Staurosporine and rebeccamycin aglycones are assembled by the oxidative action of StaP, StaC, and RebC on chromopyrrolic acid

Article abstract of DOI:10.1021/ja063898m

In the biosynthesis of the antitumor indolocarbazoles rebeccamycin and staurosporine by streptomycetes, assembly of the aglycones involves a complex set of oxidative condensations. Overall formation of aglycones K252c and arcyriaflavin A from their biosynthetic precursor chromopyrrolic acid involves four- and eight-electron oxidations, respectively. This process is catalyzed by the remarkable enzyme StaP, with StaC and RebC acting to direct the level of oxidation in the newly formed five-membered ring. An aryl-aryl coupling reaction is integral to this transformation as well as oxidative decarboxylation of the dicarboxypyrrole moiety of chromopyrrolic acid. Herein we describe the heterologous expression of staP, staC, and rebC in Escherichia coli and the activity of the corresponding enzymes in constructing the two distinct six-ring scaffolds. StaP is a cytochrome P450 enzyme, requiring dioxygen, ferredoxin, flavodoxin NADP⁺-reductase, and NAD(P)H for activity. StaP on its own converts chromopyrrolic acid into three aglycone products, K252c, arcyriaflavin A, and 7-hydroxy-K252c; in the presence of StaC, K252c is the predominant product, while the presence of RebC directs formation of arcyriaflavin A. ¹⁸O-Labeling studies indicate that the oxygen(s) of the pyrrolinone and maleimide functionalities of the aglycones formed are all derived from dioxygen. This work allowed for the in vitro reconstitution of the full biosynthetic pathway from L-tryptophan to the staurosporine and rebeccamycin aglycones, K252c and 1,11-dichloroarcyriaflavin A.

Full text of DOI:10.1021/ja063898m

Published on Web 08/23/2006
Staurosporine and Rebeccamycin Aglycones Are Assembled
by the Oxidative Action of StaP, StaC, and RebC on
Chromopyrrolic Acid
Annaleise R. Howard-Jones and Christopher T. Walsh*
Contribution from the Department of Biological Chemistry and Molecular Pharmacology,
HarVard Medical School, Boston, Massachusetts 02115
Received June 13, 2006; E-mail: christopher_walsh@hms.harvard.edu
Abstract: In the biosynthesis of the antitumor indolocarbazoles rebeccamycin and staurosporine by
streptomycetes, assembly of the aglycones involves a complex set of oxidative condensations. Overall
formation of aglycones K252c and arcyriaflavin A from their biosynthetic precursor chromopyrrolic acid
involves four- and eight-electron oxidations, respectively. This process is catalyzed by the remarkable
enzyme StaP, with StaC and RebC acting to direct the level of oxidation in the newly formed five-mem-
bered ring. An aryl-aryl coupling reaction is integral to this transformation as well as oxidative decar-
boxylation of the dicarboxypyrrole moiety of chromopyrrolic acid. Herein we describe the heterologous
expression of staP, staC, and rebC in Escherichia coli and the activity of the corresponding enzymes in
constructing the two distinct six-ring scaffolds. StaP is a cytochrome P450 enzyme, requiring dioxygen,
+
ferredoxin, flavodoxin NADP -reductase, and NAD(P)H for activity. StaP on its own converts chromo-
pyrrolic acid into three aglycone products, K252c, arcyriaflavin A, and 7-hydroxy-K252c; in the presence of
StaC, K252c is the predominant product, while the presence of RebC directs formation of arcyriaflavin A.
18
O-Labeling studies indicate that the oxygen(s) of the pyrrolinone and maleimide functionalities of the
aglycones formed are all derived from dioxygen. This work allowed for the in vitro reconstitution of the full
biosynthetic pathway from L-tryptophan to the staurosporine and rebeccamycin aglycones, K252c and
1,11-dichloroarcyriaflavin A.
Introduction
The indolocarbazole antibiotics rebeccamycin and staurospo-
rine (Figure 1) have attracted much attention due to their broad-
spectrum antitumor activity. Despite their striking structural
similarity, these two compounds exert their biological activity
via two distinct pathways. Rebeccamycin and its analogues
bearing simple N-glycosidic linkages act to stabilize the DNA-
topoisomerase I cleavable complex, with a minimum inhibitory
1
concentration of 1.75 µM. By contrast, staurosporine is one of
the strongest known inhibitors of protein kinases, with IC50s in
2
the 1-20 nM range. Its interaction with the ATP binding pocket
of these enzymes explains the ability of staurosporine to interact
with most known protein kinases, an activity shared by its bis-
linked sugar-indolocarbazole analogues. The presence of the
carbohydrate moiety is important for recognition by these
cellular targets, as is the planar six-ring scaffold.
Figure 1. Chemical structures of rebeccamycin 1 and staurosporine 2.
flat scaffold intercalates between the DNA base pairs, and the
sugar moiety forms hydrogen bonding interactions in the major
3
Crystal structures have been solved of rebeccamycin analogue
groove. In the staurosporine-protein kinase structure, the
3
SA315F in complex with DNA and human topoisomerase I,
aglycone portion fits snugly into the ATP-binding pocket,
roughly superimposing with the position of the purine moiety
of ATP, and the indole nitrogens and sugar hydroxyls form key
and of staurosporine in complex with the catalytic subunit of
4
cAMP-dependent protein kinase. In the case of SA315F, the
4
hydrogen-bonding interactions. These structures clearly dem-
(
1) Moreau, P.; Anizon, F.; Sancelme, M.; Prudhomme, M.; Bailly, C.; Severe,
onstrate the importance of the planar, aromatic aglycone core
in directing these indolocarbazoles toward their cellular targets.
D.; Riou, J.-F.; Fabbro, D.; Meyer, T.; Aubertin, A.-M. J. Med. Chem.
1
999, 42, 584-592.
(2) Tamaoki, T.; Nomoto, H.; Takahashi, I.; Kato, Y.; Morimoto, M.; Tomita,
F. Biochem. Biophys. Res. Commun. 1986, 135, 397-402.
(
3) Staker, B. L.; Feese, M. D.; Cushman, M.; Pommier, Y.; Zembower, D.;
Stewart, L.; Burgin, A. B. J. Med. Chem. 2005, 48, 2336-2345.
(4) Prade, L.; Engh, R. A.; Girod, A.; Kinzel, V.; Huber, R.; Bossemeyer, D.
Structure 1997, 5, 1627-1637.
10.1021/ja063898m CCC: $33.50 © 2006 American Chemical Society
J. AM. CHEM. SOC. 2006, 128, 12289-12298
9
12289

A R T I C L E S
Howard-Jones and Walsh
Scheme 1. Overall Biosynthetic Route to Rebeccamycin 1 and Staurosporine 2
The rebeccamycin and staurosporine indolocarbazole cores
are derived from L-tryptophan (L-Trp) via a series of oxidative
transformations (Scheme 1). Gene disruption studies combined
with isolation of putative intermediates led to an initial
description of the biosynthetic pathways to rebeccamycin and
is only now emerging. Halogenation of L-Trp by RebH (and its
8
flavin reductase partner protein RebF) has been demonstrated
and shown to involve formation of HOCl in the active site.9
The formation of chromopyrrolic acid (CPA), a key biosynthetic
intermediate, through the joint action of RebO, an L-amino acid
5
-7
10
staurosporine.
Unsurprisingly, the pathways to the two
oxidase, and RebD (or StaD), a heme-containing oxidase, has
1
1-13
aglycones follow very similar routes, differing only by (1)
chlorination at the 7-position of L-Trp en route to rebeccamycin
and (2) the oxidation state of the pyrrole-derived five-membered
ring, taking the form of a maleimide in rebeccamycin and a
pyrrolinone in the staurosporine aglycone. Subsequent N-gly-
cosylation and tailoring modifications follow divergent pathways
toward rebeccamycin and staurosporine.
also been investigated
and shown to proceed via the
condensation of two molecules of indole-3-pyruvic acid imine.1
1
Most recently, in vivo studies of the N-glycosyltransferase
1
4
RebG and in vitro characterization of RebM, the methyl-
(
8) Yeh, E.; Garneau, S.; Walsh, C. T. Proc. Natl. Acad. Sci. U.S.A. 2005,
102, 3960-3965.
(
9) Yeh, E.; Cole, L. J.; Barr, E. W.; Bollinger, J. M., Jr.; Ballou, D. P.; Walsh,
C. T. Biochemistry 2006, 45, 7904-7912.
In vitro reconstitution and analysis of the biosynthetic
enzymes involved in rebeccamycin and staurosporine formation
(10) Nishizawa, T.; Aldrich, C. C.; Sherman, D. H. J. Bacteriol. 2005, 187,
2084-2092.
(
(
11) Howard-Jones, A. R.; Walsh, C. T. Biochemistry 2005, 44, 15652-15663.
12) Nishizawa, T.; Gr u¨ schow, S.; Jayamaha, D.-H. E.; Nishizawa-Harada, C.;
Sherman, D. H. J. Am. Chem. Soc. 2005, 128, 724-725.
(
(
(
5) S a´ nchez, C.; Butovich, I. A.; Bra n˜ a, A. F.; Rohr, J.; M e´ ndez, C.; Salas, J.
A. Chem. Biol. 2002, 9, 519-531.
(13) Asamizu, S.; Kato, Y.; Igarashi, Y.; Furumai, T.; Onaka, H. Tetrahedron
Lett. 2006, 47, 473-475.
6) Onaka, H.; Taniguchi, S.-I.; Igarashi, Y.; Furumai, T. J. Antibiot. 2002,
5
5, 1063-1071.
(14) Zhang, C.; Albermann, C.; Fu, X.; Peters, N. R.; Chisholm, J. D.; Zhang,
G.; Gilbert, E. J.; Wang, P. G.; Van Vranken, D. L.; Thorson, J. S. Chem.
BioChem. 2006, 7, 795-804.
7) Onaka, H.; Taniguchi, S.-I.; Igarashi, Y.; Furumai, T. Biosci. Biotechnol.
Biochem. 2003, 67, 127-138.
12290 J. AM. CHEM. SOC.
9
VOL. 128, NO. 37, 2006

Staurosporine and Rebeccamycin Aglycone Assembly
A R T I C L E S
at 0.5 mL‚min- [0 f 100% acetonitrile in water, with 0.1% formic
1
transferase that catalyzes the final step in rebeccamycin bio-
synthesis, were reported.14,15
acid] in positive ion mode. NMR spectra were recorded on a Varian
00 MHz Fourier Transform NMR spectrometer. UV/visible spectra
6
To date, no biochemical studies on the enzymes involved in
aglycone ring construction, StaP and StaC (or RebP and RebC),
have been reported. Gene knockouts of rebC produced three
were collected on a Cary 50 Bio UV-visible spectrophotometer. DNA
sequencing was performed by the Biopolymers Facility, Department
of Genetics, Harvard Medical School, Boston, and by the Molecular
Biology Core Facility, Dana-Farber Cancer Institute, Boston. Chemical
ionization mass spectrometry was performed by the Mass Spectrometry
Service, Department of Chemistry and Chemical Biology, Harvard
University, Boston.
distinct aglycone products, 1,11-dichloro-K252c, 1,11-dichloro-
7
7
-hydroxy-K252c, and 1,11- dichloroarcyriaflavin A 7. Simi-
larly, constructs bearing rebO, rebD, and rebP (or staP) resulted
in the formation of three productssK252c 4, 7-hydroxy-K252c
5
, and arcyriaflavin A 6swhile the additional presence of staC
DNA Isolation and Manipulation. LecheValieria aerocolonigenes
(ATCC 39243) and Streptomyces longisporoflaVus (DSM 10189) were
grown in yeast-malt extract medium (3 g‚L yeast extract; 5 g‚L
or rebC led to exclusive formation of K252c 4 or arcyriaflavin
A 6, respectively (Scheme 1).16
-1
-1
malt extract; pH 7.0) for preparation of chromosomal genomic DNA.
The cultures were incubated at 26 °C and 250 rpm for 18 h prior to
isolation of genomic DNA. Standard methods for DNA isolation and
manipulation were performed as described by Sambrook et al. DNA
fragments were isolated from agarose gels using a QIAGEN gel
extraction kit.
RebP and StaP have been annotated as cytochrome P450
enzymes and display the appropriate signature motifs, including
a (FXXGXHXCXG) motif bearing the putative axial cysteine
for heme-iron ligation. RebC and StaC have been annotated as
flavin adenine dinucleotide (FAD) monooxygenases and contain
the characteristic (GXGXXG) motif of the Rossman âRâ
dinucleotide binding fold,17 as well as the GD motif believed
to be important for hydrogen bonding to the 3′-hydroxyl of the
FAD ribosyl moiety.18 The RebP and StaP proteins share 54%
identity at the amino acid level, and RebC and StaC are 65%
identical. Previous work on enzymes from the reb and sta
systemssnamely, RebO, RebD, and RebGshas shown that they
2
2
Cloning, Expression, and Purification of StaP, StaC, and
RebC. Primers for staP (5′-GGAGAGCATATGCCATCCGCGAC-
GCTGC-3′ and 5′-GTCAAGCTTGGGGTGGCTGGCCGAGGG-3′),
staC (5′-GGAGAGCATATGACGCATTCCGGTGAGCGGACC-3′
and 5′- GTCAAGCTTGCCCCGCGGCTCACGGGGCGCGGC-3′) and
rebC (5′- GGAGAGCATATGAACGCGCCCATCGAA-3′ and 5′-
GTCAAGCTTTCACGCGGCACCCCTCAC- 3′) contained NdeI and
HindIII sites (italicized) and were used to amplify the relevant genes
of interest. These orfs were cloned into the corresponding NdeI/HindIII
sites of pET28a (rebC) and pET22b (staP and staC). The sequences
of the cloned expression vectors were confirmed by DNA sequencing.
Some disparities were observed with the published sequences of S.
are relatively promiscuous in terms of accepting substrates from
the alternative (Sta) system.10,11,14
The overall reaction catalyzed by StaP is intriguing as it
involves both aryl-aryl coupling and double oxidative decar-
boxylation to facilitate net four- to eight-electron oxidation.
While aryl-aryl coupling reactions are not uncommon in
biosynthetic pathways, few enzymes mediating such steps have
been examined in biochemical detail.19-21 The opportunity to
study this class of enzymatic reactions, as well as the intriguing
oxidative decarboxylation process, lent the StaP/StaC (and RebP/
RebC) system much appeal as a focus of biosynthetic and
mechanistic investigations.
23
longisporoflaVus staP and staC; these are described in Supporting
Information and have been deposited at GenBank under accession
numbers DQ861904 and DQ861905, respectively.
E. coli BL21(DE3) cells overexpressing staP were grown in Luria-
Bertani medium supplemented with ampicillin (100 µg‚mL-1). After
22 h growth at 15 °C, OD600 ) 0.4 was reached, at which point
δ-aminolevulinic acid (1 mM) and ferrous ammonium sulfate (32 mM)
were added. The cells were induced with 100 µM isopropyl â-D-
thiogalactoside (IPTG) and grown for a further 14 h at 15 °C. E. coli
BL21(DE3) cells overexpressing staC were grown in Luria-Bertani
Experimental Section
-
1
medium supplemented with ampicillin (100 µg‚mL ). Cells were
grown at 15 °C to OD600 ) 0.5 and then induced with 100 µM IPTG
and grown for a further 72 h at 15 °C. E. coli BL21(DE3) cells
overexpressing rebC were grown in Luria-Bertani medium supple-
Amino acids, spinach ferredoxin, and other general reagents were
18
purchased from Sigma-Aldrich; ( O)-labeled water was obtained from
1
8
Cambridge Isotope Laboratories; and ( O)-labeled oxygen was pur-
1
1
chased from Chemgas. CPA was prepared as previously described.
-
1
+
mented with kanamycin (50 g‚mL ). Cells were grown at 15 °C to
OD600 ) 0.4 and then induced with 100 µM IPTG and grown for a
further 72 h at 15 °C.
E. coli flavodoxin NADP -reductase was overproduced and purified
(construct kindly provided by Prof. Rowena Matthews), giving 10 mg
protein per liter culture. E. coli TOP10 and BL21(DE3) cells were
obtained from Invitrogen. pET28a and pET22b plasmids were purchased
from Novagen. Restriction enzymes were obtained from New England
Biolabs. Nickel-nitrilotriacetic acid-agarose (Ni-NTA) gel was
obtained from QIAGEN. The HiLoad 26/60 Superdex 200 column used
was obtained from GE Healthcare. Electrospray mass spectra were
collected by liquid chromatography-mass spectrometry (LCMS), using
a Shimadzu LCMS-QP8000R equipped with a Higgins Analytical
Cells were harvested by centrifugation (20 min at 3000g) and
resuspended in 20 mM Tris-HCl (pH 8.0), 300 mM NaCl, 5 mM
imidazole. Resuspended cells were lysed (2 passes at 10000-15000
psi, Avestin EmulsiFlex-C5 high-pressure homogenizer), and the cell
debris was removed by centrifugation (35 min at 95000g).
6 6 6
C-His -tagged-StaP, C-His -tagged-StaC, and N-His -tagged-RebC
were purified by Ni-NTA affinity chromatography. Eluted protein
fractions were purified further by passage through a HiLoad 26/60
Superdex 200 gel filtration column; fractions containing the protein of
interest were collected and combined. (RebC could be reconstituted
with FAD to 55% occupancy by incubating with 10 equiv of FAD at
(Mountain View, CA) Sprite Targa C18 column (20 × 2.1 mm) running
(
(
15) Zhang, C.; Weller, R. L.; Thorson, J. S.; Rajski, S. R. J. Am. Chem. Soc.
2
006, 128, 2760-2761.
16) S a´ nchez, C.; Zhu, L.; Bra n˜ a, A. F.; Salas, A. P.; Rohr, J.; M e´ ndez, C.;
Salas, J. A. Proc. Natl. Acad. Sci. U.S.A. 2005, 102, 461-466.
4
°C for 2 h prior to gel filtration; however, the sample prior to
(
(
17) Rossman, M. G.; Moras, D.; Olsen, K. W. Nature 1974, 250, 194-199.
reconstitution (with 33% FAD) showed higher activity and was used
18) Eppink, M. H. M.; Schreuder, H. A.; Van Berkel, W. J. H. Protein Sci.
1
997, 6, 2454-2458.
(
19) Funa, N.; Funabashi, M.; Ohnishi, Y.; Horinouchi, S. J. Bacteriol. 2005,
(22) Sambrook, J.; Russell, D. W. Molecular cloning: a laboratory manual,
3rd ed.; Cold Spring Harbor Laboratory Press: Cold Spring Harbor, NY,
2000.
(23) Schupp, T.; Engel, N.; Bietenhader, J.; Toupet, C.; Pospiech, A. U.S. Pat.
6,210,935; 2001.
1
87, 8149-8155.
(
(
20) Zhao, B.; et al. J. Biol. Chem. 2005, 280, 11599-11607.
21) Zerbe, K.; Woithe, K.; Li, D. B.; Vitali, F.; Bigler, L.; Robinson, J. A.
Angew. Chem., Int. Ed. 2004, 43, 6709-6713.
J. AM. CHEM. SOC.
9
VOL. 128, NO. 37, 2006 12291

A R T I C L E S
Howard-Jones and Walsh
+
for activity assays.) The proteins were flash-frozen in liquid nitrogen
and stored in 25 mM HEPES, pH 7.5, 10% glycerol (with 150 mM
NaCl, in the case of StaC and RebC) at -80 °C. Protein concentrations
were determined using the method of Bradford, with bovine serum
spinach ferredoxin (20 µM), E. coli flavodoxin NADP -reductase (1
µM), and NAD(P)H (5 mM) in 75 mM HEPPS buffer, pH 8.0.
Reactions were quenched by addition of methanol (2 volumes), and
the protein was removed by centrifugation. HPLC assays were run on
a Beckman System Gold (Beckman Coulter) with a Vydac C18 column
24
-1
albumin (BSA) as a standard. The total yield of protein was 3 mg‚L
culture for C-His
mg‚L for N-His
6
-StaP, 11 mg‚L- culture for C-His
1
6
-StaC and 18
(250 × 10 mm) at 1 mL‚min or a Higgins analytical C18 column
-1
-
1
-1
6
-RebC.
(50 × 4.6 mm) at 3 mL‚min using a gradient of 0 f 60% acetonitrile
Biochemical Characterization of StaP, StaC, and RebC. The
molecular masses of the proteins were determined by gel filtration,
using known standards (Sigma-Aldrich) as reference. To determine the
reduced carbon monoxide difference spectrum of StaP, sodium dithion-
ite was added (to a final concentration of 1 mM) to an anaerobic cuvette
containing StaP (21 µM), after which CO (g) was bubbled through the
solution for 5 min. UV/visible spectra (200-800 nm) were taken before
and after this procedure, and the difference spectrum calculated. The
binding curve of CPA for StaP was determined by measuring a
difference UV/visible absorption spectrum following addition of varying
quantities of CPA to a solution containing 37 µM StaP in 75 mM
HEPES, pH 7.5 (difference spectra measured relative to solution with
no CPA). The difference absorbance value at 423 nm was subtracted
from the value at 388 nm, and the resulting parameter (dA) was plotted
in 0.1% trifluoroacetic acid (TFA). The elution profiles were monitored
at 280, 289, or 315 nm.
Anaerobic Experiments. Anaerobic assays were performed in a
Unilab glovebox (Mbraun, Stratham, NH), with oxygen levels main-
tained at or below 2 ppm. All reagents and solutions were degassed by
bubbling argon for 10 min prior to their introduction into the glovebox,
then equilibrated overnight to remove residual traces of oxygen.
Incubations were quenched immediately prior to removal from the
glovebox by addition of two volumes of methanol.
18O-Incorporation Experiments. 18O-Incorporation studies were
used to determine the origin of the oxygen atom(s) in the aglycone
products. CPA (150 µM) was incubated with StaP (1 µM), StaC or
+
RebC (5 µM), ferredoxin (20 µM), flavodoxin NADP -reductase (1
-
1
µM), NADPH (5 mM), and BSA (1 mg.mL ) in 75 mM HEPPS, pH
8.0 (50 µL), for 15 h in the presence of either O-labeled dioxygen or
O-labeled water. After quenching with methanol (2 volumes), the
supernatant was purified by HPLC, and the peaks corresponding to
the relevant aglycone products were collected and analyzed by
electrospray mass spectrometry (positive ion mode).
In Vitro Reconstitution of Complete Pathway to the Rebecca-
mycin Aglycone 1,11-Dichloroarcyriaflavin A 7. 7-Chloro-L-tryp-
tophan (5 mM) was incubated with RebO (1 µM), RebD (3 µM), StaP
(1 µM), and RebC (5 µM) in the presence of NADH (5 mM), ferredoxin
(20 µM), flavodoxin NADP -reductase (1 µM), and BSA (1 mg‚mL )
in HEPES, pH 7.5, at room temperature for 46 h, with further addition
of NADH (10 mM), ferredoxin (20 µM), flavodoxin NADP -reductase
(1 µM), StaP (1 µM), and RebC (5 µM) after 22 h and additional NADH
(10 mM) at 25 h. The product 7 was purified by reversed phase HPLC
and analyzed by chemical ionization mass spectrometry (positive ion
mode). The compound is observed by mass spectrometry as the (ring
opened) maleamic acid form, its isotope pattern consistent with that of
18
against CPA concentration to obtain the K
D
for CPA.
1
8
The RebC cofactor was identified as FAD by HPLC and UV/visible
spectroscopy. A sample of protein was denatured (100 °C, 30 min)
and the supernatant analyzed by HPLC, with comparison to standard
samples of FAD and flavin mononucleotide (FMN). The FAD content
of the protein was determined using the known extinction coefficient
-
1
-1
for FAD (ꢀ450 11 300 M ‚cm ). Apo-RebC was prepared using the
His-tag immobilization method of Hefti et al.2
5
The reduced â-nicotinamide adenine dinucleotide 2′-phosphate
+
-1
(
NADPH) and NADH oxidase activity of RebC and StaC (10 µM)
were determined by following the consumption of NADPH (or NADH)
200 µM initial concentration) over time in the presence of FAD in 75
mM HEPPS, pH 8.0. NAD(P)H kinetic parameters were determined
at 400 µM FAD; K values for FAD were determined at 200 µM
NADH. NADPH and NADH consumption rates were calculated using
340 values, with reference to a standard curve, and corrected for
+
(
m
A
background air oxidation of NADPH or NADH. Reaction progress was
accompanied by an appropriate decrease in the absorbance at 450 nm,
+
1,11-dichloroarcyriaflavin A 7: m/z [M + H] 412 (97%), 413 (22%),
consistent with the reduction of FAD to FADH
2
.
414 (100%), 415 (20%).
Biochemical StaP/StaC Pull-Down Experiments. To facilitate co-
transformation into E. coli BL21(DE3) cells and pull-down experiments,
the staP and staC genes were subcloned into pRSF-1b (as N-terminal
Synthesis of Aglycone Standards, Arcyriaflavin A 6, and 7-Hy-
droxy-K252c 5. Arcyriaflavin A 6 was synthesized from arcyriarubin
A 9 by palladium(II)-mediated aryl-aryl coupling, according to the
method of Harris et al.,26 to give a yellow solid with spectral properties
6
His -tagged constructs) and into pET22b (as untagged constructs). Thus,
consistent with those in the literature.2 H NMR (600 MHz, d
7 1
-acetone)
6
), 9.16 (d, 2H, J 7.5
staP and staC were subcloned from pET28a into pRSF-1b using NcoI
and HindIII restriction sites, and from pET28a into pET22b using NdeI
and HindIII sites. NHis-staP-pRSF-1b and staC-pET22b were co-
transformed into E. coli BL21(DE3) cells; NHis-staC-pRSF-1b and
staP-pET22b were co-transformed similarly. The resulting E. coli
BL21(DE3) cells overproducing NHis-StaP/StaC or NHis-StaC/StaP
6
δ 11.10 (br s, 2H, H12 and H13), 9.79 (br s, 1H, H
Hz, H
4
and H
8
), 7.72 (d, 2H, J 8.0 Hz, H
1
and H11), 7.55 (t, 2H, J 7.5
+
Hz, H
2
and H10), 7.37 (t, 2H, J 7.5 Hz, H
+
3
and H
9
) ppm; m/z (ES ) 326
[M + H] .
The synthesis of 7-hydroxy-K252c 5 was achieved by lithium
(
2 L each) were grown in Luria-Bertani broth supplemented with
aluminum hydride-mediated reduction of arcyriaflavin A 6, as described
in the literature.2 H NMR (600 MHz, d
-acetone) δ 10.91, 10.77 (2
6 1
δ-aminolevulinic acid (1 mM), ferrous ammonium sulfate (40 µM),
ampicillin (100 µg‚mL ), and kanamycin (50 µg‚mL ). After 17 h,
cultures reached OD600 ) 0.7, at which point they were induced with
6
-
1
-1
× br s, 2H, H and H ), 9.33 (d, 1H, J 7.9 Hz, 1 of H or H ), 8.50
12
13
4
8
(d, 1H, J 7.9 Hz, 1 of H
8.3 Hz, 1 of H or H11), 7.63 (d, 1H, J 8.2 Hz, 1 of H
1H, J 7.6 Hz, 1 of H or H10), 7.43 (t, 1H, J 7.6 Hz, 1 of H
7.31 (t, 1H, J 7.3 Hz, 1 of H
4
or H
8
), 7.76 (br s, 1H, H
6
), 7.68 (d, 1H, J
or H11), 7.46 (t,
or H10),
6
0 µM IPTG. Growth was continued for a further 28 h at 15 °C. Cells
1
1
were harvested and lysed and the debris pelleted as described above.
The lysate was loaded on pre-equilibrated Ni-NTA resin and eluted
with increasing concentrations of imidazole (5 to 200 mM) in 20 mM
Tris, pH 8, 300 mM NaCl.
2
2
3
or H
9
), 7.27 (t, 1H, J 7.2 Hz, 1 of H
3
+
or H
9
), 6.58 (s, 1H, H
7
), 5.27 (br s, 1H, 7-OH) ppm; m/z (ES ) 328
+
[M + H] .
HPLC Activity Assays. Conversion of CPA 3 to aglycones 4, 5,
and/or 6 was examined by analytical reversed phase HPLC. A typical
assay examined the turnover of 150 µM CPA by StaP (1 µM) and
Results
Purification and Characterization of StaP, StaC, and
RebC. StaP was overproduced and purified as a C-terminal His6-
-
1
StaC (5 µM) or RebC (5 µM), in the presence of BSA (1 mg‚mL ),
(
24) Bradford, M. M. Anal. Biochem. 1976, 72, 248-254.
25) Hefti, M. H.; Milder, F. J.; Boeren, S.; Vervoort, J.; Van Berkel, W. J. H.
Biochim. Biophys. Acta 2003, 1619, 139-143.
(26) Harris, W.; Hill, C. H.; Keech, E.; Malsher, P. Tetrahedron Lett. 1993, 34,
8361-8364.
(27) Bergman, J.; Pelcman, B. J. Org. Chem. 1989, 54, 824-828.
(
12292 J. AM. CHEM. SOC.
9
VOL. 128, NO. 37, 2006

Staurosporine and Rebeccamycin Aglycone Assembly
A R T I C L E S
RebC was overproduced and purified at 18 mg‚L- culture
as an N-terminal His6-tagged 59 kDa protein (Figure 2B). It
contained 33% bound FAD on isolation and could be recon-
stituted to 55% bound FAD, though the protein as isolated (with
1
33% FAD) was used in activity assays due to improved turnover
at this level of FAD incorporation, as discussed below. StaC
was overproduced and purified as a C-terminal His6-tagged 61
-
1
kDa protein at 11 mg‚L culture (Figure 2C). It contained no
bound cofactors and could not be reconstituted with FAD. RebC
ran predominantly as a monomer on gel filtration, whereas StaC
showed the elution profile of a monomer/dimer mixture, with
the dimer being the predominant oligomeric state.
The kinetic profile of StaC and RebC in oxidizing NAD-
(P)H in the presence of FAD and FMN was examined. No
activity was observed in the presence of FMN or, in the case
of RebC, in the presence of FAD and NADPH. Saturation could
not be achieved under the assay conditions for the StaC-induced
turnover of NAD(P)H and FAD due to high Km values for NAD-
Figure 2. SDS-PAGE gel of (A) StaP (48 kDa); (B) RebC (59 kDa); and
(C) StaC (61 kDa).
(P)H and the saturation limits of the detector; kcat/Km values
were extracted by linear regression analysis of the line generated
by a plot of velocity versus NAD(P)H concentration. StaC
-
1
-1
displayed a kcat/Km of 30 ( 3 mM min with NADH and
2
-
1
-1
.3 ( 0.2 mM min with NADPH; in the case of RebC,
kinetic parameters could only be extracted for NADH oxidase
-
1
activity (kcat 8.3 ( 1.4 min ; Km 138 ( 58 µM; kcat/Km 60 (
-
1
-1
2
7 mM min ). NADH is thus the preferred reductant for
StaC- and RebC-mediated turnover of FAD.
As saturation of NADH could not be examined under these
assay conditions, an apparent Km for FAD was calculated at
2
00 µM NADH. The Km(apparent) for FAD was 140 ( 30 µM
for StaC and 70 ( 20 µM for RebC. The relatively high Km
values for FAD are not surprising given the low observed
binding affinity for FAD (especially in the case of StaC).
Nonetheless, these data collectively suggest that RebC and StaC
are capable of reducing FAD in situ and may therefore function
as flavin- dependent monooxygenases.
To investigate the possibility of complex formation between
StaP and StaC, biochemical pull-down experiments were
conducted using cells overproducing both StaP and StaC
simultaneously, one as a His6-tagged construct and the other as
an untagged protein. These studies showed no evidence for a
stable complex between StaP and StaC. The His6-tagged protein
bound to Ni-NTA resin and required high imidazole concentra-
tions for its elution, while its untagged partner eluted in the
initial flow-through.
Aglycone Formation by StaP, StaC, and RebC. The
catalytic cycle of a cytochrome P450 typically involves the four-
electron reduction of dioxygen to water coupled to a two-
electron substrate oxidation, the additional two electrons coming
from an external redox partner system. The most commonly
observed redox partner system for cytosolic prokaryotic P450s
is a two-component ferredoxin/ferredoxin reductase (or fla-
vodoxin/flavodoxin reductase) system; eukaryotic P450s are
often membrane-bound proteins that interact with a membrane-
Figure 3. (A) Reduced carbon monoxide binding spectrum of StaP in active
blue) and inactive (red) forms; (B) type I substrate binding curve; and (C)
dissociation curve for interaction of CPA with StaP.
(
tagged protein of molecular mass 48 kDa (3 mg‚L-1 culture)
2
8
bound cytochrome P450 reductase. NAD(P)H is the ultimate
electron source for both of these systems. As StaP displayed
all of the characteristic traits of a cytochrome P450 enzyme, a
(Figure 2A). The protein displayed the typical reduced carbon
monoxide difference spectrum of a cytochrome P450, in contrast
to alternative (inactive) preparations of the enzyme, which gave
the P420 form (Figure 3A). A type I binding spectrum was
observed on interaction with its putative substrate CPA (Figure
(28) McLean, K. J.; Sabri, M.; Marshall, K. R.; Lawson, R. J.; Lewis, D. G.;
Clift, D.; Balding, P. R.; Dunford, A. J.; Warman, A. J.; McVey, J. P.;
Quinn, A. -M.; Sutcliffe, M. J.; Scrutton, N. S.; Munro, A. W. Biochem.
Soc. Trans. 2005, 33, 796-801.
3
B), with a KD of 17.2 ( 1.1 µM (Figure 3C).
J. AM. CHEM. SOC.
9
VOL. 128, NO. 37, 2006 12293

A R T I C L E S
Howard-Jones and Walsh
ferredoxin/flavodoxin NADP+ reductase system was investi-
gated as the potential redox partner for this enzymatic reaction.
A control assay omitting StaC and RebC from the reaction
showed no difference in rate between 0 and 0.5 µM FAD, but
significant rate suppression at 10 µM FAD (data not shown).
Indeed, incubation of StaP with CPA, NADH (or NADPH),
+
+
A 20-fold increase in flavodoxin NADP -reductase concentra-
spinach ferredoxin, and E. coli flavodoxin NADP -reductase
tion restored 40% of the activity lost due to the increase in FAD
concentration to 10 µM; increases in ferredoxin concentration
gave no such rate enhancement (data not shown). This indicates
that the rate enhancement at low FAD concentrations (up to
under aerobic conditions led to the formation of several products
(Figure 4A), the major three coeluting with authentic standards
for compounds 4, 5, and 6. These were formed in an approx-
imately 1:7:1 ratio. The identities of these products were further
confirmed by comparison of the UV/visible absorbance spectra
and mass spectra with authentic materials. The inclusion of StaC
in this reaction mixture led primarily to the formation of K252c
0
.5 µM) is probably due to interaction of FAD with StaC (or
RebC), whereas the inhibitory effect of FAD at high concentra-
+
tions is likely due to a flavodoxin NADP -reductase dependent
effect. We postulate that this may be due to reduction of FAD
4
; the presence of RebC gave predominantly arcyriaflavin A 6.
in solution by the reductase or due to equilibration of internally
Incubation of CPA with StaC (or RebC), NAD(P)H, ferredoxin,
+
+
generated FADH in the flavodoxin NADP -reductase active
and flavodoxin NADP -reductase in the absence of StaP gave
2
site with free FAD in solution. Therefore the optimal FAD
no observable turnover; similarly, omission of NAD(P)H, ferre-
+
concentration for this assay betrays a fine balance between
competing components of the redox system involved in this
reaction.
doxin, or flavodoxin NADP -reductase from the StaP/StaC reac-
tion mixture prevented the formation of any aglycone products.
+
Spinach ferredoxin NADP -reductase (Sigma) was also able
to reduce spinach ferredoxin and hence facilitate the StaP-
Attempted Turnover of Possible Biosynthetic Intermedi-
ates. In order to probe the overall reaction pathway, we
investigated the ability of the StaP/StaC and StaP/RebC systems
to turn over possible biosynthetic intermediates to the final
aglycone products. Incubation of arcyriarubin A 9 with RebC
and/or StaP under the previously described conditions gave no
observable formation of aglycones 4, 5, or 6 (Figure 4C-i).
Similarly, incubation of K252c 4 or 7- hydroxy-K252c 5 with
StaP and/or RebC gave no formation of any other aglycone
products (Figure 4C-ii and -iii). Incubation of arcyriaflavin A
mediated reaction, however due to supply difficulties, E. coli
+
flavodoxin NADP -reductase was used instead. Cytochrome
P450 reductase could not substitute for the ferredoxin/flavodoxin
+
NADP -reductase system. The reaction was optimized in terms
+
of pH, NAD(P)H, ferredoxin, and flavodoxin NADP -reductase
concentrations and StaP:StaC (or StaP:RebC) ratio.
The assay was also carried out under anaerobic conditions
to assess the oxygen-dependence of this reaction. Aglycone 4
was formed on introduction of oxygen (Figure 4B, +O2 trace),
as previously discussed, but no aglycone formation was observed
under anaerobic conditions (Figure 4B, -O2 trace). New peaks
were observed to form that were not seen in aerobic incubation
mixtures; however, these were also formed in control anaerobic
6
with StaC and/or StaP gave no conversion to K252c 4 or 7-
hydroxy-K252c 5 (Figure 4C-iv). Thus it seems that arcyriarubin
A 9 is not a biosynthetic intermediate in the conversion of CPA
to aglycones 4, 5, and 6. Furthermore, aglycones 4, 5, and 6
are not interconvertible, but appear to diverge from a common
biosynthetic intermediate.
+
incubations of ferredoxin and flavodoxin NADP -reductase in
HEPPS buffer (Figure 4B). These peaks therefore arise inde-
pendently of any interaction with CPA, StaP, or StaC.
18O-Incorporation Experiments. Incubations of CPA with
StaP/StaC and StaP/RebC were conducted in the presence of
O-labeled dioxygen or O-labeled water to determine the
In addition to the turnover of CPA to aglycones 4, 5, and 6,
the complete reconstitution of the biosynthetic pathway from
1
8
18
origin of the oxygen atoms in K252c 4 and arcyriaflavin A 6.
The mass spectra of the aglycone products thus obtained were
determined, demonstrating that the amide oxygen of K252c 4
and both oxygens of arcyriaflavin A 6 are derived from dioxygen
7
-chloro-L-tryptophan (7-Cl-L-Trp) to the rebeccamycin agly-
cone 1,11-dichloroarcyriaflavin A 7 was also demonstrated.
Thus, incubation of 7-Cl-L-Trp with RebO, RebD,11 StaP, and
RebC, along with NADH, ferredoxin, and flavodoxin NADP -
reductase, resulted in the formation of 1,11-dichloroarcyriaflavin
+
(Figure 6).
A 7. This result demonstrates the ability of StaP to accept the
Incubation of CPA with StaP and StaC in the presence of
O-labeled water and unlabelled dioxygen gave a product with
18
1,11-dichloro-CPA analogue, produced as an intermediate in
this assay by the action of RebO and RebD. Similar promiscuity
has also been observed in the behavior of earlier enzymes in
the rebeccamycin and staurosporine clusters.
m/z 312, identical to the standard K252c 4. To the contrary,
incubation in the presence of unlabelled water and 18O-labeled
dioxygen led to an increase of two mass units in the K252c
product obtained. This mass shift is consistent with the
10-12
Dependence of Reaction Rate on FAD. The StaP/StaC and
StaP/RebC reactions showed complex biphasic behavior with
respect to their dependence on FAD concentration (Figure 5).
While exogenous FAD was not required for the StaP/StaC
reaction, a 3.6-fold enhancement of kcat was observed on
inclusion of 0.5 µM FAD (0.1 equiv per StaC) in the reaction
mixture. As FAD concentrations were increased beyond this
level, an inhibitory effect was observed, with the rate at 10 µM
FAD dropping more than 6.5-fold relative to the rate observed
in the absence of FAD. FAD gave a similar inhibition profile
in the reaction of StaP and RebC with CPA, the optimal FAD
concentration being at 0.33 equiv per RebC molecule. This ratio
corresponds to the FAD content isolated in the native protein.
1
8
incorporation of oxygen from O-dioxygen into K252c 4.
Similar assays were conducted with CPA in the presence of
StaP and RebC to look for oxygen-18 incorporation into
arcyriaflavin A 6. Product 6 obtained on incubation with
unlabelled oxygen gave m/z 326, as seen for the native product.
18
Incubation with O-labeled dioxygen and unlabelled water gave
a mass shift of +4 (m/z 330), in accord with the incorporation
1
8
of two molecules of O from the labeled dioxygen. Thus it
seems that the two maleimide oxygens of 6 and the one amide
oxygen of 4 are directly derived from dioxygen.
Steady-State Kinetic Studies of StaP/StaC and StaP/RebC
Turnover. The kinetics of formation of aglycones 4 and 6 by
12294 J. AM. CHEM. SOC.
9
VOL. 128, NO. 37, 2006

Staurosporine and Rebeccamycin Aglycone Assembly
A R T I C L E S
Figure 5. Effect of FAD concentration on observed rate of formation of
(
6
A) K252c 4 by StaP (1 µM) and StaC (5 µM) and of (B) arcyriaflavin A
by StaP (1 µM) and RebC (5 µM) from CPA 3 in the presence of 75 mM
-
1
HEPPS, pH 8, 1 mg‚mL BSA, 5 mM NADPH, 20 µM ferredoxin, and 1
µM flavodoxin NADP -reductase.
+
Table 1. Steady-State Kinetic Parameters for Turnover of CPA 3
by StaP (1 µM) and StaC (5 µM) or RebC (5 µM) in the Presence
-
1
of 75 mM HEPPS, pH 8.0, 1 mg‚mL BSA, 5 mM NAD(P)H, 20
+
µM Ferredoxin, 1 µM Flavodoxin NADP -Reductase, and 0.5 µM
FAD (StaP/StaC Assays) or 1.7 µM FAD (StaP/RebC Assays).
StaP/StaC
NADPH
StaP/RebC
NADH NADPH
NADH
kcat (min-1)
Km (mM)
3.6 ( 0.9 0.61 ( 0.05 4.1 ( 0.6 1.3 ( 0.1
76 ( 28
0.047
35 ( 9
0.017
-
26 ( 6
0.15
26 ( 6
0.050
-
kcat/Km (mM min-1
-1
)
Ki (mM)
93 ( 35
84 ( 19
the StaP/StaC and StaP/RebC systems, respectively, were
studied under steady-state conditions, and the relevant param-
eters are given in Table 1. Interestingly, both the StaP/StaC and
StaP/RebC systems showed a marked preference for NADH
over NADPH as the ultimate reductant in this reaction. For both
the StaC and RebC reactions, a 3-fold increase in catalytic
efficiency is seen for CPA turnover in the presence of NADH
versus that seen in the presence of NADPH. In contrast, the
rate of aglycone formation by StaP in isolation does not change
if NADH is substituted for NADPH. This overall rate effect
therefore seems directly attributable to a StaC/RebC-catalyzed
reaction.
In the presence of NADH, the Km values of StaP/StaC and
StaP/RebC for CPA differ by approximately 3-fold, but the kcat
values are comparable (Table 1). This gives a 3-fold increase
in catalytic efficiency in the formation of arcyriaflavin A 6 by
StaP/RebC, relative to the StaP/StaC-mediated formation of
K252c 4. Due to substantial substrate inhibition observed in
both cases above approximately 70 µM, it is important to note
Figure 4. HPLC traces (280 nm) of products formed on incubation of (A)
CPA (150 µM) with StaP (1 µM) and StaC (5 µM)sor RebC (5 µM)sin
-
1
the presence of 75 mM HEPPS, pH 8, 1 mg‚mL BSA, 5 mM NADPH,
+
2
0 µM ferredoxin, and 1 µM flavodoxin NADP -reductase for 3 h; (B) in
the presence and absence of dioxygen (anaerobic control incorporating only
ferredoxin (20 µM) and flavodoxin NADP -reductase (1 µM) in 75 mM
HEPPS, pH 8, reveal that new peaks are not CPA-derived); and (C) on
+
incu-bation of StaP (1 µM) in the presence of 75 mM HEPPS, pH 8, 1
-
1
mg‚mL BSA, 5 mM NADPH, 20 µM ferredoxin, and 1 µM flavodoxin
+
NADP -reductase with (i) arcyriarubin A 9 (100 µM), (ii) 7-hydroxy-K252c
5
(150 µM), (iii) K252c 4 (150 µM), and (iv) arcyriaflavin A 6. Traces for
parts A and C were obtained using a Higgins analytical C18 column (50
mm × 4.6 mm); part B traces were obtained using a Vydac C18 column
(
250 mm × 10 mm). (Identical traces were seen on incubation of each of
the aglycones with RebC (5 µM) or StaC (5 µM) in the presence or absence
of StaP.)
J. AM. CHEM. SOC.
9
VOL. 128, NO. 37, 2006 12295

A R T I C L E S
Howard-Jones and Walsh
that the calculated kcat values are never actually realized. The
rates of formation of K252c 4 and arcyriaflavin A 6 by StaP in
isolation are an order of magnitude lower than those seen for
the formation of 4 and 6 in the presence of StaC or RebC,
respectively (data not shown). Thus, while StaP is catalytically
competent to turn over CPA 3 to the aglycone products 4, 5,
and 6, the presence of StaC or RebC give significant boosts in
reaction rates as well as providing oxidative control.
Discussion
The extensive oxidative transformations required to generate
the rebeccamycin and staurosporine aglycone scaffolds from
L-Trp (or 7-Cl-L-Trp) are catalyzed by only four enzymes. The
formation of the planar six-ring aglycone core involves a 10-
to 14-electron net oxidation of the starting pair of L-Trp (or
7-Cl-L-Trp) molecules. At the crux of this biosynthetic machin-
ery is StaP, which alone can mediate a four- to eight-electron
oxidation of its substrate, CPA 3.
StaP, a cytochrome P450 enzyme, catalyzes an aryl-aryl
bond-forming reaction to give a six-ring indolocarbazole scaf-
fold, as well as mediating decarboxylation and oxidation of the
putative dicarboxypyrrole moiety. This action presumably
requires two to four cycles of net two-electron substrate oxi-
dation at the catalytic heme center. StaP is remarkable, not
only due to the extensive oxidation processes it catalyzes, but
also because it produces three distinct products, differing in
oxidation level. For the production of K252c 4 from CPA 3,
a net four-electron oxidation is required; the generation of
arcyriaflavin A 6 from 3 requires an overall eight-electron
oxidation. StaP is thus unusual in the apparent lack of oxi-
dative control it possesses over the outcome of its catalytic
turnover.
Control of the overall oxidation route is provided by a second
enzyme StaC, which imparts the net effect of directing the
oxidation level of the pyrrole-derived ring. While StaP in
isolation gives three aglycone forms, StaP and StaC turn over
CPA to give only a single product, K252c 4. Similarly, RebC
guides the turnover of CPA 3 toward the more highly oxidized
maleimide-bearing aglycone, arcyriaflavin A 6. This in vitro
behavior of StaP, StaC, and RebC exactly mirrors the previously
reported in vivo results.7
+
Figure 6. (A) Mass spectra (ES ) of products derived from reaction of
CPA 3 (150 µM) with StaP (1 µM) and StaC (5 µM) or RebC (5 µM) in
-1
the presence of 75 mM HEPPS, pH 8, 1 mg‚mL BSA, 5 mM NADPH,
+
18
20 µM ferredoxin, and 1 µM flavodoxin NADP -reductase, in O-labeled
O2 or water; and (B) biosynthetic route showing incorporation of oxygen
atoms from dioxygen.
,16
The interaction between StaC and StaP has been investigated
using biochemical pull-down experiments and gel filtration),
but no compelling evidence for a stable complex has been
observed to date. Although the formation of a weakly associating
catalytic complex cannot be ruled out, it seems that these two
proteins do not have high affinity for one another.
Exogenous FAD is not required for the catalytic activity of
StaC, but does give a 3.6-fold rate enhancement when 0.1 equiv
are included in the assay mixture. (A similar effect was observed
for RebC, with optimal activity occurring at 0.33 equiv of FAD
per RebC molecule.) Since the assays contained in this work
(
+
utilize flavodoxin NADP -reductase, which contains bound
The ability of StaC and RebC to oxidize NAD(P)H in the
presence of FAD suggests that these enzymes may be FAD-
dependent monooxygenases, in line with their previous annota-
tion.7 Given these results, it is interesting to note that StaC
does not form a tightly bound complex with FAD, but RebC
can be reconstituted up to 0.55 equiv of FAD. The high Kms
for FAD seen in NADH oxidation by StaC and RebC further
suggest that the FAD/StaC and FAD/RebC complexes may not
be strongly associating. The disparity in the FAD-binding
capacities of StaC and RebC is intriguing given the high overall
homology between these two proteins and the presence of
characteristic flavin-binding motifs in both. The relevance of
this difference in FAD-binding affinity to the catalytic abilities
of these enzymes remains to be determined.
FAD, it is possible that the baseline activity seen in the absence
of exogenous FAD derives from competition by StaC or RebC
for the reductase-bound flavin.
,16
At higher concentrations of FAD, aglycone formation is
significantly inhibited, in the presence or absence of StaC (or
RebC). That this inhibition can be partially ameliorated by
+
increasing the concentration of flavodoxin-NADP -reductase
suggests that it is due, at least in part, to an interaction between
+
FAD and flavodoxin NADP -reductase. The flavodoxin reduc-
tase mechanism involves the transient generation of reduced
29
flavin to facilitate reduction of ferredoxin. Excess FAD in
solution may compete with reduced flavin in the flavodoxin
(29) Wan, J. T.; Jarrett, J. T. Arch. Biochem. Biophys. 2002, 406, 116-126.
12296 J. AM. CHEM. SOC.
9
VOL. 128, NO. 37, 2006

Staurosporine and Rebeccamycin Aglycone Assembly
A R T I C L E S
Scheme 2. Possible Biosynthetic Routes to K252c 4, 7-Hydroxy-K252c 5, and Arcyriaflavin A 6
+
NADP -reductase active site or may be itself reduced by
lation or (b) oxidative decarboxylation followed by an aryl-
aryl coupling step (Scheme 2). The failure of StaP and/or RebC
to turn over arcyriarubin A 9 suggests that this is not a
biosynthetic intermediate on the pathway to arcyriaflavin A 6.
Hence the likely overall pathway involves initial aryl-aryl
coupling to intermediate 8 with subsequent decarboxylation and
oxidation. We suggest that StaP is the key aryl-aryl coupling
catalyst since StaP is absolutely required for turnover, and StaC
and RebC are incapable of reacting with CPA 3. This oxidative
coupling likely proceeds in one-electron steps, following initial
reduction of the heme center by reduced ferredoxin and oxygen
binding to generate a reactive ferryl-oxo intermediate, according
30
flavodoxin reductase, thus disrupting the redox chain necessary
for StaP reactivity. The optimal ratios of 1:10 for FAD:StaC
and 1:3 for FAD:RebC observed in this assay almost certainly
result from two competing equilibria between FAD/StaC (or
FAD/RebC) and FAD/flavodoxin reductase, rather than being
a direct indication of the optimal FAD:StaC (or FAD:RebC)
binding ratio for catalysis.
The derivation of the oxygen atoms in K252c 4 and
arcyriaflavin A 6 had not been previously investigated. Although
mechanistic logic would favor the incorporation of oxygen at
these positions from dioxygen, this had not been unequivocally
proven. In this report, we demonstrate the direct inclusion of
oxygen atoms from dioxygen at these positions through the use
of O-labeled water and dioxygen. The mass shift of +2 for
K252c 4 and +4 for arcyriaflavin A 6 upon incubation with
O-labeled dioxygen (relative to incubations with atmospheric
dioxygen) indicates that O2 is the source of the oxygens of the
pyrrolinone and maleimide rings of these indolocarbazoles. Due
to a high degree of uncoupling between the ferredoxin/flavo-
3
1
to the typical cytochrome P450 mechanism.
The putative intermediate 8 has not been isolated and is likely
quite unstable. Given its highly conjugated nature, decarboxy-
lation may occur spontaneously and, in an aerobic environment,
could be accompanied by autoxidation. Such a nonenzymatic
decarboxylation/oxidation pathway may go some way to
explaining the suite of products formed in the absence of StaC
and RebC.
We note that no combination of StaP, StaC, and RebC is
able to interconvert any of the three aglycones 4, 5, and 6,
suggesting that these aglycones are not on pathway to one
another. Thus, it seems that these three aglycones share a
common precursor earlier in the biosynthetic pathway and
diverge separately from this point. This result is somewhat
surprising, since it would seem particularly likely that arcyri-
aflavin A 6 arises by a two-electron oxidation of 7-hydroxy-
K252c 5, which may in turn be derived from K252c 4.
The increased rate of product formation observed in the
presence of StaC or RebC (relative to that seen with StaP alone)
18
1
8
+
doxin NADP -reductase system and StaP, it was not possible
to correlate rates of oxygen consumption with product formation.
However, based on the stoichiometry of the reaction, it seems
likely that four molecules of oxygen are consumed per molecule
of arcyriaflavin A 6 formed, and two oxygen molecules utilized
for the production of one molecule of K252c 4.
Our investigations into potential biosynthetic intermediates
have allowed us to narrow the possibilities for the overall
reaction pathway from CPA 3 to aglycones 4, 5, and 6. One
might envisage two overall pathways for the StaP-catalyzed
reaction: (a) initial aryl-aryl coupling followed by decarboxy-
(30) Uyeda, K.; Rabinowitz, J. C. J. Biol. Chem. 1971, 246, 3111-3119.
(31) Ullrich, V. Angew. Chem. Int. Ed. 1972, 11, 701-712.
J. AM. CHEM. SOC.
9
VOL. 128, NO. 37, 2006 12297

A R T I C L E S
Howard-Jones and Walsh
suggests that these enzymes play physiologically relevant roles
in the final steps of aglycone formation, in accord with in vivo
reports.7 The most remarkable difference between StaC and
RebC is that these two homologous NADH-flavin oxidoreduc-
tases catalyze overall flux of the reaction to two distinct
products, aglycones 4 and 6. These products are identical except
for a four-electron difference in the oxidation state at one
position in the pyrrole-derived ring.
StaC appears to direct the product flux such that the second
decarboxylation follows a non-oxidative pathway. By contrast,
RebC enables an oxidative pathway (e.g., by providing an
hydroxyl radical or cation equivalent) and then presumably
mediates a subsequent oxidation to give the maleimide ring of
RebC-mediated production of 6 in the presence of NADH
appears to be due to a direct interaction between NADH and
StaC or RebC. In addition, the dependence of the reaction rate
on FAD concentration suggests the involvement of StaC. A 3.5-
fold increase in catalytic efficiency for the StaP/StaC reaction
is observed when the FAD concentration is increased from 0
to 0.5 µM, but no rate enhancement is seen in the absence of
StaC or RebC. Taken together, these results suggest that the
enhanced reaction rates seen in the presence of StaC and RebC
are FAD- and NADH-dependent effects that involve StaC or
RebC in a catalytic manner.
Since StaC and RebC are not catalytically competent to turn
over CPA 3, it seems likely that the physiological role of StaP
is to form a key intermediate through mediating initial aryl-
aryl coupling. Decarboxylation of the highly conjugated six-
ring scaffold 8 would follow, perhaps via a nonenzymatic
process, with StaP, StaC, or RebC mediating aglycone formation
from this intermediate stage.
,16
6. Given the observed lack of turnover of 7-hydroxy-K252c 5,
this second oxidation (on compound 5) would have to occur
before its irreversible release to solvent. In this manifold, RebC
but not StaC would have oxygenase activity itself or function
to elicit that reactivity mode from StaP.
In summary, this work demonstrates reconstitution of the full
staurosporine and rebeccamycin biosynthetic pathways from
L-Trp (or 7-Cl-L-Trp) to the staurosporine aglycone K252c 4
and the rebeccamycin aglycone 1,11-dichloroarcyriaflavin A 7.
The results described herein specifically address the biosynthetic
route from CPA 3 to the staurosporine- and rebeccamycin-
related aglycones, K252c 4, 7-hydroxy-K252c 5, and arcyri-
aflavin A 6. We postulate that the overall reaction is initiated
by StaP-mediated aryl-aryl coupling of chromopyrrolic acid
to give a dicarboxylated six-ring intermediate 8 (Scheme 2).
Enzyme-mediated aryl-aryl coupling steps are poorly under-
stood processes. The robust aryl-aryl coupling reaction cata-
lyzed by StaP, as well as the subsequent oxidative decarbox-
ylation to form the rebeccamycin and staurosporine aglycone
scaffolds, lend the StaP/StaC enzyme system much appeal. The
significant oxidative changes mediated by StaP are particularly
striking given that these transformations require only a single
enzyme. Furthermore, the catalytic roles of StaC and RebC in
enhancing rates of aglycone formation and controlling the oxi-
dative outcome of the reaction are intriguing. This work repre-
sents a starting point for future mechanistic studies into the en-
zymes responsible for this complex and intriguing catalytic
cycle.
It is still unclear, however, as to whether StaC and RebC
exert their catalytic effects through direct action on an inter-
mediate or whether they act indirectly via StaP. These enzymes
may function directly as FAD monooxygenases, converting the
CPA-derived dicarboxypyrrole ring to the pyrrolinone and
maleimide rings of the aglycone products, 4 and 6. It is however
surprising that these indolocarbazole- producing actinomycetes
would have evolved two parallel (and fundamentally different)
systems for catalyzing the same reaction. Alternatively, they
may act indirectly, either through feeding electrons into the StaP
catalytic center or by exerting some allosteric effect on StaP,
rather than effecting substrate oxidation directly.
If StaC and RebC act only through modulating the activity
of StaP, the mechanism of this process is still unclear. For
example, it has not been possible to discount the option that
StaC and RebC simply function to feed electrons into the StaP
heme center, which would in turn effect catalytic turnover.
However, StaC and RebC cannot substitute for the ferredoxin/
+
flavodoxin NADP -reductase redox system, which is absolutely
required for catalysis, and hence it is difficult to rationalize an
obligate role for two redox feeder systems for the heme iron
center. Furthermore, there are no known redox partners bearing
FAD that directly interact with cytochrome P450 enzymes.
Rather, iron-sulfur clusters (e.g., ferredoxins) and FMN (e.g.,
cytochrome P450 reductases and flavodoxins) are preferred as
Acknowledgment. We thank Prof. Rowena Matthews for
+
provision of a flavodoxin NADP -reductase construct and Prof.
2
8
the immediate electron donors for the P450 heme center.
Robert S. Phillips for providing an authentic sample of 7-Cl-
L-Trp. We thank Dr. Wendy L. Kelly for helpful discussions.
This work was supported by National Institutes of Health Grant
GM020011.
Either way, however, it seems that the increased specificity
and efficiency of aglycone formation in the presence of StaC
and RebC are FAD- and NADH-dependent effects. In the
presence of both StaC and RebC (but not in their absence), an
increase in the rate of aglycone formation is observed in the
presence of NADH, relative to that seen with NADPH. NADH
is the preferred reductant for FAD oxidation by StaC, with a
Supporting Information Available: Sequences for staC and
staP from S. longisporoflaVus (DSM 10189) and complete ref
2
0. This information is available free of charge via the Internet
at http://pubs.acs.org.
1
3-fold enhancement in kcat/Km relative to that seen with
+
NADPH; flavodoxin- NADP -reductase shows the inverse
JA063898M
preference, with a 5-fold greater catalytic efficiency with
NADPH versus NADH.32 Therefore, the 3-fold kcat/Km improve-
(32) Leadbeater, C.; Mclver, L.; Campopiano, D. J.; Webster, S. P.; Baxter, R.
L.; Kelly, S. M.; Price, N. C.; Lysek, D. A.; Noble, M. A.; Chapman, S.
K.; Munro, A. W. Biochem. J. 2000, 352, 257-266.
ment for the StaP/StaC-mediated formation of 4 and the StaP/
12298 J. AM. CHEM. SOC.
9
VOL. 128, NO. 37, 2006