Selenocysteine as a Substrate, an Inhibitor and a Mechanistic Probe for Bacterial and Fungal Iron-Dependent Sulfoxide Synthases

Article abstract of DOI:10.1002/chem.201903898

Sulfoxide synthases are non-heme iron enzymes that participate in the biosynthesis of thiohistidines, such as ergothioneine and ovothiol A. The sulfoxide synthase EgtB from Chloracidobacterium thermophilum (CthEgtB) catalyzes oxidative coupling between the side chains of N-α-trimethyl histidine (TMH) and cysteine (Cys) in a reaction that entails complete reduction of molecular oxygen, carbon–sulfur (C?S) and sulfur–oxygen (S?O) bond formation as well as carbon–hydrogen (C?H) bond cleavage. In this report, we show that CthEgtB and other bacterial sulfoxide synthases cannot efficiently accept selenocysteine (SeCys) as a substrate in place of cysteine. In contrast, the sulfoxide synthase from the filamentous fungus Chaetomium thermophilum (CthEgt1) catalyzes C?S and C?Se bond formation at almost equal efficiency. We discuss evidence suggesting that this functional difference between bacterial and fungal sulfoxide synthases emerges from different modes of oxygen activation.

Full text of DOI:10.1002/chem.201903898

A Journal of
Accepted Article
Title: Selenocysteine as a substrate, an inhibitor and a mechanistic
probe for bacterial and fungal iron-dependent sulfoxide
synthases.
Authors: Kristina V. Goncharenko, Sebastian Flückiger, Cangsong
Liao, David Lim, Anja R. Stampfli, and Florian P. Seebeck
This manuscript has been accepted after peer review and appears as an
Accepted Article online prior to editing, proofing, and formal publication
of the final Version of Record (VoR). This work is currently citable by
using the Digital Object Identifier (DOI) given below. The VoR will be
published online in Early View as soon as possible and may be different
to this Accepted Article as a result of editing. Readers should obtain
the VoR from the journal website shown below when it is published
to ensure accuracy of information. The authors are responsible for the
content of this Accepted Article.
To be cited as: Chem. Eur. J. 10.1002/chem.201903898
Link to VoR: http://dx.doi.org/10.1002/chem.201903898
Supported by

10.1002/chem.201903898
Chemistry - A European Journal
Selenocysteine as a substrate, an inhibitor and a mechanistic probe
for bacterial and fungal iron-dependent sulfoxide synthases.
Kristina V. Goncharenko1, Sebastian Flückiger1, Cangsong Liao1, David Lim1, Anja R. Stampfli1, Flo-
rian P. Seebeck1*
1 Department of Chemistry, University of Basel, Mattenstrasse 24a, Basel 4002, Switzerland
ABSTRACT: Sulfoxide synthases are non-heme iron enzymes that participate in the biosynthesis of thiohistidines such as ergothi-
oneine and ovothiol A. The sulfoxide synthase EgtB from Chloracidobacterium thermophilum (CthEgtB) catalyzes oxidative cou-
pling between the side chains of N-a-trimethyl histidine (TMH) and cysteine (Cys) in a reaction that entails complete reduction of
molecular oxygen, carbon-sulfur (C-S) and sulfur-oxygen (S-O) bond formation and carbon-hydrogen (C-H) bond cleavage. In this
report we show that bacterial sulfoxide synthases cannot efficiently turnover selenocysteine (SeCys) as an alternative substrate
because the sulfur-to-selenium substitution. In contrast, the sulfoxide synthase from the filamentous fungus Chaetomium thermoph-
ilum (CthEgt1) catalyzes C-S and C-Se bond formation at almost equal efficiency. We discuss evidence suggesting that this differ-
ence emerges from different modes of oxygen-activation.
1
This article is protected by copyright. All rights reserved.

10.1002/chem.201903898
Chemistry - A European Journal
INTRODUCTION
Ergothioneine (1, Figure 1) is a ubiquitous bacterial and fungal natural product that may also play an important role in human health.
The unique redox activity of the 2-mercaptoimidazole ring suggest that 1 is likely involved in cellular redox homeostasis.[1] The
quest to identify specific physiological roles of 1 has received an interesting parallel with the discovery of the selenium containing
analogue selenoneine (2, Figure 1). This compound was discovered as the major organoselenide in marine fish.[2] Questions
regarding the physiological activity of 2 are also relevant to human health because it accumulates in individuals with a seafood rich
diet.[3]
One way to study functions of natural products is to examine their biosynthetic origin and the environmental conditions that make
their production advantageous for individual cells or cellular communities. For example, discovery of ergothioneine biosynthetic
genes in mycobacteria enabled the first definitive demonstration that this antioxidant protects Mycobacterium tuberculosis against
oxidative stress and supports drug susceptibility and virulence.[4] Many bacteria, some archaea and most fungi produce 1 in an
oxygen-dependent reaction cascade that attaches a sulfur atom to the imidazole ring of N-a-trimethylhistidine (TMH, Figure 1).[5]
Oxidative coupling of Cys or g-glutamyl cysteine (g-GCys) to TMH is catalyzed by iron-dependent sulfoxide synthases in a complex
reaction involving carbon-sulfur (C-S) bond formation, oxygen atom transfer to sulfur (S-O), carbon-hydrogen (C-H) bond cleavage
and complete reduction of molecular oxygen (O2, Figure 1). The sequence of these elementary steps and the structure of the involved
intermediates are not yet fully understood despite a number of empirical and theoretical investigations.[6]. Less is known about the
biosynthetic origin of 2. Given that most enzymes involved in sulfur metabolism do not discriminate against selenium,[7] it is possible
that 2 emerges from the same chemistry as 1 (Figure 1). This pathway could start with the entry of inorganic selenium into the sulfur
assimilation pathway leading to production of selenocysteine (SeCys) by Cys-biosynthetic enzymes. Support for this idea comes
from the demonstration that Schizosaccharomyces pombe grown in selenium-enriched media produces small but detectable amounts
of 2 in a sulfoxide synthase-dependent manner.[8] The limited accessibility of selenoneine by fermentation or by isolation from
marine fish is a major roadblock for exploring and exploitation of selenoneine as a therapeutic.[3c, 8b, 9] A strategy for chemical
synthesis has been published just recently.[10]
In this report we examined the efficiency of sulfoxide synthase-catalyzed carbon-selenium (C-Se) bond formation. This study
revealed that bacterial sulfoxide synthases are highly inefficient catalysts for this reaction. In contrast, the type of sulfoxide synthases
that only occurs in fungi catalyzes C-Se and C-S bond formation with similar efficiency. These observations suggest that naturally
occurring 2 is either exclusively produced by fungal sulfoxide synthases, or by an entirely different set of enzymes. Kinetic
characterization of the sulfoxide synthase from Chloracidobacterium thermophilum (CthEgtB) revealed that substituting Cys with
2
This article is protected by copyright. All rights reserved.

10.1002/chem.201903898
Chemistry - A European Journal
SeCys reduces the enzymes efficiency of O2 activation and introduces a primary substrate kinetic isotope effect that is obfuscated in
the native reaction. These findings provide novel insight into the catalytic mechanism of sulfoxide synthases.
O
O
NH3
O
S
N
EgtB
HN
NH
S
NH
iron (II)
N
O
NH
N
O
N
O
Cys, O2
HO
O
N
3
1
O
O
O
O
O
NH3
O
NH3
Se
N
EgtB
HN
NH
Se
Se
NH
NH
iron (II)
N
N
O
NH
N
O
N
HO
SeCys, O2
O
O
O
N
N
O
O
5
4
2
O
O
Se
H3N
O3Se
H3N
S
Se
O
O
O
OH
O
OH
H3N
H3N
O
O
O
O
6
7
8
9
O
O
NH3
O
O
O
NH
O
O
H
N
H3N
O
S
O
S
S
N
N
O
N
NH
N
H3N
H
O
O
H3N
O
N
O
O
12
10
11
Figure 1. CthEgtB catalyzes the production of sulfoxide 3 which is converted to ergothioneine (1) by a PLP-dependent b-lyase.[5b, 6f, 11] In presence of SeCys
CthEgtB can catalyze the formation of the selenoether 4.[8a] Compounds 5 – 12 are discussed in the text.
Results and Discussion
SeCys is a poor substrate for CthEgtB. We have previously described the crystal structure of CthEgtB and shown that this type of
bacterial sulfoxide synthase accepts TMH and Cys as substrates.[6f, 12] At neutral pH and at room temperature CthEgtB produces
sulfoxide 3 as the main product (80 %) and cysteine dioxide as a minor product (6, 20 %, Figure 1). Under steady-state conditions
consumption of Cys is characterized by a kcat of 0.14 s-1 and apparent KM values of 27 uM for Cys and 65 uM for TMH.[6f]
Quantification of SeCys consumption by the same assay was complicated by the considerable reactivity of selenols towards O2 and
towards reducing agents. SeCys can react with O2 to form the corresponding diselenide (7), the seleninic acid (8) or the selenonic
acid (9). The latter two can undergo syn b-elimination causing irreversible destruction of SeCys. Oxidation to the diselenide is
reversible in the presence of the reducing agent tris(2-carboxyethyl)phosphine (TCEP). Therefore, diselenide formation does not
contribute to the observed consumption of SeCys. On the other hand, reductive deselenation by TCEP is also a potential avenue for
irreversible destruction of SeCys.[13] Because of these possible side reactions we first needed to establish the rate of SeCys
3
This article is protected by copyright. All rights reserved.

10.1002/chem.201903898
Chemistry - A European Journal
consumption in the absence of CthEgtB. In a reaction containing 0.5 mM SeCys, 0.5 mM TMH, 2 mM TCEP, in a 100 mM phosphate
buffer at pH 8.0 we observed SeCys consumption at an initial rate of 4 ± 1 uM/min (Figure S1). Addition of 1, 3 or 10 uM enzyme
did not increase this rate, suggesting that CthEgtB-catalyzed consumption of SeCys cannot be faster than 2 x 10-3 s-1. We also found
no measurable indication that CthEgtB stimulates oxidation of TCEP or ascorbate (Figure S1), suggesting that enzyme-catalyzed O2
activation is at least 10-fold slower when Cys is substituted with SeCys.
Careful analysis of SeCys-containing reactions by HPLC and high-resolution electrospray ionization mass spectrometry revealed
the selenoether 4 as a genuine enzyme-dependent product (HR ESI-MS: m/z obs. 409.0360, calc. 409.0362, C12H19N4O4SeNa2,
Figure S2). The same compound has been implicated as an intermediate in the biosynthesis of 2 by S. pombe.[8a] In the presence of
0.5 mM SeCys and 0.5 mM TMH CthEgtB produces 4 at a rate of (1.2 ± 0.2) x 10-3 s-1 (Figure S3) which is roughly 200-fold slower
than formation of sulfoxide 3 under the same conditions.[6f] We did not observe selenoxide 5 which is not surprising because
selenoxides are rapidly reduced in the presence of selenols, ascorbate or TCEP.[14]
Figure 2. Lineweaver–Burk plots for catalytic rates recorded at 26°C in the presence of 0.2 mM TMH, variable [Cys] (left); or in the presence of 0.1 mM
Cys and variable [TMH] (right). [SeCys] was kept constant at 0 mM (n), 0.05 mM (l), 0.1 mM (▲), 0.2 mM (▼). The reactions contained 100 mM HEPES
buffer at pH 8.0, 100 mM NaCl, 2 mM ascorbate, 2 mM TCEP, 4 µM FeSO4, and 1.3 µM CthEgtB. The production rates of 3 were measured by HPLC. The
corresponding Michaelis–Menten plots are shown in Figure S4.
SeCys is a Cys-competitive inhibitor. Enzyme-catalyzed C-Se bond formation may be slow because SeCys is a poor ligand for
CthEgtB. To test this idea, we examined the ability of SeCys to inhibit turnover of Cys. First, we recorded rates of enzyme-catalyzed
4
This article is protected by copyright. All rights reserved.

10.1002/chem.201903898
Chemistry - A European Journal
formation of 3 in the presence of 0.2 mM TMH and variable concentrations of Cys and SeCys (Figure 2, Figure S4). The Lineweaver–
Burk plot of the measured rates reveals that increasing the concentration of SeCys increases KM,Cys but does not affect kcat, suggesting
that SeCys acts as a Cys-competitive inhibitor. We calculated an inhibitory constant (Ki) for SeCys of 17 ± 4 µM using the following
relation:
Secondly, we determined the production rates of 3 in the presence of 0.1 mM Cys and variable concentrations of TMH (Figure 2,
Figure S4). The corresponding Lineweaver–Burk plot showed that increasing the concentration of SeCys decreases both kcat and
KM,TMH, revealing SeCys as an uncompetitive inhibitor with respect to TMH, meaning that SeCys does not bind to the enzyme in the
absence of TMH. The two inhibition patterns suggest that the ternary complex (CthEgtB:TMH:Cys) forms through an obligatory
binding sequence with TMH as the leading substrate.[15] This binding sequence is consistent with the crystal structure of CthEgtB in
complex with TMH (PDB: 6QKJ, Figure 3). The binding sites for TMH and iron are located at the bottom of a deep tunnel. A model
of the ternary complex (CthEgtB:TMH:Cys, Figure 3) shows that Cys binds on top of TMH and partially blocks the access to the
TMH binding site (Figure 3).[6a, 6f]
Arg106
Arg103
Cys(modelled)
Phe416
His62
Cys(modelled)
TMH
Cleq
TMH
Phe415
His156
Figure 3. Active Site of CthEgtB. Left: Stick and cartoon representation of the active site of CthEgtB located between the N-terminal domain (blue) and the
C-terminal domain (turquoise). Iron (II) (brown sphere) is coordinated by His62, His 152 and His 156 and the imidazole ring of TMH. The equatorial chloride
ligand (green, Cleq) indicates the position of the presumed O2-binding site. A second Cl in an axial position is not shown. Instead, Cys was modelled to bind
to iron (II) through the axial coordination site.[6f] Active site loop 1 (red, residues 93-99) and active site loop 2 (orange, residues 378-386) are disordered in
the native structure (PDB: 6QKI) but can fold over the active site when TMH is present (PDB: 6QKJ). Right: CthEgtB (gray surface) in complex with TMH
(yellow) and Cys (brown) viewed from the active site entrance. Cys partially blocks access to the TMH binding site.
5
This article is protected by copyright. All rights reserved.

10.1002/chem.201903898
Chemistry - A European Journal
Interestingly, this binding sequence is not apparent in the absence of SeCys. Measuring the catalytic rates at variable [TMH] and
constant [Cys], or constant [TMH] and variable [Cys] produced two Lineweaver–Burk plots with parallel lines (Figure S5). This
pattern is typical for enzymes that catalyze two independent half-reactions in a ping-pong mechanism. However, such at mechanism
is difficult to reconcile with the fact that oxidative coupling requires the simultaneous presence of Cys and TMH. More likely, the
observed pattern emerges from a sequential mechanism where binding of O2 to the CthEgtB:TMH or the CthEgtB:TMH:Cys
complex is irreversible, for example because [O2] (0.25 mM) is saturating.[16] The kinetic behavior alone does not directly reveal as
to whether O2 binds before or after Cys. However, previous work on non-heme iron enzymes established that O2-binding is usually
the last step in the formation of the reactive complex.[17] Indeed, the crystal structure of EgtB from Mycobacterium thermoresistibile
(MthEgtB) showed that both substrates - TMH and g-GCys - can bind prior to O2,[6a] suggesting that sulfoxide synthases follow a
similar substrate-binding mechanism (O2 last) as most non-heme iron enzymes (Scheme 1).
O2
TMH
Cys
E:TMH:Cys:O2
E
E:TMH
SeCys
E:TMH:SeCys
E:TMH:Cys
Scheme 1.
Comparison to cysteine dioxygenase. The observation that the SeCys-bound CthEgtB can hardly produce oxidation products
suggests that this complex cannot efficiently activate O2. A similar observation has been made with cysteine dioxygenase (CDO).
This enzyme too fails to oxidize SeCys.[18] Although CDOs are unrelated to sulfoxide synthases, their active sites are remarkably
similar. Both enzymes contain an iron (II) center coordinated by a three-histidine facial triad.[6a, 6f, 19] Both enzymes coordinate their
substrates directly to the metal via one sulfur and one nitrogen-ligand, reserving a sixth coordination site for O2 binding. The
structural similarity between the two enzyme types is also reflected by similar activity. CthEgtB and other sulfoxide synthases have
significant thiol dioxygenase activity.[6e, 6f, 20] Spectroscopic characterization showed that CDO binds SeCys in the same way as Cys,
but that the CDO:SeCys complex is entirely unreactive with O2. Computational models confirmed that the S-to-Se substitution
dramatically reduces the ability of the iron (II) center to bind O2.[18b] The catalytic activity of CthEgtB towards SeCys might be
affected by the same problem. On the other hand, since this limitation is not fundamental it is conceivable that sulfoxide synthases
with slightly different active sites might indeed catalyze C-Se bond formation more efficiently.
Comparison with other sulfoxide synthases. With this thought in mind we examined the C-Se bond forming activities of other
sulfoxide synthases. Bacterial and eukaryotic homologs with known activities can be classified into four types distinguished by
significant differences in their active site geometries and their substrate preferences.[6f] Briefly, CthEgtB represents type II sulfoxide
6
This article is protected by copyright. All rights reserved.

10.1002/chem.201903898
Chemistry - A European Journal
synthases. Type I EgtBs such as MthEgtB consume g-GCys instead of Cys as the sulfur donor.[6a] Type III enzymes are fungal
sulfoxide synthases that use Cys as the sulfur donor.[21] In our study, type III enzymes are represented by the sulfoxide synthase from
the thermophilic filamentous fungus Chaetomium thermophilum (CthEgt1, Type III). Most type IV sulfoxide synthases, such as
OvoA from Erwinia tasmaniensis (EtaOvoA), produce sulfoxide 11 as a biosynthetic intermediate of ovothiol (12, Figure 1).[22]
However, when presented with TMH as a sulfur acceptor EtaOvoA can also produce the same product as type II and III enzymes
(sulfoxide 3).[23] In fact, cyanobacterial OvoA homologs have adapted to produce sulfoxide 3 which further underscores the
functional and mechanistic similarity between type IV and I – III enzymes.[24] The C-S and C-Se bond forming abilities of these
representative type I – IV sulfoxide synthases were assayed in reactions containing 1 mM TMH and 1 mM Cys or SeCys. The
activity of MthEgtB was assayed using g-GCys or g-glutamyl selenocysteine (g-GSeCys) as the second substrate (Table 1). This
comparison showed that the bacterial enzymes catalyze TMH selenation about two orders of magnitude slower than sulfurization
(Table 2). In contrast, the fungal enzyme CthEgt1 produced selenoether 4 almost as fast as sulfoxide 3. Because of the robust activity
we could determine the Michaelis-Menten parameters for CthEgt1-catalyzed formation of selenoether 4 and sulfoxide 3 (kcat,Cys
=
0.07 ± 0.01; KM,Cys = 38 ± 3 uM; kcat,SeCys = 0.04± 0.01; KM,Cys = 33 ± 6 uM, Figure S6). These parameters show that CthEgt1 accepts
SeCys and Cys with very similar efficiency.
The difference in S vs. Se selectivities between bacterial and fungal enzymes is interesting for two reasons. First, these observations
suggest that either only fungi can produce 2, or that this oranoselenide emerges from an entirely different biosynthetic process. The
similar catalytic parameters associated with CthEgt1-catalyzed production of 3 and 4 suggest that the production rates of 1 and 2 in
vivo may be directly related to the cellular concentrations of Cys and SeCys. In most organisms the concentration of SeCys is kept
very low to avoid oxidative stress and inadvertent incorporation in place of Cys during ribosomal protein synthesis.[25] In fact,
specific ribosomal insertion of SeCys into selenoproteins usually depends on sophisticated co-translational systems that do not
require free SeCys as a substrate.[26] Hence, free SeCys appears as an unlikely building block for specific biosynthesis of
selenometabolites. Since Cys-biosynthetic enzymes cannot efficiently distinguish between sulfur and selenium,[27] high
concentrations of selenium in the environment can increase the concentration of SeCys relative to that of Cys.[7b, 27] Therefor,
production of selenoneine by S. pombe may simply be a unspecific response to high selenium concentrations in the growth
medium.[8] The physiological relevance of this unspecific response is unclear. If certain organisms do produce selenoneine as a
specific metabolite with a specific purpose, they are likely to use a different set of enzymes than for ergothioneine biosynthesis and
a selenium-donor that is less problematic than SeCys.
7
This article is protected by copyright. All rights reserved.

10.1002/chem.201903898
Chemistry - A European Journal
Secondly, our observations identify fungal sulfoxide synthases as interesting targets for protein crystallography. Since CthEgtB and
CthEgt1 both bind SeCys efficiently, the catalytic differences between the two enzymes must lie in the O2 binding pocket.
Apparently, fungal enzymes bind and activate O2 in a way that is largely insensitive to the S-to-Se substitution. Comparison between
the active sites of fungal and bacterial enzymes will be important to identify the structural determinants for these different behaviour.
Unfortunately, the sequence alignment of CthEgt1 and the two crystallized homologs CthEgtB and MthEgtB shows that the protein
segment participating in Cys- and O2-binding is highly variable, precluding the construction of a reliable homology model of fungal
enzymes (Figure S7).[6a, 6f] Hence, a search for structural differences that enable C-Se bond formation can only begin once a crystal
structure of type III sulfoxide synthases becomes available.
Table 1. Comparison between C-S and C-Se bond formation activity of type I – IV sulfoxide synthases.[a]
enzyme
type
kobs,s
[s-1]
0.75
0.2
kobs,se
kobs,S/kobs,Se
DVapp,S /DVapp,Se
[s-1]
MthEgtB
CthEgtB
CthEgt1
EtaOvoA
I
0.012
0.0012
0.04
60
n.d.
II
170
2
0.9 ± 0.1/3.1 ± 0.1
1.0 ± 0.1/1.2 ± 0.1
n.d.
III
IV
0.07
0.03
0.00007
470
[a] Reaction conditions: 100 mM phosphate buffer pH 8.0, 100 mM NaCl, 2 mM ascorbate, 2 mM TCEP, 4 µM FeSO4, 0.5 mM TMH, 0.5 mM gGSeCys or
0.5 mM SeCys, 26°C.
SeCys as a mechanistic probe. CthEgtB and CthEgt1 differ by one more aspect that might provide new insight into the catalytic
mechanism of sulfoxide synthases. The slow production of selenoether 4 by CthEgtB is subject to a large substrate kinetic isotope
effect of 3.1 ± 0.1 (Table 1, Figure S8). This effect is an observed kinetic isotope effect (DVapp,SeCys [28] determined by comparing the
)
production rate of 4 in reactions containing saturating concentrations of SeCys (0.5 mM) and TMH (0.5 mM) or C2-2H-TMH, the
isotopolog of TMH deuterated at C2 of the imidazole ring. An isotope effect of this magnitude is attributable to a primary isotope
effect – related to C-H bond cleavage. Secondary isotope effects and binding isotope effects are usually much smaller.[6e, 29] In
contrast, the Cys-containing reaction is characterized by a substrate isotope effect near unity (DVapp,Cys = 0.9 ± 0.1, Table 1, Figure
S8). Similar experiments with MthEgtB[20b] (gGCys) and EtaOvoA (Cys),[6e, 30] produced kinetic isotope effects near unity, suggesting
that C-H cleavage generally has little effect on the overall rate of the native reactions. Interestingly, CthEgt1-catalyzed consumption
of C2-2H-TMH/TMH is also characterized by an isotope effects near unity with either Cys or SeCys as substrates (Table 1, Figure
S9).
8
This article is protected by copyright. All rights reserved.

10.1002/chem.201903898
Chemistry - A European Journal
R
H
X
R
H
X
R
H
X
COO
HN
O2
FeIII
N
O
FeIII
FeIII
O
H
N
R
O
R
N
N
O
OH2
O
HN
HN
Tyr
X = SeCys
a
b
E:TMH:X:O2
X = Cys or SeCys
X = Cys
a
E:TMH:X:O2 a, b, c
d
3
R
R
X
O
R
X
N
X
FeIII
O
O
N
H
FeIII
NH
O
OOC
N
N
O
R
OH
N
H
R
H
N
H
Tyr
H
3
d
c
Figure 4. Left: Proposed mechanism for sulfoxide synthase-catalyzed oxidative coupling of TMH to Cys or SeCys. Right: Qualitative potential-energy
landscape for this reaction with Cys (black) or SeCys (orange) as substrate. Black arrows indicate suggested changes in the potential-energy landscape
effected by the S to Se substitution.
The large isotope effect for CthEgtB-catalyzed turnover of SeCys is the first probe available to directly examine C-H bond cleavage,
and provides qualitative information about the energy landscape of the entire reaction coordinate. Four potential mechanisms for
this complicated reaction have been proposed. Three mechanisms for the sulfoxide synthase reaction have been proposed based on
computation. These models provide valuable estimations of barrier highs associated with elementary steps.[6b-d] However, as we
have discussed elsewhere,[6f] these models are not consistent with all observables, such as the stereochemistry of the product, and
predict isotope effects that are not observed. Therefor we base the following discussion on a fourth mechanism which has not yet
been examined by computation, but has so far survived all experimental scrutiny (Figure 4). In this mechanism the iron-bound
superoxide in the reactive complex (a) is protonated by Tyr93 (CthEgtB) or Tyr377 (MthEgtB) and reduced by the iron-coordinated
Cys (or gGCys). The resulting species b reacts to c via C-S bond formation, followed immediately by oxidation of the sulfur atom
(S-O bond formation). Proton abstraction from carbon 2 (e) produces the rearomatized product sulfoxide 3. This mechanism can be
segmented into three phases: i) Formation of the reactive complex (E:TMH:Cys:O2), ii) C-S and/or S-O bond formation, and c) C-
H bond cleavage. To examine as to why the S-to-Se substitution increases the observed kinetic isotope effect on C-H cleavage for
CthEgtB but not for CthEgt1 we need to consider the possible effects of this substitution on each of these three reaction phases. The
expression of an intrinsic kinetic isotope effect (DkC-H) on the observed isotope effect (DV) is influenced by three kinetic functions:[31]
i) a function termed forward ratio of catalysis (Rf), comparing the barrier height of the isotope-sensitive step forward with those of
the remaining forward steps; ii) the reverse commitment to catalysis (Cf), describing the extent to which the isotope sensitive step is
equilibrating; and iii) a function that encompasses all equilibria that link formation of the enzyme:substrate complex with the isotope
sensitive step (forward equilibration preceding catalysis, Ef). These functions relate DV and DkC-H as follows:
9
This article is protected by copyright. All rights reserved.

10.1002/chem.201903898
Chemistry - A European Journal
DV = (DkC-H + Rf/ Ef + Cf x Keq)/(1 + Rf/ Ef + Cf)
This comprehensive relation simplifies if the reaction includes irreversible steps: C-H cleavage (d à 3, Figure 4) is coupled to
irreversible aromatization of the imidazole ring.[6b, 6c] Irreversibility of this step reduces Cf to zero. S-O bond formation (c à d) is
also a strongly exothermic reaction, and most likely irreversible. Under these conditions Ef approaches unity.[6c] Hence, the relation
simplifies to
DV = (DkC-H + Rf)/(1 + Rf)
meaning that the expression of the intrinsic isotope effect likely depends on the forward net rate constants across the preceding or
succeeding steps.[32] Perturbations that slow down steps other than the isotope sensitive step would decrease the expression of DkC-H
on the observed rate. As discussed above, the S-to-Se substitution reduces the O2 activation efficiency of CthEgtB by at least 10 and
more likely 100-fold. Making O2 activation more difficult should – if anything – decrease the expression of DkC-H on the maximal
velocity. Since the S-to-Se substitution increases the expression of DkC-H for CthEgtB we conclude that O2 activation is not the only
affected step.
In addition, C-H bond cleavage (d à 3) may become more difficult. Indeed, there is published evidence that protons a to
selenoethers can be less acidic than protons a to thioethers.[33] Less is known about protons a to selenoxides or sulfoxides. As a
model reaction we compared the rates of proton to deuterium exchange (H/D exchange) at the methyl groups of dimethylsulfoxide
(DMSO) and dimethylselenoxide (DMSeO, Figure S10). In reactions containing deuterium oxide as solvent and
tetramethylammonium hydroxide as base incubated at 60°C we observed 4-fold faster H/D exchange in DMSO than in DMSeO.
Hence the S-to-Se substitution can indeed slow down C-H cleavage (d à 3). However, since the rate of O2 activation is reduced by
at least 10, the reduction of the rate for C-H cleavage would have to be much larger in order to increase the observable isotope effect.
On the other hand, a dramatic increase of the intrinsic barrier height associated with C-H cleavage is inconsistent with the observation
that CthEgt1-catalyzed turnover of SeCys and Cys occur at similar rates, and with the same isotope effects near unity.
The most convincing explanation for the observed isotope effects is that the S-to-Se substitution accelerates chemical steps preceding
C-H cleavage. Indeed, organoselenides are usually characterized with a lower redox potential and higher polarizability than their
10
This article is protected by copyright. All rights reserved.

10.1002/chem.201903898
Chemistry - A European Journal
sulfur isologs.[3d, 34] This increased reactivity could lower the reaction barriers for oxidation (b à c), C-Se bond formation (c à d)
and oxygen transfer to Se (d à e). A qualitative energy landscape drawn for the Cys- and SeCys-containing reactions illustrates
how reduction of these barriers could increase the expression of the intrinsic substrate kinetic isotope effect on the observed rate of
the SeCys reaction (Figure 5). Increase of the barrier associated with C-H cleavage (as discussed above) would influence the isotope
effect in the same direction.
In this interpretation, the reactive complex with selenium (b) appears as a more reactive species with regard to the forward reaction
than the same complex with sulfur. Therefore, we suggest that careful characterization of the geometric and electronic structures of
these two species by computational methods could help to identify the precise basis for the increased reactivity, and the precise
nature of the reaction forward. The active site of CthEgt1 appears to mitigate the chemical differences between the sulfur and
selenium versions of b. Characterization of the substrate binding pockets of this enzyme by high-resolution crystallography is
therefore another important objective en route to understanding the catalytic mechanism of sulfoxide synthases.
Conclusions
In this report we described the catalytic activity of bacterial and fungal sulfoxide synthases with selenol instead of thiol substrates.
Bacterial enzymes catalyze oxidative coupling of TMH to SeCys or gGSeCys two orders of magnitude less efficiently than coupling
to Cys or gGCys. Based on these observations we conclude that bacteria are not likely to produce selenoneine (2) from SeCys in a
sulfoxide synthase-dependent reaction. The substrate-binding mechanism, the inhibitory activity of SeCys and the slow turnover of
SeCys observed for CthEgtB provide evidence that the S-to-Se substitution reduces the ability of the enzyme:substrate complex to
bind and/or activate O2. The contrasting behavior of CthEgt1 points to significant differences in the O2-binding mechanism between
fungal and bacterial enzymes. Furthermore, we described a mechanistic interpretation of the observed substrate kinetic isotope
effects suggesting that the S-to-Se substitution affects more than one elementary step along the reaction coordinate. We hope that
these findings inspire additional empirical and computational enquiries into the nature of this complicated enzyme reaction.
ACKNOWLEDGMENT
K.V.G is a recipient of a Swiss Government Fellowship for Excellence; C.L. is a recipient of a SSSTC & CSC postdoctoral
fellowship; F.P.S. is supported by the “Professur für Molekulare Bionik” and by a starting grant from the European Research Council
(ERC-2013- StG 336559).
11
This article is protected by copyright. All rights reserved.

10.1002/chem.201903898
Chemistry - A European Journal
REFERENCES
[1]
aK. D. Asmus, R. V. Bensasson, J. L. Bernier, R. Houssin, E. J. Land, Biochem. J. 1996, 315, 625-629; bC. Stoffels, M.
Oumari, A. Perrou, A. Termath, W. Schlundt, H. G. Schmalz, M. Schäfer, V. Wewer, S. Metzger, E. Schömig, D. Gründemann,
Free Radic Biol Med. 2017, 113, 385 - 394; cB. Halliwell, I. K. Cheah, R. M. Y. Tang, FEBS Lett. 2018, 592, 3357 - 3366; dL.
Servillo, N. DʼOnofrio, M. L. Balestrieri, J Cardiovasc Pharmacol. 2017, 69, 183 - 191.
[2]
[3]
aY. Yamashita, M. Yamashita, J. Biol. Chem. 2010, 285, 18134-18138; bY. Yamashita, M. Yamashita, H. Iida, Nutrients 2013,
5, 388 - 395.
aM. Yamashita, Y. Yamashita, T. Ando, J. Wakamiya, S. Akiba, Biol. Trace. Elem. Res. 2013, 156, 36 - 44; bM. P. Rayman,
Lancet 2012, 379, 1256 - 1268; cJ. Masuda, C. Umemura, M. Yokozawa, K. Yamauchi, T. Seko, M. Yamashita, Y. Yamashita,
Nutrients 2018, 10, E1380; dA. S. Rahmanto, M. J. Davies, IUBMB Life 2012, 64, 863 - 871.
aB. M. Cumming, K. C. Chinta, V. P. Reddy, A. J. C. Steyn, Antioxid. Redox Signal. 2018, 28, 431 - 444; bH. S. Saini, B. M.
Cumming, L. Guidry, D. A. Lamprecht, J. H. Adamson, V. P. Reddy, K. C. Chinta, J. H. Mazorodze, J. N. Glasgow, M.
Richard-Greenblatt, A. Gomez-Velasco, H. Bach, Y. Av-Gay, H. Eoh, K. Rhee, A. J. C. Steyn, Cell Rep. 2016, 14, 572 - 585.
aA. Askari, D. B. Melville, J. Biol. Chem. 1962, 237, 1615-&; bF. P. Seebeck, J. Am. Chem. Soc. 2010, 132, 6632-6633.
aK. V. Goncharenko, A. Vit, W. Blankenfeldt, F. P. Seebeck, Angew. Chem. Int. Ed. Engl. 2015, 54, 2821 - 2824; bW. J. Wei,
P. E. Siegbahn, R. Z. Liao, Inorg. Chem. 2017, 56, 3589 - 3599; cA. S. Faponle, F. P. Seebeck, S. P. de Visser, J. Am. Chem.
Soc. 2017, 139, 9259 - 9270; dG. Tian, H. Su, Y. Liu, ACS Catalysis 2018, 8, 5875 - 5889; eL. Chen, N. Naowarojna, H. Song,
S. Wang, J. Wang, Z. Deng, C. Zhao, L. P., J. Am. Chem. Soc. 2018, 140, 4604 - 4612; fA. Stampfli, K. V. Goncharenko, M.
Meury, B. Dubey, T. Schirmer, F. P. Seebeck, J Am Chem Soc 2019, 141, 5275 - 5285.
[4]
[5]
[6]
[7]
[8]
[9]
aT. C. Stadtman, Annu Rev Biochem. 1990, 59, 111 - 127; bC. Arnaudguilhem, K. Bierla, L. Ouerdane, H. Preud'homme, A.
Yiannikouris, R. Lobinski, Anal Chim Acta. 2012, 757, 26 - 38.
aT. Pluskal, M. Ueno, M. Yanagida, PLoS One 2014, 9, e97774; bN. G. Turrini, N. Kroepfl, K. B. Jensen, T. C. Reiter, K. A.
Francesconi, T. Schwerdtle, W. Kroutil, D. Kuehnelt, Metallomics 2018, 10, 1532 - 1538.
aI. Rohn, N. Kroepfl, M. Aschner, J. Bornhorst, D. Kuehnelt, T. Schwerdtle, J Trace Elem Med Biol. 2019, 55, 78 - 81; bI.
Rohn, N. Kroepfl, J. Bornhorst, D. Kuehnelt, T. Schwerdtle, Mol Nutr Food Res. 2019, 63, e1900080; cR. Alhasan, M. J.
Nasim, C. Jacob, c. Gaucher, Curr Pharmacol Rep 2019, 5, 163 – 173.
[10]
[11]
[12]
D. Lim, D. Gründemann, F. P. Seebeck, Angew Chem Int Ed Engl 2019, in press.
H. Song, W. Hu, N. Naowarojna, A. S. Her, S. Wang, R. Desai, L. Qin, X. Chen, P. Liu, Sci Rep. 2015, 5, 11870.
A. M. Gamage, C. Liao, I. K. Cheah, Y. Chen, D. R. X. Lim, J. W. K. Ku, R. S. L. Chee, M. Gengenbacher, F. P. Seebeck, B.
Halliwell, Y. H. Gan, FASEB, J 2018, 11:fj201800716.
[13]
[14]
S. Dery, P. S. Reddy, L. Dery, R. Mousa, R. N. Dardashti, N. Metanis, Chem. Sci. 2015, 6, 6207 - 6212.
aH. J. Reich, R. J. Hondal, ACS Chem Biol. 2016, 11, 821 841; bV. De Silva, M. M. Woznichak, K. L. Burns, K. B. Grant, S.
W. May, J. Am. Chem. Soc. 2004, 126, 2409 - 2413.
[15]
aH. J. Fromm, V. Zewe, J. Biol. Chem. 1962, 237, 3027 - 3032; bD. L. Purich, Enzyme Kinetics, Academic Press, London,
2010.
[16]
[17]
[18]
H. J. Fromm, Biochim. Biophys. Acta 1967, 39, 221 - 230.
E. I. Solomon, S. Goudarzi, K. D. Sutherlin, Biochemistry 2016, 55, 6363 - 6374.
aJ. D. Gardner, B. S. Pierce, B. G. Fox, T. C. Brunold, Biochemistry 2010, 49, 6033-6041; bE. Blaesi, J. D. Gardner, B. G. Fox,
T. C. Brunold, Biochemistry 2013, 52, 6040 - 6051.
[19]
[20]
C. A. Joseph, M. J. Maroney, Chem. Commun. 2007, 3338-3349.
aH. Song, A. S. Her, F. Raso, Z. Zheng, Y. Huo, P. Liu, Org. Lett. 2014, 16, 2122-2125; bK. V. Goncharenko, F. P. Seebeck,
Chem. Commun. 2016, 52, 1945 - 1948.
[21]
aK. J. Sheridan, B. E. Lechner, G. O. Keeffe, M. A. Keller, E. R. Werner, H. Lindner, G. W. Jones, H. Haas, S. Doyle, Sci Rep.
12
This article is protected by copyright. All rights reserved.

10.1002/chem.201903898
Chemistry - A European Journal
2016, 6, 35306; bW. Hu, H. Song, A. Sae Her, D. W. Bak, N. Naowarojna, S. J. Elliot, L. Qin, X. Chen, P. Liu, Org. Lett. 2014,
16, 5382 - 5385; cM. H. Bello, V. Barrera-Perez, D. Morin, L. Epstein, Fungal Genet. Biiol 2012, 49, 160-172.
aA. Braunshausen, F. P. Seebeck, J. Am. Chem. Soc. 2011, 133, 1757-1759; bI. Castellano, F. P. Seebeck, Nat. Prod. Rep.
2018, 35, 1241 - 1250; cN. Naowarojna, R. Cheng, L. Chen, M. Quill, M. Xu, C. Zhao, P. Liu, Biochemistry 2018, 57, 3309 -
3325.
[22]
[23]
[24]
[25]
H. Song, M. Leninger, N. Lee, P. Liu, Org. Lett. 2013, 15, 4854 - 4857.
C. Liao, F. P. Seebeck, ChemBioChem 2017, 18, 2115 - 2118.
aJ. Lu, A. Holmgren, J. Biol. Chem. 2009, 284, 723 - 727; bM. Lazard, M. Dauplais, S. Blanquet, P. Plateau, Biomol Concepts
2017, 8, 93 - 104; cS. Wang, A. T. Al-Soodani, G. C. Thomas, B. A. Buck-Koehntop, K. J. Woycechowsky, Bioconjug. Chem.
2018, 29, 2332 - 2342.
[26]
[27]
[28]
[29]
V. M. Labunskyy, D. L. Hatfield, V. N. Gladyshev, Physiol Rev. 2014, 94, 739 - 777.
M. Birringer, S. Pilawa, L. Flohé, Nat Prod Rep. 2002, 19, 693 - 718.
D. B. Northrop, Annu Rev Biochem. 1981, 50, 103 - 131.
aR. A. Caldwell, H. Misawa, J. Am. Chem. Soc. 1987, 109, 6869 - 6870; bK. W. Rickert, J. P. Klinman, Biochemistry 1999, 38,
12218 - 12228.
[30]
[31]
[32]
[33]
[34]
G. T. Mashabela, F. P. Seebeck, Chem. Commun. 2013, 49, 7714 - 7716.
D. B. Northrop, Biochemistry 1981, 20, 4056 - 4061.
W. W. Cleland, Biochemistry 1975, 14, 3220 - 3224.
H. J. REICH, W. W. WILLIS, J. Org. Chem. 1980, 45, 5227 - 5229.
aR. J. Hondal, S. M. Marino, V. N. Gladyshev, Antiox. Redox Signal. 2013, 18, 1675 - 1689; bC. Jacob, G. I. Giles, N. M.
Giles, H. Sies, Angew Chem Int Ed Engl 2003, 42, 4742 - 4758.
13
This article is protected by copyright. All rights reserved.