Structural and Mechanistic Basis of an Oxepin-CoA Forming Isomerase in Bacterial Primary and Secondary Metabolism

Article abstract of DOI:10.1021/acschembio.9b00742

Numerous aromatic compounds are aerobically degraded in bacteria via the central intermediate phenylacetic acid (paa). In one of the key steps of this widespread catabolic pathway, 1,2-epoxyphenylacetyl-CoA is converted by PaaG into the heterocyclic oxepin-CoA. PaaG thereby elegantly generates an α,β-unsaturated CoA ester that is predisposed to undergo β-oxidation subsequent to hydrolytic ring-cleavage. Moreover, oxepin-CoA serves as a precursor for secondary metabolites (e.g., tropodithietic acid) that act as antibiotics and quorum-sensing signals. Here we verify that PaaG adopts a second role in aromatic catabolism by converting cis-3,4-didehydroadipoyl-CoA into trans-2,3-didehydroadipoyl-CoA and corroborate a δ³,δ²-enoyl-CoA isomerase-like proton shuttling mechanism for both distinct substrates. Biochemical and structural investigations of PaaG reveal active site adaptations to the structurally different substrates and provide detailed insight into catalysis and control of stereospecificity. This work elucidates the mechanism of action of unusual isomerase PaaG and sheds new light on the ubiquitous enoyl-CoA isomerases of the crotonase superfamily.

Full text of DOI:10.1021/acschembio.9b00742

Subscriber access provided by The University of Melbourne Libraries
Article
Structural and Mechanistic Basis of an Oxepin-CoA Forming
Isomerase in Bacterial Primary and Secondary Metabolism
Melanie Spieker, Raspudin Saleem-Batcha, and Robin Teufel
ACS Chem. Biol., Just Accepted Manuscript • DOI: 10.1021/acschembio.9b00742 • Publication Date (Web): 05 Nov 2019
Downloaded from pubs.acs.org on November 10, 2019
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 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 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.
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 21
ACS Chemical Biology
1
2
3
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
Figure 1. (A) Bacterial phenylacetic acid (1) catabolic pathway (2), see text for details. PaaG catalyzes two
isomerization steps of structurally-distinct CoA-esters 2/3 and 9, which affords 4 and 10, respectively
(purple box). In addition, non-native substrate 12 (in grey) is converted to 10 by PaaG. If labile 5 is not
immediately oxidized to 6, spontaneous condensation (indicated by dashed arrows) generates 7 that is likely
the sought precursor for primary and secondary metabolites (dashed box). Ultimately, 1 is degraded to two
acetyl-CoA and one succinyl-CoA. The red and blue colored oxygen atoms derive from O2 and H2O,
respectively. (B) Proposed PaaG catalytic mechanism for both native substrates 3 (tautomer 2 not shown)
and 9 involving the catalytic aspartate side chain and thioester enolate transition states stabilized by
hydrogen bonding (residue numbering from T. thermophilus). The shuttled proton is highlighted in green.
288x131mm (300 x 300 DPI)
ACS Paragon Plus Environment

ACS Chemical Biology
Page 2 of 21
1
2
3
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
Figure 2. UPLC-HRMS data of enzyme reactions with 6 as substrate. Traces with the same color belong to
the same assay (respective enzyme compositions shown to the right). Shown are the extracted ion
chromatograms (EICs) for [M+H]+ of 936.165 (compound 6), [M+H]+ of 894.155 (compounds 9 and 10)
and [M+H]+ of 912.165 (compound 11), respectively. Only in presence of PaaGPY2, 10 is formed in
significant amounts, thus verifying the Δ3,Δ2-enoyl-CoA isomerase activity of PaaG. Trace amounts of 11
formed in the absence of PaaG can be explained by the known low isomerase side activity of enoyl-CoA
hydratases (PaaF). Note that 9 and 10 have identical masses but slightly different retention times and can
be further distinguished by the characteristic UV-Vis spectra (Supplementary Figure 2). If not further
converted, unstable 9 spontaneously decomposes.
176x54mm (300 x 300 DPI)
ACS Paragon Plus Environment

Page 3 of 21
ACS Chemical Biology
1
2
3
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
Figure 3. Polder OMIT maps (contoured at 3σ above the mean) of (A) PaaGTT-D136N•4 (PDB ID: 6SLA), (B)
PaaGTT•10/12 (PDB ID: 6SLB), and (C) apo-PaaGTT (PDB ID: 6SL9). The unbiased maps allowed placement
of the ligands within the electron density. For unsoaked apo-PaaGTT, water molecules and spurious density
could be observed. Note that the distance and orientation of residue 136 to the C2 and C4 atoms of the
ligand in 6SLA differs due to replacement of the catalytic aspartate residue with asparagine, which adopts a
different conformation.
292x66mm (300 x 300 DPI)
ACS Paragon Plus Environment

ACS Chemical Biology
Page 4 of 21
1
2
3
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
Figure 4. Overall structure and active site of PaaGTT. (A) Superposition of apo-PaaGTT (grey), PaaGTT-
D136N•4 chain A (green), PaaGTT•10/12 (blue). (B, C, D) Inlets showing the respective enlarged active
sites. Catalytic residues and ligands are represented as stick models. Distances, e.g., between the catalytic
Asp136 and the C2 and C4 atoms of the ligands are indicated in Å. The mobile Tyr80 closes the active site
upon substrate binding and expels water. To accommodate 4, loop β3-α3 opens up and thereby provides
space for the bulky ring.
162x170mm (300 x 300 DPI)
ACS Paragon Plus Environment

Page 5 of 21
ACS Chemical Biology
1
2
3
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
Figure 5. (A) Possible conformations of PaaG’s native substrate 9 in enzyme free environment (top). The
thioester enolates, following deprotonation, represent the transition states (middle), which are converted to
the products (bottom). If the transition state adopts a bent conformation (red boxes), the undesired cis-2,3
product would be formed rather than the observed trans-2,3 (blue boxes). R = CH2COOH. (B, C) PaaGTT
suppresses the bent conformation of the transition state through a confined active site, thereby steering the
exclusive formation of trans-2,3-didehydroadipoyl-CoA (10). Modeled bent cis-1,2, cis-3,4 transition state
binding-modes are shown as yellow sticks (B) and modeled bent trans-1,2, cis-3,4 transition state binding-
modes as green sticks (C). Both transition state models are aligned along the CoA-moieties and C1-C2 axes
with the experimentally observed 10/12 (pink sticks). Red spheres represent steric clashes of transition
states with the protein. The C2 and C4 atoms are highlighted as spheres with indicated distances to the
catalytic D136 side chain.
135x154mm (300 x 300 DPI)
ACS Paragon Plus Environment

ACS Chemical Biology
Page 6 of 21
1
2
3
4
5
6
Structural and Mechanistic Basis of an Oxepin-CoA Forming Isomerase in
Bacterial Primary and Secondary Metabolism
7
8
9
Authors: Melanie Spieker†,‡,§, Raspudin Saleem-Batcha†,‡,§, & Robin Teufel†,‡,*
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
Affiliations:
†ZBSA, Center for Biological Systems Analysis, University of Freiburg, 79104 Freiburg,
Germany.
‡Faculty of Biology, University of Freiburg, 79104 Freiburg, Germany.
§These authors contributed equally
*Correspondence: robin.teufel@zbsa.uni-freiburg.de
1
Abstract
2
3
4
5
6
7
8
9
10
11
12
13
14
15
Numerous aromatic compounds are aerobically degraded in bacteria via the central
intermediate phenylacetic acid (paa). In one of the key steps of this widespread catabolic
pathway, 1,2-epoxyphenylacetyl-CoA is converted by PaaG into the heterocyclic oxepin-
CoA. PaaG thereby elegantly generates an α,β-unsaturated CoA ester that is predisposed
to undergo β-oxidation subsequent to hydrolytic ring-cleavage. Moreover, oxepin-CoA
serves as precursor for secondary metabolites (e.g., tropodithietic acid) that act as
antibiotics and quorum sensing signals. Here we verify that PaaG adopts a second role in
aromatic catabolism by converting cis-3,4-didehydroadipoyl-CoA into trans-2,3-
didehydroadipoyl-CoA and corroborate a Δ3,Δ2-enoyl-CoA isomerases-like proton
shuttling mechanism for both distinct substrates. Biochemical and structural
investigations of PaaG reveal active site adaptations to the structurally different
substrates and provide detailed insight into catalysis and control of stereospecificity. This
work elucidates the mechanism of action of unusual isomerase PaaG and sheds new light
on the ubiquitous enoyl-CoA isomerases of the crotonase superfamily.
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
Aromatic compounds are highly abundant in nature and serve as important carbon and
energy sources for microorganisms. Commonly, peripheral pathways convert structurally
diverse aromatics into a small number of central intermediates such as benzoic acid or
phenylacetic acid (paa, compound 1, Figure 1A) before further degradation via central pathways
takes place (1). Depending on the availability of dioxygen (O2), different strategies are
employed to overcome the high resonance energy of the aromatic ring (1). Under aerobic
conditions, ring activation through hydroxylation is followed by oxygenolytic ring-cleavage. In
contrast, in the absence of O2, aromatics such as benzoic acid are activated by CoA thioester
formation and subsequently reduced, before hydrolytic ring-cleavage yields open-chain CoA-
thioesters that are degraded via β-oxidation-like steps (1). In addition, “hybrid strategies”
(likely preferably employed under microaerobic conditions) combine features of both
paradigms, i.e. CoA-thioester formation, ring activation through oxygenation (i.e. epoxidation),
hydrolytic ring cleavage, and β-oxidation-like steps (2–8,1,9). For instance, numerous
phylogenetically diverse bacteria (e.g., Escherichia coli K-12 or Thermus thermophilus) and
some archaea employ a hybrid pathway for degradation of 1 (2,10,11,6,3,7), which is also
relevant for virulence of Burkholderia cenocepacia in cystic fibrosis (12) (Figure 1A).
In the phenylacetic acid degradation pathway (2), 1 is first ligated with coenzyme A by
PaaK to phenylacetyl-CoA (13,2) and then epoxidation by the diiron-dependent
multicomponent monooxygenase PaaABCE results in 1,2-epoxyphenylacetyl-CoA (2)
(10,2,14). Isomerase PaaG next produces an α,β-unsaturated CoA-thioester motif by
isomerizing 2 (or its spontaneously formed tautomer 2-(oxepin-yl)acetyl-CoA (3)) into the
1
ACS Paragon Plus Environment

Page 7 of 21
ACS Chemical Biology
1
2
3
4
5
6
7
8
9
37
38
39
40
unusual enolether (Z)-2-(oxepin-2(3H)-ylidene)acetyl-CoA (“oxepin-CoA”, 4), which is prone
to undergo hydrolytic ring cleavage catalyzed by the hydratase domain of the bifunctional
fusion protein PaaZ (2,11). The resulting open-chain aldehyde (5) is further oxidized to the
stable 3-oxo-5,6-didehydrosuberoyl-CoA (6) by the NAD(P)+-dependent aldehyde
dehydrogenase domain of PaaZ. If not directly oxidized, highly reactive 5 spontaneously
undergoes rapid intramolecular aldol condensation followed by dehydration (Knoevenagel
condensation), thereby affording 2-hydroxycyclohepta-1,4,6-triene-1-carboxyl-CoA (7)
featuring a seven-membered carbon cycle (11). Compound 7 represents a shunt product of the
catabolic pathway, but most likely serves as precursor for unusual ω-cycloheptyl fatty acids
(15) as well as tropone-based secondary metabolites including, e.g., tropoditheitic acid (16–25)
(8) or various roseobacticides (Figure 1A) (26–28). In contrast, further catabolism of 6 proceeds
via β-oxidation-like steps catalyzed by β-ketothiolase (PaaJ) (2) that generates acetyl-CoA and
presumably cis-3,4-didehydroadiypl-CoA (9), which is further isomerized to trans-2,3-
didehydroadiypl-CoA (10). Final β-oxidation-like steps are catalyzed by enoyl-CoA hydratase
(PaaF), alcohol dehydrogenase (PaaH) (2), and once more PaaJ (2,29), in total producing two
acetyl-CoA and one succinyl-CoA (Figure 1A) (2).
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
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
PaaG and PaaF both belong to the enoyl-CoA hydratase/isomerase (“crotonase”)
superfamily (2,30,31) that comprise well-investigated enzymes (30,32–37). As a universal
feature during catalysis, a negative charge develops in the transition state in the form of a
thioester enolate, which is stabilized by a dedicated pocket (i.e. the “oxyanion hole”) in the
active site via hydrogen-bonding interactions. Hence all enzymes of this superfamily strictly
depend on CoA-thioester substrates, while the attached acyl groups may differ (30). Enoyl-CoA
hydratases employ two conserved acidic amino acid residues for the activation and
incorporation of water at the β–position (32), whereas isomerases apparently utilize only a
single acidic side-chain (Glu or Asp) as proton shuttle (34–36). In the case of PaaG, a similar
mechanism can be envisaged and a second PaaG-mediated isomerization reaction in the
pathway, i.e. the conversion of 9 to 10, has been proposed before without experimental
validation (2). Previously, PaaG from Pseudomonas sp. Y2 (henceforth referred to as PaaGPY2
)
was shown to catalyze the isomerization of 2/3 to 4 by in vitro assays, mass spectrometry
including 18O-labelling, as well as nuclear magnetic resonance (NMR) studies with 13C-labelled
precursors (2). In addition, crystal structures of apo-PaaG from Thermus thermophilus (PaaGTT)
(38) and apo-PaaG in complex with apo-PaaF from E. coli K-12 (31) were reported previously.
In this work, we characterize PaaG in detail and provide evidence for a Δ3,Δ2-enoyl-CoA
isomerase-like activity by combining mechanistic and structural studies with native substrates,
thereby confirming PaaG’s unusual dual function, while providing general insight into
structural requisites for isomerase catalysis.
72
73
74
75
76
77
78
79
80
81
82
Figure 1. (A) Bacterial phenylacetic acid (1) catabolic pathway (2), see text for details. PaaG catalyzes two
isomerization steps of structurally-distinct CoA-esters 2/3 and 9, which affords 4 and 10, respectively (purple box).
In addition, non-native substrate 12 (in grey) is converted to 10 by PaaG. If labile 5 is not immediately oxidized
to 6, spontaneous condensation (indicated by dashed arrows) generates 7 that is likely the sought precursor for
primary and secondary metabolites (dashed box). Ultimately, 1 is degraded to two acetyl-CoA and one succinyl-
CoA. The red and blue colored oxygen atoms derive from O2 and H2O, respectively. (B) Proposed PaaG catalytic
mechanism for both native substrates 3 (tautomer 2 not shown) and 9 involving the catalytic aspartate side chain
and thioester enolate transition states stabilized by hydrogen bonding (residue numbering from T. thermophilus).
The shuttled proton is highlighted in green.
83
84
Results and Discussion
PaaG Catalyzes Two Distinct Isomerization Steps in Phenylacetatic Acid Catabolism
The unusual conversion of 2/3 to 4 catalyzed by PaaG can be envisaged by a Δ3,Δ2-
enoyl-CoA isomerase-like mechanism, consistent with amino acid sequence-based functional
85
86
2
ACS Paragon Plus Environment

ACS Chemical Biology
Page 8 of 21
1
2
3
4
5
6
7
8
9
87
88
89
90
91
92
93
94
95
96
97
98
99
predictions (Figure 1B). Typically, Δ3,Δ2-enoyl-CoA isomerases catalyze the conversion of
enoyl-CoA esters with cis-3,4 or trans-3,4 configuration into trans-2,3-enoyl-CoAs.
Previously, we speculated that PaaG may thus also be involved in a further downstream step by
isomerizing 9 to 10, albeit other candidates (e.g., PaaJ) could not be ruled out (2). To further
investigate this, we first sought to acquire the proposed PaaG substrate, i.e. 3,4-
didehydroadiypl-CoA (9) that is produced via PaaJ-mediated thiolytic cleavage of precursor 4
and most likely retains the cis-configuration of the double bond (Figure 1A). As cis-3,4-
didehydroadipate cannot be purchased commercially, we first attempted to produce 9
enzymatically. However, 9 proved to be highly unstable during purification, most likely due to
CoA-ester elimination via spontaneous intramolecular acid anhydride formation and could not
be isolated in sufficient amounts for enzymatic assays. Instead, we enzymatically generated and
purified the adequately stable PaaZ product 6, which was then used as substrate for in vitro
coupled enzyme assays to investigate the postulated Δ3,Δ2-isomerization step that affords 10.
The assays contained 6, HS-CoA, as well as different combinations of PaaJ, PaaG and enoyl-
CoA hydratase PaaF (that converts 10 into 3-hydroxyadipoyl-CoA (11)) (Figure 2). Indeed,
these experiments confirmed that PaaGPY2 catalyzes the formation of 10, as verified by ultra
performance liquid chromatography high-resolution mass spectrometry (UPLC-HRMS)
(Figure 2).
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
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
Figure 2. UPLC-HRMS data of enzyme reactions with 6 as substrate. Traces with the same color belong to the
same assay (respective enzyme compositions shown to the right). Shown are the extracted ion chromatograms
(EICs) for [M+H]+ of 936.165 (compound 6), [M+H]+ of 894.155 (compounds 9 and 10) and [M+H]+ of 912.165
(compound 11), respectively. Only in presence of PaaGPY2, 10 is formed in significant amounts, thus verifying the
Δ3,Δ2-enoyl-CoA isomerase activity of PaaG. Trace amounts of 11 formed in the absence of PaaG can be explained
by the known low isomerase side activity of enoyl-CoA hydratases (PaaF). Note that 9 and 10 have identical
masses but slightly different retention times and can be further distinguished by the characteristic UV-Vis spectra
(Supplementary Figure 2). If not further converted, unstable 9 spontaneously decomposes.
In addition, 6 was used as substrate for coupled enzyme assays to estimate the specific
activity by UPLC-HRMS for the conversion of 9 into 10 for both PaaGPY2 and PaaGTT (Table
1). Because of the instability of the native substrate 9, we chemically synthesized and purified
isomer trans-3,4-didehydroadipoyl-CoA (12) (for details, see experimental section) for detailed
structural and further mechanistic studies. As expected, 12 proved to be significantly more
stable and was converted to the same product 10, as verified by UPLC-HRMS (Supplementary
Figure 1). The kinetic parameters for the PaaGPY2/PaaGTT-mediated conversion of 12 into 10
were then determined spectrophotometrically at 30 °C (Supplementary Figure 2 and Table 1).
In addition, we determined the specific activities for the second native substrate 2/3. For that,
HPLC assays were employed due to the instability of 2 and because spectrophotometric assays
were impeded by NADPH, [2Fe-2S]-clusters, and FAD (required by or part of PaaABCE)
(Supplementary Figure 3). Overall, PaaGTT appeared to be more adapted to the conversion of
the bulky ring-substrate 2/3, whereas PaaGPY2 readily converted both linear substrates (9 & 12)
as well as 2/3. Interestingly, chemically synthesized 2-cyclohexenylacetyl-CoA that is
structurally similar to 2/3 was not accepted by either enzyme, as determined
spectrophotometrically, by HPLC, and by X-ray crystallography (see below). The enzyme
kinetics are summarized in Table 1.
A Catalytic Aspartate Side Chain Acts as Proton Relay in PaaG
134
135
136
137
138
Δ3,Δ2-enoyl-CoA isomerases are proposed to shuttle a proton between C4 and C2 of the
respective enoyl-CoA substrates in a reversible reaction (Figure 1B). In PaaG, an aspartate side
chain (D136) is found in the position of the proposed catalytic residue (2,31,38). To further
examine the role of D136, we constructed a PaaGTT-D136N variant, which was crystallized
along with the wild type (wt). The overall structure of PaaGTT-D136N was highly similar to
3
ACS Paragon Plus Environment

Page 9 of 21
ACS Chemical Biology
1
2
3
4
5
6
7
8
9
139
140
141
142
143
144
145
146
PaaGTT wt and showed no significant changes, as further confirmed using circular dichroism
(CD) spectroscopy (that was also conducted for other enzyme variants discussed below)
(Supplementary Figure 4). As expected, PaaGTT-D136N functionality was drastically affected,
allowing the enzyme to operate at < 0.1 % relative activity for the conversion of 2 to 4 compared
to PaaGTT wt (Table 1), while activity for 12 was completely abolished, thus strongly supporting
a catalytic role for D136. This critical function of D136 was further corroborated by the
complex structures of PaaGTT wt and PaaGTT-D136N with native ligands (see crystallography
section below).
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
Table 1. Overview of PaaG kinetic parameters. Parameters for PaaGTT wt (and variants) and PaaGPY2 were determined
for 1,2-epoxyphenylacetyl-CoA (2), trans-3,4-didehydroadipoyl-CoA (12) and 2-cyclohexenylacetyl-CoA. Different
methods for measurement were applied: 1: measured in coupled assay by HPLC; 2: measured with spectrophotometer; 3:
measured by UPLC-MS. Abbreviations: n.d.: (activity) not detected; -: not measured.
Kinetics of PaaGTT and PaaGPY2
1,2-
epoxyphenyl-
cis-3,4-
didehydro-
adipoyl-
2-cyclohexenyl -
acetyl-CoA 1 & 2
trans-3,4-didehydroadipoyl-CoA (12) 2
acetyl-CoA
CoA (9) 3
(2) 1
specific
activity
specific
activity
specific
activity
Catalytic
efficiency
Enzyme
form
Km
kcat
specific activity
in U mg-1
in µM
in s-1
in U mg-1
in U mg-1
in U mg-1
in s-1 µM-1
138 ± 11
3 ± 1
(7.6 ± 0.9) 10-3
564 ± 149
123 ± 23
n.d.
(3.5 ± 0.4) 10-3
(4.5 ± 0.3) 10-4
n.d.
(6 ± 3) 10-6
n.d.
wild type
99 ± 6
-
(9.1 ± 0.6) 10-4
n.d.
(4 ± 1) 10-6
n.d.
-
-
variant Y80F
PaaGTT
variant
D136N
0.05 ± 0.01
0.15 ± 0.01
182 ± 10
-
variant
Y80F/D136N
-
n.d.
n.d.
n.d.
n.d.
-
PaaGPY2
105 ± 16
116 ± 4
142 ± 13
(60 ± 0.1) 10-3
(4 ± 2) 10-4
n.d.
wild type
147
148
Structural Elucidation of PaaG in Complex with Native Ligands
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
To gain insight into the substrate binding mode and catalytic mechanism, we next
attempted to crystallize PaaGTT with the native ligands. After modification of previously
reported conditions, diffracting polygonal shaped crystals of the PaaGTT wildtype and PaaGTT-
D136N were obtained. Co-crystallization with native ligands, however, did not meet with
success. Soaking of PaaGTT crystals with enzymatically prepared 9 also failed due to its
instability. Hence, we focused our efforts on crystal soaking with more stable 4 and 12 (that
can be converted in the PaaG active site to the on-pathway product 10). In addition, 2-
cyclohexenylacetyl-CoA was used that structurally resembles 4 (Table 1). For soaking
experiments with 4, the PaaGTT-D136N variant was employed because of the reversibility of
isomerization of 2 to 4 and the innate instability of 2. Compared to the previously published
apo-PaaGTT (form I; PDB ID: 3HRX (38)), we obtained two distinct crystal forms II and III of
PaaGTT. Molecular replacement provided phasing in the cubical P213 space group with one
chain in the asymmetric unit (form II for apo-PaaGTT and complex structures of PaaGTT·10/12)
and a monoclinic P21 space group with six chains in the asymmetric unit (form III for PaaGTT-
D136N·4) (Table 2). According to PISA (40) prediction, both the apo and complex structures
(forms II and III) form a stable homotrimeric biological assembly, consistent with previous
reports (38). Data statistics and refinements are shown in Supplementary Table 1.
4
ACS Paragon Plus Environment

ACS Chemical Biology
Page 10 of 21
1
2
3
4
5
6
7
8
9
166
167
Table 2. Overview of obtained crystal structures of PaaGTT (homotrimeric in its biologically active form)
and local displacements.
Structures
PDB
codes
Crystal
Forms
Space
group
Resolution Ligand Monomers
Oligomeric
state
loop
β3-α3
Tyr80
(IN or
OUT)
Distance:
Tyr80-
Asp136/Asn136
(in Å)
(Å)
in asu
Apo-PaaGTT
(this study)
PaaGTT·10/12
(this study)
6SL9
6SLB
6SLA
II
II
P213
P213
P21
1.27
no
10/12
4
1
homotrimer
homotrimer
homotrimer
Ο
Ο
OUT
IN
13.1
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
1.88
2.55
1
6
5.6
PaaGTT
-
III
A: Ø
B: Ø
C: Ø
D: Ø
E: Ø
F: Ø
A: Ο
B: Ο
C: Ο
D: Ο
E: Ο
F: Ø
A: IN
B: IN
5.9
5.8
D136N·4
(this study)
(chains A-F)
C: IN
5.6
D: IN
5.5
E: IN
6.4
F: IN
5.7
*Apo-PaaGTT
3HRX
I
P212121
1.85
no
6
homotrimer
A: IN
6.4
(chains A-F)
B: OUT
C: OUT
D: OUT
E: OUT
F: IN
12.9
13.0
13.1
13.1
6.1
168
169
(Abbreviations/symbols: asu – asymmetric unit; O – ordered, Ø – disordered; IN – Tyr80 orienting towards residue 136, OUT – Tyr80
orienting away from residue 136). *Data taken from reference (38)
170
171
172
173
174
175
176
177
178
179
180
181
182
183
184
185
186
187
188
189
190
191
192
193
194
195
196
197
198
Extensive rebuilding was required for portions of the structure encompassing loop β3-
α3 and adjacent structural elements, where the apo-PaaGTT structure differed from the phasing
model. PaaG monomers adopted the canonical crotonase fold of the superfamily with two β-
sheets enclosed by six α-helices, as well as a C-terminal helical domain, as previously reported
(38,31). In the protein·substrate complex structures, well-defined electron densities were
observed indicative of the binding of each CoA-ester substrate; these electron densities were
absent in apo-PaaGTT (Figure 3). However, initially-calculated Fo-Fc omit maps for PaaGTT-
D136N·4 displayed poor electron density as a result of the conformational heterogeneity at the
terminal moieties of the ligands. To improve the model of the bound ligands, we calculated
both 2Fo-Fc maps and POLDER omit maps (41) and fitted the ligands into the respective
electron densities, which allowed unambiguous identification of substrate positioning. A
superimposition of apo-PaaGTT with PaaGTT·10/12 and PaaGTT D136N·4 provided root-mean-
square deviation values for Cα atoms of 0.42 Å and 0.60 Å, respectively (Supplementary Figure
5). Except for structural elements of the protein near the bound substrates (see below), the
overall fold of the all the conformers were virtually identical to apo-PaaGTT. The protein
substrate complexes furthermore indicated subtle differences of active site residues upon
binding of ligands 4 or 10/12. Notably, the previously unknown C2-C3 double configuration of
4 could be inferred as (Z) from our complex structure. Ligand interactions are shown in
Supplementary Tables 2 & 3. Interestingly, soaking with stable 2-cyclohexenylacetyl-CoA did
not result in significant electron density within the ligand binding site of PaaGTT.
The CoA moieties of both ligands in our complex structures adopt a typical U-form as
observed in other Δ3,Δ2-enoyl-CoA isomerases with the respective side chains embedded in a
hydrophobic site. In contrast to typical Δ3,Δ2-enoyl-CoA isomerase substrates such as 9, the
conversion of 2 is exceptional but can be rationalized by considering the known spontaneous
and reversible electrocyclic rearrangement of 1,2-epoxybenzene moieties to their oxepin
tautomers (42). Hence, 3 is most likely spontaneously formed from 2 (Figure 1A), which is then
enzymatically converted to 4. In vitro, the observed reversible PaaG-mediated equilibrium lies
on the side of 4 over 2/3 (2). Although physiological conditions (i.e. watery solution with high
dielectric constant) and the lack of a conjugated α-substitution should favor 2 over 3 (42), PaaG
5
ACS Paragon Plus Environment

Page 11 of 21
ACS Chemical Biology
1
2
3
4
5
6
7
8
9
199
200
201
202
203
204
205
206
207
may stabilize and process the oxepin tautomer in the active site, as the geometry of the planar
3 backbone more closely resembles typical 3,4-enoyl-CoA substrates. Notably, based on our
kinetic analyses, PaaGTT (in contrast to PaaGPY2) appears to be significantly better adapted for
processing of 2/3 rather than 9. While a structure of PaaGPY2 for direct comparison between the
two homologous enzymes is currently lacking, we speculate that this substrate preference in
PaaGTT proved advantageous in the course of evolution due to the instability and toxicity of 2,
which is exacerbated under thermophilic conditions. It cannot be ruled out, however, that
another isomerase is responsible for the isomerization of 9 into 10 in T. thermophilus.
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
208
209
210
211
212
Figure 3. Polder OMIT maps (contoured at 3σ above the mean) of (A) PaaGTT-D136N•4 (PDB ID: 6SLA), (B)
PaaGTT•10/12 (PDB ID: 6SLB), and (C) apo-PaaGTT (PDB ID: 6SL9). The unbiased maps allowed placement of
the ligands within the electron density. For unsoaked apo-PaaGTT, water molecules and spurious density could be
observed. Note that the distance and orientation of residue 136 to the C2 and C4 atoms of the ligand in 6SLA
differs due to replacement of the catalytic aspartate residue with asparagine, which adopts a different conformation.
213
Critical Interactions and Active Site Adaptations of the PaaG Complex Structures
214
215
216
217
218
219
220
221
222
223
224
225
226
227
228
229
230
231
232
233
234
235
236
237
238
239
240
241
242
243
244
245
246
247
248
In both protein·substrate complexes, the adenine of the CoA moiety stacks (π-π
interaction) with the phenolic ring of the highly-conserved Phe243 (38) from the adjacent
monomer and is located at the bottom of a well-defined open binding pocket (site 1)
(Supplementary Figure 6). CoA binding also involves a salt bridge of the diphosphate moiety
with Arg55, as well as hydrogen bonds with Ala59, Glu61 and Leu63 and other non-hydrogen
bond interactions (Supplementary Figure 6). In contrast, the acyl moieties of the ligands (both
4 and 10/12) were oriented towards a shallow active site (site 2), which is mostly enclosed by
hydrophobic residues. Notably, the main chain amide groups of Gln61 and Ala105 at site 2
participate in H-bonding with the thioester carbonyl of both ligands and are thus ideally
positioned to stabilize the thioester enolate transition state (Figure 3). Our mutagenesis studies
furthermore suggested a catalytic role for D136 in PaaGTT. Indeed, this residue is positioned in
close proximity to C2 and C4 of both ligands 4 and 10/12, with distances between 2.8 Å and
4.1 Å (Figure 4), fully consistent with its envisaged role as proton relay. The carboxylic group
of 10/12 weakly hydrogen bonds with the Tyr73 side chain (distance of 3.5 Å), thus promoting
ligand binding in the hydrophobic site 2.
A comparison of the active-sites of apo-PaaGTT, PaaGTT-D136N·4 and PaaGTT·10/12
via superimposition of the structures revealed a significant plasticity of the loop β3-α3 residues
(residues 66-74). In PaaGTT·10/12, this loop moved inward compared to the apo form and
participates in substrate binding, whereas the density for this loop including Tyr73 was missing
in PaaG D136N·4 and thereby provided space for the bulky oxepin ring (Figure 4,
Supplementary Figure 6). In addition to loop β3-α3 flexibility, a displacement of the Tyr80 side
chain by around 10 Å resulted in two distinct conformations, as previously observed (38).
Interestingly, while the Tyr80 “OUT” conformation was found for apo-PaaGTT, the Tyr80
moved to the “IN” conformation for both protein·substrate complexes, thereby forming a lid
over the ligands and furthermore establishing a hydrogen-bond with the side-chain of Gln61
(Figure 3). This closing of the active site was accompanied by the expulsion of water
(Supplementary Figure 6). When replaced with phenylalanine in PaaGTT-Y80F, the enzyme
showed reduced activity for the conversion of 12 and 2/3, whereas a dual variant PaaGTT-
Y80F/D136N exhibited only residual activity for 2/3 and did not convert 12 (Table 1). Hence,
mutagenesis, enzyme kinetics, and structural data imply an important role for the mobile Tyr80.
Ligand binding (4 as well as 12) triggers a conformational change of Tyr80 that closes up the
active site and displaces water. We speculate that a water-free active site is likely relevant to
prevent undesired hydratase activity. Such a side reaction seems plausible, as the functionality
of both enoyl-CoA isomerases, as well as structurally-related hydratases, relies on the
stabilization of thioester enolate transition states via an oxyanion hole. Consequently, the β-
6
ACS Paragon Plus Environment

ACS Chemical Biology
Page 12 of 21
1
2
3
4
5
6
7
8
9
249
250
251
252
253
254
255
256
257
258
259
260
261
carbons of product 10 or other 2,3-enoyl-CoAs are susceptible to nucleophilic attack when
bound in the active site. By excluding water, PaaG may thus avoid unspecific hydration to (R)-
3-hydroxyadipoyl-CoA that cannot be further processed by (S)-specific 3-hydroxyacyl-CoA
dehydrogenases (PaaH in the 1 catabolic pathway). In addition to this functionality, Tyr80 acts
as H-bond donor for the carbonyl of the Glu61 side chain (Figure 3), whose main chain NH
group stabilizes the enolate transition state. Hence, the H-bonding interaction of Glu61 with
Tyr80 likely promotes the active site closure and water expulsion.
These structural adaptations, i.e. the independent movements of two segments (Tyr80
side chain and loop β3-α3) of the active center, apparently achieve the relaxed substrate
specificity and facilitate catalysis. From an evolutionary point of view, PaaG may derive from
a conventional Δ3,Δ2-enoyl-CoA isomerase recruited from fatty acid catabolism and may have
conceivably been selected for aromatic degradation because of flexible structural elements that
allow the accommodation of bulkier ring substrates.
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
262
263
264
265
266
267
Figure 4. Overall structure and active site of PaaGTT. (A) Superposition of apo-PaaGTT (grey), PaaGTT-D136N•4
chain A (green), PaaGTT•10/12 (blue). (B, C, D) Inlets showing the respective enlarged active sites. Catalytic
residues and ligands are represented as stick models. Distances, e.g., between the catalytic Asp136 and the C2 and
C4 atoms of the ligands are indicated in Å. The mobile Tyr80 closes the active site upon substrate binding and
expels water. To accommodate 4, loop β3-α3 opens up and thereby provides space for the bulky ring.
268
PaaG Restricts the Motional Flexibility of Bound Ligands for Stereochemical Control
269
270
271
272
273
274
275
276
277
278
279
280
281
282
283
284
285
286
287
288
289
To the best of our knowledge, the structural requisites that govern the stereochemistry
of Δ3,Δ2-enoyl-CoA isomerases have not been clearly established. We thus looked into
structural factors in PaaGTT that may control the stereochemistry of the isomerization of 9 to 10
resulting in a trans-2,3 configuration, as observed for other Δ3,Δ2-enoyl-CoA isomerases. We
propose that PaaGTT and related isomerases may constrain the motional flexibility of their
substrates and thereby preclude formation of cis-2,3 side products, which would arise from (re-
)protonation of a bent thioester enolate transition state (Figure 5A). To investigate that, we
modelled bent transition state intermediates into the active site of PaaGTT and aligned the
respective CoA-moieties and C1–C2 axes with the experimentally observed PaaGTT•10/12
complex structure, while rotating the C2–C3 bonds. Distances between the catalytic D136 side
chains and the respective displaced C4 atoms of the conformers remained modest (3.7 – 4.4 Å).
On the other hand, the modelling clearly revealed steric clashes with the protein backbone, thus
allowing only the elongated conformers to bind (Figure 5BC). A modeling in the active site of
multifunctional enzyme, type-1 (MFE1) (33) revealed similar steric clashes with the protein
backbone (Supplementary Figure 7). These results suggest that a confined active site may be a
key determinant for stereochemical control by Δ3,Δ2-enoyl-CoA isomerases in order to steer
the formation of trans-2,3-enoyl-CoAs, which are also energetically favored over the respective
cis-isomers. In addition, the placement of the catalytic aspartate (D136) may contribute to the
regiospecificity, although distances between the D136 side chain and the C4-atom of the
substrates in the hypothetical bent states are only marginally different to the experimentally
observed ones (Figure 4).
290
291
292
293
294
295
296
297
298
Figure 5. (A) Possible conformations of PaaG’s native substrate 9 in enzyme free environment (top). The thioester
enolates, following deprotonation, represent the transition states (middle), which are converted to the products
(bottom). If the transition state adopts a bent conformation (red boxes), the undesired cis-2,3 product would be
formed rather than the observed trans-2,3 (blue boxes). R = CH2COOH. (B, C) PaaGTT suppresses the bent
conformation of the transition state through a confined active site, thereby steering the exclusive formation of
trans-2,3-didehydroadipoyl-CoA (10). Modeled bent cis-1,2, cis-3,4 transition state binding-modes are shown as
yellow sticks (B) and modeled bent trans-1,2, cis-3,4 transition state binding-modes as green sticks (C). Both
transition state models are aligned along the CoA-moieties and C1-C2 axes with the experimentally observed 10/12
7
ACS Paragon Plus Environment

Page 13 of 21
ACS Chemical Biology
1
2
3
4
5
299
300
(pink sticks). Red spheres represent steric clashes of transition states with the protein. The C2 and C4 atoms are
highlighted as spheres with indicated distances to the catalytic D136 side chain.
6
7
301
Conclusion
302
303
304
305
306
307
308
309
310
311
312
313
314
315
316
317
318
319
320
321
322
323
324
325
326
327
328
329
330
331
332
333
334
335
336
In this work we biochemically and structurally characterized the unusual isomerase
PaaG and verify a Δ3,Δ2-enoyl-CoA isomerases-like mechanism. Our mechanistic and
structural studies of the wild type and the enzyme variants furthermore strongly suggest a
catalytic aspartate (D136) side chain that is ideally situated to act as proton shuttle, similar to
structurally related isomerases. The installment of an α-β double bond in 4 by PaaG
subsequently enables facile β-hydration, ring cleavage and further β-oxidation-like steps.
Hence, PaaG acts as crucial mediator between aromatic degradation and β-oxidation. In
addition to the previously confirmed conversion of 2/3 into 4 (2), we verify a second role for
PaaG in the 1 catabolic pathway, where PaaG catalyzes the isomerization of 9 into 10 and thus
processes two structurally-distinct substrates. Our data furthermore grants insight into how
control of trans-2,3 stereospecificity is exerted by PaaGTT. We show that PaaGTT (and related
enzymes) feature a confined substrate binding site and thereby preclude formation of cis-2,3
side products that would result from bent thioester enolate transition states.
Interestingly, the PaaG product 4 is not only degraded to acetyl-CoA and succinyl-CoA,
but most likely also serves as precursor for unusual primary metabolites (ω-cycloheptyl fatty
acids (15)) as well as secondary metabolites (11) known to have antimicrobial, anticancer, and
antiviral activities, i.e. tropone natural products such as 8 (20,16,43,21,19) or roseobacticides
(28,27,26)) (Figure 1A). Tropones and derivatives are produced by numerous bacteria including
predominant marine Roseobacter sp. and are also detected among the bacterial volatiles (44–
46)). Moreover, 8 has been shown to not only act as antibiotic, but also as bacterial quorum-
sensing signal (47,17). These compounds furthermore appear to play a central role in marine
symbiotic relationships of bacterial producers with oysters, sponges, algae and corals by
warding off pathogens (27,22,48,23) and likely have a pivotal role in structuring coral-
associated bacterial communities (48). Moreover, 8 is a promising antibiotic, e.g., for use in
aquacultures, as resistance mechanisms develop slowly, are conferred on a low level, and
disappear rapidly (23,49,50). PaaG and the extraordinary isomerization of 2/3 to 4 is thus a key
functionality for formation of these ecologically and pharmaceutically interesting metabolites.
In summary, we characterized the bacterial PaaG in detail and for the first time
elucidated structures of a Δ3,Δ2-enoyl-CoA isomerase in complex with its native ligands. Our
data provide a rationale for the observed substrate flexibility as well as the enzymatic control
of stereospecificity. Hence, our findings also shed new light onto the general Δ3,Δ2-enoyl-CoA
isomerase type, which are most commonly involved in the β-oxidation of unsaturated fatty acids
that carry a double bond at an odd position. In the future, PaaG may be a prime candidate for
biotechnological applications and exploited for in vitro or in vivo production and bioengineering
of tropone natural products.
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
337
338
Methods
339
340
341
342
343
344
345
346
347
Cloning and Mutagenesis
To heterologously produce PaaG from Thermus thermophilus, the gene was amplified by PCR
(oligonucleotides in Supplementary Table 1) from genomic DNA. The PCR product was
isolated and cloned into the pET His6 TEV LIC cloning vector (Addgene). For site-directed
mutagenesis, primers with single-point-mutation (Supplementary Table 4) were used to create
the mutants PaaGTT Y80F, D136N and Y80F/D136N. Sequence mutations were confirmed by
DNA sequencing (Eurofins Genomics).
Protein Expression and Purification
8
ACS Paragon Plus Environment

ACS Chemical Biology
Page 14 of 21
1
2
3
4
5
6
7
8
9
348
349
350
351
352
353
354
355
356
357
358
359
360
361
362
363
364
365
366
367
368
369
370
371
372
373
374
375
376
377
378
379
380
381
382
383
384
385
386
387
388
389
390
391
392
393
394
395
396
397
398
PaaG from Pseudomonas sp. Y2 was produced in E. coli BL21 (DE3) and purified as described
previously (2). The isolated protein was stored with 30 % glycerol at -20°C. For overexpression,
the recombinant plasmid pET His6 TEV LIC-paaGTT was transformed into E. coli BL21 (DE3)
(Thermo Fisher Scientific). Liquid cultures were grown shaking at 37°C until an OD600 of 0.5-
0.6 was reached. Then, the temperature was lowered to 18°C and protein production was
induced by adding IPTG to a final concentration of 100 µM. After 19-22 hours, the cells were
harvested by centrifugation at 3220 g and 4°C for 15 minutes. Cell pellets were stored at –20°C
or directly resuspended in buffer A (20 mM Tris-HCl, 100 mM KCl, 1 mg mL-1 DNase I, pH
8.0) for protein isolation. The suspension was sonicated 3 times (0.5 sec on, 2.5 sec off, 40 %
amplitude, 5 minutes pause) on ice, incubated for 15 minutes in a 65°C water bath, and
centrifuged (100000 x g) at 4°C for 1 h. The supernatant was directly applied to a FF HisTrap
5 mL column (GE Healthcare) by FPLC (Äkta Pure, GE Healtcare), washed and eluted with
buffer B (20 mM Tris-HCl, 100 mM KCl, 500 mM imidazole, pH 8.0). After elution, the
fractions with high protein concentration were pooled and desalted with a HiTrap-column (GE
Healthcare). For crystallization, the polyhistidine-tags of PaaGTT wt and the PaaGTT-D136N
variant were cleaved off by digestion with TEV protease. For that, the protein fractions were
pooled and concentrated using a 10 kDa MW cut-off MACROSEP spin column (Pall
Cooperation) to a volume of 4 mL and supplemented with 1 mM dithiothreitol (DTT), 0.5 mM
EDTA (pH 8.0) and 1.5 mg TEV-His protease. The mixture was incubated over night at 4°C
and then applied to a FF HisTrap column. The flow-through was collected that contained the
protein without tag. The purity of the protein was then increased by size exclusion
chromatography utilizing a HiLoad 16/600 Superdex 200 pg column (GE Healthcare)
equilibrated with buffer C (20 mM Tris-HCl, 150 mM NaCl, 1 mM DTT pH 8.0) and connected
to a FPLC system. The eluted protein was stored with 20 % (v/v) glycerol at -20 °C. The purity
of the eluted protein was confirmed by SDS-PAGE.
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
Specific Activities for Unstable 1,2-Epoxyphenylacetyl-CoA (2) and cis-3,4-
Didehyroadipoyl-CoA (9).
A solution of 50 mM Tris-HCl pH 8.0, 0.3 mM phenylacetyl-CoA, 1 mM NADPH and 1 mg
mL-1 PaaABCE was prepared. The premix was incubated at 30°C for 3 minutes. The
isomerization reaction was then started by the addition of PaaG (1 nM PaaGTT wt, 2 nM PaaGTT
Y80F, 20 nM PaaGTT D136N, 10 nM PaaGTT Y80F/D136N or 0.5 nM PaaGPY2 wt). To calculate
the specific activity, the formed 4 over time was quantified by HPLC using a standard curve.
Accordingly, samples were withdrawn at varying time points and quenched with 100 % (v/v)
ice cold MeOH. After centrifugation for 20 min at 18,000 × g, MeOH was removed by speed
vac and samples analyzed by HPLC. For 9, the specific activity was also determined via a
coupled assay and quantified by UPLC-HRMS. A solution of 50 mM Tris-HCl pH 8.0, 0.3 mM
3-oxo-5,6-didehydrosuberyl-CoA, 0.45 mM CoA and 0.2 µM PaaJ was prepared and incubated
at 30°C for 10 minutes. The isomerization reaction was then started by the addition of PaaGTT
wt (800 nM, 400 nM, 160 nM) or PaaGPY2 wt (25 nM, 10 nM, 2,5 nM). After 1 min, the reaction
was stopped by addition of 200 % (v/v) ice-cold MeOH. After centrifugation for 2 x 20 min at
18,000 × g, MeOH was removed by speed vac and samples were analyzed by UPLC-HRMS.
Photometric Assays for trans-3,4-Didehyroadipoyl-CoA
The kinetic parameters for trans-3,4-didehyroadipoyl-CoA were determined photometrically
using a UV-1650PC Shimadzu spectrophotometer. The isomerization of trans-3,4-
didehyroadipoyl-CoA to trans-2,3-didehyroadipoyl-CoA was quantified by measuring the
increase of absorption at 260 nm because of different extinction coefficients (16.4 cm-1 mM-1
for saturated acyl-CoAs; 22.4 cm-1 mM-1 for unsatured enoyl-CoAs) (39). Reaction rates were
then calculated according to Beer–Lambert law. Because of the absorption limit of the detector,
a 1 mm cuvette (Type 110-QS, 1 mm, Hellma) was used for concentrations of 200 µM substrate
9
ACS Paragon Plus Environment

Page 15 of 21
ACS Chemical Biology
1
2
3
4
5
6
7
8
399
400
401
and above and a 10 mm cuvette (Type 105.205-QS, Hellma) for 100 µM and below. The
reaction mixtures contained 10 µM, 20 µM, 50 µM, 100 µM, 200 µM or 1 mM trans-3,4-
didehyroadipoyl-CoA in 50 mM Tris-HCl pH 8.0. The reactions were started by addition of a
sufficient quantity of PaaG that allowed measuring the linear increase in absorption. The slope
of the absorption was read out by the software (UV Probe 2.43, Shimadzu).
402
403
404
405
406
407
408
409
410
411
412
413
414
415
416
417
418
419
420
421
422
423
424
425
426
427
428
429
430
431
432
433
434
435
436
437
438
439
440
441
442
443
444
445
446
447
448
449
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
Chemical and Enzymatic Synthesis of CoA-Esters
Trans-3,4-didehydroadipoyl-CoA was chemically synthesized from free acid and CoA via
activation of the carboxyl group by N,N’-dicyclohexylcarbodiimide and esterification to the
respective succinimide ester intermediates (51). Accordingly, trans-2-butene-1,4-dicarboxylic
acid (abcr, Germany) (0.75 g, 5 mmol) and 0.58 g N-hydroxysuccinimid were dissolved in 25
mL water-free dioxane. A solution of dicyclohexylcarbodiimid (1.55 g, 7.5 mmol) in 5 mL
dioxane was dripped to the starter mixture over 30 min and the reaction mixture stirred
overnight. The filtrate was freeze dried and stored at -20 °C. For the esterification, 20 mL
hydrogencarbonat solution (100 mM, pH 8) was mixed with 50 mg succinimid ester and
Coenzyme A (50 mg, 64 µmol) under anaerobic conditions. The consumption of free CoA was
checked with Ellmans’ reagant. After completion of the reaction the pH was adjusted to 3.5
with 20 % acetic acid. After reduction of the volume, the CoA-esters were purified via HPLC.
Phenylacetyl-CoA was synthesized via the same method beginning with phenylacetic-
succinimid (inherited from Georg Fuchs) and free CoA. Cyclohexenyliden-CoA was
chemically synthesized of 2-(cyclohex-1-en-1-yl)acetic acid (Enamine, Ukraine) (3.8 mg, 26.8
µmol), 1-Hydroxybenzotriazole (3.6 mg, 26.8 µmol), tetramethyluronium tetrafluoroborate
(8.6 mg, 26.8 µmol), N-ethyldiisopropylamine (16 µL, 12.2 mg, 94 µmol) and Coenzyme A
trilithium salt (20 mg, 24.4 µmol) according to (52). All chemical syntheses were conducted at
room temperature. Compounds 4 and 6 were enzymatically synthesized in vitro in large scale
in 20 mL assays and purified by semi-preparative HPLC. For 4 formation, assays contained 20
µg PaaABCE, 0.4 µg PaaGPY2 wt, 1 mM NADPH, 0.5 mM Pa-CoA in 50 mM Tris-HCl pH 8.0.
Compounds were then purified via semi-preparative RP-HPLC and lyophilized.
Ultra/High Performance Liquid Chromatography
Compounds were isolated using reverse-phase High Performance Liquid Chromatography (RP-
HPLC) on a 1100 series chromatographic system (Agilent Technologies) with a semi-
preparative C18-E column (Nucleodur C18e 250/10, Machery-Nagel). For activity tests, the
column was developed at a flow rate of 4.5 mL min-1 by a linear gradient from 2 % acetonitrile
in 10 mM ammonium acetate (pH 6.8) to 100% acetonitrile within 20 min, after 2 min at
isocratic conditions with 2% acetonitrile. For isolation of chemically produced compounds or
enzymatically produced 6, a flow rate of 3.5 mL min-1 with the same gradient but with
ammonium acetate at pH 4.5 was used.
Liquid Chromatography-Mass Spectrometry
For identification of intermediates, enzymatic assays were analyzed by Waters Acquity I-class
UPLC (Waters C-18 HSS T3 column, 2.1 mm x 100 mm, 1.8 μm particle size) coupled to a
Waters Acquity photo diode array detector (Waters) and a Waters Synapt G2-Si HDMS
electrosprayionization (ESI)/quadrupole time-of-flight (Q-TOF) system. Two minutes after
injection, the initial concentration of 2 % acetonitrile in ammonium acetate (pH 4.5) at a flow
rate of 0.2 mL min-1 was increased via a linear gradient to 30 % acetonitrile in 9 min. CoA
esters were measured in MS positive mode with a capillary voltage of 1.5 kV, 100 °C source
temperature, 300°C desolvation temperature, 600 L min-1 N2 desolvation gas flow. Collision
induced dissociation of precursor ions was performed using a collision energy ramp from 10 to
40 V.
10
ACS Paragon Plus Environment

ACS Chemical Biology
Page 16 of 21
1
2
3
4
5
6
7
8
9
450
451
452
453
454
455
456
457
458
459
460
461
462
463
464
465
466
467
468
469
470
471
472
473
474
475
476
477
478
479
480
481
482
483
484
485
486
487
488
489
490
491
492
493
494
495
496
497
498
499
Circular Dichroism Spectroscopy
Circular dichroism (CD) of PaaGTT wt and variants were recorded in a 0.2 mm path length
cuvette at 25°C with a Jasco J-810 spectropolarimeter (Jasco, Easton, MD, USA) equipped with
a Peltier temperature controller. Protein samples were dissolved in 10 mM sodium phosphate
buffer (pH 8.0). Protein concentration was determined by the Bradford method. Instrumental
parameters for measurement of the CD spectra were: 190-300 nm measurement range, 1 nm
band width, standard sensitivity, 200 nm/min of scanning speed. Analysis method: Three
spectra for each enzyme variant as well as the buffer were measured and averaged. To receive
the CD spectra, the buffer was subtracted and the received CD spectra were normalized to mean
residue ellipticity θMRW (deg x cm2 x dmol-1).
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
Crystallization, X-ray Data Collection and Processing
Crystallization screens were conducted using the sitting drop vapor diffusion method matching
the published conditions. After screening and optimization with various precipitants,
crystallizations were set up manually using the vapor diffusion sitting drop method with 48-
well plates. The reservoir solution contained 50 mM KH2PO4, pH 4.5, 20-22% PEG 3350, 3%
(w/v) non-detergent sulfobetain (NDSB-201) and 20% glycerol (v/v) at 25°C. The drops were
set up at a 1:1 ratio of protein to reservoir solution and incubated at 25 °C. Co-crystallization
trials with substrates oxepin-CoA (4) and trans-3,4-didehydroadipoyl-CoA (11) were
unsuccessful, hence crystal soaking was implemented. For PaaG D136N·4 complex, respective
crystals were soaked in a 10 mM solution of 4 for a short period of 2 hours to 4 hours. For
PaaG·10/12 complex, a pinch of the compound was sprinkled over the crystals for soaking
between 2 and 16 hours. Each crystal was transferred into the reservoir solution, which also
acted as cryobuffer, and subsequently flash-cooled to 100 K in a cold nitrogen-gas stream.
Crystals diffracted up to 1.27 Å for the apo protein, 1.88 Å for PaaG·10/12, and up to 2.55 Å
for PaaG D136N·4 and were recorded at the Swiss Light source (Villigen, Switzerland) on
beam station PXI at 100 K. Cryo cooling was carried out prior to X-ray data collection, after
stabilizing the crystals in mother liquor cryoprotectant. The datasets were processed by XDS in
P213 (a = b = c = 88.02 for the 1.27 Å dataset and a = b = c = 88.45 for the 1.88 Å dataset) and
P21 (a = 76.16 b = 73.05 c = 130.30, β = 92.5° for the 2.55 Å dataset). The X-ray diffraction
reflections were merged and scaled using SCALA.
Structure Determination, Model Building and Refinement
Initial phases for the apo structure at 1.27 Å and PaaG·10/12 complex at 1.88 Å (form II), and
PaaG·4 complex at 2.55 Å (form III) were calculated by molecular replacement using PDB
code 3HRX chain A of apo PaaG (space group P212121, form I) as the search model and
PHASER as implemented in PHENIX. The missing residues from 66 to 74 in the original
published PaaG structure were built based on the electron density map of the PaaG apo structure
at 1.27 Å. Electron density for ligands 4 and 10/12 were observed at the active site in their
respective complexes and these ligands were modeled. To ensure that stereochemical restraints
imposed on the ligands during the refinement process had not biased the geometry of the
resulting models in the active site, restraints were removed from the ligands and a subsequent
round of simulated annealing refinement was carried out in PHENIX (53). For omit map
calculations, the Polder OMIT maps tool was used. Further refinement and model building were
performed using the PHENIX software suite and COOT (54). Statistics for data reduction,
model refinement, and model quality are summarized in Table 1. The model is complete for
residues 1–254 with the exception of residues 66–74, in external loops between helices α2 to
α3 that could not be modeled because of conformational heterogeneity.
Figure Generation and Data Deposition
11
ACS Paragon Plus Environment

Page 17 of 21
ACS Chemical Biology
1
2
3
4
5
6
7
8
9
500
501
502
Graphic images were generated using PyMOL(55) and representations of substrate binding
were generated using the academic free version of Discovery Studio Visualizer (56).
503
504
505
506
507
508
509
510
511
512
513
514
515
516
517
518
519
520
521
522
523
524
525
526
527
528
529
530
531
532
533
534
535
536
537
538
539
540
541
542
543
544
545
546
547
548
549
550
551
Accession codes
The atomic coordinates and structure factors (PDB codes 6SL9, 6SLA, 6SLB) were deposited
in the Protein Data Bank.
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 Available: This material is available free of charge via the Internet.
Additional data on enzyme kinetics, CD spectroscopy, X-ray crystallography (including
refinement statistics), and list of oligonucleotides (PDF).
Author Information
Corresponding Author: *E-mail: robin.teufel@zbsa.uni-freiburg.de.
ORCID:
Robin Teufel: 0000-0001-5863-7248
Author Contributions: M.S. planned and conducted mechanistic experiments and analyzed
and interpreted data; R.S.B. planned and conducted X-ray crystallography experiments and
analyzed and interpreted data; R.T. designed research, analyzed and interpreted data and wrote
the manuscript. M.S. and R.S.B. contributed equally to the work.
Notes: The authors have no conflict of interest to declare.
Acknowledgements: We thank G. Fuchs in whose group this project was initiated and S.
Lüdeke for assisting with CD spectroscopy. We thank T. Takashi from the Swiss Light Source
(SLS), Villigen (CH) for technical support.
Funding: This work was supported by a DFG grant awarded to R.T. (TE 931/2-1). The research
leading to these results has received funding from the European Union’s Horizon 2020 research
and innovation programme under grant agreement n.° 730872, project CALIPSOplus.
References
1. Fuchs, G., Boll, M., and Heider, J. (2011) Microbial degradation of aromatic compounds — from one strategy
to four, Nat. Rev. Microbiol. 9, 803–816.
2. Teufel, R., Mascaraque, V., Ismail, W., Voss, M., Perera, J., Eisenreich, W., Haehnel, W., and Fuchs, G.
(2010) Bacterial phenylalanine and phenylacetate catabolic pathway revealed, Proc. Natl. Acad. Sci. U S A 107,
14390–14395.
3. Olivera, E. R., Miñambres, B., García, B., Muñiz, C., Moreno, M. A., Ferrández, A., Díaz, E., García, J. L.,
and Luengo, J. M. (1998) Molecular characterization of the phenylacetic acid catabolic pathway in Pseudomonas
putida U: the phenylacetyl-CoA catabolon, Proc. Natl. Acad. Sci. U S A 95, 6419–6424.
4. Navarro-Llorens, J. M., Patrauchan, M. A., Stewart, G. R., Davies, J. E., Eltis, L. D., and Mohn, W. W. (2005)
Phenylacetate catabolism in Rhodococcus sp. strain RHA1: a central pathway for degradation of aromatic
compounds, J. Bacteriol. 187, 4497–4504.
5. Luengo, J. M., García, J. L., and Olivera, E. R. (2001) The phenylacetyl-CoA catabolon: a complex catabolic
unit with broad biotechnological applications, Mol. Microbiol. 39, 1434–1442.
6. Ismail, W., El-Said Mohamed, M., Wanner, B. L., Datsenko, K. A., Eisenreich, W., Rohdich, F., Bacher, A.,
and Fuchs, G. (2003) Functional genomics by NMR spectroscopy. Phenylacetate catabolism in Escherichia coli,
Eur. J. Biochem. 270, 3047–3054.
12
ACS Paragon Plus Environment

ACS Chemical Biology
Page 18 of 21
1
2
3
4
5
6
7
8
9
552
553
554
555
556
557
558
559
560
561
562
563
564
565
566
567
568
569
570
571
572
573
574
575
576
577
578
579
580
581
582
583
584
585
586
587
588
589
590
591
592
593
594
595
596
597
598
599
600
601
602
603
604
605
606
607
608
609
610
611
7. Ferrández, A., Miñambres, B., García, B., Olivera, E. R., Luengo, J. M., García, J. L., and Díaz, E. (1998)
Catabolism of phenylacetic acid in Escherichia coli. Characterization of a new aerobic hybrid pathway, J. Biol.
Chem. 273, 25974–25986.
8. Bartolomé-Martín, D., Martínez-García, E., Mascaraque, V., Rubio, J., Perera, J., and Alonso, S. (2004)
Characterization of a second functional gene cluster for the catabolism of phenylacetic acid in Pseudomonas sp.
strain Y2, Gene 341, 167–179.
9. Rather, L. J., Knapp, B., Haehnel, W., and Fuchs, G. (2010) Coenzyme A-dependent aerobic metabolism of
benzoate via epoxide formation, J. Biol. Chem. 285, 20615–20624.
10. Teufel, R., Friedrich, T., and Fuchs, G. (2012) An oxygenase that forms and deoxygenates toxic epoxide,
Nature 483, 359–362.
11. Teufel, R., Gantert, C., Voss, M., Eisenreich, W., Haehnel, W., and Fuchs, G. (2011) Studies on the
mechanism of ring hydrolysis in phenylacetate degradation: a metabolic branching point, J. Biol. Chem. 286,
11021–11034.
12. Pribytkova, T., Lightly, T. J., Kumar, B., Bernier, S. P., Sorensen, J. L., Surette, M. G., and Cardona, S. T.
(2014) The attenuated virulence of a Burkholderia cenocepacia paaABCDE mutant is due to inhibition of
quorum sensing by release of phenylacetic acid, Mol. Microbiol. 94, 522–536.
13. Erb, T. J., Ismail, W., and Fuchs, G. (2008) Phenylacetate metabolism in thermophiles: characterization of
phenylacetate-CoA ligase, the initial enzyme of the hybrid pathway in Thermus thermophilus, Curr. Microbiol.
57, 27–32.
14. Grishin, A. M., Ajamian, E., Tao, L., Zhang, L., Menard, R., and Cygler, M. (2011) Structural and functional
studies of the Escherichia coli phenylacetyl-CoA monooxygenase complex, J. Biol. Chem. 286, 10735–10743.
15. Moore, B. S., Walker, K., Tornus, I., Handa, S., Poralla, K., and Floss, H. G. (1997) Biosynthetic Studies of
omega-Cycloheptyl Fatty Acids in Alicyclobacillus cycloheptanicus. Formation of Cycloheptanecarboxylic Acid
from Phenylacetic Acid, J. Org. Chem. 62, 2173–2185.
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
16. Bentley, R. (2008) A fresh look at natural tropolonoids, Nat. Prod. Rep. 25, 118–138.
17. Berger, M., Neumann, A., Schulz, S., Simon, M., and Brinkhoff, T. (2011) Tropodithietic acid production in
Phaeobacter gallaeciensis is regulated by N-acyl homoserine lactone-mediated quorum sensing, J. Bacteriol.
193, 6576–6585.
18. Berger, M., Brock, N. L., Liesegang, H., Dogs, M., Preuth, I., Simon, M., Dickschat, J. S., and Brinkhoff, T.
(2012) Genetic analysis of the upper phenylacetate catabolic pathway in the production of tropodithietic acid by
Phaeobacter gallaeciensis, Appl. Environ. Microbiol. 78, 3539–3551.
19. Brock, N. L., Nikolay, A., and Dickschat, J. S. (2014) Biosynthesis of the antibiotic tropodithietic acid by the
marine bacterium Phaeobacter inhibens, Chem. Commun. 50, 5487–5489.
20. Chen, X., Xu, M., Lü, J., Xu, J., Wang, Y., Lin, S., Deng, Z., and Tao, M. (2018) Biosynthesis of Tropolones
in Streptomyces spp.: Interweaving Biosynthesis and Degradation of Phenylacetic Acid and Hydroxylations on
the Tropone Ring, Appl. Environ. Microbiol. Epub 2018. DOI: 10.1128/AEM.00349-18.
21. Geng, H., Bruhn, J. B., Nielsen, K. F., Gram, L., and Belas, R. (2008) Genetic dissection of tropodithietic
acid biosynthesis by marine roseobacters, Appl. Environ. Microbiol. 74, 1535–1545.
22. Harrington, C., Reen, F. J., Mooij, M. J., Stewart, F. A., Chabot, J.-B., Guerra, A. F., Glöckner, F. O.,
Nielsen, K. F., Gram, L., Dobson, A. D. W., Adams, C., and O'Gara, F. (2014) Characterisation of non-
autoinducing tropodithietic Acid (TDA) production from marine sponge Pseudovibrio species, Mar. Drugs 12,
5960–5978.
23. Zhao, W., Dao, C., Karim, M., Gomez-Chiarri, M., Rowley, D., and Nelson, D. R. (2016) Contributions of
tropodithietic acid and biofilm formation to the probiotic activity of Phaeobacter inhibens, BMC Microbiol. 16,
1.
24. Kintaka, K., Ono, H., Tsubotani, S., Harada, S., and Okazaki, H. (1984) Thiotropocin, a new sulfur-
containing 7-membered-ring antibiotic produced by a Pseudomonas sp, J. Antibiot. 37, 1294–1300.
25. Guo, H., Roman, D., and Beemelmanns, C. (2018) Tropolone natural products, Nat. Prod. Rep.
26. Seyedsayamdost, M. R., Carr, G., Kolter, R., and Clardy, J. (2011) Roseobacticides: small molecule
modulators of an algal-bacterial symbiosis, J. Am. Chem. Soc. 133, 18343–18349.
27. Seyedsayamdost, M. R., Wang, R., Kolter, R., and Clardy, J. (2014) Hybrid biosynthesis of roseobacticides
from algal and bacterial precursor molecules, J. Am. Chem. Soc. 136, 15150–15153.
28. Wang, R., Gallant, É., and Seyedsayamdost, M. R. (2016) Investigation of the Genetics and Biochemistry of
Roseobacticide Production in the Roseobacter Clade Bacterium Phaeobacter inhibens, mBio 7, e02118.
29. Nogales, J., Macchi, R., Franchi, F., Barzaghi, D., Fernández, C., García, J. L., Bertoni, G., and Díaz, E.
(2007) Characterization of the last step of the aerobic phenylacetic acid degradation pathway, Microbiology
(Reading, England) 153, 357–365.
30. Holden, H. M., Benning, M. M., Haller, T., and Gerlt, J. A. (2001) The crotonase superfamily: divergently
related enzymes that catalyze different reactions involving acyl coenzyme a thioesters, Acc. Chem. Res. 34, 145–
157.
13
ACS Paragon Plus Environment

Page 19 of 21
ACS Chemical Biology
1
2
3
4
5
6
7
8
9
612
613
614
615
616
617
618
619
620
621
622
623
624
625
626
627
628
629
630
631
632
633
634
635
636
637
638
639
640
641
642
643
644
645
646
647
648
649
650
651
652
653
654
655
656
657
658
659
660
661
662
663
664
665
666
667
668
669
670
671
672
31. Grishin, A. M., Ajamian, E., Zhang, L., Rouiller, I., Bostina, M., and Cygler, M. (2012) Protein-protein
interactions in the β-oxidation part of the phenylacetate utilization pathway: crystal structure of the PaaF-PaaG
hydratase-isomerase complex, J. Biol. Chem. 287, 37986–37996.
32. Engel, C. K., Kiema, T. R., Hiltunen, J. K., and Wierenga, R. K. (1998) The crystal structure of enoyl-CoA
hydratase complexed with octanoyl-CoA reveals the structural adaptations required for binding of a long chain
fatty acid-CoA molecule, J. Mol. Biol. 275, 847–859.
33. Kasaragod, P., Schmitz, W., Hiltunen, J. K., and Wierenga, R. K. (2013) The isomerase and hydratase
reaction mechanism of the crotonase active site of the multifunctional enzyme (type-1), as deduced from
structures of complexes with 3S-hydroxy-acyl-CoA, FEBS J. 280, 3160–3175.
34. Partanen, S. T., Novikov, D. K., Popov, A. N., Mursula, A. M., Kalervo Hiltunen, J., and Wierenga, R. K.
(2004) The 1.3Å Crystal Structure of Human Mitochondrial Δ3-Δ2-Enoyl-CoA Isomerase Shows a Novel Mode
of Binding for the Fatty Acyl Group, J. Mol. Biol. 342, 1197–1208.
35. Zhang, D., Yu, W., Geisbrecht, B. V., Gould, S. J., Sprecher, H., and Schulz, H. (2002) Functional
characterization of Delta3,Delta2-enoyl-CoA isomerases from rat liver, J. Biol. Chem. 277, 9127–9132.
36. Mursula, A. M., van Aalten, D. M.F., Hiltunen, J.K., and Wierenga, R. K. (2001) The crystal structure of Δ3-
Δ2-enoyl-CoA isomerase, J. Mol. Biol. 309, 845–853.
37. Crooks, G. P., and Copley, S. D. (1994) Purification and characterization of 4-chlorobenzoyl CoA
dehalogenase from Arthrobacter sp. strain 4-CB1, Biochemistry 33, 11645–11649.
38. Kichise, T., Hisano, T., Takeda, K., and Miki, K. (2009) Crystal structure of phenylacetic acid degradation
protein PaaG from Thermus thermophilus HB8, Proteins 76, 779–786.
39. Peter, D. M., Vögeli, B., Cortina, N. S., and Erb, T. J. (2016) A Chemo-Enzymatic Road Map to the
Synthesis of CoA Esters, Molecules (Basel, Switzerland) 21, 517.
40. Krissinel, E., and Henrick, K. (2007) Inference of macromolecular assemblies from crystalline state, J. Mol.
Biol. 372, 774–797.
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
41. Liebschner, D., Afonine, P. V., Moriarty, N. W., Poon, B. K., Sobolev, O. V., Terwilliger, T. C., and Adams,
P. D. (2017) Polder maps: improving OMIT maps by excluding bulk solvent, Acta Crystallogr. D Struct. Biol.
73, 148–157.
42. Vogel, E., and Günther, H. (1967) Benzene Oxide-Oxepin Valence Tautomerism, Angew. Chem. Int. Ed.
Engl. 6, 385–401.
43. Cane, D. E., Wu, Z., and van Epp, J. E. (1992) Thiotropocin biosynthesis. Shikimate origin of a sulfur-
containing tropolone derivative, J. Am. Chem. Soc. 114, 8479–8483.
44. Rost, R., Haas, S., Hammer, E., Herrmann, H., and Burchhardt, G. (2002) Molecular analysis of aerobic
phenylacetate degradation in Azoarcus evansii, Mol. Genet. Genomics 267, 656–663.
45. Thiel, V., Brinkhoff, T., Dickschat, J. S., Wickel, S., Grunenberg, J., Wagner-Döbler, I., Simon, M., and
Schulz, S. (2010) Identification and biosynthesis of tropone derivatives and sulfur volatiles produced by bacteria
of the marine Roseobacter clade, Org. Biomol. Chem. 8, 234–246.
46. Citron, C. A., Rabe, P., and Dickschat, J. S. (2012) The scent of bacteria: headspace analysis for the
discovery of natural products, J. Nat. Prod. 75, 1765–1776.
47. Wilson, M. Z., Wang, R., Gitai, Z., and Seyedsayamdost, M. R. (2016) Mode of action and resistance studies
unveil new roles for tropodithietic acid as an anticancer agent and the γ-glutamyl cycle as a proton sink, Proc.
Natl. Acad. Sci. U S A 113, 1630–1635.
48. Raina, J.-B., Tapiolas, D., Motti, C. A., Foret, S., Seemann, T., Tebben, J., Willis, B. L., and Bourne, D. G.
(2016) Isolation of an antimicrobial compound produced by bacteria associated with reef-building corals, PeerJ
4, e2275.
49. Rasmussen, B. B., Grotkjær, T., D'Alvise, P. W., Yin, G., Zhang, F., Bunk, B., Spröer, C., Bentzon-Tilia, M.,
and Gram, L. (2016) Vibrio anguillarum Is Genetically and Phenotypically Unaffected by Long-Term
Continuous Exposure to the Antibacterial Compound Tropodithietic Acid, Appl. Environ. Microbiol. 82, 4802–
4810.
50. Porsby, C. H., Webber, M. A., Nielsen, K. F., Piddock, L. J. V., and Gram, L. (2011) Resistance and
tolerance to tropodithietic acid, an antimicrobial in aquaculture, is hard to select, Antimicrob. Agents Chemother.
55, 1332–1337.
51. Schachter, D., and Taggart, J. V. (1953) Benzoyl coenzyme A and hippurate synthesis, J. Biol. Chem. 203,
925–934.
52. Willistein, M., Haas, J., Fuchs, J., Estelmann, S., Ferlaino, S., Müller, M., Lüdeke, S., and Boll, M. (2018)
Enantioselective Enzymatic Naphthoyl Ring Reduction, Chemistry 24, 12505–12508.
53. Adams, P. D., Afonine, P. V., Bunkóczi, G., Chen, V. B., Davis, I. W., Echols, N., Headd, J. J., Hung, L.-W.,
Kapral, G. J., Grosse-Kunstleve, R. W., McCoy, A. J., Moriarty, N. W., Oeffner, R., Read, R. J., Richardson, D.
C., Richardson, J. S., Terwilliger, T. C., and Zwart, P. H. (2010) PHENIX: a comprehensive Python-based
system for macromolecular structure solution, Acta Crystallogr. D Biol. Crystallogr. 66, 213–221.
54. Emsley, P., and Cowtan, K. (2004) Coot: model-building tools for molecular graphics, Acta Crystallogr. D
Biol. Crystallogr. 60, 2126–2132.
14
ACS Paragon Plus Environment

ACS Chemical Biology
Page 20 of 21
1
2
3
4
5
6
673
674
675
55. The PyMOL Molecular Graphics System, Version 2.0 Schrödinger, LLC.
56. Dassault Systèmes BIOVIA, Discovery Studio Visualizer, v 17.2.0, San Diego: Dassault Systèmes, 2017
676
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
15
ACS Paragon Plus Environment

Page 21 of 21
ACS Chemical Biology
1
2
3
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
Proton shuttling by PaaG
(graphical abstract)
159x103mm (300 x 300 DPI)
ACS Paragon Plus Environment