An Unusual Flavin-Dependent Halogenase from the Metagenome of the Marine Sponge Theonella swinhoei WA

Article abstract of DOI:10.1021/acschembio.6b01115

Uncultured bacteria from sponges have been demonstrated to be responsible for the generation of many potent, bioactive natural products including halogenated metabolites.1 The identification of gene clusters from the metagenomes of such bacterial communities enables the discovery of enzymes that mediate new and useful chemistries and allows insight to be gained into the biogenesis of potentially pharmacologically important natural products. Here we report a new pathway to the keramamides (krm); the first functional evidence for the existence of a distinct producer in the Theonella swinhoei WA chemotype is revealed, and a key enzyme on the pathway, a unique flavin-dependent halogenase with a broad substrate specificity, with potential as a useful new biocatalytic tool, is described.

Full text of DOI:10.1021/acschembio.6b01115

Subscriber access provided by University of Florida | Smathers Libraries
Article
An unusual flavin-dependent halogenase from the
metagenome of the marine sponge Theonella swinhoei WA
Duncan R.M. Smith, Agustinus R. Robert Uria, Eric J. N. Helfrich, Daniela
Milbredt, Karl-Heinz van-Pee, Joern Piel, and Rebecca J. M. Goss
ACS Chem. Biol., Just Accepted Manuscript • DOI: 10.1021/acschembio.6b01115 • Publication Date (Web): 15 Feb 2017
Downloaded from http://pubs.acs.org on February 17, 2017
Just Accepted
“Just Accepted” manuscripts have been peer-reviewed and accepted for publication. They are posted
online prior to technical editing, formatting for publication and author proofing. The American Chemical
Society provides “Just Accepted” as a free service to the research community to expedite the
dissemination of scientific material as soon as possible after acceptance. “Just Accepted” manuscripts
appear in full in PDF format accompanied by an HTML abstract. “Just Accepted” manuscripts have been
fully peer reviewed, but should not be considered the official version of record. They are accessible to all
readers and citable by the Digital Object Identifier (DOI®). “Just Accepted” is an optional service offered
to authors. Therefore, the “Just Accepted” Web site may not include all articles that will be published
in the journal. After a manuscript is technically edited and formatted, it will be removed from the “Just
Accepted” Web site and published as an ASAP article. Note that technical editing may introduce minor
changes to the manuscript text and/or graphics which could affect content, and all legal disclaimers
and ethical guidelines that apply to the journal pertain. ACS cannot be held responsible for errors
or consequences arising from the use of information contained in these “Just Accepted” manuscripts.
ACS Chemical Biology 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 11
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
KrmI, a halogenase with broad substrate specificity, identified from the metagenome of a sponge and linked
to a novel keramamide biosynthetic pathway
254x142mm (72 x 72 DPI)
ACS Paragon Plus Environment

ACS Chemical Biology
Page 2 of 11
1
2
3
4
5
6
7
8
An unusual flavin-dependent halogenase from the metagenome of the marine sponge
Theonella swinhoei WA
9
Duncan R. M. Smitha, d, Agustinus R. Uriab, e, Eric J. N. Helfrichb, Daniela Milbredtc, Karl-Heinz van Péec, Jörn
Piel*b, Rebecca J. M. Goss*a
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
aSchool of Chemistry, University of St Andrews, St Andrews, KY16 9ST, United Kingdom
b Institute of Microbiology, Eidgenössische Technische Hochschule Zürich, Vladimir-Prelog-Weg 4, 8093 Zurich, Switzerland
c Technische Universität Dresden, 01062 Dresden, Germany
d currently researching at The Manchester Institute of Biotechnology(MIB) The University of Manchester, M1 7DN
e currently researching at The Research and Development Center for Marine and Fisheries Product Competitiveness and Biotechnology.
Jalan KS Tubun Petamburan VI, Jakarta 10260, Indonesia.
rjmg@st-andrews.ac.uk
RECEIVED DATE (automatically inserted by publisher);
Abstract: Uncultured bacteria from sponges have been
demonstrated to be responsible for the generation of many
potent, bioactive natural products including halogenated
metabolites.1 The identification of gene clusters from the
metagenomes of such bacterial communities enables the
discovery of enzymes that mediate new and useful
chemistries and allows insight to be gained into the
biogenesis of potentially pharmacologically important
natural products. Here we report a new pathway to the
keramamides (krm); the first functional evidence for the
existence of a distinct producer in the Theonella swinhoei WA
chemotype is revealed, and a key enzyme on the pathway, a
unique flavin dependent halogenase with a broad substrate
specificity, and with potential as a useful new biocatalytic
tool is described.
there is an increasing demand for the discovery and
development of halogenases as new tools for (bioꢀ)catalysis.
In contrast to synthetic approaches, halogenases are able to
regioselectively introduce halogens under mild and green
conditions without generating harmful waste. As a result,
halogenases are attracting much attention.7–15
Introduction: Sponges have been shown to be a rich source of
bioactive molecules, including halometabolites, and an
exciting treasure trove in which to search for new enzymes.
Halogenation represents an important strategy for tuning and
improving bioavailability and bioactivity of molecules, and is
attracting increasing interest from academia and industry
alike. Over 4000 halogenated natural products have been
identified to date, many of which exhibit potent
bioactivities.1 Keramamides and related compounds (Figure
Figure 1 Symbiontꢀderived natural products containing the
haloꢀhydroxytryptophan motif (highlighted in blue) that have
been identified from extracts of T. swinhoei chemotypes.
Two distinct halohydroxytryptophan motifs are observed: 2ꢀ
bromoꢀ5ꢀhydroxytryptophan
hydroxytryptophan. Keramamide H (
konbamide ( ), orbiculamide A ( ).
and
6ꢀchloroꢀ5ꢀ
),
1
) are a unique series of halogenated marine nonꢀribosomal
peptide natural products that have been shown to be
cytotoxic against murine and human cancer cell lines.2
1
), keramamide A (
2
3
4
Due to conferring favorable properties, halogenation is used
prolifically by both agrochemical and pharmaceutical
industries, with around 20% of small molecule drugs and
30% of agrochemical compounds being halogenated.3 In
2013, the topꢀrevenue drug in the USA was the 2,3ꢀ
Keramamides contain rare halohydroxytryptophans motifs.
Notably, the marine sponge Theonella swinhoei has been
identified as the source of several natural products that
contain this moiety, including konbamide,14 and
dichlorophenylꢀcontaining
antipsychotic
aripiprazole
orbiculamide15 (e.g., 1–4
, Figure 1) as well as the
(Abilify).4 Furthermore, the introduction of halogens into
molecules can provide sites for selective functionalization
through crossꢀcoupling chemistries.5 Synthetic conditions for
halogenation are often harsh and nonꢀregioselective, utilize
noxious reagents, or generate undesirable byꢀproducts6 and
keramamides.2,16 Intriguingly, distinct variants of these
compounds occur in different chemotypes of the sponge.
Orbiculamide
A
(
4
)
(containing
a
2ꢀbromoꢀ5ꢀ
hydroxytryptophan moiety) is found in the chemotype T.
swinhoei Y, which is named to reflect the characteristic
ACS Paragon Plus Environment

Page 3 of 11
ACS Chemical Biology
yellow color of the sponge interior (Figure 1).17 T. swinhoei
chemotype WA (white interior from Hachijo Island, Japan)
was not previously reported to contain such peptidic natural
products.
matched those of known keramamides, others are unique and
observed for the first time.
1
2
3
4
5
6
RESULTS AND DISCUSSION
7
8
A new krm gene cluster from T. swinhoei WA
By identifying the biosynthetic genes, it was recently
demonstrated that keramamideꢀtype peptides and many other
natural products isolated from T. swinhoei Y are produced by
an asꢀyetꢀuncultured bacterial symbiont of the sponge,
Candidatus ‘Entotheonella factor’ TSY1.17 The sponge T.
swinhoei WA examined in the current study was also
previously shown to contain filamentous ‘Entotheonella’
symbionts.17 In this sponge, ‘Entotheonella serta’ TSWA1 is
the source of the homodimeric polyketide actin inhibitor
misakinolide A.18 Here, using an enriched ‘Entotheonella’
cell fraction prepared from T. swinhoei WA, PCR analysis
with primers specific for adenylation (A) domainꢀencoding
regions of nonribosomal peptide synthetase (NRPS) genes
revealed the presence of homologous genes. These were
closely related to those of the keramamide (ker) biosynthetic
genes previously identified in ‘E. factor’ associated with T.
swinhoei Y. To isolate the corresponding genomic locus, a
fosmid library was constructed from the enriched
‘Entotheonella’ cell fraction of the WA chemotype and
screened, resulting in the isolation of fosmid pADU2a.
Sequencing of his fosmid confirmed the presence of NRPS
and polyketide synthase (PKS) genes sharing high similarity
and architectural features with the ker cluster (Figures 2 and
S1, Table S3). Notable differences of the krm biosynthetic
gene cluster to the ker system, included krmR, and krmI.
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. (A) The previously identified keramamide (ker
)
biosynthetic gene cluster from T. swinhoei Y. (B) Newly
identified krm keramamideꢀtype gene cluster from T.
swinhoei WA containing the unusual 5ꢀhydroxyꢀ6ꢀ
halotryptophan gene, (C) Biosynthetic model for the ker
pathway. (D) Predicted function of the krm pathway, in
which the halogenase KrmI functionalizes the 6 position of
the indole. (See table S1 for further information on the genes
Keramamides identified from T. swinhoei WA exhibit an
methylserine extension as present in keramamide H (
Figure 1), and krmR is predicted to encode an NRPS module
activating and incorporating ꢀmethylserine. The Cꢀterminal
Oꢀ
1
,
identified in the krm pathway.14
)
O
region of KrmI resembles flavinꢀdependent tryptophan
halogenases (FDTHs). FDHs are typically found encoded
within a natural product biosynthetic pathway gene cluster,
and many are known to halogenate indole, phenol, and
pyrrole motifs. Though halotryptophans occur fairly
frequently in natural products, such halohydroxytryptophans
are far less common.
The detection of more than a dozen different keramamide
congeners in both chemotypes might indicate an inherent
flexibility of the krmꢀencoded adenylation domains, often
observed for NRPS pathways, which seem to accept series of
similar amino acids. From the comparative MS network
analysis, the position of halogenation cannot be determined
Challenged by the limited amounts of material available, as
NMR analysis of the WA keramamides would not be a
possible method for us to determine the regiochemistry of the
halogenation, we next set out to directly investigate the
unusual associated halogenase KrmI.
Challenged by the low concentration of keramamides in the
sponge tissue, which is itself of limited availability, and the
current unculturable nature of ‘Entotheonella’ from either T.
swinhoei WA or Y1, we set out to use comparative MS
analysis to analyze the chemistry produced by the sponge
microbiomes and to glean evidence for the activity of the
newly identified krm gene cluster. Strikingly, our analysis
revealed the presence of numerous keramamide congeners
from each system. Figures S2 and S3 show a molecular
network analysis of extracts from both chemotypes. Whilst
the MS2 characteristics of some T. swinhoei WA compounds
Bioinformatic analysis of krmI
Bioinformatic analysis of the gene krmI at the translated
amino acid level using BLAST19 and HHPred20
(SI Tables
S2 and S3) indicated that the Cꢀterminus of the protein KrmI
is homologous to known halogenases including RebH (30%
identity, 48% similarity over homologous region), PrnA
(30% / 47%), and CmdE (21% / 35%), and contains the wellꢀ
characterized sequence motifs for FDTHs ꢀ GxGxxG and
ACS Paragon Plus Environment

ACS Chemical Biology
Page 4 of 11
WxWxIP ꢀ as well as mechanistically relevant residues the presence of a polyhistidine tag at either terminus had only
1
2
3
4
5
6
7
8
(Lys443 and Glu696 align to Lys79 and Glu346 from PrnA)
(SI figure S3).21–23 Unusually, KrmI appears to include a
second, Nꢀterminal domainꢀlike region with distant
homology to members of the ThiF enzyme family (Pfam
family ThiF PF00899). Typically, these enzymes utilize ATP
to activate carboxylic acids by adenylation.24,25 Whilst
sequence identity to characterized protein sequences is low,
the alignment revealed that in KrmI, this domain is truncated
and the majority of residues involved in substrate and coꢀ
factor binding are likely to be absent, although we must be
cautious making definite claims based on alignments of such
distantly related proteins. Only one other sequence
displaying this unusual combination of truncated ThiFꢀlike
and FDTH domains has been deposited in the NCBI nonꢀ
redundant protein sequence database; this hypothetical
protein is uncharacterized and no information on the context
of this gene is available.26 The ThiF family is similar to the
E1ꢀlike proteins that have been implicated in the specific
binding of peptide substrates in ribosomal peptide natural
product biosynthesis,27,28 and it is possible that the Cꢀ
terminal domain serves a similar function here.
a slightly detrimental effect on solubility. A significant
proportion of the soluble protein did not bind to nickel
nitrilotriacetic acid (NiNTA) resin during immobilized metal
affinity purification (IMAC), indicating that the polyhistidine
tag may have low accessibility. IMAC purification of this
enzyme is lowꢀyielding (approximately 1 mg protein per L E.
coli culture). We demonstrated, however, that the crude cell
lysate (CCL) containing KrmI could be readily utilized as a
biocatalyst, without any detrimental effect on the reaction,
and therefore KrmI was used as a component of the CCL
throughout the majority of the work reported here.
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
Halogenation assays & product characterization
The gene was shown to encode a functional halogenase as in
vitro assays demonstrated that the fullꢀlength protein was
able to convert 5ꢀhydroxytryptophan (5HTP, 5) to a single
new product when the enzyme was utilized as a component
of a CCL (Scheme 1). LCꢀHRMS was used to identify the
accurate mass of this product, suggesting a chemical formula
of C11H11N2O2Cl and indicating that 5HTP (
5) had been
regioselectively chlorinated by KrmI ([M+H]+
=
C11H12N2O3Cl calculated: 255.0536; measured: 255.0532 ꢁ:
+0.2 ppm). Use of the truncated protein resulted in only trace
conversion to 6Cl5HTP (Figure 4C), and the purified
enzymes also showed greatly reduced activity (Figure 4D
and 4G) compared to the fullꢀlength protein. Hence it was
decided to utilize KrmI as a component of the CCL for the
rest of this study.
Figure 3. A graphical representation of KrmI with its two
enzymatic domains in grey (top, numbered by translated
amino acid position), and homologous regions of Pfam
protein families PF00899 (ThiF, green) and PF04820
(PrnA Trp Halogenase, red) that were identified using
HHpred.20 Similarity between KrmI domains and the PFAM
representatives to which they are compared are indicated by a
greyꢀscale similarity representation with black showing high
levels of similarity and light grey showing low levels of
similarity.
Scheme 1. KrmI converts 5HTP (5) to 6Cl5HTP (6). A
Production of KrmI
simplified catalytic cycle for the required coꢀfactor FAD/H2,
which is generated in situ by the flavin reductase PrnF, is
shown.
To determine whether krmI encodes an active halogenase,
and whether both domains are necessary for activity, two
expression constructs were prepared. The first encoded the
fullꢀlength gene, whilst the second encoded only the Cꢀ
terminal domain (truncated to the first inꢀframe ATG codon
preceding the FDTH domain). Both constructs produced little
soluble protein when expressed in Escherichia coli unless the
chaperone proteins GroEL/S, DnaK/J and GrpE were coꢀ
expressed. Though the production of many other halogenases
in heterologous expression hosts do not require the use of
chaperones;29–32 chaperones GroEL/S has previously been
reported to improve the solubility and yield of RebH when
produced in E. coli 11
.
Coexpression of genes encoding GroEL/S, DnaK/J and GrpE
along with KrmI resulted in adequate levels of production of
soluble KrmI (Figures S6 and S7). Attempts to produce
either protein as an Nꢀterminal maltose binding protein
(MBP) fusion resulted in complete loss of solubility, whilst
ACS Paragon Plus Environment

Page 5 of 11
ACS Chemical Biology
6Cl5HTP (
keramamide A (
6
) (Scheme 1), a likely biosynthetic precursor to
) and other congeners.
1
2
2
3
4
Halide and substrate scope
Having determined that KrmI is able to chlorinate 5HTP, we
next sought to explore whether it could also catalyze
bromination or iodination of the same substrate. As the
enzyme was used in the form of a CCL, it was not practical
to completely eliminate endogenous sodium chloride from
the preparation; however, we found that the enzyme was able
to tolerate the addition of salt to the CCL. Therefore, we
performed assays in the presence of additional sodium
chloride, sodium bromide, and potassium iodide. As is
typical for FDTH,30,33,35 KrmI is able to utilize chloride and
bromide but not iodide, with reduced activity when
incorporating bromide (Figures 5 and 6).
5
6
7
8
9
10
11
12
13
14
15
16
17
18
Figure 4. Halogenation of 5HTP (5) achieved utilizing
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
various KrmI preparations. The conversion of 5HTP to
6Cl5HTP, as monitored by UPLC, at an absorbance at 300
nm is depicted. Assay conditions are described in the SI. (A)
5HTP (5) standard, (B) Fullꢀlength KrmI as a component of a
CCL (0.6 mU/mL). (C) truncated Cꢀterminal domain of
KrmI as a component of the CCL (trace activity). (D)
Purified fullꢀlength KrmI (20 µM, 0.1 mU/mL). (E) Purified
fullꢀlength KrmI + E. coli CCL (20 µM, 0.3 mU/mL), (F)
Purified fullꢀlength KrmI + catalase + SOD (20 µM, 0.6
mU/mL). (G) Purified Cꢀterminal domain (20 µM, <0.1
mU/mL). (H) Purified Cꢀterminal domain + catalase + SOD
(20 µM, <0.1 mU/mL). 1 mU is the amount of enzyme that
converts 1 nmol of 5HTP (
under assay conditions.
5) to 6Cl5HTP (6) per minute
Figure 5. UPLC chromatograms (UV absorbance at 300 nm)
monitoring the products of 5HTP (5) halogenation by KrmI
Exclusively in the case of the fullꢀlength purified protein, it
was possible to restore activity by supplementing with either
a CCL prepared from E. coli expressing an empty vector
in the presence of NaCl (A), NaBr (B), and KI (C). A small
amount of NaCl is present in all assays, leading to the
production of 6Cl5HTP. In the presence of bromide, an
additional new product is formed, which was determined to
be 6Br5HTP.
(
Figure 4E) or with catalase and superoxide dismutase
(SOD) (Figure 4F), implying that the protein may be harmed
by reactive oxygen species (ROS) generated in situ. Other
FDTHs reported in the literature have been active upon
purification without the need for the inclusion of any further
additives to counter the effect of ROS.29,30,33
In order to fully characterize the product and identify the
regiochemistry of halogenation, the reaction was scaled up to
10 mg, which allowed the product to be purified and
1
1
analyzed by H NMR. As the aromatic region of the H and
COSY NMR spectra were devoid of coupling, it could be
concluded that halogenation has taken place at position 6 and
the three observed singlet aromatic resonances correspond to
the protons at positions 2, 4, and 7 of the indole ring. 6ꢀ
chloroꢀ5ꢀhydroxytryptophan (6Cl5HTP) has previously been
described in the literature,34 and the 1H NMR
characterization agrees with that reported here. The single
product generated by the KrmIꢀcatalyzed regioselective
Figure 6. Product distribution of 5HTP (5) halogenation by
KrmI in the presence of varying additional concentrations of
Clꢀ and Brꢀ. UPLC peak areas (300 nm) are plotted as
normalized values. 6Cl5HTP: filled circles, 6Br5HTP: open
circles, total 6ꢀhaloꢀ5ꢀHTP: crosses & dashed line. Error bars
indicate the standard deviation of triplicate measurements
and are in many cases obscured by the data markers.
chlorination of 5HTP (5) was therefore confirmed as
ACS Paragon Plus Environment

ACS Chemical Biology
Page 6 of 11
1
2
3
4
5
6
7
8
The brominated product was isolated from a scaledꢀup
reaction and characterized by H NMR, which indicated that
1
regioselectivity had been retained and the product was 6ꢀ
bromoꢀ5ꢀhydroxytryptophan (6Br5HTP). By combining
sodium bromide and sodium chloride in varying ratios in the
assay, we were able to determine that bromination occurred
in preference to chlorination despite chloride being present in
higher concentrations and a decrease in overall conversion of
the starting material (Figure 6).
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
FDTH are known to halogenate a range of substrates,11,12,32
typically tryptophan derivatives and other aromatic moieties.
To initiate a study of the substrate scope for this unusual
halogenase we first explored whether the enzyme might
halogenate tryptophan and serotonin, as they are closely
related to 5HTP. Indeed, KrmI was able to halogenate
serotonin and 5HTP with approximately equal efficiency in
terms of time and conversion levels, whilst tryptophan was
less well tolerated as a substrate (Figure 7). To our
knowledge, the only other FDTH with a preference for a
substituted tryptophan is KtzR, which installs chlorine at the
6ꢀposition of 7ꢀchlorotryptophan in the biosynthesis of
kutzneride.36
Encouraged by these initial results, we extended our study of
the substrate scope of KrmI. Again we utilized crude lysate
containing the fullꢀlength protein in incubations with a
variety of potential substrates (Figure 8). We observed good
levels of conversion for the substrates Lꢀ5HTP (
5
), serotonin
Figure 7 UPLC chromatograms of the halogenation of 5HTP
(( ), 300 nm), tryptophan (( ), 280 nm), and serotonin ((C),
300 nm). The identities of the labelled peaks have been
confirmed by LCꢀHRMS and NMR.
(
10), and 5ꢀ methoxyindole (18); and lower levels of
A
B
conversion for Lꢀtryptophan (8), Dꢀtryptophan (12) and 5ꢀ
aminoindole (19). We determined that Lꢀ7ꢀfluorotryptophan
13), Lꢀ7ꢀhydroxytryptophan (14), 4ꢀhydroxyindole (15), 5ꢀ
(
hydroxyindole (16), 5ꢀhydroxyindoleꢀ3ꢀacetic acid (17), 5ꢀ
formylindole (20), and 7ꢀazaindole (21) could also be
processed by KrmI, but at trace level as indicated by LCꢀMS
measurements. All substrates were converted to a single new
monoꢀchlorinated product, with the exception of 5ꢀ
aminoindole (19), which was converted into two halogenated
products (LCꢀHRMS analysis of all halogenated substrates is
included in the SI). This indicates that KrmI is able to
regioselectively halogenate a range of substrates, although
due to low levels of product, the position of halogenation has
not been confirmed in all cases. The acceptance of 5ꢀ
Conclusions
Through metabolite extraction and MS network analysis we
show that T. swinhoei WA (white chemotype), like its yellow
counterpart, T. swinhoei Y produces keramamides. Whilst
the keramamide series of compounds produced by T.
swinhoei WA and Y partially overlap, they also show distinct
chemotypeꢀspecific characteristics. The discovery of the new
biosynthetic gene cluster (e.g. krmI) from the metagenome of
the sponge T. swinhoei WA explains the difference in the
suites of compounds produced in each system. The krm
genes, though showing homology to the previously identified
ker gene cluster, exhibit distinct differences; a key difference
is the presence of krmI. Biooinformatic analysis of krmI
indicated it to be a flavinꢀdependent halogenase. Functional
hydroxyindole
(
16), 5ꢀmethoxyindole
(
18
)
and 5ꢀ
aminoindole (19) but not of 5ꢀfluoroindole (30) or 5ꢀ
fluorotryptophan (generated in situ from 30) may suggest
that specificity within the active site could arise from the
requirement for a hydrogen bond acceptor within the ring,
rather than steric requirements, as 5ꢀfluorotryptophan is
isosteric to the preferred substrate 5HTP. There is clearly
considerable substrate flexibility as 5ꢀformylindole (20) and
analysis of the encoded protein product, revealed
a
catalytically active halogenase with an in vitro substrate
preference for 5ꢀhydroxytryptophan over tryptophan,
generating 6ꢀhaloꢀ5ꢀhydroxytryptophan as
a
single
4ꢀhydroxyindole
(
15
)
were
also
accepted.
The
regioisomer, indicating that this halogenase operates to
generate the 6ꢀhalo containing keramamides. It is likely that
a second, as yet unidentified halogenase acts to generate the
accommodation of both Lꢀ and Dꢀtryptophan (
8
and 12
)
demonstrates that this hydrogen bonding interaction is not an
absolute requirement, and that interactions with the amino
acid moiety are flexible.
ACS Paragon Plus Environment

Page 7 of 11
ACS Chemical Biology
2ꢀhalo containing keramamides observed within the T.
1
2
swinhoei Y system.
3
4
5
6
7
8
9
KrmI is the first halogenase identified to use Lꢀ5ꢀ
hydroxytryptophan as its preferred substrate. Furthermore
another unique characteristic is the unusual Nꢀterminal
region of the protein. Though no function has been assigned
to the Nꢀterminal domain, in our hands, the monoꢀdomain
truncation demonstrated considerably lower activity
compared to its fullꢀlength counterpart. Our initial
investigations of KrmI as a potential tool for biocatalysis
demonstrate that it has promisingly broad substrate
specificity. KrmI can accommodate a range of indoleꢀderived
substrates. Future work will include crystallographic analysis
which may provide insight into substrate flexibility and
informed engineering of the enzyme to further develop it for
biotechnological applications. KrmI is an efficient catalyst
for the production of halogenated 5HTP and serotonin, and
due to its inherent substrate flexibility presents a promising
starting point for the generation of a range of chlorinated and
brominated products following protein engineering.
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
It is notable that halogenation was almost exclusively
observed at the 6ꢀposition of the indole ring in these studies,
whilst in the keramamide congeners halogenation is observed
at both the 2ꢀ and the 6ꢀpositions. It seems likely, therefore,
that there is either another halogenase in the keramamide
gene clusters that is responsible for halogenation at the 2ꢀ
position of the indole ring, or KrmI possesses an activity in
vivo which is not observed in vitro. It is tempting to postulate
that the Cꢀterminal domain may be responsible to binding a
peptide substrate, and that the true activity of KrmI would
not be observed when the enzyme acts on a free amino acid
substrate rather than a peptide substrate. The instability of 2ꢀ
halotryptophans, in which loss of halogen is promoted by the
free carboxylate or even amine positioning over the indole
ring37,38 makes it likely that halogenation at the 2ꢀposition
does not occur on free tryptophan substrates. Such
halogenation is more likely to occur once the tryptophan
substrate is integrated into a peptide, where both Nꢀ and Cꢀ
termini are functionalized as amides.
Figure 8. Substrate specificity of KrmI. Substrates shown
above the bold line were halogenated, whilst those shown
below the bold line were not. Substrates 13 and 14 were
generated in situ from the corresponding indole by the native
E. coli tryptophan synthase. The substrates that were
This study adds further functional evidence to the role of
symbiotic bacteria in the production of spongeꢀderived
natural products. Whilst keramamides have previously been
assigned to ‘E. factor’ in the chemotype T. swinhoei Y,17 the
source of the halogenated congeners in T. swinhoei WA
remains unknown. In another study of the WA chemotype,
the polyketide misakinolide A was assigned to the symbiont
‘E. serta’ TSWA1 by singleꢀcell genome sequencing;18
however, analysis of that sequence data did not reveal the
presence of keramamide genes, suggesting that these
compounds are likely to be produced by either a different
‘Entotheonella’ phylotype or another symbiotic bacterium
found in the sponge.
halogenated include: Lꢀ5HTP
serotonin (10), Dꢀtryptophan (12), 7ꢀfluorotryptophan (13),
7ꢀhydroxytryptophan 14), 4ꢀhydroxyindole 15), 5ꢀ
(
5
), Lꢀtryptophan
(8),
(
(
hydroxyindole (16), 5ꢀhydroxyindoleꢀ3ꢀacetic acid (17), 5ꢀ
methoxyindole (18), 5ꢀaminoindole (19), 5ꢀformylindole
(
20), and 7ꢀazaindole (21). The enzyme did not process: Lꢀ
phenylalanine (22), Lꢀtyrosine (23), DLꢀmetatyrosine (24),
Lꢀhistidine (25), indole (26), 6ꢀhydroxyindole (27), 7ꢀ
hydroxyindole (28), 4ꢀfluoroindole (29), 5ꢀfluoroindole (30),
6ꢀfluoroindole (31), 7ꢀfluoroindole (32), 4ꢀchloroindole (33),
5ꢀchloroindole (34), 6ꢀchloroindole (35), 7ꢀchloroindole (36),
2ꢀformylindole (37), 7ꢀformylindole (38), 5ꢀnitroindole (39
and 7ꢀnitroindole (40).
)
METHODS
ACS Paragon Plus Environment

ACS Chemical Biology
Isolation of krm genes from the metagenome of T. swinhoei
Page 8 of 11
1
2
3
4
5
6
7
8
WA
Full experimental methods, supporting figures S1 to S9,
supporting tables S1 to S4 and characterization of
isolated products
NMR spectra
LC-HRMS chromatograms
DNA was isolated from
a
previously described17
‘Entotheonella’ꢀenriched bacterial fraction of T. swinhoei
WA following a procedure adapted from Syn and Swarup.
PCR using degenerate primers was used to amplify gene
fragments corresponding to NRPS A domains, which were
purified, ligated into pBlueScript II (SKꢀ) (Stratagene) and
sequenced. A second round of PCR using specific primers
enabled identification of the fosmid pADU2a containing the
krm cluster.
9
Acknowledgements
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
We thank S Matsunaga for providing a sample of T. swinhoei
WA; and the University of St Andrews United College
scholarship (DRMS), EU BlueGenics (Seventh Framework
Programme, Collaborative project “BlueGenics”, Grant no.
311848 RJMG and JP), the SNF (Grant no.
205321_165695 to JP), the Helmut Horten Foundation
(JP), and ERAIB (Grant no. 031A338A KHVP and RJMG)
for funding.
Metabolic analysis
Metabolites were extracted from T. swinhoei Y and WA
tissue samples and analyzed by LCꢀMSꢀMS using a Thermo
QExactive mass spectrometer operating in positive ionization
mode and using dataꢀdependent acquisition of fragmentation
spectra. This data was processed using the Global Natural
Products Social Molecular Networking web server
(www.gnps.ucsd.edu). Novel keramamide analogues were
assigned based on spectral similarity to known keramamide
analogues.
References
(1) Gribble, G. W. (2010) Naturally Occurring
Organohalogen Compounds ꢀ A Comprehensive Update.
Springer, Vienna.
Vector preparation
(2) Kobayashi, J., Itagaki, F., Shigemori, I., Takao, T., and
Shimonishi, Y. (1995) Keramamides E, G, H, and J, new
cyclic peptides containing an oxazole or a thiazole ring
from a Theonella sponge. Tetrahedron 51, 2525–2532.
(3) HerreraꢀRodriguez, L. N., Khan, F., Robins, K. T., and
Meyer, H.ꢀP. (2011) Perspectives on biotechnological
halogenation Part I: Halogenated products and enzymatic
halogenation. Chim. Oggi 29, 31–33.
A PCR product containing krmI and appropriate restriction
sites was ligated into a vector derived from pET21a using
standard techniques. The sequence of vectors was confirmed
by restriction digest and sequencing.
Protein production
E. coli BL21ꢀDE3, expressing chaperone genes from the
plasmid pGꢀKJE8 (TaKaRa Bio) was transformed with the
vector encoding KrmI. An overnight culture of this was used
to inoculate LB medium containing appropriate antibiotics
and chaperone inducers. This was grown at 37 °C to an OD
of 0.6, then cooled to 16 °C and induced by addition of
IPTG. The cells were harvested by centrifugation and lysed
by sonication. The crude lysate was clarified by
centrifugation. The protein was purified using NiꢀNTA
affinity chromatography, eluting with imidazole. Proteins
were analyzed by denaturing polyacrylamide gel
electrophoresis according to standard methods.
4 Top 100 Drugs for 2013 by Sales ꢀ U.S. Pharmaceutical
Statistics,
(accessed July 2015).
https://www.drugs.com/stats/top100/sales,
(5) Mahoney, K. P. P., Smith, D. R. M., Bogosyan, E. J. A.,
and Goss, R. J. M. (2014) Access to High Value Natural
and Unnatural Products through Hyphenating Chemical
Synthesis and Biosynthesis. Synthesis 46, 2122–2132.
(6) Eissen, M., and Lenoir, D. (2008) Electrophilic
Bromination of Alkenes: Environmental, Health and Safety
Aspects of New Alternative Methods. Chem. – Eur. J. 14,
9830–9841.
(7) Grüschow, S., Smith, D. R. M., Gkotsi, D. S., and Goss,
R. J. M. (2014) Halogenases, in Biocatalysis in Organic
Synthesis 3 (Faber, K., Fessner, W.ꢀD., and Turner, N. J.,
Eds.), pp 313–359. Thieme, Stuttgart.
(8) Smith, D. R. M., Grüschow, S., and Goss, R. J. M.
(2013) Scope and Potential of Halogenases in Biosynthetic
Applications. Curr. Opin. Chem. Biol. 17, 276–283.
(9) Brown, S., and O’Connor, S. E. (2015) Halogenase
Engineering for the Generation of New Natural Product
Analogues. ChemBioChem 16, 2129–2135.
Halogenation assays
Assays containing purified KrmI or CCL containing KrmI
along with PrnF, FAD, substrate, NaCl, and NADH,
(optionally also superoxide dismutase and catalase) in a final
volume of 0.1 mL, were incubated at 28 °C for 90 minutes
then quenched by addition of acidic methanol. Samples were
analyzed by UPLC or LCMS. For preparation of purified
substrates, the assays were scaled up to 10 mL and purified
by semiꢀpreparative HPLC.
(10) Timmins, A., and de Visser, S. P. (2015) Enzymatic
Halogenases and Haloperoxidases: Computational Studies
on Mechanism and Function. Adv. Protein Chem. Struct.
Biol. 100, 113–151.
(11) Payne, J. T., Andorfer, M. C., and Lewis, J. C. (2013)
Regioselective Arene Halogenation using the FADꢀ
Supporting Information
The Supporting Information is available free of charge on the
ACS Publications website at DOI:
ACS Paragon Plus Environment

Page 9 of 11
ACS Chemical Biology
Dependent Halogenase RebH. Angew. Chem. Int. Ed. 52, (25) Duda, D. M., Walden, H., Sfondouris, J., and
1
2
3
4
5
6
7
8
5271–5274.
Schulman, B. A. (2005) Structural Analysis of Escherichia
coli ThiF. J. Mol. Biol. 349, 774–786.
(12) Frese, M., Guzowska, P. H., Voß, H., and Sewald, N.
(2014) Regioselective Enzymatic Halogenation of
Substituted Tryptophan Derivatives using the FADꢀ
Dependent Halogenase RebH. ChemCatChem 6, 1270–
1276.
(13) Frese, M., and Sewald, N. (2015) Enzymatic
Halogenation of Tryptophan on a Gram Scale. Angew.
Chem. Int. Ed. 54, 298–301.
(26) Shih, P. M., Wu, D., Latifi, A., Axen, S. D., Fewer, D.
P., Talla, E., Calteau, A., Cai, F., Marsac, N. T. de, Rippka,
R., Herdman, M., Sivonen, K., Coursin, T., Laurent, T.,
Goodwin, L., Nolan, M., Davenport, K. W., Han, C. S.,
Rubin, E. M., Eisen, J. A., Woyke, T., Gugger, M., and
Kerfeld, C. A. (2013) Improving the coverage of the
cyanobacterial phylum using diversityꢀdriven genome
sequencing. Proc. Natl. Acad. Sci. 110, 1053–1058.
(27) Burkhart, B. J., Hudson, G. A., Dunbar, K. L., and
Mitchell, D. A. (2015) A Prevalent PeptideꢀBinding
Domain Guides Ribosomal Natural Product Biosynthesis.
Nat. Chem. Biol. 11, 564–570.
(28) Koehnke, J., Mann, G., Bent, A. F., Ludewig, H.,
Shirran, S., Botting, C., Lebl, T., Houssen, W. E., Jaspars,
M., and Naismith, J. H. (2015) Structural Analysis of
Leader Peptide Binding Enables Leaderꢀfree Cyanobactin
Processing. Nat. Chem. Biol. 11, 558–563.
(29) Keller, S., Wage, T., Hohaus, K., Hölzer, M.,
Eichhorn, E., and van Pée, K.ꢀH. (2000) Purification and
Partial Characterization of Tryptophan 7ꢀHalogenase
(PrnA) from Pseudomonas fluorescens. Angew. Chem. Int.
Ed. 39, 2300–2302.
(30) Zehner, S., Kotzsch, A., Bister, B., Süssmuth, R. D.,
Méndez, C., Salas, J. A., and van Pée, K.ꢀH. (2005) A
Regioselective Tryptophan 5ꢀHalogenase Is Involved in
Pyrroindomycin Biosynthesis in Streptomyces rugosporus
LLꢀ42D005. Chem. Biol. 12, 445–452.
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
(14) Kobayashi, J. ’ichi, Sato, M., Murayama, T., Ishibashi,
M., Wälchi, M. R., Kanai, M., Shoji, J., and Ohizumi, Y.
(1991) Konbamide, a Novel Peptide with Calmoduiin
Antagonistic Activity from the Okinawan Marine Sponge
Theonella sp. J. Chem. Soc. Chem. Commun. 1050–1052.
(15) Fusetani, N., Sugawara, T., Matsunaga, S., and Hirota,
H. (1991) Orbiculamide A: A Novel Cytotoxic Cyclic
Peptide from a Marine Sponge Theonella sp. J. Am. Chem.
Soc. 113, 7811–7812.
(16) Kobayashi, J., Sato, M., Ishibashi, M., Shigemori, H.,
Nakamura, T., and Ohizumi, Y. (1991) Keramamide A, A
Novel Peptide from the Okinawan Marine Sponge
Theonella sp. J. Chem. Soc. [Perkin 1] 2609–2611.
(17) Wilson, M. C., Mori, T., Rückert, C., Uria, A. R.,
Helf, M. J., Takada, K., Gernert, C., Steffens, U. A. E.,
Heycke, N., Schmitt, S., Rinke, C., Helfrich, E. J. N.,
Brachmann, A. O., Gurgui, C., Wakimoto, T., Kracht, M.,
Crüsemann, M., Hentschel, U., Abe, I., Matsunaga, S.,
Kalinowski, J., Takeyama, H., and Piel, J. (2014) An
environmental bacterial taxon with a large and distinct
metabolic repertoire. Nature 506, 58–62.
(18) Ueoka, R., Uria, A. R., Reiter, S., Mori, T., Karbaum,
P., Peters, E. E., Helfrich, E. J. N., Morinaka, B. I., Gugger,
M., Takeyama, H., Matsunaga, S., and Piel, J. (2015)
Metabolic and Evolutionary Origin of ActinꢀBinding
Polyketides from Diverse Organisms. Nat. Chem. Biol. 11,
705–712.
(19) Altschul, S., Gish, W., Miller, W., Myers, E., and
Lipman, D. (1990) Basic Local Alignment Search Tool. J.
Mol. Biol. 215, 403–410.
(31) Bitto, E., Huang, Y., Bingman, C. A., Singh, S.,
Thorson, J. S., and Phillips, G. N. (2008) The Structure of
FlavinꢀDependent Tryptophan 7ꢀhalogenase RebH.
Proteins Struct. Funct. Bioinforma. 70, 289–293.
(32) Shepherd, S. A., Karthikeyan, C., Latham, J., Struck,
A.ꢀW., Thompson, M. L., Menon, B. R. K., Styles, M. Q.,
Levy, C., Leys, D., and Micklefield, J. (2015) Extending
the Biocatalytic Scope of Regiocomplementary Flavinꢀ
Dependent Halogenase Enzymes. Chem. Sci. 6, 3454–3460.
(33) Yeh, E., Garneau, S., and Walsh, C. T. (2005) Robust
in vitro Activity of RebF and RebH, a TwoꢀComponent
Reductase/Halogenase, Generating 7ꢀchlorotryptophan
during rebeccamycin biosynthesis. Proc. Natl. Acad. Sci.
U. S. A. 102, 3960–3965.
(20) Söding, J., Biegert, A., and Lupas, A. N. (2005) The
HHpred Interactive Server for Protein Homology Detection
and Structure Prediction. Nucleic Acids Res. 33, W244–
W248.
(21) Zhu, X., De Laurentis, W., Leang, K., Herrmann, J.,
Ihlefeld, K., van Pée, K.ꢀH., and Naismith, J. H. (2009)
Structural Insights into Regioselectivity in the Enzymatic
Chlorination of Tryptophan. J. Mol. Biol. 391, 74–85.
(22) Flecks, S., Patallo, E. P., Zhu, X., Ernyei, A. J.,
Seifert, G., Alexander, Dong, C., Naismith, J. H., and van
Pée, K.ꢀH. (2008) New Insights into the Mechanism of
Enzymatic Chlorination of Tryptophan. Angew. Chem. Int.
Ed. Engl. 47, 9533–9536.
(23) Dong, C., Flecks, S., Unversucht, S., Haupt, C., Pée,
K.ꢀH. van, and Naismith, J. H. (2005) Tryptophan 7ꢀ
Halogenase (PrnA) Structure Suggests a Mechanism for
Regioselective Chlorination. Science 309, 2216–2219.
(24) Taylor, S. V., Kelleher, N. L., Kinsland, C., Chiu, H.ꢀ
J., Costello, C. A., Backstrom, A. D., McLafferty, F. W.,
and Begley, T. P. (1998) Thiamin Biosynthesis in
Escherichia coli. J. Biol. Chem. 273, 16555–16560.
(34) Zhang, P., Liu, R., and Cook, J. M. (1995) An
Enantiospecific synthesis of L(−)ꢀ and D(+)ꢀ6ꢀChloroꢀ5ꢀ
Hydroxytryptophan: an Unusual Amino acid Residue from
the Cyclic Hexapeptide Keramamide a. Tetrahedron Lett.
36, 7411–7414.
(35) Dorrestein, P. C., Yeh, E., GarneauꢀTsodikova, S.,
Kelleher, N. L., and Walsh, C. T. (2005) Dichlorination of
a
PyrrolylꢀSꢀcarrier Protein by FADH2ꢀdependent
Halogenase PltA during Pyoluteorin Biosynthesis. Proc.
Natl. Acad. Sci. U. S. A. 102, 13843–13848.
(36) Heemstra, J. R., and Walsh, C. T. (2008) Tandem
Action of the O2ꢀ and FADH2ꢀDependent Halogenases
KtzQ and KtzR Produce 6,7ꢀDichlorotryptophan for
Kutzneride Assembly. J. Am. Chem. Soc. 130, 14024–
14025.
ACS Paragon Plus Environment

ACS Chemical Biology
Page 10 of 11
(37) Phillips, R. S., and Cohen, L. A. (1983) Synthesis of
2ꢀBromoꢀLꢀtryptophan and 2ꢀChloroꢀLꢀtryptophan.
Tetrahedron Lett. 24, 5555–5558.
(38) Hinman, R. L., and Bauman, C. P. (1964) Reactions of
NꢀBromosuccinimide and Indoles. A Simple Synthesis of
3ꢀBromooxindoles. J. Org. Chem. 29, 1206–1215.
(39) Syn, C. K. C. and Swarup, S. (2000) A Scalable
Protocol for the Isolation of LargeꢀSized Genomic DNA
within an Hour from Several Bacteria. Anal. Biochem. 278,
86–90.
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
ACS Paragon Plus Environment

Page 11 of 11
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
ACS Paragon Plus Environment