Monovalent Cation Activation of the Radical SAM Enzyme Pyruvate Formate-Lyase Activating Enzyme

Article abstract of DOI:10.1021/jacs.7b04883

Pyruvate formate-lyase activating enzyme (PFL-AE) is a radical S-adenosyl-l-methionine (SAM) enzyme that installs a catalytically essential glycyl radical on pyruvate formate-lyase. We show that PFL-AE binds a catalytically essential monovalent cation at its active site, yet another parallel with B₁₂ enzymes, and we characterize this cation site by a combination of structural, biochemical, and spectroscopic approaches. Refinement of the PFL-AE crystal structure reveals Na⁺ as the most likely ion present in the solved structures, and pulsed electron nuclear double resonance (ENDOR) demonstrates that the same cation site is occupied by ²³Na in the solution state of the as-isolated enzyme. A SAM carboxylate-oxygen is an M⁺ ligand, and EPR and circular dichroism spectroscopies reveal that both the site occupancy and the identity of the cation perturb the electronic properties of the SAM-chelated iron-sulfur cluster. ENDOR studies of the PFL-AE/[¹³C-methyl]-SAM complex show that the target sulfonium positioning varies with the cation, while the observation of an isotropic hyperfine coupling to the cation by ENDOR measurements establishes its intimate, SAM-mediated interaction with the cluster. This monovalent cation site controls enzyme activity: (i) PFL-AE in the absence of any simple monovalent cations has little-no activity; and (ii) among monocations, going down Group 1 of the periodic table from Li⁺ to Cs⁺, PFL-AE activity sharply maximizes at K⁺, with NH₄⁺ closely matching the efficacy of K⁺. PFL-AE is thus a type I M⁺-activated enzyme whose M⁺ controls reactivity by interactions with the cosubstrate, SAM, which is bound to the catalytic iron-sulfur cluster.

Full text of DOI:10.1021/jacs.7b04883

Subscriber access provided by UNIV LAVAL
Article
Monovalent Cation Activation of the Radical SAM
Enzyme Pyruvate Formate-Lyase Activating Enzyme
Krista A. Shisler, Rachel U. Hutcheson, Masaki Horitani, Kaitlin S. Duschene, Adam V.
Crain, Amanda S. Byer, Eric M. Shepard, Ashley Rasmussen, Jian Yang, William E.
Broderick, Jessica L. Vey, Catherine L. Drennan, Brian M. Hoffman, and Joan B. Broderick
J. Am. Chem. Soc., Just Accepted Manuscript • DOI: 10.1021/jacs.7b04883 • Publication Date (Web): 02 Aug 2017
Downloaded from http://pubs.acs.org on August 6, 2017
Just Accepted
“Just Accepted” manuscripts have been peer-reviewed and accepted for publication. They are posted
online prior to technical editing, formatting for publication and author proofing. The American Chemical
Society provides “Just Accepted” as a free service to the research community to expedite the
dissemination of scientific material as soon as possible after acceptance. “Just Accepted” manuscripts
appear in full in PDF format accompanied by an HTML abstract. “Just Accepted” manuscripts have been
fully peer reviewed, but should not be considered the official version of record. They are accessible to all
readers and citable by the Digital Object Identifier (DOI®). “Just Accepted” is an optional service offered
to authors. Therefore, the “Just Accepted” Web site may not include all articles that will be published
in the journal. After a manuscript is technically edited and formatted, it will be removed from the “Just
Accepted” Web site and published as an ASAP article. Note that technical editing may introduce minor
changes to the manuscript text and/or graphics which could affect content, and all legal disclaimers
and ethical guidelines that apply to the journal pertain. ACS cannot be held responsible for errors
or consequences arising from the use of information contained in these “Just Accepted” manuscripts.
Journal of the American Chemical Society is published by the American Chemical
Society. 1155 Sixteenth Street N.W., Washington, DC 20036
Published by American Chemical Society. Copyright © American Chemical Society.
However, no copyright claim is made to original U.S. Government works, or works
produced by employees of any Commonwealth realm Crown government in the course
of their duties.

Page 1 of 13
Journal of the American Chemical Society
1
2
3
4
5
6
7
8
9
Monovalent Cation Activation of the Radical SAM Enzyme Py-
ruvate Formate-Lyase Activating Enzyme.
Krista A. Shisler†, Rachel U. Hutcheson†, Masaki Horitani‡¥, Kaitlin S. Duschene†, Adam V. Crain†,
Amanda S. Byer†, Eric M. Shepard†, Ashley Rasmussen†, Jian Yang†, William E. Broderick†, Jessica L.
Vey§#, Catherine L. Drennan§, Brian M. Hoffman‡, and Joan B. Broderick†*.
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
†Department of Chemistry and Biochemistry, Montana State University, Bozeman, MT 59717, ‡Department of Chemistry,
Northwestern University, Evanston, IL 60208, #Department of Chemistry and Biochemistry, California State University
Northridge, Northridge, CA 91330, and the §Departments of Chemistry and Biology and the Howard Hughes Medical Insti-
tute, Massachusetts Institute of Technology, Cambridge, MA 02139.
KEYWORDS. radical S-adenosyl-L-methionine, iron-sulfur cluster, glycyl radical, monovalent cation stimulation.
ABSTRACT: Pyruvate formate-lyase activating enzyme (PFL-AE) is a radical S-adenosyl-L-methionine (SAM) enzyme that in-
stalls a catalytically essential glycyl radical on pyruvate formate-lyase. We show that PFL-AE binds a catalytically significant
monovalent cation at its active site, yet another parallel with B12 enzymes, and we characterize this cation site by a combination of
structural, biochemical, and spectroscopic approaches. Refinement of the PFL-AE crystal structure reveals Na+ as the most likely
ion present in the solved structures, and pulsed electron nuclear double resonance (ENDOR) demonstrates that the same cation site
is occupied by 23Na in the solution state of the as-isolated enzyme. A SAM carboxylate-oxygen is an M+ ligand, and EPR and circu-
lar dichroism spectroscopies reveal that both the site occupancy and the identity of the cation perturb the electronic properties of the
SAM-chelated iron-sulfur cluster. ENDOR studies of the PFL-AE/[13C-methyl]-SAM complex show that the target sulfonium posi-
tioning varies with the cation, while the observation of an isotropic hyperfine coupling to the cation by ENDOR measurements es-
tablishes its intimate, SAM-mediated interaction with the cluster. This monovalent cation site controls enzyme activity: (i) PFL-AE
in the absence of any simple monovalent cations has little-no activity; and (ii) among mono-cations, going down Group 1 of the
periodic table from Li+ to Cs+ PFL-AE activity sharply maximizes at K+, with NH4+ closely matching the efficacy of K+. PFL-AE is
thus a type I M+-activated enzyme whose M+ controls reactivity by interactions with the co-substrate, SAM, which is bound to the
catalytic iron-sulfur cluster.
gen atom from glycine 734 of pyruvate formate-lyase (PFL),
activating it for the catalytic conversion of pyruvate and coen-
INTRODUCTION
Pyruvate formate-lyase activating enzyme (PFL-AE) is a
member of the large and diverse radical S-adenosyl-L-
methionine (SAM) superfamily, members of which use an
iron-sulfur cluster and SAM to initiate difficult radical trans-
formations in all kingdoms of life.1-3 Radical SAM enzymes
share a common CX3CX2C motif or variation thereof, and the
conserved cysteines coordinate three irons of a [4Fe-4S] clus-
ter, while SAM coordinates the fourth iron through its amino
and carboxylate moieties.4-5 Radical SAM catalysis is thought
to be initiated by electron transfer from the reduced [4Fe-4S]+
cluster to the sulfonium of SAM, inducing S-C(5') bond cleav-
age to give methionine and a 5'-deoxyadenosyl radical (5'-
dAdo•) intermediate that abstracts a hydrogen atom from sub-
strate. While the 5'-dAdo• intermediate has never been ob-
served for PFL-AE nor for any other radical SAM enzyme,
recent results reveal the involvement of a novel organometallic
intermediate in which the SAM-derived dAdo moiety is direct-
ly bound via the 5'-C to the unique iron of the [4Fe-4S] clus-
ter.6 This organometallic intermediate is proposed to undergo
homolytic Fe-C bond cleavage to yield the 5'-dAdo•.6 In the
case of PFL-AE, the 5'-dAdo• radical then abstracts a hydro-
zyme A (CoA) to formate and acetyl-CoA under anaerobic
conditions (Figure 1).7-8 Careful studies of the PFL-AE/PFL
system allowed quantitative determination of the product of H-
atom abstraction, the stable glycyl radical on PFL, demonstrat-
ing the 1:1 stoichiometry between electron loss from the [4Fe-
4S]+ cluster and product formation on PFL.9
Two crystal structures of PFL-AE have been reported: one
of PFL-AE/SAM and another of PFL-AE/SAM with a bound
7-mer PFL peptide substrate analog of PFL containing the
target glycine residue.10 The structures reveal a partial (β/α)6
triosephosphate isomerase (TIM) barrel with a wide opening
able to accommodate the large substrate PFL, a domain of
which is proposed to bind in the active site of PFL-AE.10
SAM is bound through the amino and carboxylate moieties to
the unique iron of the cluster, a binding mode first revealed by
electron nuclear double resonance (ENDOR) spectroscopy.5
Two motifs conserved in the radical SAM superfamily interac-
tion with SAM: the GGE motif (PFL-AE residues G77, G78,
and E79) directly binds the methionine portion of SAM, while
the GXIXGXXE motif (G167VIVGXXE175 in PFL-AE) stabi-
lizes the adenine moiety binding site. The PFL peptide sub-
ACS Paragon Plus Environment

Journal of the American Chemical Society
Page 2 of 13
1
2
3
4
5
6
7
8
strate is bound across the lateral opening of the partial TIM
this cation, we have carried out detailed structural refinement
barrel primarily by contacts between the peptide backbone and
PFL-AE side chains. One of these interactions involves the
highly conserved DGXGXR motif located on PFL-AE loop A
(residues 10-20), and upon peptide binding, this loop under-
goes a large conformational change by swinging up into the
active site to establish these interactions. The movement of
this loop is thought to be essential to the activation of PFL,
possibly by orienting the glycine loop of PFL in the active
site.10
and analysis, EPR, ENDOR, and circular dichroism (CD)
spectroscopic studies of cation binding, and studies of the var-
iations in enzymatic function with choice of cation. Instances
where a monovalent metal ion interacts with a protein-bound
metal center14 and even more importantly, is essential for
enzymatic catalytic function,15-16 are rare. However, in yet
another parallel between the radical-SAM and B12 radical en-
zymes, assays of PFL-AE activity reveal that its catalytic func-
tion is controlled by the presence and identity of the bound
monovalent cation, requiring a K+ ion bound in the active site
for optimal activity, as is the case for the B12 enzyme diolde-
hydrase.16-17
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
MATERIALS AND METHODS
All chemicals were obtained from commercial sources and
were of the highest purity commercially available. The en-
zymes citrate synthase and malic dehydrogenase were pur-
chased from Sigma.
Refinement of Metal Ions/Water Molecule in the
Putative Metal Site
The previously reported PFL-AE crystallization conditions
consisted of 100 mM HEPES, 3.5 M sodium formate and 0.2
mM 2,6-dimethyl-4-heptyl-beta-d-maltopyranoside, pH 6.8,
and the resulting peptide-PFL-AE model was refined in
CNS.10 To further investigate a putative metal ion binding site,
ten additional cycles of refinement of the deposited structure
(PDB entry 3CB8) were carried out in parallel in REFMAC518
in the CCP4i program suite19 with the putative site modeled as
a water molecule, sodium ion, magnesium ion, potassium ion
or calcium ion (note that other than water, of these ions, only
sodium was known to be present in the crystallization condi-
tions). Additionally, a round of refinement was carried out in
CNS20 using the same starting model with water, sodium,
magnesium, potassium or calcium modeled in the metal site,
this time, refining the occupancy of metal site with the site’s
B-factor set to 76.00 (the average B-factor of the surrounding
atoms). The B-factors, R-factors, occupancies, distances to
nearby atoms (Table S1-5), and difference density maps (Fig-
ure S1) of the resulting models were examined to judge likeli-
hood of each atom type’s presence in the site in question.
Figure 1. The reaction catalyzed by PFL-AE. The PFL-AE
substrate: PFL, is activated by H-atom abstraction from G734,
concomitant with conversion of SAM to methionine and 5'-
deoxyadenosine.
The PFL substrate of PFL-AE is a homodimer wherein
each monomer consists of a 10-stranded α/β barrel with the
active site located at the center.11-12 Within the active site are
two catalytically essential cysteine residues, C418 and C419,
as well as the site of the glycyl radical, G734. This active site
is buried 8 Å from the surface of the protein; however, G734
must be directly accessed by the active site of PFL-AE in or-
der for the stereospecific H-atom abstraction that activates
PFL. Taken together, the structures of PFL and PFL-AE
thereby suggest that a significant PFL conformational change
is required for the glycine-containing loop to be accessible to
the active site of PFL-AE. Experimental evidence for such a
conformational change has been provided by a study utilizing
fluorescence and electron paramagnetic resonance (EPR) spec-
troscopies and activity assays that together point to an open
conformation of PFL in the presence of PFL-AE.13 In the
proposed open conformation of PFL, the radical domain con-
taining G734 flips out of the center of PFL and becomes ac-
cessible to PFL-AE.13
Expression and Purification of PFL and PFL-AE
and Generation of Variants
PFL-AE wild-type and variants were expressed and puri-
fied as previously described5, 21-24 with modifications as de-
scribed here and in the SI. The pCAL-n-EK plasmid contain-
ing the PFL-AE gene was transformed into E. coli
BL21(DE3)pLysS (Stratagene) cells for overexpression. Mu-
tagenesis was carried out using the QuikChange site-directed
mutagenesis kit from Stratagene and plasmids containing the
variant genes were transformed into BL21(DE3)pLysS cells.
Mutagenesis primers to introduce the D104A change (CAT
TCA TAC CTG TCT GGC CAC CAA CGG TTT TGT TC)
and the D129A change (CGT TCA TCT GTT TGA GAG
CGA GCA TTA CCA GGT C) were synthesized by Integrated
DNA Technologies. Purification of “metal-free” PFL-AE in-
volved the omission of alkali metal cations in all purification
buffers, in order to minimize occupancy of the monovalent
cation site. The Tris buffers were made by using Tris base, and
adjusting the pH with HCl. In addition, no alkali metal salts
Our studies of PFL-AE over many years provided a de-
tailed understanding of the EPR spectroscopic properties of
the active site [4Fe-4S]+ in the absence and presence of SAM.
We were therefore surprised when we began to see significant
batch-batch variations in these spectra, and we ultimately drew
a connection between these spectroscopic differences and
changes in buffer salt composition. Subsequently, close in-
spection of the structure of PFL-AE in complex with the PFL
model 7-mer peptide revealed the presence of a bound mono-
valent cation in the active site.
To identify this cation, to characterize the cation binding
site, to determine the likely identity of the in vivo metal occu-
pying that site, and to discover the functional significance of
ACS Paragon Plus Environment

Page 3 of 13
Journal of the American Chemical Society
1
2
3
4
5
6
7
8
were added to the PFL-AE purification buffers. Excess mono-
valent cations were removed from the PFL-AE substrate,
namely PFL, by dialysis. All further details are provided in the
SI.
of 2 mM and incubated for ~5 minutes before freezing in liq-
uid nitrogen.
EPR spectra were recorded on a Bruker EMX X-band spec-
trometer equipped with a liquid helium cryostat and tempera-
ture controller from Oxford Instruments. Typical experi-
mental parameters were 12 K and 9.37 GHz, with 1.0 mW
microwave power, 100 kHz modulation frequency, and 10 G
modulation amplitude. Each spectrum is the average of four
scans. The EPR spectra in the presence of SAM were simulat-
ed using Easy Spin version 5.1.11.28
PFL-AE Activity Assays
PFL-AE activity was assayed in a Unilab MBraun anaero-
bic chamber containing ≤ 1 ppm O2 using a modified version
of the coupled enzyme assays previously published.21, 25-26 For
assaying the effect of each metal ion on activity, a 100 mM
Tris pH 7.6 buffer containing 100 mM of monovalent cation
salt (NaCl, KCl, NH4Cl, RbCl, CsCl, or LiCl) or 1 mM of
divalent cation salt (CaCl2, MgCl2, or Zn(NO3)2) was used. A
control lacking any salt addition was also conducted with each
assay. To carry out the assay, 450 μL of activation solution
was mixed containing a final concentration of 0.05 µM PFL-
AE purified in the absence of M+, 5 µM PFL, 0.1 mM SAM,
10 mM oxamate (or 10 mM pyruvate for assays conducted
with CaCl2), 8 mM DTT and 25 µM 5-deazariboflavin (added
last in the dark) in a 100 mM Tris, pH 7.6 buffer containing
different cations as described above (all concentrations given
as final concentrations). This mixture was added to a short-
ened, acid-washed NMR tube and PFL-AE was photoreduced
by illumination with a 300 W halogen lamp and the protein
was kept in a 30˚C ± 2˚C water bath for 5 minutes and then
covered with foil to prevent further reduction. To assay the
activity, 5 µL of the resulting activation solution was placed
on a lid of an anaerobic cuvette containing 895 µL of coupling
solution (3 mM NAD+, 55 µM CoA, 0.05 mg/mL BSA, 10
mM pyruvate, 10 mM malate, 2 U/mL citrate synthase, 30
U/mL malic dehydrogenase, and 10 mM DTT in a 100 mM
Tris, pH 8.1 buffer) prior to sealing the cuvette and removing
it from the chamber. The two solutions were mixed by invert-
ing the cuvette just prior to placing in a thermostatted (at
30°C) Cary 60 UV-Vis spectrophotometer. The production of
NADH was monitored at 340 nm for 90 seconds. Rates were
calculated from the slope of absorbance vs. time from 18 sec-
onds to 90 seconds, when the slopes stabilized and were the
most linear. One unit of PFL activity corresponds to the pro-
duction of one µmole of pyruvate per minute, and 35 units of
PFL is equivalent to 1 nmole of PFL active sites.27 The defini-
tion of one unit of PFL-AE activity is the amount that catalyz-
es the production of 1 nmole of active PFL per minute.21 All
assays were conducted in triplicate.
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
Circular Dichroism
CD experiments were conducted under anaerobic condi-
tions using a Jasco-710 spectropolarimeter at ambient temper-
ature (23˚C ± 2˚C). Measurements were collected in a 1 cm
path length cuvette from 300-800 nm with a sensitivity of 100
millidegrees, 0.1 nm data pitch, continuous scan mode with a
speed of 100 nm/min, a response time of 1 s, 1.0 nm band
width, and accumulation of 3 scans. PFL-AE was added to an
anaerobic cuvette in an MBraun anaerobic chamber to a final
concentration of 50 μM containing 1 mM DTT in a 100 mM
Tris, pH 7.6 buffer and either no cation or 100 mM of NaCl,
KCl, NH4Cl, RbCl, CsCl, or LiCl. After initial scans without
SAM were conducted, SAM was added to a final concentra-
tion of 500 μM. For determination of the KD for K+, a final
concentration of 50 μM PFL-AE and 500 μM SAM was used
prior to KCl addition. KCl was titrated to final concentrations
of 0 mM, 2 mM, 3 mM, 6 mM, 12 mM, 21 mM, 35 mM, 49
mM, 101 mM, and 177 mM, 301 mM, and 500 mM.
ENDOR measurements
Q-band CW ENDOR spectra were collected as described.29-
30 Pulsed ENDOR spectra were collected at 2 K on a previous-
ly described spectrometer.30 The Mims pulse sequence (π/2-τ-
π/2-T-π/2-τ-echo, RF applied during interval T) was used for
small hyperfine couplings.31-32
For a nucleus with hyperfine coupling, A, Mims pulsed
ENDOR has a response R which depends on the product, Aτ,
according to the equation,
R ~ [1 - cos(2π A)]
(1)
The response function R (eq 1) equals zero, corresponding to
minima in the Mims ENDOR response (hyperfine “suppres-
sion holes”), when Aτ = n, n = 0, 1, …, and maxima at Aτ =
(2n+1)/2; n = 0, 1,… .33 The “holes” at n = 1, 2, 3… can be
adjusted by varying τ, but the ‘central’, n = 0, hole at ν = νN
persists regardless. This can be of significance in distinguish-
ing a tensor that is dominated by anisotropic interactions from
one that is dominated by isotropic ones. The latter would nev-
er lead to ENDOR intensity near νN; the former does so for
certain orientations, but the ν = 0 Mims hole tends to diminish
the differences between the two cases. Simulation of the
ENDOR patterns employed the program, ENDORSIM.34
To examine the dependence of PFL-AE activity on the
concentration of K+, PFL-AE was assayed with 500 mM, 300
mM, 200 mM, 100 mM, 75 mM, 50 mM, 25 mM, 10 mM, or
0 mM KCl in a 100 mM Tris, pH 7.6 buffer with 1 mM DTT.
The ionic strength was maintained at a constant value by add-
ing choline chloride for those with less than 100 mM KCl and
activity was monitored.
EPR Sample Preparation and Spectroscopy
In an MBraun anaerobic chamber with ≤1 ppm O2, 300 µM
PFL-AE, 100 µM 5-deazariboflavin (added last in the dark), 5
mM DTT in a 100 mM Tris, pH 7.6 buffer containing either
no cation, or 100 mM of monovalent cation salt (NaCl, KCl,
NH4Cl, RbCl, CsCl, or LiCl), or 1 mM of divalent cation salt
(CaCl2, MgCl2, or Zn(NO3)2) was placed in an EPR tube. Pho-
toreduction was accomplished by illumination with a 300 W
halogen lamp in an ice bath for one hour, and samples were
either frozen as-is or SAM was added to a final concentration
Results
Refinement of Metal Ions/Water Molecule in the
Putative Metal Site
The X-ray crystal structure of PFL-AE/SAM/7-mer PFL
peptide10 reveals a previously unreported putative metal bind-
ing site in the vicinity of the active site. Two aspartic acid side
chains (D104 and D129), two carbonyl oxygens from residues
ACS Paragon Plus Environment

Journal of the American Chemical Society
Page 4 of 13
1
2
3
4
5
6
7
8
T105 and M127, and one oxygen from the carboxylate of
with the crystallographic data, we set the occupancies of these
SAM point towards the unidentified electron density in a dis-
torted trigonal bipyramidal geometry, with each of the poten-
tial oxygen ligands within 3 Å of the center of the site (Fig-
ures 2 and S1). We thus carried out further refinements di-
rected towards characterizing this putative metal site, using an
approach similar to that described previously.35
ions/molecule to 100%, and compared the B-factors for each
following refinement (Tables S3 and S4). A resulting B-
factor lower than the average of the surrounding area indicates
that the ions/molecule has too little scattering to account for
the density, whereas if the B-factor is higher, that is an indica-
tion of too much scattering. The best fit is the ion/molecule
with a B-factor that is most similar to the average in that area.
We find that water has significantly lower than the average B-
factor compared to the surrounding atoms, calcium and potas-
sium are significantly higher, and sodium and magnesium are
the closest to average (Table S4). In addition, more extensive
positive Fo-Fc electron density is present in this site when
modeled with water than when modeled with sodium or mag-
nesium (Figure S2), and negative Fo-Fc electron density ap-
pears when the site is modeled as potassium or calcium (Fig-
ure S2). These observations are consistent with presence of
sodium or magnesium in this site.
The most likely candidate ion to occupy this site in the
crystal structure is Na+, as this cation is present in the crystal-
lization buffer as well as throughout the protein purification.
The protein also is exposed to Mg2+ in early purification steps
and so this cation was considered as well. In addition, we
considered the possibility that a water molecule might account
for the electron density observed in this site. For complete-
ness and comparison, we also carried out refinements with the
species in this site assigned as K+ and Ca2+.
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
As an additional approach, we set the B-factor of the
ions/molecule to the average of the surrounding area, allowed
the occupancies to refine, and compared the resulting occu-
pancies (Table S5). An occupancy value of higher than 100%
would indicate that the modeled species has too little scatter-
ing, whereas a lower occupancy would reflect that the mod-
eled species has too much scattering power. Whereas water
had too high an occupancy, the other ligands had occupancies
of less than 100%. As expected, sodium and magnesium had
the most reasonable occupancy values. Again, these observa-
tions are consistent with presence of sodium or magnesium in
this site, with sodium more likely because it was known to be
present in our crystallization buffer at high concentrations (3.5
M sodium formate) and exposure of the protein to magnesium
was limited to early purification steps.
We also considered coordination geometry and the average
distance of the oxygen ligands (Tables S1 and S2) in our as-
signment of the likely ion/molecule to occupy this site. Based
on these criteria, water is not a good candidate, as it would not
be capable of forming the number (5) and the geometry (dis-
torted trigonal bipyramidal) of interactions observed. In con-
trast, trigonal bipyramidal is a common geometry for metal
ions. The average metal-oxygen distance (2.5 Å) agrees with
known sodium ion sites in protein structures as well as the
Cambridge Structural Database.36 Our conclusion is therefore
that Na+ occupies this metal binding site in the PFL-
AE/SAM/peptide crystal structure (Figure 2). However, a
different metal ion might occupy this site in vivo.
Figure 2. Putative metal site with sodium modeled (purple
sphere). Protein amino acids are shown in grey, a portion of
SAM is shown in green, and [4Fe-4S] cluster is shown in
yellow and rust. (a) 2Fo-Fc electron density map contoured
at 0.75 σ and displayed as a blue mesh within 2 Å of the so-
dium ion. Refined distances to sodium are also displayed. (b)
Angles between metal ligands. (c) Stereoview of the putative
metal site with sodium modeled.
EPR and the Effect of Cations on the Electronic
Structure of PFL-AE
The proximity of the monovalent cation site to the [4Fe-4S]
cluster suggested the cation might affect the electronic and/or
metrical structure of the active site of PFL-AE. This was in-
vestigated using EPR spectroscopy of PFL-AE in the presence
of a variety of monovalent and divalent cations, and with and
without SAM.
In the absence of SAM, the signal from the [4Fe-4S]+ clus-
ter changes with the presence and identity of the cation (Fig-
ure S3). The M+-free signal is slightly heterogeneous, and the
heterogeneity varies with the identity of M+.
The locations (Figure S2) and B-factors (Table S4) of the
four protein-derived oxygen ligands do not change significant-
ly after refinement in REFMAC, regardless of whether water,
sodium, potassium, magnesium, or calcium was modeled into
the site (Figure S2), though small shifts do occur to accom-
modate the different atoms’ radius (for example, compare the
distances listed in Tables S1 and S2). To examine whether
the presence of water or a metal ion provides better agreement
As has been previously shown,9, 37 the addition of SAM to
PFL-AE causes a marked change in the cluster signal. The
ACS Paragon Plus Environment

Page 5 of 13
Journal of the American Chemical Society
1
2
3
4
5
6
7
8
majority rhombic component of the SAM-free signal converts
To further analyze how different cations affect the active
to a majority axial signal, with g|| = 2.00, g┴ = 1.89, with a
minority rhombic component (see Table S6 for all g-values),
and an intensification of the [4Fe-4S]+ EPR signal.
site in PFL-AE, CD spectroscopy in the visible region was
performed on the M+-free purified protein in the presence and
absence of different monovalent cations and SAM. CD spec-
troscopy in the visible region is used to analyze d-d transitions
of metals in chiral environments such as the [4Fe-4S] clusters
in PFL-AE. In the absence of SAM, the M+-free spectrum is
unaffected by the addition of a cation (Figure S6). On the
other hand, when SAM is present, new spectral features ap-
The effects of the monovalent cation on the EPR signal are
even more pronounced in the presence of SAM, which arrang-
es the spectra in order of increase M+ ionic radius (Figure 3).
The EPR signals for M+-free PFL-AE in the absence of SAM
and for all the PFL-AE samples in the presence of different
monovalent cations and SAM were simulated as a sum of
spectra from two different enzyme states (Figure S4). One
state has a near-axial g-tensor, g1 = [2.01, 1.89, 1.88], similar
to those previously reported for photoreduced PFL-AE9, 37 and
independent of the added cation; the second state has a rhom-
bic g-tensor, but with precise values that depend slightly on
the identity of the cation: g2 ≈ [2.01, 1.92, 1.87] (Table S6).
The relative populations of the two states varies strongly with
cation. For both the smallest (Li+) and largest (Cs+) cations the
near-axial signal is the majority component of the signal. In-
deed, those spectra are little different from the spectrium in the
absence of M+, although ENDOR spectroscopy shows that
both cations do bind (see below). The signal from the sample
with K+ has by far the largest contribution from the rhombic
component, with g2 = 1.920; activity measurements presented
below suggest this represents the active conformation.
9
+
pear in the presence of K+, NH4 , Rb+, and Na+ (Figure S7),
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
with the most prominent at approximately 400 nm and 500
nm. These features grow in with increasing concentrations of
K+ (Figure S8), plateauing after 50 mM KCl. A fit of these
spectral changes yielded, KD = 7.6 + 1.2 mM (Figure S8).
Above ~180 mM KCl the intensity drops off drastically, which
we attribute to perturbation of the active site due to the high
ionic strength.
Altering the Cation Binding Site
To further characterize the cation binding site in PFL-AE,
the two conserved aspartate (D104 and D129) residues coordi-
nating the monovalent cation were changed to alanine (D104A
and D129A, respectively). D129 of PFL-AE hydrogen bonds
to the ribose hydroxyls of SAM in addition to coordinating the
monovalent cation;10 the corresponding residue in the radical
SAM enzyme biotin synthase is a catalytically essential aspar-
agine (N153) that also hydrogen bonds to a ribose hydroxyl of
SAM but does not coordinate a monovalent cation.38-39
PFL-AE activity assays in the presence of 100 mM KCl re-
vealed that the D104A variant had very low activity (5.5 ± 0.2
U/mg, compared to 54.2 ± 1.5 for WT enzyme). EPR analysis
of this variant confirms that D104 is important for cation bind-
ing. Firstly, none of the monovalent cations had a significant
effect on the EPR signals for [4Fe-4S]+-PFL-AE(D104A) or
[4Fe-4S]+-PFL-AE(D104A)/SAM, with the EPR spectra in all
Figure 3. EPR spectra of PFL-AE in the presence of SAM
and the indicated monovalent cation. Spectra are stacked in
order from smallest (Li+) to largest (Cs+) cation. Solid verti-
cal lines are drawn at g values of the nearly axial signal
(2.005 and 1.889), and dashed vertical lines are drawn at
typical g-values of the more rhombic signal (2.007, 1.920,
1.868). The origin of the modest variation in signal intensity
is unknown. Each sample contained 300 µM PFL-AE in the
presence of 100 mM of the cation indicated and 2 mM SAM.
EPR parameters: 12 K, 9.37 GHz, 1.0 mW.
g-value
Figure 4. EPR spectra of the PFL-AE wild type (solid line)
and variants D104A (dashed line) and D129A (dotted line)
in the presence of SAM, and in the absence or presence of
100 mM K+ or Na+. EPR parameters are as for Figure 3.
cases looking very similar to the buffer exchanged M+-free
D104A variant (Figure S9). In addition, the EPR signal ob-
served for D104A in the presence of SAM is similar to D104A
M+-free in the absence of SAM, suggesting SAM does not
bind well to in this variant (Figure 4).
EPR spectra of PFL-AE in the presence of the Ca2+ or Mg2+
were essentially identical to those for the M+-free PFL-AE
both for SAM-free and SAM-bound enzyme (Figure S5), and
activity measurements (see below) support the suggestion that
unlike the alkali metal cations, divalent cations do not bind.
ACS Paragon Plus Environment

Journal of the American Chemical Society
Page 6 of 13
1
2
3
4
5
6
7
8
The PFL-AE(D129A) variant, on the other hand, retains the
for 23Na+ corresponds to only ~ 3 × 10-5 of a spin in the 23Na 3s
orbital.
ability to bind cations. The EPR spectra of PFL-AE(D129A)
vary with the presence and identity of cation, particularly in
the presence of SAM (Figure 4), giving rise to spectra that are
very similar to those of the WT enzyme, indicating that this
variant binds M+ and SAM in a manner similar to WT. Fur-
ther, in the presence of K+ this variant exhibits only moderate-
ly lower enzymatic activity than WT enzyme (42.7 ± 3.4 vs
54.2 ± 1.5 U/mg).
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
ENDOR Studies of Cation Binding Near the Clus-
ter
To identify and further characterize cation binding to the
newly identified site, we carried out ENDOR studies charac-
terizing the interactions between the cation nuclear spin and
the electron spin of the [4Fe-4S]+ cluster of PFL-AE. To our
knowledge, the only previous study showing hyperfine cou-
pling to a monocation located nearby an FeS cluster involved a
Cs+ ion bound to a residue near the S = ½ [Fe3S4]+ cluster of
the hyperthermophilic archaeon Pyrococcus furiosus.14 The
present ENDOR studies both directly examine the cation
bound in the active site of PFL-AE, and probe how the pres-
ence and identity of the cation influences the positioning of
SAM.
M+ Binding: The first result of this approach is confirma-
tion that Na+ occupies this metal binding site in the PFL-
AE/SAM complex in a solution of enzyme isolated with buff-
ers that contain Na+. A 35 GHz Mims ENDOR spectrum col-
lected from such a sample (Figure 5A) exhibits a doublet cen-
tered at the 23Na Larmor frequency, split by a hyperfine cou-
pling, A = 0.14 MHz. As shown, this doublet is barely detecta-
ble in a sample purified in the absence of 23Na+ (or any alkali
metal cation) added to the buffer. The full 23Na+ intensity is
restored by 23Na+ loading, and the weak signal is completely
eliminated upon treatment with, and replacement of 23Na+, by
a range of other monovalent cations, as illustrated for Tl+.
Clearly, studies to date have invariably involved enzyme with
23Na+ bound in the active site, since standard purification pro-
cedures involve buffers containing NaCl. Indeed, in hindsight,
this signal was initially seen, although not assigned, in our
early studies of PFL-AE.23
Figure 5B shows the 2D pattern of 23Na+ Mims ENDOR
spectra collected across the EPR envelope of the PFL-
AE/Na+/SAM complex. The field-dependent features are well
described by an axial hyperfine-coupling matrix, A = [0.19, -
0.14, -0.14] (MHz), which can be decomposed into an axial
anisotropic interaction, T = [2T, -T, -T], T = 0.11 MHz, and,
surprisingly an isotropic coupling, aiso = -0.03 MHz. The pres-
ence of this latter coupling requires that the spin-density delo-
calized onto the SAM carboxylate ligand to the unique Fe is
further delocalized onto Na+. Whether this delocalization in-
volves weak covalency or merely ‘Pauli delocalization’40-43
through intimate contact between carboxylate-O and Na+, its
presence does require direct interaction/contact between Na+
and this ligand, with spin delocalization as indicated in
Scheme 1. Incidentally, the isotropic coupling to a monocation
equals the spin density in its ‘valence’ s-orbital, ρs multiplied
Figure 5. 35 GHz Mims 23Na ENDOR spectra of PFL-
AE/SAM: (A) 23Na ENDOR spectra at g2 (g~1.89) for PFL-
AE/SAM/as-isolated (black), Na+ treated (red), M+-free
(brown) and Tl+ treated (green) (B) 2D field-frequency pattern
of 23Na ENDOR spectra for PFL-AE/SAM/as-isolated (black).
Experimental fields as indicated. ENDOR simulations are
shown in blue. Experimental conditions: microwave frequency
~34.8 GHz; microwave pulse length (π/2) = 50 ns; τ = 1250
ns; and T = 2 K. Simulation parameters: aiso = -0.03 MHz, T =
0.11 MHz, Euler angle β = 30°.
The dipolar constant, T, is somewhat larger (+0.03 MHz)
than would be expected for the through-space interaction spin
on the unique Fe ion that is 6.5 Å away from Na+ (or all Fe
ions on the cluster), which can be explained by additional in-
teraction with the small spin densities on the carboxylate. The
dipolar interaction with a center bearing spin-density ρ at dis-
Scheme 1. M+-SAM-cluster interaction. M+ is
shown in purple, iron in rust, and sulfur in yellow.
M+
+
by the coupling for a single electron in that orbital, As, aiso
=
Asρs, and use of a tabulated value for As44 shows that the aiso
ACS Paragon Plus Environment

Page 7 of 13
Journal of the American Chemical Society
1
2
3
4
5
6
7
8
tance r from the nucleus scales as T ~ ρ/r3, and the difference
between the measured T and that calculated for ρ = 1 at 6.5 Å
is equivalent to an interaction with ρ = 1 % of a spin on the
carboxylate-O at r = 2.4 Å.
a tight ion-pair with the carboxylate-O, as indicated by the
presence of a small isotropic contribution to the hyperfine
interaction, Figure S10. The coupling to Tl+ is ~50 times larg-
er than those to the other ions, A ~ 2 MHz, simply because As,
the coupling per s-orbital electron for this high atomic-number
ion is much larger than those for the other ions.44 For compari-
son, we note that Tl+ actually incorporated into the [Fe3S4]+
cluster of the hyperthermophilic archaeon Pyrococcus furiosus
to form a S = ½ [TlFe3S4]+ cluster displays a hyperfine cou-
pling of A ~ 370 MHz.14 whereas ENDOR of the radical-pair
intermediate trapped during reaction of the B12 enzyme diol-
dehydrdratase shows no hyperfine coupling to Tl+ bound in the
monocation site of this enzyme.45
Finally, the Li+ and Cs+ spectra show additional intensity at
the Larmor frequency associated with ‘distant’ ions, presuma-
bly present in the solvent-accessible cavity at the active site.
ENDOR Study of M+ Influence on SAM Binding
Geometry
We next examined the samples treated with the other M+
ions. Hyperfine coupling between the reduced cluster and the
cation is observed for Li+, Cs+, and Tl+, as well as Na+ (Figure
6), demonstrating that these three also bind to the cation site.
Figure S10 shows the 2D field-frequency pattern of Mims
ENDOR spectra for Li+, Cs+, and Tl+ ions, and Table S8 con-
tains the hyperfine-tensor parameters and ns-orbital spin den-
sities derived from their simulation. The spin densities for the
other three ions bound to PFL-AE are within a factor of ~3 of
that for Na+. Although no signal was observed for K+ or Rb+,
their presence abolished the residual 23Na+ signal; that, and the
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
To test for a possible influence of the cation on the posi-
tioning of the sulfonium target for SAM cleavage relative to
the PFL-AE [4Fe-4S] cluster, we used ENDOR to study the
PFL-AE/SAM complex with the sulfonium methyl labeled
with 13C. We had previously studied the CH3 positioning for
the Na+-bound PFL-AE, although the presence of the cation
was not recognized at the time.23 Analysis of 2D field-
frequency 13C ENDOR patterns revealed an isotropic hyper-
fine contribution, which requires that SAM lies in direct con-
tact with the cluster. From the ENDOR-derived anisotropic
2
hyperfine coupling contributions for H and 13C we estimated
the distance from the closest methyl proton of SAM to the
closest iron of the cluster to be ~3.0-3.8 Å, while the distance
from the methyl carbon to the nearest iron to be ~4-5 Å.23 The
subsequent X-ray structure determination confirmed, and
made precise this picture.10
Figure 7 shows 13C Mims ENDOR spectra for PFL-
AE\methyl-13C-SAM\M+ collected at g2, arranged according to
the ionic radius of M+, as suggested by reactivity measure-
ments described immediately below. For each M+, the spec-
trum is a doublet with hyperfine coupling, A ~ 0.3 MHz. How-
ever, the shapes of the 13C doublets clearly vary with M+, and
do so in a fashion that roughly reflects the M+ ionic radius.
Thus, for Li+, the smallest M+, the individual peaks of the
doublet are smooth and unstructured. The peaks for Na+ clear-
ly instead show structure, and this structure is most evident in
the spectrum for K+, which ion also appears to cause a slight
increase in the doublet splitting, suggesting a slightly stronger
interaction with the cluster spin. The structure persists through
Cs+; it not only vanishes with Tl+, but the individual peaks are
the narrowest of all for this ion. For completeness, the includ-
ed spectrum with no M+ is also structured.
This ‘structure’ in the 13C-methyl ENDOR spectra is clear-
ly related to the heterogeneity in the EPR signals (Figure 3).
We interpret it to mean that the ENDOR spectra are a super-
position of distinct 13C signals from the two major active-site
conformations associated with the two major components of
the EPR signal, States 1 and 2 (Figure 3). This in turn implies
that the position of the sulfonium methyl relative to the cluster
Fe, and thus its ENDOR spectrum, is different in in these two
conformations K+ stands out among the ions in Group 1 of the
periodic table in the high occupancy of the ‘rhombic’ configu-
Figure 6. (top) Mims ENDOR spectra at g2 are centered at each
nuclei Larmor frequency. (bottom) CW ENDOR spectrum at
first g2 (g ~ 1.94). Larmor frequencies for 203Tl (30 % natural
abundance) and 205Tl (70 % natural abundance) are 31.93 and
32.23 MHz, respectively. Conditions: For Mims ENDOR, same
as Figure 5. For CW ENDOR, microwave frequency ~ 34.97
GHz, microwave power 1 mW, 100 kHz modulation amplitude
0.3 G, rf sweep rate 2 MHz/s and T = 2 K.
influences of those ions on the 13C ENDOR of 13CH3-SAM
(see below) show that they likewise occupy the cation binding
site. The major isotope of each of these two nuclei has an ex-
tremely small nuclear g-factor; it is the resulting low Larmor
frequencies and electron-nuclear hyperfine interactions that
render their signals unobservable.
The 2D field-frequency patterns of ENDOR spectra were
collected for each of the three ENDOR-observable alternative
ions (Li+, Cs+, and Tl+), and analyzed as for 23Na+. The data
reveal that all four cations, whether small (Li+), large (Cs+,
Tl+), or in between (Na+), are able to bind in the cation site as
ACS Paragon Plus Environment

Journal of the American Chemical Society
Page 8 of 13
1
2
3
4
ration and the corresponding ‘structure’ in the 13C-methyl
ENDOR spectrum.
ons bind in the metal site. In the presence of divalent cations
(Ca2+, Mg2+, Zn2+) PFL-AE showed little-to-no activity (Fig-
ure 8 and Table S7).
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
Figure 8. PFL-AE specific activity (U/mg) is influenced by the
ionic radius of the cation present in the activity assay and thus
the active site. Monovalent cations stimulate PFL-AE activity
while divalent cations (blue box) inhibit PFL-AE activity rela-
tive to that in the absence of these cations (green-dashed line).
The black dotted line is at zero activity. The red line is provided
to guide the eye. For these assays, a concentration of 5 µM PFL,
0.05 µM PFL-AE, and 0.1 mM SAM in the presence of 100 mM
of the specified monovalent cation or 1 mM of divalent cation
was used. Data are the average of at least three different trials
with each trial done in triplicate. “M+-free” is PFL-AE purified
in the absence of any alkali metal cation and no cation is added
to the assay.
By carrying out the PFL-AE assay at a range of K+ concen-
trations, we were able to estimate a KD of 44. + 11. mM for K+
binding to the PFL-AE/SAM complex under assay conditions
(Figure S8). Although this value is higher than that estimated
via the CD titrations described above (7.6 + 1.2 mM), the ex-
perimental conditions vary (including temperature, presence
and concentrations of various buffer components, presence of
the PFL-AE substrate PFL in the activity assays) by necessity
between these two experiments, and we attribute the differing
values to these differences. We previously used EPR spectros-
copy to estimate KD = 6.5 mM for K+ binding to PFL-
AE/SAM,46 a value in line with the CD measurement reported
herein. Regardless, KD in the range of 7 – 44 mM puts the
binding affinity in a range that would provide near maximal
PFL-AE activity at the typical in vivo K+ concentrations in E.
coli of ~200 mM.47
Figure 7. 35 GHz 13C Mims ENDOR spectra at g2 for PFL-
AE/methyl-13C SAM/M+. Experimental conditions: As for
Figure 5 except τ = 600 ns.
Effect of Cations on PFL-AE Activity
The cation-dependence of the EPR spectrum of PFL-
AE/(M+)/SAM, the ion-pairing of the cation with the carbox-
ylate-ligand of the iron-sulfur cluster, and the cation-
dependence of the SAM positioning in the PFL-AE active site
as shown by ENDOR, all suggested a possible role for this
cation in the enzyme-catalyzed reaction. To test this, enzyme
activity assays were carried out in the presence of each of six
+
monovalent cations (Na+, K+, NH4 , Rb+, Cs+, and Li+) and
three divalent cations (Ca2+, Mg2+, and Zn2+), as well as in the
absence of any monatomic or ammonium cations. These ex-
periments were carried out with PFL-AE that had been puri-
fied in the absence of such cations, using a Tris-Cl buffer with
no added alkali metal salt. The results reveal a strong depend-
ence of PFL-AE activity on the identity of the cation, with K+
providing the highest activity (Figure 8 and Table S7).
Among mono-cations, going down Group 1 of the periodic
Table from Li+ to Cs+ Figure 8 shows a sharp maximum at the
ionic radius of K+, with NH4+ closely matching the efficacy of
K+. The 13C ENDOR measurements correspondingly suggest
that K+ likewise has a distinctive effect on the positioning of
the sulfonium target for SAM cleavage (Figure 7). The lack of
stimulation by Li+ and Cs+ is not due to an absence of cation
binding, as ENDOR provides evidence that both of these cati-
Discussion
Close examination of the X-ray crystal structure revealed
the presence of a monovalent cation in the PFL-AE active site
coordinated by 5 oxygens: two carboxylate oxygens from con-
served aspartate residues (D104 and D129), two backbone
carbonyls from T105 and M127, and one carboxyl oxygen of
SAM. This cation binding site is located at the back of the
active site against the barrel, and given its interactions with
both conserved amino acid residues and SAM, we hypothe-
sized that this cation likely played an important role in active
site structure, SAM binding, and enzymatic activity.
To probe monovalent cation binding in solution, we used
EPR and ENDOR spectroscopies. Clear perturbations in the
ACS Paragon Plus Environment

Page 9 of 13
Journal of the American Chemical Society
1
2
3
4
5
6
7
8
EPR lineshapes and g-values were observed that depended on
that monovalent alkali metal ions stimulate PFL-AE activity in
the presence and identity of the cation, with the most dramatic
effects occurring when SAM is also present. These observa-
tions are consistent with the binding of monovalent cations in
the PFL-AE active site as identified by X-ray crystallography,
with SAM providing one of the ligands to the monovalent
cation. The maximum in activity at K+ seen in going down
Group 1 of the periodic table (Figure 8) correlates with the
maximum contribution of a ‘rhombic’ component to the EPR
of the enzyme with bound SAM/K+ (Figure 3). 13C ENDOR
spectroscopy correspondingly shows that the location of the
target sulfonium of SAM is different in the ‘rhombic’ confor-
mation (Figure 7), which suggests that the size of the cation
plays a key role in the optimal positioning of SAM relative to
the [4Fe-4S] cluster (Figure 7), and that the positioning in the
‘rhombic’ conformation enhances reactivity.
a pattern that correlates with their position in Group 1 of the
periodic table. Figure 8 suggests this correlates with a tight
activity optimum at the ionic radius of K+, with gradual falloff
to lower radius, a sharp falloff to higher. Potassium is general-
ly the most abundant alkali metal cation in E. coli cells, with
concentrations typically ~200 mM.47 Half-maximal activation
was obtained at [K+] = 44 ± 10.5 mM, reflecting a binding
constant in the physiologically relevant range.
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
Divalent cations do not activate PFL-AE, and relative to
enzyme isolated in the absence of monatomic or ammonium
cations, divalents appear to be slightly inhibitory (Figure 8).
However, although our data appears to show that in the ab-
sence of monovalent cation PFL-AE retains a small amount
(~10%) of activity, it is instead likely that this residual activity
is due to trace amounts of monovalent cations present even in
“metal-free” protein. This conclusion is supported by the weak
but unambiguous Na+ ENDOR signal from nominally ‘M+-
free’ PFL-AE, Figure 5. Thus, we suggest that the divalent
cations should be viewed as being non-activating, rather than
inhibitory. Indeed, there is no direct evidence that these diva-
lent cations really bind.
ENDOR spectroscopy has also been used to probe the elec-
tronic coupling of the monovalent cations to the [4Fe-4S] clus-
ter. As shown in Figure 6, an ENDOR signal due to the inter-
action between the nuclear spin of the cation and the electron
spin of the [4Fe-4S]+ cluster is observed for cations that bind
in the metal site identified by X-ray crystallography. Such a
signal is observed for Li+, Na+, Cs+, and Tl+, indicating that
these cations are indeed bound at this metal-binding site in
close proximity to the [4Fe-4S]+ cluster. While we do not ob-
serve ENDOR signals in the case of K+ and Rb+, this is at-
tributed to the low Larmor frequencies and electron-nuclear
hyperfine interactions for these nuclei rather than to lack of
binding; in fact, addition of either K+ or Rb+ abolishes the Na+
ENDOR signal for PFL-AE purified in Na+-containing buff-
ers, providing indirect confirmation that these cations bind in
the same site. The existence of an isotropic hyperfine coupling
of the spin on the [4Fe-4S] cluster to the ions with observed
ENDOR response shows the existence of intimate contact
between the Fe-bound carboxylate-O of SAM and the M+ ion,
as indicated in Scheme 1.
The monovalent cation binding site in PFL-AE involves
two conserved aspartic acid residues, D104 and D129. Chang-
ing either of these amino acids to alanine decreased activity,
although the decrease was modest for D129A and dramatic for
D104A. Analysis of the EPR spectra of the two aspartate vari-
ants suggest that D104 is required for cation binding while
D129 plays a less critical role. First, changing the monovalent
cation induces almost no change in the EPR signal for D104A,
suggesting cations are not binding effectively to this site.
Second, the EPR signal of D104A is not significantly altered
in the presence of SAM, indicating that the lack of cation
binding negatively impacts SAM binding. In contrast, EPR
spectroscopic studies of the D129A variant indicate that the
monovalent cations bind, and SAM coordinates in a manner
similar to that of the WT.
Focusing on the strong dependence of steady-state rates on
the presence and identity of the monovalent cation, they show
ACS Paragon Plus Environment

Journal of the American Chemical Society
Page 10 of 13
1
2
3
4
5
6
7
8
Observation of an active-site monovalent cation in PFL-AE
screening other enzymes for cation binding. Understanding the
that directly interacts with SAM and the [4Fe-4S] cluster and
affects catalytic activity is an unexpected and significant find-
ing. To evaluate the prevalence of such a cation site in the
radical SAM superfamily, we performed an alignment of PFL-
AE with other radical SAM enzymes, including other glycyl
radical enzyme-activating enzymes (GRE-AEs), with a focus
on the DTX21MXD motif that spans the ligands to the mono-
valent cation in PFL-AE (Figure 9). The key acidic residues
that serve as ligands to the monovalent cation in PFL-AE are
conserved in only a subset of the activating enzymes, and are
absent in other radical SAM sequences examined, although in
some cases alternative potential ligands replace the aspartate
residues in the DTX21MXD motif (Figure 9). A third ligand
to the cation site in PFL-AE is a backbone carbonyl of T105;
interestingly, this residue is conserved in many of the radical
SAM sequences, including several for which the acidic residue
ligands are not conserved. Based on the limited conservation
of the acidic residue ligands to the monovalent cation, it ap-
pears that this monovalent cation is present in many, but not
all, of the GRE-AE enzymes in the radical SAM superfamily.
ENDOR spectroscopy appears to provide a means of rapidly
detailed mechanism by which the monovalent cation serves to
enhance catalytic activity of PFL-AE, and presumably several
other GRE-AEs, awaits further studies.
Enzymes activated by metal cations such as Na+ or K+, in-
cluding the many enzymes whose catalytic mechanism does
not directly incorporate a metal center, can be classified as
cofactor-like (type I) or allosteric (type II).48 In type I en-
zymes, the M+ is absolutely required for enzyme activity and
helps to anchor the substrate in the active site, often acting in
tandem with a divalent cation. In type II enzymes, no direct
interaction between M+ and substrate occurs and the M+ is not
absolutely required for activity or substrate binding. Our ob-
servation that PFL-AE most likely is inactive in the absence of
a monovalent cation and that the cation binds in the active site
and is coordinated by SAM, a co-substrate with PFL in this
reaction, plus the evidence from EPR spectroscopy that mono-
valent cation binding is required for SAM binding, points to
PFL-AE being a type I M+ activated enzyme. It is interesting
to note, however, that as yet there are few reported cases of
metalloenzymes that are activated by a monovalent cation
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
Figure 9. Sequence alignment of PFL-AE with other activating enzymes and selected radical SAM enzymes. This alignment focuses on
the PFL-AE DTX21MXD motif, and has been truncated to show only the region between beta2 to alpha4 for clarity. In the sequences
that have been characterized structurally, beta strands are highlighted in yellow and alpha helices are highlighted in cyan. The residues
of the PFL-AE DTX21MXD motif are highlighted in pink and shown in bold text. Residues of this motif that are conserved in the other
sequences are highlighted in pink, and conservative substitutions of residues in the motif are highlighted in maroon with bold white
lettering. Note that at the positions corresponding to D104 and D129, acidic residues are highlighted in pink, and any other amino acid
side chain that could act as a hard metal ligand is highlighted as a conservative substitution. The sequences aligned are: E. coli pyruvate
formate-lyase activating enzyme (UniProtKB ID P0A9N4), T. aromatica Benzylsuccinate synthase activating enzyme (O87941), P.
difficile 4-hydroxyphenylacetate decarboxylase activating enzyme (C9XIS7), C. butyricum Glycerol dehydratase activator (Q8GEZ7),
E. coli Anaerobic ribonucleoside-triphosphate reductase-activating protein (P0A9N8), C. perfringens Anaerobic sulfatase-maturating
enzyme (Q0TTH1), E. coli Oxygen-independent coproporphyrinogen III oxidase (P32131), E. coli biotin synthase (P12996), S. aureus
Molybdenum cofactor biosynthesis protein A (P65388), C. subterminale lysine 2,3-aminomutase (Q9XBQ8), C. hydrogenoformans
HydG protein (Q3ABV3), T. maritima [FeFe]-Hydrogenase maturase (Q9X0Z6), B. circulans BtrN (Q8G907), E. coli Ribosomal RNA
large subunit methyltransferase N (P36979), and G. thermodenitrificans spore photoproduct lyase (A4IQU1).
ACS Paragon Plus Environment

Page 11 of 13
Journal of the American Chemical Society
1
2
3
4
5
6
7
8
binding in the active site in proximity to the catalytic metal
cation. This behavior links this radical SAM enzyme to the
center. Among these are the Mn(II) fosfomycin resistance
enzyme, FosA,15 and the B12–dependent dioldehydrase.16 We
speculate that this may well change.
B12-dependent dioldehydratase, yet another parallel between
the two classes of radical-generating enzymes, while the mul-
tiple approaches to the characterization of the PFL-AE cation
provide even greater depth to its characterization. The cation’s
location, catalytic importance, and modulation of substrate
position all indicate its central mechanistic role. Indeed, this
role might even be the same in both enzymes, control of an M-
C bond cleavage to generate the 5’-dAdo radical.
The monovalent cation might be playing one or more of a
variety of roles. It clearly plays an important role in SAM
binding and positioning in the active site of PFL-AE. The
bridging coordination of the SAM carboxylate between the
unique iron of the [4Fe-4S] cluster and the monovalent cation
could affect the SAM-cluster interaction and thus reactivity.
This would be consistent with the proposed role for the K+
bound in the B12-dependent dioldehydrase, where the binding
energy of the cation is thought to induce a conformational
change that stimulates homolytic Co-C bond cleavage.16 We
have recently shown that PFL-AE catalysis involves an organ-
ometallic intermediate in which the 5'-C of the 5'-
deoxyadenosine is directly bound to an iron of the [4Fe-4S]
cluster;6 this intermediate is analogous to the B12 cofactor,
which has a bond between the 5'-C of 5'-deoxyadenosine and
the cobalt of cobalamin. By extension then, one might suggest
that the monovalent cation in the active site of PFL-AE pro-
motes Fe-C bond cleavage in this intermediate. Regardless, the
observation that both the radical-generating B12 enzyme diol-
dehydrase and the radical SAM enzyme PFL-AE incorporate a
monovalent cation in the active site, near the catalytic metal
and in contact with substrate, potentially with the role of influ-
encing conformation and promoting metal-carbon bond cleav-
age, provides yet one more intriguing parallel between radical
SAM and B12 enzymes.
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
ASSOCIATED CONTENT
Supporting Information
Supporting Information document includes details on protein
production and purification, as well as Figures S1 – S11 and Ta-
bles S1 – S8.
The Supporting Information is available free of charge via the
Internet at http://pubs.acs.org.
AUTHOR INFORMATION
Corresponding Author
* jbroderick@chemistry.montana.edu, (406) 994-6160.
Present Addresses
¥ Department of Applied Biochemistry and Food Science, Saga
University, Saga 840-8502, Japan
The role of the monocation could also include interacting
with the methionine after SAM cleavage, perhaps orchestrat-
ing its departure from the active site. The presence of the
monovalent cation and its interaction with SAM affects the
charge balance in the active site, and how this might affect
catalysis is unclear. Finally, and at the most basic level it does
appear that the monovalent cation confers some stability to
PFL-AE: the M+-free enzyme is more susceptible to stresses
such as freeze-thaw events than the enzyme purified in the
presence of monovalent cations.
Author Contributions
The manuscript was written through contributions of all authors.
All authors have given approval to the final version of the manu-
script.
Funding Sources
This work was funded by the National Institutes of Health
(GM054608 to J.B.B. and GM111097 to B.M.H.), and the Na-
tional Science Foundation (MCB-0543833 to C.L.D.).
Conclusion
We describe here for the first time the presence of a catalyt-
ically important monovalent cation in the active site of a radi-
cal SAM enzyme, PFL-AE. This monovalent cation is ob-
served in the crystal structures of PFL-AE and identified as
Na+ in these structures. We have probed the solution state of
the enzyme using EPR and ENDOR spectroscopies, and have
shown that a wide range of monovalent cations, including all
but the largest alkali metals, is able to bind in this site, but
going down Group 1 of the periodic Table from Li+ to Cs+
there is a sharp activity maximum at K+, suggesting a precise
requirement on the ionic radius. Indeed, the monovalent cation
is coordinated by a carboxylate oxygen of SAM, with the oth-
er carboxylate oxygen of SAM coordinated to the unique iron
of the [4Fe-4S] cluster, in a direct linkage between the mono-
valent cation and the catalytic cluster that allows spin transfer
from the cluster to the mono-cation.
ACKNOWLEDGMENT
This work was supported by the National Institutes of Health
grants GM54608 (to J.B.B.) and GM111097 (to B.M.H.), and by
the National Science Foundation grant MCB-0543833 (to C.L.D.).
C.L.D. is a Howard Hughes Medical Institute Investigator.
ABBREVIATIONS
PFL, pyruvate formate-lyase; PFL-AE, pyruvate formate-lyase
activating enzyme; SAM, S-adenosyl-L-methionine; 5’-dAdo•,
5’-deoxyadenosyl radical; EPR, electron paramagnetic reso-
nance; ENDOR, electron nuclear double resonance; CD, circu-
lar dichroism
REFERENCES
The presence of a bound cation appears to be essential for
catalytic function, indicating that PFL-AE is a type I M+ acti-
vated enzyme with K+ being the physiologically most-active
1.
F.; Miller, N. E. Nucleic Acids Res. 2001, 29, 1097-1106.
2. Frey, P. A.; Hegeman, A. D.; Ruzicka, F. J. Crit. Rev.
Biochem. Mol. Biol. 2008, 43, 63-88.
Sofia, H. J.; Chen, G.; Hetzler, B. G.; Reyes-Spindola, J.
ACS Paragon Plus Environment

Journal of the American Chemical Society
Page 12 of 13
1
2
3.
Broderick, J. B.; Duffus, B. R.; Duschene, K. S.; Shepard,
26.
Brush, E. J.; Lipsett, K. A.; Kozarich, J. W. Biochemistry
E. M. Chem. Rev. 2014, 114, 4229-4317.
1988, 27, 2217-2222.
4.
Krebs, C.; Broderick, W. E.; Henshaw, T. F.; Broderick, J.
27.
Külzer, R.; Pils, T.; Kappl, R.; Hüttermann, J.; Knappe, J.
3
4
5
6
B.; Huynh, B. H. J. Am. Chem. Soc. 2002, 124, 912-913.
J. Biol. Chem. 1998, 273, 4897-4903.
28.
29.
Chem. Soc. 1991, 113, 1533-1538.
30.
5.
B.; Hoffman, B. M. J. Am. Chem. Soc. 2002, 124, 11270-11271.
6. Horitani, M.; Shisler, K. A.; Broderick, W. E.; Hutcheson,
Walsby, C. J.; Ortillo, D.; Broderick, W. E.; Broderick, J.
Stoll, S.; Schweiger, A. J. Magn. Res. 2006, 178, 42-55.
Werst, M. M.; Davoust, C. E.; Hoffman, B. M. J. Am.
7
R. U.; Duschene, K. S.; Marts, A. R.; Hoffman, B. M.; Broderick, J.
Davoust, C. E.; Doan, P. E.; Hoffman, B. M. J. Magn.
8
9
B. Science 2016, 352, 822-825.
Reson. 1996, 119, 38-44.
7.
Gänzler, M. Proc. Natl. Acad. Sci U.S.A. 1984, 81, 1332-1335.
8. Wagner, A. F. V.; Frey, M.; Neugebauer, F. A.; Schäfer,
W.; Knappe, J. Proc. Natl. Acad. Sci. USA 1992, 89, 996-1000.
9. Henshaw, T. F.; Cheek, J.; Broderick, J. B. J. Am. Chem.
Soc. 2000, 122, 8331-8332.
Knappe, J.; Neugebauer, F., A.; Blaschkowski, H. P.;
31.
208-214.
32.
Doan, P. E.; Hoffman, B. M. Chem. Phys. Let. 1997, 269,
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
Schweiger, A.; Jeschke, G., Principles of Pulse Electron
Paramagnetic Resonance. Oxford University Press: Oxford, UK,
2001.
33.
Doan, P. E.; Lees, N. S.; Shanmugam, M.; Hoffman, B. M.
10.
Vey, J. L.; Yang, J.; Li, M.; Broderick, W. E.; Broderick, J.
Appl. Magn. Res. 2010, 37, 763-779.
B.; Drennan, C. L. Proc. Natl. Acad. Sci. U.S.A. 2008, 105, 16137-
16141.
34.
Doan, P. E., The past, present, and future of orientation-
selected ENDOR analysis: Solving the challenges of dipolar-coupled
nuclei. In Paramagnetic Resonance of Metallobiomolecules, Telser,
J., Ed. American Chemical Society: Washington, D. C., 2003; pp 55-
81.
11.
Becker, A.; Fritz-Wolf, K.; Kabsch, W.; Knappe, J.;
Schultz, S.; Wagner, A. F. V. Nat. Struct. Biol. 1999, 6, 969-975.
12.
40042.
13.
Becker, A.; Kabsch, W. J. Biol. Chem. 2002, 277, 40036-
35.
Wong, C.; Fujimori, D. G.; Walsh, C. T.; Drennan, C. L. J.
Peng, Y.; Veneziano, S. E.; Gillispie, G. D.; Broderick, J.
Am .Chem. Soc. 2009, 131, 4872-4879.
B. J. Biol. Chem. 2010, 285, 27224-27231.
14. Fu, W.; Telser, J.; Hoffman, B. M.; Smith, E. T.; Adams,
M. W. W.; Finnegan, M. G.; Conover, R. C.; Johnson, M. K. J. Am
.Chem. Soc. 1994, 116, 5722-5729.
36.
Zheng, H.; Chruszcz, M.; Lasota, P.; Lebioda, L.; Minor,
W. J. Inorg. Biochem. 2008, 102, 1765-1776.
37.
Walsby, C. J.; Ortillo, D.; Yang, J.; Nnyepi, M.; Broderick,
W. E.; Hoffman, B. M.; Broderick, J. B. Inorg. Chem. 2005, 44, 727-
741.
15.
Rife, C. L.; Pharris, R. E.; Newcomer, M. A.; Armstrong,
R. N. J. Am .Chem. Soc. 2002, 124, 11001-11003.
38.
2458.
39.
Farrar, C. E.; Jarrett, J. T. Biochemistry 2009, 48, 2448-
16.
7301.
17.
Schwartz, P. A.; Frey, P. A. Biochemistry 2007, 46, 7293-
Berkovitch, F.; Nicolet, Y.; Wan, J. T.; Jarrett, J. T.;
Shibata, N.; Masuda, J.; Tobimatsu, T.; Toraya, T.; Suto,
Drennan, C. L. Science 2004, 303, 76-79.
K.; Morimoto, Y.; Yasuoka, N. Structure 1999, 7, 997-1008.
18. Murshudov, G. N.; Skubák, P.; Lebedev, A. A.; Pannu, N.
S.; Steiner, R. A.; Nicholls, R. A.; Winn, M. D.; Long, F.; Vagin, A.
40.
Horitani, M.; Byer, A. S.; Shisler, K. A.; Chandra, T.;
Broderick, J. B.; Hoffman, B. M. J. Am .Chem. Soc. 2015, 137, 7111-
7121.
A. Acta Crystallogr. D Biol. Crystallogr. 2011, 67, 355-67.
41.
Horitani, M.; Offenbacher, A. R.; Carr, C. A. M.; Yu, T.;
19.
Crystallogr. D Biol. Crystallogr. 2003, 59, 1131-7.
20. Brünger, A. T.; Adams, P. D.; Clore, G. M.; DeLano, W.
Potterton, E.; Briggs, P.; Turkenburg, M.; Dodson, E. Acta
Hoeke, V.; Cutsail, G. E., III; Hammes-Schiffer, S.; Klinman, J. P.;
Hoffman, B. M. J. Am .Chem. Soc. 2017, 139, 1984-1997.
42.
Foner, S. N.; Cochran, E. L.; Bowers, V. A.; Jen, C. K. J.
L.; Gros, P.; Grosse-Kunstleve, R. W.; Jiang, J. S.; Kuszewski, J.;
Nilges, M.; Pannu, N. S.; Read, R. J.; Rice, L. M.; Simonson, T.;
Warren, G. L. Acta. Crystallogr. D Biol. Crystallogr. 1998, 54, 905-
21.
Chem. Phys. 1960, 32, 963-971.
43.
Morton, J. R.; Preston, K. F.; Strach, S. J.; Adrian, F. J.;
Jette, A. N. J. Chem. Phys. 1979, 70, 2889-2893.
44.
Weil, J. A.; Bolton, J. R.; Wertz, J. E., Electron
21.
Broderick, J. B.; Henshaw, T. F.; Cheek, J.; Wojtuszewski,
Paramagnetic Resonance: Elementary Theory and Practical
Applications. John Wiley & Sons: New York, 1994.
K.; Smith, S. R.; Trojan, M. R.; McGhan, R. M.; Kopf, A.; Kibbey,
M.; Broderick, W. E. Biochem. Biophys. Res. Comm. 2000, 269, 451-
456.
45.
Schwartz, P. A.; LoBrutto, R.; Reed, G. H.; Frey, P. A.
Protein Sci. 2007, 16, 1157-1167.
22.
Nnyepi, M. R.; Peng, Y.; Broderick, J. B. Arch. Biochem.
46.
Hutcheson, R. A. U. Mechanistic and spectroscopic
Biophys. 2007, 459, 1-9.
investigations of pyruvate formate-lyase activating enzyme. Montana
State University, 2012.
23.
Walsby, C. J.; Hong, W.; Broderick, W. E.; Cheek, J.;
Ortillo, D.; Broderick, J. B.; Hoffman, B. M. J. Am. Chem. Soc. 2002,
124, 3143-3151.
47.
T.; Shabala, S. Environ. Microbiol. 2009, 11, 137-148.
48. Page, M. J.; Di Cera, E. Physiol. Rev. 2006, 86, 1049-1092.
Shabala, L.; Bowman, J.; Brown, J.; Ross, T.; McMeekin,
24.
Broderick, J. B. J. Am. Chem. Soc. 2000, 122, 12497-12506.
25. Conradt, H.; Hohmann-Berger, M.; Hohmann, H.-P.;
Krebs, C.; Henshaw, T. F.; Cheek, J.; Huynh, B.-H.;
Blaschkowski, H. P.; Knappe, J. Arch. Biochem. Biophys. 1984, 228,
133-142.
ACS Paragon Plus Environment

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