science.abq7333

Download as pdf or txt
Download as pdf or txt
You are on page 1of 10

RES EARCH

◥ BGC (Act) in Streptomyces coelicolor M145.


RESEARCH ARTICLE SUMMARY After conjugations, we obtained pCap-Act and
observed production enhancement of actino-
NATURAL PRODUCTS rhodin in the Act-mobilized mutant, demon-
strating ACTIMOT’s feasibility.
Autologous DNA mobilization and multiplication Expanding our approach, we applied ACTIMOT
to the 48-kb target DNA region Sav11 from
expedite natural products discovery from bacteria S. avidinii DSM40526, containing two unknown
nonribosomal peptide synthetase (NRPS) BGCs.
Feng Xie†, Haowen Zhao†, Jiaqi Liu, Xiaoli Yang, Markus Neuber, Amay Ajaykumar Agrawal, Using the improved single-plasmid ACTIMOT
Amninder Kaur, Jennifer Herrmann, Olga V. Kalinina, Xiaoyi Wei, Rolf Müller*, Chengzhang Fu* devised by merging pRel and pCap, we enhanced
efficiency and indeed identified two previously
unknown peptide families: avidistatins (1 to 7)
INTRODUCTION: The escalation of antimicrobial mulTiplication of BGCs) to mobilize, relocate, and avidilipopeptides (8 to 22). Additionally, we
resistance as a global health threat is driven by and multiply BGCs within bacterial cells. We used applied ACTIMOT to the 67-kb Sar13, housing a
highly effective genetic spreading mechanisms. clustered regularly interspaced short palindromic cryptic “ladderane”-NRPS BGC from S. armeniacus
Antibiotic resistance gene transmission in- repeats (CRISPR) and the CRISPR-associated DSM19369. The corresponding BGC activation

Downloaded from https://www.science.org at Baylor College of Medicine on December 13, 2024


volves a multistep process: mobilization through protein 9 (Cas9) to liberate large target DNA re- led to a 40-fold production increase in mobili-
insertion sequences or integrons, relocation gions and translocated the liberated BGCs onto peptin (23 to 27) production, including their
from chromosomes to mobile genetic elements, a plasmid through homologous recombination, cryptic cyclic peptide precursor, providing in-
and lastly, horizontal gene transfer among mi- enabling its multiplication within the same cell. sights into the biosynthesis of the “ladderane”-
croorganisms. We aimed to artificially mimic This BGC amplification holds immense potential NRPS BGC family.
this mechanism to facilitate the “mobilization- to enhance natural product production, thus Furthermore, applying ACTIMOT to the 149-kb
relocation-transfer” process for large DNA frag- leading to the discovery of unknown compounds. Sav17 from S. avidinii DSM40526 led to the
ments in bacteria, such as biosynthetic gene discovery of benzoxazole-containing actimotins
clusters (BGCs) responsible for producing nat- RESULTS: We designed two sets of plasmids: (29 to 40). Actimotins are produced by a cur-
ural products. the pRel series for BGC mobilization and the rently unpredictable BGC located within Sav17,
pCap series for BGC relocation and multipli- highlighting ACTIMOT’s potential for unravel-
RATIONALE: We developed ACTIMOT (Advanced cation. As a proof of concept, we engineered ing unrecognized pathways. Although an initial
Cas9-mediaTed In vivo MObilization and two plasmids targeting the 24-kb actinorhodin panel of cell-based bioactivity assays did not
detect substantial activity for most compounds,
actimotin J exhibited decent transthyretin-
stabilizing activity in a range similar to that
Target DNA Region of the approved drug tafamidis, suggesting
(TDR) that additional activity screening assays may
uncover bioactivities of compounds discovered
Cas9 Cas9
by ACTIMOT.
Mobilization
Avidilipopeptins CONCLUSION: Using ACTIMOT, we uncovered
Released TDR
four uncharacterized natural compound classes
Capture without the need for altering native BGCs.
plasmid
backbone ACTIMOT effectively mobilizes and multiplies
BGCs, directly augmenting compound yields
Relocation
within native species through the gene dosage
effect, as demonstrated by the discoveries of
Avidistatins mobilipeptins and actimotins. When BGCs are
Plasmid repressed in the native strains, relocated BGCs
with TDR can be transferred to genetically tractable hosts
for production, exemplified by avidistatins and
Multiplication
avidilipopeptins. Future research will explore
ACTIMOT’s adaptability in other species and
Mobilipeptins its potential to unlock the vast genomic poten-
tial within natural product factories. Overall,
this work highlights the promise of ACTIMOT
in accelerating natural product discovery.

The list of author affiliations is available in the full article online.
Actimotins *Corresponding author. Email: rolf.mueller@helmholtz-hips.de
(R.M.); chengzhang.fu@helmholtz-hips.de (C.F.)
†These authors contributed equally to this work.
ACTIMOT facilitates the exploration of cryptic biosynthetic potential in bacteria. ACTIMOT leverages
Cite this article as F. Xie et al., Science 386, eabq7333
CRISPR-Cas9 to efficiently mobilize and relocate the target DNA region harboring BGC(s) onto a multicopy (2024). DOI: 10.1126/science.abq7333
plasmid backbone, enabling BGC(s) multiplication. This process allows product yield enhancement within
native host cells. Using ACTIMOT has led to the discovery of four classes of previously unknown compounds. READ THE FULL ARTICLE AT
[Figure partially created with BioRender.com] https://doi.org/10.1126/science.abq7333

Xie et al., Science 386, 1242 (2024) 13 December 2024 1 of 1


RES EARCH

◥ erate large target DNA regions (TDRs) such


RESEARCH ARTICLE as BGCs from the bacterial chromosome (Fig.
1B). Subsequently, the Cas9-cleaved TDR is cap-
NATURAL PRODUCTS tured by a multicopy plasmid within the cell,
mimicking the relocation process (Fig. 1B). The
Autologous DNA mobilization and multiplication presence of the multicopy replicon enables the
multiplication of the target BGC, resembling
expedite natural products discovery from bacteria another crucial mechanism of AMR occurrence,
namely gene duplication and amplification
Feng Xie1,2,3†, Haowen Zhao1,2,3†, Jiaqi Liu1,2, Xiaoli Yang1,2, Markus Neuber1, (28). ACTIMOT allows for a cloning process
Amay Ajaykumar Agrawal1, Amninder Kaur1,2,3, Jennifer Herrmann1,3, Olga V. Kalinina1,3,4, Xiaoyi Wei5, independent of genomic DNA isolation and
Rolf Müller1,2,3,6,7*, Chengzhang Fu1,2,3* intermediate cloning hosts, which enhances
the efficiency of relocating the BGC into a plas-
The transmission of antibiotic-resistance genes, comprising mobilization and relocation events, mid within the native cell, thereby facilitating
orchestrates the dissemination of antimicrobial resistance. Inspired by this evolutionarily successful the discovery of the final products through the
paradigm, we developed ACTIMOT, a CRISPR-Cas9–based approach to unlock the vast chemical diversity gene amplification effect. In cases where gene
concealed within bacterial genomes. ACTIMOT enables the efficient mobilization and relocation of dosage does not increase production, the mo-

Downloaded from https://www.science.org at Baylor College of Medicine on December 13, 2024


large DNA fragments from the chromosome to replicative plasmids within the same bacterial cell. bilized BGCs are transferred into heterologous
ACTIMOT circumvents the limitations of traditional molecular cloning methods involving handling and hosts, thus circumventing negative regulation
replicating large pieces of genomic DNA. Using ACTIMOT, we mobilized and activated four cryptic in the native host. In this study, we present
biosynthetic gene clusters from Streptomyces, leading to the discovery of 39 compounds across four the development and successful application
distinct classes. This work highlights the potential of ACTIMOT for accelerating the exploration of of ACTIMOT, directly enabling the discov-
biosynthetic pathways and the discovery of natural products. ery of four classes of previously unknown NPs
without additional pathway engineering.

T
he antimicrobial resistance (AMR) crisis ments, a fundamental requirement for nu- The principle and design of ACTIMOT
has been identified by the World Health merous potential applications. For instance, In contrast to conventional molecular cloning,
Organization (WHO) as one of the great- biosynthetic gene clusters (BGCs)—composed ACTIMOT eliminates the need for isolating ge-
est global threats (1). Although AMR un- of multiple genes (varying from several kb to nomic DNA from the source species and rep-
doubtedly poses a major health threat, >100 kb) forming a pathway to produce small licating recombinant DNA in the host species.
its inherent evolutionary mechanism in gene molecules and related congeners—hold im- Instead, ACTIMOT enables the leap of TDRs
transmission appears to be a successful strat- mense value for chemical entities exhibiting from the bacterial chromosome to plasmids
egy in nature. Recent studies have indicated extremely diverse biological activities. Such within the same bacterial cell. We developed
that the spread of antibiotic-resistance genes natural products (NPs) have contributed to the prototype of ACTIMOT as comprising two
(ARGs) occurs through the transfer of AMR 66.8% of all US Food and Drug Administra- distinct Streptomyces–Escherichia coli shuttle
genes from the chromosome to mobile genetic tion (FDA)–approved small-molecule drugs in plasmids, facilitating TDR mobilization and
elements (MGEs) such as multicopy plasmids the last four decades, making them vital for relocation. The release plasmid (pRel) used in
(2, 3). This process generally includes two suc- the treatment of various diseases (9). How- the TDR mobilization step consists of a codon-
cessive events: the mobilization of ARGs typ- ever, technical challenges and high rediscovery optimized Cas9 gene, a single-guide RNA (sgRNA)
ically achieved by either insertion sequences rates have led to a decline in interest from the cassette, and a Streptomyces replicon [from pSG5
(4, 5) or integrons (6, 7) and the subsequent pharmaceutical industry in NP research (10). (29)]. The capture plasmid (pCap) for BGC relo-
relocation of the mobilized ARGs to MGEs (8). Nevertheless, recent genomic investigations cation features a bacterial artificial chromosome
Subsequently, horizontal gene transfer (HGT) have unveiled the heavily underestimated po- (BAC) backbone and a Streptomyces-specific rep-
facilitates the transmission of ARGs among tential of untapped secondary metabolites in licon [either from pSG5 or pIJ101 (30)] (Fig. 2A,
bacteria (2, 3) (Fig. 1A). bacteria, even in extensively studied taxa such figs. S1 and S2, and table S2). The option of the
Inspired by the “mobilization-relocation- as Actinobacteria (11). Although prior efforts in pIJ101 replicon in pCap ensures compatibility
transfer” process naturally occurring during the field developed different strategies to un- with pRel when used together in Streptomyces.
AMR spread, we contemplated the possibility cover NPs from BGCs (12–21), the discon- To mobilize a specific TDR from the bacterial
of artificially simulating the molecular mecha- nection between the enormous biosynthetic chromosome, a functional pRel carrying two
nisms to mobilize and transmit large DNA frag- potential of BGCs and our limited knowledge sgRNA spacers targeting the two ends of TDR is
of encoded chemical entities highlights the required to ensure accurate cleavage on the chro-
1
need for alternative approaches to accelerate mosome (Fig. 1B). Simultaneously, a functional
Helmholtz Institute for Pharmaceutical Research Saarland
(HIPS), Helmholtz Centre for Infection Research (HZI),
the discovery of what are now known as “cryptic” pCap harboring a specific spacer–protospacer
Saarbrücken, Germany. 2Helmholtz International Lab for Anti- NPs (22). adjacent motif (PAM) cassette between the
Infectives, Helmholtz Centre for Infection Research, To accomplish this objective, we have de- pair of homologous arms (HAs) is cleaved by
Braunschweig, Germany. 3PharmaScienceHub, Saarbrücken,
Germany. 4Faculty of Medicine, Saarland University,
vised an approach inspired by ARG transmis- sgRNA-Cas9, exposing the HAs, CapL, and CapR.
Homburg, Germany. 5Key Laboratory of South China sion called ACTIMOT (Advanced Cas9-mediaTed The linearized pCap, the destination for BGC
Agricultural Plant Molecular Analysis and Genetic In vivo MObilization and mulTiplication of relocation, captures the mobilized TDR frag-
Improvement, South China Botanical Garden, Chinese
BGCs). To simulate the mechanisms of gene ment through homologous recombination.
Academy of Sciences, Guangzhou, China. 6Department of
Pharmacy, Saarland University, Saarbrücken, Germany. mobilization, we have leveraged the power Meanwhile, the Cas9-triggered double-strand
7
German Center for Infection Research (DZIF), of clustered regularly interspaced short pal- break on the chromosome is repaired through
Braunschweig, Germany. indromic repeats (CRISPR) and the CRISPR- homology-directed repair by HAs (DelL and DelR)
*Corresponding author. Email: rolf.mueller@helmholtz-hips.de
(R.M.); chengzhang.fu@helmholtz-hips.de (C.F.) associated protein 9 (Cas9) (23–27), a cutting-edge cloned in the functional pRel. The relocated BGC
†These authors contributed equally to this work. genome-editing technology, to artificially lib- on the multicopy plasmid undergoes replication

Xie et al., Science 386, eabq7333 (2024) 13 December 2024 1 of 9


RES EARCH | R E S E A R C H A R T I C L E

panel, and fig. S6), indicating an enhanced


production of the pigment actinorhodins in
these mutants, likely due to the multiplication
of the corresponding BGC. Moreover, the evi-
dent actinorhodin color observed in S. lividans
DYA10 mutants expressing pCap-Act provides
further support for the robustness and stability
of ACTIMOT (Fig. 2B, lower panel). This proof-
of-concept study demonstrates the feasibility of
ACTIMOT and its potential to uncover hidden
NPs from actinomycetes.

Discovery of NPs through improved


single-plasmid ACTIMOT
Despite decades of exploration, the genus
Streptomyces remains a prolific source of un-
known BGCs (11, 12, 15, 32). Conventional ap-

Downloaded from https://www.science.org at Baylor College of Medicine on December 13, 2024


proaches, including activity-based random
screening, have become less effective in dis-
covering previously unidentified compounds.
To address this challenge and explore the poten-
tial of ACTIMOT for uncovering uncharacterized
NPs, we selected two type strains with reported
compounds previously discovered by traditional
methods, namely S. avidinii DSM40526 (33)
and S. armeniacus DSM19369 (34, 35).
The initial successful activation of the 24-kb
Act prompted us to gradually increase the size
of the TDRs in subsequent studies. The 48-kb
TDR Sav11 in the chromosome of S. avidinii
DSM40526 appeared to be a promising candi-
date. It encompasses two adjacent unknown
nonribosomal peptide synthetase (NRPS) BGCs,
avl and avs (Fig. 3A, fig. S7, and table S5). The
Fig. 1. The principle and design of ACTIMOT. (A) A representative spread model of antimicrobial organization of the NRPS genes in a tail-to-tail
resistance genes (ARGs, depicted in blue): (i) ARGs located on the bacterial chromosome are mobilized manner and the presence of modular thioes-
by either insertion sequences or integrons; (ii) the mobilized ARGs are transferred to mobile genetic elements terase domains indicate two separate BGCs.
(MGEs), such as plasmids; and (iii) the spread of ARGs is achieved through horizontal gene transfer. BGC avs comprises two NRPS genes with five
(B) Mobilization, relocation, and multiplication of target DNA regions (TDRs). The simulation of the modules, whereas BGC avl consists of multiple
transmission of ARGs involves the mobilization of TDRs in bacterial chromosomes through CRISPR-Cas9– discrete NRPS and precursor biosynthesis
mediated cuts. The relocation of released TDR into a cleaved multicopy plasmid and the repair of the genes. Further bioinformatic analyses indicated
chromosome are accomplished through homology-directed recombination. The TDR is then multiplied that both BGCs may yield uncharacterized
through the replication of the plasmid. [Figure created with BioRender.com] compounds despite their broad distribution in
different actinomycete strains (table S6 and
data S1 to S4).
dependent on the replicon of the pCap plas- arms through polymerase chain reaction (PCR) To mobilize Sav11, we constructed two func-
mid, potentially leading to the activation or verification (fig. S5, A to C), suggesting the suc- tional plasmids, pCap-101-Apr-SAV11-LR and
enhancement of BGC expression through the cessful relocation of Act from the chromosome pRel-Hyg-SAV11-dsp, and then introduced them
gene dosage. into the pCap plasmid within the cells. into S. avidinii using conjugation. PCR verifi-
To confirm the successful BGC relocation, cation of the exconjugants confirmed the pres-
Proof of concept of ACTIMOT we electroporated three PCR-verified plasmid ence of the correct plasmid carrying the relocated
To validate the concept of ACTIMOT, we se- samples into E. coli DH10B. Subsequent re- Sav11 (pCap-Sav11) (fig. S8A). As expected, upon
lected the BGC of actinorhodins (Act), pro- striction digestion of the extracted plasmids transforming E. coli with plasmids extracted
ducing well-studied pigment compounds (31). revealed the correct presence of pCap-Act, from verified S. avidinii exconjugants, we iden-
Two working plasmids, namely pCap-101-Hyg- harboring the 24-kb Act BGC, in four of the tified two colonies carrying the correct pCap-
ACT-LR and pRel-ACT-dsp, were constructed six randomly picked E. coli colonies (fig. S5, Sav11 from 24 randomly selected transformants
(figs. S3 and S4) and introduced into S. coelicolor D and E). Furthermore, the successful recovery (fig. S8B). We further validated the completeness
M145 through a two-step conjugation process. of pCap-Act depended on the presence of the of the mobilized Sav11 through short-read se-
To examine the occurrence of the Act BGC functional pRel in the M145 strain, highlight- quencing (fig. S9A). After meticulously ex-
mobilization from chromosome to plasmid, we ing the crucial role of CRISPR-Cas9–mediated amining the remaining 22 E. coli colonies,
isolated plasmids from 15 randomly picked cleavage in the mobilization process (fig. S5F). we identified the original empty pCap in every
M145 exconjugants. All samples showed the The M145 exconjugants carrying the pCap-Act colony (fig. S8B). We hypothesized that the
coverage of the two boundaries at the vector plasmid exhibited a darker color compared Cas9 protein expressed by pRel might not com-
and the TDR sequence beyond the homologous with the wild-type (WT) strain (Fig. 2B, upper pletely cleave the independent high-copy pCap

Xie et al., Science 386, eabq7333 (2024) 13 December 2024 2 of 9


RES EARCH | R E S E A R C H A R T I C L E

Downloaded from https://www.science.org at Baylor College of Medicine on December 13, 2024


Fig. 2. The proof of concept of ACTIMOT. (A) Two sets of fundamental tool plasmids for ACTIMOT, pCap, and pRel. Detailed information about plasmid variants
(different combinations of replicons and selection markers) is provided in fig. S1. (B) Production improvement of actinorhodin (ACT) by ACTIMOT in either
native Streptomyces coelicolor M145 strain (up, I and II) or heterologous host (down, III). I (fermentation broth) and II (culture on agar plates) show S. coelicolor
M145 WT (a) and two independent M145 mobilization mutants carrying pCap-Act [(b) and (c)]. III shows S. lividans DYA10 WT (a), DYA10 carrying empty pCap-Hyg-
101-ACT-LR (b), and DYA10 carrying pCap-Act (c). [Figure partially created with BioRender.com]

in vivo, resulting in residual empty pCap in the level production of dozens of compounds with- residue present in 1, as indicated by their
mutants. To further improve the efficiency of out any genetic modifications (Fig. 3, B, D, MS/MS data (fig. S18). In avidistatins 5 to 7,
ACTIMOT, we postulated that a CRISPR-Cas9– and E). We isolated the main compounds and the 2,3-Dab residues were substituted with Dhb
mediated self-cleavage could enhance the ef- structurally elucidated two types of peptides by residues (Fig. 3C and fig. S18), which presum-
ficiency of cutting multicopy plasmids within nuclear magnetic resonance (NMR) and ultra- ably undergo conversion through a C domain–
cells. Therefore, we fused pCap-101-Apr-SAV11-LR high-performance liquid chromatography–high- mediated online dehydration from a threonyl
and pRel-Hyg-SAV11-dsp (materials and methods), resolution mass spectrometry (UHPLC-HRMS) residue (39). This observation suggests the po-
generating the single-plasmid pRelCap-SAV11- (figs. S10 to S13, S50 to S104, and tables S14 to tential promiscuity of the adenylation (A) do-
dsp for the mobilization of Sav11. Similarly to S17). The stereochemistry of amino acid residues main within the second module (fig. S21). We
the dual-plasmid approach, we easily obtained in these compounds was assigned by advanced confirmed the presence of (2S,3R)-Dab using
and verified the correct exconjugants harbor- Marfey’s method (38) (figs. S14 to S17). Addi- the advanced Marfey’s method (fig. S14). The
ing the pCap-Sav11 (fig. S8C). All E. coli colonies tionally, the structures of unisolated derivatives genes avsC, avsD, and avsE are implicated in
were confirmed to carry the correct pCap-Sav11 were elucidated through tandem MS (MS/MS) the formation of (2S,3R)-Dab because they show
after transformation with the plasmids ex- analysis (figs. S18 and S19). similarities to the previously reported (2S,3R)-
tracted from Streptomyces exconjugants (fig. We named the peptide families avidistatins Dab biosynthesis cassette (40). Recent studies
S8D), indicating improved efficiency of the single- (1 to 7) and avidilipopeptins (8 to 22) (Fig. 3, C have corroborated that analogous genes are re-
plasmid ACTIMOT versus the dual-plasmid and F, and fig. S19). Their corresponding BGCs sponsible for the synthesis of (2S,3R)-Dab from
system for BGC mobilization (fig. S8E). were confirmed through gene deletion on the threonine (41, 42), supporting their roles in pro-
These two NRPS BGCs remained silent after pCap-Sav11 (fig. S20). Avidistatin A1 (1) is a lin- viding free substrate available for avidistatin
the successful mobilization in S. avidinii be- ear acetylated pentapeptide composed of unusual assembly. Furthermore, the alignment of the
cause we did not detect prominent new com- nonproteogenic amino acid residues, includ- majority of amino acid building blocks in
pounds in chromatographic analysis in the ing (2S,3R)-2,3-diamino-butyric acid [(2S,3R)- avidistatin A with the A domain–specificity pre-
correct mutants. We hypothesized that the ex- Dab] and (Z)-dehydrobutyrine (Dhb) (Fig. 3C). diction in the avs pathway (table S7) strength-
pression of Sav11 might still be repressed in Avidistatin A2 (2) was identified as an isomer ens the proposed biosynthetic pathway of
the native host (36). Therefore, we transferred of 1, on the basis of their identical molecular avidistatins (fig. S21).
pCap-Sav11 into the heterologous host S. albus formulas and MS/MS data (fig. S18), whereas Avidilipopeptins are linear lipopeptides, in-
Del14 (37) to escape the potential repression avidistatin B1 (3) and B2 (4) were charac- cluding acylated pentapeptides 8 to 12, 17 to
effect in the native strain, which led to the high- terized by the absence of the terminal serine 22, and acylated dipeptides 13 to 16 (Fig. 3F).

Xie et al., Science 386, eabq7333 (2024) 13 December 2024 3 of 9


RES EARCH | R E S E A R C H A R T I C L E

Downloaded from https://www.science.org at Baylor College of Medicine on December 13, 2024


Fig. 3. Discovery of peptides by the improved single-plasmid ACTIMOT. (A) The organization of BGCs avl and avs in TDR Sav11. NRPS, nonribosomal peptide
synthetase. (B) Extracted ion chromatograms (EICs) of avidistatins in crude extracts of S. albus Del14/pCap-Sav11 (i) and S. albus Del14 (ii). The targeted
mass/charge ratio (m/z) values are shown. (C) The structures of the identified avidistatins. (D) EICs of complete-length avidilipopeptins in crude extracts of
S. albus Del14/pCap-Sav11 (i) and S. albus Del14 (ii). The targeted m/z values are shown. (E) EICs of truncated avidilipopeptins in crude extracts of S. albus
Del14/pCap-Sav11 (i) and S. albus Del14 (ii). The targeted m/z values are shown. (F) The structures of the identified avidilipopeptins.

All characterized avidilipopeptins share the first covery of armeniaspirols and streptopyrroles cinnamoyl-containing compounds, which is
two residues (D-arginine and L-serine), suggest- (34, 35), underscoring the untapped biosyn- corroborated by the linear polyene moieties
ing a potential pretermination step after the thetic potential to be explored by using different present in ishigamide (43) and colibrimycins
second NRPS module in the avl pathway (fig. strategies. Among these, the 67-kb TDR Sar13 (44). These BGCs were categorized into various
S22). The diverse structures of avidistatins and emerged as a promising target, containing a groups: 6 mop BGCs, 12 ishigamide BGCs, 4
avidilipopeptins highlight the strategies used by cryptic “ladderane”-NRPS BGC mop (Fig. 4A, colibrimycin BGCs, and 12 other unknown BGCs
NRPS pathways to diversify the final products, fig. S23, and table S8). The “ladderane”-NRPS (table S10, and data S5 and S6), suggesting
including variable building blocks introduced BGCs are known to produce polyene-peptide structural diversity encoded within this family.
by the promiscuous A domains and different hybrid compounds, such as ishigamide (43) and Exploring the products and biosynthesis of mop
chain lengths resulting from the optional colibrimycins (44), or cinnamoyl-containing will deepen our knowledge of the structural
chain release from the assembly line (figs. S21 cyclodepsipeptides, such as kitacinnamycins features concealed within this BGC family.
and S22). (45) and skyllamycins (46, 47). BGC mop shares To mobilize and activate Sar13, we constructed
more similarity with ishigamide BGC and the plasmid pRelCap-SAR13-dsp and introduced
BGC multiplication facilitates the discovery of colibrimycin BGCs than with those produc- it into S. armeniacus. Following the same val-
unusual lipopeptides ing cinnamoyl-containing cyclodepsipeptides. idation procedure described in the Sav11 mo-
Building upon the success of the improved ClusterBlast analysis (48) identified 34 similar bilization, we verified the successful Sar13
single-plasmid ACTIMOT, we expanded its ap- BGCs distributed in various Actinobacteria, mobilization in 10 out of 11 plasmids derived
plication to activate other BGCs from different featuring a set of highly reducing type II poly- from four randomly picked starting Strepto-
Streptomyces strains. Despite many BGCs pres- ketide synthase (PKS), NRPS modules, and 2,3- myces exconjugants (fig. S24), revealing a 90.9%
ent in the genome of S. armeniacus DSM19369, diaminopropionic acid (2,3-Dap) formation success rate for Sar13 mobilization. Short-read
previous screening efforts by the pharma- genes. These BGCs lack genes for benzene ring sequencing of pCap-Sar13 further corroborated
ceutical industry have only resulted in the dis- formation (49), deviating their products from the integrity of the entire plasmid (fig. S9B).

Xie et al., Science 386, eabq7333 (2024) 13 December 2024 4 of 9


RES EARCH | R E S E A R C H A R T I C L E

Fig. 4. The mop BGC multiplica-


tion unveils lipopeptides.
(A) The organization of BGC mop
in TDR Sar13. PKS, polyketide
synthase. (B) Production
enhancement of mobilipeptins
through the mobilization and multi-
plication of Sar13 in S. armeniacus.
Chromatograms at ultraviolet
wavelength of 336 nm are shown:
(i) S. armeniacus/pCap-Sar13
cultured in M2 medium
plus 50 mg/ml apramycin,
(ii) S. armeniacus/pCap-Sar13
cultured in M2 medium
without apramycin, and
(iii) S. armeniacus cultured

Downloaded from https://www.science.org at Baylor College of Medicine on December 13, 2024


in M2 medium. (C) The structures
of mobilipeptin A (23) and
B (24). (D) Detection of
mobilipeptins D (26), E (27), and
A/B (23/24) in S. armeniacus/
pCap-Sar13 (i) and S. albus
Del14/pCap-Sar13-int (ii). EICs
are shown as depicted in the
figure. The data presented
in this panel are from the
sampling with highest yield
of 26 and 27 (day 2). (E) The
structures of mobilipeptins D
(26) and E (27).

The mobilization and multiplication of Sar13 the unusual type II PKS genes (50–52) present mobilipeptin E) at retention times of 6.7 and
promoted the discovery of a series of metab- in the “ladderane” region of Sar13, whereas 7.5 min, respectively (Fig. 4D and fig. S30). These
olites that we named mobilipeptins. The yield the dipeptide is installed by one of the NRPS “transient” compounds exhibited their highest
of mobilipeptins in the mutants carrying mo- genes with the starter condensation (Cs) do- yield on day 2 when cultured in ISP4 medium
bilized Sar13 increased nearly 40-fold (Fig. 4B main (mopI). The substrates predicted for the but faded quickly in the subsequent fermen-
and fig. S25). Discovering mobilipeptins di- two modules in MopI correspond closely with tation (fig. S30). We next purified the tetracyclic
rectly from the WT of S. armeniacus would the observed residues (table S9). However, peptide 27 from a 2-day fermentation of the
have been challenging because of their pro- the sole TE domain in mop is situated after the Sar13 mobilized mutant in ISP4. However, 27
duction in trace amounts that are only barely glycine-recognizing NRPS module in MopJ exhibits extremely poor solubility in various
detectable with selective ion monitoring after (tables S8 and S9), hinting at the potential pres- solvents, including methanol, water, chloroform,
the target mass is identified by using ACTI- ence of precursor compounds with more resi- and dimethyl sulfoxide, making the NMR mea-
MOT (Fig. 4B and fig. S25A). The absence of dues present in the pathway. We hypothesized surement and the ensuing bioactivity test
selection antibiotics did not affect the produc- that this hybrid BGC initially produces precursor challenging. Unexpectedly, we found that 27
tion of mobilipeptins (Fig. 4B and fig. S25B), molecules with a tetrameric cyclic peptide core. exhibited the desired solubility in acetone;
indicating the replication stability of the plas- This hypothesis is supported by the identifica- however, the MS analysis of the solution showed
mid in the cells. This feature shows the potential tion of an additional mobilipeptin congener 25, 27 to convert to a new derivative 28 (fig. S31).
of using ACTIMOT for increasing large-scale featuring an extra glycine residue coupled to the Subsequently, we solved the structure of 28
economic production of high-value NPs. a-amino group of L-2,3-Dap (fig. S29). through HRMS and NMR (figs. S26, S32, S118
Structural elucidation by NMR, UHPLC-HRMS, To explore the possibility of a structurally to S123, and table S19), which suggested that
and Marfey’s method revealed that mobilipeptin different primary product, we conducted fer- 28 is an acetone adduct of 27, which itself
A (23) and B (24) are lipopeptides containing mentation under varying conditions. Numerous most likely exhibits a structure in line with the
a linear conjugated polyene chain coupled with attempts resulted in the detection of short-lived in silico prediction of substrate specificity of
a L-2,3-Dap and an O-methyl-D-tyrosine residue products from the Sar13-activated mutant with all the four A domains in mop (table S9). The
(Fig. 4C; figs. S26 to S28, S106 to S117; and table corresponding molecular ions of [M+H]+ = deduced structure of 27 is a linear polyene chain
S18). The polyene chain is likely derived from 571.29 (26, mobilipeptin D) and 553.28 (27, coupled with tetracyclic peptide comprising

Xie et al., Science 386, eabq7333 (2024) 13 December 2024 5 of 9


RES EARCH | R E S E A R C H A R T I C L E

Downloaded from https://www.science.org at Baylor College of Medicine on December 13, 2024


Fig. 5. Biosynthetic “dark matter” uncovered by ACTIMOT. (A) The scheme of the organization of TDR Sav17. (B) The activation of actimotins through ACTIMOT.
EICs of representative actimotins in crude extracts of S. avidinii/pCap-Sav17 (i), S. avidinii (ii), S. albus Del14/pCap-Sav17 (iii), and S. albus Del14 (iv). The base
peak chromatograms (BPCs) are also shown in traces iii and iv. (C) The structures of the identified actimotins.

two L-2,3-Dap residues, an O-methyl-D-tyrosine mop by ACTIMOT significantly enhanced the S. avidinii mutants carrying pCap-Sav17, we
residue, and a glycine residue (Fig. 4E). More- yield of mobilipeptins at different biosynthesis identified a series of highly yield-improved
over, the structure of 26 was determined by MS stages (Fig. 4D), which exemplified the distinc- compounds that were easy to neglect in the
fragmentation to be a degradation product of tive capability of ACTIMOT. WT strain because of their meager production
27, resulting from the hydrolysis of the amide (Fig. 5B). This observation indicated the activa-
bond between O-methyl-D-tyrosine and the third Biosynthetic “dark matter” revealed by ACTIMOT tion of a cryptic BGC within Sav17. Subsequent-
residue (L-2,3-Dap) (fig. S29), further support- In our pursuit to uncover hidden biosynthetic ly, the heterologous expression of pCap-Sav17
ing the structure assignment of 27. Hence, we potential, we embarked on mobilizing the co- in S. albus Del14 showed a much higher pro-
propose that 27 is the direct product released lossal TDR Sav17 from the chromosome of duction of the compounds found in S. avidinii
and cyclized from the NRPS of mop, which can S. avidinii DSM40526. This expansive segment and multiple new compounds (Fig. 5B). Through
be hydrolyzed into 26, followed by stepwise encompasses a massive 149-kb region predicted comprehensive analysis using NMR and UHPLC-
cleavage into 23 to 25 (fig. S34). to harbor a giant NRPS BGC (Fig. 5A). This HRMS/MS, we purified and structurally elu-
Although direct degradation of 27 remains BGC comprises 28 NRPS modules with a Cs cidated nine compounds (29 to 36, and 38),
a plausible scenario, the identification of spe- domain in the initial module, suggesting the unveiling a family of benzoxazole-containing
cific intermediates with different numbers of potential production of large lipopeptides (fig. compounds that we named actimotins (Fig.
amino acid residues suggests that the tetra- S35). A thorough ClusterBlast search on this 5C; figs. S38 to S40, S124 to S176; and tables
cyclic peptide precursor could also undergo BGC revealed a lack of similar BGCs within the S20 to S23). The stereochemical configurations
stepwise degradation by an as-yet-unknown current public database (data S7 and S8). After of 31 and 35 were determined with Mosher
mechanism (fig. S34). This process could in- the mobilization procedure using the plasmid ester analysis (fig. S39 and tables S24 and S25),
volve rare enzyme-mediated cleavage, similar pRelCap-SAV17-dsp targeting the two ends of and the stereochemistry of 29 and 30 was
to that observed in the biosynthetic pathways Sav17 on the chromosome (fig. S36), we eval- assigned through density-functional theory
of other peptides (53–55). The discovery of the uated four exconjugants for the TDR mobili- simulation (figs. S41 to S45). MS/MS analysis
cyclic mobilipeptin and a series of processing zation efficiency. Subsequently, five out of eight facilitated the identification of the structures
products of mobilipeptins has not only iden- random E. coli colonies obtained by transform- of three additional derivatives (37, 39, and
tified the likely authentic products of the mop ing plasmids from the starting four S. avidinii 40) (fig. S46). One notable structural feature
pathway but also shed light on the biosynthe- exconjugants were found to harbor the plas- of actimotins, alongside their rare metasubsti-
sis and potential degradation mechanisms of mid pCap-Sav17 that contains the correct 149-kb tuted benzoxazole pattern, is the (1S,2R,4S)-2-
mobilipeptins. Furthermore, the mobilipeptin DNA region (fig. S37). Further short-read se- aminocyclohexane-1,4-diol (ACHD) moiety in
biosynthetic pathway implies that the biosyn- quencing of the plasmids confirmed the in- several congeners (29 to 31).
thesis of ishigamide might also involve a cryptic tegrity of the mobilized Sav17, suggesting the We did not find the reported benzoxazole
cyclic peptide precursor. The direct activation of high fidelity of ACTIMOT (fig. S9C). In the BGC in sav17 (56). The only plausible candidate

Xie et al., Science 386, eabq7333 (2024) 13 December 2024 6 of 9


RES EARCH | R E S E A R C H A R T I C L E

is an upstream 20-kb region enriched with var- which contain two 3,4-AHBA units (Fig. 5C). The the most common variant linked to TTR amy-
ious biosynthetic genes, including discrete A do- incorporation rate was low (fig. S48, A and B), loidosis (66). To evaluate a potential similar
main genes, which we initially presumed to which likely results from the use of aspartic acid activity of actimotins, we developed a thioflavin
be part of the giant NRPS BGC. Through gene in primary metabolism and additional catalytic T-based assay to stabilize the TTR-V30M, on
deletions and heterologous expression of the steps required before incorporation into acti- the basis of a previous study (67). In our assays
modified sav17 (Fig. 5A, fig. S47, and table S11), motin (59). A similarly low incorporation rate with 10 mM TTR-V30M, tafamidis, used as posi-
we confirmed that this cryptic 16-gene region is was observed in actimotin A (29) and B (30) tive control, showed activity with a median ef-
responsible for actimotin biosynthesis, distinct (fig. S48, C and D), suggesting that the ACHD fective concentration (EC50) value of 9.8 mM
from the giant NRPS BGC. Thus, ACTIMOT moiety originates from 3,4-AHBA through step- (fig. S50A), which is consistent with the previous
paved the way for discovering the actimotin wise reductions, although alternative path- report (68). Among the isolated actimotins,
biosynthetic pathway, which was not identi- ways cannot be excluded. Further feeding with actimotin J (38) exhibited activity in a similar
fied by in silico prediction owing to the limited 13 15
L-cysteine- C3, N revealed a 4 Da difference range, with an EC50 value of 67.8 mM (fig. S50B).
understanding of the biosynthesis of metasub- between control and isotope-labeled 29 and 30 Actimotin J shares the same structure with a
stituted benzoxazoles. During our study on (fig. S48, E and F), indicating that L-cysteine previously isolated metasubstituted benzoxa-
actimotin biosynthesis, the first BGC encod- serves as the precursor for their N-acetylcysteinyl zole from Nocardiopsis lucentensis DSM44048,
ing for metasubstituted benzoxazoles was pub- moieties. A previous study on grixazole bio- which was also reported to show no cell-based
lished (57). However, genome mining efforts synthesis showed that N-acetylcysteine (NAC) bioactivity (69). Our finding holds promise for

Downloaded from https://www.science.org at Baylor College of Medicine on December 13, 2024


by these authors using the respective BGC did nonenzymatically links to an o-quinone imine discovering further actimotin derivatives ex-
not provide any hint at the actimotin pathway derivative, oxidized from 3-amino-4-hydroxy- hibiting TTR amyloidogenesis inhibition ac-
(57), probably because of low sequence sim- benzaldehyde by a tyrosinase-like enzyme, GriF tivity from the identified amo-like BGCs. The
ilarity (table S12). Even the recently updated (60). Although amo lacks a GriF-like oxidase, observation of the TTR-stabilizing activity of
antiSMASH (v7.0) did not predict the actimotin we hypothesize that NAC incorporation into actimotin J implies that these compounds war-
pathway, further confirming its distinctiveness. the actimotin pathway might be facilitated by rant further evaluation in various assays, partic-
Amo5 shows 38.9% amino acid identity to a different oxidoreductase, leading to the pro- ularly non–cell based assays, to better understand
GriI, and Amo6 exhibits 39.1% identity to GriH duction of 29 and 30 (fig. S49). The structural their potential and biological function.
from the grixazone BGC. GriI and GriH have diversity among actimotins suggests the in-
been shown to be responsible for the biosyn- volvement of various tailoring genes that mod- Conclusions
thesis of 3-amino-4-hydroxybenzoic acid (3,4- ify the core benzoxazole structure, warranting As a technological advancement, ACTIMOT
AHBA) from L-aspartate-4-semialdehyde and further investigation (fig. S49). integrates BGC cloning and activation stages
dihydroxyacetone phosphate (58), and we sug- We subsequently conducted a comprehen- through one-step manipulation–triggered mo-
gest that Amo5 and Amo6 play similar roles in sive survey to explore the distribution of amo- bilization, relocation, and multiplication within
3,4-AHBA formation in actimotin biosynthesis. like BGCs featuring analogous genes to amo5, native bacterial strains. This approach stream-
The BGC amo encodes three phenylacetate- amo6, amo7, and amo4/amo8, which are core lines the process, requiring only a few basic
CoA ligase family proteins—Amo4, Amo8, and benzoxazole biosynthetic genes. Among the cloning steps for small DNA fragments in E. coli.
Amo10—which exhibit low sequence identities 383 hits harboring all the four essential genes It holds the potential to directly enhance com-
to ClxA and ClxC (table S12), implying their uncovered, 364 were identified as full-length pound yields from target BGCs without excessive
involvement in activating and linking residues BGCs and the remaining 19 were fragmented manipulation. ACTIMOT outperforms a cloning-
in the actimotin biosynthetic pathway. The because of sequence quality issues (data S9). independent approach that uses integrase-
amidohydrolase Amo7, which is homologous These 364 BGCs include two actimotin BGCs triggered site-specific recombination in native
to ClxD, likely catalyzes the formation of the and 45 closoxazole BGCs (clx), along with two bacterial cells (70), which is hampered by the
oxazole moiety through heterocyclization. In BGCs analogous to amo and 35 BGCs similar to complexity of three-step conjugations and
actimotin derivatives featuring a third benzoyl clx. The remaining 280 BGCs, each containing the tedious screening required at each step.
residue (29 to 31, 36, and 37), this residue is one to four phenylacetate-CoA ligases, diverge ACTIMOT’s efficiency and simplicity make it
attached to the 3-amino group of the benzoxazole from both known categories. This variety indi- a promising candidate for integration with
scaffold (Fig. 5C). This linkage pattern differs cates a potential structural diversity of amo-like other approaches, such as BGC engineering
from the one observed in closoxazoles (57), sug- BGC products in nature. through promoter refactoring and heterologous
gesting the absence of a ClxC-like ligase in amo. Lastly, in our exploration of the bioactivity expression in suitable hosts. Recent studies
Furthermore, actimotins exhibit additional of the four compound classes obtained in this (71, 72) have introduced intriguing BGC ex-
structural diversity, primarily arising from the work, only actimotin G (35) exhibited weak pression strategies, with a landing pad (LP)–
variation of the second and the third residues activity toward E. coli DtolC, whereas all other mediated bacteria domestication method play-
(Fig. 5C and fig. S49). The promiscuity of cor- compounds showed no appreciable bioactivity ing an important role in host domestication
responding phenylacetate-CoA ligases likely against different microorganisms and a hu- and efficient BGC expression. These LP-based
explains these variations. man cell line (table S13).We noticed that the domestication-expression frameworks pres-
To investigate whether the ACHD moiety actimotins share the benzoxazole structural ent opportunities to enhance the versatility of
shares the same biosynthetic origin with 3,4- feature with tafamidis, a drug used medicinally ACTIMOT. Our research underscores ACTIMOT’s
AHBA, we conducted a feeding experiment as a transthyretin (TTR) stabilizer to treat TTR prowess by unearthing four previously un-
using isotope-labeled L-aspartic acid-13C4,15N amyloidosis-related diseases (61, 62). TTR is a known compound categories without altering
on the mutant carrying pCap-Sav17. Because homotetramic transport protein in human the native BGCs. Particularly noteworthy is the
L-aspartate-4-semialdehyde is derived from plasma responsible for transporting thyroxine unexpected identification of actimotins, whose
L-aspartate (59), this approach allowed us to (T4) and retinol (63). TTR amyloid fibrils are BGC defies prediction by current in silico tools,
trace the incorporation of the labeled precur- associated with various diseases, including hinting at ACTIMOT’s potential for uncovering
sor. Our results indicated a double incorporation familial amyloidotic polyneuropathy (FAP) (64) unpredictable pathways in future investigations.
of the labeled precursor skeleton in actimotin and familial amyloid cardiomyopathy (FAC) Additionally, the identification of the hidden
J (38) and D (32) (fig. S48, A and B), both of (65). The TTR mutant V30M (TTR-V30M) is cyclic mobilipeptin precursor demonstrates the

Xie et al., Science 386, eabq7333 (2024) 13 December 2024 7 of 9


RES EARCH | R E S E A R C H A R T I C L E

power of ACTIMOT in boosting NP discovery. 13. L. Huo et al., Heterologous expression of bacterial natural 35. C. Fu et al., Armeniaspirol antibiotic biosynthesis: Chlorination
The achievements of ACTIMOT in Streptomyces product biosynthetic pathways. Nat. Prod. Rep. 36, 1412–1436 and oxidative dechlorination steps affording spiro4.4non-
(2019). doi: 10.1039/C8NP00091C; pmid: 30620035 8-ene. ChemBioChem 20, 764–769 (2019). doi: 10.1002/
fuel optimism about its potential applicability 14. K. Scherlach, C. Hertweck, Mining and unearthing hidden cbic.201800791; pmid: 30556942
in rare actinomycetes, owing to compatible rep- biosynthetic potential. Nat. Commun. 12, 3864 (2021). 36. B. Aigle, C. Corre, Waking up Streptomyces secondary
licons (73). Although the current application doi: 10.1038/s41467-021-24133-5; pmid: 34162873 metabolism by constitutive expression of activators or genetic
15. A. G. Atanasov, S. B. Zotchev, V. M. Dirsch, C. T. Supuran, disruption of repressors. Methods Enzymol. 517, 343–366 (2012).
range of ACTIMOT is limited and depends on International Natural Product Sciences Taskforce, Natural doi: 10.1016/B978-0-12-404634-4.00017-6; pmid: 23084947
the genetic tractability of target strains, it shows products in drug discovery: Advances and opportunities. 37. M. Myronovskyi et al., Generation of a cluster-free
potential in other genetically manipulable and Nat. Rev. Drug Discov. 20, 200–216 (2021). doi: 10.1038/ Streptomyces albus chassis strains for improved heterologous
s41573-020-00114-z; pmid: 33510482 expression of secondary metabolite clusters. Metab. Eng.
BGC-rich bacterial species, such as Proteobacteria 16. F. Panter, C. D. Bader, R. Müller, Synergizing the potential of 49, 316–324 (2018). doi: 10.1016/j.ymben.2018.09.004;
and Firmicutes (11). To broaden its utility, bacterial genomics and metabolomics to find novel antibiotics. pmid: 30196100
further efforts are needed to develop compati- Chem. Sci. 12, 5994–6010 (2021). doi: 10.1039/D0SC06919A; 38. P. Marfey, Determination of D-amino acids. II. Use of a
pmid: 33995996 bifunctional reagent, 1,5-difluoro-2,4-dinitrobenzene.
ble genetic elements, thereby expanding the Carlsberg Res. Commun. 49, 591–596 (1984). doi: 10.1007/
17. A. Bauermeister, H. Mannochio-Russo, L. V. Costa-Lotufo,
scope of ACTIMOT’s applicability. These future A. K. Jarmusch, P. C. Dorrestein, Mass spectrometry-based BF02908688
applications will allow for a more comprehen- metabolomics in microbiome investigations. Nat. Rev. 39. S. Wang et al., Discovery and biosynthetic investigation of a
Microbiol. 20, 143–160 (2022). doi: 10.1038/s41579-021- new antibacterial dehydrated non-ribosomal tripeptide.
sive assessment of its broader impacts. In sum- Angew. Chem. Int. Ed. 60, 3229–3237 (2021). doi: 10.1002/
00621-9; pmid: 34552265
mary, this study demonstrates the potential 18. M. H. Medema, T. de Rond, B. S. Moore, Mining genomes anie.202012902; pmid: 33107670
of ACTIMOT as a strategy to discover previously to illuminate the specialized chemistry of life. Nat. Rev. 40. C. Müller et al., Sequencing and analysis of the biosynthetic

Downloaded from https://www.science.org at Baylor College of Medicine on December 13, 2024


Genet. 22, 553–571 (2021). doi: 10.1038/s41576-021-00363-7; gene cluster of the lipopeptide antibiotic Friulimicin in
unkown NPs across diverse microorganisms. Actinoplanes friuliensis. Antimicrob. Agents Chemother. 51,
pmid: 34083778
ACTIMOT stands poised to unlock the vast ge- 19. A. F. Rosenzweig, J. Burian, S. F. Brady, Present and future 1028–1037 (2007). doi: 10.1128/AAC.00942-06; pmid: 17220414
nomic potential within these NP factories and outlooks on environmental DNA-based methods for antibiotic 41. R. Li, M. S. Lichstrahl, T. A. Zandi, L. Kahlert, C. A. Townsend,
discovery. Curr. Opin. Microbiol. 75, 102335 (2023). The dabABC operon is a marker of C4-alkylated monobactam
to expand known biochemical space. biosynthesis and responsible for (2S,3R)-diaminobutyrate
doi: 10.1016/j.mib.2023.102335; pmid: 37327680
20. B. C. Covington, F. Xu, M. R. Seyedsayamdost, A natural production. iScience 27, 109202 (2024). doi: 10.1016/
Methods summary product chemist’s guide to unlocking silent biosynthetic j.isci.2024.109202; pmid: 38433893
clusters. Annu. Rev. Biochem. 90, 763–788 (2021). 42. S. Sakata et al., Identification of the cirratiomycin biosynthesis
A detailed description of the methods, includ- gene cluster in Streptomyces cirratus: Elucidation of the
doi: 10.1146/annurev-biochem-081420-102432; pmid: 33848426
ing plasmid design and production, BGC mo- 21. M. Miethke et al., Towards the sustainable discovery and biosynthetic pathways for 2,3-diaminobutyric acid and
bilization, bioinformatics analysis, bacterial development of new antibiotics. Nat. Rev. Chem. 5, 726–749 hydroxymethylserine. Chemistry 30, e202400271 (2024).
doi: 10.1002/chem.202400271; pmid: 38456538
growth conditions, natural product isolation (2021). doi: 10.1038/s41570-021-00313-1
43. D. Du et al., Production of a novel amide-containing polyene by
22. R. H. Baltz, Natural product drug discovery in the genomic era:
and characterization, and biological activity activating a cryptic biosynthetic gene cluster in Streptomyces
Realities, conjectures, misconceptions, and opportunities.
testing, can be found in the supplementary J. Ind. Microbiol. Biotechnol. 46, 281–299 (2019). doi: 10.1007/ sp. MSC090213JE08. ChemBioChem 17, 1464–1471 (2016).
doi: 10.1002/cbic.201600167; pmid: 27311327
materials. s10295-018-2115-4; pmid: 30484124
23. M. Jinek et al., A programmable dual-RNA-guided DNA 44. L. Prado-Alonso et al., Colibrimycins, novel halogenated
endonuclease in adaptive bacterial immunity. Science 337, hybrid polyketide synthase-nonribosomal peptide synthetase
RE FE RENCES AND N OT ES (PKS-NRPS) compounds produced by Streptomyces sp. strain
816–821 (2012). doi: 10.1126/science.1225829; pmid: 22745249
1. World Health Organization, “Antimicrobial resistance,” 24. L. Cong et al., Multiplex genome engineering using CRISPR/ CS147. Appl. Environ. Microbiol. 88, e0183921 (2022).
21 November 2023; https://www.who.int/news-room/fact- Cas systems. Science 339, 819–823 (2013). doi: 10.1126/ doi: 10.1128/AEM.01839-21; pmid: 34669429
sheets/detail/antimicrobial-resistance. science.1231143; pmid: 23287718 45. J. Shi et al., Genome mining and biosynthesis of kitacinnamycins
2. D. G. J. Larsson, C.-F. Flach, Antibiotic resistance in the 25. P. Mali et al., RNA-guided human genome engineering via Cas9. as a STING activator. Chem. Sci. 10, 4839–4846 (2019).
environment. Nat. Rev. Microbiol. 20, 257–269 (2022). Science 339, 823–826 (2013). doi: 10.1126/science.1232033; doi: 10.1039/C9SC00815B; pmid: 31160959
doi: 10.1038/s41579-021-00649-x; pmid: 34737424 pmid: 23287722 46. J. Bracegirdle et al., Skyllamycins D and E, non-ribosomal
3. M. A. Cook, G. D. Wright, The past, present, and future of 26. R. E. Cobb, Y. Wang, H. Zhao, High-efficiency multiplex cyclic depsipeptides from lichen-sourced Streptomyces
antibiotics. Sci. Transl. Med. 14, eabo7793 (2022). genome editing of Streptomyces species using an engineered anulatus. J. Nat. Prod. 84, 2536–2543 (2021). doi: 10.1021/acs.
doi: 10.1126/scitranslmed.abo7793; pmid: 35947678 CRISPR/Cas system. ACS Synth. Biol. 4, 723–728 (2015). jnatprod.1c00547; pmid: 34490774
4. M. Razavi, E. Kristiansson, C.-F. Flach, D. G. J. Larsson, The doi: 10.1021/sb500351f; pmid: 25458909 47. S. Pohle, C. Appelt, M. Roux, H.-P. Fiedler, R. D. Süssmuth,
association between insertion sequences and antibiotic 27. Y. Tong, P. Charusanti, L. Zhang, T. Weber, S. Y. Lee, CRISPR- Biosynthetic gene cluster of the non-ribosomally synthesized
resistance genes. MSphere 5, e00418-20 (2020). doi: 10.1128/ Cas9 based engineering of actinomycetal genomes. ACS Synth. cyclodepsipeptide skyllamycin: Deciphering unprecedented ways
mSphere.00418-20; pmid: 32878926 Biol. 4, 1020–1029 (2015). doi: 10.1021/acssynbio.5b00038; of unusual hydroxylation reactions. J. Am. Chem. Soc. 133,
5. Y. Che et al., Conjugative plasmids interact with insertion pmid: 25806970 6194–6205 (2011). doi: 10.1021/ja108971p; pmid: 21456593
sequences to shape the horizontal transfer of antimicrobial 28. L. Sandegren, D. I. Andersson, Bacterial gene amplification: 48. K. Blin et al., antiSMASH 7.0: New and improved predictions for
resistance genes. Proc. Natl. Acad. Sci. U.S.A. 118, Implications for the evolution of antibiotic resistance. Nat. Rev. detection, regulation, chemical structures and visualisation.
e2008731118 (2021). doi: 10.1073/pnas.2008731118; Microbiol. 7, 578–588 (2009). doi: 10.1038/nrmicro2174; Nucleic Acids Res. 51, W46–W50 (2023). doi: 10.1093/nar/
pmid: 33526659 pmid: 19609259 gkad344; pmid: 37140036
6. M. Gillings et al., The evolution of class 1 integrons and the rise 29. G. Muth, W. Wohlleben, A. Pühler, The minimal replicon of the 49. J. Shi et al., In vitro reconstitution of cinnamoyl moiety
of antibiotic resistance. J. Bacteriol. 190, 5095–5100 (2008). Streptomyces ghanaensis plasmid pSG5 identified by reveals two distinct cyclases for benzene ring formation.
doi: 10.1128/JB.00152-08; pmid: 18487337 subcloning and Tn5 mutagenesis. Mol. Gen. Genet. 211, J. Am. Chem. Soc. 144, 7939–7948 (2022). doi: 10.1021/
7. M. Razavi et al., Discovery of the fourth mobile sulfonamide 424–429 (1988). doi: 10.1007/BF00425695; pmid: 18297777 jacs.2c02855; pmid: 35470672
resistance gene. Microbiome 5, 160 (2017). doi: 10.1186/ 30. Z. X. Deng, T. Kieser, D. A. Hopwood, “Strong incompatibility” 50. D. Du, Y. Katsuyama, K. Shin-Ya, Y. Ohnishi, Reconstitution of a
s40168-017-0379-y; pmid: 29246178 between derivatives of the Streptomyces multi-copy type II polyketide synthase that catalyzes polyene formation.
8. S. Castañeda-Barba, E. M. Top, T. Stalder, Plasmids, a plasmid pIJ101. Mol. Gen. Genet. 214, 286–294 (1988). Angew. Chem. Int. Ed. 57, 1954–1957 (2018). doi: 10.1002/
molecular cornerstone of antimicrobial resistance in the One doi: 10.1007/BF00337723; pmid: 3070352 anie.201709636; pmid: 29265713
Health era. Nat. Rev. Microbiol. 22, 18–32 (2024). doi: 10.1038/ 31. S. Okamoto, T. Taguchi, K. Ochi, K. Ichinose, Biosynthesis of 51. D. Du et al., Structural basis for selectivity in a highly
s41579-023-00926-x; pmid: 37430173 actinorhodin and related antibiotics: Discovery of alternative reducing type II polyketide synthase. Nat. Chem. Biol. 16,
9. D. J. Newman, G. M. Cragg, Natural products as sources of routes for quinone formation encoded in the act gene 776–782 (2020). doi: 10.1038/s41589-020-0530-0;
new drugs over the nearly four decades from 01/1981 to cluster. Chem. Biol. 16, 226–236 (2009). doi: 10.1016/ pmid: 32367018
09/2019. J. Nat. Prod. 83, 770–803 (2020). doi: 10.1021/ j.chembiol.2009.01.015; pmid: 19246012 52. G. L. C. Grammbitter et al., An uncommon type II PKS
acs.jnatprod.9b01285; pmid: 32162523 32. O. Genilloud, Actinomycetes: Still a source of novel antibiotics. catalyzes biosynthesis of aryl polyene pigments. J. Am. Chem.
10. K. Lewis, The science of antibiotic discovery. Cell 181, 29–45 Nat. Prod. Rep. 34, 1203–1232 (2017). doi: 10.1039/ Soc. 141, 16615–16623 (2019). doi: 10.1021/jacs.8b10776;
(2020). doi: 10.1016/j.cell.2020.02.056; pmid: 32197064 C7NP00026J; pmid: 28820533 pmid: 30908039
11. A. Gavriilidou et al., Compendium of specialized metabolite 33. E. O. Stapley, J. M. Mata, I. M. Miller, T. C. Demny, H. B. Woodruff, 53. C. Fu et al., Biosynthetic studies of telomycin reveal new
biosynthetic diversity encoded in bacterial genomes. Antibiotic MSD-235. I. Production by Streptomyces avidinii and lipopeptides with enhanced activity. J. Am. Chem. Soc. 137,
Nat. Microbiol. 7, 726–735 (2022). doi: 10.1038/s41564-022- Streptomyces lavendulae. Antimicrob. Agents Chemother. 161, 7692–7705 (2015). doi: 10.1021/jacs.5b01794; pmid: 26043159
01110-2; pmid: 35505244 20–27 (1963). pmid: 14274896 54. J. Masschelein et al., A combination of polyunsaturated fatty
12. F. Hemmerling, J. Piel, Strategies to access biosynthetic 34. C. Dufour et al., Isolation and structural elucidation of acid, nonribosomal peptide and polyketide biosynthetic
novelty in bacterial genomes for drug discovery. Nat. Rev. Drug armeniaspirols A-C: Potent antibiotics against gram-positive machinery is used to assemble the zeamine antibiotics.
Discov. 21, 359–378 (2022). doi: 10.1038/s41573-022-00414-6; pathogens. Chemistry 18, 16123–16128 (2012). doi: 10.1002/ Chem. Sci. 6, 923–929 (2015). doi: 10.1039/C4SC01927J;
pmid: 35296832 chem.201201635; pmid: 23143837 pmid: 29560178

Xie et al., Science 386, eabq7333 (2024) 13 December 2024 8 of 9


RES EARCH | R E S E A R C H A R T I C L E

55. Y.-M. Shi et al., Global analysis of biosynthetic gene clusters reveals Americans. N. Engl. J. Med. 336, 466–473 (1997). thank A. F. Stewart, A. Keller, C. Beemelmanns, and A. A. Gurevich
conserved and unique natural products in entomopathogenic doi: 10.1056/NEJM199702133360703; pmid: 9017939 for their valuable discussions. Funding: C.F. and R.M. acknowledge
nematode-symbiotic bacteria. Nat. Chem. 14, 701–712 (2022). 66. M. L. Soares et al., Haplotypes and DNA sequence variation support from the Helmholtz International Lab (Helmholtz
doi: 10.1038/s41557-022-00923-2; pmid: 35469007 within and surrounding the transthyretin gene: Genotype- Association, InterLabs0007) and the German Federal Ministry of
56. M. Lv, J. Zhao, Z. Deng, Y. Yu, Characterization of the phenotype correlations in familial amyloid polyneuropathy Education and Research (BMBF, 01DO22003). R.M. acknowledges
biosynthetic gene cluster for benzoxazole antibiotics A33853 (V30M) in Portugal and Sweden. Eur. J. Hum. Genet. 12, funding from the German Research Foundation (DFG) and German
reveals unusual assembly logic. Chem. Biol. 22, 1313–1324 225–237 (2004). doi: 10.1038/sj.ejhg.5201095; Center for Infection Research (DZIF). Author contributions:
(2015). doi: 10.1016/j.chembiol.2015.09.005; pmid: 26496684 pmid: 14673473 Conceptualization: C.F., F.X., and R.M. Methodology: C.F., F.X., H.Z.,
57. T. Horch et al., Alternative benzoxazole assembly discovered in 67. T. Yokoyama, Y. Kosaka, M. Mizuguchi, Crystal structures and X.W. Investigation: C.F., F.X., H.Z., J.L., X.Y., M.N., A.A.A., A.K.,
anaerobic bacteria provides access to privileged heterocyclic of human transthyretin complexed with glabridin. J. Med. J.H., and O.V.K. Visualization: C.F., F.X., H.Z., and X.W. Funding
scaffold. Angew. Chem. Int. Ed. 61, e202205409 (2022). Chem. 57, 1090–1096 (2014). doi: 10.1021/jm401832j; acquisition: C.F. and R.M. Project administration: C.F., R.M., and
doi: 10.1002/anie.202205409; pmid: 35656913 pmid: 24422526 F.X. Supervision: C.F. and R.M. Writing – original draft: C.F., F.X., and
58. H. Suzuki, Y. Ohnishi, Y. Furusho, S. Sakuda, S. Horinouchi, 68. C. E. Bulawa et al., Tafamidis, a potent and selective H.Z. Writing – review and editing: C.F., F.X., R.M., H.Z., and X.W.
Novel benzene ring biosynthesis from C(3) and C(4) primary transthyretin kinetic stabilizer that inhibits the amyloid Competing interests: The authors declare that they have no
metabolites by two enzymes. J. Biol. Chem. 281, 36944–36951 cascade. Proc. Natl. Acad. Sci. U.S.A. 109, 9629–9634 (2012). competing interests. Data and materials availability: All tool
(2006). doi: 10.1074/jbc.M608103200; pmid: 17003031 doi: 10.1073/pnas.1121005109; pmid: 22645360 plasmids developed in this study will be deposited in Addgene
59. R. E. Viola, The central enzymes of the aspartate family of 69. M. Sun, X. Zhang, H. Hao, W. Li, C. Lu, Nocarbenzoxazoles (ID nos. 227507, 227508, and 227579 to 227583). All DNA sequences
amino acid biosynthesis. Acc. Chem. Res. 34, 339–349 (2001). A–G, benzoxazoles produced by halophilic Nocardiopsis of TDRs have been deposited in GenBank under accession nos.
doi: 10.1021/ar000057q; pmid: 11352712 lucentensis DSM 44048. J. Nat. Prod. 78, 2123–2127 (2015). OR339702 (Sav11), OR339701 (Sar13), and OR339703 (Sav17). All
60. H. Suzuki, Y. Furusho, T. Higashi, Y. Ohnishi, S. Horinouchi, doi: 10.1021/acs.jnatprod.5b00031; pmid: 26270803 DNA sequences of ACTIMOT toolkit plasmids have been deposited
A novel o-aminophenol oxidase responsible for formation 70. D. Du et al., Genome engineering and direct cloning of in GenBank under accession nos. OR264833 (pRel-Hyg),
of the phenoxazinone chromophore of grixazone. J. Biol. Chem. antibiotic gene clusters via phage ϕBT1 integrase-mediated OR264834 (pRel), OR264835 (pHelp), OR264836 (pCap-SG5-Hyg),

Downloaded from https://www.science.org at Baylor College of Medicine on December 13, 2024


281, 824–833 (2006). doi: 10.1074/jbc.M505806200; site-specific recombination in Streptomyces. Sci. Rep. 5, 8740 OR264837 (pCap-SG5-Apr), OR264838 (pCap-101-Hyg), and
pmid: 16282322 (2015). doi: 10.1038/srep08740; pmid: 25737113 OR264839 (pCap-101-Apr). The MS and NMR data for new
61. T. Coelho et al., Long-term effects of tafamidis for the 71. G. Wang et al., CRAGE enables rapid activation of biosynthetic molecules are deposited at Dryad (74). All other data supporting
treatment of transthyretin familial amyloid polyneuropathy. gene clusters in undomesticated bacteria. Nat. Microbiol. 4, this study’s findings are available in the main text and the
J. Neurol. 260, 2802–2814 (2013). doi: 10.1007/ 2498–2510 (2019). doi: 10.1038/s41564-019-0573-8; supplementary materials. License information: Copyright © 2024
s00415-013-7051-7; pmid: 23974642 pmid: 31611640 the authors, some rights reserved; exclusive licensee American
62. Y. Ando et al., Effects of tafamidis treatment on transthyretin 72. J. R. Patel, J. Oh, S. Wang, J. M. Crawford, F. J. Isaacs, Association for the Advancement of Science. No claim to original
(TTR) stabilization, efficacy, and safety in Japanese patients Cross-kingdom expression of synthetic genetic elements US government works. https://www.science.org/about/science-
with familial amyloid polyneuropathy (TTR-FAP) with promotes discovery of metabolites in the human microbiome. licenses-journal-article-reuse
Val30Met and non-Val30Met: A phase III, open-label study. Cell 185, 1487–1505.e14 (2022). doi: 10.1016/j.cell.2022.03.008;
J. Neurol. Sci. 362, 266–271 (2016). doi: 10.1016/ pmid: 35366417
j.jns.2016.01.046; pmid: 26944161 73. G. Muth, The pSG5-based thermosensitive vector family for SUPPLEMENTARY MATERIALS
63. L. Bartalena, J. Robbins, Thyroid hormone transport proteins. genome editing and gene expression in actinomycetes. science.org/doi/10.1126/science.abq7333
Clin. Lab. Med. 13, 583–598 (1993). doi: 10.1016/S0272-2712 Appl. Microbiol. Biotechnol. 102, 9067–9080 (2018). Materials and Methods
(18)30427-X; pmid: 8222576 doi: 10.1007/s00253-018-9334-5; pmid: 30191290 Figs. S1 to S176
64. M. J. Saraiva, P. P. Costa, D. S. Goodman, Biochemical 74. F. Xie et al., Autologous DNA mobilization and multiplication Tables S1 to S25
marker in familial amyloidotic polyneuropathy, Portuguese expedite natural products discovery from bacteria, Version 4, References (75–110)
type. Family studies on the transthyretin (prealbumin)- Dryad (2024); https://doi.org/10.5061/dryad.2280gb60t. MDAR Reproducibility Checklist
methionine-30 variant. J. Clin. Invest. 76, 2171–2177 (1985). Data S1 to S9
doi: 10.1172/JCI112224; pmid: 3908483 ACKN OWLED GMEN TS
65. D. R. Jacobson et al., Variant-sequence transthyretin We thank the Helmholtz Institute for Pharmaceutical Research Submitted 22 August 2023; accepted 10 October 2024
(isoleucine 122) in late-onset cardiac amyloidosis in black Saarland (HIPS) for use of the MS and NMR platforms. We also 10.1126/science.abq7333

Xie et al., Science 386, eabq7333 (2024) 13 December 2024 9 of 9

You might also like

pFad - Phonifier reborn

Pfad - The Proxy pFad of © 2024 Garber Painting. All rights reserved.

Note: This service is not intended for secure transactions such as banking, social media, email, or purchasing. Use at your own risk. We assume no liability whatsoever for broken pages.


Alternative Proxies:

Alternative Proxy

pFad Proxy

pFad v3 Proxy

pFad v4 Proxy