 (48), 9647−9657.
34) St. Maurice, M.; Mera, P.; Park, K.; Brunold, T. C.; Escalante-
Semerena, J. C.; Rayment, I. Biochemistry 2008, 47 (21), 5755−5766.
35) Mera, P. E.; St Maurice, M.; Rayment, I.; Escalante-Semerena, J.
C. Biochemistry 2009, 48 (14), 3138−3145.
36) Stich, T. A.; Buan, N. R.; Escalante-Semerena, J. C.; Brunold, T.
C. J. Am. Chem. Soc. 2005, 127 (24), 8710−8719.
37) Park, K.; Mera, P. E.; Escalante-Semerena, J. C.; Brunold, T. C.
Biochemistry 2008, 47 (34), 9007−9015.
38) Park, K.; Mera, P. E.; Escalante-Semerena, J. C.; Brunold, T. C.
Inorg. Chem. 2012, 51, 4482−4494.
39) Pallares, I. G.; Moore, T. C.; Escalante-Semerena, J. C.; Brunold,
T. C. Biochemistry 2014, 53, 7969−7982.
40) Liptak, M. D.; Fleischhacker, A. S.; Matthews, R. G.; Telser, J.;
Brunold, T. C. J. Phys. Chem. B 2009, 113 (15), 5245−5254.
41) Park, K.; Mera, P. E.; Moore, T. C.; Escalante-Semerena, J. C.;
Brunold, T. C. Angew. Chem., Int. Ed. 2015, 54 (24), 7158−7161.
42) Buan, N. R.; Suh, S.; Escalante-Semerena, J. C. Microbiology
(
1) Sheppard, D. E.; Penrod, J. T.; Bobik, T.; Roth, J. R.; Kofoid, E. J.
Bacteriol. 2004, 186 (22), 7635−7644.
(
(
1
(
2
2) Buan, N. R.; Suh, S.; Escalante-Semerena, J. C. J. Bacteriol. 2004,
86 (17), 5708−5714.
3) Mera, P. E.; Escalante-Semerena, J. C. Appl. Microbiol. Biotechnol.
010, 88 (1), 41−48.
4) Stich, T. A.; Brooks, A. J.; Buan, N. R.; Brunold, T. C. J. Am.
Chem. Soc. 2003, 125 (19), 5897−5914.
5) Conrad, K. S.; Brunold, T. C. Inorg. Chem. 2011, 50 (18), 8755−
766.
6) Pallares, I. G.; Brunold, T. C. Inorg. Chem. 2014, 53 (14), 7676−
691.
7) Stich, T. A.; Buan, N. R.; Brunold, T. C. J. Am. Chem. Soc. 2004,
26 (31), 9735−9749.
8) Mori, K.; Bando, R.; Hieda, N.; Toraya, T. J. Bacteriol. 2004, 186
20), 6845−6854.
9) Banerjee, R.; Ragsdale, S. W. Annu. Rev. Biochem. 2003, 72, 209−
47.
10) Yeates, T. O.; Crowley, C. S.; Tanaka, S. Annu. Rev. Biophys.
010, 39, 185−205.
11) Chowdhury, C.; Sinha, S.; Chun, S.; Yeates, T. O.; Bobik, T. A.
Microbiol. Mol. Biol. Rev. 2014, 78 (3), 438−468.
12) Kerfeld, C. A.; Heinhorst, S.; Cannon, G. C. Annu. Rev.
Microbiol. 2010, 64, 391−408.
13) Sampson, E. M.; Bobik, T. A. J. Bacteriol. 2008, 190 (8), 2966−
971.
14) Thiennimitr, P.; Winter, S. E.; Winter, M. G.; Xavier, M. N.;
(
(
(
(
(
8
(
7
(
1
(
(
(
2
(
2
(
(
(
(
(
(
2004, 186 (17), 5708−5714.
(43) Blommel, P. G.; Fox, B. G. Protein Expression Purif. 2007, 55 (1),
53−68.
(44) Nilges, M. J. Thesis in Chemistry, University of Illinois, Urbana-
Champaign, IL, 1979.
(
(45) Pratt, J. M. Inorganic Chemistry of Vitamin B12; Academic Press
Inc: London, 1972.
(46) Brooks, A. J.; Vlasie, M.; Banerjee, R.; Brunold, T. C. J. Am.
Chem. Soc. 2005, 127 (47), 16522−16528.
(47) Stich, T. A.; Yamanishi, M.; Banerjee, R.; Brunold, T. C. J. Am.
Chem. Soc. 2005, 127 (21), 7660−7661.
(
2
(
Tolstikov, V.; Huseby, D. L.; Sterzenbach, T.; Tsolis, R. M.; Roth, J.
R.; Baumler, A. J. Proc. Natl. Acad. Sci. U. S. A. 2011, 108 (42), 17480−
7485.
1
(
(
15) Garsin, D. A. Nat. Rev. Microbiol. 2010, 8 (4), 290−295.
16) Shibata, N.; Tamagaki, H.; Hieda, N.; Akita, K.; Komori, H.;
Shomura, Y.; Terawaki, S.-I.; Mori, K.; Yasuoka, N.; Higuchi, Y.;
Toraya, T. J. Biol. Chem. 2010, 285 (34), 26484−26493.
(
17) Buan, N. R.; Escalante-Semerena, J. C. J. Biol. Chem. 2006, 281
(
25), 16971−16977.
18) Tsoy, O.; Ravcheev, D.; Mushegian, A. J. Bacteriol. 2009, 191
23), 7157−7164.
19) Warren, M. J.; Raux, E.; Schubert, H. L.; Escalante-Semerena, J.
C. Nat. Prod. Rep. 2002, 19 (4), 390−412.
20) Escalante-Semerena, J. C.; Suh, S. J.; Roth, J. R. J. Bacteriol.
990, 172 (1), 273−280.
21) Jorge-Finnigan, A.; Aguado, C.; San
Richard, E.; Merinero, B.; Gamez, A.; Banerjee, R.; Desviat, L. R.;
Ugarte, M.; Perez, B. Hum. Mutat. 2010, 31 (9), 1033−1042.
22) Padovani, D.; Labunska, T.; Palfey, B. A.; Ballou, D. P.;
Banerjee, R. Nat. Chem. Biol. 2008, 4 (3), 194−196.
23) Fonseca, M. V.; Escalante-Semerena, J. C. J. Bacteriol. 2000, 182
15), 4304−4309.
24) Liptak, M. D.; Brunold, T. C. J. Am. Chem. Soc. 2006, 128 (28),
144−9156.
25) Faure, D.; Lexa, D.; Savea
Interfacial Electrochem. 1982, 140 (2), 297−309.
26) Mera, P. E.; Escalante-Semerena, J. C. J. Biol. Chem. 2010, 285
5), 2911−2917.
27) Moore, T. C.; Mera, P. E.; Escalante-Semerena, J. C. J. Bacteriol.
014, 196 (4), 903−910.
28) Buan, N. R.; Escalante-Semerena, J. C. J. Biol. Chem. 2005, 280
49), 40948−40956.
29) Mayhe, S. G.; Foust, G. P.; Massey, V. J. Biol. Chem. 1969, 244
3), 803−810.
30) Hoover, D. M.; Jarrett, J. T.; Sands, R. H.; Dunham, W. R.;
Ludwig, M. L.; Matthews, R. G. Biochemistry 1997, 36 (1), 127−138.
31) Mera, P. E.; St. Maurice, M.; Rayment, I.; Escalante-Semerena, J.
C. Biochemistry 2007, 46 (48), 13829−13836.
32) Fonseca, M. V.; Buan, N. R.; Horswill, A. R.; Rayment, I.;
Escalante-Semerena, J. C. J. Biol. Chem. 2002, 277 (36), 33127−33131.
(
(
(
(
1
(
́
chez-Alcudia, R.; Abia, D.;
́
́
(
(
(
(
9
(
́
nt, J.-M. J. Electroanal. Chem.
(
(
(
2
(
(
(
(
(
(
(
K
DOI: 10.1021/jacs.5b11708
J. Am. Chem. Soc. XXXX, XXX, XXX−XXX