Bacterial Alkaloid Biosynthesis: Structural Diversity via a Minimalistic Nonribosomal Peptide Synthetase

Article abstract of DOI:10.1016/j.chembiol.2018.02.013

Chemical and biochemical analyses of one of the most basic nonribosomal peptide synthetases (NRPS) from a Pseudomonas fluorescens strain revealed its striking plasticity. Determination of the potential substrate scope enabled us to anticipate novel secondary metabolites that could subsequently be isolated and tested for their bioactivities. Detailed analyses of the monomodular pyreudione synthetase showed that the biosynthesis of the bacterial pyreudione alkaloids does not require additional biosynthetic enzymes. Heterologous expression of a similar and functional, yet cryptic, NRPS of Pseudomonas entomophila was successful and allowed us to perform a phylogenetic analysis of their thioesterase domains. Nonribosomal peptides are a diverse class of ecologically and clinically relevant natural products. Here, Klapper et al. study one of the simplest bacterial nonribosomal peptide synthetases of different Pseudomonas strains. A single gene, encoding the monomodular pyreudione synthetase, leads to the production of a variety of bioactive alkaloids.

Full text of DOI:10.1016/j.chembiol.2018.02.013

Brief Communication
Bacterial Alkaloid Biosynthesis: Structural Diversity
via a Minimalistic Nonribosomal Peptide Synthetase
Graphical Abstract
Authors
Martin Klapper, Daniel Braga,
Gerald Lackner, Rosa Herbst,
Pierre Stallforth
H
O
N
OH
O
Correspondence
pierre.stallforth@leibniz-hki.de
O
N
OH
H
In Brief
C
A
T
TE
Nonribosomal peptides are a diverse
class of ecologically and clinically
relevant natural products. Here, Klapper
et al. study one of the simplest bacterial
nonribosomal peptide synthetases of
different Pseudomonas strains. A single
gene, encoding the monomodular
pyreudione synthetase, leads to the
production of a variety of bioactive
alkaloids.
O
HO
H
HS
O
OH
HO
N
NH
O
OH
O
NH OH
H
O
N
O
OH
Highlights
d
d
d
d
The pyreudione synthetase is a monomodular NRPS that
displays striking plasticity
Heterologous expression of the pyreudione synthetase was
achieved in various hosts
A functional pyreudione synthetase homolog was found in
Pseudomonas entomophila
We performed a phylogenetic analysis of different NRPS
thioesterase domains
Klapper et al., 2018, Cell Chemical Biology 25, 1–7
June 21, 2018 ª 2018 Elsevier Ltd.
https://doi.org/10.1016/j.chembiol.2018.02.013

Please cite this article in press as: Klapper et al., Bacterial Alkaloid Biosynthesis: Structural Diversity via a Minimalistic Nonribosomal Peptide Synthe-
tase, Cell Chemical Biology (2018), https://doi.org/10.1016/j.chembiol.2018.02.013
Cell Chemical Biology
Brief Communication
Bacterial Alkaloid Biosynthesis: Structural
Diversity via a Minimalistic
Nonribosomal Peptide Synthetase
Martin Klapper, Daniel Braga, Gerald Lackner, Rosa Herbst, and Pierre Stallforth1,3,*
1
2
2
1
1
Junior Research Group Chemistry of Microbial Communication, Leibniz Institute for Natural Product Research and Infection Biology, HKI,
Beutenbergstrasse 11a, 07745 Jena, Germany
2JuniorResearchGroupSyntheticMicrobiology,LeibnizInstituteforNaturalProductResearchandInfectionBiology,HKI,Beutenbergstrasse11a,
0
7745 Jena, Germany
3Lead Contact
Correspondence: pierre.stallforth@leibniz-hki.de
*
https://doi.org/10.1016/j.chembiol.2018.02.013
SUMMARY
The C domain then catalyzes the amide bond formation
between two amino acids or, in the case of a Cstarter domain
Chemical and biochemical analyses of one of the (Rausch et al., 2007), also between a fatty acid and an amino
most basic nonribosomal peptide synthetases acid. Plasticity of an NRPS can thus arise from the substrate
promiscuity or substrate multispecificity of the
A and
(
NRPS) from a Pseudomonas fluorescens strain
C domains. Eventually the peptide is released, typically via a
C-terminal reductase or a thioesterase (TE) domain (Du and
Lou, 2010). Different modes of product release have been
described, including macrolactonization, macrolactamization,
hydrolysis, and Dieckmann cyclization (DC).
revealed its striking plasticity. Determination of the
potential substrate scope enabled us to anticipate
novel secondary metabolites that could subse-
quently be isolated and tested for their bioactivities.
Detailed analyses of the monomodular pyreudione
synthetase showed that the biosynthesis of the
bacterial pyreudione alkaloids does not require addi-
Giant multimodular NRPSs have been shown to incorporate, for
instance, up to 15 amino acids in the case of kolossin A (Bode
et al., 2015). Accordingly, the respective 1.8 MDa enzyme harbors
tional biosynthetic enzymes. Heterologous expres- 15 individual modules, which are required to generate this large
sion of a similar and functional, yet cryptic, NRPS peptide. While many reports of huge bacterial multimodular
of Pseudomonas entomophila was successful and NRPSs exist (Hur et al., 2012; Marahiel, 2016), very little is known
about the most basic system: the stand-alone monomodular
NRPS (S u€ ssmuth and Mainz, 2017). This simple system can incor-
allowed us to perform a phylogenetic analysis of their
thioesterase domains.
porate and thus modify in principle a single amino acid. Recently,
we showed that one such monomodular NRPS, the pyreudione
INTRODUCTION
synthetase Pys from Pseudomonas fluorescens HKI0770,
enables the formation of the pyreudiones (e.g., pyreudione A, 1,
Many therapeutically used drugs contain nonribosomal peptides Figure 1A), a class of potent, bacterially produced amoebicidal
NRPs) or derivatives thereof as active ingredients (Sieber and alkaloids (Klapper et al., 2016); however, details of their biosyn-
(
Marahiel, 2005; S u€ ssmuth and Mainz, 2017; Walsh and thesis have remained elusive.
Fischbach, 2010). Daptomycin, for instance, is an antibacterial
nonribosomal cyclic lipopeptide in clinical use against methi- RESULTS AND DISCUSSION
cillin-resistant Staphylococcus aureus (Baltz, 2014; Baltz et al.,
2
005). Ciclosporin, on the other hand, is a calcineurin inhibitor, The NRPS Pys Is Sufficient for Pyreudione Biosynthesis
indispensable in the prevention of organ rejection in transplant Compared with the 1.8 MDa mega enzyme that assembles
patients (Liu et al., 1991). These peptides are biosynthesized kolossin A, the minuscule 142 kDa pyreudione synthetase is
by modular enzyme complexes, so-called nonribosomal peptide basically a combination of a starter and a termination module
synthetases (NRPSs), which link amino acids in an assembly line with a C-A-T-TE architecture (Figure 1A). Bioinformatics analysis
fashion. Each module can be subdivided into individual domains of the C domain using NaPDoS (Ziemert et al., 2012) clearly
that catalyze specific steps of the elongation process. Three confirmed that it belongs to the Cstarter family, typically involved
domains are particularly crucial: the adenylation (A), the thiola- in linkage of an ACP-bound fatty acid to the downstream amino
tion (T), and the condensation (C) domain. acid (Rausch et al., 2007). This finding is in line with our proposed
The A domain recruits a specific amino acid monomer and biosynthetic model (Klapper et al., 2016).
activates it as an aminoacyl adenylate. The sequential order
To test if the pys gene is sufficient for the assembly of pyreu-
of A domains along the assembly line thus determines, in the diones from simple precursors, we performed the heterologous
majority of cases, the primary sequence of the final peptide. expression of pys in closely related Pseudomonas protegens
Upon activation, the amino acid is loaded onto a T domain. Pf-5, as well as in phylogenetically unrelated Escherichia coli
Cell Chemical Biology 25, 1–7, June 21, 2018 ª 2018 Elsevier Ltd.
1

Please cite this article in press as: Klapper et al., Bacterial Alkaloid Biosynthesis: Structural Diversity via a Minimalistic Nonribosomal Peptide Synthe-
tase, Cell Chemical Biology (2018), https://doi.org/10.1016/j.chembiol.2018.02.013
A
B
Figure 1. The Monomodular Nonribosomal Peptide Synthetase Pys Catalyzes Pyreudione Formation
A) Pyreudione A (1) biosynthesis: (i) adenylation of L-proline and loading on the T domain, amide bond formation between secondary amine and activated fatty
(
acid; (ii) offloading via DC. (B) HPLC profiles of the native pyreudione producer P. fluorescens HKI0770 (top, solid line), and of culture extracts of E. coli HM0079
sfp pMQ72pys (i.e., heterologously expressed pys, dotted line) and P. protegens Pf-5 pMQ72pys (dashed line), control: P. protegens Pf-5 wild-type (bottom,
dot-dashed line), at l = 280 nm. * Indicates pyreudiones.
32
HM0079 sfp, which harbors
phosphopantethein transferase necessary for NRPS activation assay. Thus, a C-A didomain protein bearing an N-terminal
Gruenewald et al., 2004). To this end, we ligated the biosynthetic His-tag was recombinantly produced. All canonical amino acids
a chromosomally integrated of the Pys A domain using an ATP/[ P]pyrophosphate exchange
(
gene with an engineered ribosomal binding site into the arabi- and L-ornithine were subdivided into five pools. Every pool was
nose-inducible vector pMQ72 (Shanks et al., 2006). Upon trans- then individually subjected to the exchange assay in the presence
32
formation with the resulting plasmid, we cultured the respective of the C-A didomain. Incorporation of [ P] into ATP was
strain in the presence of arabinose. Indeed, the high-perfor- observed when pool II was added to the exchange assay (Fig-
mance liquid chromatography (HPLC) profiles of P. protegens ure 2A). The A domain was selective for L-proline, when tested
Pf-5 pMQ72pys extract and E. coli HM0079 sfp pMQ72pys individually. No other constituent of pool II led to the incorporation
32
extract resembled that of P. fluorescens HKI0770 (Figure 1B), of [ P] into ATP, nor did D-proline serve as substrate for the
with E. coli producing even higher amounts of pyreudiones A domain. We then proceeded to investigate if the A domain
than the native producer. Importantly, pyreudione production would accept amino acids bearing some structural similarity to
in E. coli indicates that no other biosynthetic genes but pys proline, such as pipecolic acid (2), azetidine 2-carboxylic acid
were required for the biosynthesis of the pyreudiones. Activated (3), pyrrole-3-carboxylic acid (4), pyrrole-2-carboxylic acid
acyl precursors thus stem from fatty acid primary metabolism, (5), L-hydroxyproline (6), and L-oxoproline (7). Interestingly, most
since the E. coli strain used for expression lacks biosynthetic of these substrates were accepted by the Pys A domain
genes that could be present in Pseudomonas sp. and provide (Figure 2B).
acyl precursors.
Since many microorganisms have the capacity to generate
L-pipecolic acid (Chang and Adams, 1971), we wondered
if the corresponding pyreudione congener, indolizidine
dione 8 (Figure 2C), could be endogenously produced by
Substrate Multispecificity of the Pys A Domain Enables
Access to Novel Pyreudiones
The simplicity of this monomodular NRPS, which effectively P. fluorescens HKI0770. Importantly, genome mining of
activates a single amino acid, is appealing for investigating P. fluorescens HKI0770 allowed the identification of D1-piperi-
fundamental mechanistic questions, such as the plasticity of an deine-2-carboxylate (P2C) reductase, which catalyzes the
isolated NRPS module. While substrate promiscuity with respect reduction of P2C into pipecolic acid via the P2C route
to the C domain was already reported (Klapper et al., 2016), the (Chang and Adams, 1971). Liquid chromatography-mass spec-
A domain could in principle also contribute to product diversifica- trometry-based identification of a compound with m/z 280.2
+
tion, as shown for instance in the cyclohexadepsipeptide [M + H] gave a first indication that the anticipated novel alka-
destruxins (Wang et al., 2012) or in the rapamycin biosynthesis loid may be present, albeit in low abundance. Chemical synthe-
(
Khaw et al., 1998). We thus analyzed the substrate specificity sis of the corresponding standard using a previously estab-
of the A domain. NRPS substrates can often be predicted by lished synthetic approach (Klapper et al., 2016), starting from
aligning ten conserved residues (A domain signature) in the sub- L-pipecolic acid (Figure 2D), confirmed endogenous production
strate-binding pockets of A domains (Stachelhaus et al., 1999). of pyreudione E (8), the indolizidine analog of pyreudione A (1),
Based on our previous biosynthetic model (Klapper et al., by the native producer.
2
016), the Pys A domain selects and activates proline. However,
Due to their low production rates, isolation of indolicidine con-
the A domain signature of Pys (D-M-L-V-M-G-V-F-A-K) bears no geners was impossible under standard conditions. Increasing
resemblance to the consensus A domain signature for proline the concentration of L-pipecolic acid in the culture medium by
described by Stachelhaus et al. (1999) (D-V-Q-L-I-A-H-V-V-K). exogenous supplementation, however, led to an increase in
Other NRPS analysis programs also failed to predict the A the production of pyreudione E (8) and other indolizidine diones.
domain specificity (Bachmann and Ravel, 2009; R o¨ ttig et al., Eventually, we were able to isolate pyreudiones E to H (8–11)
2
011). This led us to biochemically determine the substrate scope from P. fluorescens HKI0770 in sufficient quantities for
2
Cell Chemical Biology 25, 1–7, June 21, 2018

Please cite this article in press as: Klapper et al., Bacterial Alkaloid Biosynthesis: Structural Diversity via a Minimalistic Nonribosomal Peptide Synthe-
tase, Cell Chemical Biology (2018), https://doi.org/10.1016/j.chembiol.2018.02.013
A
B
C
D
Figure 2. Substrate Multispecificity of the Pys A Domain Enables Access to Novel Pyreudiones
3
A and B) Substrate specificities of the Pys A domain determined by the ATP/[ P]pyrophosphate exchange assay. The substrate with highest activity was set to
2
(
100%. The average of three replicates is shown. Error bars represent SD. (A) The canonical amino acids (including L-ornithine) were divided into five pools that
were individually tested. Of these substrates, only L-proline was accepted as substrate. (B) Substrates bearing structural similarity with L-proline were accepted
by the A domain.
(
C) Chemical structures of novel pyreudiones derived from L-pipecolic acid (8–11) and L-hydroxyproline (12–14).
ꢀ
(
D) Synthesis of pyreudione E (8). Reagents and conditions: a. Ac O, THF, NaHCO , RT; b. NEt , EDCI, N,O-dimethyl hydroxylamine, CH Cl , 0 C; c. LHMDS,
2 3 3 2 2
ꢀ
THF, ꢁ78 C; d. EDCI, CH
2 2 3 2 6 2
Cl , CH (CH ) CO H, DMAP, RT. Overall yield over four steps from L-pipecolic acid 11%. See Supplemental Information for detailed
procedures.
their structure elucidation. Nuclear magnetic resonance (NMR) vaccae with minimum inhibitory concentration (MIC)
=
spectroscopic data for the main metabolite matched those of 12.5 mg/mL (45 mM), whereas pyreudione I (12) did not show
synthesized 8, while the structures of the other congeners any antimicrobial activity (MIC >100 mg/mL). Short-chain ana-
were determined using one- and two-dimensional NMR logs of pyreudione A (1) with six or fewer carbon atoms in the
spectroscopy and high-resolution mass spectrometry (HRMS). alkyl side chain also did not show any antimicrobial activity
Interestingly, we could not identify any azetidine-based (Klapper et al., 2016). Notably, pyreudione A derivative contain-
pyreudiones by supplementing azetidine-2-carboxylic acid to ing a C12 alkyl chain displayed strongly increased activity
the culture medium. Feeding either L-hydroxy-or L-oxoproline, against M. vaccae and M. aurum with MIC = 0.4 mg/mL
however, led in both cases to the production of pyreudiones I (1.2 mM). The influence of the compound’s lipophilicity on
to
P. fluorescens HKI0770 (Figure 2C).
K (12–14), which were otherwise not produced by bioactivity due to variations in alkyl chain lengths or the pres-
ence of an additional hydroxyl group (pyreudione I, 12), needs
to be addressed in future structure-activity relationship and
mode-of-action studies.
Antimicrobial Activity
The main congeners pyreudione E (8) and pyreudione I (12)
were tested for amoebicidal activity, as well as for antimicrobial Expression Analysis of pys-Homolog pysPENT from
activity. Interestingly, only naturally occurring pyreudione E (8) Pseudomonas entomophila
displayed amoebicidal activity comparable to previously The prerequisite for the production of a diversity of pyreu-
reported pyreudiones (1–11 mg/mL or 3–41 mM, respectively, diones is not only a multispecific A domain, but the TE domain
Klapper et al., 2016) with half maximal inhibitory concentration has to promote the DC irrespective of the amino acid loaded
(
IC50) (Dictyostelium discoideum) = 6 mg/mL (21 mM). Further on the T domain. In order to decipher the evolutionary history
antimicrobial testing revealed an antibiotic activity of pyreu- of the Pys TE domain, we investigated the presence of C-A-T-
dione E (8) against Mycobacterium aurum and Mycobacterium TE-type monomodular NRPS in other bacterial species using a
Cell Chemical Biology 25, 1–7, June 21, 2018
3

Please cite this article in press as: Klapper et al., Bacterial Alkaloid Biosynthesis: Structural Diversity via a Minimalistic Nonribosomal Peptide Synthe-
tase, Cell Chemical Biology (2018), https://doi.org/10.1016/j.chembiol.2018.02.013
Figure 3. Expression Analysis of
Homolog from P. entomophila (pysPENT
A) HPLC profiles of the native pyreudione producer
a pys-
A
B
)
(
P. fluorescens HKI0770 (top, solid line), and of cul-
ture extracts of E. coli HM0079 sfp pMQ72pysPENT
(
(
(
dotted line) and P. protegens Pf-5 pMQ72pysPENT
dashed line), control: P. entomophila wild-type
bottom, dot-dashed line) at l = 280 nm, * indicates
pyreudiones.
B) RT-qPCR expression analysis of pysPENT of
(
P. entomophila (DSMZ 28517) and pys of
P. fluorescens HKI0770, respectively. Both strains
were cultured in SM/5 medium to OD600 = 1.
Expression level of pysPENT was set to 1. The respective rpoD gene was used for internal normalization. The average of three independent biological replicates is
shown. Three technical replicates were conducted for each biological replicate. Error bars represent SD of biological replicates.
protein-protein BLAST search of Pys against the non-redun- Lou, 2010). For the described monomodular NRPSs, however,
dant database of NCBI. We found a variety of bacteria that a DC yields an E/Z mixture of the tetramic acids (with the Z
harbor similar C-A-T-TE-type NRPSs (Table S2); however, isomer being the major form). While it remains unclear whether
none of the genes had been linked to a secondary metabolite the TE domain catalyzes the cyclization during the pyreudione
so far.
biosynthesis, TE-like domains with DC activity have been
The monomodular NRPS of the insect pathogen Pseudomonas described. TE-like domains related to the heat-stable
entomophila L48 (PysPENT) showed 81% similarity to that of antifungal factor synthetase from Lysobacter enzymogenes
P. fluorescens HKI0770. Moreover, the A domain signatures of (Du and Lou, 2010; Yu et al., 2006) and free-standing
both NRPSs were identical, suggesting a similar substrate TE-like domains similar to LipX2 of the lipomycin pathway of
scope of Pys and PysPENT
sequenced P. entomophila (DSMZ 28517) appeared to be a ever, neither of the two pyreudione synthetase TE domains
good candidate to search for the production of pyreudiones or shows close phylogenetic relationship to the above-
. Due to this high similarity, Streptomyces aureofaciens for instance were reported. How-
a
analogs thereof. We cultured P. entomophila in various media mentioned groups. Despite intensive efforts, we could not
and extracted the cultures with ethyl acetate. Unfortunately, we detect any specific motif shared by all TE domains that
could not detect the presence of any compounds with a yield DC products. Interestingly, the pyreudione synthetase
similar UV absorbance spectrum to the pyreudiones in these TE domains are more closely related to the TE of serrawettin
culture extracts (Figure 3A). Transcriptional analysis of pysPENT synthetase SwrW, which exhibits the identical C-A-T-TE
in comparison to pys by quantitative reverse transcription domain architecture. Yet, this TE supports macrocyclization,
PCR (RT-qPCR) showed that the former was 30-fold less ex- rather than DC. Still, it is unusual in that it catalyzes a
pressed in SM/5 medium (Figure 3B), a medium that led to partic- dimerization, most likely via two following transesterification
ularly good production rates of pyreudiones in P. fluorescens steps (Li et al., 2005). Notably, tetramic acid formation by a
HKI0770.
TE domain can alternatively proceed via lactamization as
Finally, we aimed at expressing pysPENT in P. protegens Pf-5 proposed for the hybrubins (HbnA) (Zhao et al., 2016).
and E. coli HM0079 sfp (using pMQ72pysPENT). Eventually, this
approach led to the production of pyreudiones A–D (Klapper
SIGNIFICANCE
et al., 2016) as produced by P. fluorescens HKI0770 (Figure 3A).
Thus, despite being fully functional, pysPENT is silent in
P. entomophila. Since transformation of P. entomophila with
Nonribosomal peptides (NRPs) are important antibiotics,
anticancer agents, or immunomodulators.
A detailed
pMQ72pysPENT led to the production of the same set of
pyreudiones, it is likely that P. entomophila requires specific
environmental cues in order to activate pysPENT expression.
The established expression system is thus suitable for express-
ing other homologous monomodular NRPS with so far unknown
products in future studies.
biosynthetic analysis of a strongly bioactive class of NRPs
enabled the identification and isolation of novel alkaloids
that would otherwise have been overlooked. The amoe-
bicidal pyreudione alkaloids are formed by the sole action
of the monomodular nonribosomal peptide synthetase
(
(
NRPS) Pys. A homologous NRPS in P. entomophila
PysPENT), which has 81% similarity to Pys, appears silent
Phylogenetic Analyses of TE-like Domains of
Bacterial NRPS
under different culture conditions. Heterologous expres-
sion of pys and pysPENT in E. coli and different Pseudomonas
With the amino acid sequences of Pys and PysPENT, as well as species, however, led to the production of the same set of
4 TE domain sequences of other representative NRPSs, we pyreudiones as seen in P. fluorescens HKI0770. The
2
constructed a maximum likelihood tree (Figure 4). As ex- simplicity and plasticity of this stand-alone monomodular
pected, both TE domains of Pys and PysPENT were phyloge- NRPS is a great system for subsequent bioengineering
netically closely related. In NRPS, TE domains typically of NRPSs in order to access a variety of new natural
catalyze the ‘‘canonical’’ release of the NRP via hydrolysis or products. Using the established expression system,
macrocyclization, as seen for cyclic lipopeptides (Du and the substrate scopes of other monomodular NRPSs,
4
Cell Chemical Biology 25, 1–7, June 21, 2018

Please cite this article in press as: Klapper et al., Bacterial Alkaloid Biosynthesis: Structural Diversity via a Minimalistic Nonribosomal Peptide Synthe-
tase, Cell Chemical Biology (2018), https://doi.org/10.1016/j.chembiol.2018.02.013
Figure 4. Phylogenetic Analysis of 26 TE-like Domains of Bacterial NRPS
The observed mode of release, such as macrolactonization, DC, or hydrolysis, is indicated. The scale bar represents 50 substitutions per 100 amino acids.
Bootstrap values higher than 45% are shown. The respective proteins and the resulting nonribosomal peptides are shown in Table S3.
considering both the C domain and A domain specificity, STAR+METHODS
can now be investigated. Importantly, this class of
enzymes seems to be widely spread across the bacterial Detailed methods are provided in the online version of this paper
kingdom and their biosynthetic products await discovery. and include the following:
Identification of the presumed environmental cues for
d KEY RESOURCES TABLE
their activation would enhance our knowledge of their
d CONTACT FOR REAGENT AND RESOURCE SHARING
ecological functions.
Cell Chemical Biology 25, 1–7, June 21, 2018
5

Please cite this article in press as: Klapper et al., Bacterial Alkaloid Biosynthesis: Structural Diversity via a Minimalistic Nonribosomal Peptide Synthe-
tase, Cell Chemical Biology (2018), https://doi.org/10.1016/j.chembiol.2018.02.013
d EXPERIMENTAL MODEL AND SUBJECT DETAILS
B Dictyostelium discoideum AX2
B Pseudomonas Strains
Baltz, R.H. (2014). Combinatorial biosynthesis of cyclic lipopeptide antibiotics:
a model for synthetic biology to accelerate the evolution of secondary metab-
olite biosynthetic pathways. ACS Synth. Biol. 3, 748–758.
B Escherichia coli Strains
Baltz, R.H., Miao, V., and Wrigley, S.K. (2005). Natural products to drugs:
daptomycin and related lipopeptide antibiotics. Nat. Prod. Rep. 22, 717–741.
d METHOD DETAILS
B General Biological Methods
Bode, H.B., Brachmann, A.O., Jadhav, K.B., Seyfarth, L., Dauth, C., Fuchs,
S.W., Kaiser, M., Waterfield, N.R., Sack, H., Heinemann, S.H., and Arndt,
H.D. (2015). Structure elucidation and activity of kolossin A, the d-/l- pentade-
capeptide product of a giant nonribosomal peptide synthetase. Angew. Chem.
Int. Ed. 54, 10352–10356.
B Instrumentation
B Heterologous Expression of pys and pysPENT
B Pys C-A Didomain Expression Vector
B Heterologous Protein Production&Purification
Bradford, M.M. (1976). A rapid and sensitive method for the quantitation of
microgram quantities of protein utilizing the principle of protein-dye binding.
Anal. Biochem. 72, 248–254.
3
2
B ATP-[ P]pyrophosphate Exchange Assay
B LC-MS-Based Metabolomic Profiling
B General Synthetic Methods
Chang, Y.F., and Adams, E. (1971). Induction of separate catabolic pathways
for L- and D-lysine in Pseudomonas putida. Biochem. Biophys. Res. Commun.
45, 570–577.
B Synthesis of Pyreudione E
B Feeding Experiments
B Isolation of Pyreudiones E – H, 8 – 11
B Isolation of Pyreudiones I – K, 12 – 14
B Analytical Data of Pyreudiones E – K, 8 – 14
B Growth Inhibition Assay
Clinical and Laboratory Standards Institute. (2006). Methods for Dilution
Antimicrobial Susceptibility Tests for Bacteria that Grow Aerobically,
Approved Standard, Vol. 26, Fourth Edition (Clinical and Laboratory
Standards Institute).
B Antimicrobial and MIC Tests
Du, L., and Lou, L. (2010). PKS and NRPS release mechanisms. Nat. Prod.
Rep. 27, 255–278.
B Protein BLAST
B Expression Analyses of pys and pysPENT
B Phylogenetic Analyses of TE Domains
d QUANTIFICATION AND STATISTICAL ANALYSIS
Edgar, R.C. (2004). MUSCLE: multiple sequence alignment with high accuracy
and high throughput. Nucleic Acids Res. 32, 1792–1797.
Gervais, A.L., Marques, M., and Gaudreau, L. (2010). PCRTiler: automated
design of tiled and specific PCR primer pairs. Nucleic Acids Res. 38,
W308–W312.
SUPPLEMENTAL INFORMATION
Gomi, S., Imamura, K., Yaguchi, T., Kodama, Y., Minowa, N., and Koyama, M.
(
1994). PF1018, a novel insecticidal compound produced by Humicola sp.
Supplemental Information includes two figures, four tables, and one data
file and can be found with this article online at https://doi.org/10.1016/j.
chembiol.2018.02.013.
J. Antibiot. (Tokyo) 47, 571–580.
Gruenewald, S., Mootz, H.D., Stehmeier, P., and Stachelhaus, T. (2004). In vivo
production of artificial nonribosomal peptide products in the heterologous host
Escherichia coli. Appl. Environ. Microbiol. 70, 3282–3291.
ACKNOWLEDGMENTS
Hur, G.H., Vickery, C.R., and Burkart, M.D. (2012). Explorations of catalytic
domains in non-ribosomal peptide synthetase enzymology. Nat. Prod. Rep.
We thank A. Perner, H. Heinecke, and Dr. S. Schieferdecker (Hans Kn o¨ ll
Institute, Jena) for MS and NMR measurements; Dr. M. Roth, K. Willing,
M. Steinacker, J. Sch o¨ nemann, and P. Berthel for large-scale fermentations;
C. Weigel for antimicrobial activity testing; and H. Kries for providing E. coli
HM0079 sfp. This work was funded by the DFG (STA 1431/2-1, and CRC
2
9, 1074–1098.
Huson, D.H. (1998). SplitsTree: analyzing and visualizing evolutionary data.
Bioinformatics 14, 68–73.
Jeong, Y.C., Bikadi, Z., Hazai, E., and Moloney, M.G. (2014). A detailed study
of antibacterial 3-acyltetramic acids and 3-acylpiperidine-2,4-diones.
ChemMedChem 9, 1826–1837.
1
127 ChemBioSys) We are also grateful for financial support from the
Leibniz Association, Carl Zeiss Foundation (D.B., G.L.), an Aventis Founda-
tion PhD fellowship (M.K.), and for a fellowship of the Daimler and Benz
Foundation (P.S.).
Khaw, L.E., B o¨ hm, G.A., Metcalfe, S., Staunton, J., and Leadlay, P.F. (1998).
Mutational biosynthesis of novel rapamycins by a strain of Streptomyces
hygroscopicus NRRL 5491 disrupted in rapL, encoding a putative lysine cyclo-
deaminase. J. Bacteriol. 180, 809–814.
AUTHOR CONTRIBUTIONS
P.S., M.K., D.B., R.H., and G.L. planned experiments; M.K., D.B., and G.L.
Klapper, M., G o¨ tze, S., Barnett, R., Willing, K., and Stallforth, P. (2016).
i
performed experiments (protein isolation, ATP-PP exchange assay and
Bacterial alkaloids prevent amoebal predation. Angew. Chem. Int. Ed. 55,
phylogeny: D.B., G.L.); P.S. and M.K. wrote the manuscript; D.B., G.L., and
R.H. edited the manuscript.
8
944–8947.
Le, S.Q., and Gascuel, O. (2008). An improved general amino acid replacement
matrix. Mol. Biol. Evol. 25, 1307–1320.
DECLARATION OF INTERESTS
Li, H., Tanikawa, T., Sato, Y., Nakagawa, Y., and Matsuyama, T. (2005).
Serratia marcescens gene required for surfactant serrawettin W1 production
encodes putative aminolipid synthetase belonging to nonribosomal peptide
synthetase family. Microbiol. Immunol. 49, 303–310.
The authors declare no competing interests.
Received: December 13, 2017
Revised: January 26, 2018
Accepted: February 22, 2018
Published: March 29, 2018
Liu, J., Farmer, J.D., Lane, W.S., Friedman, J., Weissman, I., and Schreiber,
S.L. (1991). Calcineurin is a common target of cyclophilin-cyclosporin A and
FKBP-FK506 complexes. Cell 66, 807–815.
Livak, K.J., and Schmittgen, T.D. (2001). Analysis of relative gene expression
REFERENCES
ꢁDDC
data using real-time quantitative PCR and the 2
02–408.
T
method. Methods 25,
4
Bachmann, B.O., and Ravel, J. (2009). Methods for in silico prediction of
microbial polyketide and nonribosomal peptide biosynthetic pathways from
DNA sequence data. Methods Enzymol. 458, 181–217.
Marahiel, M.A. (2016). A structural model for multimodular NRPS assembly
lines. Nat. Prod. Rep. 33, 136–140.
6
Cell Chemical Biology 25, 1–7, June 21, 2018

Please cite this article in press as: Klapper et al., Bacterial Alkaloid Biosynthesis: Structural Diversity via a Minimalistic Nonribosomal Peptide Synthe-
tase, Cell Chemical Biology (2018), https://doi.org/10.1016/j.chembiol.2018.02.013
Murray, A., Proctor, G.R., and Murray, P.J. (1996). An efficient route to
the pyrrolizidine ring system via an N-acyl anion cyclisation process.
Tetrahedron 52, 3757–3766.
Tamura, K., Stecher, G., Peterson, D., Filipski, A., and Kumar, S. (2013).
MEGA6: molecular evolutionary genetics analysis version 6.0. Mol. Biol.
Evol. 30, 2725–2729.
Pizzorno, M.T., and Albonico, S.M. (1977). Novel synthesis of 5,6,7,8-tetrahy-
Walsh, C.T., and Fischbach, M.A. (2010). Natural products version 2.0:
droindolizine. J. Org. Chem. 42, 909–910.
connecting genes to molecules. J. Am. Chem. Soc. 132, 2469–2493.
Rausch, C., Hoof, I., Weber, T., Wohlleben, W., and Huson, D.H. (2007).
Phylogenetic analysis of condensation domains in NRPS sheds light on their
functional evolution. BMC Evol. Biol. 7, 78.
Wang, B., Kang, Q., Lu, Y., Bai, L., and Wang, C. (2012). Unveiling the biosyn-
thetic puzzle of destruxins in Metarhizium species. Proc. Natl. Acad. Sci. USA
109, 1287–1292.
R o¨ ttig, M., Medema, M.H., Blin, K., Weber, T., Rausch, C., and Kohlbacher, O.
Wiegand, I., Hilpert, K., and Hancock, R.E. (2008). Agar and broth dilution
methods to determine the minimal inhibitory concentration (MIC) of antimicro-
bial substances. Nat. Protoc. 3, 163–175.
(2011). NRPSpredictor2—a web server for predicting NRPS adenylation
domain specificity. Nucleic Acids Res. 39, W362–W367.
Shanks, R.M., Caiazza, N.C., Hinsa, S.M., Toutain, C.M., and O’Toole, G.A.
Yu, F., Zaleta-Rivera, K., Zhua, X., Huffman, J., Millet, J.C., Harris, S.D., Yuen,
G., Li, X.C., and Du, L. (2006). Structure and biosynthesis of heat-stable
antifungal factor (HSAF), a broad-spectrum antimycotic with a novel mode
of action. Antimicrob. Agents Chemother. 51, 64–72.
(
2006). Saccharomyces cerevisiae-based molecular tool kit for manipulation
of genes from gram-negative bacteria. Appl. Environ. Microbiol. 72,
027–5036.
5
Sieber, S.A., and Marahiel, M.A. (2005). Molecular mechanisms underlying
nonribosomal peptide synthesis: approaches to new antibiotics. Chem. Rev.
Zhao, Z., Shi, T., Xu, M., Brock, N.L., Zhao, Y.-L., Wang, Y., Deng, Z., Pang, X.,
and Tao, M. (2016). Hybrubins: bipyrrole tetramic acids obtained by crosstalk
between a truncated undecylprodigiosin pathway and heterologous tetramic
acid biosynthetic genes. Org. Lett. 18, 572–575.
105, 715–738.
Stachelhaus, T., Mootz, H.D., and Marahiel, M.A. (1999). The specificity-
conferring code of adenylation domains in nonribosomal peptide synthetases.
Chem. Biol. 6, 493–505.
Ziemert, N., Podell, S., Penn, K., Badger, J.H., Allen, E., and Jensen, P.R.
(
2012). The natural product domain seeker NaPDoS: a phylogeny based
S u€ ssmuth, R.D., and Mainz, A. (2017). Nonribosomal peptide synthesis-
principles and prospects. Angew. Chem. Int. Ed. 56, 3770–3821.
bioinformatic tool to classify secondary metabolite gene diversity. PLoS One
7, e34064.
Cell Chemical Biology 25, 1–7, June 21, 2018
7

Please cite this article in press as: Klapper et al., Bacterial Alkaloid Biosynthesis: Structural Diversity via a Minimalistic Nonribosomal Peptide Synthe-
tase, Cell Chemical Biology (2018), https://doi.org/10.1016/j.chembiol.2018.02.013
STAR+METHODS
KEY RESOURCES TABLE
REAGENT or RESOURCE
Bacterial and Virus Strains
Pseudomonas fluorescens HKI0770 Dclp
Escherichia coli Top10
SOURCE
IDENTIFIER
Klapper et al., 2016
N/A
ThermoFisher
Cat#C404010
N/A
Escherichia coli ET12567 pUZ8002
Pseudomonas protegens Pf-5
Pseudomonas entomophila DSM28517
Escherichia coli HM0079 sfp
Escherichia coli KRX
Kothe Lab
Clardy Lab
N/A
DSMZ
DSM28517
N/A
Kries Lab, Gruenewald et al., 2004
Promega
Cat#L3002
N/A
Bacillus subtilis 6633
HKI microbial strain collection
HKI microbial strain collection
HKI microbial strain collection
HKI microbial strain collection
HKI microbial strain collection
HKI microbial strain collection
HKI microbial strain collection
HKI microbial strain collection
HKI microbial strain collection
HKI microbial strain collection
Staphylococcus aureus SG511
Staphylococcus aureus SG511
Escherichia coli SG458
N/A
N/A
N/A
Pseudomonas aeruginosa K799/61
Mycobacterium vaccae 10670
Mycobacterium aurum SB66
Sporobolomyces salmonicolor 549
Candida albicans C.A.
N/A
N/A
N/A
N/A
N/A
Penicillum notatum JP36
Biological Samples
N/A
gDNA of Pseudomonas fluorescens HKI0770
gDNA of Pseudomonas entomophila DSM28517
Chemicals, Peptides, and Recombinant Proteins
CutSmartꢀ Buffer
Klapper et al., 2016
This study
N/A
N/A
New England Biolabs
New England Biolabs
New England Biolabs
New England Biolabs
VWR
Cat#B7204S
XmaI
Cat#R0180S
SalI-HFꢀ
Cat#R3138S
HindIII-HFꢀ
Cat#R3104S
L-rhamnose
Cat#ICNA0210280980;CAS:10030-85-0
Cat#1B1473;CAS:5328-37-0
Cat#I0001;CAS:288-32-4
Cat#87153.0006;CAS:865-49-6
Cat#2475.1;CAS:1405-41-0
Cat#T832.1;CAS:25389-94-0
Cat#3886.2;CAS:56-75-7
Cat#K029.1
L-arabinose
VWR
Imidazole
TCI
Chloroform, deuterated
VWR Chemicals
Carl Roth
Gentamicinsulfate solution (gentamicin)
Kanamycinsulfate (kanamycin)
Chloramphenicol
Carl Roth
Carl Roth
Ampicillin sodium salt (ampicillin)
SM/5
Carl Roth
Formediumꢁ
Macherey-Nagel
GE Healthcare
This study
Cat#SMB50102
Protino nickel-nitrilo acetic resin
PD-10 desalting column
Cat#745400.500
Cat#17-0851-01
Pys N-His
3
6
C-A didomain
N/A
2
[
P]pyrophosphate
Perkin Elmer
Carl Roth
Cat#NEX019001MC
Cat#0016.3
Scintillation liquid
L-proline
TCI
Cat#P0481;CAS:147-85-3
Cat#A10682;CAS:344-25-2
Cat#CC10002;CAS:3105-95-1
Cat#CC10003;CAS:1723-00-8
Cat#093784;CAS:2133-34-8
D-proline
Alfa Aesar
L-pipecolic acid 2
D-pipecolic acid 2
L-azetidine 2-carboxylic acid 3
Carbolution Chemicals
Carbolution Chemicals
Fluorochem Ltd
(Continued on next page)
e1 Cell Chemical Biology 25, 1–7.e1–e9, June 21, 2018

Please cite this article in press as: Klapper et al., Bacterial Alkaloid Biosynthesis: Structural Diversity via a Minimalistic Nonribosomal Peptide Synthe-
tase, Cell Chemical Biology (2018), https://doi.org/10.1016/j.chembiol.2018.02.013
Continued
REAGENT or RESOURCE
D-azetidine 2-carboxylic acid 3
Pyrrole-2-carboxylic acid 5
Pyrrole-3-carboxylic acid 4
L-hydroxyproline 6
SOURCE
IDENTIFIER
Fluorochem Ltd
Sigma Aldrich
Sigma Aldrich
Sigma Aldrich
Alfa Chemistry
ThermoFisher Scientific
Cat#040006;CAS:7729-30-8
CAT#P73609;CAS:634-97-9
CAT#689416;CAS:931-03-3
CAT#H54409;CAS:51-35-4
Cat#ACM4347186;CAS:4347-18-6
Cat#60108-301
L-oxoproline 7
TM
HyperSep 1g C18 column
Critical Commercial Assays
QIAprepꢀ Spin Miniprep Kit
GeneJET Gel Extraction Kit
QIAamp DNA Mini Kit
Qiagen
Cat#27106
Cat#K0692
Cat#51304
Cat#M0492S
Cat#K1082
Cat#E2611S
Cat#E0544S
Cat#74104
Cat#76506
Cat#AM1907
Cat#EP0442
Cat#A25742
N/A
Thermo Fisher Scientific
Qiagen
Q5ꢀ High-Fidelity 2x Master Mix
DreamTaq Green PCR Master Mix 2x
Gibson assemblyꢀ Master Mix
Q5ꢀ Site-Directed Mutagenesis Kit
RNeasyꢀ Mini Kit
New England Biolabs
Thermo Scientific
New England Biolabs
New England Biolabs
Qiagen
RNAprotectꢀ Bacteria Reagent
TM
TURBO DNA-free kit
Qiagen
Thermo Fisher Scientific
Thermo Fisher Scientific
Thermo Fisher Scientific
LGC Genomics, Berlin
RevertAid Reverse Transcriptase
PowerUpTM SYBRꢀ Green Master Mix
DNA Sanger sequencing
Deposited Data
WP_064118616, Pys protein sequence
NCBI
NCBI
https://www.ncbi.nlm.nih.gov/
protein/ WP_064118616
WP_011533552, PysPENT protein sequence
https://www.ncbi.nlm.nih.gov/
protein/WP_011533552
Experimental Models: Cell Lines
Dictyostelium discoideum AX2
dictybase.org
N/A
Experimental Models: Organisms/Strains
Pseudomonas fluorescens HKI0770
Escherichia coli Top10 pMQ72pys
Klapper et al., 2016
This study
This study
This study
This study
This study
This study
This study
This study
This study
This study
N/A
N/A
N/A
N/A
N/A
N/A
N/A
N/A
N/A
N/A
N/A
Escherichia coli Top10 pMQ72pysPENT
Escherichia coli ET12567 pUZ8002 pMQ72pysPENT
Escherichia coli ET12567 pUZ8002 pMQ72pysPENT
Pseudomonas protegens Pf-5 pMQ72pys
Pseudomonas protegens Pf-5 pMQ72pysPENT
Pseudomonas entomophila pMQ72pysPENT
Escherichia coli HM0079 sfp pMQ72pys
Escherichia coli HM0079 sfp pMQ72pysPENT
Escherichia coli KRX pBK01
Oligonucleotides
Primers for all cloning procedures: c.f. Table S1
Invitrogen ThermoFisher
Scientific (Darmstadt)
N/A
Random Hexamers
Recombinant DNA
pMQ97_gfp
Thermo Fisher Scientific
Cat#N8080127
Shanks et al., 2006
Shanks et al., 2006
This study
N/A
N/A
N/A
N/A
N/A
pMQ72
pMQ72rbs
pMQ72pys
This study
pMQ72pysPENT
This study
(Continued on next page)
Cell Chemical Biology 25, 1–7.e1–e9, June 21, 2018 e2

Please cite this article in press as: Klapper et al., Bacterial Alkaloid Biosynthesis: Structural Diversity via a Minimalistic Nonribosomal Peptide Synthe-
tase, Cell Chemical Biology (2018), https://doi.org/10.1016/j.chembiol.2018.02.013
Continued
REAGENT or RESOURCE
pET28a(+)
SOURCE
IDENTIFIER
Cat#69864
N/A
Novagen/Merck
This study
pBK01
Software and Algorithms
Gibson Assembly
TM
New England Biolabs
New England Biolabs
NCBI
http://nebuilder.neb.com
NEBaseChanger
https://nebasechanger.neb.com/
NCBI Protein Blast
PCRTiler
https://blast.ncbi.nlm.nih.gov/
Blast.cgi?PAGE=Proteins
PCRTiler v1.42,
http://pcrtiler.genap.ca/PCRTiler/
Gervais et al., 2010
ChemBioDraw Ultra 13.0
Geneious 11
PerminElmer
Biomatters
http://www.cambridgesoft.com/
https://www.geneious.com/
https://www.bruker.com/
Topspin 3.2
Bruker
GraphPad Prism, Version 5.03
Graphpad Software, Inc.
https://www.graphpad.com/
CONTACT FOR REAGENT AND RESOURCE SHARING
Further information and requests for resources and reagents should be directed to and will be fulfilled by the Lead Contact, Dr. Pierre
Stallforth (pierre.stallforth@leibniz-hki.de).
EXPERIMENTAL MODEL AND SUBJECT DETAILS
Dictyostelium discoideum AX2
Axenic Dictyostelium discoideum AX2 cells (dictybase.org) were cultured in HL/5 liquid medium (Formediumꢁ, UK). An aliquot
(
After five days at 22 C, when confluency was reached, the cells were detached by gentle pipetting and split (1/10) into a new cell
5–10 mL) of a D. discoideum spore glycerol stock was added to 10 mL HL5 medium in a cell culture plate (Sarstedt, Germany).
ꢀ
ꢀ
culture plate. Adherent cells were detached and used to inoculate a continuous shaking culture (22 C, 135 rpm). Cultures were split
after 3 – 4 days to keep the cells in their exponential growth phase. Cultures were renewed monthly from spore stocks.
Pseudomonas Strains
Pseudomonas fluorescens HKI0770, Pseudomonas fluorescens HKI0770 Dclp, Pseudomonas entomophila DSM28517 and
Pseudomonas protegens Pf-5 were propagated on SM/5 solid agar or grown in SM/5 liquid medium (Formediumꢁ, UK) on a gyratory
shaker at 160 rpm at 22 C unless otherwise noted.
ꢀ
When required, media were supplemented with gentamicin (75 mg/mL for P. entomophila DSM28517 mutants, or 50 mg/mL
for P. protegens Pf-5 mutants, Carl Roth, Germany), kanamycin (50 mg/mL, Carl Roth, Germany), chloramphenicol (50 mg/mL,
Carl Roth, Germany), and ampicillin (100 mg/mL, Carl Roth, Germany).
Escherichia coli Strains
Escherichia coli strains were cultured at 37 C in LB liquid or on solid agar medium (Carl Roth, Germany). When required, media were
ꢀ
supplemented with gentamicin (15 mg/mL, Carl Roth, Germany), kanamycin (50 mg/mL, Carl Roth, Germany), and chloramphenicol
(
50 mg/mL, Carl Roth, Germany).
METHOD DETAILS
General Biological Methods
Glycerol stocks of bacterial strains were prepared by mixing 1 mL overnight culture of the respective strain with 0.5 mL of 60% (v/v)
ꢀ
aq. glycerol and stored at –80 C.
Plasmid constructions were carried out using the Gibson Assembly method (New England Biolabs, software: http://nebuilder.neb.
com). Plasmids were isolated using QIAprepꢀ Spin Miniprep Kit (Qiagen), and PCR reaction products were purified for subsequent
cloning and sequencing by gel extraction using GeneJET Gel Extraction Kit (Thermo Scientific). PCR reactions were carried out using
Q5ꢀ High-Fidelity2x Master Mix(New England Biolabs)forapplications requiringhigh fidelity such as sequencing and cloning, for other
applications (e.g. colony PCR) DreamTaq Green PCR Master Mix 2x (Thermo Scientific, Darmstadt) was used. Restriction digests were
carried out using NEB restriction enzymes and CutSmartꢀ Buffer (New England Biolabs). Genomic DNA was isolated from 3 mL bac-
terial overnight culture in SM/5 medium using the QIAamp DNA Mini Kit (Qiagen). Oligonucleotides used in this study were purchased
from Invitrogen (ThermoFisher Scientific, Darmstadt). DNA Sanger sequencing was performed by LGC Genomics, Berlin.
e3 Cell Chemical Biology 25, 1–7.e1–e9, June 21, 2018

Please cite this article in press as: Klapper et al., Bacterial Alkaloid Biosynthesis: Structural Diversity via a Minimalistic Nonribosomal Peptide Synthe-
tase, Cell Chemical Biology (2018), https://doi.org/10.1016/j.chembiol.2018.02.013
Instrumentation
1
All 1D ( H, C) and 2D NMR spectra ( H- H COSY, HSQC, HMBC, NOESY, homonuclear J-resolved H) were recorded in deuterated
13
1
1
1
chloroform (Carl Roth, Germany) on Bruker AVANCE III 300, 500 and 600 MHz (equipped with a Bruker Cryo Platform) instruments.
1
= 7.26/77.0 for H and
The chemical shifts are reported in parts per million (d) relative to the resonance of the residual solvent (d CHCl
3
1
3
C spectra, respectively). Coupling constants (J) are reported in Hertz (Hz). HRESI-MS measurements were conducted on a Thermo
Fisher Exactive Orbitrap either by direct injection or in combination with a Thermo Accela HPLC system. The system is equipped with
an electrospray ion source and a Betasil 100-3 C18 column (150 3 2.1 mm). The following elution gradient was used: solvent A: H O +
.1% HCOOH, solvent B: acetonitrile, gradient: 5% B for 1 min, 5% to 98% B in 15 min, 98% B for 15 min, flow rate: 0.2 mL/min,
2
0
injection: 5 mL. UHPLC-MS was performed on a Shimadzu System (LC-30AD, SPD-M30A, LCMS-2020). The system is equipped
˚
with an electrospray ion source and a Kinetexꢀ C18 column (50 3 2.1 mm, 1.7 mm, 100 A, Phenomenex). The following elution
gradient was used: solvent A: H
2
O + 0.1% HCOOH, solvent B: acetonitrile + 0.1% HCOOH, gradient: 10% B for 0.5 min, 10% to
1
00% B in 8 min, 100% B for 3 min, flow rate: 0.7 mL/min, injection: 10 mL. Semi-preparative HPLC purification was conducted
˚
on a Shimadzu HPLC System (LC-20AD, SPD-M20A) using a Lunaꢀ C18(2) column (250 3 10 mm, 5 mm, 100 A, Phenomenex).
Acetonitrile + 0.1% HCOOH and H
using a 0.5 dm cuvette on a JASCO P-1020 polarimeter at 25 C (unless otherwise noted).
2
O + 0.1% HCOOH was used as mobile phase. Optical rotation measurements were performed
ꢀ
Heterologous Expression of pys and pysPENT
For heterologous gene expression, Pseudomonas protegens Pf-5 and E. coli HM0079 sfp were used as host. Suitability of the pMQ
expression vector series (Shanks et al., 2006) was tested by introducing pMQ97_gfp into P. protegens Pf-5 (via biparental mating with
E. coli ET12567 pUZ8002 pMQ97_gfp) and chemically competent E. coli HM0079 sfp. After induction with 100 mM L-arabinose (18 h),
GFP was produced and fluorescent bacterial cells were visualized using a fluorescence microscope (Leika DM4500 B).
R
A strategy based on a gentamicin resistance (gent ) selection marker was used to introduce pMQ72 shuttle vectors, encoding the
full pys NRPS of P. fluorescens HKI0770 or pysPENT of P. entomophila (sequences available at the National Center for Biotechnology
Information website NCBI, WP_064118616 and WP_011533552). The corresponding pMQ72 expression vectors (Shanks et al.,
2
006) were constructed using the Gibson Assembly method. First, pMQ72pys was assembled, upon which the engineered ribosomal
binding site (RBS, sequence AGGAGA) and spacer region present in pMQ97 was introduced, while keeping the XmaI site intact.
The parent plasmid pMQ72 was linearized by XmaI and HindIII restriction. The NRPS gene pys was PCR amplified from the genomic
DNA of P. fluorescens HKI0770 using the primer pair pMQ72pysRBS_F/R (c.f. Table S1). These primers include a sequence of about
20 – 25 bp that overlap with the linearized vector, and the forward primer additionally contains the RBS and spacer (the primers were
designed using: http://nebuilder.neb.com). The PCR product was ligated into the pMQ72 expression vector using the standard
Gibson Assembly protocol to yield expression vector pMQ72pys. The vector pMQ72pys was transformed into chemically competent
ꢀ
E. coli Top10 cells via heat shock at 42 C. Plasmids were purified from overnight cultures (LB 15 mg/mL gentamicin) of single colonies
using the QIAprepꢀ Spin Miniprep Kit and sequenced.
From pMQ72pys, we generated a mutagenized plasmid, which allows for convenient expression of other pys homologs.
We designed the expression vector to contain a cutting site 7 bp downstream of the RBS and concomitantly, we removed
the pys ORF. To this end, pMQ72pys plasmid containing an RBS and a spacer region was mutagenized using the Q5ꢀ
Site-Directed Mutagenesis Kit (NEB), to replace the pys gene (+1 bp upstream) by a short sequence containing a SalI and HindIII
TM
cutting site (GTCGACCTGCAGGCATGCAAGCT). Primers were designed using NEBaseChanger
(setting: substitution).
Following the manufacturers protocol, the vector pMQ72rbs was PCR amplified from pMQ72pys template using primers
Q5SDM_pMQ72_F and Q5SDM_pMQ72_R (c.f. Table S1), subsequently treated with KLD enzyme mix (kinase, ligase, DpnI)
ꢀ
and transformed into chemically competent E. coli Top10 cells via heat shock at 42 C. Plasmids were purified from overnight
cultures of single colonies using the QIAprepꢀ Spin Miniprep Kit, sequenced, and subsequently linearized by SalI and HindIII
restriction.
The NRPS gene pysPENT was PCR amplified from the genomic DNA of P. entomophila using the primer pair pMQ72pysPENT_F/R
c.f. Table S1). These primers include a sequence of about 20 – 25 bp that overlap with the linearized vector pMQ72rbs (the primers
(
were designed using: http://nebuilder.neb.com). The vector pMQ72pysPENT was transformed into chemically competent E. coli
ꢀ
Top10 cells via heat shock at 42 C. Plasmids were purified from overnight cultures (LB 15 mg/mL gentamicin) of single colonies using
the QIAprepꢀ Spin Miniprep Kit and sequenced. For heterologous expression in E. coli, chemically competent E. coli HM0079 sfp
bearing a chromosomally installed phosphopantethein transferase (Gruenewald et al., 2004) was transformed with the respective
expression vectors. For intergenic conjugation, chemically competent E. coli ET12567 pUZ8002 was transformed with the respective
expression vectors (LB medium supplemented with 15 mg/mL gentamicin, 50 mg/mL kanamycin, and 50 mg/mL chloramphenicol).
Biparental mating was performed using a standard protocol (Shanks et al., 2006). Briefly, overnight donor (E. coli ET12567
pUZ8002 pMQ72pys/pysPENT) and acceptor (P. protegens Pf-5 or P. entomophila) strain cultures were mixed in a 3:1 ratio (v/v)
ꢀ
and washed with sterile, deionized water. Mating spots (25 mL) were spotted on well-dried LB agar plates and incubated at 28 C
overnight. The mating spots were then suspended in LB medium (200 mL) and plated on LB plates (100 mL, 50 mg/mL gentamicin
and 100 mg/mL ampicillin). Single transformants of P. protegens Pf-5 pMQ72pys and P. protegens Pf-5 pMQ72pysPENT were selected
and used to inoculate LB medium (50 mg/mL gentamicin and 100 mg/mL ampicillin) overnight cultures, of which glycerol stocks were
prepared. The generation of P. entomophila pMQ72pysPENT was performed analogously, however the concentration of gentamicin
was 75 mg/mL.
Cell Chemical Biology 25, 1–7.e1–e9, June 21, 2018 e4

Please cite this article in press as: Klapper et al., Bacterial Alkaloid Biosynthesis: Structural Diversity via a Minimalistic Nonribosomal Peptide Synthe-
tase, Cell Chemical Biology (2018), https://doi.org/10.1016/j.chembiol.2018.02.013
For heterologous expression, transformants were inoculated in LB medium with 100 mM L-arabinose and 15 mg/mL (E. coli) or
0 mg/mL (P. protegens Pf-5, P. entomophila) gentamicin, respectively. E. coli HM0079 sfp and P. protegens Pf-5 were cultivated
5
as negative control. After 1 – 2 days, the cultures were extracted with ethyl acetate. The organic phase was dried over Na
2 4
SO
and the solvent was removed in vacuo. Samples were analyzed via LC-MS using the standard protocol and selected ion monitoring
+
mode for m/z 266 ([M+H] pyreudione A) and m/z 294 ([M+H] pyreudione B) to monitor pyreudione production.
+
Pys C-A Didomain Expression Vector
The genomic sequence encoding P. fluorescens HKI0770 pyreudione synthetase (Pys) is available at the National Center for
Biotechnology Information website (NCBI, WP_064118616). An artificial ORF encoding Pys C-A didomain (979 aa, 106.95 kDa
including the hexa-histidine leader sequence) was amplified from genomic DNA of P. fluorescens HKI0770 using the primers
oDB01 and oDB02 (c.f. Table S1). The amplicon was assembled in the expression vector pET28a(+) (Novagen) using Gibson
assembly kit (NEB). The resulting plasmid, pBK01, harbored a 2937-bp reading frame for the N-terminal hexa-histidine fusion protein
(
979 aa, 106.95 kDa).
Heterologous Protein Production&Purification
Escherichia coli KRX (Promega) was transformed with plasmid pBK01. An overnight pre-culture (5 mL LB, 100 mg/mL kanamycin,
ꢀ
3
7 C, 210 rpm) was grown to saturation and used to inoculate the production culture (400 mL LB, identical conditions). With an optical
density of OD600 = 0.7 reached, gene expression was induced by supplementation of L-rhamnose to a final concentration of 0.1% (m/v)
ꢀ
ꢀ
and carried out for 18 hours at 16 C. Cells were harvested by centrifugation (4 C, 3,200 3 g, 30 min) and suspended in lysis buffer
(
50 mM NaH
2
PO $H
4
2
O, 300 mM NaCl, pH 8.0) amended with 20 mM imidazole. Disruption was carried out with a Sonopuls ultrasonic
ꢀ
sonifier (Bandelin) and was followed by centrifugation (4 C, 17,000 3 g, 20 min) for the removal of cell debris. Protein purification was
performed by nickel affinity chromatography on a Protino nickel-nitrilo acetic resin (Macherey-Nagel) subjecting the sample to a step-
wise imidazole gradient (20-500 mM) in lysis buffer. The purified Pys N-His
GE Healthcare) and eluted with reaction buffer (100 mM Tris, pH 7.5, 25 mM MgCl
tration were verified and determined by SDS-PAGE and Bradford’s assay (Bradford, 1976), respectively (c.f. Figure S1).
6
C-A didomain was desalted on a PD-10 desalting column
(
2
, 2.5 mM EDTA). Protein purification and concen-
3
2
ATP-[ P]pyrophosphate Exchange Assay
3
2
The ATP-[ P]pyrophosphate exchange assay was performed to examine the activity and substrate predilection of the recombinant
C-A didomain of Pys. The reaction mixture was composed of reaction buffer (c.f. Heterologous protein production&purification), 1 mM
3
2
substrate, 100 nM purified enzyme, and 0.1 mM [ P]pyrophosphate (50 Ci/mmol). The substrates were assayed either as pools (pool I:
Gly, L-Ala, L-Val, L-Leu, L-Ile; pool II: L-Cys, L-Met, L-Ser, L-Thr, L-Pro; pool III: L-His, L-Phe, L-Tyr, L-Trp; pool IV: L-Asp, L-Asn, L-Glu,
L-Gln; pool V: L-Lys, L-Arg, L-Orn), or individual pure compounds (D-proline, L/D-pipecolic acid 2, L/D-azetidine 2-carboxylic acid 3,
pyrrole-2-carboxylic acid 5, pyrrole-3-carboxylic acid 4, L-hydroxyproline 6, L-oxoproline 7. 100 mL reactions were carried out in trip-
ꢀ
licates at 20 C. After 30 min incubation the reaction was quenched with 500 mL of an aqueous mixture of 4.6% (w/v) tetrasodium
pyrophosphate and 3.5% (v/v) perchloric acid amended with 1.6% (w/v) activated charcoal. The charcoal was captured in filter paper
and washed twice with quenching solution, following by suspension in 3 mL of scintillation liquid. A PerkinElmer TriCarb 2910TR
scintillation counter was used for the quantification of radionuclide exchange.
LC-MS-Based Metabolomic Profiling
Potential pyreudione E (8) production by Pseudomonas fluorescens HKI0770 was analyzed after being cultured in SM/5 medium for
+
4 h via UHPLC-MS (Shimadzu). Selected ion monitoring was conducted on the calculated compound mass [M+H] 280.2. Potential
2
pyreudione production by Pseudomonas entomophila was analyzed using different media (LB, HL5, SM/5, SIH) to culture the strain.
Liquid cultures (5 mL) were cultured for 1 to 3 d and extracted with 10 mL ethyl acetate. The organic phase was dried over sodium
sulfate and solvents were removed using a speedvac. LC-MS samples were prepared by dissolving the crude bacterial extracts in
2
00 mL methanol and filtration through PTFE filter membranes (0.2 mm, Carl Roth Germany). LC-MS analysis was performed on a
˚
Kinetexꢀ C18 column (50 3 2.1 mm, 1.7 mm, 100 A, Phenomenexꢀ) using a linear gradient of acetonitrile in water with 0.1% (v/v)
formic acid and a flow rate of 0.7 mL/min: 10% ACN for 0.5 min, 10% to 100% ACN over 8 min, 100% ACN for 3 min. The column
was equilibrated prior to each injection with 10% ACN for 2.5 min. LC-MS results were analyzed using the LabSolutions Postrun and
LabSolutions Browser software (v.5.60). Potential pyreudione E (8) production by Pseudomonas fluorescens HKI0770 could be de-
+
tected ([M+H] = 280.2, tret = 6.6 min (identical mass and retention time detected for synthetic pyreudione E), whereas no pyreudione
production could be detected for P. entomophila in any medium tested.
General Synthetic Methods
All chemicals used were reagent grade and used as supplied except where noted (Sigma Aldrich, TCI, Carl Roth, VWR, etc.). Solvents
for chromatography and workup procedures were obtained from VWR (Chromanorm, HPLC grade). Reactions were performed under
an argon atmosphere except where noted. Analytical thin-layer chromatography was performed on E. Merck silica gel 60 F254 plates
(0.25 mm). Compounds were visualized by UV-light at 254 nm and by dipping the plates in ninhydrin solution, cerium sulfate ammo-
nium molybdate (CAM) solution, or permanganate stain. Liquid chromatography was performed using forced flow of the indicated
solvent on Normasil 60 silica gel 40-63 mm (VWR, Dresden).
e5 Cell Chemical Biology 25, 1–7.e1–e9, June 21, 2018

Please cite this article in press as: Klapper et al., Bacterial Alkaloid Biosynthesis: Structural Diversity via a Minimalistic Nonribosomal Peptide Synthe-
tase, Cell Chemical Biology (2018), https://doi.org/10.1016/j.chembiol.2018.02.013
Synthesis of Pyreudione E
(S)-N-Acetylpipecolic Acid 15
(
S)-N-acetylpipecolic acid (15) was prepared according to a slightly modified procedure (Klapper et al., 2016). Briefly, 1 eq.
L-pipecolic acid (2, 0.5 g, 3.9 mmol) and 2.5 eq. acetic anhydride were suspended in tetrahydrofuran (25 mL) and saturated aqueous
sodium hydrogen carbonate solution was added until all starting material dissolved (1 mL). The reaction mixture was stirred at room
temperature for 2 h, after which it was quenched by the addition of 1 M aqueous HCl. The aqueous phase was extracted with CH
and the combined organic extracts were dried over Na SO . The solvent was removed in vacuo and the crude product purified by
silica gel column chromatography (CH Cl with MeOH, 0 to 10% v/v) to yield (S)-N-acetylpipecolic acid (15, 0.39 g, 2.3 mmol, 59%) as
2 2
Cl
2
4
2
2
clear oil. All spectroscopic data were in agreement with reported values (Pizzorno and Albonico, 1977).
Weinreb Amide 16
(
(
S)-N-acetyl-N’-methoxy-N’-methylpiperidine-2-carboxamide (16) was prepared according to a slightly modified procedure
Klapper et al., 2016; Murray et al., 1996). To a solution of (S)-N-acetylpipecolic acid (15, 160 mg, 0.94 mmol, 1.0 eq.) in
ꢀ
CH
2
Cl
2
(10 mL) at 0 C was added 1-ethyl-3-(3-dimethylaminopropyl) carbodiimide hydrochloride (0.54 g, 2.8 mmol, 3 eq.) and
ꢀ
triethylamine (0.39 mL, 2.8 mmol, 3 eq.). The mixture was stirred at 0 C for 15 min after which N,O-dimethylhydroxylamine
ꢀ
hydrochloride (0.27 g, 2.8 mmol, 3.0 eq.) was added. The reaction mixture was stirred for 1 h at 0 C and another 2 h at room
temperature after which it was quenched by the addition of water. The aqueous phase was extracted with CH
bined organic extracts were dried over Na SO . The solvent was removed in vacuo and the crude product purified by silica gel
column chromatography (CH Cl with MeOH 0% to 3% v/v) to yield (S)-N-acetylproline-N’-methoxy-N’-methyl amide (16, 0.15 g,
.69 mmol, 74%) as clear oil. ½aꢂ = ꢁ 7:0 (c = 0.3 in MeOH); H NMR (300 MHz, CDCl
2 2
Cl and the com-
2
4
2
2
2
4
1
0
3
4
3
) d 5.41 (bs, 1H), 3.77 (m, 4H),
.65 (m, 1H), 3.13 (bs, 3H), 2.07 (bs, 3H), 1.95 (m, 1H), 1.34–1.77 (m, 5H); C NMR (75 MHz, CDCl ) d 172.7, 171.4, 61.2,
+H] 215.1390, found 215.1386. NMR spectra are
D
13
3
+
18 2 3
9.0, 44.3, 31.8, 26.1, 25.1, 21.9, 19.3; HRMS (ESI+) calcd for [C10H N O
shown in Data S1.
Synthesis of Pyreudione E 8
Pyreudione E (2-octacyl-indolizidine-1,3-dione, 8) was prepared according to a reported literature procedure, which was slightly
modified (Klapper et al., 2016). First, the indolizidine-1,3-dione core structure (17) was prepared. To a solution of (S)-N-acetyl-N’-
ꢀ
methoxy-N’-methylpiperidine-2-carboxamide (16, 69 mg, 0.32 mmol, 1 eq.) in anhydrous THF (5 mL) at –78 C was added drop-
wise 1.3 M lithiumhexamethyl disilazide THF solution (0.5 mL, 0.64 mmol, 2 eq.) over a period of 45 min. The reaction mixture was
ꢀ
stirred for 3.5 h after which it was left to warm to –30 C. After another 30 min, the reaction mixture was quenched by the addition
2 2 2 4
of 2 M HCl and extracted with CH Cl . The combined organic layers were dried over Na SO and concentrated in vacuo to provide
a crude solution of (S)-indolizidine-1,3-dione (17). Due to thermal instability, the solution was directly used for the subsequent acyl-
ation step (Klapper et al., 2016; Jeong et al., 2014). To a solution of n-octanoic acid (56 mL, 0.35 mmol, 1.1 eq.) in anhydrous di-
chloromethane (5 mL) were added EDCI (68 mg, 0.35 mmol, 1.1 eq.), DMAP (47.6 mg, 0.39 mmol, 1.2 eq.), and after 10 min crude
(
ature after which it was left for 15 h at 2 C (to complete acyl migration). The reaction mixture was quenched with saturated
S)-indolizidine-1,3-dione (17, 1.0 eq.) solution (approximately 10 mL). The reaction mixture was stirred 2 hours at room temper-
ꢀ
aqueous ammonium chloride solution. The aqueous phase was extracted with CH
dried over Na SO . The solvent was removed in vacuo and the crude product purified by silica gel column chromatography
CH Cl with methanol 0 to 5% v/v) and semi-preparative HPLC on a Lunaꢀ C18(2) column (250 3 10 mm, 5 mm, 100 A, Phenom-
enex) using the following elution gradient: solvent A: H O + 0.1% HCOOH, solvent B: acetonitrile + 0.1% HCOOH, gradient: 75% B
for 14 min, 75% to 100% B in 0.5 min, 100% B for 3 min, flow rate = 5 mL/min to obtain 22.1 mg pyreudione E (8) as red oil. t
2 2
Cl and the combined organic extracts were
2
4
˚
(
2
2
2
R
=
2
D
5
1
1
3
1
2.8 – 13.7 min; 25% yield; ½aꢂ = ꢁ 1:2 (c = 0.3 in MeOH); Z:E = 6:1; H NMR (500 MHz, CDCl
3
) d Z-form 4.29 (dd, 13.4, 4.8, 1H),
.57 (dd, 12.1, 4.1, 1H), 2.82 (m, 2H), 2.78 (m, 1H), 2.16 (m, 1H), 1.99 (m, 1H), 1.77 (m, 1H), 1.66 (p, 7.6, 2H), 1.51 (m, 1H), 1.18–
.45 (m, 10 H), 0.88 (t, 7.1, 3H); E-form 4.37 (dd, 13.4, 5.1, 1H), 3.73 (dd, 11.9, 4.1, 1H), 2.93 (m, 2H), 2.73 (dt, 13.0, 3.6, 1H), 2.16
13
(
m, 1H), 1.99 (m, 1H), 1.77 (m, 1H), 1.66 (m, 2H), 1.51 (m, 1H), 1.18–1.45 (m, 10 H), 0.88 (t, 7,1, 3H); C NMR (125 MHz, CDCl
3
)
d Z-form 194.4, 187.9, 171.0, 101.0, 63.3, 38.5, 32.6, 31.6, 29.2, 28.9, 27.7, 26.0, 25.0, 23.3, 22.6, 14.0; E-form 200.6, 191.3,
+
1
64.4, 104.2, 60.3, 38.3, 32.9, 31.6, 29.3, 28.9, 27.8, 26.0, 25.0, 23.4, 22.6, 14.0; HRMS (ESI+) calcd for [C16
80.1907, found 280.1905. NMR spectra are shown in Data S1.
3
H25NO +H]
2
Feeding Experiments
Stock solutions (100 mg/mL) of amino acid substrates (L-pipecolic acid 2, L/D-azetidine 2-carboxylic acid 3, L-hydroxyproline 6, and L-
oxoproline 7) as well as D-pipecolic acid (2) in water were sterile filtered and added to SM/5 medium for 0.1 g/L, 0.5 g/L, 1.0 g/L and
3
.0 g/L final concentration of the respective amino acid. Overnight culture of Pseudomonas fluorescens HKI0770 was used for inoc-
ꢀ
ulation. After 24 h of growth at 22 C, 180 rpm, the cultures were extracted and analyzed as previously described (c.f. LC-MS-
based metabolomic profiling). While no increase in pyreudione E (8, or derivatives thereof) production was observed upon addition
of D-pipecolic acid, exogenous supply of 0.5 g/L L-pipecolic (2) acid increased production of pyreudione E (8) and analogues (9 – 11)
strongly, enabling their isolation. Addition of L/D-azetidine 2-carboxylic acid (3) did not led to the production of respective analogues.
Exogenous supply of either L-hydroxyproline (6) or L-oxoproline (7) led in both cases to the production of pyreudione I (12) and
analogues (13, 14), with 3 g/L L-hydroxyproline (6) in sufficient amounts for their isolation.
Cell Chemical Biology 25, 1–7.e1–e9, June 21, 2018 e6

Please cite this article in press as: Klapper et al., Bacterial Alkaloid Biosynthesis: Structural Diversity via a Minimalistic Nonribosomal Peptide Synthe-
tase, Cell Chemical Biology (2018), https://doi.org/10.1016/j.chembiol.2018.02.013
Isolation of Pyreudiones E – H, 8 – 11
P. fluorescens HKI0770 Dclp was cultured in 3 L SM/5 medium with 0.5 g/L L-pipecolic acid (2). After 24 h, the culture supernatant was
extracted with an equal volume of ethyl acetate and the two phases were separated. The organic phase was dried over Na
2 4
SO and
the solvent was removed in vacuo. The crude extract was fractionated by reverse-phase column chromatography using a
TM
HyperSep
1g C18 column (ThermoFisher). The C18-column was washed prior use with 50 mL acetonitrile and equilibrated
with 50 mL 15% acetonitrile (v/v) in water. The crude extract was dissolved in 0.5 mL MeOH and 2.85 mL water was added (final
concentration of MeOH in water = 15% v/v). The solution was applied onto the conditioned C18 column. After loading, a stepwise
acetonitrile-water gradient was applied to the column (15, 30, 50, 75, 100% v/v acetonitrile in water) and each fraction was collected
separately. Solvents were removed using a speedvac system.
UHPLC-MS (Shimadzu System LC-30AD, SPD-M30A, LCMS-2020) analysis showed that the 75% fraction contained the highest
amounts of indolizidine dione derivatives and was further purified using a semi-preparative HPLC (Shimadzu) equipped with a Lunaꢀ
˚
C18(2) column (250 3 10 mm, 5 mm, 100 A, Phenomenex), flow rate = 5 mL/min.
The 75% acetonitrile/water fraction (45 mg) was dissolved in MeOH and purified using the following elution gradient: solvent A:
H
2
O + 0.1% HCOOH, solvent B: acetonitrile + 0.1% HCOOH, gradient: 80% B for 19 min, 80% to 100% B in 0.5 min, 100% B
for 3 min, to yield 21.3 mg pyreudione E (8, t = 10.0 min), 3.1 mg pyreudione F (9, t = 18.1 min), 2.5 mg pyreudione G
10, t = 16.7 min), and 1.7 mg pyreudione H (11, t = 21.8min). Pyreudione G had to be repurified due to minor impurities with
pyreudione D (Klapper et al., 2016). Pyreudione G was dissolved in MeOH and repurified using the following elution gradient:
solvent A: H O + 0.1% HCOOH, solvent B: acetonitrile + 0.1% HCOOH, gradient: 62% B for 58 min, 62% to 100% B in 0.1 min,
00% B for 3 min, to yield 0.7 mg pyreudione G (10, t = 57.1 min).
R
R
(
R
R
2
1
R
Isolation of Pyreudiones I – K, 12 – 14
P. fluorescens HKI0770 Dclp was cultured in 20 L SM/5 medium with 3 g/L L-4-hydroxyproline (6). A 200 mL pre-culture in the same
medium was used for inoculation. After 20 h, the culture supernatant was extracted with an equal volume of ethyl acetate and the two
2 4
phases were separated. The organic phase was dried over Na SO and the solvent was removed in vacuo. The crude extract was
fractionated by reverse-phase column chromatography using 15 g octadecyl-functionalized silica gel. The C18-column was washed
prior use with 50 mL acetonitrile and equilibrated with 50 mL 15% acetonitrile (v/v) in water. The crude extract was dissolved in 3 mL
MeOH and 16.8 mL water was added (final concentration of MeOH in water = 15% v/v). The solution was applied onto the conditioned
C18 column. After loading, a stepwise acetonitrile-water gradient was applied to the column (15, 30, 50, 75, 100% v/v acetonitrile in
water) and each fraction was collected separately. Solvents were removed using a speedvac system.
UHPLC-MS (Shimadzu System LC-30AD, SPD-M30A, LCMS-2020) analysis showed that the 50% fraction contained the highest
amounts of hydroxylated pyrrolizidine dione derivatives and was further purified using a semi-preparative HPLC (Shimadzu) equip-
˚
ped with a Lunaꢀ C18(2) column (250 3 10 mm, 5 mm, 100 A, Phenomenex), flow rate = 5 mL/min.
The 50% acetonitrile/water fraction (353 mg) was dissolved in MeOH and purified using the following elution gradient: solvent A:
H
2
O + 0.1% HCOOH, solvent B: acetonitrile + 0.1% HCOOH, gradient: 50% B for 14 min, 50% to 60% B in 0.5 min, 60% B for 5.5 min,
0% to 100% B in 0.5 min, 100% B for 5 min. Pyreudione I (tret = 10 – 12 min) was repurified using the following elution gradient: 48%
B for 19 min, 48% to 100% B in 0.5 min, 100% B for 5 min to yield 9.6 mg pyreudione I (12, t = 16.1 min). Pyreudione J and K (tret
9.2 – 19.8 min) were repurified using the following elution gradient: 57.5% B for 19 min, 57.5% to 100% B in 0.5 min, 100% B for
.5 min to yield 0.8 mg pyreudione J (13, t = 19.1 min) and 1.3 mg pyreudione K (14, t = 14.2 min).
6
R
=
1
3
R
R
Analytical Data of Pyreudiones E – K, 8 – 14
The diastereomeric ratios were obtained from the integrals of proton 8a for pyreudione E analogues and 7a for pyreudione I
1
analogues in the corresponding H NMR spectra, since both diastereomers showed a well-separated peak for this single proton.
For pyreudione E, the diastereomeric ratios match those of the respective synthetic compound (c.f. Synthesis of pyreudione E).
+
+H] 280.1907, found 280.1906.
2
D
5
pyreudione E (8): Z: E = 6:1, ½aꢂ = ꢁ 3:4 (c = 0.3 in MeOH); HRMS (ESI+) calcd for [C16
H
25NO
3
NMR signals for both diastereomers are listed in Table S4. NMR spectra are shown in Data S1.
25
+
+H] 308.2220, found 308.2221.
pyreudione F (9): Z: E = 6:1, ½aꢂ = ꢁ 2:8 (c = 0.3 in MeOH); HRMS (ESI+) calcd for [C18
H
29NO
3
D
NMR signals for both diastereomers are listed in Table S4. NMR spectra are shown in Data S1.
2
D
5
+
+H] 306.2064, found 306.2069.
pyreudione G (10): Z:E = 9:2, ½aꢂ = + 0:3 (c = 0.2 in MeOH); HRMS (ESI+) calcd for [C18
H
The coupling constant between proton H2’ and H3’ (J = 15.7/15.9 Hz) suggests an E-configuration of the corresponding double bond
27NO
3
(for comparison of NMR data of a related conjugated tetramic acid including a crystal structure see: Gomi et al., 1994). NMR signals
for both diastereomers are listed in Table S4. NMR spectra are shown in Data S1.
2
D
5
+
pyreudione H (11): Z:E = 7:1, ½aꢂ = + 3:5 (c = 0.2 in MeOH); HRMS (ESI+) calcd for [C20
3
H31NO +H] 334.2377, found 334.2383.
The NOESY correlation and coupling constant between proton H5’ and H6’ (J = 10.8 Hz / 10.9 Hz) suggest a Z-configuration of the
corresponding double bond. NMR signals for both diastereomers are listed in Table S4. NMR spectra are shown in Data S1.
2
D
5
+
+H] 282.1700, found 282.1700.
pyreudione I (12): Z:E = 8:3, ½aꢂ = ꢁ 25:3 (c = 0.2 in MeOH); HRMS (ESI+) calcd for [C15
H23NO
4
NMR signals for both diastereomers are listed in Table S4. NMR spectra are shown in Data S1.
25
+
+H] 308.1856, found 308.1855.
pyreudione J (13): Z:E = 3:1, ½aꢂ = ꢁ 64:3 (c = 0.2 in MeOH); HRMS (ESI+) calcd for [C17
H
25NO
4
D
The coupling constant between proton H2’ and H3’ (J = 15.6 Hz/15.8 Hz) suggests an E-configuration of the corresponding double
e7 Cell Chemical Biology 25, 1–7.e1–e9, June 21, 2018

Please cite this article in press as: Klapper et al., Bacterial Alkaloid Biosynthesis: Structural Diversity via a Minimalistic Nonribosomal Peptide Synthe-
tase, Cell Chemical Biology (2018), https://doi.org/10.1016/j.chembiol.2018.02.013
bond (for comparison of NMR data of a related conjugated tetramic acid including a crystal structure see: Gomi et al., 1994). NMR
signals for both diastereomers are listed in Table S4. NMR spectra are shown in Data S1.
2
D
5
+
+H] 308.1856, found 308.1858.
pyreudione K (14): Z:E = 3:1, ½aꢂ = ꢁ 32:5 (c = 0.2 in MeOH); HRMS (ESI+) calcd for [C17
H25NO
4
1
The NOESY correlation and coupling constant between proton H3’ and H4’ (J = 10.6 Hz, see 2D homonuclear J-resolved H) suggest
a Z-configuration of the corresponding double bond. NMR signals for both diastereomers are listed in Table S4. NMR spectra are
shown in Data S1.
Growth Inhibition Assay
3000 D. discoideum cells (AX2) were cultured in 96-well plates (Sarstedt) in 200 mL HL5 medium containing 1% DMSO (Carl Roth) and
ꢀ
a specified amount of compound (starting from 100 mg/mL down to 0.1 mg/mL; 2-fold serial dilution) in triplicates at 22 C. The positive
growth control contained only DMSO (1%). After 72 h the cell concentration was determined according to the manufacturers
ꢀ
instruction using a CASY Cell Counter + Analyser System (Model TT, Roche Innovatis AG) equipped with a 60 mm capillary and
the evaluation cursor set to 7.5 – 17.5 mm. The viable cell concentration was plotted against the logarithmic concentration of the
compound and the IC50 value was determined using PRISM (GraphPad, Version 5.03). The assay was repeated once and the given
value is the average of both experiments.
-1
The amoebicidal activity of pyreudione E was determined with IC50 = 5.7 ± 1.4 mg mL , whereas no activity (IC50 > 100 mg/mL) was
found for pyreudione I.
Antimicrobial and MIC Tests
Antimicrobial activity was evaluated against B. subtilis 6633; S. aureus SG511, E. coli SG458, P. aeruginosa K799/61, M. vaccae
1
0670, S. salmonicolor 549, C. albicans C.A., and P. notatum JP36 by measuring the inhibition zone in mm (disc diffusion assay)
according to the NCCLS (National Committee for Clinical Laboratory Standards, Clinical and Laboratory Standards Institute,
006) and subsequent MIC testing was performed according to Wiegand et al. (2008) against M. vaccae 10670 and M. aurum SB66.
2
Protein BLAST
A search for homologs of the pyreudione synthetase Pys was performed using NCBI protein BLAST based on the full amino acid
sequence of Pys (accession number WP_064118616, https://www.ncbi.nlm.nih.gov/). The result of this search is shown in Table
S2. Only fully sequenced bacterial strains were considered. The strain available from DSMZ with the most similar pys homologue
(
P. entomophila) was used for further studies (heterologous expression, expression analyses).
Expression Analyses of pys and pysPENT
For gene expression analyses, P. fluorescens HKI0770 and P. entomophila were cultured in SM/5 medium due to P. fluorescens’
large production levels of pyreudiones in SM/5 medium. Bacterial overnight cultures of P. fluorescens HKI0770 and
P. entomophila were inoculated in 5 mL SM/5 medium to OD600 = 0.1 and grown to OD600 = 1.0 (highest pys RNA expression levels
were observed at OD600 = 1.0). RNA isolation was performed as described in the RNeasyꢀ Mini Kit (Qiagen) manual using
8
RNAprotectꢀ Bacteria Reagent and 2 mL exponentially grown culture (OD600 = 1.0, ꢃ2x10 cells). Residual DNA was removed
TM
with TURBO
DNase (Thermo Fisher Scientific) and RNA was reverse-transcribed into cDNA using RevertAid Reverse
ꢀ
ꢀ
ꢀ
Transcriptase (Thermo Fisher Scientific, 25 C 5 min, 42 C 60 min, 70 C 5 min) and Random Hexamers (Thermo Fisher Scientific).
A negative control of RNA from the DNAse digestion was kept to test for complete removal of residual gDNA via qPCR. cDNA
was purified by ethanol precipitation and resuspension, 50 ng of which were used for quantitative RT-PCR with primers amplifying
rpoDHKI0770/pysHKI0770 or rpoDPENT/pysPENT (c.f. Table S1, primers were designed using PCRTiler (Gervais et al., 2010), 400 nM final
primer concentration, except for HKI0770rpoD_qPCR4_R – 800 nM, primer concentrations were optimized for equal amplification
ꢀ
ꢀ
ꢀ
effiency) and PowerUpTM SYBRꢀ Green Master Mix (Applied Biosystems, 95 C 2 min, then 40 times 95 C 3 sec and 62 C 30
sec, melting curve: 95 C 15 sec, 60 C 1 min, then stepwise +0.3 C to 95 C) on a StepOnePlus Real-Time PCR System (Applied
ꢀ
ꢀ
ꢀ
ꢀ
Biosystems).
The fold change in expression was calculated using the 2
-DDCt
method (Livak and Schmittgen, 2001), comparing the gene
expression of pys in P. fluorescens HKI0770 with pysPENT in P. entomophila in three independent biological replicates. The respective
rpoD gene was used for internal normalization. For RT-qPCR, three technical replicates were conducted for each biological replicate.
The relative expression of pysPENT in P. entomophila was set to 1.
Phylogenetic Analyses of TE Domains
Protein sequences were retrieved from public databases (for source organisms and accession numbers c.f. Table S3) and aligned
using the MUSCLE algorithm (Edgar, 2004). In case of multi-domain proteins, only the TE domain was analyzed. Maximum likelihood
trees were computed using Mega6 (Tamura et al., 2013) as follows: The best evolutionary model was determined to be the LG model
(
Le and Gascuel, 2008) with a gamma distribution of rates (G), invariable sites (LG + G + I + F). Therefore, this model was chosen to
compute the maximum likelihood tree. A discrete Gamma distribution was used to model evolutionary rate differences among sites
5 categories (+G, parameter = 4.1228)). The rate variation model allowed for some sites to be evolutionarily invariable ([+I], 2.1950%
(
sites). Gaps were treated as complete deletions leaving a total of 159 positions in the final dataset. The percentage of trees in which
the associated taxa clustered together is shown next to the branches (bootstrap values). 100 bootstrap replicates were run to assess
Cell Chemical Biology 25, 1–7.e1–e9, June 21, 2018 e8

Please cite this article in press as: Klapper et al., Bacterial Alkaloid Biosynthesis: Structural Diversity via a Minimalistic Nonribosomal Peptide Synthe-
tase, Cell Chemical Biology (2018), https://doi.org/10.1016/j.chembiol.2018.02.013
consistency of branches. The TE domain of the erythromycin synthase (EryA3) was used as outgroup to root the tree. The tree is
drawn to scale, with branch lengths measured in the number of substitutions per site. The tree with the highest log likelihood
(-7392.1610) is shown (Figure 4). Construction of phylogenetic networks was performed with SplitsTree 4.14 (Huson, 1998) using
the neighbor-net algorithm (Figure S2).
QUANTIFICATION AND STATISTICAL ANALYSIS
For growth inhibition assays, the viable cell concentration was plotted against the logarithmic concentration of the compound
and the IC50 value was determined using PRISM (GraphPad, Version 5.03). The assay was repeated once and the given value is
the average of both experiments.
e9 Cell Chemical Biology 25, 1–7.e1–e9, June 21, 2018