Rapid advancements in sequencing technologies have revolutionized our view on microbial communities. While amplicon sequencing provides information on community composition and diversity, shotgun and whole-genome sequencing allow us to reliably anticipate evolutionary and ecological relationships between microbes and to obtain functional information on communities. Computational models assessing the metabolic capacity of individual members, or an entire consortium, have become very popular and powerful (Zengler and Palsson, 2012; Gu et al., 2019; García-Jiménez et al., 2021). The major focus of such modeling approaches is typically on the primary metabolism of bacteria, as genes involved in core metabolic pathways are highly conserved and can be identified with relative ease (Gu et al., 2019; Colarusso et al., 2021). Conversely, analysis of the secondary metabolites has attracted less attention, even though they include compounds such as antibiotics, toxins, siderophores, biosurfactants, all known to have important implications for microbial community assembly (Scherlach and Hertweck, 2018; Trivedi et al., 2020) and to be important sources for pharmaceutical discoveries (Price-Whelan et al., 2006; Thirumurugan et al., 2018; Chaturvedi et al., 2012).

There are multiple challenges that currently prevent a detailed unraveling of secondary metabolism of bacteria based on genome data (Penn et al., 2009; Yee et al., 2023; Lautru et al., 2005). First, most secondary metabolites are produced by pathways comprised of modular enzymes such as non-ribosomal peptide synthetases (NRPSs) or polyketide synthases (PKS) (Andryukov et al., 2019; Kautsar et al., 2021). Locating complete biosynthesis clusters and identifying all enzyme-encoding genes is challenging from highly fragmented metagenomic sequences or draft genomes with a high number of contigs. Second, functional predictions for coding regions within a cluster rely on homologous comparisons with experimentally characterized genes. Such information is often restricted to a limited number of model organisms, meaning that only a small portion of the existing secondary metabolism pathways is covered by current databases. Finally, given the complex multi-modular biosynthesis machineries, it is challenging to precisely predict the secondary metabolites produced even with accurately annotated NRPS or PKS clusters. The main challenge is that a large pool of non-proteinogenic amino acids is used as substrates and the specificity of an enzyme’s A domain, connecting these unusual amino acids, is often poorly understood (He et al., 2023). As a result, new computational methods are needed to accurately reconstruct bacterial secondary metabolism from sequence data.

Here, we present a new bioinformatic pipeline that overcomes these challenges. We specifically focus on a particular class of secondary metabolites (iron-scavenging siderophores) as a case study to develop a bioinformatic workflow that predicts the chemical structure of the produced metabolites with nearly 100% accuracy. Our pipeline is based on improved gene annotation combined with a phylogeny- and feature sequence-based substrate prediction techniques (Figure 1). In comparison with the currently available databases and bioinformatic tools (Blin et al., 2023), the main advancement of our method is the more accurate prediction of synthesized products based on non-ribosomal peptide synthetase (NRPS) clusters identified in genome data. In addition, we show that our workflow is capable of correcting previous misannotations and is extendable to other secondary metabolites (e.g. toxins, biosurfactants, virulence factors) and microbial genera.

Figure 1

Download asset Open asset

Among siderophores, we focus on NRPS machineries that are responsible for the biosynthesis of pyoverdines, a class of chemically diverse siderophores with high iron affinity, produced by Pseudomonas bacteria (Ringel and Brüser, 2018; Hopkinson and Morel, 2009). While each Pseudomonas strain produces a single type of pyoverdine, an enormous structural diversity has been described across strains and species (Cobessi et al., 2005; Diggle and Whiteley, 2020; Meyer et al., 1997; Bodilis et al., 2009). Pyoverdine types differ in their peptide backbone, meaning that the diversity should be mirrored in NRPS enzyme diversity and their selectivity for the different amino acid substrates (Hopkinson and Morel, 2009). Based on this knowledge, our pipeline entails the following steps (Figure 1): (i) identification of the complete sequences of pyoverdine synthetase genes from fragmented draft genomes, (ii) building the pyoverdine biosynthesis machinery in silico by extracting the feature sequences for substrate specificity from motif-standardized NRPSs, and (iii) predicting the precise chemical structure of pyoverdines followed by empirical verification.

An additional element of iron acquisition involves specific receptors that recognize and trigger the uptake of iron-loaded siderophores. Pyoverdine receptors are annotated as FpvA and it is known that receptor diversity matches pyoverdine diversity (Bodilis et al., 2009; Kümmerli, 2023). Moreover, FpvA belongs to the family of TonB-dependent receptors and a single Pseudomonas species often has multiple gene copies encoding different receptors. This poses an additional bioinformatic challenge: how to find the gene encoding the specific pyoverdine receptor among several potential receptor genes? To overcome this, we develop an algorithm that focuses on receptor sequence regions involved in pyoverdine recognition and translocation across the outer membrane with supervised learning methods that reliably locate fpvA genes in fragmented genomes based on these regions (Figure 1). Not only did our method unveil a substantial number of previously unrecognized pyoverdine receptors, but it also revealed that certain receptor groups are widely distributed among Pseudomonas strains, opening the possibility of pyoverdine sharing and exploitation across the strain and species boundaries. Altogether, our bioinformatic pipeline uses knowledge-guided insights empowered by supervised learning to construct a first systematic sequence-to-function mapping of a family of secondary metabolites (pyoverdine) and their corresponding receptors. Our analysis unveils a yet unrecognized extraordinary diversity of iron-scavenging machineries in pseudomonads.

The rapid expansion of sequencing data offers exciting opportunities for microbiology (Almeida et al., 2019; Schloss and Handelsman, 2005; Handelsman, 2004). One key challenge of current research in the field is to infer biological functions of microbial communities from genome sequence data (Tsilimigras and Fodor, 2016; Weiss et al., 2016; Faust and Raes, 2012). While this endeavor is increasingly successful for biological functions involving the primary metabolism and the associated complex metabolic flux, reconstructing aspects of the secondary metabolism is much more challenging. The main issue is that neither the function of a secondary metabolite enzyme nor the resulting metabolite can be precisely predicted from gene sequence data. In our study, we tackled this challenge and developed a bioinformatic pipeline to reconstruct the complete secondary metabolism pathway of pyoverdines, a class of iron-scavenging siderophores produced by Pseudomonas spp. These secondary metabolites are synthetized by a series of non-ribosomal peptide synthetases and require a specific receptor (FpvA) for uptake. We combined knowledge-guided learning with phylogeny-based methods to predict with high accuracy: (i) the full pyoverdine assembly line, (ii) the substrate specificity for each enzyme within the assembly lines, (iii) the complete chemical structure of pyoverdines, and (iv) the FpvA receptors from genome sequences. After validation, we tested our pipeline with sequence data from 1664 phylogenetically distinct Pseudomonas strains and were able to determine 18,292 enzymatic A domains involved in pyoverdine biosynthesis, to reliably predict 97.8% of their substrates, and to identify 188 different pyoverdine molecule structures and 4547 FpvA receptor variants belonging to 94 distinct groups. The uncovered diversity is stunning and goes far beyond currently known levels of variation (73 pyoverdines and 3 FpvA groups).

We show that knowledge-guided learning is an extremely powerful tool to predict enzyme, metabolite, and receptor properties. The establishment of our entire pipeline is based on only 101 previously known enzymatic A domains (from 13 known pyoverdine assembly lines) and 21 FpvA receptor sequences. Even with this limited amount of information, we were able to predict the substrates of almost all the 18,292 enzymatic A domains and to identify 4547 FpvA receptors from the sequence data. A key insight from our knowledge-guided learning is that comparisons based on the full gene sequences (e.g. for pyoverdine synthetase or receptor) are typically non-informative and unsuitable for obtaining functional information. This is because overall diversity does not stand for functional diversity, meaning that A domains recognizing the same substrate can diverge substantially in their full sequences. The same holds true for receptor sequences: whole-sequence alignments can neither accurately identify FpvA receptors nor reliably separate them into functional groups. Instead, it is imperative to extract informative feature sequences that are defined as sequence stretches within a gene whose diversity is tightly linked to variation in its functioning. We successfully extracted and applied feature sequence comparisons for both A domain substrate prediction and FpvA identification. It is important to note that a knowledge-guided pipeline does not have to be perfect right from the start. For example, our pipeline for pyoverdine structure prediction returned unknowns for several amino acid positions within the PEP. Our experimental verifications then revealed new substrates such as valine and citrulline. This information can then be used to refine our prediction algorithm in a feedback loop.

Another main advantage of our bioinformatic pipeline is that it can be applied to draft genomes. This reflects a major improvement compared to existing annotation tools such as antiSMASH (Blin et al., 2019), which typically has difficulties in recognizing NRPS structures in fragmented genome assemblies. However, draft genomes are the most common data source in microbiology. While our pipeline shows high performance, we need to acknowledge that we still lose many genomes (6087 out of 9599, 63.4%). The reason for the loss is that the pyoverdine biosynthesis machinery is large, which increases the probability that it is positioned at the end of a contig. We decided to exclude those cases because the annotated biosynthesis machinery might be truncated and thus incomplete. Thus, the high loss rate of draft genomes is rather due to limitations in sequence quality (too many short contigs) and not due to a limitation of our bioinformatic pipeline. We believe that this limitation may disappear in the future as long-read sequencing technologies are quickly becoming cheaper and more reliable.

We further show that knowledge-guided learning combined with a phylogeny-focused approach is a powerful tool for predicting the substrate specificity of A domains of synthetases. It outperforms currently known bioinformatics prediction tools of NRPS substrates such as antiSMASH (Blin et al., 2019). Most current algorithms (Khayatt et al., 2013; Minowa et al., 2007; Prieto et al., 2012; Röttig et al., 2011; Zierep et al., 2020) perform poorly when applied to pyoverdines, particularly when encountering non-proteogenic amino acids. The high accuracy of our algorithm can largely be attributed to our reference set, composing only 13 pyoverdines from Pseudomonas spp., yet capturing most of the substrate diversity. Similarly accurate predictions based on a handful of known substrates among closely related species were observed in several fungal NRPS systems (Fan et al., 2022). It is worth noting that when the algorithm output is ‘unknown’, it signifies the presence of uncharacterized A domains that are not yet incorporated into the reference dataset. This should prompt researchers to pay attention to these A domains, and like in our case, subject them to further experimental investigation. This approach helped us discover new substrates (valine, citrulline) and link them to the corresponding A domains. The novel substrates identified through our structural assessment and mass spectrometry experiment can subsequently be used to enhance the precision of our phylogeny-centered substrate prediction technique in the future, creating a progressive feedback loop of expanding knowledge. Taken together, supervised learning based on a few known compounds produced by species from the same genus can outperform generalized prediction algorithms trained on many products from a diverse set of microbes for NRPS substrate predictions.

Our results show that both pyoverdine and receptor diversity has been vastly underestimated. While considerable pyoverdine diversity (n=73) has already been captured in previous studies, here we predicted 151 putatively new variants. On the receptor side, the uncovered novel diversity is more dramatic. One reason for this is that research on receptors has mainly focused on the pathogen P. aeruginosa (Cobessi et al., 2005; Diggle and Whiteley, 2020; Meyer et al., 1997; Bodilis et al., 2009; Smith et al., 2005). For this species, three different pyoverdine types were described (Meyer et al., 1997) together with three structurally different FpvA receptor types that each recognize one of the pyoverdine types (Bodilis et al., 2009). While our study confirmed that P. aeruginosa strains (n=554) indeed have only 3 pyoverdine-receptor types, we further discovered 91 new FpvA groups among environmental Pseudomonas spp. Our findings raise the question why there are so many different pyoverdine and receptor variants. One potential explanation is that the benefit of specific siderophores could be context-dependent and locally adapted to the multitude of different environmental conditions pseudomonads are exposed to. For example, experimental work has revealed that pyoverdines can be cooperatively shared among strains with matching receptors (Kramer et al., 2020; Gu et al., 2020), or conversely, pyoverdines can serve as competitive agents by locking away iron from species that have non-matching receptors (Figueiredo et al., 2022). Given that bioavailable iron is limited in most natural and host-associated habitats (Andrews et al., 2003; Boyd and Ellwood, 2010; Emerson et al., 2012), the unraveled functional diversity is likely a direct evolutionary consequence of the struggle and competition of microbes for iron. While experimental work is often restricted to a low number of strains, we propose that our bioinformatic pipeline can be used to predict pyoverdine-mediated interaction networks across thousands of strains and across different habitats. We have addressed this point in a complementary study (Gu et al., 2024).

We are confident that our pipeline can be expanded to study iron competition in multi-species communities and perhaps also in plant-microbe ecosystems, as siderophores exist ubiquitously and are shared among microbes and certain hosts (Ruolin, 2023). To move further, a key question is whether our knowledge-guided approach can be applied to other microbes and important secondary metabolites, such as antibiotics, toxins, biosurfactants, and pigments? Our preliminary investigation, focused on the curated dataset in Burkholderiales as a test case, provides a highly promising response to this query. The outcomes from this evaluation predominantly showcased the method’s potential for extension to diverse microbial genera and metabolites, given sufficiently accurate dataset. Meanwhile, we observed a notable decline in accuracy (from 94% to 83%) when compared to pyoverdine. This indicates that our algorithm possibly needs refinement, particularly concerning specific secondary metabolites or microbial genera. As soon as sufficient case-by-case knowledge on a specific system is available, the annotation strategies together with the feature sequence extraction and the phylogeny-focused approach developed in our paper can be applied. It is worth noting that our current algorithm does not predict tailoring reactions. While tailoring reactions are crucial for predicting the final NRPS product structure, this challenge is not addressed by any existing pipelines and remains an area for future research. In our study, it is important to note that the specificity of pyoverdines is mainly determined by the backbone composition, with tailoring reactions playing a relatively minor role (Ringel and Brüser, 2018). For most of the secondary metabolites listed above, there are no receptors as the compounds have purely extracellular functions, which substantially simplifies the development of bioinformatic pipelines. In the long run, it will certainly be possible to automate the steps implemented in our workflow so that the algorithms can be applied to a large set of secondary metabolites when fed with an appropriate training set.

The source code and parameters used are available in the supplementary material. The sequences of the 20 Pseudomonas strains are available in ENA under accession number PRJEB76792. The raw data and code can be accessed at the Dryad repository (https://doi.org/10.5061/dryad.jdfn2z3kw) and on GitHub (https://github.com/zhiyuanLab/Pyoverdine-genome-mining copy archived at Gu, 2024).

1. 1. Maréchal S
   2. Kümmerli R

   (2024) European Nucleotide Archive

   ID PRJEB76792. From sequence to molecules: Feature sequence-based genome mining uncovers the hidden diversity of bacterial siderophore pathways.

   https://www.ebi.ac.uk/ena/browser/view/PRJEB76792
2. 1. Gu S
   2. Li Z

   (2024) Dryad Digital Repository

   Data from: Feature sequence-based genome mining uncovers the hidden diversity of bacterial siderophore pathways.

   https://doi.org/10.5061/dryad.jdfn2z3kw

1. 1. Almeida A
   2. Mitchell AL
   3. Boland M
   4. Forster SC
   5. Gloor GB
   6. Tarkowska A
   7. Lawley TD
   8. Finn RD

   (2019) A new genomic blueprint of the human gut microbiota

   Nature 568:499–504.

   https://doi.org/10.1038/s41586-019-0965-1

   • PubMed
   • Google Scholar
2. 1. Andrews SC
   2. Robinson AK
   3. Rodríguez-Quiñones F

   (2003) Bacterial iron homeostasis

   FEMS Microbiology Reviews 27:215–237.

   https://doi.org/10.1016/S0168-6445(03)00055-X

   • PubMed
   • Google Scholar
3. 1. Andryukov B
   2. Mikhailov V
   3. Besednova N

   (2019) The biotechnological potential of secondary metabolites from marine bacteria

   Biology 01:0063.v1.

   https://doi.org/10.20944/preprints201905.0063.v1

   • Google Scholar
4. 1. Asnicar F
   2. Thomas AM
   3. Beghini F
   4. Mengoni C
   5. Manara S
   6. Manghi P
   7. Zhu Q
   8. Bolzan M
   9. Cumbo F
   10. May U
   11. Sanders JG
   12. Zolfo M
   13. Kopylova E
   14. Pasolli E
   15. Knight R
   16. Mirarab S
   17. Huttenhower C
   18. Segata N

   (2020) Precise phylogenetic analysis of microbial isolates and genomes from metagenomes using PhyloPhlAn 3.0

   Nature Communications 11:2500.

   https://doi.org/10.1038/s41467-020-16366-7

   • PubMed
   • Google Scholar
5. 1. Bateman A
   2. Coin L
   3. Durbin R
   4. Finn RD
   5. Hollich V
   6. Griffiths-Jones S
   7. Khanna A
   8. Marshall M
   9. Moxon S
   10. Sonnhammer ELL
   11. Studholme DJ
   12. Yeats C
   13. Eddy SR

   (2004) The Pfam protein families database

   Nucleic Acids Research 32:D138–D141.

   https://doi.org/10.1093/nar/gkh121

   • PubMed
   • Google Scholar
6. 1. Blin K
   2. Shaw S
   3. Steinke K
   4. Villebro R
   5. Ziemert N
   6. Lee SY
   7. Medema MH
   8. Weber T

   (2019) antiSMASH 5.0: updates to the secondary metabolite genome mining pipeline

   Nucleic Acids Research 47:W81–W87.

   https://doi.org/10.1093/nar/gkz310

   • PubMed
   • Google Scholar
7. 1. Blin K
   2. Shaw S
   3. Augustijn HE
   4. Reitz ZL
   5. Biermann F
   6. Alanjary M
   7. Fetter A
   8. Terlouw BR
   9. Metcalf WW
   10. Helfrich EJN
   11. van Wezel GP
   12. Medema MH
   13. Weber T

   (2023) antiSMASH 7.0: new and improved predictions for detection, regulation, chemical structures and visualisation

   Nucleic Acids Research 51:W46–W50.

   https://doi.org/10.1093/nar/gkad344

   • PubMed
   • Google Scholar
8. 1. Bodilis J
   2. Ghysels B
   3. Osayande J
   4. Matthijs S
   5. Pirnay J-P
   6. Denayer S
   7. De Vos D
   8. Cornelis P

   (2009) Distribution and evolution of ferripyoverdine receptors in Pseudomonas aeruginosa

   Environmental Microbiology 11:2123–2135.

   https://doi.org/10.1111/j.1462-2920.2009.01932.x

   • PubMed
   • Google Scholar
9. 1. Boyd PW
   2. Ellwood MJ

   (2010) The biogeochemical cycle of iron in the ocean

   Nature Geoscience 3:675–682.

   https://doi.org/10.1038/ngeo964

   • Google Scholar
10. 1. Butaitė E
    2. Baumgartner M
    3. Wyder S
    4. Kümmerli R

    (2017) Siderophore cheating and cheating resistance shape competition for iron in soil and freshwater Pseudomonas communities

    Nature Communications 8:414.

    https://doi.org/10.1038/s41467-017-00509-4

    • PubMed
    • Google Scholar
11. 1. Butaitė E
    2. Kramer J
    3. Wyder S
    4. Kümmerli R

    (2018) Environmental determinants of pyoverdine production, exploitation and competition in natural Pseudomonas communities

    Environmental Microbiology 20:3629–3642.

    https://doi.org/10.1111/1462-2920.14355

    • PubMed
    • Google Scholar
12. Preprint

    1. Chan DCK
    2. Burrows LL

    (2022) FpvB Is a high-affinity transporter for xenosiderophores ferrichrome and ferrioxamine B

    bioRxiv.

    https://doi.org/10.1101/2022.09.20.508722

    • Google Scholar
13. 1. Chaturvedi KS
    2. Hung CS
    3. Crowley JR
    4. Stapleton AE
    5. Henderson JP

    (2012) The siderophore yersiniabactin binds copper to protect pathogens during infection

    Nature Chemical Biology 8:731–736.

    https://doi.org/10.1038/nchembio.1020

    • PubMed
    • Google Scholar
14. 1. Chen H
    2. Zhong L
    3. Zhou H
    4. Bai X
    5. Sun T
    6. Wang X
    7. Zhao Y
    8. Ji X
    9. Tu Q
    10. Zhang Y
    11. Bian X

    (2023) Biosynthesis and engineering of the nonribosomal peptides with a C-terminal putrescine

    Nature Communications 14:6619.

    https://doi.org/10.1038/s41467-023-42387-z

    • PubMed
    • Google Scholar
15. 1. Cobessi D
    2. Celia H
    3. Folschweiller N
    4. Schalk IJ
    5. Abdallah MA
    6. Pattus F

    (2005) The crystal structure of the pyoverdine outer membrane receptor FpvA from Pseudomonas aeruginosa at 3.6 angstroms resolution

    Journal of Molecular Biology 347:121–134.

    https://doi.org/10.1016/j.jmb.2005.01.021

    • PubMed
    • Google Scholar
16. 1. Colarusso AV
    2. Goodchild-Michelman I
    3. Rayle M
    4. Zomorrodi AR

    (2021) Computational modeling of metabolism in microbial communities on a genome-scale

    Current Opinion in Systems Biology 26:46–57.

    https://doi.org/10.1016/j.coisb.2021.04.001

    • Google Scholar
17. 1. Diggle SP
    2. Whiteley M

    (2020) Microbe profile: Pseudomonas aeruginosa: opportunistic pathogen and lab rat

    Microbiology 166:30–33.

    https://doi.org/10.1099/mic.0.000860

    • PubMed
    • Google Scholar
18. 1. Emerson D
    2. Roden E
    3. Twining BS

    (2012) The microbial ferrous wheel: iron cycling in terrestrial, freshwater, and marine environments

    Frontiers in Microbiology 3:383.

    https://doi.org/10.3389/fmicb.2012.00383

    • PubMed
    • Google Scholar
19. Preprint

    1. Fan J
    2. Ren J
    3. He R
    4. Wei P-L
    5. Li Y
    6. Li W
    7. Chen D
    8. Druzhinina IS
    9. Li Z
    10. Yin W-B

    (2022) Biosynthetic diversification of peptaibol mediates fungus-mycohost interactions

    bioRxiv.

    https://doi.org/10.1101/2022.06.05.494846

    • Google Scholar
20. 1. Faust K
    2. Raes J

    (2012) Microbial interactions: from networks to models

    Nature Reviews. Microbiology 10:538–550.

    https://doi.org/10.1038/nrmicro2832

    • PubMed
    • Google Scholar
21. 1. Felnagle EA
    2. Jackson EE
    3. Chan YA
    4. Podevels AM
    5. Berti AD
    6. McMahon MD
    7. Thomas MG

    (2008) Nonribosomal peptide synthetases involved in the production of medically relevant natural products

    Molecular Pharmaceutics 5:191–211.

    https://doi.org/10.1021/mp700137g

    • PubMed
    • Google Scholar
22. 1. Figueiredo ART
    2. Özkaya Ö
    3. Kümmerli R
    4. Kramer J

    (2022) Siderophores drive invasion dynamics in bacterial communities through their dual role as public good versus public bad

    Ecology Letters 25:138–150.

    https://doi.org/10.1111/ele.13912

    • PubMed
    • Google Scholar
23. 1. García-Jiménez B
    2. Torres-Bacete J
    3. Nogales J

    (2021) Metabolic modelling approaches for describing and engineering microbial communities

    Computational and Structural Biotechnology Journal 19:226–246.

    https://doi.org/10.1016/j.csbj.2020.12.003

    • PubMed
    • Google Scholar
24. 1. González J
    2. Salvador M
    3. Özkaya Ö
    4. Spick M
    5. Reid K
    6. Costa C
    7. Bailey MJ
    8. Avignone Rossa C
    9. Kümmerli R
    10. Jiménez JI

    (2021) Loss of a pyoverdine secondary receptor in Pseudomonas aeruginosa results in a fitter strain suitable for population invasion

    The ISME Journal 15:1330–1343.

    https://doi.org/10.1038/s41396-020-00853-2

    • PubMed
    • Google Scholar
25. 1. Greenwald J
    2. Nader M
    3. Celia H
    4. Gruffaz C
    5. Geoffroy V
    6. Meyer J-M
    7. Schalk IJ
    8. Pattus F

    (2009) FpvA bound to non-cognate pyoverdines: molecular basis of siderophore recognition by an iron transporter

    Molecular Microbiology 72:1246–1259.

    https://doi.org/10.1111/j.1365-2958.2009.06721.x

    • PubMed
    • Google Scholar
26. 1. Gu C
    2. Kim GB
    3. Kim WJ
    4. Kim HU
    5. Lee SY

    (2019) Current status and applications of genome-scale metabolic models

    Genome Biology 20:121.

    https://doi.org/10.1186/s13059-019-1730-3

    • PubMed
    • Google Scholar
27. 1. Gu S
    2. Wei Z
    3. Shao Z
    4. Friman V-P
    5. Cao K
    6. Yang T
    7. Kramer J
    8. Wang X
    9. Li M
    10. Mei X
    11. Xu Y
    12. Shen Q
    13. Kümmerli R
    14. Jousset A

    (2020) Competition for iron drives phytopathogen control by natural rhizosphere microbiomes

    Nature Microbiology 5:1002–1010.

    https://doi.org/10.1038/s41564-020-0719-8

    • PubMed
    • Google Scholar
28. Software

    1. Gu S

    (2024) Pyoverdine-genome-mining, version swh:1:rev:0753663ba33d2fe449b973f91a173a4f4a4bba1e

    Software Heritage.

    https://archive.softwareheritage.org/swh:1:dir:ec13a0ec91efa902ea941f0a4d9a20fb0a6fa2c4;origin=https://github.com/ZhiyuanLab/Pyoverdine-genome-mining;visit=swh:1:snp:3286a57f0132680fabf898a6d47e2676818da587;anchor=swh:1:rev:0753663ba33d2fe449b973f91a173a4f4a4bba1e
29. Preprint

    1. Gu S
    2. Shao Z
    3. Qu Z
    4. Zhu S
    5. Shao Y
    6. Zhang D
    7. Allen R
    8. He R
    9. Shao J
    10. Xiong G
    11. Jousset A
    12. Friman V-P
    13. Wei Z
    14. Kuemmerli R
    15. Li Z

    (2024) Siderophore synthetase-receptor gene coevolution reveals habitat- and pathogen-specific bacterial iron interaction networks

    bioRxiv.

    https://doi.org/10.1101/2023.11.05.565711

    • Google Scholar
30. 1. Handelsman J

    (2004) Metagenomics: application of genomics to uncultured microorganisms

    Microbiology and Molecular Biology Reviews 68:669–685.

    https://doi.org/10.1128/MMBR.68.4.669-685.2004

    • PubMed
    • Google Scholar
31. 1. He R
    2. Zhang J
    3. Shao Y
    4. Gu S
    5. Song C
    6. Qian L
    7. Yin W-B
    8. Li Z

    (2023) Knowledge-guided data mining on the standardized architecture of NRPS: Subtypes, novel motifs, and sequence entanglements

    PLOS Computational Biology 19:e1011100.

    https://doi.org/10.1371/journal.pcbi.1011100

    • PubMed
    • Google Scholar
32. 1. Hopkinson BM
    2. Morel FMM

    (2009) The role of siderophores in iron acquisition by photosynthetic marine microorganisms

    Biometals 22:659–669.

    https://doi.org/10.1007/s10534-009-9235-2

    • PubMed
    • Google Scholar
33. 1. Jin Z
    2. Li J
    3. Ni L
    4. Zhang R
    5. Xia A
    6. Jin F

    (2018) Conditional privatization of a public siderophore enables Pseudomonas aeruginosa to resist cheater invasion

    Nature Communications 9:1383.

    https://doi.org/10.1038/s41467-018-03791-y

    • PubMed
    • Google Scholar
34. 1. Kautsar SA
    2. Blin K
    3. Shaw S
    4. Weber T
    5. Medema MH

    (2021) BiG-FAM: the biosynthetic gene cluster families database

    Nucleic Acids Research 49:D490–D497.

    https://doi.org/10.1093/nar/gkaa812

    • PubMed
    • Google Scholar
35. 1. Keller NP

    (2019) Fungal secondary metabolism: regulation, function and drug discovery

    Nature Reviews. Microbiology 17:167–180.

    https://doi.org/10.1038/s41579-018-0121-1

    • PubMed
    • Google Scholar
36. 1. Khayatt BI
    2. Overmars L
    3. Siezen RJ
    4. Francke C

    (2013) Classification of the adenylation and acyl-transferase activity of NRPS and PKS systems using ensembles of substrate specific hidden Markov models

    PLOS ONE 8:e62136.

    https://doi.org/10.1371/journal.pone.0062136

    • PubMed
    • Google Scholar
37. 1. Kramer J
    2. Özkaya Ö
    3. Kümmerli R

    (2020) Bacterial siderophores in community and host interactions

    Nature Reviews. Microbiology 18:152–163.

    https://doi.org/10.1038/s41579-019-0284-4

    • PubMed
    • Google Scholar
38. 1. Kümmerli R

    (2023) Iron acquisition strategies in pseudomonads: mechanisms, ecology, and evolution

    Biometals 36:777–797.

    https://doi.org/10.1007/s10534-022-00480-8

    • PubMed
    • Google Scholar
39. 1. Lautru S
    2. Deeth RJ
    3. Bailey LM
    4. Challis GL

    (2005) Discovery of a new peptide natural product by Streptomyces coelicolor genome mining

    Nature Chemical Biology 1:265–269.

    https://doi.org/10.1038/nchembio731

    • PubMed
    • Google Scholar
40. 1. Meyer J-M
    2. Stintzi A
    3. De Vos D
    4. Cornelis P
    5. Tappe R
    6. Taraz K
    7. Budzikiewicz H

    (1997) Use of siderophores to type pseudomonads: the three Pseudomonas aeruginosa pyoverdine systems

    Microbiology 143:35–43.

    https://doi.org/10.1099/00221287-143-1-35

    • PubMed
    • Google Scholar
41. 1. Meyer JM

    (2000) Pyoverdines: pigments, siderophores and potential taxonomic markers of fluorescent Pseudomonas species

    Archives of Microbiology 174:135–142.

    https://doi.org/10.1007/s002030000188

    • PubMed
    • Google Scholar
42. 1. Minowa Y
    2. Araki M
    3. Kanehisa M

    (2007) Comprehensive analysis of distinctive polyketide and nonribosomal peptide structural motifs encoded in microbial genomes

    Journal of Molecular Biology 368:1500–1517.

    https://doi.org/10.1016/j.jmb.2007.02.099

    • PubMed
    • Google Scholar
43. 1. Penn K
    2. Jenkins C
    3. Nett M
    4. Udwary DW
    5. Gontang EA
    6. McGlinchey RP
    7. Foster B
    8. Lapidus A
    9. Podell S
    10. Allen EE
    11. Moore BS
    12. Jensen PR

    (2009) Genomic islands link secondary metabolism to functional adaptation in marine Actinobacteria

    The ISME Journal 3:1193–1203.

    https://doi.org/10.1038/ismej.2009.58

    • PubMed
    • Google Scholar
44. 1. Price-Whelan A
    2. Dietrich LEP
    3. Newman DK

    (2006) Rethinking “secondary” metabolism: physiological roles for phenazine antibiotics

    Nature Chemical Biology 2:71–78.

    https://doi.org/10.1038/nchembio764

    • PubMed
    • Google Scholar
45. 1. Prieto C
    2. García-Estrada C
    3. Lorenzana D
    4. Martín JF

    (2012) NRPSsp: non-ribosomal peptide synthase substrate predictor

    Bioinformatics 28:426–427.

    https://doi.org/10.1093/bioinformatics/btr659

    • PubMed
    • Google Scholar
46. 1. Rehm K
    2. Vollenweider V
    3. Kümmerli R
    4. Bigler L

    (2022) A comprehensive method to elucidate pyoverdines produced by fluorescent Pseudomonas spp. by UHPLC-HR-MS/MS

    Analytical and Bioanalytical Chemistry 414:2671–2685.

    https://doi.org/10.1007/s00216-022-03907-w

    • PubMed
    • Google Scholar
47. 1. Ringel MT
    2. Brüser T

    (2018) The biosynthesis of pyoverdines

    Microbial Cell 5:424–437.

    https://doi.org/10.15698/mic2018.10.649

    • PubMed
    • Google Scholar
48. 1. Röttig M
    2. Medema MH
    3. Blin K
    4. Weber T
    5. Rausch C
    6. Kohlbacher O

    (2011) NRPSpredictor2--a web server for predicting NRPS adenylation domain specificity

    Nucleic Acids Research 39:W362–W367.

    https://doi.org/10.1093/nar/gkr323

    • PubMed
    • Google Scholar
49. Preprint

    1. Ruolin H

    (2023) SIDERITE: unveiling hidden siderophore diversity in the chemical space through digital exploration

    bioRxiv.

    https://doi.org/10.1101/2023.08.31.555687

    • Google Scholar
50. 1. Schalk IJ
    2. Mislin GLA
    3. Brillet K

    (2012) Structure, function and binding selectivity and stereoselectivity of siderophore-iron outer membrane transporters

    Current Topics in Membranes 69:37–66.

    https://doi.org/10.1016/B978-0-12-394390-3.00002-1

    • PubMed
    • Google Scholar
51. 1. Scherlach K
    2. Hertweck C

    (2018) Mediators of mutualistic microbe-microbe interactions

    Natural Product Reports 35:303–308.

    https://doi.org/10.1039/c7np00035a

    • PubMed
    • Google Scholar
52. 1. Schloss PD
    2. Handelsman J

    (2005) Metagenomics for studying unculturable microorganisms: cutting the Gordian knot

    Genome Biology 6:229.

    https://doi.org/10.1186/gb-2005-6-8-229

    • PubMed
    • Google Scholar
53. 1. Smith EE
    2. Sims EH
    3. Spencer DH
    4. Kaul R
    5. Olson MV

    (2005) Evidence for diversifying selection at the pyoverdine locus of Pseudomonas aeruginosa

    Journal of Bacteriology 187:2138–2147.

    https://doi.org/10.1128/JB.187.6.2138-2147.2005

    • PubMed
    • Google Scholar
54. 1. Süssmuth RD
    2. Mainz A

    (2017) Nonribosomal peptide synthesis-principles and prospects

    Angewandte Chemie 56:3770–3821.

    https://doi.org/10.1002/anie.201609079

    • PubMed
    • Google Scholar
55. 1. Terlouw BR
    2. Blin K
    3. Navarro-Muñoz JC
    4. Avalon NE
    5. Chevrette MG
    6. Egbert S
    7. Lee S
    8. Meijer D
    9. Recchia MJJ
    10. Reitz ZL
    11. van Santen JA
    12. Selem-Mojica N
    13. Tørring T
    14. Zaroubi L
    15. Alanjary M
    16. Aleti G
    17. Aguilar C
    18. Al-Salihi SAA
    19. Augustijn HE
    20. Avelar-Rivas JA
    21. Avitia-Domínguez LA
    22. Barona-Gómez F
    23. Bernaldo-Agüero J
    24. Bielinski VA
    25. Biermann F
    26. Booth TJ
    27. Carrion Bravo VJ
    28. Castelo-Branco R
    29. Chagas FO
    30. Cruz-Morales P
    31. Du C
    32. Duncan KR
    33. Gavriilidou A
    34. Gayrard D
    35. Gutiérrez-García K
    36. Haslinger K
    37. Helfrich EJN
    38. van der Hooft JJJ
    39. Jati AP
    40. Kalkreuter E
    41. Kalyvas N
    42. Kang KB
    43. Kautsar S
    44. Kim W
    45. Kunjapur AM
    46. Li Y-X
    47. Lin G-M
    48. Loureiro C
    49. Louwen JJR
    50. Louwen NLL
    51. Lund G
    52. Parra J
    53. Philmus B
    54. Pourmohsenin B
    55. Pronk LJU
    56. Rego A
    57. Rex DAB
    58. Robinson S
    59. Rosas-Becerra LR
    60. Roxborough ET
    61. Schorn MA
    62. Scobie DJ
    63. Singh KS
    64. Sokolova N
    65. Tang X
    66. Udwary D
    67. Vigneshwari A
    68. Vind K
    69. Vromans SPJM
    70. Waschulin V
    71. Williams SE
    72. Winter JM
    73. Witte TE
    74. Xie H
    75. Yang D
    76. Yu J
    77. Zdouc M
    78. Zhong Z
    79. Collemare J
    80. Linington RG
    81. Weber T
    82. Medema MH

    (2023) MIBiG 3.0: a community-driven effort to annotate experimentally validated biosynthetic gene clusters

    Nucleic Acids Research 51:D603–D610.

    https://doi.org/10.1093/nar/gkac1049

    • PubMed
    • Google Scholar
56. Book

    1. Thirumurugan D
    2. Cholarajan A
    3. Raja S
    4. Vijayakumar R

    (2018) An introductory chapter: secondary metabolites

    Academia.

    https://doi.org/10.5772/intechopen.79766

    • Google Scholar
57. 1. Trivedi P
    2. Leach JE
    3. Tringe SG
    4. Sa T
    5. Singh BK

    (2020) Plant-microbiome interactions: from community assembly to plant health

    Nature Reviews. Microbiology 18:607–621.

    https://doi.org/10.1038/s41579-020-0412-1

    • PubMed
    • Google Scholar
58. 1. Tsilimigras MCB
    2. Fodor AA

    (2016) Compositional data analysis of the microbiome: fundamentals, tools, and challenges

    Annals of Epidemiology 26:330–335.

    https://doi.org/10.1016/j.annepidem.2016.03.002

    • PubMed
    • Google Scholar
59. 1. Weiss S
    2. Van Treuren W
    3. Lozupone C
    4. Faust K
    5. Friedman J
    6. Deng Y
    7. Xia LC
    8. Xu ZZ
    9. Ursell L
    10. Alm EJ
    11. Birmingham A
    12. Cram JA
    13. Fuhrman JA
    14. Raes J
    15. Sun F
    16. Zhou J
    17. Knight R

    (2016) Correlation detection strategies in microbial data sets vary widely in sensitivity and precision

    The ISME Journal 10:1669–1681.

    https://doi.org/10.1038/ismej.2015.235

    • PubMed
    • Google Scholar
60. 1. Winsor GL
    2. Griffiths EJ
    3. Lo R
    4. Dhillon BK
    5. Shay JA
    6. Brinkman FSL

    (2016) Enhanced annotations and features for comparing thousands of Pseudomonas genomes in the Pseudomonas genome database

    Nucleic Acids Research 44:D646–D653.

    https://doi.org/10.1093/nar/gkv1227

    • PubMed
    • Google Scholar
61. 1. Xu Z
    2. Park TJ
    3. Cao H

    (2023) Advances in mining and expressing microbial biosynthetic gene clusters

    Critical Reviews in Microbiology 49:18–37.

    https://doi.org/10.1080/1040841X.2022.2036099

    • PubMed
    • Google Scholar
62. 1. Yee DA
    2. Niwa K
    3. Perlatti B
    4. Chen M
    5. Li Y
    6. Tang Y

    (2023) Genome mining for unknown-unknown natural products

    Nature Chemical Biology 19:633–640.

    https://doi.org/10.1038/s41589-022-01246-6

    • PubMed
    • Google Scholar
63. 1. Zengler K
    2. Palsson BO

    (2012) A road map for the development of community systems (CoSy) biology

    Nature Reviews. Microbiology 10:366–372.

    https://doi.org/10.1038/nrmicro2763

    • PubMed
    • Google Scholar
64. 1. Zierep PF
    2. Ceci AT
    3. Dobrusin I
    4. Rockwell-Kollmann SC
    5. Günther S

    (2020) SeMPI 2.0-A web server for pks and nrps predictions combined with metabolite screening in natural product databases

    Metabolites 11:13.

    https://doi.org/10.3390/metabo11010013

    • PubMed
    • Google Scholar

Author details

1. Shaohua Gu

   Center for Quantitative Biology, Academy for Advanced Interdisciplinary Studies, Peking University, Beijing, China

   Contribution

   Conceptualization, Data curation, Formal analysis, Funding acquisition, Validation, Investigation, Visualization, Methodology, Writing – original draft, Project administration, Writing – review and editing

   Contributed equally with

   Yuanzhe Shao

   Competing interests

   No competing interests declared

   "This ORCID iD identifies the author of this article:" 0000-0003-3071-5747

2. Yuanzhe Shao

   Peking-Tsinghua Center for Life Sciences, Academy for Advanced Interdisciplinary Studies, Peking University, Beijing, China

   Contribution

   Data curation, Software, Formal analysis, Investigation, Visualization, Methodology

   Contributed equally with

   Shaohua Gu

   Competing interests

   No competing interests declared

   "This ORCID iD identifies the author of this article:" 0009-0003-6739-4198

3. Karoline Rehm

   University of Zurich, Department of Chemistry, Zurich, Switzerland

   Contribution

   Resources, Data curation, Validation

   Competing interests

   No competing interests declared

4. Laurent Bigler

   University of Zurich, Department of Chemistry, Zurich, Switzerland

   Contribution

   Resources, Data curation, Validation

   Competing interests

   No competing interests declared

5. Di Zhang

   Center for Quantitative Biology, Academy for Advanced Interdisciplinary Studies, Peking University, Beijing, China

   Contribution

   Software, Formal analysis, Visualization, Methodology

   Competing interests

   No competing interests declared

6. Ruolin He

   Center for Quantitative Biology, Academy for Advanced Interdisciplinary Studies, Peking University, Beijing, China

   Contribution

   Software, Methodology

   Competing interests

   No competing interests declared

7. Ruichen Xu

   School of Life Science, Shandong University, Qingdao, China

   Contribution

   Investigation

   Competing interests

   No competing interests declared

8. Jiqi Shao

   Center for Quantitative Biology, Academy for Advanced Interdisciplinary Studies, Peking University, Beijing, China

   Contribution

   Investigation, Methodology

   Competing interests

   No competing interests declared

9. Alexandre Jousset

   Jiangsu Provincial Key Lab for Organic Solid Waste Utilization, Key Lab of Organic-based Fertilizers of China, Nanjing Agricultural University, Nanjing, China

   Contribution

   Writing – review and editing

   Competing interests

   No competing interests declared

10. Ville-Petri Friman

    University of Helsinki, Department of Microbiology, Helsinki, Finland

    Contribution

    Writing – review and editing

    Competing interests

    No competing interests declared

11. Xiaoying Bian

    Helmholtz International Lab for Anti-infectives, State Key Laboratory of Microbial Technology, Shandong University, Qingdao, China

    Contribution

    Resources, Investigation

    Competing interests

    No competing interests declared

12. Zhong Wei

    Jiangsu Provincial Key Lab for Organic Solid Waste Utilization, Key Lab of Organic-based Fertilizers of China, Nanjing Agricultural University, Nanjing, China

    Contribution

    Conceptualization, Resources, Supervision, Funding acquisition, Validation, Methodology, Project administration, Writing – review and editing

    For correspondence

    weizhong@njau.edu.cn

    Competing interests

    No competing interests declared

    "This ORCID iD identifies the author of this article:" 0000-0002-7967-4897

13. Rolf Kümmerli

    University of Zurich, Department of Quantitative Biomedicine, Zurich, Switzerland

    Contribution

    Conceptualization, Resources, Supervision, Funding acquisition, Validation, Methodology, Writing – original draft, Project administration, Writing – review and editing

    For correspondence

    rolf.kuemmerli@uzh.ch

    Competing interests

    No competing interests declared

    "This ORCID iD identifies the author of this article:" 0000-0003-4084-6679

14. Zhiyuan Li

    1. Center for Quantitative Biology, Academy for Advanced Interdisciplinary Studies, Peking University, Beijing, China
    2. Peking-Tsinghua Center for Life Sciences, Academy for Advanced Interdisciplinary Studies, Peking University, Beijing, China

    Contribution

    Conceptualization, Resources, Data curation, Software, Formal analysis, Supervision, Funding acquisition, Investigation, Methodology, Writing – original draft, Project administration, Writing – review and editing

    For correspondence

    zhiyuanli@pku.edu.cn

    Competing interests

    No competing interests declared

    "This ORCID iD identifies the author of this article:" 0000-0001-6662-2636

• 1,189

  views

• 144

  downloads

• 1

  citations

Views, downloads and citations are aggregated across all versions of this paper published by eLife.