CfbA promotes insertion of cobalt and nickel into ruffled tetrapyrroles: In vitro

Article abstract of DOI:10.1039/c9dt03601f

The nickel chelatase CfbA is the smallest member of the chelatase family, but the mechanism by which this enzyme inserts nickel into sirohydrochlorin is unknown. In order to gain mechanistic insight, metal binding, tetrapyrrole binding, and enzyme activity were characterized for a variety of substrates using several spectroscopic and computational approaches. Mass spectrometery and magnetic circular dichroism experiments revealed that CfbA binds an octahedral, high-spin metal substrate. UV/Vis absorption spectroscopy demonstrated that the enzyme binds a wide range of tetrapyrrole substrates and perturbs their electronic structures. Based upon activity assays, CfbA promotes insertion of cobalt and nickel into several tetrapyrroles, including cobalt insertion into protopophyrin IX. Finally, density functional theory models were developed which strongly suggest that observed spectral changes upon binding to the enzyme can be explained by tetrapyrrole ruffling, but not deprotonation or saddling. The observation of an octahedral, high-spin metal bound to CfbA leads to a generalization for all class II chelatases: these enzymes bind labile metal substrates and metal desolvation is not a rate-limiting step. The conclusion that CfbA ruffles its tetrapyrrole substrate reveals that the CfbA mechanism is different from that currently proposed for ferrochelatase, and identifies an intriguing correlation between metal substrate specificity and tetrapyrrole distortion mode in chelatases.

Full text of DOI:10.1039/c9dt03601f

Dalton
Transactions
View Article Online
View Journal
PAPER
CfbA promotes insertion of cobalt and nickel into
ruffled tetrapyrroles in vitro†
Cite this: DOI: 10.1039/c9dt03601f
Ariel E. Schuelke-Sanchez, Alissa A. Stone and Matthew D. Liptak
*
The nickel chelatase CfbA is the smallest member of the chelatase family, but the mechanism by which
this enzyme inserts nickel into sirohydrochlorin is unknown. In order to gain mechanistic insight, metal
binding, tetrapyrrole binding, and enzyme activity were characterized for a variety of substrates using
several spectroscopic and computational approaches. Mass spectrometery and magnetic circular dichro-
ism experiments revealed that CfbA binds an octahedral, high-spin metal substrate. UV/Vis absorption
spectroscopy demonstrated that the enzyme binds a wide range of tetrapyrrole substrates and perturbs
their electronic structures. Based upon activity assays, CfbA promotes insertion of cobalt and nickel into
several tetrapyrroles, including cobalt insertion into protopophyrin IX. Finally, density functional theory
models were developed which strongly suggest that observed spectral changes upon binding to the
enzyme can be explained by tetrapyrrole ruffling, but not deprotonation or saddling. The observation of
an octahedral, high-spin metal bound to CfbA leads to a generalization for all class II chelatases: these
enzymes bind labile metal substrates and metal desolvation is not a rate-limiting step. The conclusion that
CfbA ruffles its tetrapyrrole substrate reveals that the CfbA mechanism is different from that currently pro-
posed for ferrochelatase, and identifies an intriguing correlation between metal substrate specificity and
tetrapyrrole distortion mode in chelatases.
Received 6th September 2019,
Accepted 13th December 2019
DOI: 10.1039/c9dt03601f
rsc.li/dalton
metals, so there is significant motivation to discover efficient
routes to a broad range of metal tetrapyrroles.
Introduction
Metal tetrapyrroles catalyse a myriad of important chemical
In biological systems, a family of enzymes known as chela-
transformations both in biological and abiological environ- tases are involved in the biosynthesis of metal tetrapyrroles.
ments. A fundamental reason why metal tetrapyrroles can cata- In vivo each of these enzymes must catalyse the insertion of a
1
0
lyse diverse chemistry is the incorporation of a transition specific metal into a specific tetrapyrrole ligand. Chelatases
1
1
metal that can access a larger number of reaction mechanisms have been divided into three classes. Class I chelatases
than a main group element. In biological systems, the combi- require ATP and are involved in the biosynthesis of chlorophyll
1
2,13
nation of iron and porphyrin, or heme, catalyses a variety of and cobalamin.
Class II chelatases do not require
1
oxygen-dependent reactions, including hydroxylation of additional substrates beyond metal and tetrapyrrole. These
alkanes, degradation of peroxides to alcohols, and reduction enzymes either insert iron into protoporphyrin IX (PPIX), or
of molecular oxygen to water. Vitamin B12 and cofactor F430 a 3d metal into the more reduced sirohydrochlroin (SHC)
Finally, class III chelatases are multifunctional
methyl group transfer, organic radical rearrangements, and enzymes involved in siroheme biosynthesis.
The smallest member of the chelatase family is CfbA (pre-
pyrroles have shown promise catalysing reactions critical for viously annotated as CbiX ), and the enzymatic mechanism for
next-generation alternative energy applications, such as oxygen this nickel chelatase has remained elusive. Elucidation of the
2
3
14
4
1
5–17
are cobalt and nickel tetrapyrroles, respectively, that catalyse ring.
5
6
18
7
methanogenesis. From an abiological perspective, metal tetra-
S
8
8
9
reduction, hydrogen evolution, and methane oxidation. All CfbA mechanism is of particular importance because this
of these transformations rely upon earth-abundant transition enzyme is proposed to be the ancestor of all class II chela-
1
1
tases. In support of this proposal, random in-frame muta-
genesis of CfbA can produce synthetic proteins that resemble
either the cobalt chelatase CbiK or the iron chelatase HemH.
With regards to the substrate scope of CfbA, this enzyme has
1
9
Department of Chemistry, University of Vermont, Burlington, Vermont 05405, USA.
E-mail: matthew.liptak@uvm.edu
11,19
11,19,20
been shown to insert iron,
cobalt,
and nickel into
†
Electronic supplementary information (ESI) available: DNA sequencing data,
1
1,17,21
SHC (Fig. 1).
Very little is known regarding the mecha-
FPLC trace, SDS-PAGE gels, additional UV/Vis Abs spectra, and Cartesian coordi-
nates. See DOI: 10.1039/c9dt03601f
nism of CfbA. Based upon a structural overlay with CbiK, CfbA
This journal is © The Royal Society of Chemistry 2019
Dalton Trans.

View Article Online
Paper
Dalton Transactions
2
3
plasmid encoding IsdG.
Lastly, David Waugh (National
r
Cancer Institute) provided a pRK793 (Amp ) plasmid encoding
2
4,25
S219V TEV protease.
A pET15b vector encoding CfbA with a TEV-cleavable
N-terminal His tag was prepared. The SalI and BamHI restric-
6
tion enzymes (New England Biolabs) were used to digest the
pIDTSMART plasmid encoding CfbA and the pET15b plasmid
encoding IsdG. Next, the cfbA and pET15b DNA fragments
were isolated via gel electrophoresis and recovered using the
Fig. 1 Cobalt and nickel insertion into five porphyrin substrates by QIAquick gel extraction kit (Qiagen). The cfbA gene was ligated
CfbA will be carefully assessed to gain insight into the enzymatic to pET15b by overnight incubation with T4 DNA ligase (New
mechanism. The porphyrin substrates have different side-chain substi-
tutions from one another.
England Biolabs) followed by purification using a QIAprep
spin mini-prep kit (Qiagen). DNA sequencing by the Vermont
Cancer Center DNA analysis facility confirmed the sequence of
r
the resulting CfbA-encoding pET15b (Amp ) vector (Table S1†).
2
0
is proposed to bind a 3d metal via two His residues. CfbA is
Recombinant CfbA and S219V TEV protease were over-
also proposed to bind a distorted SHC based upon the crystal
expressed in BL21-GOLD(DE3) cells (Strategene). BL21-GOLD
2
2
structure of product-bound enzyme.
We anticipate that
r
(
DE3) cells containing the CfbA-encoding pET15b (Amp )
insight into the CfbA mechanism will facilitate development of
artificial chelatases that catalyse the formation of non-natural
metal tetrapyrroles.
Metal and tetrapyrrole binding by CfbA, along with chela-
tase activity, for a diverse set of substrates has been assessed
in order to elucidate the enzymatic mechanism. This approach
is essential because some combinations of metal and tetrapyr-
role can combine spontaneously resulting in misleading
vector were grown at 37 °C in Luria–Bertani (LB) medium con-
taining 100 μg ampicillin using a ThermoScientific MaxQ incu-
bator shaker, which is similar to a previously reported pro-
20
cedure. CfbA expression was induced at an OD600 of 0.6 a.u.
by the addition of isopropyl-β-D-1-thiogalactopyranoside
(
IPTG), followed by growth at 30 °C for at least five hours.
These cells were harvested by centrifugation at 10 000g and
°C for 10 minutes using a ThermoScientific Sorvall Legend
4
17
specific activity data. Furthermore, binding of only one sub-
strate (metal or tetrapyrrole) could promote insertion in vitro.
This mechanism is exceedingly unlikely in vivo where metal
centrifuge. The second enzyme used in this work, S219V TEV
protease, was expressed and purified as previously
24,25
described.
1
0,17,21
and tetrapyrrole concentrations are regulated.
Metal
6
N-Terminal His -tagged CfbA was purified using affinity
binding by CfbA was measured with electrospray ionization
mass spectrometry (ESI-MS) and magnetic circular dichroism
MCD) spectroscopy. Tetrapyrrole binding by CfbA was charac-
chromatography. Cell pellets containing overexpressed CfbA
were resuspended in 50 mM 4-(2-hydroxyethyl)-1-piperazi-
neethanesulfonic acid (HEPES) pH 7.5 with 500 mM NaCl,
(
terized using UV/Vis absorption (Abs) spectroscopy. Finally,
chelatase activity for different combinations of metal and tetra-
pyrrole substrates was monitored using a combination of UV/
Vis Abs spectroscopy and ESI-MS. By interpreting these data
within the framework of density functional theory (DFT) and
time-dependent DFT (TDDFT) calculations, we gained insight
into the enzymatic mechanism of CfbA.
5
mM imidazole, and 5% glycerol (v/v). Enough phenylmetha-
nesulfonyl fluoride and lysozyme were added to this suspen-
sion to reach final concentrations of 1.0 mM and 0.5 mg mL
respectively, prior to sonication. A Branson Ultasonics sonifier
lysed the resuspended cells to release overexpressed CfbA from
the E. coli cells. The resulting lysate was clarified by centrifu-
gation at 15 000g and 4 °C followed by vacuum filtration with a
−1
,
0.45 μm membrane (Millipore). The clarified lysate was loaded
onto a nickel HiTrap HP 5 mL column (GE Healthcare) equili-
brated with 50 mM HEPES pH 7.5, 500 mM NaCl, 5 mM imid-
azole, and 5% glycerol (v/v) using an Äkta pure chromato-
graphy system (GE Healthcare). The loaded column was
washed with an imidazole gradient from 5 to 500 mM over
Experimental
Unless otherwise noted, we purchased all materials in this
work from Fisher Scientific and used them without further
purification.
2
00 mL. Pure, N-terminal His -tagged CfbA eluted at 500 mM
6
imidazole (Fig. S1†).
Cloning, expression, and purification of CfbA
6
The N-terminal His tag was removed from purified CfbA
DNA plasmids encoding Archaeoglobus fulgidus CfbA, S219V using S219V TEV protease. S219V TEV protease was added to
tobacco etch virus (TEV) protease, and Staphylococcus aureus His -tagged CfbA in an A280 ratio of 3 : 100 with sufficient
IsdG were obtained. We designed a synthetic gene encoding dithiothreitol and ethylenediaminetetraacetic acid to reach
6
2
5
CfbA that was optimized for Escherichia coli codon usage. final concentrations of 1 mM and 0.5 mM, respectively. The
Integrated DNA Technologies (Coralville, IA) synthesized this solution was dialyzed twice against 10 mM HEPES pH 7.5,
r
cfbA gene and inserted it into pIDTSMART (Kan ). Eric Skaar 500 mM NaCl at 4 °C for three hours each. After increasing the
r
(
Vanderbilt University) gifted a pET15b (Amp , Novagen) imidazole concentration to 75 mM, the dialyzed solution was
Dalton Trans.
This journal is © The Royal Society of Chemistry 2019

View Article Online
Dalton Transactions
Paper
loaded onto a HisPur Co-NTA gravity column equilibrated with copper/quartz sample cells and flash frozen in liquid nitrogen.
5
5
0 mM HEPES pH 7.5, 500 mM NaCl, 5 mM imidazole, and The samples were equilibrated at temperatures of 2, 5, 10, and
% glycerol (v/v). Following protein loading, the column was 20 K along with magnetic fields of ±1, 3, 5, and 7 T using an
washed with 50 mM HEPES pH 7.5, 500 mM NaCl, 75 mM Oxford SM4000-8T Spectromag. MCD data between 800 and
imidazole, and 5% glycerol (v/v) to elute pure, untagged CfbA 300 nm were acquired under these conditions using a Jasco
−1
(
Fig. S2†). The flow-throughs from the loading and wash steps J-815 CD spectrometer with a scanning speed of 200 nm min ,
containing pure, untagged CfbA were combined and the a bandwidth of 1 nm, a digital integration time of 0.25 s, and a
volume was reduced to 10 mL using Amicon stirred cell ultra- data pitch of 0.5 nm. For all MCD data, the negative magnetic
filtration cells (Millipore). The concentrated CfbA solution was field spectrum was subtracted from the positive field data and
dialyzed twice against 10 mM HEPES pH 7.5, 500 mM NaCl at the result was divided by 2 to remove the CD contribution.
4
°C for three hours each yielding pure, cleaved CfbA. CfbA
was >99% pure according to SDS-PAGE gel electrophoresis Tetrapyrrole binding
Fig. S3†). Based upon a Bradford assay using bovine serum
albumin (Pierce) as the standard, the CfbA yield was 30 mg per
L of LB medium.
(
Tetrapyrrole binding by CfbA was assessed using UV/Vis Abs
spectroscopy. Uroporphyrin I, uroporphyrin III, copropor-
phyrin I, and coproporphyrin III were purchased from Frontier
Scientific and dissolved in 50 mM Tris pH 8.0, 150 mM NaCl.
Protoporphyrin IX was also purchased from Frontier Scientific,
but dissolved in methanol prior to preparation of a solution
sample in 50 mM Tris pH 8.0, 150 mM NaCl, and 10% (v/v)
1
CfbA characterization
The molecular weight of pure, untagged CfbA was determined
using ESI-MS. 143 μM CfbA in 10 mM HEPES pH 7.5, 500 mM
NaCl was prepared. This sample was loaded onto a C18 guard
column equilibrated with 2% acetonitrile with 0.1% formic
acid (v/v) in water (v/v). A gradient from 2% to 98% acetonitrile
was used to separate CfbA prior to ESI-MS analysis using an
methanol. These tetrapyrrole solutions were diluted to 6 μM
based upon reported extinction coefficients,
26–30
wrapped in alu-
minium foil to prevent photodegradation, and used the same day
as preparation to mitigate aggregation (Fig. S4†). CfbA-bound tet-
rapyrrole samples were prepared by combining tetrapyrrole and
CfbA solutions in a 1 : 1 molar ratio. UV/Vis Abs spectra in the
AB-Sciex 4000
Q Trap LCMS/MS spectrometer (Applied
Biosystems). ESI-MS data for 600–1700 m/z was acquired in
positive ion mode and deconvoluted using BioAnalyst 1.5.
The secondary structure of CfbA was assessed using circular
dichroism (CD) spectroscopy. A 10 μM CfbA sample was
exchanged into 10 mM potassium phosphate (KP ) pH 7.5
using a PD-10 column (GE Healthcare) and loaded into a
800–350 nm region were recorded for protein-free and CfbA-
bound tetrapyrroles using a Cary 100 Bio spectrophotometer with
−1
a scan rate of 600 nm min , a bandwidth of 2 nm, an averaging
time of 0.1 s, and a 0.5 nm data interval.
i
1
mm path length quartz cuvette (Starna). The CD spectrum
Activity assays
was collected from 250 to 190 nm with a scan rate of 200 nm
−
1
The rates of metal insertion into uroporphyrin and copropor-
phyrin by CfbA were quantified using UV/Vis Abs spectroscopy.
The rates of cobalt and nickel insertion in the presence of
CfbA were measured in aqueous solution with 5–6 μM CfbA
and tetrapyrrole along with either 200 μM CoCl2 or 200 μM
min and a 0.5 nm data interval on a Jasco J-815 CD spectro-
meter. The secondary structure was estimated from the result-
ing data using the PLS method in the CD multivariate SSE
program (Jasco).
Metal binding
4
NiSO in 50 mM Tris pH 8.0 with 150 mM NaCl. The rates of
metal insertion in the absence of CfbA were measured in
nearly identical solutions that lacked enzyme. In order to
monitor the formation of metal tetrapyrrole, UV/Vis Abs spectra
were acquired using the same instrument and parameters as
described above. For metal insertion into the uroporphyrins,
and cobalt insertion into the coproporphyrins, new UV/Vis
spectra were acquired every minute for a total of 0.5–30 h. For
nickel insertion into the coproporphyrins, new spectra were
acquired every hour for 30–72 h. The assays were performed in
triplicate. The pseudo-first order rate constants, and their stan-
dard deviations, for CfbA-promoted and enzyme-free metal
insertion into uroporphyrin and coproporphyrin were deter-
mined by fitting the data to eqn (1) with Graph Pad Prism.
Metal binding by CfbA was evaluated using ESI-MS. Stock solu-
tions of 20 mM Co(II) or Ni(II) in 50 mM Tris pH 8.0, 150 mM
NaCl were prepared by adding sufficient CoCl2 or NiSO4,
respectively, to aqueous buffer. Next, the Co(II) or Ni(II) solu-
tion was combined with purified CfbA yielding 5 μM CfbA in
5
0 mM Tris pH 8.0, 150 mM NaCl, 200 μM CoCl or NiSO .
2
4
4
CfbA was loaded onto a C column equilibrated with 2% aceto-
nitrile with 0.1% formic acid (v/v) in water (v/v). The same
ESI-MS protocol described above for apoprotein characteriz-
ation was employed for the metal-loaded CfbA samples. The
resulting data was deconvoluted with the BioAnalyst 1.5 soft-
ware program.
The metal coordination site was characterized using MCD
spectroscopy. Sufficient CoCl
fied, concentrated CfbA to yield 400 μM CfbA in 50 mM KP pH
2
and glycerol were added to puri-
A ¼ ðA0 ꢀ BÞeꢀ þ B
kx
ð1Þ
i
8
.0, 50 mM NaCl, 400 μM CoCl
2
, and 60% glycerol (v/v). In where A is the absorbance intensity, A
0
is the absorbance at
addition, a sample of 400 μM CoCl
2
in the same aqueous time (x) equal to zero, B is the absorbance intensity at infinite
buffer was prepared. These two samples were loaded into time, and k is the rate constant.
This journal is © The Royal Society of Chemistry 2019
Dalton Trans.

View Article Online
Paper
Metal insertion into protoporphyrin IX by CfbA was charac-
Dalton Transactions
The ground- and excited-state potential energy surfaces
terized with UV/Vis Abs spectroscopy and ESI-MS. 200 μM along the lowest energy ruffling and saddling normal modes
cobalt insertion into 5 μM protoporphyrin IX, both in the pres- were computed. Geometric coordinates for distortions along
ence and absence of 5 μM CfbA, was monitored every <100 cm− vibrational modes for the BLYP/def2-TZVP model of
1
1
5 minutes in 50 mM Tris pH 8.0, 150 mM NaCl, and 10% octamethylporphyrin described above were extracted using
4
0,41
methanol (v/v) for 18 hours by UV/Vis spectroscopy. Similarly, gOpenMol.
These coordinates were analysed by the
2
00 μM nickel insertion into 5 μM protoporphyrin IX, both in normal-coordinate structural decomposition program to ident-
4
2,43
the presence and absence of 5 μM CfbA, was monitored every ify and categorize out-of-plane deformation modes.
hour for 72 hours. In addition, 300–1500 m/z ESI-MS data for
cobalt insertion reactions were acquired after 72 hours using a the b2u saddling distortion was identified as a 22.2 cm
The
1
b
1u ruffling distortion was found to be a 42.6 cm− mode and
−
1
C18 guard column using the same instrument, solvent gradi- vibration. The ground- and excited-state potential energy sur-
ent, and analysis software described above. In order to identify faces along these two modes were predicted using the MTR
the binding and oxidation state of the reaction product, Co(III) subprogram of ORCA. For each mode, 50 dimensionless
protoporphyrin IX was purchased from Frontier Scientific and normal-coordinate displacements of 0.05 a.u. in both the posi-
dissolved in 50 mM Tris pH 8.0, 150 mM NaCl, and 10% tive and negative directions from the equilibrium geometry
methanol (v/v). A Co(II) protoporphyrin IX solution was pre- were computed.
pared by reducing a deoxygenated Co(III) protoporphyrin IX
solution with excess sodium dithionite under anaerobic con-
ditions. A 10 μM Co(III) protoporphyrin IX solution was pre-
pared based upon a published extinction coefficient, half of
31
Results
the solution was combined with 10 μM CfbA, and UV/Vis Recombinant CfbA is properly folded
spectra were acquired after 18 hours. Finally, an anaerobic
ESI-MS and CD data confirmed that the recombinant CfbA
characterized in this work is identical to the native enzyme.
The molecular weight of purified, recombinant CfbA is 16 349
Da according to ESI-MS (Fig. 2). Considering experimental
error, this is consistent with the expected isotopically averaged
mass of 16 347 Da for monomeric CfbA. Next, the secondary
structure of CfbA was estimated using CD spectroscopy. The
secondary structure of CfbA in solution is estimated to be 29 ±
1
0 μM Co(II) protoporphyrin IX solution was prepared, half of
that solution was combined with deoxygenated 10 μM CfbA
under anaerobic conditions, and UV/Vis spectra of these two
samples were acquired using screw top cuvettes (Starna).
Computational characterization
DFT models of octamethylporphyrin and doubly deprotonated
octamethylporphyrin were prepared to aid interpretation of
the experimental data. The initial structural model was gener-
ated in ChemDraw (PerkinElmer), and hydrogen atoms were
added in ArgusLab (Planaria Software). All electronic structure
calculations were performed in the ORCA 4.0.0.2 software
package on the Bluemoon cluster of the Vermont Advanced
1
% α-helix and 19 ± 1% β-sheet based upon analysis of the CD
spectrum (Fig. S5†). This secondary structure is similar to that
observed in the X-ray crystal structure of monomeric A. fulgidus
CfbA (PDB ID 1TJN, 28% α-helix and 18% β-sheet). The CD
spectrum is less consistent with the dimeric X-ray crystal struc-
ture (PDB ID 2DJ5, 32% α-helix and 21% β-sheet). Thus, CfbA
is likely a monomeric protein with a solution structure similar
to its monomeric solid-state structure.
20
3
2
Computing Core. Geometry optimizations were completed
3
3,34
35
36
37
using the BLYP,
B3LYP, PBE, and PBE0 density func-
3
8
tionals in combination with a def2-TZVP basis set, very tight
SCF convergence criteria, and an implicit solvation model. The
conductor-like polarizable continuum model (CPCM) was used
to model the influence of methanol solvation on the geometric
and electronic structure of octamethylporphyrin and its
dianion. For each optimized DFT model, vibrational fre-
quency calculations were performed to ascertain whether the
structure represented a true minimum on the potential energy
surface.
3
9
TDDFT was employed to predict the UV/Vis Abs spectra of
octamethylporphryin and its dianion. These calculations uti-
lized the same density functionals, basis set, and solvation
model described above for geometry optimizations. The
TDDFT calculations predicted the first 30 electronic states
using an expansion space of up to six vectors per root. Finally,
the UV/Vis Abs spectra were simulated based upon the TDDFT
data by convoluting Gaussian-shaped bands with full width at
Fig. 2 ESI-MS spectra of CfbA (black line), nickel-bound CfbA (NiCfbA,
dashed blue line), and cobalt-bound CfbA (CoCfbA, dashed red line).
The observed masses of CfbA (16 349 Da), NiCfbA (16 404 Da), and
CoCfbA (16 405 Da) are consistent with the formation of a 1 : 1 complex
between CfbA and a metal substrate.
−
1
half-maximum bandwidths of 1500 cm
.
Dalton Trans.
This journal is © The Royal Society of Chemistry 2019

View Article Online
Dalton Transactions
Paper
CfbA binds cobalt or nickel at an octahedral site
511 nm. These data indicated that aqueous Co(II) is five-coordi-
nate in the 50 mM KP pH 8.0, 50 mM NaCl, and 60% glycerol
i
ESI-MS data demonstrate metal binding to CfbA. The mole-
cular weights of CfbA following incubation with 40-fold molar
excesses of CoCl2 or NiSO4 are 16 405 Da and 16 404 Da,
respectively (Fig. 2). The most abundant isotopes of cobalt and
nickel have molecular weights of 59 and 58 Da, respectively, so
these mass changes are consistent with binding a single metal
ion considering experimental error. Alternatively, the 1 Da dis-
crepancy between the expected and observed masses could be
due to deprotonation of a single amino acid, such as histidine,
upon metal binding. Based upon ESI-MS intensities, CfbA is
almost fully cobalt-bound and about 10% nickel-bound under
the conditions employed here. Thus, further spectroscopic
characterization of metal binding by CfbA focused on cobalt-
bound CfbA.
4
6
(
v/v) buffer used for MCD studies (Fig. S6†). In order to deter-
mine the spin-state of CfbA-bound Co(II), the intensity of the
95 nm MCD band was plotted as a function of sample temp-
4
erature and applied magnetic field (Fig. 3B). The “nesting” of
the isotemperature curves is unambiguous evidence for a high-
4
7
spin (S = 3/2) Co(II) ion. Thus, CfbA binds high-spin Co(II)
with an octahedral coordination site comprised of nitrogen-
and/or oxygen-based ligands.
CfbA binds a diverse set of tetrapyrroles
UV/Vis Abs data indicate that all five isomers of copropor-
phyrin, uroporphyrin, and protoporphyrin used in this study
bind to CfbA. Upon binding to CfbA, the Soret bands of copro-
porphyrin I, uroporphyrin I, and uroporphyrin III decrease in
intensity and broaden (Fig. 4A and S7†). These changes are
most consistent with increased conformational heterogeneity
for coproporphyrin I, uroporphyrin I, and uroporphyrin III
upon binding to CfbA. On the other hand, the coproporphyrin
4,45 III Soret band intensity increases upon binding to CfbA. This
MCD spectroscopy revealed that CfbA-bound cobalt is an
octahedral, high-spin Co(II) species. The MCD spectrum of
CfbA-bound cobalt is dominated by a temperature-dependent,
negatively-signed band at 495 nm (Fig. 3A). This is consistent
with an octahedral Co(II) complex lacking cysteine ligands,
4
where one or more of the ligands may be solvent-derived.
suggests that coproporphyrin III may adopt a more well-
defined structure upon binding to the enzyme (Table S2†).
Similarly, the Soret band intensity of PPIX significantly
In contrast, the MCD spectrum of CoCl2 in the absence of
CfbA has three negatively-signed bands at 623, 590, and
Fig. 3 5 K, 7 T MCD spectra of CoCl
in 50 mM KPi pH 8.0, 50 mM NaCl, 60% (v/v) glycerol (A). 2, 5, 10, and
0 K saturation magnetization curves for the 495 nm band of CoCfbA
constructed from MCD data acquired at 0, 1, 3, 5, and 7 T (B). The CoCl
spectrum is consistent with a five-coordinate species, and the CoCfbA
MCD data indicate the presence of a six-coordinate, S = 3/2 Co(II) ion.
2
(blue, dashed) and CoCfbA (black)
Fig. 4 UV/Vis Abs spectra of 6 μM UPI in 50 mM Tris pH 8.0 and
150 mM NaCl, both in the absence (solid black line) and presence
(dashed blue line) of equimolar CfbA (A). UV/Vis Abs spectra of 6 μM
PPIX in 50 mM Tris pH 8.0, 150 mM NaCl, and 10% (v/v) methanol
without (solid black line) and with (dashed red line) equimolar CfbA (B).
CfbA binding perturbs the electronic structures of both porphyrins.
2
2
This journal is © The Royal Society of Chemistry 2019
Dalton Trans.

View Article Online
Paper
Dalton Transactions
increases upon binding to CfbA (Fig. 4B). In this case, the tet-
rapyrrole is only sparingly soluble in aqueous solution and its
ill-advised to interpret the spectral changes as evidence for
anything more than protein binding. Nevertheless, the UV/Vis
Abs data do indicate that all five tetrapyrroles bind to CfbA
and this event increases the conformational heterogeneity of
coproporphyrin I, uroporphyrin I, and uroporphyrin III.
Binding to CfbA significantly shifts the energy of the Soret
band, and introduces smaller perturbations to the Q band
energies, for all five tetrapyrroles. The Soret band of copropor-
−
1
phyin I exhibits the largest red-shift of 1100 cm
2
from
6 900 cm− to 25 800 cm upon binding to CfbA (Table 1).
1
−1
Somewhat surprisingly, the other isomer of coproporphyrin
studied here, copropophyrin III, undergoes a 400 cm− blue-
shift upon binding to CfbA. The Soret bands of both uropor-
1
phyrin isomers and PPIX all moderately red-shift by
00–400 cm− upon binding to CfbA. On the other hand, the Q
1
3
−1
band components in the 15 000 to 21 000 cm region do not
consistently red- or blue-shift upon tetrapyrrole binding to
CfbA. In the cases of coproporphyrin I and uroporphyrin I,
CfbA binding does not induce any Q-band shifts that are
larger than experimental error. For the other three tetrapyr-
roles, some Q-band components red- or blue-shift, but in none
of these cases do all four Q-band components consistently red-
or blue-shift (Table S3†). Thus, the Soret band energies appear
to be a useful spectroscopic probe for CfbA binding, but the Q
band energies do not.
CfbA promotes metal insertion into several tetrapyrroles
CfbA increases the rate of cobalt insertion into coproporphyrin
I, uroporphyrin I, and uroporphyrin III. Most notably, 5 μM
uroporphyrin I was fully converted to cobalt-uroporphyrin I in
under one minute by 5 μM CfbA in 50 mM Tris pH 8.0,
1
(
50 mM NaCl, and 200 μM CoCl2 at room temperature
Fig. 5A). This represents a significant rate increase compared
to the enzyme-free reaction, which converted 5 μM uropor-
phyrin I to cobalt-uroporphyrin I with a pseudo-first order rate Fig. 5 UV/Vis Abs spectra of CfbA, UPI, and excess CoCl
2
in 50 mM Tris
pH 8.0, 150 mM NaCl immediately following mixing (blue line), after
–29 min (dashed grey lines), and after 30 min (black line, A). UV/Vis Abs
spectra of CfbA, PPIX, and excess CoCl in 50 mM Tris pH 8.0, 150 mM
NaCl, 10% methanol (v/v) immediately following mixing (blue line), after
−1
constant of 0.095 ± 0.003 min (Table 2). CfbA also promoted
the insertion of cobalt into coproporphyrin I and uroporphyrin
III. In these cases, the enzyme-promoted rates were 2.8- and
1
2
4
.2-fold greater than the CfbA-free reactions, respectively. In 1–6 h (dashed grey lines), and after 72 h (black line, B). UV/Vis Abs
in 50 mM Tris pH 8.0, 150 mM
contrast, the rate of cobalt insertion into coproporphyrin III spectra of CfbA, UPIII, and excess NiSO
4
NaCl immediately following mixing (blue line), after 1–119 min (dashed
grey lines), and after 120 min (black line, C). CfbA promotes conversion
of UPI and PPIX to Co–UPI and Co–PPIX, respectively. CfbA also pro-
motes conversion of UPIII to Ni–UPIII.
was unchanged by the addition of CfbA (Fig. S8†).
Nevertheless, these data do clearly indicate that CfbA promotes
cobalt insertion into three of the four isomers of copropor-
phyrin and uroporphyrin investigated here.
Table 1 Soret band shifts for porphyrin substrates upon binding to
CfbA
Table 2 Pseudo-first order rates for Co insertion into porphyrin sub-
strates (min−
1
)
Porphyrin
Without CfbA
With CfbA
% Increase
CPI
CPIII
UPI
UPIII
PPIX
0.0273 ± 0.0005
0.078 ± 0.009
0.095 ± 0.003
0.110 ± 0.002
0
0.077 ± 0.002
0.0796 ± 0.0003
≫1
0.46 ± 0.02
>0
180%
2%
≫900%
320%
∞
Dalton Trans.
This journal is © The Royal Society of Chemistry 2019

View Article Online
Dalton Transactions
Paper
CfbA also promotes the insertion of cobalt into PPIX. Over electronic structure of the porphyrin π system, it is necessary
the course of 72 hours in 50 mM Tris pH 8.0, 150 mM NaCl, to validate potential DFT models using experimental UV/Vis
and 10% methanol (v/v), CfbA converts CoCl
product with an intense UV/Vis Abs maximum of 424 nm optimized models of octamethylporphyrin using the BLYP,
Fig. 5B). Under these same conditions, but in the absence of PBE, B3LYP, and PBE0 density functionals (Fig. S12†). The
CfbA, there is no observable product formation and instead BLYP and PBE TDDFT-predicted Soret band energies are more
2
and PPIX to a Abs data prior to analysis. Thus, we developed fully geometry
(
4
8
PPIX partially precipitates out of solution.
employed to assess whether the product with an intense lations with the B3LYP or PBE0 functionals, which indicates
24 nm UV/Vis Abs maximum was CoPPIX (Fig. S9†). Indeed, that the generalized gradient approximation functionals
ESI-MS was similar to experiment than those predicted by TDDFT calcu-
4
the reaction generates a product with a mass of 619.5 Da, produce more accurate models of the porphyrin valence
which is consistent with formation of CoPPIX. Finally, samples orbital energies and the configuration interaction mixing
4
9
of Co(III)PPIX and Co(II)PPIX were prepared non-enzymatically between the first two excited states. The BLYP TDDFT calcu-
and characterized using UV/Vis Abs spectroscopy to determine lation generates the most accurate Soret to Q band intensity
the oxidation state of the enzyme product (Fig. S10†). Based ratio, which means that this is the most accurate model for
upon the results of these experiments, CfbA converts CoCl
2
valence orbital energies and configuration interaction mixing.
3
1
and PPIX to enzyme-bound Co(III)PPIX. Thus, formation of However, the BLYP TDDFT-predicted Q-band region only
CfbA-bound Co(III)PPIX from Co(II) and PPIX is likely a two- exhibited three bands compared to the four observed experi-
step reaction where the initial Co(II)PPIX product is rapidly oxi- mentally. This is not necessarily surprising because the
5
0
dized by molecular oxygen following metal insertion.
TDDFT model has neglected vibronic coupling, and a wide
Finally, CfbA also promotes nickel insertion into uropor- range of theoretical models have failed to accurately reproduce
5
1
phyrins I and III. Enzyme activity assays were carried out and this aspect of the porphyrin UV/Vis Abs spectrum. Thus, the
analysed as described above for cobalt, only with nickel as the BLYP model will be used to elucidate CfbA-induced pertur-
metal substrate. In the presence of 5 μM CfbA, 5 μM uropor- bations of the tetrapyrrole substrate, and our analysis will
phyrin III was converted to nickel-uroporphyrin III in 50 mM emphasize changes observed for the Soret band.
Tris pH 8.0, 150 mM NaCl, and 200 μM NiSO
4
over the course
The BLYP model developed in this work reveals that the
of two hours at room temperature (Fig. 5C). CfbA-catalysed CfbA-induced changes to the tetrapyrrole UV/Vis Abs spectra
Ni(II) insertion into uroporphyrin I was six-fold slower under do not arise from substrate deprotonation. This is an impor-
these same conditions (Table 3). Unlike the Co(II) insertion tant discovery because small molecule studies have shown that
assays, no Ni(II) insertion into either uroporphyrin was observed porphyrin deprotonation does induce a red-shift of the Soret
in the absence of CfbA (Fig. S11†). As for the coproporphyrin band as seen here for four of the five tetrapyrrole substrates
5
2
and protoporphyrin substrates, nickel insertion was not upon binding to CfbA. Furthermore, small molecule studies
observed in either the presence or absence of CfbA. have shown that porphyrin deprotonation promotes metal che-
5
3
Nevertheless, the data for coproporphyrin III merits further lation by generating four pyrrole nitrogen lone pairs.
comment. Over the course of 30 hours, the Soret band for Deprotonation of the BLYP model for octamethylporphyrin fol-
coproporphyrin III shifted from 389 to 403 nm in the presence lowed by geometry optimization and TDDFT prediction of the
of CfbA. However, the conversion of the Q band from four com- UV/Vis Abs spectrum does predict an 1800 cm− red-shift of
ponents to two components as is typically associated with tetr- the Soret band (Fig. S13†). However, the TDDFT calculations
1
49,50
pyrrole metalation did not occur.
Thus, we conclude that also predict a decrease in the number of Q bands to two,
coproporphyrin III binds to CfbA, but Ni(II) is not inserted. To which has been observed previously for porphyrin structural
5
0
summarize the nickel activity assay data, CfbA catalyses nickel changes that increase molecular symmetry. The TDDFT-pre-
insertion into uroporphryins and Ni(II) insertion into uropor- dicted red-shift of the Soret band upon porphyrin deprotona-
phyrin III is six times faster than insertion into uroporphyrin I.
tion is larger than any of the band-shifts observed in this
study. Also, no significant changes to the shape of the Q-band
region were observed experimentally for any of the five tetra-
CfbA ruffles uroporphyrin and protoporphyrin
A DFT model was developed in order to gain further insight pyrroles. Thus, we conclude that the CfbA-induced changes to
into how CfbA promotes metal chelation. Due to the complex the tetrapyrrole UV/Vis Abs spectra are not due to substrate
deprotonation and must arise from a different structural
change.
Table 3 Pseudo-first order rates for Ni insertion into porphyrin sub-
The DFT model developed here also reveals that the UV/Vis
Abs changes observed upon binding substrate to CfbA cannot
be attributed to a saddling deformation of the tetrapyrrole.
The saddling deformation has been previously correlated with
−1
strates (min
)
Porphyrin
Without CfbA
With CfbA
% Increase
5
4–57
CPI
CPIII
UPI
UPIII
PPIX
0
0
0
0
0
0
0
0%
0%
∞
ferrochelatase activity,
and this distortion could explain
the disorder observed in the X-ray crystal structure of CfbA-
0.0029 ± 0.0002
0.0171 ± 0.0002
0
2
2
bound metal sirohydrochlorin (PDB ID 2XWQ). Thus, the
computational model of octamethylporphyrin was distorted
∞
0%
This journal is © The Royal Society of Chemistry 2019
Dalton Trans.

View Article Online
Paper
Dalton Transactions
Fig. 6 BLYP TDDFT-predicted influence of a b2u out-of-plane saddling
deformation on the Soret band transition energies. Two TDDFT tran-
sitions (black and blue diamonds) contribute to the observed UV/Vis Abs
band. The model predicts a 10–25 cm− red-shift of the Soret band per
Fig. 7 BLYP TDDFT-predicted influence of a b1u out-of-plane ruffling
deformation on the Soret band transition energies. Two TDDFT tran-
sitions (black and blue diamonds) contribute to the observed UV/Vis Abs
1
−1
band. The model predicts a 20–100 cm red-shift of the Soret band per
0
.1 Å of saddling from the equilibrium value of 0.15 Å. Porphyrin sad-
0.1 Å of ruffling from the equilibrium value of 0.12 Å. An increase of por-
phyrin ruffling to 1.4, 1.6, and 1.6 Å, respectively, can explain the UPI,
UPIII, and PPIX Soret band red-shifts observed upon binding to CfbA.
dling cannot explain the porphyrin Soret bandshifts observed upon
binding to CfbA.
along its lowest energy saddling normal mode and the UV/Vis 300–400 cm− red-shifts of the uroporphyrin and PPIX Soret
Abs spectrum was predicted by TDDFT at each point (Fig. 6). bands are fully consistent with a ruffling deformation. Based
These calculations revealed that saddling induces a minor red- upon the spectro-structural model described in this work,
1
−
1
shift of the Soret band by 10–25 cm per 0.1 Å for small dis- CfbA increases uroporphyrin I, uroporphyrin III, and PPIX
tortions from the equilibrium value of 0.15 Å. This minor red- ruffling to 1.4, 1.6, and 1.6 Å, respectively. Thus, the CfbA-
shift is consistent with previous computational studies which induced Soret band red-shifts for uroporphyrin and PPIX are
concluded that saddling does not significantly red-shift the tet- too small to be attributed to pyrrole deprotonation, and too
5
8,59
rapyrrole Soret band.
Due to the minor dependence of the large to be attributed to a saddling deformation, but
−
1
Soret band energy on saddling, the 300–1100 cm red-shifts “just right” for enzyme-induced ruffling of the tetrapyrrole
for coproporphyrin I, uroporphyrin, and PPIX could only be substrate.
explained by saddling if CfbA-binding increases tetrapyrrole
saddling to greater than 1.9 Å. In addition, the 400 cm− blue-
1
shift of the coproporphyrin III Soret band could only be
explained by saddling if CfbA-binding decreases the magni-
tude of this deformation to less than 0 Å! Since none of these
Discussion
Class II chelatases bind labile transition metals
large structural changes are consistent with the X-ray crystal The MCD data presented in this work reveals that a labile
structure of CfbA-bound metal sirohydrochlorin (PDB ID metal binding site appears to be common to class II chela-
2
XWQ), we conclude that saddling is not the origin of the UV/ tases. Specifically, the MCD data clearly demonstrates that
Vis Abs changes observed here.
CfbA-bound cobalt is an octahedral, high-spin Co(II) species
Instead of deprotonation or saddling, the computational (Fig. 3). Based upon fundamental coordination chemistry prin-
model developed here identifies ruffling as the most likely ciples, high-spin Co(II) is a labile metal that will undergo rapid
origin of the CfbA-induced spectral changes. A series of ligand substitution reactions to form the most thermo-
TDDFT-predicted UV/Vis Abs spectra as a function of distor- dynamically-favourable complex. Importantly the ligand field
tion along the lowest energy ruffling mode were generated electronic structure extracted from the Co-bound CfbA MCD
(Fig. 7). These data predict that the tetrapyrrole Soret band will data strongly suggests that Ni-bound CfbA will have an octa-
moderately red-shift by 20–100 per cm per 0.1 Å with increased hedral, high-spin Ni(II) ion. Octahedral, high-spin Ni(II) is also
ruffling from the equilibrium value of 0.12 Å. Even with this labile, whereas a square planar Ni(II) ion would be low-spin
moderate dependence upon ruffling, the large Soret band and kinetically inert. Interestingly, previous studies have
shifts observed for CPI and CPIII upon binding to CfbA shown that the metal substrate of ferrochelatase is high-spin
5
4,60
cannot be readily explained. Most likely, CfbA distorts copro- Fe(II).
porphyrin in a different fashion. However, the CfbA-induced control over the coordination geometry and oxidation state of
Thus, ferrochelatase also appears to exert important
Dalton Trans.
This journal is © The Royal Society of Chemistry 2019

View Article Online
Dalton Transactions
Paper
its metal substrate, iron, since high-spin Fe(II) is labile, but
low-spin Fe(III) and Fe(II) are inert. As discussed in the final
section, the observation that class II chelatases bind labile
metal substrates has important implications for their enzy-
matic mechanism.
CfbA ruffles its tetrapyrrole substrate
Analysis of the UV/Vis Abs data presented here within the
framework of DFT calculations strongly suggests that CfbA
induces a ruffling out-of-plane distortion of its tetrapyrrole
substrate in contrast to the saddling deformation reported for
−
1
ferrochelatase. CfbA binding triggers a 300–400 cm red-shift
of the uroporphyrin and PPIX Soret bands (Fig. 4 and Table 1).
DFT calculations have demonstrated that this is consistent
with increased tetrapyrrole ruffling, but not saddling or depro-
tonation (Fig. 6 and 7). The choice of ruffling as an out-of-
plane distortion by CfbA may have important implications for
Fig. 8 Side-chain substitutions influence the degree of porphyrin out-
of-plane deformation upon binding to CfbA and the rate enhancement
its native function as a nickel chelatase. Ruffled nickel tetra- for Co insertion. The side-chains highlighted in blue are identical to
those of the native substrate, sirohydrochlorin (SHC).
pyrroles are known to have an additional bonding interaction
between the Ni 3dxy and porphyrin a2u orbitals, which means
5
9
that ruffling of a tetrapyrrole substrate may stabilize a nickel
tetrapyrrole product and thermodynamically drive the reaction spectroscopic data presented here, and analysis of data avail-
5
4,60
to completion. On the other hand, ferrochelatase has been able in the literature,
shown to saddle its tetrapyrrole substrate.
both CfbA and ferrochelatase bind a
5
4,56
The saddling labile metal. This means that the metal substrates of these
out-of-plane deformation drives the pyrrole nitrogen lone pairs enzymes will rapidly exchange ligands within the enzyme
out of the macrocyclic plane, which is proposed to promote active sites and there is no kinetic barrier associated with
6
1
iron chelation. Ruffling does not drive the pyrrole nitrogen metal desolvation. A major difference between CfbA and ferro-
4
2
lone pairs out of the macrocyclic plane, and we will discuss chelatase is the manner in which these enzymes bind a tetra-
the potential mechanistic implications of this deformation in pyrrole substrate. Whereas ferrochelatase binds a saddled
5
4,56
the final section.
PPIX substrate,
the spectroscopic and computational data
The ruffling deformation induced by CfbA depends upon presented here strongly suggests that CfbA binds a ruffled SHC
the identity and isomer of the tetrapyrrole substrate. The UV/ substrate (Fig. 7). PPIX saddling is a key component of the pro-
Vis Abs spectral changes observed upon binding copropor- posed ferrochelatase mechanism since this out-of-plane distor-
phryins to CfbA are inconsistent with deprotonation, saddling, tion drives the pyrrole nitrogen lone pairs out of the macro-
or ruffling of these tetrapyrroles. The spectral data suggest that cyclic plane and minimizes the kinetic barrier to iron–nitrogen
6
1
coproporphyrins are strongly perturbed upon binding to CfbA, bond formation. The ruffling deformation induced by CfbA
4
2
but may not be properly positioned within the enzyme active does not push the pyrrole nitrogen lone pairs out-of-plane,
site. On the other hand, uroporphyrins and PPIX exhibit mod- so this enzyme must follow a unique mechanism. For CfbA,
erately red-shifted Soret bands upon binding to CfbA, which is pushing the pyrrole nitrogen lone pairs out-of-plane may
consistent with ruffling of these substrates. A closer examin- simply be unnecessary due to the smaller ionic radius of Ni(II).
ation of the data for the uroporphyrin isomers reveals that the Instead, the ruffling deformation may offer a thermodynamic
magnitude of the ruffling deformation depends upon the exact driving force for tetrapyrrole metalation by destabilizing the
pattern of side-chains around the tetrapyrrole core (Fig. 8). organic reactant and stabilizing the inorganic product
5
9
Uroporphyrin III, which has the same distribution of car- (Fig. 9).
boxylic acid side-chains as the native SHC substrate, is more
It is worth noting that the data presented in this manu-
ruffled than UPI. This strongly suggests that CfbA triggers a script suggests that CfbA preferentially inserts nickel into its
ruffling deformation of its native SHC substrate. With these native tetrapyrrole substrate, SHC. Of the five tetrapyrrole sub-
details regarding nickel and SHC binding by CfbA now avail- strates investigated in this work, SHC is most similar to UPIII
able, a mechanistic proposal emerges that has similarities to (Fig. 8). SHC and UPIII have the same distribution of car-
and differences from that of ferrochelatase.
boxylic acid side-chains, but pyrrole rings A and B are further
reduced in SHC. UPIII is tied with PPIX as the most ruffled
CfbA substrate investigated here (Table 1). Importantly, the
UPIII analogue of SHC is more ruffled than UPI. Finally, CfbA-
The novel mechanism of CfbA is an ideal starting point for
development of synthetic chelatases
The spectroscopic and computational data presented in this catalyzed nickel insertion is six-times faster than nickel inser-
manuscript reveals that the mechanism of metal insertion by tion into UPI (Table 3). This is a striking reversal of tetrapyr-
CfbA is distinct from that of ferrochelatase. Based upon the role substrate preference compared to cobalt insertion
This journal is © The Royal Society of Chemistry 2019
Dalton Trans.

View Article Online
Paper
Dalton Transactions
lvation is facile, and this is not the rate-limiting step of the
reaction. The tetrapyrrole binding site of CfbA induces a
ruffling deformation of this substrate. This ruffling distorition
is distinct from the critical saddling deformation observed in
5
4,56
ferrochelatase,
which means that the CfbA mechanism
6
1
must be different from that proposed for ferrochelatase.
These data reveal a fascinating correlation between native
metal substrate identity and tetrapyrrole distortion.
Conflicts of interest
There are no conflicts to declare.
Fig. 9 Proposed mechanism for Ni(II) insertion into SHC by CfbA based
upon the data presented in this article. Upon binding to CfbA, high-spin
octahedral Ni(II) is labile and a rapid equilibrium will be established
between six- and five-coordinate Ni(II). This strongly suggests that CfbA
follows a dissociative substitution mechanism in order to insert Ni(II) into
a ruffled SHC substrate.
Acknowledgements
M. D. L. thanks the University of Vermont and the National
Science Foundation (DMR-1506248) for financial support.
A. A. S. thanks St. Lawrence University for financial support.
The authors thank Andrea Lee (University of Vermont) for
assistance with cloning, Alexis Haley (St. Lawrence University),
and Bruce O’Rourke (University of Vermont) for assistance
with ESI-MS data acquisition.
(Table 2). More definitive conclusions regarding the preference
of CfbA for SHC will require future studies employing SHC
itself, but these data identify an intriguing correlation between
the identities of the tetrapryrrole side-chains and nickel inser-
tion activity.
The mechanism of CfbA merits further investigation since
this is the first enzyme shown to promote metal insertion into
both porphyrins and bacteriochlorins. Previous studies have
References
1 T. L. Poulos, Chem. Rev., 2014, 114, 3919–3962.
2 P. R. Ortiz de Montellano, Cytochrome P450: Structure,
Mechanism, and Biochemistry, Springer International
Publishing, Switzerland, 2015.
3 E. Raven and B. Dunford, Heme Peroxidases, Royal Society
of Chemistry, Cambridge, UK, 2016.
4 S. Yoshikawa and A. Shimada, Chem. Rev., 2015, 115, 1936–
1989.
5 R. G. Matthews, M. Koutmos and S. Datta, Curr. Opin.
Struct. Biol., 2008, 18, 658–666.
1
1,19
11,19,20
demonstrated that CfbA can insert Fe(II),
Co(II),
and
1
1,17,21
Ni(II)
into SHC based upon both in vitro and in vivo
assays. Here, we have reported CfbA-promoted Co(II) insertion
into coproporphyrin, uroporphyrin, and PPIX based upon
careful spectroscopic characterization of the reaction (Fig. 5).
We have also observed CfbA-catalysed Ni(II) insertion into uro-
porphyrin. Co–PPIX is a particularly valuable product because
this species has been shown to catalyse two reactions of par-
6
2–69
ticular relevance for the development of fuel cells.
Also,
coproporphryin, uroporphyrin, PPIX, and SHC are a reasonably
broad representation of tetrapyrrole structures and electronic
structures. The number of propionate side-chains in this set of
substrates ranges from two (PPIX) to eight (SHC), and the
number of electrons in the tetrapyrrole π system ranges from
6 R. Banerjee, Chem. Rev., 2003, 103, 2083–2094.
7 R. K. Thauer, Biochemistry, 2019, DOI: 10.1021/
acs.biochem.9b00164.
8 W. Zhang, W. Lai and R. Cao, Chem. Rev., 2017, 117, 3717–
3797.
1
8 in SHC to 22 in the porphyrins. This broad substrate scope
9 T. J. Lawton and A. C. Rosenzweig, J. Am. Chem. Soc., 2016,
138, 9327–9340.
means that CfbA is a good starting point for design of syn-
thetic chelatases via iterative improvement of existing func- 10 D. Osman, M. A. Martini, A. W. Foster, J. Chen,
tion. The mechanistic insights revealed by this study represent
a critical first step in this process.
A. J. P. Scott, R. J. Morton, J. W. Steed, E. Lurie-Luke,
T. G. Huggins, A. D. Lawrence, E. Deery, M. J. Warren,
P. T. Chivers and N. J. Robinson, Nat. Chem. Biol., 2019, 15,
241–249.
1
1 A. A. Brindley, E. Raux, H. K. Leech, H. L. Schubert and
M. J. Warren, J. Biol. Chem., 2003, 278, 22388–22395.
Conclusions
In summary, this work has shown that CfbA has the broadest 12 M. Hansson, J. Lundqvist, N. Sirijovski and S. Al-
substrate scope of all characterized chelatases and has pro-
vided critical insight into the enzymatic mechanism. The
metal binding site of CfbA accommodates an octahedral, high-
spin Co(II) ion. Thus, the metal substrate is labile, metal deso-
Karadaghi, Chlorophyll, photosynthesis and bio-inspired
energy, in Handbook of Porphyrin Science: with applications
to chemistry, physics, materials science, engineering, biology,
and medicine, ed. G. C. Ferreira, K. M. Kadish, K. M. Smith
Dalton Trans.
This journal is © The Royal Society of Chemistry 2019

View Article Online
Dalton Transactions
Paper
and R. Guilard, World Scientific, Hackensack, NJ, USA, 33 A. D. Becke, Phys. Rev. A, 1988, 38, 3098–3100.
2
013, vol. 28.
34 C. Lee, W. Yang and R. G. Parr, Phys. Rev. B: Condens.
Matter Mater. Phys., 1988, 37, 785–789.
1
1
3 E. Deery, S. Schroeder, A. D. Lawrence, S. L. Taylor,
A. Seyedarabi, J. Waterman, K. S. Wilson, D. Brown, 35 A. D. Becke, J. Chem. Phys., 1993, 98, 5648–5652.
M. A. Geeves, M. J. Howard, R. W. Pickersgill and 36 J. P. Perdew, K. Burke and M. Ernzerhof, Phys. Rev. Lett.,
M. J. Warren, Nat. Chem. Biol., 2012, 8, 933–940.
1996, 77, 3865–3868.
4 H. A. Dailey, T. A. Dailey, C.-K. Wu, A. E. Medlock, 37 J. P. Perdew, M. Ernzerhof and K. Burke, J. Chem. Phys.,
K.-F. Wang, J. P. Rose and B.-C. Wang, Cell. Mol. Life Sci.,
000, 57, 1909–1926.
5 E. Raux, C. Thermes, P. Heathcote, A. Rambach and
M. J. Warren, J. Bacteriol., 1997, 179, 3202–3212.
6 H. K. Leech, E. Raux-Deery, P. Heathcote and M. J. Warren,
Biochem. Soc. Trans., 2002, 30, 610–613.
1996, 105, 9982–9985.
38 F. Weigend and R. Ahlrichs, Phys. Chem. Chem. Phys., 2005,
7, 3297–3305.
39 V. Barone and M. Cossi, J. Phys. Chem. A, 1998, 102, 1995–
2001.
40 A. Laaksonen, J. Mol. Graphics, 1992, 10, 33–34.
2
1
1
1
1
7 K. Zheng, P. D. Ngo, V. L. Owens, X.-P. Yang and 41 D. L. Bergman, L. Laaksonen and A. Laaksonen, J. Mol.
S. O. Mansoorabadi, Science, 2016, 354, 339–342. Graphics, 1997, 15, 301–306.
8 G. Layer and M. J. Warren, Biochemistry of tetrapyrroles 42 W. Jentzen, X.-Z. Song and J. A. Shelnutt, J. Phys. Chem. B,
part II), in Handbook of Porphyrin Science: with applications 1997, 101, 1684–1699.
to chemistry, physics, materials science, engineering, biology, 43 A. B. Graves, M. T. Graves and M. D. Liptak, J. Phys. Chem.
(
and medicine, ed. K. M. Kadish, K. M. Smith and R. Guilard,
World Scientific, Hackensack, NJ, USA, 2012, vol. 19.
9 A. Pisarchik, R. Petri and C. Schmidt-Dannert, Protein Eng.,
Des. Sel., 2007, 20, 257–265.
0 J. Yin, L. X. Xu, M. M. Cherney, E. Raux-Deery,
A. A. Bindley, A. Savchenko, J. R. Walker, M. E. Cuff,
B, 2016, 120, 3844–3853.
44 A. J. McCaffery, P. J. Stephens and P. N. Schatz, Inorg.
Chem., 1967, 6, 1614–1625.
45 E. Decaneto, T. Vasilevskaya, Y. Kutin, H. Ogata,
M. Grossman, I. Sagi, M. Havenith, W. Lubitz, W. Thiel and
N. Cox, Phys. Chem. Chem. Phys., 2017, 19, 30316–30331.
1
2
M. J. Warren and M. N. G. James, J. Struct. Funct. Genomics, 46 J. A. Larrabee, C. M. Alessi, E. T. Asiedu, J. O. Cook,
2
006, 7, 37–50.
K. R. Hoerning, L. J. Klingler, G. S. Okin, S. G. Santee and
2
2
1 S. J. Moore, S. T. Sowa, C. Schuchardt, E. Deery,
A. D. Lawrence, J. V. Ramos, S. Billig, C. Birkemeyer, 47 F. Neese and E. I. Solomon, Inorg. Chem., 1999, 38, 1847–1865.
P. T. Chivers, M. J. Howard, S. E. J. Rigby, G. Layer and 48 L. M. Scolaro, M. Castriciano, A. Romeo, S. Patané,
M. J. Warren, Nature, 2017, 543, 78–82.
T. L. Volkert, J. Am. Chem. Soc., 1997, 119, 4182–4196.
E. Cefalí and M. Allegrini, J. Phys. Chem. B, 2002, 106,
2453–2459.
2 C. V. Romão, D. Ladakis, S. A. L. Lobo, M. A. Carrondo,
A. A. Brindley, E. Deery, P. M. Matias, R. W. Pickersgill, 49 M. Gouterman, J. Chem. Phys., 1959, 30, 1139–1161.
L. M. Saraiva and M. J. Warren, Proc. Natl. Acad. 50 T. Hashimoto, Y.-K. Choe, H. Nakano and K. Hirao, J. Phys.
Sci. U. S. A., 2011, 108, 97–102.
Chem. A, 1999, 103, 1894–1904.
2
2
2
3 R. Wu, E. P. Skaar, R. Zhang, G. Joachmiak, P. Gornicki, 51 D. Sundholm, Phys. Chem. Chem. Phys., 2000, 2, 2275–2281.
O. Schneewind and A. Joachmiak, J. Biol. Chem., 2005, 280, 52 J. E. Falk, Porphyrins and Metalloporphyrins, Elsevier,
2
840–2846.
Amsterdam, 1964.
4 R. B. Kapust, J. Tözsér, J. D. Fox, D. E. Anderson, S. Cherry, 53 D. T. Nam, Y. B. Ivanova, S. G. Puhovskaya, M. M. Kruk and
T. D. Copeland and D. S. Waugh, Protein Eng., 2001, 14,
93–1000.
5 S. Nallamsetty, R. B. Kapust, J. Tözsér, S. Cherry,
J. E. Tropea, T. D. Copeland and D. S. Waugh, Protein
Expression Purif., 2004, 37, 108–115.
S. A. Syrbu, RSC Adv., 2015, 5, 26125–26131.
54 Y. Lu, A. Sousa, R. Franco, A. Mangravita, G. C. Ferreira,
I. Moura and J. A. Shelnutt, Biochemistry, 2002, 41, 8253–
8262.
55 Z. Shi, R. Franco, R. Haddad, J. A. Shelnutt and
G. C. Ferreira, Biochemistry, 2006, 45, 2904–2912.
56 A. E. Medlock, L. Swartz, T. A. Dailey, H. A. Dailey and
W. N. Lanzilotta, Proc. Natl. Acad. Sci. U. S. A., 2007, 104,
1789–1793.
9
26 L. Bogorad, J. Biol. Chem., 1958, 233, 501–509.
27 C. Rimington, Biochem. J., 1960, 75, 620–623.
28 R. Giovannetti and V. Bartocci, J. Liq. Chromatogr. Relat.
Technol., 1998, 21, 2607–2617.
2
3
3
9 E. Kazutoshi and I. Shukuro, Bull. Chem. Soc. Jpn., 1995, 57 R. Franco, S. Al-Karadaghi and G. C. Ferreira, J. Porphyrins
8, 3085–3090. Phthalocyanines, 2011, 15, 357–363.
0 C. Jullian, J. E. Brunet, V. Thomas and D. M. Jameson, 58 A. K. Wertsching, A. S. Koch and S. G. DiMagno, J. Am.
Biochim. Biophys. Acta, 1989, 997, 206–210. Chem. Soc., 2001, 123, 3932–3939.
1 I. Barr, S. H. Weitz, T. Atkin, P. Hsu, M. Karayiorgou, 59 H. Ryeng and A. Ghosh, J. Am. Chem. Soc., 2002, 124, 8099–
6
J. A. Gogos, S. Weiss and F. Guo, Chem. Biol., 2015, 22,
93–802.
2 F. Neese, Wiley Interdiscip. Rev.: Comput. Mol. Sci., 2012, 2,
3–78.
8103.
7
60 R. Franco, J. J. G. Moura, I. Moura, S. G. Lloyd,
B. H. Huynh, W. S. Forbes and G. C. Ferreira, J. Biol. Chem.,
1995, 270, 26352–26357.
3
7
This journal is © The Royal Society of Chemistry 2019
Dalton Trans.

View Article Online
Paper
Dalton Transactions
6
1 S. Al-Karadaghi, R. Franco, M. Hansson, J. A. Shelnutt, 66 J. P. Collman, P. Denisevich, Y. Konai, M. Marrocco,
G. Isaya and G. C. Ferreira, Trends Biochem. Sci., 2006, 31,
35–142.
C. Koval and F. C. Anson, J. Am. Chem. Soc., 1980, 102,
6027–6036.
1
6
6
6
6
2 J. G. Kleingardner, B. Kandemir and K. L. Bren, J. Am. 67 R. R. Durand, C. S. Bencosme, J. P. Collman and
Chem. Soc., 2014, 136, 4–7.
F. C. Anson, J. Am. Chem. Soc., 1983, 105, 2710–
3 B. Kandemir, S. Chakraborty, Y. X. Guo and K. L. Bren,
Inorg. Chem., 2016, 55, 467–477.
4 D. J. Sommer, M. D. Vaughn and G. Ghirlanda, Chem.
Commun., 2014, 50, 15852–15855.
2718.
68 Y. L. Mest, C. Inisan, A. Laouenan, M. L’Her, J. Talarmin,
E. M. Khalifa and J.-Y. Sailard, J. Am. Chem. Soc., 1997, 119,
6095–6106.
5 D. J. Sommer, M. D. Vaughn, B. C. Clark, J. Tomlin, A. Roy 69 C. J. Chang, Y. Deng, C. Shi, C. K. Chang, F. C. Anson and
and G. Ghirlanda, Biochim. Biophys. Acta, 2016, 1857, 598–603.
D. G. Nocera, Chem. Commun., 2000, 15, 1355–1356.
Dalton Trans.
This journal is © The Royal Society of Chemistry 2019