Charting the Lipopeptidome of Nonpathogenic Pseudomonas

Abstract

A major source of pseudomonad-specialized metabolites is the nonribosomal peptide synthetases (NRPSs) assembling siderophores and lipopeptides. Cyclic lipopeptides (CLPs) of the Mycin and Peptin families are frequently associated with, but not restricted to, phytopathogenic species. We conducted an in silico analysis of the NRPSs encoded by lipopeptide biosynthetic gene clusters in nonpathogenic Pseudomonas genomes, covering 13 chemically diversified families. This global assessment of lipopeptide production capacity revealed it to be confined to the Pseudomonas fluorescens lineage, with most strains synthesizing a single type of CLP. Whereas certain lipopeptide families are specific for a taxonomic subgroup, others are found in distant groups. NRPS activation domain-guided peptide predictions enabled reliable family assignments, including identification of novel members. Focusing on the two most abundant lipopeptide families (Viscosin and Amphisin), a portion of their uncharted diversity was mapped, including characterization of two novel Amphisin family members (nepenthesin and oakridgin). Using NMR fingerprint matching, known Viscosin-family lipopeptides were identified in 15 (type) species spread across different taxonomic groups. A bifurcate genomic organization predominates among Viscosin-family producers and typifies Xantholysin-, Entolysin-, and Poaeamide-family producers but most families feature a single NRPS gene cluster embedded between cognate regulator and transporter genes. The strong correlation observed between NRPS system phylogeny and rpoD-based taxonomic affiliation indicates that much of the structural diversity is linked to speciation, providing few indications of horizontal gene transfer. The grouping of most NRPS systems in four superfamilies based on activation domain homology suggests extensive module dynamics driven by domain deletions, duplications, and exchanges. IMPORTANCE Pseudomonas species are prominent producers of lipopeptides that support proliferation in a multitude of environments and foster varied lifestyles. By genome mining of biosynthetic gene clusters (BGCs) with lipopeptide-specific organization, we mapped the global Pseudomonas lipopeptidome and linked its staggering diversity to taxonomy of the producers, belonging to different groups within the major Pseudomonas fluorescens lineage. Activation domain phylogeny of newly mined lipopeptide synthetases combined with previously characterized enzymes enabled assignment of predicted BGC products to specific lipopeptide families. In addition, novel peptide sequences were detected, showing the value of substrate specificity analysis for prioritization of BGCs for further characterization. NMR fingerprint matching proved an excellent tool to unequivocally identify multiple lipopeptides bioinformatically assigned to the Viscosin family, by far the most abundant one in Pseudomonas and with stereochemistry of all its current members elucidated. In-depth analysis of activation domains provided insight into mechanisms driving lipopeptide structural diversification.

Keywords: NMR fingerprint matching; Pseudomonas; biosynthetic gene clusters (BGCs); cyclic lipopeptides (CLPs); genome mining; lipopeptidome; nonribosomal peptide synthetases (NRPSs); phylogeny; rpoD.

FIG 1

Structural and genomic diversity of LPs in nonpathogenic Pseudomonas. (A) Generic representation of the multidomain architecture of the NRPS enzymes synthesizing the majority of cyclic and linear LPs in Pseudomonas. Biosynthesis is initiated by NRPSⁱ, continued by NRPS^m (if present), and completed by NRPS^t. Assembly starts with acylation of the first AA by the C_s-A-T module. Synthesis proceeds in a collinear fashion, matching the order of the consecutive C-A-T modules, by progressive extension of T domain-attached LP intermediates with A domain-selected residues. The linear or cyclic product (peptide length n = 1+a+b+c) is released by the thioesterase activity of the tandem TE domains. Unlike in the biosynthesis of Mycin-type CLPs, standalone NRPS or tailoring enzymes are not required. (B) Genomic architecture of the LP BGCs of different families (Table S1) with the NRPS genes flanked by one or two LuxR-family regulatory genes and three genes encoding the tripartite PleABC export system. An additional type of ATP-dependent transporter (SyrD) and an aminotransferase (Dab) for synthesis of 2,4-diaminobutyric acid substrate are encoded in the Tolaasin family BGCs. A conserved secondary transporter gene of the major facilitator superfamily (MFS) is associated with the thanafactin NRPS genes. The number of AA-specifying modules is indicated for each NRPS gene. In addition to the predominant contiguous BGC organization, a splitted cluster (represented by a dotted connector line) is typical for the Poaeamide, Entolysin, and Xantholysin families. Both types of organization occur in the Viscosin family.

FIG 2

Global distribution of LP families in nonpathogenic Pseudomonas. The rpoD cladogram comprises previously characterized LP producers and related type strains, using P. aeruginosa as outgroup. This set is complemented with the LP producers identified in this study (marked with a black or red star to distinguish monoproducers and dual producers, respectively). Type strains highlighted with bold species names represent established and predicted LP producers. No (tentative) species names are included for nontype strains. All strains belong to either the Fluorescens, Putida, or Syringae group (green, orange, and purple branches in the cladogram, respectively). The subgroups constituting these lineages are delineated with different background colors as specified in the legend panel. The circumjacent annotation represents the capacity to produce an LP of a particular family, with families (colored names) differentiated by symbols of corresponding colors. Closed symbols refer to chemically identified LPs and open symbols designate predicted LP family members. Empty dots indicate that a LP BGC is absent (strains with published genome sequences) or that no evidence for LP production is available (strains without published genome sequences). The genomic BGC organization, contiguous or bifurcated, is specified by squares and triangles, respectively. The corresponding rpoD-based phylogenetic tree with bootstrap values is provided in Fig. S8 at https://doi.org/10.48804/KKFWS9.

FIG 3

Affiliation of LP producers with Pseudomonas taxonomic (sub)groups. The demonstrated secretion of individual LPs is plotted by family-specific colored dots against an rpoD-derived phylogenetic tree of the corresponding producer strains, with the indication of the taxonomic (sub)group they belong to (bootstrapped tree in Fig. S9 at https://doi.org/10.48804/KKFWS9). All strains belong to either the Putida, Fluorescens, or Syringae group (orange, green, and purple tree branches, respectively). The subgroups constituting these lineages are delineated with different background colors as specified in the legend panel. Bold species names highlight LP-producing type strains. An empty colored dot indicates that an additional BGC of a specified family is present but that the predicted LP product awaits characterization. Selected producers of Mycin and Peptin CLPs are included for comparison.

FIG 4

Quantitative distribution of LP BGCs in genomes of nonpathogenic Pseudomonas. The 3D bars represent numbers of known and predicted LP BGCs assigned to the respective known families. Within the families, LPs are further grouped along the z axis (1 through 6) according to peptide sequence if such variants occur. Stacked bars represent chemically characterized LPs (upper segment(s)) and predicted LPs (lower segment) as indicated in the legend box. In the Viscosin family, the upper bar segments correspond to the l-Leu5-containing LPs viscosin (peptide sequence variant 1), viscosinamide (variant 2), and massetolide (variant 3); the middle bar segments show the corresponding d-Leu5 stereoisomers WLIP (variant 1), pseudodesmin (variant 2), and pseudophomin (variant 3). Respectively, six and four sequence subtypes are shown for the Amphisin family (variants 1 to 6: lokisin/anikasin; milkisin/stechlisin/tensin; arthrofactin; amphisin; oakridgin; nepenthesin) and the Bananamide family (variants 1 to 4: bananamide A–C; bananamide D–G; MDN-0066; prosekin). Three sequence subtypes are shown for the Tolaasin family (variants 1 to 3: tolaasin I, sessilin, tolaasin F). The Poaeamide family comprises the sequence variants PPZPM (1) and poaeamide A (2). Minor congeners of a particular peptide sequence variant, if any, may be produced due to relaxed substrate specificity of a module, but these are not specified. The corresponding producer strains and accession numbers are compiled in Table S4.

FIG 5

Diversity of NRPS systems in the Viscosin family. The NRPS-based cladogram comprises LP systems encoded by known and predicted Viscosin-family BGCs retrieved from (putative) LP producers. Type strains are highlighted with bold species names. Strain numbers are provided for all other isolates. The taxonomic (sub)group affiliations of the strains are specified with different background colors. The branches of the Putida G and Fluorescens G are highlighted in orange and green, respectively. Clusters of strains with NRPS concatenates that share at least 90% AA-identity are marked with a dark gray dot. The circular annotation of strains represents the capacity, either known or predicted, to produce a particular LP (specified in the legend panel). Different symbols are used for LPs containing l-Leu5 (squares) or d-Leu5 (triangles). The strains marked with a black strip carry a contiguous genomic BGC organization. A gray strip refers to such organization predicted by concatenation of two contigs, inferred from synteny with an orthologous system. The LP systems characterized in this study are marked with a black star. Strains that have the capacity to coproduce tolaasin (I or F) are highlighted with a red star. The corresponding tree inferred from an alignment of concatenated NRPS sequences, with bootstrap values added, is provided in Fig. S10 at https://doi.org/10.48804/KKFWS9.

FIG 6

Relationships within and between LP families of nonpathogenic Pseudomonas revealed by A domain homology. Each A domain is represented by a box specifying its position in the NRPS (between 1 and 18) and selected AA. The resulting AA sequence(s) represent(s) the peptide sequence(s) found in the respective LPs (names shown on the left), assembled by collinear activity of the corresponding C-A-T modules in two or three consecutively acting NRPS enzymes (Fig. 1). A domains from modules present in different NRPSs constituting an LP system are separated by dots. The corresponding NRPS genes can be part of a contiguous BGC (marked with black/gray filled dots) or the first NRPS gene can occur separately (bifurcated BGC with split marked by an additional black/gray open dot). For systems with predicted LPs, dash-lined boxes and gray dots are used. To distinguish among domains that activate the same AA but belong to different homology clusters, box-filling colors are differentiated as defined in the respective phylogenetic trees for substrate AAs (Fig. S12). Boxes of A domains with relaxed substrate specificity are additionally labeled (plus sign, Ile-Val-Leu; diamond, Asp-Glu; triangle, Hse-Gly). AAs that occur only in the Tolaasin family are shown in a different color on a white background, except for Dhb (phylogenetically poorly separable from Thr-specific domains). Chemical families comprising multiple LPs are delineated with solid right brackets. Dashed right brackets with encircled numbers delineate four major superfamilies encompassing known and predicted LP systems sharing largely overlapping A domain homology profiles. LPs for which in this study the BGC of a previously reported producer strain is identified or the BGC of a novel producer strain/species is described, are shown with names in orange and green, respectively. LPs first described in this study are highlighted in pink (prediction validated by structure elucidation) or blue (prediction awaiting validation). A single A domain profile is used to represent a particular LP that is synthesized by orthologous NRPS systems in related species. Conversely, to indicate the involvement of sequence-diversified NRPS systems in biosynthesis of a particular LP by taxonomically distant producers, these are shown separately (pseudodesmin, WLIP). In the Amphisin family, peptide sequences that only differ by stereochemistry are combined (anikasin/lokisin; tensin[stechlisin]/milkisin). The stereochemical differentiation among members of the Viscosin family (by Leu5) and between Tolaasin CLPs (by Val9) is highlighted by using yellow font for the l-configured AAs. Colored vertical bars delineate the taxonomic (sub)group affiliation of the LP producers (specified in the taxonomic clade legend). Putative events evolutionarily linking certain known or predicted LPs (discussed in the text) are marked with red symbol-labeled dashed arrows representing module/domain addition (α), exchange (ε) or deletion (δ).