Insights into the programmed ketoreduction of partially reducing polyketide synthases: Stereo- and substrate-specificity of the ketoreductase domain

Article abstract of DOI:10.1039/c4ob01777c

One of the hallmarks of iterative polyketide synthases (PKSs) is the programming mechanism which is essential for the generation of structurally diverse polyketide products. In partially reducing iterative PKSs (PR-PKSs), the programming mechanism is mainly dictated by the ketoreductase (KR) domain. The KR domain contributes to the programming of PR-PKSs through selective reduction of polyketide intermediates. How the KR domain achieves the selective ketoreduction remains to be fully understood. In this study, we found that the KR domain of the (R)-mellein-synthesizing PR-PKS SACE5532 functions as a B-type KR domain to generate (R)-hydroxyl functionalities. Comparative studies of the KR domains of SACE5532 and NcsB suggested that the two KR domains have distinct substrate preferences towards simple N-acetylcysteamine thioester (SNAC) substrates. We further found that the substrate preference of KR_SACE5532 can be switched by swapping several motifs with KR_NcsB, and that swapping of the same motifs in the full length SACE5532 resulted in a reprogramming of the PKS. Together, the results advance our understanding of the programming of iterative PR-PKSs by providing new support to the hypothesis that the programmed ketoreduction is accomplished by differential recognition of polyketide intermediates. This journal is

Full text of DOI:10.1039/c4ob01777c

Organic &
Biomolecular Chemistry
View Article Online
View Journal | View Issue
PAPER
Insights into the programmed ketoreduction of
partially reducing polyketide synthases: stereo-
and substrate-specificity of the ketoreductase
domain†
Cite this: Org. Biomol. Chem., 2014,
2, 8542
1
a
a
b
a
a
Ishin Soehano,‡ Lifeng Yang,‡ Feiqing Ding,‡ Huihua Sun, Zhen Jie Low,
b
a
Xuewei Liu and Zhao-Xun Liang*
One of the hallmarks of iterative polyketide synthases (PKSs) is the programming mechanism which is
essential for the generation of structurally diverse polyketide products. In partially reducing iterative PKSs
(PR-PKSs), the programming mechanism is mainly dictated by the ketoreductase (KR) domain. The KR
domain contributes to the programming of PR-PKSs through selective reduction of polyketide intermedi-
ates. How the KR domain achieves the selective ketoreduction remains to be fully understood. In this
study, we found that the KR domain of the (R)-mellein-synthesizing PR-PKS SACE5532 functions as a
B-type KR domain to generate (R)-hydroxyl functionalities. Comparative studies of the KR domains of
SACE5532 and NcsB suggested that the two KR domains have distinct substrate preferences towards
simple N-acetylcysteamine thioester (SNAC) substrates. We further found that the substrate preference of
KRSACE5532 can be switched by swapping several motifs with KRNcsB, and that swapping of the same motifs
in the full length SACE5532 resulted in a reprogramming of the PKS. Together, the results advance our
understanding of the programming of iterative PR-PKSs by providing new support to the hypothesis that
the programmed ketoreduction is accomplished by differential recognition of polyketide intermediates.
Received 20th August 2014,
Accepted 3rd September 2014
DOI: 10.1039/c4ob01777c
www.rsc.org/obc
1
2
bacterium Saccharopolyspora erythraea. Despite the fact that
SACE5532 shares the same domain composition and a head-
Introduction
Iterative PKSs are multifunctional enzymes that assemble to-tail sequence homology with NcsB, AziB and 6-MSAS,
structurally diverse products using a single module of catalytic SACE5532 synthesizes (R)-mellein, an aromatized pentaketide
domains. The use of a single module of catalytic domains product that is likely to be generated by a distinct program-
makes the iterative PKSs mechanistically distinct from the ming mechanism. SACE5532 was postulated to selectively
1
–3
modular PKSs. Iterative PKSs are categorized into non-redu- reduce the β-keto group of the diketide and tetraketide inter-
cing (NR), highly reducing (HR) and partially reducing (PR)- mediates to generate hydroxyl groups at C2 and C6 positions
PKSs according to the composition of catalytic domains and (Fig. 1). The ketoreduction pattern is different from those of
2
,4
oxidation levels of the polyketide intermediates. The best 6-MSAS and NcsB, with 6-MSAS only reducing the triketide
known PR-PKSs are the bacterium or fungus-derived 6-MSAS, intermediate and NcsB reducing the triketide and pentaketide
1
2
NcsB and AziB, which synthesize 6-methylsalicylic acid intermediates.
5
–8
9,10
(6-MSA),
2-hydroxyl-5-methyl-naphthoic acid (hmNPA)
The structural diversity of PR-PKS products arises from the
and 5-methyl-naphthoic acid (mNPA) respectively. In a recent programming of the C–C bond forming and processing cataly-
effort to explore the diversity of PR-PKSs, we discovered a tic domains, mainly through controlling carbon-chain length,
1
1
(
R)-mellein-synthesizing PR-PKS (SACE5532) from the soil ketoreduction and cyclization pattern. The carbon-chain
length of the products is most likely to be determined by the
substrate-binding pocket of the ketosynthase (KS) domain, as
1
3–15
a
observed in other PKSs,
likely to be governed by the positions of the hydroxyl and keto
functionalities without using product template (PT)
whereas the cyclization pattern is
School of Biological Sciences Nanyang Technological University, 60 Nanyang Drive,
37551, Singapore. E-mail: zxliang@ntu.edu.sg; Fax: (+)65 6791 3856
6
b
School of Mathematics and Physics, Nanyang Technological University, 21 Nanyang
a
Link, 637371, Singapore
16
domain. While the NADPH-dependent ketoreduction is cata-
lysed by the KR domain, how the ketoreduction pattern is
determined remains to be fully understood.
†
Electronic supplementary information (ESI) available. See DOI: 10.1039/
c4ob01777c
These authors contributed equally to this work.
7
,17–19
‡
Despite the
8542 | Org. Biomol. Chem., 2014, 12, 8542–8549
This journal is © The Royal Society of Chemistry 2014

View Article Online
Organic & Biomolecular Chemistry
Paper
Fig. 1 Postulated selective ketoreduction in SACE5532 and NcsB.
SACE5532 reduces the keto group of the diketide and tetraketide inter-
mediates, whereas NcsB reduces the keto group of the triketide and
pentaketide intermediates. In addition to the essential ketosynthase (KS),
acyltransferase (AT) and acyl carrier protein (ACP) domains, the PR-PKSs
Fig. 2 Stereo-specific ketoreduction of the KRSACE5532 domain. (a)
Partial sequence comparison of the KR domains of PR-PKSs and rep-
resentative type-A and B KR domains from modular PKSs. The motifs
containing the signature “LDD” motifs for type-B KR domains are shown
in the frame (Azi26_KR of AziB_KR: KR domain of AziB; PokM1_KR: KR
domain of PokM1; MdpB_KR: KR domain of MdpB; ChlB1_KR: KR
domain of ChlB1; Ery_KR1: KR domain of the first module of erythromy-
cin PKS; Pik_KR1: KR domain of the first module of pikromycin PKS;
Tyl_KR1: KR domain of the first module of tylosin PKS; Ery_KR2: KR
domain of the second module of erythromycin PKS; Pik_KR5: KR
20
also contain a dehydratase-like thiohydrolase (TH) domain
ketoreductase (KR) domain.
and a
fact that PR-PKSs are some of the first PKSs to be discovered
7
,17–19
and characterized,
the function of individual catalytic
domains of PR-PKSs is difficult to dissect and much remains
to be understood. Based on our recent observation that the domain of the fifth module of pikromycin PKS). (b) Chiral HPLC analysis
ketoreductase (KR) domain of SACE5532, but not the KR of the product from the reduction of acetoacetyl-SNAC by KRSACE5532
.
1
2
domain of NcsB, is able to reduce diketide substrates, we
proposed that the ketoreduction patterns of SACE5532 and
NcsB are solely governed by the KR domains through differen-
tial substrate recognition. In this study, the stereo- and sub-
strate-specificity of the KRSACE5532 domain were further probed
to yield fresh evidence supporting the hypothesis that differen-
tial substrate recognition is central for achieving the pro-
grammed ketoreduction.
hydroxyl groups. To establish the stereo-preference of the
KRSACE5532 domain, we cloned the stand-alone KRSACE5532
domain (1170–1638 aa of SACE5532) according to newly
assigned domain boundaries and synthesized the two N-acetyl-
cysteamine thioester (SNAC) enantiomers 1d-(R) and 1d-(S). As
revealed by chiral HPLC analysis, incubation of the diketide
substrate acetoacetyl-SNAC (1c) with KRSACE552 and NADPH
produced 1d-(R) (Fig. 2b), suggesting that KRSACE5532 is a
B-type KR domain. Generation of the (R)-configured diketide
product is fully consistent with the (R)-configured chiral centre
12
Results and discussion
A large body of knowledge about the stereo-specificity of the in (R)-mellein. Moreover, a survey of the sequences of more
KR domains of fatty acid synthases, type II PKS and type I than 100 putative PR-PKSs revealed that although the signature
2
1–38
modular PKSs has been accumulated.
Enzymatically active motif “LDD” of B-type KR domains is not strictly conserved in
KR domains are classified into A-type and B-type according to most PR-PKSs, a similar motif (V/L/Y)(N/S/D/E)(D/N/S) can still
the stereochemistry of the β-hydroxyl group generated by be found at the same position (Fig. 2a). Considering that the
2
6,31–35,39–41
ketoreduction.
KR domains have been identified. However, because the signa- phan residue found in A-type KR domains
Signature motifs of the A- or B-type KR domains of PR-PKSs also lack a highly conserved trypto-
2
6,32,39
we propose
ture motifs found in the KR domains of modular PKSs are that 6-MSAS-like PR-PKSs contain B-type KR domains that
usually absent in the KR domains of most PR-PKSs (Fig. 2a), it generate (R)-configured hydroxyl groups.
is not immediately evident from sequence analysis whether
Having established the stereochemistry of the KR domains,
the KR domains of PR-PKSs produce (R)- or (S)-configured we next tested whether KRSACE5532 selectively reduces the
This journal is © The Royal Society of Chemistry 2014
Org. Biomol. Chem., 2014, 12, 8542–8549 | 8543

View Article Online
Paper
Organic & Biomolecular Chemistry
diketide (1a) and tetraketide (3a) intermediates and KRNcsB KRNcsB (2.8 × 10 M s−1) favours a mechanism whereby cycli-
2
−1
prefers the triketide (2b) and pentaketide (4b) intermediates zation of the polyketide intermediate occurs only after the
(Fig. 3a). To probe the substrate preference of the KR domains, ketoreduction. The only substrate that did not show the antici-
we started by synthesizing a series of tri-, tetra- and pentake- pated preference is the tetraketide 3c, with KRNcsB showing
tide-SNAC substrates. Our effort to prepare the structurally moderately higher activity. Enzymatic assays of the KR activity
faithful acyl-SNAC mimics of 2b, 3a and4b was not successful with full-length SACE5532 and NcsB showed similar results,
because the products were highly prone to spontaneous cycli- with the two KR domains exhibiting strong preference for 1c,
zation. To circumvent the problem, we synthesized the simple 2c and 4c (data not shown). Despite the drawback associated
tri-, tetra- and pentaketide-SNAC substrates 2c, 3c and 4c with the use of the simple SNAC substrates, the ability to
instead (Fig. 3b). The SNAC-substrates 2c, 3c and 4c share the differentiate the SNAC substrates indicates that the carbon-
same carbon chain length with the postulated PKS intermedi- chain length of the polyketide intermediate is likely an impor-
ates 2b, 3a and 4b, but lack the hydroxyl or keto groups tant structural factor in substrate recognition. For 3c, we
beyond the β-position. In addition to KRSACE5532, KRNcsB speculate that the lack of differentiation is due to the absence
(1162–1641 aa of NcsB) was cloned and shown to be enzymati- of γ- and δ-hydroxyl/keto groups or the tethering phosphopan-
cally active using the non-specific substrate trans-1-deca- tetheinyl. Although it can be argued that the ACP domain may
1
2,22
lone.
By in vitro enzymatic assays, we found that play a role in substrate recognition and that the observed sub-
KRSACE5532 and KRNcsB exhibited clear preferences towards strate preferences may not reflect the reduction of ACP-teth-
three of the four SNAC substrates. KRSACE5532 reduced the dike- ered substrates, the ACP domain does not affect substrate
of 2.1 M−
1
s , in contrast to specificity because the ACPSACE5532 and ACPNcsB domains are
−1
tide substrate 1c with a kcat/K
M
the negligible activity for KRNcsB. In contrast, KRNcsB exhibited interchangeable, which we will discuss later on. Hence,
higher activity than KRSACE5532 towards the triketide 2c and the despite the limitation of the simple SNAC substrates used in
pentaketide 4c (Fig. 3b). Note that the facile reduction of 4c by the study, the observations suggest that the KR domains have
distinct substrate preferences.
The ketoreduction catalysed by KR domains is believed to
4
0,42
be assisted by a triad of Tyr, Ser and His residues.
How the
ACP-tethered substrates are bound and recognized by the KR
domains is less clear. To identify the residue and regions
important for substrate recognition, we constructed a homo-
logy model for KR
using a type-B KR structure (PDB ID:
SACE5532
4
0
2FR0 ) as the template. The structural model features a cleft
for accommodating NADPH and likely acyl substrates (Fig. 4a).
Judging from their locations, we identified all potential resi-
dues that are likely to be involved in direct binding of the acyl
substrates. The identified 51 residues are contained in four
fragments or motifs. Motif I consists of a short loop preceding
the essential catalytic residue Y1536. The region occupied by
motif I has been shown to be important for controlling the
stereochemistry of the KR domain of the PKS in hypothemycin
4
3
biosynthesis. Motif II consists of a potentially flexible C-
terminal motif. Motifs I and II form the walls of a narrow cleft
in which the nicotinamide moiety of NADPH and the β-keto
group of the acyl substrate are likely to be bound. Motif III
encompasses a short helix and a loop situated on the top of
motif I, whereas motif IV is located next to motif II and is
probably involved in the binding of ACP and the phosphopan-
tetheinyl moiety. We asked the question of whether the sub-
strate preference of KRSAC3352 can be altered by replacing some
of the residues from the four motifs with the corresponding
residues from KRNcsB. The corresponding residues in KRNcsB
were identified by sequence alignment and secondary struc-
Fig. 3 Differential recognition of polyketide substrates by KRSACE5532 ture analysis (Fig. S1†). Substrates 1c and 4c were used as
and KRNcsB domains. (a) Putative polyketide intermediates for SACE5532
and NcsB. Note that only the intermediates that bear a β-keto group are
shown. The shaded keto groups are hypothesized to be reduced by the
KR domain of SACE5532 or NcsB. (b) Comparison of the efficiency of
probes to report the switching of substrate preference. A
switch of substrate preference would be indicated by a loss of
activity towards 1c and a concomitant gain of activity towards
4c. The KRSACE5532 variants harbouring different number of
the two KR domains in reducing the four acyl-SNAC substrates. The
assay conditions are described in ESI.†
residue replacements were constructed using site-directed
8544 | Org. Biomol. Chem., 2014, 12, 8542–8549
This journal is © The Royal Society of Chemistry 2014

View Article Online
Organic & Biomolecular Chemistry
Paper
activity towards 4c, with KRI2B showing a notable 16% activity
relative to KRNcsB. Comparable activity (∼17%) was observed
when the entire motif II was swapped (KRII). The chimera
(
KRI4/II) with motifs I and II swapped exhibited further
increased enzymatic activity (24%) for 4c, whereas the replace-
ment of the residues in motif III alone or in combination with
other changes did not seem to improve the activity. The last
chimera (KRI/II/III/IV) with all four motifs swapped (a total of 51
replacements) is a reasonably efficient ketoreductase domain
that exhibited 56% activity towards 4c (Fig. 4b). A second-order
2
−1 −1
rate constant of 1.6 × 10 M
S
was obtained for KRI/II/III/IV,
2
−1 −1
compared to the rate constant of 2.8 × 10 M
S
for KRNcsB.
In conjunction with the observations that none of the single,
double or triple mutants exhibited significant activity towards
4c, the results indicate that the substrate specificity is collec-
tively determined by a large number of residues in the vicinity
of the active site. Some residues, such as the ones from motif
IV, may play indirect yet important roles by positioning sub-
strate and cofactor-binding residues. These observations seem
to suggest that evolution of new substrate specificity for the KR
domains by protein engineering may require the replacement
of more than a few residues. On an encouraging note, the
remarkable tolerance of KRSACE5532 to residue replacement
indicates that the substrate-binding pocket is highly plastic
Fig. 4 Identification of the regions in KRSACE5532 that govern substrate and evolvable. Such evolvability is likely attributed to the facts
specificity. (a) Coloured are the four motifs that are likely to govern sub- that the four motifs are composed of mainly surface-exposed
strate specificity. The α-carbon of the residues from the motifs I–IV that
are different between KRSACE5532 and KRNcsB are represented by the
is mainly governed by other parts of the protein. Such de-
balls. NADPH is shown as sticks. The active site is indicated by the aster-
segments and that the overall stability of the protein scaffold
coupling of stability and function has been observed in other
isk. (b) Enzymatic activities of the chimeric KR towards substrates 1c and
4
c in relation to KRSACE5532 and KRNcsB. The chimeras were derived from enzymes with high adaptability and is consistent with the per-
KRSACE5532. The residues that are replaced in the mutants are listed in petual need of nature to evolve KR domains with new substrate
Table S5.† The assay conditions are described in ESI.†
47,48
specificity.
Having shown that substrate specificity can be altered by
motif-swapping in the KRSACE5532 domain, we finally tested
mutagenesis and overlapping PCR methods as described in how the alteration of substrate specificity affects the ketoreduc-
4
6
the Experimental section.
tion pattern and product formation in the full-length PKS. We
We first focused on the residues from motif I because of prepared the full PKS construct (SACE5532(KRI/II/III/IV)) with
their proximity to the active site and reported roles in substrate the KR domain replaced by the KRI/II/III/IV domain. Ketoreduc-
4
3
recognition. Only four residues are different for motif I tion assays showed that the integrated KRI/II/III/IV domain is
between KRSACE5532 and KRNcsB. The two double mutants KRI2A enzymatically active with anticipated preference for substrates
(T1530L, T1531S) and KRI2B (V1534A, T1535S) exhibited sig- 2c and 4c (data not shown). Instead of producing (R)-mellein
nificantly lower enzymatic activity than KRSACE5532, whereas or hmNPA, SACE5532(KRI/II/III/IV
)
produced triacetic acid
the replacement of the four residues resulted in 96% loss of lactone (TAL) and 6-MSA as the major products (Fig. 5a). In
activity towards 1c (Fig. 4b). Although the replacement of indi- contrast to SACE5532, the production of TAL and 6-MSA indi-
vidual residues in motifs II and III did not have significant cates that the diketide intermediate is no longer reduced by
effects, their enzymatic activity was substantially decreased the KRI/II/III/IV domain. The production of 6-MSA suggests that
when the entire motif II or III was swapped. It is not surprising a portion of the triketide intermediate is now reduced, con-
that low enzymatic activity towards 1c was also observed for firming that the ketoreduction pattern has indeed been altered
the variants that contain two, three or all four swapped motifs. (Fig. 5b). The results also suggest that further optimization is
Differential scanning fluorometry (DSF) measurements needed for the KRI/II/III/IV domain to suppress the production
showed that the chimeras have similar unfolding tempera- of TAL that is derived from the unreduced triketide. However,
tures, indicating that the loss of enzymatic activity was not due we found that further improvement by swapping the residues
to a change of protein stability (data not shown). Importantly, situated beyond the four motifs is challenging because the
when the enzymatic activity of the variants towards 4c was swapping often led to a loss of protein stability. It will be
measured, we noticed that the replacement of the residues crucial in the future to have the high-resolution crystal struc-
V1534 and T1535 from motif I led to a gain of enzymatic ture of KRSACE5532 or KRNcsB to guide the optimization.
This journal is © The Royal Society of Chemistry 2014
Org. Biomol. Chem., 2014, 12, 8542–8549 | 8545

View Article Online
Paper
Organic & Biomolecular Chemistry
PKSs are interchangeable. Given the modest decrease in
mellein production, the ACP domain is likely to play a simple
role enhancing the efficiency of intermediate shuttling
4
9–54
through electrostatic-guided docking.
And lastly, we found
that the SACE5532(TH–KRNcsB) construct generated by swap-
ping the TH-KR didomain produced 6-MSA and a negligible
amount of TAL. The didomain-swapping result not only con-
firms that the KR domain controls the ketoreduction pattern
but also indicates that the TH-KR domain interface is impor-
tant for the function of PR-PKSs.
In summary, we have established that the KR domain of the
mellein-synthesizing SACE5532, and likely other PR-PKSs,
functions as a B-type KR domain to generate (R)-hydroxyl func-
tionalities. We showed that the KR domains of SACE5532 and
NcsB have distinct substrate preferences towards simple SNAC
substrates. We found that a KRSACE5532 variant that contains
motifs grafted from KRNcsB exhibited altered substrate speci-
ficity and can cause a reprogramming of the full length PKS.
The motif-swapping results also demonstrate the plasticity of
the KR active site and the potential of reprogramming the
PKSs by motif-grafting. In contrast to the domain-swapping
5
5–59
method
that often disrupts domain–domain interface and
causes a loss of function, the motif-swapping or grafting
method may hold some advantages for PKS engineering. The
observations lend further support to the hypothesis that the
KR domain of PR-PKSs achieves programmed ketoreduction by
differential recognition of polyketide intermediates. Further
work, such as the determination of the crystal structures of
enzyme–substrate complexes, will be required to fully establish
the molecular basis of the substrate preferences exhibited by
the KR domains.
Fig. 5 Domain swapping between SACE5532 and NcsB resulted in the
alteration of the ketoreduction pattern. (a) Product analysis by HPLC.
Experimental section
Reaction conditions are described in the Experimental section and Molecular cloning and DNA manipulation
44,45
described elsewhere.
observed products.
(b) Postulated mechanism for the formation of
The cloning of KRSACE5532 and KRNcsB domains was described
1
2
previously. The single, double and quadruple mutants were
generated by PCR using KAPAHiFi™ DNA polymerase with the
Meanwhile, the absence of the hexaketide product hmNPA primers listed in Table S2.† The templates used for the double
for SACE5532(KRI/II/III/IV) indicates that further extension of the and quadruple mutants were pET-KRSACE5532 and the syn-
carbon chain is hindered, most likely because the tetraketide thesized double mutant pET-KRSACE5532 KRI2A respectively.
1
5
intermediate is not recognized by the KSSACE5532 domain.
Similarly, mutations were introduced into other expression
The release of 6-MSA by the THSACE5532 domain also seems to plasmids such as pET-KRII to yield KRI2A/II. The PCR con-
suggest that the TH domain is able to tolerate some degree of ditions used for site-directed mutagenesis are: denaturing at
structural change in the product. It should be noted that we 95 °C for 5 minutes, 30 cycles at 98 °C for 50 seconds, gradient
did not observe the production of 6-MSA by SACE5532(KRNcsB), annealing temperatures (60–85 °C) for 50 seconds, 72 °C for
a chimera generated by replacing the KR domain with KRNcsB
.
5 minutes, and 10 minutes at 72 °C. To construct the
One plausible reason for the complete lack of 6-MSA is the KRSACE5532 variants that contain the motifs from KRNcsB (e.g.
alteration of the quaternary PKS structure and dynamics KRIII and KRI4/II/III), an overlapping PCR reaction was per-
caused by the disruption of TH-KR domain interface. We also formed. The three-step PCR reaction eventually generated DNA
found that the replacement of the ACP domain by ACPNcsB did fragments which were subsequently used for ligation into the
not affect the ketoreduction pattern and production of expression vector pET28(b)+. For the PCR reactions, specific
mellein, indicating that the ACP domain is not involved in primers with 15 bases of overlapping sequences were used
differential substrate recognition. The result is consistent with (Table S3†). Briefly, the first round of PCR used the forward
recent studies which showed that the ACP domains of iterative primer A containing the NdeI restriction site and the reverse
8546 | Org. Biomol. Chem., 2014, 12, 8542–8549
This journal is © The Royal Society of Chemistry 2014

View Article Online
Organic & Biomolecular Chemistry
Paper
primer B containing the replacement motif sequences. The ml− kanamycin. The culture was kept in a shaking incubator
1
second round of PCR generated the DNA fragment consisting overnight at 37 °C at 180 rpm. An 8 ml aliquot was transferred
−1
of the replacement motif with the remaining C-terminal to 800 ml of LB medium supplemented with 50 μg ml kana-
sequences. The forward primer that contains the replacement mycin. The culture was kept at 37 °C at 200 rpm till the OD600
motif sequences and the reverse primer that contains an XhoI reached ∼0.6 (∼3 hours) before it was cooled down to 16 °C
restriction site were utilized in the second round of PCR. The and induced with 0.2 mM IPTG. After incubation at 16 °C for
PCR conditions are: denaturing at 95 °C for 5 minutes, 30 an additional ∼20 hours at 160 rpm, the cells were harvested.
cycles at 98 °C for 50 seconds, gradient annealing tempera- The cell pellet was re-suspended in the lysis buffer (50 mM
tures (50–75 °C) for 50 seconds, and 72 °C for 1 minute, and NaH
minutes at 72 °C. The products of the first and second PCR (v/v) glycerol) and lysed by sonication for 10 minutes. After
were gel-purified separately. The third round of PCR employed centrifugation at 20 000 rpm for 30 minutes at 4 °C, the super-
0 ng each of PCR products as templates with the forward natant was filtered using a 0.45 μm membrane and loaded
2 4
PO (pH 8.0), 300 mM NaCl, 20 mM imidazole, and 10%
7
7
primer A and the reverse primer D (see Table S3†). The 1.5 kb onto a Ni-NTA column. The column was washed using the lysis
PCR product was gel-purified and digested with NdeI and XhoI buffer and wash buffer supplemented with 10, 50 and 200 mM
restriction enzymes. The digested DNA fragment was gel-puri- of imidazole before the protein was eluted out with the elution
fied and ligated into pET28b(+). The resulted plasmid was buffer containing 500 mM imidazole. Proteins were desalted
then transformed into the Top10 competent cells and later and exchanged into the Tris buffer (50 mM Tris-HCl (pH 8.0),
into Rosetta (DE3) cells for protein expression. The genes that 150 mM NaCl, and 10% (v/v) glycerol) using the PD10 desalt-
encode the KRSACE5532 variants KRII and KRI/II/III/IV were ing column. The protein was concentrated, flash frozen in
obtained from GenScript Corporation (NJ, USA) with the DNA liquid nitrogen, and stored in −80 °C freezer.
sequence optimized for protein expression in E. coli. The
genes were provided as pUC57-based plasmids with NdeI and described previously and the expression and purification of
XhoI restriction sites added to the N- and C-termini and sub- NcsB and chimeric PKS proteins were carried out using a
The expression and purification of SACE5532 were
1
2
cloned into pET28b(+).
similar procedure. By co-transformation, the PKSs were co-
To construct the full length SACE5532(KRI/II/III/IV), Gibson expressed with the 4′-phosphopantetheinyl transferase Sfp to
Assembly™ Cloning kit (New England BioLabs) was used. facilitate the phosphopantetheinylation of the ACP domain.
Using this technique, various DNA fragments and the vector
Synthesis of SNAC substrates
backbone were assembled using overlapping primers that
include 15–25 nucleotide overlapping sequences and 15–25 The synthesis of acetoacetyl-SNAC (1c) was described pre-
1
2
nucleotide gene-specific sequences (Table S4†). The PCR- viously. The synthetic method for 1d(S) and 1d(R) was
amplified DNA fragments and the linearized vector pET28(b)+ adopted from a previously reported procedure and the struc-
were incubated with the Gibson assembly master mix at 50 °C tures of the products were established by MS and NMR spec-
6
0
for 45 minutes. This Gibson-assembled mixture was sub- troscopy. For the synthesis of 2c, 3c and 4c, a 0.5 M solution
sequently transformed into NEB 5α competent cells according of Meldrum’s acid (1 eq.) in CH Cl was prepared in a round
to the manufacturer’s instructions. bottomed flask at room temperature. Pyridine (2 eq.) and
2
2
To construct the ACP domain swapping variant SACE5532 DMAP (20 mol%) were added and the reaction was stirred for
ACPNcsB), the plasmids pUC57-NcsB and PUC57-SACE5532 10 minutes at room temperature. The reaction was cooled to
(
were digested with ClaI and XhoI. The 498 bp ACPNcsB frag- 0 °C, 1 eq. of the desired acid chloride was added slowly, and
ment was ligated into the linearized PUC57-SACE5532 plasmid the reaction was warmed to room temperature and stirred over-
to yield PUC57-SACE5532(ACPNcsB). The 5.2 kb PKS gene was night. The reaction was washed with 0.1 M HCl and then with
digested and ligated into the expression vector pET28b(+) to water. The organic layers were dried over Na SO , filtered, and
2
4
generate pET-SACE5532(ACPNcsB). Similarly, to replace the the solvent was removed. After workup of the reaction, the
TH-KR di-domain of SACE5532 (900–1638 aa) with that of product was purified by flash column chromatography
NcsB (886–1641 aa), the plasmid pET-SACE5532(TH-KRNcsB
was generated. To replace the KR domain of SACE5532 2ca, 3ca and 4ca (ESI Scheme 1†). This material was then
1170–1638 aa) with that of NcsB (1162–1641 aa), the refluxed in toluene (0.5 M) with 1 eq. of N-acetylcysteamine for
)
(7 : 1 hexanes–ethyl acetate) to give the synthetic intermediates
(
KRSACE5532 DNA region was replaced by the corresponding three hours. The reaction was cooled and the solvent was
DNA sequence from NcsB gene to yield the plasmid removed in vacuo. The product was purified by flash column
pET-SACE5532(KRNcsB).
chromatography using 1 : 1 CH
2
Cl
2
: ethyl acetate to give 2c, 3c
and 4c.
Recombinant protein expression and purification
1
3-Oxohexanoyl-SNAC (2c). H NMR (CDCl
3
, 400 MHz): δ 6.07
For the stand-alone KR domains, the plasmids pET-KRSACE5532 (br, 1H), 3.67 (s, 2H), 3.41–3.46 (m, 2H), 3.05–3.09 (m, 2H),
and pET-KRNcsB were transformed into E. coli Rosetta (DE3) 2.50 (t, J = 7.2 Hz, 2H), 1.95 (s, 3H), 1.56–1.65 (m, 2H),
1
3
3
competent cells. The cells were plated on LB medium sup- 0.89–0.95 (m, 3H); C NMR (CDCl , 100 MHz): δ 202.2, 192.4,
−
1
plemented with 50 μg ml kanamycin. A single colony was 170.5, 99.3, 57.2, 39.2, 29.2, 23.1, 16.9, 13.5; HRMS (ESI) m/z
+
used to inoculate the LB medium supplemented with 50 μg [M + Na] calcd for C H NO SNa 254.0827, found 254.0831.
1
0
17
3
This journal is © The Royal Society of Chemistry 2014
Org. Biomol. Chem., 2014, 12, 8542–8549 | 8547

View Article Online
Paper
Organic & Biomolecular Chemistry
1
3
-Oxooctanoyl-SNAC (3c). H NMR (CDCl
3
, 400 MHz): δ 6.07 340 nm wavelength typically for two hours. At the end of the
(br, 1H), 3.68 (s, 2H), 3.42–3.46 (m, 2H), 3.05–3.09 (m, 2H), reaction, the reaction was stopped by quenching with 20 μl of
2
1
1
2
.51 (t, J = 7.2 Hz, 2H), 1.96 (s, 3H), 1.54–1.61 (m, 2H), 1.2 M HCl and the solution was vortexed to precipitate the
1
3
.25–1.31 (m, 4H), 0.87–0.89 (m, 3H);
C NMR (CDCl
3
,
enzyme. The reaction mixture was extracted twice with an
00 MHz): δ 202.4, 192.4, 170.5, 99.1, 57.2, 39.2, 31.1, 29.2, equal volume of ethyl acetate. The mixture was spun at
+
3.2, 23.1, 22.4, 13.9; HRMS (ESI) m/z [M + Na] calcd for 14 800 rpm for 15 min and the organic layer was separated and
C
12
H
21NO
3
SNa 282.1140, found 282.1143.
subjected to HPLC analysis. The conditions for HPLC analysis
1
44,45,61
3
-Oxodecanoyl-SNAC (4c). H NMR (CDCl , 400 MHz): δ 6.07 are similar to the ones described previously.
3
(br, 1H), 3.65 (s, 2H), 3.41–3.46 (m, 2H), 3.05–3.09 (m, 2H),
2
1
1
2
.51 (t, J = 7.2 Hz, 2H), 1.96 (s, 3H), 1.56–1.65 (m, 2H), Construction of a structural model for KRSACE5532
1
3
.20–1.30 (m, 8H), 0.89–0.95 (m, 3H);
00 MHz): δ 202.4, 192.4, 170.5, 99.1, 57.2, 39.2, 31.6, 31.5,
9.2, 29.0, 28.9, 23.4, 22.6, 14.0; HRMS (ESI) m/z [M + Na]
C NMR (CDCl3,
The structural model of KRSACE5532 was constructed using the
first KR domain of erythromycin synthase DEBS (PDB ID:
+
40,62
63,64
2FR0) as the template
and the software Rosetta Suite.
calcd for C H NO SNa 310.1453, found 310.1451.
1
4
25
3
Energy minimization was subsequently performed using the
Rosetta protein design.
Enzymatic assays
The assays of the ketoreductase activity were conducted using
a Shimadzu UV-Vis 1650 spectrophotometer following the pro-
Acknowledgements
4
4,45
cedures described previously.
The assays were performed
in 100 mM HEPES buffer (pH 8.5) in a total volume of 200 μl This work was supported by a Tier II grant from the Ministry
−
1
containing the following components: 1 mg ml ketoreduc- of Education (MOE) of Singapore to Z.X. Liang.
tase, 1 mM 1c and 2c or 0.1 mM 3c and 4c substrate. The reac-
tion mixtures were incubated at 25 °C and the reaction was
initiated by adding NADPH solution (0.25 mM). The progress
of the reaction was followed by monitoring at 341 nm wave-
Notes and references
length continuously. Initial velocity (V ) values were obtained
from the slopes of the progress curve and the extinction coeffi-
1 A. M. Hill, Nat. Prod. Rep., 2006, 23, 256–320.
2 B. Shen, Curr. Opin. Chem. Biol., 2003, 7, 285–295.
3 C. Khosla, D. Herschlag, D. E. Cane and C. T. Walsh, Bio-
chemistry, 2014, 18, 2875–2883.
4 R. J. Cox, Org. Biomol. Chem., 2007, 5, 2010–2026.
5 I. Schmitt, S. Kautz and H. T. Lumbsch, Mycol. Res., 2008,
112, 289–296.
6 L. Shao, X.-D. Qu, X.-Y. Jia, Q.-F. Zhao, Z.-H. Tian,
M. Wang, G.-L. Tang and W. Liu, Biochem. Biophys. Res.
Commun., 2006, 345, 133–139.
7 J. Beck, S. Ripka, A. Siegner, E. Schiltz and E. Schweizer,
Eur. J. Biochem., 1990, 192, 487–498.
0
cient of NADPH. The values of V
concentrations were used for the measurement of the second-
order rate constant kcat/K
0
at five different substrate
M
.
For the determination of the stereo-selectivity of the
KRSACE5532 domain, a total of 800 μl reaction mixture of 2 mM
substrate 1c, 0.3 mM NADPH and 3.5 mg KRSACE5532 in
100 mM HEPES buffer (pH 8.5) was incubated at 26 °C for
80 minutes. The reaction mixture was then supplemented with
30 μl of NADPH (10 mM) and incubated at 26 °C for
80 minutes. The reaction was quenched with 80 μl of 1.2 mM
HCl and the solution was vortexed to remove the enzyme. The
reaction mixture was extracted twice with equal volumes of ethyl
acetate. The mixture was spun at 14 800 rpm for 15 min and the
organic layer was separated and subjected to chiral HPLC analy-
sis. The two (R) and (S)-configured standards and the reaction
extract were dissolved in 100 µl isopropanol. The samples were
8 P. Lu, A. Zhang, L. M. Dennis, A. M. Dahl-Roshak,
Y. Q. Xia, B. Arison, Z. An and J. S. Tkacz, Mol. Gen. Geno-
mics, 2005, 273, 207–216.
9 W. Liu, K. Nonaka, L. P. Nie, J. Zhang, S. D. Christenson,
J. Bae, S. G. Van Lanen, E. Zazopoulos, C. M. Farnet,
C. F. Yang and B. Shen, Chem. Biol., 2005, 12, 293–302.
analysed using normal phase HPLC equipped with a Chiralpak 10 B. Sthapit, T. J. Oh, R. Lamichhane, K. Liou, H. C. Lee,
IA® column (150 × 4.60 mm) (Chiral Technologies). The HPLC C. G. Kim and J. K. Sohng, FEBS Lett., 2004, 566, 201–206.
conditions were: linear gradient; hexane to isopropanol ratio 11 Q. F. Zhao, Q. L. He, W. Ding, M. C. Tang, Q. J. Kang, Y. Yu,
−
1
(
v/v) of 90 to 10; 30 °C; 1 ml min flow rate (254 nm).
W. Deng, Q. Zhang, J. Fang, G. L. Tang and W. Liu, Chem.
Biol., 2008, 15, 693–705.
The enzymatic assays for the wild type PKSs were performed
following the procedure described previously.
4
4,61
To test the 12 H. Sun, C. L. Ho, F. Ding, I. Soehano, X.-W. Liu and
enzymatic activity of the chimeric PKSs, a reaction mixture
containing 3 μl of MgCl (1 M), 1 μl of acetyl-CoA (100 mM), 13 I. Abe, S. Oguro, Y. Utsumi, Y. Sano and H. Noguchi, J. Am.
00 μl of co-purified PKS/Sfp (15 mg ml ), 3 μl of malonyl-
CoA (100 mM), 5 μl of NADPH (10 mM) and 88 μl of reaction 14 Y. Tang, S.-C. Tsai and C. Khosla, J. Am. Chem. Soc., 2003,
buffer (50 mM Tris (pH 8.0), 150 mM NaCl and 10% (v/v) gly- 125, 12708–12709.
cerol) was incubated at room temperature. A UV-Vis spectro- 15 K. Watanabe, C. C. C. Wang, C. N. Boddy, D. E. Cane and
Z.-X. Liang, J. Am. Chem. Soc., 2012, 134, 11924–11927.
2
−
1
1
Chem. Soc., 2005, 127, 12709–12716.
photometer was used to follow the depletion of NADPH at
C. Khosla, J. Biol. Chem., 2003, 278, 42020–42026.
8548 | Org. Biomol. Chem., 2014, 12, 8542–8549
This journal is © The Royal Society of Chemistry 2014

View Article Online
Organic & Biomolecular Chemistry
Paper
1
6 J. M. Crawford, T. P. Korman, J. W. Labonte, A. L. Vagstad, 43 H. Zhou, Z. Z. Gao, K. J. Qiao, J. J. Wang, J. C. Vederas and
E. A. Hill, O. Kamari-Bidkorpeh, S.-C. Tsai and
C. A. Townsend, Nature, 2009, 461, 1139–1143.
7 J. B. Spencer and P. M. Jordan, Biochem. J., 1992, 288,
Y. Tang, Nat. Chem. Biol., 2012, 8, 331–333.
44 R. Kong, L. P. Goh, C. W. Liew, Q. S. Ho, E. Murugan, B. Li,
K. Tang and Z.-X. Liang, J. Am. Chem. Soc., 2008, 130,
8142–8143.
1
1
1
2
2
2
8
39–846.
8 C. J. Child, J. B. Spencer, P. Bhogal and P. M. 45 M. Kotada, R. Kong, I. Qureshi, Q. Ho, H. Sun, C. W. Liew,
ShoolinginJordan, Biochemistry, 1996, 35, 12267–12274.
9 Q. Zhang, B. Pang, W. Ding and W. Liu, ACS Catal., 2013, 3,
L. P. Goh, P. Cheung, Y. Mu, J. Lescar and Z.-X. Liang,
J. Biol. Chem., 2008, 284, 15739–15749.
1
439–1447.
46 R. Higuchi, B. Krummel and R. Saiki, Nucleic Acids Res.,
1988, 16, 7351–7367.
47 E. Dellus-Gur, A. Toth-Petroczy, M. Elias and D. S. Tawfik,
J. Mol. Biol., 2013, 425, 2609–2621.
48 V. Anantharaman, L. Aravind and E. V. Koonin, Curr. Opin.
Chem. Biol., 2003, 7, 12–20.
0 T. Moriguchi, Y. Kezuka, T. Nonaka, Y. Ebizuka and I. Fujii,
J. Biol. Chem., 2010, 285, 15637–15643.
1 H. Fu, S. Ebertkhosla, D. A. Hopwood and C. Khosla, J. Am.
Chem. Soc., 1994, 116, 4166–4170.
2 L. H. Østergaard, L. Kellenberger, J. Cortés, M. P. Roddis,
M. Deacon, J. Staunton and P. F. Leadlay, Biochemistry, 49 J. Bruegger, B. Haushalter, A. Vagstad, G. Shakya, N. Mih,
2
002, 41, 2719–2726.
C. A. Townsend, M. D. Burkart and S.-C. Tsai, Chem. Biol.,
2013, 20, 1135–1146.
2
3 A. T. Hadfield, C. Limpkin, W. Teartasin, T. J. Simpson,
J. Crosby and M. P. Crump, Structure, 2004, 12, 1865– 50 J. Crosby and M. P. Crump, Nat. Prod. Rep., 2012, 29,
875. 1111–1137.
4 S. Bali, H. M. O’Hare and K. J. Weissman, ChemBioChem, 51 J. Lim, R. Kong, E. Murugan, C. L. Ho, Z.-X. Liang and
006, 7, 478–484. D. Yang, PLoS One, 2011, 6, e20549.
5 R. Castonguay, W. G. He, A. Y. Chen, C. Khosla and 52 M. Leibundgut, S. Jenni, C. Frick and N. Ban, Science, 2007,
1
2
2
2
D. E. Cane, J. Am. Chem. Soc., 2007, 129, 13758–13769.
6 A. T. Keatinge-Clay, Chem. Biol., 2007, 14, 898–908.
7 T. Maier, M. Leibundgut and N. Ban, Science, 2008, 321,
316, 288–290.
2
2
53 Z. X. Liang, M. Jiang, Q. Ning and B. M. Hoffman, J. Biol.
Inorg. Chem., 2002, 7, 580–588.
1
315–1322.
54 Z. X. Liang, I. V. Kurnikov, J. M. Nocek, A. G. Mauk,
D. N. Beratan and B. M. Hoffman, J. Am. Chem. Soc., 2004,
126, 2785–2798.
2
2
3
3
3
3
3
3
3
3
3
8 C. R. Valenzano, R. J. Lawson, A. Y. Chen, C. Khosla and
D. E. Cane, J. Am. Chem. Soc., 2009, 131, 18501–18511.
9 D. H. Kwan, M. Tosin, N. Schlager, F. Schulz and 55 T. Liu, Y.-M. Chiang, A. D. Somoza, B. R. Oakley and
P. F. Leadlay, Org. Biomol. Chem., 2011, 9, 2053–2056. C. C. C. Wang, J. Am. Chem. Soc., 2011, 133, 13314–13316.
0 J. T. Zheng and A. T. Keatinge-Clay, J. Mol. Biol., 2011, 410, 56 H.-H. Yeh, S.-L. Chang, Y.-M. Chiang, K. S. Bruno,
1
05–117.
1 J. T. Zheng and A. T. Keatinge-Clay, MedChemComm, 2013,
, 34–40.
B. R. Oakley, T.-K. Wu and C. C. C. Wang, Org. Lett., 2013,
15, 756–759.
57 X. Zhu, F. Yu, X.-C. Li and L. Du, J. Am. Chem. Soc., 2006,
129, 36–37.
58 K. M. Fisch, W. Bakeer, A. A. Yakasai, Z. Song, J. Pedrick,
Z. Wasil, A. M. Bailey, C. M. Lazarus, T. J. Simpson and
R. J. Cox, J. Am. Chem. Soc., 2011, 133, 16635–16641.
4
2 J. T. Zheng, S. K. Piasecki and A. T. Keatinge-Clay, ACS
Chem. Biol., 2013, 8, 1964–1971.
3 J. T. Zheng, C. A. Taylor, S. K. Piasecki and A. T. Keatinge-
Clay, Structure, 2010, 18, 913–922.
4 P. Javidpour, A. Das, C. Khosla and S. C. Tsai, Biochemistry, 59 A. L. Vagstad, A. G. Newman, P. A. Storm, K. Belecki,
2
011, 50, 7426–7439.
J. M. Crawford and C. A. Townsend, Angew. Chem., Int. Ed.,
2013, 52, 1718–1721.
60 T.-Y. Lin, L. S. Borketey, G. Prasad, S. A. Waters and
N. A. Schnarr, ACS Synth. Biol., 2013, 2, 635–642.
5 P. Javidpour, T. P. Korman, G. Shakya and S. C. Tsai, Bio-
chemistry, 2011, 50, 4638–4649.
6 J. T. Zheng, D. C. Gay, B. Demeler, M. A. White and
A. T. Keatinge-Clay, Nat. Chem. Biol., 2012, 8, 615–621.
7 M. Häckh, M. Müller and S. Lüdeke, Chem. – Eur. J., 2013,
61 H. Sun, R. Kong, D. Zhu, W. Lu, Q. Ji, C. W. Liew, J. Lescar,
G. Zhong and Z.-X. Liang, Chem. Commun., 2009, 7399–7401.
62 T. Schwede, J. Kopp, N. Guex and M. C. Peitsch, Nucleic
Acids Res., 2003, 31, 3381–3385.
63 R. Das and D. Baker, Annu. Rev. Biochem., 2008, 77, 363–382.
64 A. Leaver-Fay, M. Tyka, S. M. Lewis, O. F. Lange, J. Thompson,
R. Jacak, K. Kaufman, P. D. Renfrew, C. A. Smith, W. Sheffler,
I. W. Davis, S. Cooper, A. Treuille, D. J. Mandell, F. Richter,
Y. E. A. Ban, S. J. Fleishman, J. E. Corn, D. E. Kim, S. Lyskov,
M. Berrondo, S. Mentzer, Z. Popovic, J. J. Havranek,
J. Karanicolas, R. Das, J. Meiler, T. Kortemme, J. J. Gray,
B. Kuhlman, D. Baker and P. Bradley, Methods Enzymol., 487:
Comp. Meth., Pt C, 2010, 545–574.
1
9, 8922–8928.
8 A. Garg, C. Khosla and D. E. Cane, J. Am. Chem. Soc., 2013,
35, 16324–16327.
1
3
4
9 A. T. Keatinge-Clay, Nat. Prod. Rep., 2012, 29, 1050–1073.
0 A. T. Keatinge-Clay and R. M. Stroud, Structure, 2006, 14,
7
37–748.
4
1 S. K. Piasecki, J. Zheng, A. J. Axelrod, M. E. Detelich and
A. T. Keatinge-Clay, Proteins: Struct., Funct., Bioinf., 2014,
8
2, 2067–2077.
2 S. Smith and S. C. Tsai, Nat. Prod. Rep., 2007, 24, 1041–
072.
4
1
This journal is © The Royal Society of Chemistry 2014
Org. Biomol. Chem., 2014, 12, 8542–8549 | 8549