Biosynthesis of Actinorhodin and Related Antibiotics: Discovery of Alternative Routes for Quinone Formation Encoded in the act Gene Cluster

Article abstract of DOI:10.1016/j.chembiol.2009.01.015

All known benzoisochromanequinone (BIQ) biosynthetic gene clusters carry a set of genes encoding a two-component monooxygenase homologous to the ActVA-ORF5/ActVB system for actinorhodin biosynthesis in Streptomyces coelicolor A3(2). Here, we conducted molecular genetic and biochemical studies of this enzyme system. Inactivation of actVA-ORF5 yielded a shunt product, actinoperylone (ACPL), apparently derived from 6-deoxy-dihydrokalafungin. Similarly, deletion of actVB resulted in accumulation of ACPL, indicating a critical role for the monooxygenase system in C-6 oxygenation, a biosynthetic step common to all BIQ biosyntheses. Furthermore, in vitro, we showed a quinone-forming activity of the ActVA-ORF5/ActVB system in addition to that of a known C-6 monooxygenase, ActVA-ORF6, by using emodinanthrone as a model substrate. Our results demonstrate that the act gene cluster encodes two alternative routes for quinone formation by C-6 oxygenation in BIQ biosynthesis.

Full text of DOI:10.1016/j.chembiol.2009.01.015

Chemistry & Biology
Article
Biosynthesis of Actinorhodin and Related
Antibiotics: Discovery of Alternative Routes for
Quinone Formation Encoded in the act Gene Cluster
Susumu Okamoto,1,3 Takaaki Taguchi,2,3 Kozo Ochi, * and Koji Ichinose *
1,
2,
1
National Food Research Institute, Kannondai, Tsukuba 305-8642, Japan
2Research Institute of Pharmaceutical Sciences, Musashino University, Shinmachi, Nishitokyo-shi, Tokyo 202-8585, Japan
3These authors contributed equally to this work.
*Correspondence: kochi@affrc.go.jp (K.O.), ichinose@musashino-u.ac.jp (K.I.)
DOI 10.1016/j.chembiol.2009.01.015
SUMMARY
PKS (Beltran-Alvarez et al., 2007); conversion of the octaketide
to the bicyclic intermediate 2 catalyzed by the ketoreductase
All known benzoisochromanequinone (BIQ) biosyn- (KR) for the C-9 position (Korman et al., 2008); and the stereo-
thetic gene clusters carry a set of genes encoding specific ketoreduction at C-3, leading to the first chiral biosyn-
thetic intermediate, 4-dihydro-9-hydroxy-1-methyl-10-oxo-3-
H-naptho-[2,3-c]-pyran-3-(S)-acetic acid, (S)-DNPA 3 (Itoh
et al., 2007). Subsequent oxygenation at the C-6 position
a two-component monooxygenase homologous to
the ActVA-ORF5/ActVB system for actinorhodin
biosynthesis in Streptomyces coelicolor A3(2).
Here, we conducted molecular genetic and biochem-
ical studies of this enzyme system. Inactivation of
actVA-ORF5 yielded a shunt product, actinoperylone
(
Figure 1A) is the key step to complete BIQ chromophore
synthesis, and biochemical (Kendrew et al., 1997) and crystallo-
graphic (Sciara et al., 2003) studies showed that the ActVA-
ORF6 protein functions as a C-6 monooxygenase, which, unusu-
ally, requires no prosthetic group, metal ion, or cofactor. At
present, however, there is no in vivo evidence for the involvement
(
ACPL), apparently derived from 6-deoxy-dihydro-
kalafungin. Similarly, deletion of actVB resulted in
accumulation of ACPL, indicating a critical role for of ActVA-ORF6 in C-6 oxygenation.
the monooxygenase system in C-6 oxygenation,
Two other complete biosynthetic gene clusters for the BIQs
a biosynthetic step common to all BIQ biosyntheses. (Figures 1B and 1C) have been cloned and sequenced to date:
granaticin (GRA) 4 (the gra cluster) (Sherman et al., 1989; Becht-
Furthermore, in vitro, we showed a quinone-forming
activity of the ActVA-ORF5/ActVB system in addition
hold et al., 1995: Ichinose et al., 1998a) and medermycin (MED) 5
(the med cluster) (Ichinose et al., 2003). Comparative analysis of
to that of a known C-6 monooxygenase, ActVA-
ORF6, by using emodinanthrone as a model
substrate. Our results demonstrate that the act
gene cluster encodes two alternative routes for
quinone formation by C-6 oxygenation in BIQ biosyn-
thesis.
the three clusters revealed the unexpected lack of an actVA-
ORF6 homolog in the gra and med clusters, even though oxygen
is found at the C-6 position in both GRA 4 and MED 5. Instead,
they include oxygenase genes (gra-ORF21 and med-ORF7,
respectively) homologous to actVA-ORF5, which was initially
deduced to encode a hydroxylase for the C-8 position (Caballero
et al., 1991) based on its significant similarity (>70%) to pheA
encoding a phenol hydroxylase from Bacillus stearothermophilus
INTRODUCTION
(
Kim and Oriel, 1995). Recently, ActVA-ORF5 was reported to
Bacteria of the genus Streptomyces are one of the most impor- form a two-component monooxygenase together with a flavin:
tant groups of microorganisms in producing a huge variety of NADH oxidoreductase, ActVB, which provides reduced flavin
secondary metabolites, including antibiotics (Chater and Hop- to the actual oxygenase component conducting the C-8 oxygen-
wood, 1993). Actinorhodin (ACT) 1 is a blue-pigmented antibiotic ation in ACT biosynthesis (Valton et al., 2004, 2006). The gra and
produced by Streptomyces coelicolor A3(2), which is genetically med clusters also contain genes homologous to actVB (gra-
the most-characterized actinomycete (Bentley et al., 2002). ACT ORF34 and med-ORF13, respectively). The absence of the
belongs to a class of aromatic polyketides, the benzoisochroma- hydroxyl group at C-8 of MED 5 as well as the lack of an
nequinone (BIQ) antibiotics. Historically, ACT 1 is the first antibi- actVA-ORF6 homolog in the med cluster prompted us to
otic whose whole biosynthetic gene cluster was cloned and has propose the possible involvement of Med-ORF7, an ActVA-
served as one of the best model compounds for studying type II ORF5 homolog, in C-6 oxygenation in MED biosynthesis (Ichi-
polyketide synthase (PKS), their ancillary enzymes, and subse- nose et al., 2003). These conflicting situations raise questions
quent post-PKS tailoring enzymes. Consequently, numerous about the in vivo contributions of the ActVA-ORF5/ActVB and
studies have provided much important information on ACT ActVA-ORF6 oxygenases to ACT biosynthesis. This paper
biosynthesis. Recent advances in understanding the funda- describes a functional analysis of the actVA-ORF5, actVA-
mental processes revealed mechanistic details of the assembly ORF6, and actVB genes based on molecular genetics, metabolic
of the basic carbon skeleton accomplished by the type II minimal analysis, and biochemical characterization, revealing a novel
226 Chemistry & Biology 16, 226–236, February 27, 2009 ª2009 Elsevier Ltd All rights reserved

Chemistry & Biology
Alternative Quinone Formation for Actinorhodin
Figure 1. The Overview of BIQ Biosyntheses
(
(
(
A) The proposed later tailoring steps of the ACT biosynthetic pathway. The numbers are based on the biosynthetic order.
B) The structures of benzoisochromanequinone (BIQ) antibiotics.
C) The biosynthetic gene clusters of ACT, granaticin, and medermycin. actVA-ORF6 is shown by a dotted arrow. actVA-ORF5 and its homologs are shown by
filled arrows. The actVB homologs are shown by striped arrows.
situation in which alternative pathways are available for the later both genes by replacing them with a spectinomycin-strepto-
tailoring steps of ACT biosynthesis.
mycin-resistance (aadA) cassette, resulting in a DactVA-5,6
double mutant (see Figure S1 article online). Correct disruption
of the genes was confirmed by Southern blot analysis
RESULTS
(
Figure S1). A strain (M510) wild-type for the ACT pathway
actVA-ORF5, but Not actVA-ORF6, Is Essential
produced ACT 1 and its lactonized derivatives (ACTs): g-actino-
0 0
rhodin (g-ACT) 6 and g -actinorhodin (g -ACT) 7, as shown in
for ACT Biosynthesis
To establish the in vivo function of actVA-ORF5 and ORF6, we Figure 2. The DactVA-5,6 mutant did not produce ACTs (1, 6,
constructed deletion mutants for these genes. In the act cluster, and 7) at all. Instead, the mutant accumulated a yellowish-
0
0
the extreme 3 end of actVA-ORF5 overlaps with the 5 end of brown pigment, demonstrating the involvement of actVA-
actVA-ORF6, indicative of potential translational coupling. Inac- ORF5 and/or actVA-ORF6 in ACT biosynthesis (Figures 2A
tivation of actVA-ORF5 would have polar effects on the down- and 2C). Our previous spectral analysis elucidated its structure
stream gene, actVA-ORF6. Therefore, we initially inactivated as a novel perylenequinone-type shunt product, 7H,16H-perylo
Chemistry & Biology 16, 226–236, February 27, 2009 ª2009 Elsevier Ltd All rights reserved 227

Chemistry & Biology
Alternative Quinone Formation for Actinorhodin
Figure 2. Funcitonal Analysis of actVA-ORF5 and
ORF6
(
A) Effect of the actVA deletion mutations on ACT biosynthesis.
ꢁ
ꢂ
Strains were inoculated on R5 medium and grown at 30 C for
days. The reverse side of plates is shown.
B) Complementation of the DactVA-5,6 mutant. The actVA-
ORF5 gene was expressed by using pIJ8600+actVA5. Strains
4
(
ꢁ
ꢂ
were grown on R5 medium at 30 C for 3 days, either in
the absence (ꢁThio) or presence (+Thio) of thiostrepton
(
(
10 mg/ml).
C) HPLC chromatograms of the ACT-related metabolites from
culture medium of the strain M510 and its actVA mutants.
Strains were cultured in R5MS medium, and the ACT-related
metabolites in the culture supernatants were analyzed as
described in Experimental Procedures. Chemical structures
of the relevant compounds are shown on the right side of
the panel.
possible that a presumptive homolog of actVA-
ORF6 exists elsewhere in the genome and
compensated for the mutation. However, genomic
sequence analysis of S. coelicolor A3(2) (Bentley
et al., 2002) revealed no homologs of actVA-
ORF6, at least from the view of amino acid
sequence homology. Furthermore, it has been
demonstrated that the act cluster contains all
genetic information required for ACT biosynthesis
(
Malpartida and Hopwood, 1984). Apparently,
actVA-ORF5, but not actVA-ORF6, is essential for
ACT biosynthesis, and the C-6 oxygenation step
in ACT biosynthesis should be mainly governed
by actVA-ORF5.
In Vivo Contribution of ActVA-ORF5/ActVB
and ActVA-ORF6 as the C-6 Monooxygenase
A flavin:NADH oxidoreductase, ActVB, has been
reported to be necessary for ActVA-ORF5 to
function as a monooxygenase in vitro, and a combi-
nation of the two proteins was named the ActVA-
0
[
2,1c:8,7-c ]dipyran-7,16-dione,1,3,4,9,11,12-hexahydro-8,15- ActVB system (Valton et al., 2004, 2006). In our study, a more
dihydroxy-1R,9R-dimethyl-3S, 11S-diacetic acid, which was precise designation, the ActVA-ORF5/ActVB system, is
designated as actinoperylone (ACPL) 8 (Taguchi et al., 2008). proposed. To clarify the in vivo function of actVB, this gene
Blocking oxygenation at C-6 of 6-deoxy-dihydrokalafungin was replaced by an erythromycin-resistance (ermE) cassette
(
DDHK) 9 would result in possible tautomerization to the dihy- as described in Experimental Procedures (Figure S3). HPLC
droxynaphthalene derivative, followed by oxidative dimerization analyses of culture supernatants clearly showed that the deletion
to afford ACPL 8. Therefore, C-6 oxidation of DDHK 9 is mutant, DactVB, produced ACTs (6 and 7) and ACPL 8 simulta-
deduced to be compromised in the DactVA-5,6 mutant.
neously (Figure 3). Quantitative estimations demonstrated that
Next, we deleted actVA-ORF6 alone to construct the mutant the total amount of ACTs (6 and 7) produced was decreased to
DactVA-6 (Figure S1), but no obvious effect of the mutation on about 30% of wild-type levels made by M510. The amount of
ACT production was detected (ꢀ9% less production than the ACPL 8 was almost the same as in the DactVA-5,6 mutant.
wild-type, WT) (Figures 2A and 2C), indicating a rather limited This result indicates that deletion of actVB perturbed the
contribution of actVA-ORF6 to ACT biosynthesis. In addition, ActVA-ORF5/ActVB system and led to interruption of efficient
ectopic expression of actVA-ORF5 alone under the control of C-6 oxygenation of DDHK 9. Previously, an actVB mutant strain,
a thiostrepton-inducible promoter, tipAp, in the DactVA-5,6 B135 (Rudd and Hopwood, 1979), was reported to produce
mutant (the recombinant, DactVA-5,6/actVA-5) restored ACT kalafungin (KAL) 14 (Cole et al., 1987). Under our culture condi-
production in response to thiostrepton induction (Figure 2B; tions, however, the strain B135 was confirmed to produce g-
Figure S2). In contrast, expression of actVA-ORF6 alone in the ACT 6 and ACPL 8, consistent with the results from our DactVB
DactVA-5,6 mutant (the recombinant, DactVA-5,6/actVA-6) had mutant, and neither KAL 14 nor dihydrokalafungin (DHK) 10
no obvious effect on the metabolic profile (Figure S2). It was could be detected (data not shown).
228 Chemistry & Biology 16, 226–236, February 27, 2009 ª2009 Elsevier Ltd All rights reserved

Chemistry & Biology
Alternative Quinone Formation for Actinorhodin
mutant produced a significant amount of ACTs. Therefore, we
decided to investigate the contribution of actVA-ORF6 in this
genetic background. For this purpose, we constructed a DactVB
DactVA-ORF6 double mutant (DactVBDactVA-6). This mutant
still accumulated ACPL 8 (Figure 3), but virtually no ACTs (1, 6,
and 7), indicating that ActVA-ORF6 really catalyzes C-6 oxygen-
ation, and that this manifestation of enzymatic activity depends
on the DactVB mutation. In other words, C-6 oxygenation is
normally governed by the ActVA-ORF5/ActVB system, and the
ActVA-ORF6 protein contributes as an alternative C-6 monooxy-
genase in the actVB mutant background, in which the function of
the highly efficient ActVA-ORF5/ActVB system is compromised.
In Vitro Activity of ActVA-ORF5/ActVB
as a Quinone-Forming Oxygenase
The results described above strongly suggest that the ActVA-
ORF5/ActVB system is a quinone-forming enzyme that converts
DDHK 9 into DHK 10. To demonstrate this enzymatic activity
in vitro, we expressed and purified both ActVA-ORF5 and ActVB
as His
6
-tagged proteins (Figure S4). We also prepared His
6
-
tagged ActVA-ORF6 as
a
positive control. Because the
presumptive natural substrate, DDHK 9, is not available, we
used emodinanthrone 11 as a model substrate (Figure 4). This
compound was successfully used in previous studies of other
quinone-forming enzymes (Chen et al., 1995), and its chemical
structure is more analogous to that of DDHK 9 than tetracenomy-
cin F1 (TCMF1) 13, which was used in the biochemical studies of
ActVA-ORF6 (Kendrew et al., 1997). ActVA-ORF5 and ActVB
Figure 3. Effect of the DactVB and DactVB DactVA-ORF6 Double
Mutations on ACT Biosynthesis
HPLC profiles monitored by absorbance at 450 nm are shown. The DactVB
mutant produced ACTs (6 and 7) and ACPL 8 simultaneously, whereas the
DactVB DactVA-ORF6 double mutant produced only ACPL 8.
Given the crucial role of ActVB in the enzyme system, the proteins were incubated with emodinanthrone 11 and cofactors
DactVB mutation should be highly detrimental to the enzymatic FMN and NADH as described in Experimental Procedures. Strik-
function of the ActVA-ORF5/ActVB system. Nevertheless, the ingly, the reconstituted ActVA-ORF5/ActVB system catalyzed
Figure 4. Demonstration of an In Vitro Quinone-
Forming Monooxygenase Activity of the ActVA-
ORF5/ActVB System
(
(
A) HPLC analysis of reaction products. ActVA-ORF5
1 mM) and ActVB (100 nM) were incubated with emodin-
ꢂ
anthrone 11 (50 mM) for 10 min at 25 C as described in
Experimental Procedures. Elution profiles were monitored
by absorbance at 254 nm. Top, authentic (0.5 mg each)
emodinanthrone 11 and emodin 12; bottom, an ethyl
acetate extract from an in vitro assay mixture.
(
B) Mass spectra of emodinanthrone 11 and emodin 12
extracted from a reaction mixture. The peaks correspond-
ꢁ
ing to emodinanthrone 11 (m/z 255 as [M-H] ) and emodin
ꢁ
1
2 (m/z 269 as [M-H] ) were detected.
(
C) Kinetic properties of ActVA-ORF5/ActVB, ActVA-
ORF6, and their related enzymes.
Chemistry & Biology 16, 226–236, February 27, 2009 ª2009 Elsevier Ltd All rights reserved 229

Chemistry & Biology
Alternative Quinone Formation for Actinorhodin
Figure 5. Demonstration of a DHK Oxygen-
ation Activity of the ActVA-ORF5/ActVB
System
(
A) HPLC analysis of reaction products. Assay
conditions were essentially the same as those
used for emodinanthrone 11 (Figure 4). Top, DHK
1
0 (50 mM) was incubated with ActVA-ORF5 and
ActVB; bottom, DHK without the enzymes.
B) Mass spectra of (X), KAL 14, and (Y) extracted
(
from a reaction mixture. The peaks corresponding
ꢁ
to X, Y (m/z 315 as [M-H] ), and KAL 14 (m/z 299
ꢁ
as [M-H] ) were detected.
described in Experimental Procedures.
Although these compounds were rapidly
converted to their cognate quinone form
by autoxidation, the ActVA-ORF5/ActVB
system significantly enhanced this rate,
confirming a quinone-forming activity of
the enzyme system (Figure S5). However,
no hydroxylation activity was detected in
these assays (Figure S5).
Finally, therefore, we decided to use
the presumptive natural substrate, DHK
10, for C-8 hydroxylation to assess the
oxidation of emodinanthrone 11 to emodin 12, demonstrating involvement of the ActVA-ORF5/ActVB system in this biosyn-
a quinone-forming activity of the system (Figures 4A and 4B). thetic step. Results of assays are shown in Figure 5. DHK 10 is
We confirmed, as shown by Valton et al. (2004, 2006), that spontaneously lactonized, possibly via a hydroquinone form, to
ActVA-ORF5, the genuine oxygenase component, absolutely KAL 14 (Ichinose et al., 1998b), which was detected irrespective
required ActVB, FMN, and NADH for its enzymatic activity of the presence of the enzyme system (Figure 5A). Strikingly, the
(
data not shown). In contrast, ActVA-ORF6 did not require any reaction with the enzyme system provided new compounds
exogenously added cofactors, as described previously (Ken- X and Y (Figure 5A), and their molecular mass was shown to
drew et al., 1997). be 315 (ESI-negative) (Figure 5B), implying that these
To characterize the enzymatic properties of the ActVA-ORF5/ compounds are oxygenated molecules of KAL 14. It is likely
ActVB system, kinetic parameters were determined (Figure 4C). that these compounds are generated by C-8 or C-10 hydroxyl-
Interestingly, the K
system was significantly lower than that of ActVA-ORF6 system. Based on the significant decrease of KAL 14 in the enzy-
7.54 mM versus 211 mM) and comparable to that of AknX, matic sample, this compound is suggested to be the actual
m
for emodinanthrone 11 of this enzyme ation catalyzed by the ActVA-ORF5/ActVB monooxygenase
(
a distant homolog of ActVA-ORF6 involved in aklavinone biosyn- substrate for oxygenation to produce the compounds X and Y.
thesis in Streptomyces galilaeus (Chung et al., 2002). These Although further structural elucidation is required, we clearly
results imply more efficient oxidation of DDHK 9 by the ActVA- demonstrated that the ActVA-ORF5/ActVB system possesses
ORF5/ActVB system. Together, we unambiguously demon- an additional oxygenation activity other than the quinone-form-
strated an enzymatic activity of ActVA-ORF5/ActVB as ing activity we discovered.
a quinone-forming oxygenase.
Shunt Pathway Leading to 8 in Relation
An Additional Oxygenase Activity of the ActVA-ORF5/
ActVB System
to C-6 Oxygenation
A proposed scheme (Taguchi et al., 2008) for the formation of
The ActVA-ORF5/ActVB system has been proposed to be ACPL 8 includes oxidative coupling of DDHK 9, which is the
involved in the C-8 hydroxylation step in ACT biosynthesis postulated product of a reductase, ActVI-ORF2 (Taguchi et al.,
(
Valton et al., 2006, 2008). Therefore, we then investigated this 2000) (Figure 7). To elucidate the essential genes for the shunt
possibility. Our model substrate emodinanthrone 11 is suitable pathway to ACPL 8, we analyzed the act cluster-deficient strain,
to detect a quinone-forming activity because of its intrinsic CH999, carrying plasmid-borne copies of the act biosynthetic
slow rate of autoxidation. However, this compound 12 is unlikely genes required for the formation of (S)-DNPA 3, with actVI-
to be hydroxylated at position 4 (equivalent to C-8 of ACT), owing ORF2 in addition (CH999/pIJ5659) (Ichinose et al., 1999).
to its highly substituted nature. Thus, we used less substituted HPLC analysis of a liquid culture of CH999/pIJ5659 clearly de-
anthracenes, anthrone, and dithranol to detect quinone forma- tected ACPL 8 (Figure 6). This result proves that actVI-ORF2
tion and subsequent hydroxylation. These model substrates indeed encodes a reductase for C-15, and, most importantly,
were assessed under our standard assay conditions as formation of ACPL 8 was demonstrated by the minimal gene
230 Chemistry & Biology 16, 226–236, February 27, 2009 ª2009 Elsevier Ltd All rights reserved

Chemistry & Biology
Alternative Quinone Formation for Actinorhodin
reasonable to assume that ActVA-ORF6 catalyzes the C-6
oxygenation of DDHK 9 (Fern a´ ndez-Moreno et al., 1994). Indeed,
Kendrew et al. (1997) demonstrated its quinone-forming activity
by using TCMF1 13 as a model substrate. Furthermore, the
crystal structure of ActVA-ORF6 in complex with substrate and
product analogs provided an explanation for C-6 oxygenation,
where the active site cavity of the enzyme acts as a shield for
the substrate, DDHK 9, leaving only the C-6 carbon exposed
to molecular oxygen (Sciara et al., 2003).
Although C-6 oxygenation is a tailoring step common to all BIQ
biosyntheses, neither the gra nor med cluster possesses actVA-
ORF6 homologs (Figure 1C). This prompted us to investigate the
in vivo contribution of actVA-ORF6 to ACT production. An actVA-
ORF6 deletion mutant produced ACTs at levels only slightly less
than the wild-type, suggesting a rather limited contribution of the
gene to ACT biosynthesis (Figure 2). Instead, the gene immedi-
ately upstream, actVA-ORF5, was found to be critical for C-6
oxygenation in the present study (Figures 2 and 4). The actVA
genetic region corresponds to one of the phenotypic classes
of blocked mutants that produce diffusible brown pigments
(
Rudd and Hopwood, 1979). Most of the actVA mutations were
mapped to the actVA-ORF5 gene (Caballero et al., 1991), and
later chemical characterization of an actVA-ORF5 mutant (strain
B1) carrying the act-101 mutation (Caballero et al., 1991) identi-
fied (S)-DNPA 3 as an actVA intermediate (Cole et al., 1987). We
also analyzed the same class of actVA-ORF5 mutant (strain B9)
carrying the act-109 mutation (Caballero et al., 1991) and identi-
fied ACPL 8 as a major ACT-related metabolite, but a trace
amount of (S)-DNPA 3, though no ACTs (1, 6, and 7), was also
detected (data not shown). These results demonstrated the
critical role of actVA-ORF5 in C-6 oxygenation. Consistent with
this conclusion, the gra and med clusters carry homologs of
actVA-ORF5, gra-ORF21 (Ichinose et al., 1998a), and med-
ORF7 (Ichinose et al., 2003).
Figure 6. Formation of Actinoperylone by the Minimal Gene Set
HPLC analyses of metabolites produced by S. coelicolor CH999 recombinants
harboring either pIJ5660 or pIJ5659. Chromatograms monitored by absor-
bance at 450 nm are shown. The gene sets used for the plasmids are also
shown above each chromatogram. The recombinant strain harboring
pIJ5660 produced (S)-DNPA 3. On the other hand, the strain harboring
pIJ5659 accumulated a shunt product, ACPL 8.
The actVB gene was initially identified genetically among actV
mutants, being phenotypically similar but not identical to the
actVA mutants (Rudd and Hopwood, 1979; Fern a´ ndez-Moreno
et al., 1992). Based on the isolation of KAL 14, which is
convertible in vivo to ACT 1 (Cole et al., 1987), actVB had been
postulated to be involved in oxidative dimerization in ACT
biosynthesis. Despite an apparent lack of an equivalent biosyn-
thetic step, we discovered actVB homologs (gra-ORF34, med-
ORF13) in the other BIQ gene clusters (Ichinose et al., 1998a,
2003) (Figure 1C), indicating their role in a step common to the
biosynthesis of these undimerized compounds. The ActVB
sets for the possible formation of DDHK 9. Interruption of the C-6
oxygenation of DDHK 9 would result in possible tautomerization
to the dihydroxynaphthalene derivative, followed by oxidative
dimerization to afford ACPL 8. One of the possible candidates
involved in the coupling is a P450 gene (Zhao et al., 2007) outside
the act cluster.
DISCUSSION
Oxidative modifications catalyzed by monooxygenases are protein was shown to be a flavin:NADH oxidoreductase (Ken-
common tailoring steps in the biosynthesis of polyketide metab- drew et al., 1995). Recently, the ActVA-ORF5 and ActVB proteins
olites, including both aromatic and macrocyclic antibiotics. The were reported to constitute a two-component monooxygenase
most extensively studied examples are the cytochrome P450- system, in which ActVB donates reduced flavin to the actual oxy-
dependent monooxygenases (Munro et al., 2007). However, genase component, ActVA-ORF5 (Valton et al., 2004), providing
the 22 kb act biosynthetic gene cluster carries no genes for an explanation for the copresence of their cognate genes in all
P450-type enzymes, suggesting that distinct kinds of monooxy- BIQ pathways. Detailed biochemical studies (Valton et al.,
genases are involved in the oxygenation steps at C-6 and C-8 in 2008) proposed the involvement of the ActVA-ORF5/ActVB
ACT biosynthesis.
system in the oxygenation step at C-8 in ACT biosynthesis.
ActVA-ORF6 significantly resembles the monooxygenase However, notwithstanding the absence of the hydroxyl group
encoded by tcmH involved in tetracenomycin (TCM) biosyn- at C-8 in MED 5, highly conserved orthologs of actVA-ORF5
thesis in Streptomyces glaucescens (Shen and Hutchinson, (med-ORF7) and actVB (med-ORF13) are present in the med
1
993). Because TcmH had been shown to be a quinone-forming gene cluster. This fact implies an additional function for the
monooxygenase that converts TCMF1 13 to TCMD3, it was ActVA-ORF5/ActVB system other than as a catalyst for C-8
Chemistry & Biology 16, 226–236, February 27, 2009 ª2009 Elsevier Ltd All rights reserved 231

Chemistry & Biology
Alternative Quinone Formation for Actinorhodin
Figure 7. Proposed Later Tailoring Pathway of ACT Biosynthesis
The ACT biosynthetic pathway exploits a dual enzyme system for the C-6 oxygenation of DDHK 9. The ActVA-ORF5/ActVB system is a principal enzyme for this
reaction. The ActVA-ORF6 protein also catalyzes the same reaction, but its contribution is rather limited in the wild-type genetic background. However, the
activity of ActVA-ORF6 is manifested in the actVB genetic backgrounds. Production of ACT 1 from DHK 10 requires at least two reaction steps: dimerization
and 8-hydroxylation.
hydroxylation. Indeed, we demonstrated the involvement of even though GRA and MED biosynthesis also requires C-6
ActVA-ORF5/ActVB in the C-6 oxygenation step.
oxygenation. Perhaps the actVA-ORF6 gene has been acquired
We further investigated the in vivo contributions of the ActVA- by horizontal gene transfer from an unknown source, now
ORF5/ActVB and ActVA-ORF6 monooxygenases by construct- providing an alternative route to C-6 oxygenation.
ing an actVB deletion mutant (DactVB). This mutant was
An immediate interest would be the possible involvement of the
expected to produce ACPL 8, owing to a lack of the principal ActVA-ORF5/ActVB system not only in C-6 oxygenation, but also
C-6 monooxygenase, AtctVA-ORF5/ActVB. In fact, the DactVB in C-8 hydroxylaton, as proposed previously (Valton et al., 2006,
mutant accumulated ACPL 8 (Figure 3), but still produced 2008). Under our assay conditions, the model anthrone
a considerable amount of ACTs (6 and 7), indicating two possibil- substrates (emodinanthrone 11, anthrone, and dithranol) were
ities: (i) ActVA-ORF5 itself is partially functional without ActVB; (ii) efficiently converted to the corresponding anthraquinones, but
ActVA-ORF6 makes a significant contribution in the DactVB they did not receive any additional oxygenation. Therefore, we
genetic background. To clarify this issue, we then prepared an used the presumptive natural substrate, DHK 10. As expected,
actVB actVA-ORF6 double mutant (DactVBDactVA-6), and we this compound was readily converted to the oxygenated
confirmed that ACT production was virtually abolished in this compounds (Figure 5), which possess the molecular mass (316)
mutant, although the accumulation of ACPL 8 was still observed corresponding to hydroxylated KAL 14. Previous studies on the
(Figure 3). The results clearly indicate that ActVA-ORF6 is also ActVA-ORF5/ActVB system (Valton et al., 2006) demonstrated
functional in vivo and that its activity as a C-6 monooxygenase its in vitro oxygenating activity that catalyzes the conversion of
is manifested in the actVB mutant, in which the function of the DHK 10 to its monooxygenated derivative, DHK-OH, with
ActVA-ORF5/ActVB system is compromised.
a molecular mass of 318. This conversion was, however, less effi-
Following the in vivo experiments, we attempted to prove an cient, and the true substrate was suggested to be the hydroqui-
in vitro quinone-forming activity of the ActVA-ORF5/ActVB none form of DHK 10, which was efficiently oxygenated under
system. We could readily reconstitute the two-component the anaerobic assay conditions employed (Valton et al., 2006).
system with purified recombinant proteins and detect its enzy- Our experimental data imply that KAL 14, but not DHK 10, is
matic activity by using emodinanthrone 11 as a model substrate the actual substrate for this enzymatic reaction (Figure 5).
(
Figure 4). A quinone-forming activity of the ActVA-ORF6 protein Although further studies are required to determine the bona fide
was also confirmed in our assay system. These observations biosynthetic substrate for the hydroxylation reaction and the
reinforce our claim that the ActVA-ORF5/ActVB system is the chemical structure of the reaction products, we demonstrated
principal C-6 monooxygenase in ACT biosynthesis.
It is interesting that ActVA-ORF6 has a lower K
unambiguously an additional oxygenation activity other than
for tetraceno- the quinone-forming activity of the ActVA-ORF5/ActVB system.
m
mycin F1 (TCMF1) 13 than for a substrate closer in structure to an The shunt formation of ACPL 8 by the actVA-ORF5 or actVB
ACT intermediate, and that it is present only in the ACT system, mutants is reasonably explained by oxidative coupling of
232 Chemistry & Biology 16, 226–236, February 27, 2009 ª2009 Elsevier Ltd All rights reserved

Chemistry & Biology
Alternative Quinone Formation for Actinorhodin
DDHK 9, a postulated substrate of C-6 oxygenation. Strikingly,
the recombinant strain with the expression plasmid (pIJ5659)
carrying the minimal set of genes for the formation of DDHK 9
produced ACPL 8, clearly suggesting a direct link between inter-
ruption of C-6 oxygenation and oxidative coupling.
Construction of the DactVA-5,6, DactVA-6, and DactVB Mutants
The DactVA-5,6 Mutant
0
A 1 kb region 5 of actVA-ORF5 was amplified by PCR with primers DactVA5,
0
6
-UF and DactVA5,6-UR. A second 1.1 kb region 3 of actVA-ORF6 was ampli-
fied with primers DactVA6-DF and DactVA6-DR. The PCR fragments were
inserted between the EcoRI and HindIII sites of pUC19 by three-fragment liga-
tion, yielding pUC19+DactVA5,6. The aadA gene was amplified from pIJ2059
(Kieser et al., 2000) by using primers aadA-F and aadA-R. The amplified aadA
gene cassette was cloned in the XbaI site of pUC19+DactVA5,6, generating
pUC19+DactVA5,6::aadA. The DactVA5,6::aadA construct was excised by
digestion with EcoRI and HindIII and was ligated with pGM9 (Kieser et al.,
The combined data allow us to propose an alternative oxygen-
ation pathway in ACT biosynthesis as shown in Figure 7,
including a dual C-6 oxygenation system consisting of ActVA-
ORF5/ActVB and ActVA-ORF6. Our experimental results do
not disprove the functional assignment of the ActVA-ORF5/
ActVB system as a C-8 hydroxylase. Neither the missense
actVA-ORF5 mutant (strain B9; unpublished data) nor the
DactVA-5,6 mutant harboring the actVA-ORF6 expression
plasmid produced any ACTs (1, 6, and 7), although the alterna-
tive C-6 monooxygenase ActVA-ORF6 is highly likely to be func-
tional in these strains. Thus, it is apparent that the ActVA-ORF5
protein is involved in an additional biosynthetic step after C-6
oxygenation. Indeed, we confirmed an additional function of
the ActVA-ORF5/ActVB system other than that as a quinone-
forming enzyme. Further genetic and chemical studies are in
progress to elucidate details of the late ACT biosynthetic
pathway, including the C-8 hydroxylation step.
2
000), which had been treated with the same enzymes. The ligation mixture
was used to transform S. lividans TK21 protoplasts to yield the plasmid
pGM9+DactVA5,6::aadA. S. coelicolor M510 was transformed with
pGM9+DactVA5,6::aadA, and transformants were selected with thiostrepton
(
Thio). To obtain single-crossover recombinants, purified transformants were
ꢂ
cultured on Thio-containing plates at 37 C. Because pGM9 carries a tempera-
ture-sensitive replicon of pSG5, it cannot replicate at this temperature. Thio-
resistant single-crossover recombinants were subcultured by two rounds of
ꢂ
streaking in the absence of Thio at 37 C. Double-crossover recombinants
(
DactVA-5,6) in which the delivery plasmid was lost from cells were identified
as colonies with Thio-sensitive and streptomycin-resistant phenotypes.
The correct deletion of the genes was confirmed by PCR and Southern blot
analyses.
The DactVA-6 Mutant
0
A 1 kb region 5 of actVA-ORF6 was amplified by PCR with primers DactVA6-UF
0
and DactVA6-UR. A 1.1 kb region 3 of actVA-ORF6 was amplified as described
above. These PCR fragments were inserted between the EcoRI and HindIII
sites of pUC19 by three-fragment ligation, generating pUC19+DactVA6. The
aadA gene cassette described above was cloned in this recombinant plasmid,
yielding pUC19+DactVA6::aadA. The DactVA6::aadA construct was excised
by digestion with EcoRI and HindIII and ligated with pGM9 to obtain
pGM9+DactVA6::aadA. The resulting plasmid was introduced into S. coelicolor
M510, and the actVA-ORF6 deletion mutant (DactVA-6) was obtained as
described above.
SIGNIFICANCE
The striking conclusion from this work is that the gene
cluster for ACT biosynthesis encodes two alternative routes
for quinone formation at a critical stage in post-PKS
tailoring. One of these, catalyzed by ActVA-ORF5/ActVB, is
the major route in the wild-type, but when actVB is deleted
The DactVB Mutant
(and perhaps when the expression level of actVB is low in
0
A 1 kb region 5 of actVB was amplified with primers DactVB-UF and DactVB-
the wild-type), the alternative route, catalyzed by ActVA-
ORF6, becomes manifested. Oxygenation, including
quinone formation and hydroxylation, is one of the key steps
in secondary metabolism; thus, these findings may have
wide significance for understanding and manipulating the
oxygenation of a variety of aromatic compounds. The
possible bifunctionality of the ActVA-ORF5/ActVB system
found in this study would open the possibility of its applica-
tion to the oxygenation of various aromatic compounds.
0
UR. A second 1 kb region 3 of actVB was amplified with primers DactVB-DF
and DactVB-DR. The PCR fragments were inserted between the EcoRI and
XhoI sites of pGEM-11zf by three-fragment ligation, yielding pGEM+DactVB.
The erythromycin/lincomycin-resistance gene, ermE, was excised from
pIJ4026 (Kieser et al., 2000) by digestion with KpnI and was ligated with
pGEM+DactVB, which had been treated with the same enzyme, generating
pGEM+DactVB::ermE. The DactVB::ermE construct was recovered by diges-
tion with EcoRI and XhoI and ligated with pGM9 to obtain pGM+DactV
B::ermE. This plasmid was introduced into S. coelicolor M510, and the actVB
deletion mutant (DactVB) was screened by virtue of Thio-sensitive and linco-
mycin-resistance phenotypes. Similarly, pGM+DactVB::ermE was introduced
into the DactVA6 mutant, and the DactVB DactVA-ORF6 double mutant
EXPERIMENTAL PROCEDURES
(
DactVBDactVA6) was obtained in analogous fashion.
Bacterial Strains
S. coelicolor M510 (M145 DredD) (Floriano and Bibb, 1996) was used as the
wild-type strain throughout this study. This strain does not produce the red-
pigmented prodiginines due to the DredD mutation, allowing for easy charac-
terization of act mutants. S. coelicolor CH999 (proA1 argA1 redE60 Dact::ermE
Southern Hybridization
DNA fragments from agarose gels were transferred to Hybond-N+ nylon
membrane (GE Healthcare). Probes were prepared by PCR with the following
primer sets: actVA5,6 probe, DactVA5,6-Pro-F, and DactVA5,6-Pro-R; aadA
probe, aadA-F, and aadA-R; actVB probe, DactVB-Pro-F, and DactVB-Pro-
R; ermE probe, ermE-F, and ermE-R. These probes were nonradioactivlely
labeled by using the DIG DNA labeling kit (Roche). Hybridization, washing,
and detection were performed as recommended by the manufacturer (Roche).
ꢁ
ꢁ
SCP1 SCP2 ) (McDaniel et al., 1993) was also used in some experiments.
Streptomyces lividans TK21, Escherichia coli DH5a, and E. coli GM2163
ꢁ ꢁ
(dam dcm ) were used for routine DNA manipulation. E. coli BL21(DE3)/
pLysS was used as the host for protein expression. Streptomyces tanashiensis
JCM4086 (RIKEN) was used for the isolation of DHK.
Plasmids
DNA Manipulation
The plasmids used for complementation of the DactVA-5,6 mutant were con-
structed as follows. The actVA-ORF5 gene was amplified with the primers
actVA5-Nde and actVA5-Bam. The NdeI-BamHI fragment was inserted
into the Streptomyces expression plasmid pIJ8600 (Kieser et al., 2000) that
has the Thio-inducible tipA promoter, generating pIJ8600+actVA5. Similarly,
the actVA-ORF6 gene was amplified with the primers actVA6-Nde and
Plasmid isolation, restriction enzyme digestion, ligation, and transformation of
E. coli and Streptomyces were performed as described previously (Kieser
et al., 2000; Sambrook et al., 1989). PCR was performed with LA Taq (TaKaRa,
Japan) or KOD -Plus- (Toyobo, Japan) DNA polymerase. PCR primers used in
this study are listed in Table S1.
Chemistry & Biology 16, 226–236, February 27, 2009 ª2009 Elsevier Ltd All rights reserved 233

Chemistry & Biology
Alternative Quinone Formation for Actinorhodin
actVA6-EcoXba. After digestion with NdeI and XbaI, the PCR product was
cloned into pIJ8600, yielding pIJ8600+actVA6.
(pH 7.5), 1 mM dithiothreitol, and 20% (w/v) glycerol. Aliquots of the purified
ꢂ
protein were stored at ꢁ70 C.
The E. coli expression plasmids used for purification of recombinant
proteins were constructed as follows. The actVA-ORF5 gene was excised
from pIJ8600+actVA5 as the NdeI-BamHI fragment and cloned into pET-
His-ActVB
ꢂ
E. coli BL21(DE3)/pLysS harboring pET-actVB was grown at 37 C to an OD600
of 0.3 in 100 ml LBGAC medium. Expression was induced with 250 mM IPTG.
ꢂ
19b. The resulting plasmid, pET-actVA5, attaches the N-terminal His-tag
To minimize the formation of inclusion bodies, the culture was cooled to 25 C
sequence. Plasmid pET-actVA6 was constructed by using the DNA fragment
amplified with primers actVA6-Nde and actVA6-Xho. The NdeI-XhoI fragment
was inserted into pET-22b. The resulting plasmid, pET-actVA6, attaches the
C-terminal His-tag sequence. The actVB gene was amplified with primers
actVB-Nde and actVB-Bam, and the PCR fragment was cloned into pET-
after the addition of IPTG and then further grown for 5 hr at this temperature.
ꢂ
Cells were collected by centrifugation and were stored at ꢁ70 C. The cell
pellet was resuspended in buffer C (20 mM sodium phosphate [pH 7.4],
500 mM NaCl, 100 mM imidazole, 2 mM dithiothreitol, 0.1% CHAPS, 10%
(w/v) glycerol) containing Complete protease inhibitor cocktail and sonicated.
After centrifugation, the crude cell extract was loaded onto a 1 ml HisTrap HP
column equilibrated with buffer C. Then, the column was washed with buffer C
containing 300 mM imidazole, and the His-ActVB protein was eluted with 500
mM imidazole in the same buffer. The fractions containing His-ActVB were
pooled and dialyzed against a buffer containing 25 mM Tris-HCl (pH 7.5),
19b, yielding pET-actVB.
Culture Conditions for Production of ACT-Related Metabolites
Seed cultures (10 ml) were grown in TSB medium (Kieser et al., 2000) on
ꢂ
a rotary shaker at 220 rpm and 30 C for 2 days. An aliquot (2 ml) of the seed
0
.1% CHAPS, and 10% (w/v) glycerol. Aliquots of the purified protein were
ꢂ
culture was inoculated to 50 ml R5MS medium (consisting of 10 g/l glucose,
stored at ꢁ70 C.
5
g/l yeast extract [Difco], 0.1 g/l casamino acids, 0.25 g/l K
MgCl $6H O, 5.73 g/l TES buffer, 2 ml/l trace element solution [pH 7.2]) in
a 500 ml Erlenmyer flask, and the culture was grown with shaking (220 rpm)
2 4
SO , 10.12 g/l
2
2
Substrates for In Vitro Enzyme Assay
ꢂ
ꢁ
Anthrone and dithranol were purchased from Aldrich and SIGMA, respec-
tively. Anthraquinone (Wako Pure Chemical Industries, Ltd.) and chrysazin
at 30 C for 4 days. R5 medium (Huang et al., 2001) was used for solid
cultures.
(
Tokyo Chemical Industry Co., Ltd.) were also purchased for the authenticity
of the reaction products. Emodinanthrone 11 was prepared by reduction of
emodin 12 (SIGMA) by using excess hydrogen iodide (55%) in glacial acetic
acid under reflux for 3 hr as previously described (Chen et al., 1995).
HPLC Analysis
An aliquot of culture (1 ml) was centrifuged at 13,200 3 g for 5 min. The super-
natant (100 ml) was directly subjected to reversed-phase HPLC analysis on
a Waters Multisolvent System under the following conditions: column, TSK
ꢁ
HR-ESIMS data of the resultant emodinanthrone: m/z [M-H] 255.0662
(
11 4
calculated for C15H O , 255.0657). DHK 10 was extracted from liquid culture
ꢂ
gel ODS-100V (4.6 id 3 150 mm, TOSOH); column temperature, 40 C;
of S. tanashiensis JCM4086 (Johnson and Dietz, 1968) and purified. S. tana-
shiensis was cultured in R5MS medium under the same conditions as those
used for S. coelicolor in this work. The crude extract obtained with ethyl
acetate was purified by silica gel column chromatography eluting chloro-
form/ethyl acetate/acetic acid (80:20:1) and preparative HPLC by using
a Waters Multisolvent System under the following conditions: column, TSK
gradient elution, solvent A (0.5% acetic acid in CH
3
CN) and solvent B
(0.5% acetic acid in deionized H O); gradient profile: 0–35 min, 10%–94%
2
A; 35–40 min, 94% A; 40–45 min, 94%–10% A; flow rate, 1.0 ml/min; detec-
tion, absorption between 250 and 600 nm using a photo-diode array detector
(Waters 2996, Waters).
ꢂ
gel ODS-100S (7.8 id 3 300 mm, TOSOH); column temperature, 40 C;
Expression and Purification of Recombinant ActVA-ORF5,
ActVA-ORF6, and ActVB Proteins
His-ActVA-ORF5
solvent, 45% aq. CH
2.0 ml/min.
3
CN containing 0.5% acetic acid (isocratic); flow rate,
ꢂ
E. coli BL21(DE3)/pLysS harboring pET-actVA5 was grown at 37 C to an
OD600 of 0.7 in 100 ml LB medium containing 0.2% glucose, 50 mg/ml ampi-
cillin, and 25 mg/ml chloramphenicol (LBGAC). Expression was induced with
In Vitro Enzyme Assay
Quinone-forming activity was measured spectrophotometrically by following
the increase of absorbance at 490 nm resulting from the formation of emodin
ꢂ
1
mM isopropyl-b-D-thiogalactopyranoside (IPTG), and incubation was
ꢂ
from emodinanthrone at 25 C (Chen et al., 1995). Conversion of emodin-
continued at 37 C for an additional 3 hr. Cells were collected by centrifugation
ꢂ
anthrone to emodin also occurs nonenzymatically, especially under alkaline
conditions. Under our assay conditions (pH 7.0), however, the contribution
of the nonenzymatic conversion was slight and always less than 10% of that
of the enzymatic conversion. The assay mixture (500 ml) for the ActVA-
ORF5/ActVB system contained 500 mM NADH, 5 mM FMN, 30% (v/v) ethylene
glycol monomethyl ether, 120 U/ml catalase, 100 nM His-ActVB, 1 mM His-
ActVA-ORF5, and various amounts of emodinanthrone in 25 mM PIPES-
NaOH (pH 7.0). Emodinanthrone in ethyleneglycol monomethylether was
subjected to an assay. The assay mixture (500 ml) for ActVA-ORF6 contained
2 mM ActVA-ORF6-His, 30% (v/v) ethyleneglycol monomethylether, and
various amounts of the substrate in 25 mM PIPES-NaOH (pH 7.0). The enzy-
matic conversion of emodinanthrone to emodin was also confirmed by
LC/ESIMS analysis. A typical analytical sample was prepared from the fore-
going assay mixture as follows. The mixture was extracted with ethyl acetate
(500 ml) followed by evaporation to dryness. The residue was dissolved
in acetone (100 ml), and an aliquot of the solution (5 ml) was subjected to
and stored at ꢁ70 C. The cell pellet was resuspended in buffer A (20 mM
sodium phosphate [pH 7.4], 500 mM NaCl, 100 mM imidazole, 2 mM dithio-
threitol, 0.1% CHAPS) containing Complete protease inhibitor cocktail (Roche)
and sonicated. Insoluble material was removed by centrifugation, and the
crude cell extract was chromatographed on a 1 ml HisTrap HP column (GE
Healthcare) equilibrated with buffer A. Then, the column was washed with
buffer A containing 300 mM imidazole, and the His-ActVA-ORF5 protein was
eluted with 500 mM imidazole in the same buffer. The fractions containing
His-ActVA-ORF5 were pooled and dialyzed against a buffer containing
20 mM Tris-HCl (pH 7.5), 2 mM dithiothreitol, 0.1% CHAPS, and 10% (w/v)
glycerol. Protein concentration was determined by the Bradford procedure.
ꢂ
Aliquots of the purified protein were stored at ꢁ70 C.
ActVA-ORF6-His
ꢂ
E. coli BL21(DE3)/pLysS transformed with pET-actVA6 was grown at 37 C to
an OD600 of 0.7 in 100 ml LBGAC medium. Expression was induced with 1 mM
ꢂ
IPTG, and incubation was continued at 37 C for an additional 3.5 hr. Cells were
ꢂ
LC/ESIMS analysis. LC/ESIMS analysis was carried out on a Waters
harvested by centrifugation and stored at ꢁ70 C. The cell pellet was resus-
AllianceHT equipped with a photo-diode array detector (Waters 2995) under-
pended in buffer B (20 mM sodium phosphate [pH 7.4], 500 mM NaCl,
the following conditions: column, TSK gel ODS-100S (4.6 id 3 150 mm,
ꢂ
2
0 mM imidazole, 1 mM dithiothreitol, 10% (w/v) glycerol) containing Complete
TOSOH); column temperature, 40 C; solvent, 45% aq. CH
3
CN containing
protease inhibitor cocktail and sonicated. After centrifugation, the crude cell
extract was chromatographed on a 1 ml HisTrap HP column equilibrated
with buffer B. Then, the column was washed with buffer B containing
0.5% acetic acid (isocratic); flow rate, 1.0 ml/min; detection, absorbance
between 250 and 600 nm. Mass spectrometry data were obtained under elec-
tron spray ionization (ESI) on a Waters LCT-Premier.
1
3
00 mM imidazole, and the ActVA-ORF6-His protein was eluted with
00 mM imidazole in the same buffer. The fractions containing ActVA-ORF6-
DHK monooxygenase activity was detected by LC/ESIMS analysis as
described above. DHK (50 mM) was reacted with His-ActVA-ORF5 (1 mM)
and His-ActVB (100 nM).
His were pooled and dialyzed against a buffer containing 20 mM Tris-HCl
234 Chemistry & Biology 16, 226–236, February 27, 2009 ª2009 Elsevier Ltd All rights reserved

Chemistry & Biology
Alternative Quinone Formation for Actinorhodin
SUPPLEMENTAL DATA
Huang, J., Lih, C.J., Pan, K.H., and Cohen, S.N. (2001). Global analysis of
growth phase responsive gene expression and regulation antibiotic biosyn-
thetic pathways in Streptomyces coelicolor using DNA microarrays. Genes
Dev. 15, 3183–3192.
Supplemental Data include five figures and one table and can be found with
this article online at http://www.cell.com/chemistry-biology/supplemental/
S1074-5521(09)00040-4.
Ichinose, K., Bedford, D.J., Tornus, D., Bechthold, A., Bibb, M.J., Revill, W.P.,
Floss, H.G., and Hopwood, D.A. (1998a). The granaticin biosynthetic gene
cluster of Streptomyces violaceoruber T u¨ 22: sequence analysis and expres-
sion in a heterologous host. Chem. Biol. 5, 647–659.
ACKNOWLEDGMENTS
We thank Yutaka Ebizuka for valuable discussion and David A. Hopwood for
critical reading of the manuscript. Junya Ochiai is acknowledged for his
technical assistance. The authors are grateful for funding to the Effective
Promotion of Joint Research of Special Coordination Funds (to K.O.), Grant-
in-Aid for Young Scientists from the Ministry of Education, Culture, Sports,
Science and Technology (MEXT) (18710189 to T.T.), and Grant-in-Aid for
Scientific Research from the Japan Society for the Promotion of Science
Ichinose, K., Taguchi, T., Ebizuka, Y., and Hopwood, D.A. (1998b). Biosyn-
thetic gene clusters of benzoisochromanequinone antibiotics in Streptomyces
spp. –identification of genes involved in post-PKS tailoring steps–. Actinomy-
cetologica 12, 99–109.
Ichinose, K., Surti, C., Taguchi, T., Malpartida, F., Booker-Milburn, K.I.,
Stephenson, G.R., Ebizuka, Y., and Hopwood, D.A. (1999). Proof that the actVI
genetic region of Streptomyces coelicolor A3(2) is involved in stereospecific
pyran ring formation in the biosynthesis of actinorhodin. Bioorg. Med. Chem.
Lett. 9, 395–400.
(20510205 to K.I.). A part of this work was financially supported by the
‘‘High-Tech Research Center’’ Project for Private Universities: matching fund
subsidy from MEXT, 2004–2008, and the Uehara Memorial Foundation to K.I.
Ichinose, K., Ozawa, M., Itou, K., Kunieda, K., and Ebizuka, Y. (2003).
Cloning, sequencing, and heterologous expression of the medermycin
biosynthetic gene cluster of Streptomyces sp. AM-7161: towards comparative
analysis of the benzoisochromanequionone gene clusters. Microbiology 149,
Received: November 19, 2008
Revised: January 22, 2009
Accepted: January 28, 2009
Published: February 26, 2009
1
633–1645.
Itoh, T., Taguchi, T., Kimberley, M.R., Booker-Milburn, K.I., Stephenson, G.R.,
Ebizuka, Y., and Ichinose, K. (2007). Actinorhodin biosynthesis: structural
requirements for post-PKS tailoring intermediates revealed by functional anal-
ysis of ActVI-ORF1 reductase. Biochemistry 46, 8181–8188.
REFERENCES
Bechthold, A., Sohng, J.K., Smith, T.M., Chu, X., and Floss, H.G. (1995).
Identification of Streptomyces violaceoruber T u¨ 22 genes involved in the
biosynthesis of granaticin. Mol. Gen. Genet. 248, 610–620.
Johnson, L.E., and Dietz, A. (1968). Kalafungin, a new antibiotic produced by
Streptomyces tanashiensis strain Kala. Appl. Microbiol. 16, 1815–1821.
Beltran-Alvarez, P., Cox, R.J., Crosby, J., and Simpson, T.J. (2007). Dissecting
the component reactions catalyzed by the actinorhodin minimal polyketide
synthase. Biochemistry 46, 14672–14681.
Kendrew, S.G., Harding, S.E., Hopwood, D.A., and Marsh, E.N.G. (1995).
Identification of a flavin:NADH oxidoreductase involved in the biosynthesis
of actinorhodin. Purification and characterization of the recombinant enzyme.
J. Biol. Chem. 270, 17339–17343.
Bentley, S.D., Chater, K.F., Cerde n˜ o-T a´ rraga, A.M., Challis,G.L., Thomson, N.R.,
James, K.D., Harris, D.E., Quail, M.A., Kieser, H., Harper, D., et al. (2002).
Complete genome sequence of the model actinomycete Streptomyces coeli-
color A3(2). Nature 417, 141–147.
Kendrew, S.G., Hopwood, D.A., and Marsh, E.N.G. (1997). Identification of
a monooxygenase from Streptomyces coelicolor A3(2) involved in biosyn-
thesis of actinorhodin: purification and characeterization of the recombinant
enzyme. J. Bacteriol. 179, 4305–4310.
Caballero, J.L., Martinez, F., Malpartida, F., and Hopwood, D.A. (1991). Orga-
nisations and functions of the actVA region of the actinorhodin biosynthetic
gene cluster of Streptomyces coelicolor. Mol. Gen. Genet. 230, 401–412.
Kieser, T., Bibb, M.J., Buttner, M.J., Chater, K.F., and Hopwood, D.A. (2000).
Practical Streptomyces Genetics (Norwich, UK: The John Innes Foundation).
Chater, K.F., and Hopwood, D.A. (1993). Streptomyces. In Bacillus subtilis and
Other Gram-Positive Bacteria: Biochemistry, Physiology, and Molecular
Genetics, A.L. Sonenshein, J.A. Hoch, and R. Losick, eds. (Washington,
D.C.: American Society for Microbiology), pp. 83–89.
Kim, I.C., and Oriel, P.J. (1995). Characterization of the Baicillus stearothermo-
philusBR219phenolhydroxylase gene.Appl. Environ. Microbiol. 61, 1252–1256.
Korman, T.P., Tan, Y.H., Wong, J., Luo, R., and Tsai, S.C. (2008). Inhibition
kinetics and emodin cocrystal structure of a type II polyketide ketoreductase.
Biochemistry 47, 1837–1847.
Chen, Z.-G., Fujii, I., Ebizuka, Y., and Sankawa, U. (1995). Purification and
characterization of emodinanthrone oxygenase from Aspergiluus terreus.
Phytochemistry 38, 299–305.
Malpartida, F., and Hopwood, D.A. (1984). Molecular cloning of the whole
biosynthetic pathway of a Streptomyces antibiotic and its expression in
a heterologous host. Nature 309, 462–464.
Chung, J.Y., Harada, S., Fujii, I., Ebizuka, Y., and Sankawa, U. (2002). Expres-
sion, purification, and characterization of AknX anthrone oxygenase, which is
involved in aklavinone biosynthesis in Streptomyces galilaeus. J. Bacteriol.
McDaniel, R., Ebert-Khosla, S., Hopwood, D.A., and Khosla, C. (1993).
Engineered biosynthesis of novel polyketides. Science 262, 1546–1550.
1
84, 6115–6122.
Cole, S.P., Rudd, B.A.M., Hopwood, D.A., Chang, C.-J., and Floss, H.G.
1987). Biosynthesis of the antibiotic actinorhodin. Analysis of blocked mutants
Munro, A.W., Girvan, H.M., and McLean, K.J. (2007). Variation on a (t)heme –
novel mechanisms, redox partners and catalytic functions in the cytochrome
P450 superfamily. Nat. Prod. Rep. 24, 585–609.
(
of Streptomyces coelicolor. J. Antibiot. (Tokyo) 40, 340–347.
Rudd, B.A.M., and Hopwood, D.A. (1979). Genetics of actinorhodin biosyn-
Fern a´ ndez-Moreno, M.A., Martinez, E., Boto, L., Hopwood, D.A., and
Malpartida, F. (1992). Nucleotide sequence and deduced functions of a set
of cotranscribed genes of Streptomyces coelicolor A3(2) including the
polyketide synthease for the antibiotic actinorhodin. J. Biol. Chem. 267,
thesis by Streptomyces coelicolor A3(2). J. Gen. Microbiol. 114, 35–43.
Sambrook, J., Fritsch, E.F., and Maniatis, T. (1989). Molecular Cloning:
A Laboratory Manual, Second Edition (Cold Spring Harbor, NY: Cold Spring
Harbor Laboratory).
19278–19290.
Fern a´ ndez-Moreno, M.A., Martinez, E., Caballero, J.L., Ichinose, K.,
Hopwood, D.A., and Malpartida, F. (1994). DNA sequence and functions of
the actVI region of the actinorhodin biosynthetic gene cluster of Streptomyces
coelicolor A3(2). J. Biol. Chem. 269, 24854–24863.
Sciara, G., Kendrew, S.G., Miele, A.E., Marsh, N.G., Federici, L., Malatesta, F.,
Schimperna, G., Sanvino, C., and Vallone, B. (2003). The structure of ActVA-
Orf6, a novel type of monooxygenase involved in actinorhodin biosynthesis.
EMBO J. 22, 205–215.
Floriano, B., and Bibb, M. (1996). afsR is a pleiotropic but conditionally
required regulatory gene for antibiotic production in Streptomyces coelicolor
A3(2). Mol. Microbiol. 21, 385–396.
Shen, B., and Hutchinson, C.R. (1993). Tetracenomycin F1 monooxygenase:
oxidation of naphthacenone to a naphthacenequinone in the biosynthesis of
tetracenomycin C in Streptomyces glaucescens. Biochemistry 32, 6656–6663.
Chemistry & Biology 16, 226–236, February 27, 2009 ª2009 Elsevier Ltd All rights reserved 235

Chemistry & Biology
Alternative Quinone Formation for Actinorhodin
Sherman, D.H., Malpartida, F., Bibb, M.J., Kieser, H.M., Bibb, M.J., and
Hopwood, D.A. (1989). Structure and deduced function of the granaticin-
producing polyketide synthase gene cluster of Streptomyces violaceoruber
T u¨ 22. EMBO J. 8, 2717–2725.
Valton, J., Filisetti, L., Fontecave, M., and Nivi e` re, V. (2004). A two-component
flavin-dependent monooxygenase involved in actinorhodin biosynthesis in
Streptomyces coelicolor. J. Biol. Chem. 279, 44362–44369.
Valton, J., Fontecave, M., Douki, T., Kendrew, S.G., and Nivi e` re, V. (2006).
An aromatic hydroxylation reaction catalyzed by a two-component FMN-
dependent monooxygenase. The ActVA-ActVB system from Steptomyces
coelicolor. J. Biol. Chem. 281, 27–35.
Taguchi, T., Itou, K., Ebizuka, Y., Malpartida, F., Hopwood, D.A., Surti, C.M.,
Booker-Milburn, K.I., Stephenson, G.R., and Ichinose, K. (2000). Chemical
characterisation of disruptants of the Streptomyces coelicolor A3(2) actVI
genes involved in actinorhodin biosynthesis. J. Antibiot. (Tokyo) 53,
Valton, J., Mathevon, C., Fontecave, M., Nivi e` re, V., and Ballou, D.P. (2008).
Mechanism and regulation of the two-component FMN-dependent monooxy-
genase ActVA-ActVB from Streptomyces coelicolor. J. Biol. Chem. 283,
144–152.
1
0287–10296.
Taguchi, T., Okamoto, S., Itoh, T., Ebizuka, Y., Ochi, K., and Ichinose, K.
2008). Actinoperylone, a novel perylenequinone-type shunt product, from
(
Zhao, B., Lamb, D.C., Lei, L., Kelly, S.L., Yuan, H., Hachey, D.L., and
Waterman, M.R. (2007). Different binding modes of two flaviolin substrate
molecules in cytochrome P450 158A1 (CYP158A1) compared to CYP158A2.
Biochemistry 46, 8725–8733.
a deletion mutant of the actVA-ORF5 and ORF6 genes for actinorhodin
biosynthesis in Streptomyces coelicolor A3(2). Tetrahedron Lett. 49,
1208–1211.
236 Chemistry & Biology 16, 226–236, February 27, 2009 ª2009 Elsevier Ltd All rights reserved