Specific enrichment of nonribosomal peptide synthetase module by an affinity probe for adenylation domains

Article abstract of DOI:10.1016/j.bmcl.2013.12.082

We targeted the development of an affinity probe for adenylation (A) domains that can facilitate enrichment, identification, and quantification of A domain-containing modules in nonribosomal peptide synthetase (NRPS)-polyketide synthase (PKS) hybrids and NRPSs. A 5′-O-sulfamoyladenosine (AMS) non-hydrolyzable analogue of adenosine monophosphate (AMP) has been reported as a scaffold for the design of inhibitors exhibiting tight binding of adenylation enzymes. Here we describe the application of an affinity probe for A domains. Our synthetic probe, a biotinylated l-Phe-AMS (l-Phe-AMS-biotin) specifically targets the A domains in NRPS modules that activates l-Phe to an aminoacyladenylate intermediate in both recombinant NRPS enzyme systems and whole proteomes.

Full text of DOI:10.1016/j.bmcl.2013.12.082

Bioorganic & Medicinal Chemistry Letters 24 (2014) 865–869
Contents lists available at ScienceDirect
Bioorganic & Medicinal Chemistry Letters
journal homepage: www.elsevier.com/locate/bmcl
Specific enrichment of nonribosomal peptide synthetase module by
an affinity probe for adenylation domains
⇑
⇑
Fumihiro Ishikawa , Hideaki Kakeya
Department of System Chemotherapy and Molecular Sciences, Division of Bioinformatics and Chemical Genomics, Graduate School of Pharmaceutical Sciences, Kyoto University,
Sakyo-ku, Kyoto 606-8501, Japan
a r t i c l e i n f o
a b s t r a c t
Article history:
We targeted the development of an affinity probe for adenylation (A) domains that can facilitate
enrichment, identification, and quantification of A domain-containing modules in nonribosomal peptide
synthetase (NRPS)-polyketide synthase (PKS) hybrids and NRPSs. A 50-O-sulfamoyladenosine (AMS) non-
hydrolyzable analogue of adenosine monophosphate (AMP) has been reported as a scaffold for the design
of inhibitors exhibiting tight binding of adenylation enzymes. Here we describe the application of an
Received 21 October 2013
Revised 15 December 2013
Accepted 19 December 2013
Available online 25 December 2013
affinity probe for A domains. Our synthetic probe, a biotinylated
cally targets the A domains in NRPS modules that activates -Phe to an aminoacyladenylate intermediate
in both recombinant NRPS enzyme systems and whole proteomes.
L-Phe-AMS (L-Phe-AMS-biotin) specifi-
Keywords:
L
Nonribosomal peptide synthetase
Polyketide synthase
Adenylation domain
Affinity probe
Ó 2014 Elsevier Ltd. All rights reserved.
Gramicidin S
A number of small molecule natural products with significant
biological activity such as antimicrobial, anticancer, and immuno-
suppressant activities belong to a large class of natural products
known as polyketide (PK)–nonribosomal peptide (NRP) hybrid
molecules and NRPs.1 Originating from bacteria and fungi, these
peptide-containing bioactive natural products consist not only of
the 20 proteinogenic amino acids, but also non-proteinogenic ami-
no acids, aryl acids, and other acids, thus generating highly com-
plex chemical diversity.2 The biosynthesis of these small
molecules is performed by highly versatile and large multimodular
enzymes called nonribosomal peptide synthetase (NRPS)–polyke-
tide synthase (PKS) hybrids and NRPSs.3–5 There has been much
progress in the understanding biochemical programming and
molecular basis of adenylation (A) domain substrate specificity in
NRPS and NRPS–PKS hybrid systems. This has facilitated the pre-
diction of structural features of newly discovered NRP and NRP–
PK natural products assembled by these systems uncovered by
genomic information. Large portions of these compounds often
readily correlate to the amino acid specificity of A domains found
on their associated modular enzymes.6–8 Genome sequencing has
revealed that secondary metabolite gene clusters encoding these
enzymes are widely dispersed and largely uncharacterized.9 In
addition, complicated organisms containing symbiotic bacteria
are particularly resistant to most genetic methods that rely on
pure, culturable strains. Direct detection of biosynthetic enzymes
from bacterial proteomes compliments genetic approaches in
understanding the activity and dynamics of these enzymes in their
native proteomes. By taking advantage of the A domains of NRPS
and PKS–NRPS hybrid systems, this work aims to specifically
enrich A domain-containing modules from bacterial proteomes
with sequenced genomes using small-molecule probes and directly
link the chemotypes of expressed peptide-containing natural prod-
ucts to their biosynthetic enzymes. Chemical probes for A domains
using amino, aryl, or other acid building blocks that are found in
NRP and NRP–PK hybrid molecules would facilitate probe-guided
selective enrichment and identification of A domain-containing
modules from proteomic samples by LC–MS/MS analysis (Fig. 1).
In sequenced producers, such approaches would have applications
in monitoring the expression dynamics of NRPS modules and opti-
mizing bacterial culture conditions. In unsequenced organisms,
such studies should facilitate discovery of the expressed NRPS/
PKS–NRPS gene clusters.10,11
The affinity probe design for the A domains is based on the
reaction mechanism for amino acid loading, which is catalyzed
by the A domains as depicted in Figure 2a. The A domain recog-
nizes a specific amino, acyl, or another acid from the cellular pool,
and catalyzes the formation of a tightly bound acyladenylate inter-
mediate. This in turn transfers the acyl group onto the thiol group
of a 40-phosphopantethein present on a downstream carrier pro-
tein (CP) of NRPS assembly line.12,13 A 50-O-sulfamoyladenosine
(AMS), a non-hydrolyzable analogue of adenosine monophosphate
(AMP), has been applied widely in the design of inhibitors that
⇑
Tel.: +81 75 753 4524; fax: +81 75 753 4591.
E-mail addresses: fishika@pharm.kyoto-u.ac.jp (F. Ishikawa), scseigyo-hisyo@
pharm.kyoto-u.ac.jp (H. Kakeya).
0960-894X/$ - see front matter Ó 2014 Elsevier Ltd. All rights reserved.
http://dx.doi.org/10.1016/j.bmcl.2013.12.082

866
F. Ishikawa, H. Kakeya / Bioorg. Med. Chem. Lett. 24 (2014) 865–869
Figure 1. Methods for proteomic analysis of NRPSs using an affinity probe. Modules comprise carrier protein (CP), adenylation (A) [AF: L-Phe specific; AP: L-Pro specific; AV: L-
Val specific], epimerase (E), and condensation (C) domains.
Figure 2. (a) Biosynthesis of the gramicidin S. (b) Structures of L-Phe-AMS 1 and L-Phe-AMS-biotin 2.
Table 1
display tight binding of adenylation enzymes and functionally
related aminoacyl tRNA synthetases.14–16 A 50-O-[N-(phenylalanyl)-
The apparent Ki values of 1 and 2 for A domain of the tridomain GrsA
Kai pp (nM)
sulfamoyl] adenosine (L-Phe-AMS, 1, Fig. 2b) has been developed
Inhibitors
by Marahiel and coworkers wherein the reactive acylphosphate
linkage has been replaced by a bioisosteric and non-hydrolyzable
acylsulfamate group.15 Recently, an acylsulfamate-based probe
installed a photoreactive group has been reported for profiling of
the stand-alone aryl-acid adenylation enzyme, MbtA.17 A more
complete set of probes would be crucial, however, for full module
identification. Here, we describe the design, synthesis, and charac-
1
54.9 1.2
339 26
2*
*
The assays were performed in the presence of 0.0025% lgepal CA-630.
specific enrichment, identification, and quantification of A
domain-containing modules.
terization of a biotinylated variant of
L-Phe-AMS, referred to as
To expand the acylsulfamate scaffold to identify A domains of
modular NRPSs and NRPS-PKS hybrids from proteomic samples,
L-Phe-AMS-biotin and highlight its utility as an agent for the

F. Ishikawa, H. Kakeya / Bioorg. Med. Chem. Lett. 24 (2014) 865–869
867
our studies began with the preparation of a suitably tagged A
domain affinity probe (Fig. 2b). We chose a tridomain GrsA as an
initial proof of concept target protein (Fig. 2a). The probe design
was based on the tight-binding bisubstrate inhibitor, -Phe-AMS
L
1. A long PEG spacer arm was selected to minimize steric hindrance
between modular enzymes and avidin molecules and permit isola-
tion of low-abundance proteins.18 To this end, an affinity probe,
biotinylated L-Phe-AMS (L-Phe-AMS-biotin, 2) was synthesized
from commercially available adenosine, 4-dibromobutane,
Boc-Phe-OSu, and NHS-PEG12-Biotin (Scheme S2). In addition,
L-Phe-AMS 1 was prepared according to the literature procedure
to provide an analogue of 2 for comparative activity studies
(Scheme S1).14–16,19,20
Our biochemical studies began with an examination of the spe-
cific activities of these probes against recombinant GrsA derived
from gramicidin S biosynthetic enzymes as a model NRPS module.
GrsA is a single module peptide synthetase containing the domain
structure A (
-Phe during gramicidin S biosynthesis (Fig. 2a).21,22 For compara-
tive studies, we chose the NRPS module, TycB1. Derived from the
tyrocidine biosynthetic pathway, it incorporates -Pro and contains
the domain structure C-A (
-Pro)-CP.6 Both recombinant holo-GrsA
L-Phe)-CP-E and is responsible for the incorporation of
D
L
L
and holo-TycB1 were prepared as a C-terminal His6-tagged fusion
proteins according to the literature procedure (Fig. S1).23 Ki values
were determined by a coupled hydroxamate-MesG continuous
spectrophotometric assay (Fig. S2).24 Probes 1 and 2 displayed
tight-binding inhibition against GrsA-catalyzed formation of the
aminoacyladenylate, with calculated Kai pp values of 54.9 1.2 and
339 26 nM, respectively (Table 1 and Fig. S3). Conversely, the
TycB1-catalyzed the reaction showed no detectable inhibition by
1 (Fig. S4). These experiments here suggest that attachment of
Figure 3. In vitro characterization of
Pull-down of GrsA and competition with excess 1. GrsA (7.9 nM) was preincubated
in either the absence or presence of 10 M of 1 and treated with 1 M of 2. (b)
M of
M of
L-Phe-AMS-biotin 2 with purified proteins. (a)
l
l
Concentration dependence of 2. GrsA (7.9 nM) was incubated with 1 nM–10
l
2. (c) Competitive profiling with GrsA. GrsA (7.9 nM) was incubated with 4.25
l
2 in the presence of 0.1 nM–10
(5–0.01 g/mL) was incubated with 4.25
(1 g/mL), TycB1 (1 g/mL), and BSA (1
l
M of 1. (d) Limit of detection of purified GrsA. GrsA
M of 2. (e) Binding specificity of 2. GrsA
g/mL) were treated with 1 M of 2 in
l
l
l
l
l
l
either the absence or presence of 1. Full gels (Fig. S9) and experimental procedures
are provided in the SI.
Figure 4. Pull-down assays from E. coli cellular lysates. (a) Pull-down of overexpressed GrsA. E. coli lysates (1 mg/mL–31.3
staining. Lower: Anti-His blot. (b) Competition with excess 1. E. coli lysates (250 g/mL) with overexpressed GrsA were incubated with 4.25
presence of 100 M 1. (c) Concentration dependent competition by 1. E. coli lysates (250 g/mL) with overexpressed GrsA were preincubated with 1 (0–100
addition of 2 (4.25 M). (d, e) Binding specificity of probe 2. The lysates (250 g/mL) prepared in the presence or absence of IPTG were incubated with 4.25 M of 2 in either
the absence or presence of 100
M 1. E. coli was transformed with (d) pQE60-grsA and (e) pQE60-tycB1 plasmids. (⁄) depicts a non-specifically bound protein. Full gels
(Fig. S10) and experimental procedures are provided in the SI.
l
g/mL) were treated with 4.25
M of 2 in either the absence or
M) before
lM of 2. Upper: CBB
l
l
l
l
l
l
l
l
l

868
F. Ishikawa, H. Kakeya / Bioorg. Med. Chem. Lett. 24 (2014) 865–869
the pegylated long spacer at the 20-oxygen atom of 2 preserves the
high affinity binding of the -Phe-activating A domain of GrsA. In
addition, no-cross-reactivity was observed between the NRPS
modules. This indicates that an -Phe-AMS warhead is capable of
selectively targeting -Phe-activating A domains.
the concentrations of 1 to calculate an IC50 value of 11.0 4.2 nM
for 1 (Figs. 3c and S6). Quantitative analyses with a fixed amount
of probe 2 and streptavidin-agarose resin indicated that the limit
of detection was estimated above 100 ng of GrsA (Fig. 3d). Finally,
to ensure specific pull-down by probe 2, GrsA, TycB1, and BSA were
L
L
L
To evaluate the ability of probe 2 to pull-down recombinant
GrsA, GrsA was first incubated with probe 2 for 12 h at 4 °C at
pH 8 in the absence of Igepal CA-630 as a detergent. This was
accompanied by a weak GrsA band on the SDS–PAGE gel indicating
precipitated GrsA (Fig. S5). In contrast, pull-down assays in the
presence of the detergent affected probe function by preventing
aggregation effects. As shown in Figure 3a, a well-defined band
indicating precipitated protein was observed. When excess 1 was
incubated with GrsA for 1 h prior to addition of probe 2, no precip-
incubated with 4.25 lM 2. This resulted in preferential precipita-
tion of GrsA with virtually no detectable precipitation of TycB1
and BSA (Fig. 3e). This result, in addition to knowledge of the dif-
ferential amino acid specificity of A domains, strongly suggests
selective targeting of the A domain is guided by the amino acid
moiety attached to the probe molecule.
We next asked whether probe 2 could be used to probe endog-
enous A domains in proteome. We began with an analysis of the
conditions of probe 2 incubated with overexpressed holo-GrsA
and holo-TycB1 in whole Escherichia coli cellular lysates. E. coli cells
were grown either in the presence or absence of IPTG (Fig. S7). The
lysates containing overexpressed GrsA were incubated with
itation band was detected. L-Phe-AMS 1 competitively inhibited
the binding of probe 2. This indicates that the binding of probe 2
occurs at the active site of A domain of GrsA. To investigate the
concentration dependence of probe 2, varying concentrations of 2
4.25 lM 2 for 12 h at 4 °C at pH 8. Bound proteins were isolated
were incubated with 1
l
g of GrsA. As shown in Figure 3b, GrsA dis-
by purification with streptavidin-agarose resin. Non-specifically
bound proteins were removed by washing with ice-cold RIPA buf-
fer (50 mM Tris–HCl, pH 7.5, 150 mM NaCl, 1% NP-40, 1% sodium
deoxycholate, and 0.1% SDS). SDS–PAGE of the eluent revealed a
single band of GrsA at 120 kDa with CBB staining and Western
played an escalation in precipitation as probe concentration was
increased. Additionally, a dose-response curve was created by
varying concentrations of 1 with fixed concentrations of GrsA
and probe 2. The precipitated band intensities were plotted against
Figure 5. Pull-down assay of endogenous GrsA from Aneurinibacillus migulanus ATCC 9999 cellular lysate. (a) A. migulanus lysate (1.2 mg/mL) was treated with 9.5
Bound proteins were eluted by treatment with 1 (0, 0.5, 1, and 5 mM). (b) The lysate (1.2 mg/mL) was incubated with 9.5 M 2 in the presence of 100 M 1. Bound proteins
were eluted as described above. (c) Dynamics of expression of GrsA. The lysates at each cell concentration (OD660 = 2.5–29.5) were incubated with 9.5 M 2. Precipitated
lM 2.
l
l
l
proteins were eluted by treatment with 0.5 mM 1. (d) LC–MS/MS analysis identified the isolated band (blue box) as GrsA. Amino acids of the identified peptides are colored
either in red or blue. Full gels (Fig. S11) and experimental procedures are provided in the SI.

F. Ishikawa, H. Kakeya / Bioorg. Med. Chem. Lett. 24 (2014) 865–869
869
blotting using an anti 6 ꢀ His antibody (Fig. 4a). Pull-down of GrsA
was abrogated by treatment of the lysates with vehicle (DMSO)
(Fig. 4a). Furthermore, preincubation of the lysates with excess 1
resulted in an absence of GrsA precipitation (Fig. 4b). In contrast,
synthesis of this probe is straightforward, offering easy access to
synthetic derivatives containing different amino acids or other
acids for targeting A domains of interest. In addition, we have
determined optimal conditions for the specific pull-down of A do-
main-containing modules and their detection using a competitive
elution method and established the activity of this probe both in
purified proteins and within proteomic samples. Combinations of
this probe with other domain-specific reactive probes such as thi-
oesterase11,25 and dehydratase26 should further improve the
enrichment and detection of biosynthetic enzymes in proteomic
samples.
the observed ꢁ35 kDa band (⁄) is nonspecific, as
L-Phe-AMS 1 does
not have any effect on the intensity (Fig. 4b). Quantitative compet-
itive assays showed reduced GrsA band intensities with increasing
concentrations of 1 (Fig. 4c). Finally, to ensure specific pull-down
of GrsA by probe 2 in the proteomic environment, incubation of
the lysates with overexpressed GrsA or TycB1 with 4.25 lM 2 were
carried out. This resulted in preferential pull-down of GrsA with no
detectable pull-down of TycB1 (Fig. 4d and e). In the absence of
IPTG, GrsA is not expressed, as evidenced by the lack of precipi-
tated GrsA. Remarkably, the specific pull-down by probe 2 can be
completely inhibited by preincubation with excess 1. This is indic-
ative of the ability of probe 2 to discriminate between different
active-site structures of adenylation enzymes in crude cellular
lysates. This finding corresponds identically to protein precipitated
from recombinant GrsA, TycB1, and BSA by probe 2 (Fig. 3e). Taken
together, these experiments indicate that 2 is a true affinity probe
that shows no-cross-reactivity with other enzymes. Additionally,
the extremely high specificity of 2 shows that the sulfamoyl scaf-
fold coupled with a pegylated long spacer unit is an optimal design
for applications involving specific enrichment, identification, and
quantification of endogenous NRPS modules from crude cellular
lysates.
As an ultimate test of proteomic activity, we evaluated the
availability of probe 2 in a proteomic context by applying it to a na-
tive system. To isolate the cellular target of probe 2, the lysate from
the gramicidin S producer, Aneurinibacillus migulanus ATCC 9999
was incubated with probe 2, and the precipitated proteins were
isolated by purification with a streptavidin-immobilized resin. Pro-
teins were then eluted by treatment with 20 mM Tris–HCl buffer
(pH 8) containing 0, 0.5, 1, and 5 mM of 1 (Fig. 5a). A band (⁄) at
approximately 120 kDa appeared after eluting the resin with med-
ia containing 1 as a vehicle to release bound proteins (Fig. 5a).
More significantly, this band, affinity-purified by probe 2 can be
Acknowledgments
This work was supported by a Grant-in-Aid for Research Activ-
ity start-up 24810015 (F.I.) and a research grant from the Japan
Society of the Promotion of Science (H.K.). We thank Prof. Moham-
ed Marahiel (Philipps-Universität Marburg, Germany) for provid-
ing the GrsA and TycB1 expression constructs.
Supplementary data
Supplementary data (supporting figures; procedures for the
syntheses of 1 and 2; complete gel images; full experimental de-
tails; and copies of 1H and 13C NMR spectra) associated with this
article can be found, in the online version, at http://dx.doi.org/
10.1016/j.bmcl.2013.12.082.
References and notes
1. Walsh, C. T. Science 2004, 303, 1805.
2. Hur, G. H.; Vickery, C. R.; Burkart, M. D. Nat. Prod. Rep. 2012, 10, 1074.
3. Fischbach, M. A.; Walsh, C. T. Chem. Rev. 2006, 106, 3468.
4. Smith, S.; Tsai, S. C. Nat. Prod. Rep. 2007, 24, 1041.
5. Boettger, D.; Hertweck, C. ChemBioChem 2013, 14, 28.
6. Mootz, H. D.; Marahiel, M. A. J. Bacteriol. 1997, 179, 6843.
7. van Wageningen, A. M. A.; Kirkpatrick, P. N.; Williams, D. H.; Harris, B. R.;
Kershaw, J. K.; Lennard, N. J.; Jones, M.; Jones, S. J. M.; Solenberg, P. J. Chem. Biol.
1998, 5, 155.
8. Schümann, J.; Hertweck, C. J. Am. Chem. Soc. 2007, 129, 9564.
9. Keller, N. P.; Turner, G.; Bennett, J. W. Nat. Rev. Microbiol. 2005, 3, 937.
10. Bumpus, S. B.; Evans, B. S.; Thomas, P. M.; Ntai, I.; Kelleher, N. L. Nat. Biotechnol.
2009, 10, 951.
11. Meier, J. L.; Niessen, S.; Hoover, H. S.; Foley, T. L.; Cravatt, B. F.; Burkart, M. D.
ACS Chem. Biol. 2009, 4, 948.
12. Kleinkauf, H.; Gevers, W.; Lipmann, F. Proc. Natl. Acad. Sci. U.S.A. 1969, 62, 226.
13. Lee, S. G.; Lipmann, F. Proc. Natl. Acad. Sci. U.S.A. 1977, 74, 2343.
14. Heacock, D.; Forsyth, C.; Shiba, K.; Musier-Forsyth, K. Bioorg. Chem. 1996, 24,
273.
15. Finking, R.; Neumüller, A.; Solsbacher, J.; Konz, D.; Kretzschmar, G.; Schweitzer,
M.; Krumm, T.; Marahiel, M. A. ChemBioChem 2003, 4, 903.
16. Qiao, C.; Gupte, A.; Boshoff, H. I.; Wilson, D. J.; Bennett, E. M.; Somu, R. V.;
Barry, C. E., III; Aldrich, C. C. J. Med. Chem. 2007, 50, 6080.
17. Duckworth, B. P.; Wilson, D. J.; Nelson, K. M.; Boshoff, H. I.; Barry, C. E., III;
Aldrich, C. C. ACS Chem. Biol. 2012, 7, 1653.
18. Sato, S.; Kwon, Y.; Kamisuki, S.; Srivastava, N.; Mao, Q.; Kawazoe, Y.; Uesugi, M.
J. Am. Chem. Soc. 2007, 129, 873.
19. Cisar, J. S.; Ferreras, J. A.; Soni, R. K.; Quadri, L. E. N.; Tan, D. S. J. Am. Chem. Soc.
2007, 129, 7752.
20. Ferreras, J. A.; Stirrett, K. L.; Lu, X.; Ryu, J.-S.; Soll, C. E.; Tan, D. S. Chem. Biol.
2008, 15, 51.
21. Kratzschmar, J.; Krause, M.; Marahiel, M. A. J. Bacteriol. 1989, 171, 5422.
22. Hoyer, K. M.; Mahlert, C.; Marahiel, M. A. Chem. Biol. 2007, 14, 13.
23. Stachelhaus, T.; Hüser, A.; Marahiel, M. A. Chem. Biol. 1996, 3, 913.
24. Wilson, D. J.; Aldrich, C. C. Anal. Biochem. 2010, 404, 56.
25. Meier, J. L.; Mercer, A. C.; Burkart, M. D. J. Am. Chem. Soc. 2008, 130, 5443.
26. Ishikawa, F.; Haushalter, R. W.; Burkart, M. D. J. Am. Chem. Soc. 2012, 134, 769.
completely inhibited by preincubation with 100 lM 1 (Fig. 5b). A
sample of the excised band was submitted for LC–MS/MS protein
identification and found to contain the endogenous GrsA
(Figs. 5d and S8). We further evaluated probe 2 as a quantification
tool to gain an understanding of modular synthase expression at
each time point (Fig. 5c). Each of the lysates was incubated with
probe 2 for 12 h at 4 °C. Precipitated proteins were then eluted
by treatment with 500 lM 1. The evidence provided in Figure 5c
and d indicates that endogenous GrsA production abruptly began
within 9 h (OD660 = 6.0) after inoculation. Interestingly, at 24 h
the amount of GrsA in a proteomic context was on the same level
as 12 h, as estimated by the relative band intensity. This indicated
the slow disappearance of GrsA in the producer. Probe 2 allowed
successful monitoring of the expression dynamics and function of
GrsA. Accordingly, these results validate the use of probe 2 as a
selective isolation, identification, and quantification tool for A
domain-containing modules within a proteomic environment.
In summary, we have demonstrated proteomic applications for
the study of adenylation enzymes in NRPS modules using an affin-
ity probe. This probe is capable of selectively targeting A domains
with substrate specificity that corresponds to an attached amino
acid moiety. It is noteworthy that this probe can attain specific
enrichment of the target even from a native cellular lysate. The