Page 3 of 6
Journal of the American Chemical Society
of which catalyze the non-heme iron-dependent oxidative cleavage
or replaced with other divalent cations such as Co2+, Mg2+, Cu2+,
and Mn2+, or with Fe3+, but can be rescued by Fe2+ (Figure S10). It
is notable that no activity was observed when ChaP was incubated
with 3 in the presence of H2O2 (Figure S9), suggesting the hydrop-
eroxide anion may require flavin cofactor for delivery. In addition,
ChaP can accept aglycone 6 and monosaccharide derivative 10 as
substrates to produce 4 and 11, respectively (Figure 2D, viii−xi).
In vitro characterization of ChaP homologs. Homologs of
chaP gene are widely distributed in bacterial genomes, and are
mainly embedded in primary biosynthetic pathways (Table
S15−S22). A phylogenetic tree was constructed to include ChaP,
its homologs and other enzymes in VOC superfamily. ChaP and its
homologs were found to fall into a separate clade from other en-
zymes in VOC family (Figure S11). As none of these homologous
proteins is functionally characterized, we were curious if they cat-
alyze similar reaction. Thus, eight ChaP homologus enzymes (des-
ignated as ChaP-H1 to ChaP-H8, Figure S12) with amino acid se-
quence identities ranging from 65% to 30% were overproduced to
homogeneity (Figure S2). Intriguingly, five homologs (ChaP-H1–
ChaP-H5) can efficiently convert 3, 6 and 10 to 1, 4 and 11, respec-
tively, with comparable conversion rates to that of ChaP; whereas
three of them were inactive (Figure S13−S15). It is uncertain at this
stage why these homologous enzymes showed activities towards 3,
as no cha or similar gene clusters are encoded in their genomes.
Nonetheless, it can be speculated that these enzymes take part in
the degradation of ortho-benzoquinone-type compounds from the
environment.
1
2
3
4
5
6
7
8
of C−C bond adjacent to the vicinal hydroxyl groups in catechol
using O2 as oxidant (Figure S4).14,15 Most likely, ChaP catalyzes a
catechol 2,3-dioxygenase-like reaction to cleave the C−C bond in
ortho-benzoquinone. However, these dioxygenases require cate-
chol to coordinate with iron center, and there is no precedence that
ortho-benzoquinone, which has two electrons less than that of cat-
echol, can be accepted as substrate. We thus speculated that i) the
ortho-benzoquinone group in 3 must first be reduced by a ketore-
ductase-like enzyme to afford catechol-type compound as the sub-
2-
strate, or, ii) the two-electron rich O2 or its equivalent serves as
the oxidant instead of regular O2.
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
The first possibility can be excluded as no additional ketoreduc-
tase is required in the biotransformation from 3 to 1 as mentioned
before. For the second hypothesis, the most common O22- equiva-
lent in secondary metabolism is 4a-flavin hydroperoxide (Fl-4a-
OOH), which is derived from two successive one-electron transfers
from FlH2 to oxygen.16 We thus attempted to incubate 3 with ChaP
in the presence of NADH, FAD and flavin reductase (Fre) from E.
coli,17 followed by LC-MS analysis. Surprisingly, after 30 min of
incubation the production of 1 was observed along with two minor
compounds 7 (m/z 709.1739 [M+Na]+) and 8 (m/z 663.1675
[M+Na]+) (Figure 2D, vi). In contrast, no reaction occurred when
boiled ChaP was used (Figure 2D, ii). To obtain sufficient 7 and 8
for structure elucidation, a large-scale enzymatic reaction was per-
formed. Compound 8 was isolated and characterized as a shunt me-
tabolite (Figure 2B, Table S10). Purification of 7 was challenging
as it converted to 1 spontaneously. However, during purification, a
methyl derivative 9 with a molecular formula of C33H34O16 (m/z
723.1893 [M+Na]+) was precipitated from the methanol solution of
7. Complete 1D and 2D NMR of 9 indicated the structure of 9 is
similar to that of 1. Additional methoxyl and carboxylic acid groups
at C-2 were observed by a key HMBC correlation of O-methoxyl
group (δH 3.24) with a sp3-hybridized quaternary carbon C-2 (δC
100.6) as well as a characteristic chemical shift (δC 169.3) for a
carboxylic group. The structure of 9 was further reinforced by com-
parison of experimental and calculated 13C NMR chemical shifts
(Table S11). The structure of 7 is thus proposed to have a hemiketal
group at C-2 (Figure 2A).
Crystal structures of ChaP and two homologs. To further in-
vestigate the mechanistic details of the unusual reaction, the struc-
tures of ChaP and two homologous proteins ChaP-H1 from Rhodo-
coccus phenolicus and ChaP-H2 from Streptomyces curacoi were
solved at 1.70 Å, 2.00 Å and 1.63 Å resolution, respectively (Table
S23). ChaP-H1 and ChaP-H2 share 65% and 61% amino acid se-
quence identity with ChaP, respectively. Analysis of the crystal
structures revealed a symmetric homodimer with each monomer
featuring two similar βαββ motifs (Figure 3A−3C). These enzymes
are structurally highly similar to each other with r.m.s. deviations
of 0.8−1.0 Å (Figure 3D). The location of two narrow symmetric
active pockets, which likely accommodate the plate-shaped sub-
strate, is defined by the presence of two FeII ion binding sites in the
interface of the two monomers (Figure 3). FeII ion is coordinated
by one tyrosine residue (Y125), one glutamate residue (E119), a
pair of histidine residues (H7 and H63), and two water molecules
with an average distance of 2.23 Å (Figure S17). Moreover, one of
the histidine residues (H7) is from the other monomer.
Docking study and site-directed mutagenesis of ChaP. Three-
dimensional structure searching using the Dali server18 revealed
that ChaP homodimer shares similarity to BphC (PDB: 1KW8)
from Pseudomonas sp. in terms of FeII binding sites with r.m.s. de-
viations of 2.3 Å (Figure S18),19 albeit extremely low sequence
identity (<5%). BphC is a well characterized 2,3-catechol dioxy-
genase in VOC enzyme family, which catalyze the non-heme FeII-
dependent oxidative cleavage of catechols to 2-hydroxymuco-
naldehyde products.19 Docking of the ChaP with 10 and hydroper-
oxide anion (HOO-) was conducted by the Induced Fit Docking
workflow.20 The final docking poses were selected using BphC-
substrate-NO (nitric oxide) complex structure (PDB: 1KW8) as ref-
erence. As shown in Figure 3F and Figure S19, the substrate coor-
dinated with FeII through ortho-quinone group in a narrow groove,
and was clamped by Q91, Y92, Y109, R102 from one side, and F37
and R54 from another side; the HOO- was surrounded by D49, F37
and Y125. The big aromatic π-electron system of the substrate has
strong T-shaped π-π, amide-π, σ-π and cation-π interactions with
F37, Y109, Q91 and R102. Importantly, two hydrogen bonds were
observed between D49 and HOO-, and Y109 and O-14 of substrate
10.
The identification of 7 raised the question if it is a biosynthetic
intermediate from 3 to 1. We thus conducted a time-course analysis
for the ChaP-catalyzed reaction. The production of 7 increased at
early time point, which finally converted to 1 when substrate 3 was
consumed (Figure 2D, iii−vii). As 7 can also be converted to 1
spontaneously, we further measured the effect of ChaP concentra-
tion on the rate of reaction when 9 was used as substrate, which can
converted to 7 in aqueous solution. The production of 1 increased
with increasing the concentrations of ChaP (Figure S6). Moreover,
the conversion of 9 to 1 required the presence of all co-enzyme and
cofactors as omitting anyone of them led to dramatically decrease
conversion rates (Figure S6). Taken together, these data firmly es-
tablish that ChaP is responsible for two successive steps in conver-
sion of 3 to 1 through the intermediate 7.
Next, we investigated the effect of various co-enzymes and co-
factors for ChaP activity. Four flavin reductases (Figure S7) cloned
from the genome of S. chartreusis NA02069 were overproduced
and purified as bright yellow proteins, suggesting all of them carry
flavin cofactors FAD or FMN. The identities of these confactors
were determined by comparison of authentic FAD or FMN stand-
ards by LC/MS analysis (Figure S8). Comparable amounts of 1
were produced by replacing Fre with either of these four flavin re-
ductases (Figure S9). No products were observed when either Fre
or NADH was omitted from reaction system (Figure S9). The con-
version rate of 1 was significantly decreased when NADH was sub-
stituted with NADPH (Figure S9). Moreover, the activity of ChaP
was almost lost when metal ion in ChaP was removed by EDTA,
ACS Paragon Plus Environment
Wang, Yi Shuang
