Mapping the Mechanism of the Resorcinol Ring Formation Catalyzed by ArsB, a Type III Polyketide Synthase from Azotobacter vinelandii

Full text of DOI:10.1002/cbic.201200346

DOI: 10.1002/cbic.201200346
Mapping the Mechanism of the Resorcinol Ring Formation Catalyzed by
ArsB, a Type III Polyketide Synthase from Azotobacter vinelandii
Sarah E. Posehn, Sun Young Kim, Andrew G. H. Wee,* and Dae-Yeon Suh*[a]
Polyketides, produced by polyketide synthases (PKSs), are
a large class of secondary metabolites widely found in bacteria,
fungi, plants, and marine animals. Polyketide-based drugs in-
clude antibiotics, immunosuppressants, antiparasitics, and cho-
lesterol-lowering, anticancer and antioxidant agents. PKSs are
classified into three types according to their architecture.
Type I PKSs are large, modular, multidomain enzymes, and
type II PKSs are dissociable multienzyme complexes. In type I
and II PKSs, each domain or enzyme typically performs a dis-
crete function.[1] In contrast, type III PKSs—homodimers com-
prising subunits of about 45 kDa—are multifunctional.[2] Each
subunit iteratively condenses a starter acyl-CoA substrate with
a number of acetate units derived from malonyl-CoA (2,
Scheme 1), and cyclizes the linear polyketide intermediate to
produce polyketides with distinct ring structures.
the triketo CoA thioester (3), produced after three condensa-
tion reactions could be converted to the final 5-substituted re-
sorcinol (7) through several different pathways. Resorcinol for-
mation can either begin with hydrolysis of the thioester bond
in 3 to give the triketocarboxylate 8 (pathway A), or begin with
aldol cyclization of 3 to give the cyclized thioester 9 (path-
way B). In the “hydrolysis first” pathway A, aldol cyclization of 8
can occur concomitantly with decarboxylation (pathway A1) to
give 10. Alternatively, the aldol cyclization of 8 can occur first
(pathway A2) to give an equilibrium mixture of 11, 12, and
possibly the cyclized dianion 13.[10] The cyclization might then
be followed by sequential b-keto decarboxylation and dehy-
dration (pathway A2.1) or by coupled decarboxylation/dehy-
dration (pathway A2.2) to give the final product 7. Alternative-
ly, the nonaromatic cyclized compound (11, 12, or 13) might
undergo dehydration and aromatization to give the substitut-
ed b-resorcylic acid 16, which is then decarboxylated to 7. For
the “aldol first” pathway B, the cyclized thioester 9 can either
be hydrolyzed to give the same equilibrium mixture of 11, 12,
and 13 (pathway B1), or 9 can undergo dehydration and aro-
matization to the b-resorcylic thioester 14, which then enters
into a hydrolysis/decarboxylation sequence to form 7 (path-
way B2).
Although structurally simpler than type I and II enzymes,
type III PKSs also produce a diverse array of polyketide prod-
ucts. The diversity comes from the choice of the starter CoA
substrate, the number of condensation steps, and the cycliza-
tion mechanism. ArsB and ArsC from Azotobacter vinelandii uti-
lize long-chain fatty acyl-CoA esters (1a),[3] whereas chalcone
synthase (CHS) and stilbene synthase (STS), the two most stud-
ied plant type III PKSs, use p-coumaroyl-CoA or cinnamoyl-CoA
(1b) as the starter substrate.[4,5] Although the number of decar-
boxylative condensations catalyzed by type III PKSs varies from
one[6] to seven,[7] the majority of type III PKSs catalyze three
condensation reactions to give triketo CoA thioester intermedi-
ates (commonly called tetraketide intermediates, 3) that can
be cyclized into different six-membered ring structures
(Scheme 1).[8] CHS catalyzes the C-6!C-1 Claisen acylation to
give a phloroglucinol derivative (4), whereas ArsC catalyzes O-
acylation to produce 2’-oxoalkyl-a-pyrones (5). In these two
cyclization reactions, CoA serves as a leaving group. On the
other hand, alkylresorcylic acid synthase (ARAS) connects the
C-2 methylene carbon with the C-7 carbonyl carbon through
an aldol condensation and also hydrolyzes the CoA thioester
to afford 6-alkyl-b-resorcylic acid (6).[9] STS and ArsB also cata-
lyze an aldol cyclization, but instead produce 5-substituted re-
sorcinols (7; Scheme 1).
A few studies have addressed the mechanism of STS-cata-
lyzed resorcinol ring formation. In an elegant study using deu-
terated malonyl-CoA and mass spectrometry, Shibuya et al.[11]
demonstrated that stilbenecarboxylate (16b) is not an inter-
mediate in the STS-catalyzed resorcinol ring formation and
proposed that thioester hydrolysis precedes aldol cyclization
and decarboxylation (pathway A). Funa et al.[3] suggested that
16a is not an intermediate in the ArsB-catalyzed resorcinol
ring formation, given that 16a was not detected in the reac-
tion mixture of ArsB. A parallel observation was made with STS
and 16b by Li et al.[12] Meanwhile, Austin et al.[5] proposed
pathway A2.2 (coupled decarboxylation/dehydration of 12) as
the most likely mechanism for STS-catalyzed resorcinol ring
formation, based on solution chemistry of biomimetic poly-
ketide cyclization[13] and a computer-assisted docking study.
However, direct evidence for the “hydrolysis first” hypothesis
has been lacking. One way to elucidate the mechanism of a
multistep enzymatic reaction is to examine the putative reac-
tion intermediates. In this study, we prepared a linear triketo-
carboxylate 3,5,7-trioxoeicosanoic acid (C20-TKA, 8a) and incu-
bated it with ArsB to determine the first step of the ArsB-cata-
lyzed resorcinol ring formation.
The resorcinol ring formation catalyzed by STS and ArsB in-
volves aldol cyclization, hydrolysis of the thioester, decarboxy-
lation, dehydration and aromatization. However, the sequence
of these reactions remains unresolved. As shown in Scheme 2,
[a] S. E. Posehn, Dr. S. Y. Kim, Prof. A. G. H. Wee, Prof. D.-Y. Suh
Department of Chemistry and Biochemistry, University of Regina
3737 Wascana Parkway, Regina, SK S4S 0A2 (Canada)
E-mail: andrew.wee@uregina.ca
Diketo acids and their dipotassium salts have commonly
been synthesized by treating the methyl ester with ethanolic
KOH.[14,15] Attempts to use the same strategy to synthesize 8a
were unsuccessful, as the methyl ester of 8a was unstable and
rapidly aldol-cyclized under the basic conditions.[10] Instead, 8a
suhdaey@uregina.ca
Supporting information for this article is available on the WWW under
http://dx.doi.org/10.1002/cbic.201200346.
ChemBioChem 0000, 00, 1 – 6
ꢀ 2012 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
&
1
&
ÞÞ
These are not the final page numbers!

Scheme 1. Different cyclization reactions catalyzed by type III polyketide synthases.
was prepared by treatment of tert-butyl 3,5,7-trioxoeicosanoate
with TFA (Supporting Information). The formation of 8a was
confirmed by the absence of the tBu signal at 1.46 ppm in
1H NMR spectrum. The tBu ester derivative of 8a proved to be
more stable and did not cyclize as rapidly as the methyl ester
counterpart. This is likely due to steric hindrance of the attack
of C-2 on C-7 provided by the bulky tBu group in the intramo-
lecular aldol reaction step. Compound 8a was found to be un-
stable in solution at 48C, but was stable for at least four weeks
when stored as a solid at ꢀ208C. It was reasonably assumed
that 8a is cyclized to 6-tridecyl-b-resorcylic acid (16a; see
below) or decarboxylated to the triketone, as these reactions
are known to occur spontaneously.[16] Fast Blue salts are com-
monly used for the visualization of phenolic lipids including
alkylresorcinols.[17] We found that 8a, but not the diketo com-
pounds, was stained with Fast Blue B salt (0.1% (w/v) in water)
to give a reddish brown color that faded to beige/orange over
time. This offered a convenient way to detect 8a by TLC (Fig-
ure S1 in the Supporting Information).
9.2 min exhibited chromatographic and spectroscopic parame-
ters characteristic of an alkylresorcinol: UV (methanol): lmax
=
276, 282 nm;[18] stained violet with Fast Blue B salt (Figure S1);
1H NMR (CDCl3): d=2.49 (t, 2H), 6.17 (t, 1H), 6.24 ppm (d, 2H).
Thus, the compound with tR =9.2 min was identified as 5-tri-
decylresorcinol (C13-RL, 7a). The compound that eluted at tR =
10.3 min exhibited diagnostic parameters of an alkyl-b-resorcyl-
ic acid: UV (methanol): lmax =216, 258, 298 nm;[19] stained
1
violet with Fast Blue B salt (Figure S1); H NMR: d=2.90 (t, 2H),
6.25 (d, 1H), 6.28 ppm (d, 1H). Therefore, this compound was
determined to be 6-tridecyl-b-resorcylic acid (C13-RA, 16a). An
uncharacterized component was also detected at tR =11.3 min
(marked with an asterisk in Figures 1 and S2); this component
had a UV lmax of 266 nm, and was not stained with Fast Blue B
salt, unlike other polyketide-derived compounds.
When incubated in pH 7.8 buffer, 8a could no longer be de-
tected after 4 h (Figure 1A). Instead, 16a was formed, and was
detected typically within 10 min of incubation. Decarboxylation
of 16a to 7a was extremely slow under these conditions, and
only after 48 h incubation a small amount of 7a was detected.
The sequential conversion of 8a to 16a to 7a was evident
when the peak areas of the three compounds were plotted
against incubation time (Figure S3), in agreement with the
known solution chemistry at neutral pH.[10,16]
First, the stability of 8a in the HPLC mobile phase (acetoni-
trile/H2O/acetic acid 8:2:0.01, v/v/v) and in the enzyme reaction
buffer (0.1m potassium phosphate (KPi), pH 7.8) was investigat-
ed. Freshly prepared 8a remained relatively stable after incu-
bation in the mobile phase for 6 h, thus validating HPLC as the
analytical method of choice (Figure S2). After prolonged incu-
bation (22 h), the conversion of 8a to two other compounds
was almost complete. The compound that eluted at tR =
Next, the possible intermediacy of 8a in ArsB-catalyzed re-
sorcinol ring formation was examined by incubating 8a in the
presence of ArsB (200 mg). ArsB failed to convert 8a to 7a (Fig-
&2
&
www.chembiochem.org
ꢀ 2012 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
ChemBioChem 0000, 00, 1 – 6
ÝÝ
These are not the final page numbers!

ure 1B). The amounts of 7a
formed after 48 h of incubation
in the absence or presence of
ArsB were similar, thus indicating
that 7a was formed by nonenzy-
matic decarboxylation of 16a,
even in the presence of ArsB. In-
terestingly, the initial rate of the
aldol cyclization and aromatiza-
tion of 8a to 16a was enhanced
by ArsB (Figure 1). This rate en-
hancement was also observed
with ArsC, a type III PKS that
does not produce alkylresorci-
nols but produces alkylpyrones
(Scheme 1).[3] During the initial
10 min of incubation, both ArsB
and ArsC facilitated the forma-
tion of 16a from 8a by 13-fold
(Figure 2A, left). Furthermore,
both enzymes appeared to stabi-
lize 16a and hindered the decar-
boxylation of 16a to 7a; less 7a
was formed in the presence of
either enzyme after prolonged
incubation (12 h; Figure 2A,
right). The identical effects of
ArsB and ArsC on the conversion
of 8a to 16a to 7a strongly sug-
gested that these effects are not
due to the catalytic activity of
ArsB. It is likely that ArsB and
ArsC provided a favorable envi-
ronment, presumably at the acyl
binding site, that was conducive
to an entropically favored aldol
cyclization of 8a (cage effect).
The aldol cyclization occurs
spontaneously in the presence
of ArsB or ArsC, but, as the re-
sults indicate, relatively faster
Scheme 2. Possible mechanisms of resorcinol ring formation. The pathways that are eliminated by the results ob-
tained in this study are cross-checked and indicated in italics, the remaining candidate pathways are indicated in
bold.
than in solution. Observations that both denatured ArsB (Fig-
ure 2B) and BSA (data not shown) failed to facilitate the forma-
tion of 16a from 8a provided support for the notion of a cage
effect. Clearly, ArsB did not catalyze either the aldol cyclization
and aromatization of 8a to 16a or the decarboxylation of 16a
to 7a.
It could be thought that ArsB and ArsC, but not BSA, contain
other site(s) that bind 8a and exert a cage effect. Such non-
active-site binding could have limited the availability of 8a to
the active site, keeping it from being catalytically converted to
7a. On the other hand, if 8a is not an intermediate but the ob-
served conversion to 16a occurs at the active site, 8a would
exhibit an inhibitory effect on the ArsB reaction in a similar
manner to other structurally related compounds. Compound
8a inhibited the ArsB reaction by 44% at 0.35 mm. It was
a stronger inhibitor than stearic acid (32% inhibition at 1 mm)
Figure 1. Time-course HPLC analysis of the reaction mixture of C20-TKA (8a)
in the A) absence and B) presence of ArsB. In all chromatograms, 6-tridecyl-
b-resorcylic acid (C13-RA, 16a) is dark-shaded, and C20-TKA (8a) is light-
shaded. 5-Tridecylresorcinol (C13-RL, 7a) is indicated with an arrow, the un-
identified contaminant is indicated with an asterisk.
ChemBioChem 0000, 00, 1 – 6
ꢀ 2012 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chembiochem.org
&
3
&
ÞÞ
These are not the final page numbers!

Figure 2. HPLC analysis of the reaction mixture of C20-TKA (8a) and A) ArsB and ArsC, or B) denatured ArsB. In all chromatograms, 6-tridecyl-b-resorcylic acid
(C13-RA, 16a) is dark-shaded, and C20-TKA (8a) is light-shaded. 5-Tridecylresorcinol (C13-RL, 7a) is indicated with arrow, the unidentified contaminant is indicat-
ed with asterisk. i.s.=internal standard (olivetol), which was added to the reaction mixture prior to ethyl acetate extraction to control the extraction efficien-
cy.
(Figure 3A), but a weaker inhibitor than iodoacetamide (95%
inhibition at 0.05 mm). More importantly, although the diketo
analogue, 3,5-dioxooctadecanoic acid (C18-DKA) showed a com-
parable inhibition (40% at 0.5 mm), its methyl ester, methyl
3,5-dioxooctadecanoate, was not inhibitory at 1 mm. Because
ArsB activity was sensitive to organic solvents, it was not possi-
ble to measure the effects of the inhibitors at higher concen-
trations. Instead, the inhibitory effect of a water-soluble ana-
logue, the dipotassium salt of C18-DKA (C18-DKAS), was investi-
gated in more detail. C18-DKAS showed a concentration-depen-
dent inhibition, and its Ki value was determined to be 0.65 mm
when fitted to a competitive inhibition model (r2 =0.96; Fig-
ure 3B). The observed inhibition by 8a was mostly due to 8a
itself, as 16a, which should have been formed during the
inhibition assay, was a weaker inhibitor (40% inhibition at
1 mm). On the other hand, C18-DKA and C18-DKAS were solely
responsible for the observed inhibitory effect, as they were re-
covered unchanged after incubation (data not shown). Compa-
rable inhibition of the ArsB reaction by 8a and its structural
analogues, C18-DKA, C18-DKAS, and stearic acid, but not by
methyl 3,5-dioxooctadecanoate indicates specific binding of
8a and the analogues to ArsB. Although it cannot be com-
pletely excluded, the probability that 8a inhibited ArsB activity
by binding to a non-active site is very low. Rather, 8a and the
analogues most likely inhibited ArsB activity by competitively
binding to the enzyme active site. The relatively weak affinity
of 8a and 16a for ArsB, as evidenced by their weak inhibition
and the high Ki value of C18-DKAS, also agrees with the notion
that neither 8a nor 16a is an intermediate in the ArsB reac-
tion.
Based on the results that ArsB and ArsC, but not denatured
ArsB and BSA, facilitated the conversion of 8a to 16a and that
ArsB and ArsC hindered the decarboxylation of 16a to 7a, we
conclude that neither 8a nor 16a is an intermediate in ArsB-
catalyzed resorcinol ring formation, and that aldol cyclization
occurs prior to thioester hydrolysis. This eliminates any path-
way that involves 8a or 16a as an intermediate; that is, path-
ways A1, A2, B1.3, and B2 (Scheme 2). The remaining plausible
pathways are B1.1 and B1.2. Obtaining direct evidence for
either “aldol first” pathway might not be easy. One could con-
sider feeding a synthetic sample of the cyclized dianion 13 to
ArsB to see if ArsB directly converts it to alkylresorcinol. How-
ever, one should be careful in interpreting the results. The con-
figuration of the two chiral centers (C-2 and C-7) of synthet-
ic 13 can differ from that of enzyme-produced 13, and this can
lead to a false-negative conclusion.[4,10,20]
Figure 3. Inhibition of ArsB activity by C20-TKA and its analogues. A) ArsB-cat-
alyzed formation of 5-heptadecylresorcinol from stearoyl-CoA (50 mm) and
[2-14C]malonyl-CoA (10 mm) was measured in the presence of stearic acid
(1 mm) or C20-TKA (0.35 mm); meanꢁSEM (n=4); *P<0.05, Student’s t-test.
A representative radio thin-layer chromatogram of the products generated
by ArsB is shown right. B) Michaelis–Menten plots of the ArsB-catalyzed for-
mation of 5-tridecylresorcinol at different concentrations of C18-DKAS. Myris-
toyl-CoA was used as the substrate. ArsB produces tri- and tetraketide py-
rones as well as alkylresorcinol with medium-chain acyl-CoA starter sub-
strates, as shown in the representative radio-thin layer chromatogram.[3]
An enigma in the mechanism of aldol-cyclizing type III PKSs
such as ArsB and STS is how the enzyme controls the timing of
&4
&
www.chembiochem.org
ꢀ 2012 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
ChemBioChem 0000, 00, 1 – 6
ÝÝ
These are not the final page numbers!

oxoeicosanoate and 3,5,7-trioxoeicosanoic acid (8a, C20-TKA) are
described in the Supporting Information.
the hydrolysis of 3. Premature hydrolysis of the growing ketide
intermediates will result in mono-, di-, and triketo acids that
might be dead-end products. In other words, the aldol switch
needs to stay “off” until the final condensation reaction (for
the “hydrolysis-first” pathway) or until aldol cyclization (for the
“aldol-first” pathway). In the proposed “aldol-first” ArsB reac-
tion pathway, the formation of the ring structure could be the
switch trigger.
Enzyme purification and assay: Recombinant ArsB and ArsC were
produced as His10-tagged proteins in E. coli BL21(DE3) cells and
purified by Ni2+-affinity chromatography as described previously.[21]
Enzyme reaction was carried out either in KPi (0.1m, pH 7.8) or in
KPi buffer (0.1m, pH 7.2) containing 10% glycerol and 0.1% Triton
X-100 (assay buffer). Reaction mixture (100 mL) containing enzyme
(20–30 mg), [2-14C]malonyl-CoA (50 mm, 13.5 mCimmolꢀ1), and myr-
istoyl-CoA or stearoyl-CoA as the starter substrate (100 mm) was in-
cubated at 378C for 30 min. Reaction products were separated by
silica TLC, and enzyme activity was quantified by using a phosphor-
imaging system as described.[15]
STS might or might not share the same ring-formation
mechanism with ArsB. Similar experiments with 8b could be
devised to study STS-catalyzed resorcinol ring formation. Our
attempts on this line of work did not result in conclusive evi-
dence for the intermediacy of 8b in the STS reaction mainly
because of the chemical instability of 8b. However, STS can
produce short-chain alkylresorcinols by using, for example,
hexanoyl-CoA as the starter substrate.[21] Therefore, suitable 8b
analogues can be synthesized to test the “hydrolysis-first” hy-
pothesis for STS-catalyzed resorcinol ring formation. Austin
et al.[5] found a thiolase-like, hydrogen-bonding-activated water
molecule next to the catalytic Cys in the crystal structure of
STS and proposed that this “aldol switch” was responsible for
the hydrolysis of the tetraketide thioester intermediate (3). It is
worth noting that, using LC-MS, Tosin et al.[22] detected a dehy-
drated, possibly cyclized, tetraketide CoA ester species from
the reaction of STS and cinnamoyl-CoA. If this compound is
indeed cyclized, it would indicate an alternative “aldol-first”
pathway for STS in which the aldol cyclization of 3 is followed
sequentially by dehydration, hydrolysis, b-keto acid decarboxy-
lation and aromatization.
Enzymatic conversion of C20-TKA, and HPLC analysis of reaction
products: Freshly prepared C20-TKA (8a; 0.75 mm) was incubated
in KPi buffer (0.1m, pH 7.8) at 378C in the presence of ArsB, ArsC,
denatured ArsB (908C, 4 min) or BSA (each 100–200 mg). Aliquots
were removed at time intervals and extracted with ethyl acetate.
Extracts were dried under vacuum and redissolved in HPLC mobile
phase (acetonitrile/H2O/acetic acid 8:2:0.01, v/v/v). Reaction prod-
ucts were analyzed by HPLC by using a Waters instrument
equipped with a Phenomenex Luna 5m C5 100A column (250ꢁ
4.6 mm), a Waters 510 HPLC pump, and a Waters 717 plus auto-
sampler. Detection was made on an Agilent 1100 Series system at
280 nm. The isocratic elution program was run at 0.9 mLminꢀ1
with the mobile phase.
Inhibition of ArsB by C20-TKA and related compounds: The inhib-
itory effects of different compounds on the activity of ArsB were
tested in the enzyme assay buffer by using [2-14C]malonyl-CoA
(10 mm, 54 mCimmolꢀ1) and myristoyl-CoA or stearoyl-CoA as the
starter substrate (50 mm). Triton X-100 was added to increase the
solubility of inhibitors. The enzyme reaction profile was not affect-
ed by the addition of the detergent. After 30 min of reaction, the
remaining enzyme activity was measured as described above. In-
hibitors were dissolved in ethanol, except iodoacetamide and C18-
DKAS, which were dissolved in the assay buffer. The final concen-
tration of ethanol in the reaction mixture was 1%. For Ki determi-
nation, the production of 5-tridecylresorcinol was measured after
20 min of reaction of ArsB (20 mg) with myristoyl-CoA (10~120 mm)
and [2-14C]malonyl-CoA (10 mm) in the absence and presence of
C18-DKAS (0.4, 0.7, and 1.0 mm). The initial velocity data were fitted
to a Michaelis–Menten inhibition model provided in GraphPad
Prism v.5 (La Jolla, USA).
As a single active site is responsible for substrate selection,
chain elongation, and cyclization in type III PKS reactions, it is
not always possible to study cyclization reactions separately
without affecting other reactions through, for example, site-di-
rected mutagenesis. The short chain length of polyketide inter-
mediates of type III PKS systems as compared to type I and II
PKSs allows the intermediates to be prepared synthetically for
mechanistic studies. ArsB, ArsC and other related fatty acyl-
CoA-accepting type III PKSs such as sorghum alkylresorcinol
synthase[23] and rice ARAS[9] present good model systems for
studying the mechanisms of different polyketide cyclization re-
actions. It has been decades since the solution chemistry of
polyketo acids was delineated and the possibility of these
compounds being intermediates in polyketide biosynthesis
was discussed (reviewed in ref. [13]). This study demonstrates
(to the best of our knowledge for the first time) the feasibility
of using a triketo acid derivative to investigate the cyclization
mechanism of polyketide synthase.
Acknowledgements
This research was supported by the Natural Sciences and Engi-
neering Research Council, Canada Discovery Grants (A.G.H.W and
D.-Y.S), and the University of Regina. We thank Dr. Nobutaka
Funa (University of Tokyo) for expression plasmids of ArsB and
ArsC.
Experimental Section
Keywords: aldol reaction · biosynthesis · cyclization · enzyme
catalysis · polyketide synthases
Materials and syntheses: Expression plasmids, pET16b-ArsB and
pET16b-ArsC were provided by Dr. Nobutaka Funa (University of
Tokyo).[3] [2-14C]Malonyl-CoA (54 mCimmolꢀ1) was from PerkinElmer,
and acyl-CoA esters and other chemicals were from Sigma–Aldrich.
3,5-Dioxooctadecanoic acid (C18-DKA), its dipotassium salt (C18-
DKAS), and methyl 3,5-dioxooctadecanoate were prepared as de-
scribed.[15] The syntheses and characterization of tert-butyl 3,5,7-tri-
[1] J. Staunton, K. J. Weissman, Nat. Prod. Rep. 2001, 18, 380–416.
[2] I. Abe, H. Morita, Nat. Prod. Rep. 2010, 27, 809–838.
[3] N. Funa, H. Ozawa, A. Hirata, S. Horinouchi, Proc. Natl. Acad. Sci. USA
2006, 103, 6356–6361.
ChemBioChem 0000, 00, 1 – 6
ꢀ 2012 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chembiochem.org
&
5
&
ÞÞ
These are not the final page numbers!

[4] J.-L. Ferrer, J. M. Jez, M. E. Bowman, R. A. Dixon, J. P. Noel, Nat. Struct.
Biol. 1999, 6, 775–784.
[5] M. B. Austin, M. E. Bowman, J.-L. Ferrer, J. Schroder, J. P. Noel, Chem.
Biol. 2004, 11, 1179–1194.
[6] H. Morita, Y. Shimokawa, M. Tanio, R. Kato, H. Noguchi, S. Sugio, T.
Kohno, I. Abe, Proc. Natl. Acad. Sci. USA 2010, 107, 669–673.
[7] I. Abe, S. Oguro, Y. Utsumi, Y. Sano, H. Noguchi, J. Am. Chem. Soc. 2005,
127, 12709–12716.
[15] C. C. Colpitts, S. S. Kim, S. E. Posehn, C. Jepson, S. Y. Kim, G. Wiedemann,
R. Reski, A. G. H. Wee, C. J. Douglas, D.-Y. Suh, New Phytol. 2011, 192,
855–868.
[16] T. M. Harris, R. L. Carney, J. Am. Chem. Soc. 1967, 89, 6734–6741.
[17] A. Kozubek, J. P. H. Tyman, Chem. Phys. Lipids 1995, 78, 29–35.
[18] A. B. Ross, P. ꢂman, R. Andersson, A. Kamal-Eldin, J. Chromatogr. A
2004, 1054, 157–164.
[19] G. W. Van Eijk, Antonie van Leeuwenhoek 1969, 35, 497–504.
[20] T. M. Harris, T. T. Howarth, R. L. Carney, J. Am. Chem. Soc. 1971, 93,
2511–2515.
[21] Y. Yamazaki, D.-Y. Suh, W. Sitthithaworn, K. Ishiguro, Y. Kobayashi, M. Shi-
buya, Y. Ebizuka, U. Sankawa, Planta 2001, 214, 75–84.
[22] M. Tosin, D. Spiteller, J. B. Spencer, ChemBioChem 2009, 10, 1714–1723.
[23] D. Cook, A. M. Rimando, T. E. Clemente, J. Schrçder, F. E. Dayan, N. P. D.
Nanayakkara, Z. Pan, B. P. Noonan, M. Fishbein, I. Abe, S. O. Duke, S. R.
Baerson, Plant Cell 2010, 22, 867–887.
[8] a) H. Zhou, Y. Li, Y. Tang, Nat. Prod. Rep. 2010, 27, 839; b) T. Dairi, T. Ku-
zuyama, M. Nishiyama, I. Fujii, Nat. Prod. Rep. 2011, 28, 1054–1086.
[9] M. Matsuzawa, Y. Katsuyama, N. Funa, S. Horinouchi, Phytochemistry
2010, 71, 1059–1067.
[10] T. T. Howarth, T. M. Harris, J. Am. Chem. Soc. 1971, 93, 2506–2510.
[11] M. Shibuya, M. Nishioka, U. Sankawa, Y. Ebizuka, Tetrahedron Lett. 2002,
43, 5071–5074.
[12] T.-L. Li, D. Spiteller, J. B. Spencer, ChemBioChem 2009, 10, 896–901.
[13] T. M. Harris, C. M. Harris, Pure Appl. Chem. 1986, 58, 283–294.
[14] D. Schmidt, J. Conrad, I. Klaiber, U. Beifuss, Chem. Commun. 2006,
4732–4734.
Received: May 23, 2012
Published online on && &&, 0000
&6
&
www.chembiochem.org
ꢀ 2012 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
ChemBioChem 0000, 00, 1 – 6
ÝÝ
These are not the final page numbers!

COMMUNICATIONS
Who’s first? Aldol cyclization occurs
before hydrolysis in the resorcinol ring
formation catalyzed by the type III poly-
ketide synthase, ArsB. Synthetic C20-TKA
was not converted to alkylresorcinol by
ArsB, but rather inhibited the enzyme
activity, thus indicating that C20-TKA is
not an intermediate in ArsB-catalyzed
alkylresorcinol formation.
S. E. Posehn, S. Y. Kim, A. G. H. Wee,*
D.-Y. Suh*
&& – &&
Mapping the Mechanism of the
Resorcinol Ring Formation Catalyzed
by ArsB, a Type III Polyketide Synthase
from Azotobacter vinelandii
ChemBioChem 0000, 00, 1 – 6
ꢀ 2012 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chembiochem.org
&
7
&
ÞÞ
These are not the final page numbers!