The Xenorhabdus nematophila LrhA transcriptional regulator modulates production of γ-keto-N-acyl amides with inhibitory activity against mutualistic host nematode egg hatching

ABSTRACT

Xenorhabdus nematophila is a symbiotic Gammaproteobacterium that produces diverse natural products that facilitate mutualistic and pathogenic interactions in their nematode and insect hosts, respectively. The interplay between X. nematophila secondary metabolism and symbiosis stage is tuned by various global regulators. An example of such a regulator is the LysR-type protein transcription factor LrhA, which regulates amino acid metabolism and is necessary for virulence in insects and normal nematode progeny production. Here, we utilized comparative metabolomics and molecular networking to identify small molecule factors regulated by LrhA and characterized a rare γ-ketoacid (GKA) and two new N-acyl amides, GKA-Arg (1) and GKA-Pro (2) which harbor a γ-keto acyl appendage. A lrhA null mutant produced elevated levels of compound 1 and reduced levels of compound 2 relative to wild type. N-acyl amides 1 and 2 were shown to be selective agonists for the human G-protein-coupled receptors (GPCRs) C3AR1 and CHRM2, respectively. The CHRM2 agonist 2 deleteriously affected the hatch rate and length of Steinernema nematodes. This work further highlights the utility of exploiting regulators of host-bacteria interactions for the identification of the bioactive small molecule signals that they control.

IMPORTANCE

Xenorhabdus bacteria are of interest due to their symbiotic relationship with Steinernema nematodes and their ability to produce a variety of natural bioactive compounds. Despite their importance, the regulatory hierarchy connecting specific natural products and their regulators is poorly understood. In this study, comparative metabolomic profiling was utilized to identify the secondary metabolites modulated by the X. nematophila global regulator LrhA. This analysis led to the discovery of three metabolites, including an N-acyl amide that inhibited the egg hatching rate and length of Steinernema carpocapsae nematodes. These findings support the notion that X. nematophila LrhA influences the symbiosis between X. nematophila and S. carpocapsae through N-acyl amide signaling. A deeper understanding of the regulatory hierarchy of these natural products could contribute to a better comprehension of the symbiotic relationship between X. nematophila and S. carpocapsae.

INTRODUCTION

Xenorhabdus nematophila is a Gram-negative bacterial pathogen that forms a mutualistic symbiosis with entomopathogenic nematodes from the Steinernema genus (1, 2). When a Steinernema infective juvenile nematode invades the hemocoel of an insect host, it releases X. nematophila (3). The bacterium then produces a collection of factors that suppress the insect immune system, kill the insect host, and hinder the growth of competing microbes. The bacteria increase in biomass by consuming the insect tissue, and the entomopathogenic host nematode consumes the bacteria as its primary food source (4, 5). Following consumption of the insect cadaver by X. nematophila and Steinernema, the two re-associate and repeat the cycle (6).

The ability of X. nematophila to kill its insect host and outcompete microbes in its niche alludes to its reputation as a prolific producer of antibiotic and insecticidal secondary metabolites, among other bioactive molecules (7–9). Genome sequencing of Xenorhabdus isolates suggests that at least 6.5% of their genomes are dedicated to natural product biosynthesis (10); however, the complex mutualistic and pathogenic interactions that X. nematophila maintains with its nematode and insect hosts, respectively, require strict regulation of its secondary metabolism (11). Although a diversity of transcriptional and translational regulators exists in Xenorhabdus spp., three global regulators play central roles in multiple phases of the X. nematophila life cycle: the two-component system CpxR and CpxA, the leucine-responsive regulatory protein Lrp, and the LysR-type transcriptional regulator LrhA (11, 12).

Lrp and CpxR regulate various aspects of X. nematophila metabolism including motility, antibiotic production, and production of a lipase that contributes to nematode reproduction (12–17). Moreover, lrp, cpxR, and lrhA null mutants each exhibit defects in the killing of Manduca sexta insects (17–20). lrhA is positively regulated by both Lrp and CpxR (12), and ∆cpxR-associated virulence defects originate from transcriptional alterations of lrhA (19). lrhA mutants have the most severely attenuated virulence against M. sexta, killing fewer than 10% of insects compared with the wild type’s propensity to kill 90% to 100% of insects (20). A similar virulence attenuation has been shown in the mutants of a LrhA homolog in Photorhabdus luminescens, HexA (21). HexA is a global regulator of secondary metabolism in P. luminescens that plays key roles in bacterial pathogenicity, nematode symbiosis, and phenotypic variation between P. luminescens forms (22, 23). HexA-mediated repression of secondary metabolite production has been shown to result from the small regulatory RNA, ArcZ, binding to the mRNA encoding HexA (24). In contrast to HexA, the underlying mechanism for virulence attenuation in the X. nematophila lrhA mutant is unknown, but it may be related to growth defects (5). The lrhA mutant is unable to grow in minimal medium without supplementation with certain amino acids: a mixture of aspartate, glutamate, and leucine or pools containing aspartate-derived, aromatic, or branched-chain amino acids (5). Furthermore, the transcriptome of the lrhA mutant is dramatically different from that of wild type. Of the total average open reading frame transcripts detected by microarray, 10.6% were differentially abundant in the lrhA mutant relative to wild type, and of these, 18.9% were categorized as metabolism-related activities (5). Notably, the lrhA mutant exhibits a higher abundance of transcripts related to amino acid metabolism relative to the wild type strain, leading to the idea that LrhA directly or indirectly represses this process. LrhA had a unique effect on the glycine, serine, and threonine pathways, and both the sigma factor RpoS and LrhA have parallel impacts on transcript abundance of genes within the alanine, aspartate, glutamate, and leucine pathways (5). This is consistent with the insect circulatory fluid being an amino acid-rich environment (25). Given the strong influence of X. nematophila LrhA on both insect virulence and amino acid metabolism, we predicted that this regulator would have an impact on secondary metabolism.

In this work, we applied high-resolution mass spectrometry-based molecular networking and comparative metabolomic analysis between wild type X. nematophila (HGB800) and an isogenic lrhA mutant (∆lrhA, HGB1320) to evaluate the effect of the lrhA gene on secondary metabolism. We identified several families of known, biologically active small molecules and utilized chemical synthesis, high-performance liquid chromatography (HPLC)-guided isolation, isotope labeling, tandem mass spectrometry (MS/MS), and NMR spectroscopy to characterize two previously unreported N-acyl amide metabolites that displayed differential production in the wild type and ∆lrhA strains and featured a unique γ-ketoacyl functionality. Each N-acyl amide selectively activated a distinct human GPCR. The N-acyl amide with reduced levels in the ∆lrhA mutant deleteriously affected the hatching fidelity and length of Steinernema nematodes. Collectively, these studies further illustrate how the study of microbial regulators responsible for modulating host-bacteria interactions can uncover new host signaling responses.

RESULTS

X. nematophila LrhA impacts secondary metabolite biosynthesis

To investigate the role of LrhA on X. nematophila secondary metabolism, we performed high-resolution mass spectrometry (MS)-based comparative metabolomics on the culture supernatants of wild type and a ∆lrhA mutant (HGB1320; ΔlrhA) (16). Analysis of the metabolome data sets through the XCMS Online cloud server (26) conservatively revealed over 3300 prominent features (with a robust MS intensity >100,000 counts) between the wild type and ∆lrhA sample sets with 408 (12.4%) statistically significant and differentially regulated molecular features (fold change >2 and P-value < 0.01, Fig. 1A). Application of tandem MS on the significant features and MS/MS-based molecular networking through the Global Natural Products Social Molecular Networking (GNPS) platform (27) identified differentially regulated molecular families, including the leupeptins (28), xenematides (29), xenortides (30), xenocoumacins (31), and nematophin (32) (Fig. 1B through D). With the exception of the leupeptin protease inhibitor family, these molecular families exhibit modest antimicrobial activity and weak-to-negligible insecticidal activity (30–32). Overall, we found that the ∆lrhA mutant produced higher amounts of these compounds relative to wild type, indicating that LrhA negatively influences their production. This influence is unlikely to be a direct result of LrhA-dependent transcriptional regulation of the biosynthetic loci for these metabolites, as these loci were not detected among the differentially transcribed genes when comparing wild type with ∆lrhA (5).

Fig 1.

(A) Volcano plot analysis of the differential metabolomics experiment on ∆lrhA and wild type X. nematophila strains. Red dots represent mass features that had both a P-value below 0.01 and an absolute fold-change greater than two in the ∆lrhA strain compared with the wild type. Insert shows an enhanced view of a relevant section of the plot highlighting ions upregulated in the ∆lrhA strain. Leupeptin, xenortide, xenematide, xenocoumacin, nematophin, and N-acyl amide-related features are shown as dark-blue, yellow, light-blue, light-green, purple, and dark-green dots, respectively. (B) MS/MS-based molecular networking of structurally related compounds upregulated in the ∆lrhA strain. (C) Chemical structures of known compounds identified through MS/MS-based molecular networking. (D) Production of compounds identified via MS/MS-based molecular networking in the ∆lrhA and wild type strains. (E) Production of compounds identified via MS/MS-based molecular networking in the ∆lrhA (red) and wild type (black) strains, in the presence or absence of exogeneous leucine (10 g/L; predicted Lrp modulating conditions). Statistical analyses were performed through an unpaired two-tailed t-test; n.s. indicates a non-significant difference, * indicates a difference with a P-value less than 0.05, ** indicates a difference with a P-value less than 0.01, and *** indicates a difference with a P-value less than 0.001.

The negative impact of LrhA on secondary metabolism observed here is in contrast to another transcription factor, Lrp (leucine-responsive regulatory protein), whose expression, when induced from a tunable promoter, positively influences the production of xenematide A, xenortide A, and xenocoumacin 1, although it does have a negative impact on the abundance of nematophin (12). The inverse impact of Lrp and LrhA on these metabolites was unexpected, since Lrp positively regulates expression of lrhA (20) and, like LrhA, Lrp regulation has a profound impact on metabolism, with 13.6% of differentially expressed genes (comparing wild type and ∆lrp) being categorized as metabolic (5). In Escherichia coli, exogenous leucine supplementation affects Lrp multimerization and influences regulon gene expression (14, 15, 33–35). We therefore considered the possibility that exogenous supplementation with leucine might modulate Lrp activity in Xenorhabdus and in turn the production of secondary metabolites. Overall, leucine supplementation (10 g/L) significantly reduced the production of most antibiotics by the tested strains (Fig. 1E). With the exception of leupeptin, leucine supplementation significantly reduced antibiotic production in the wild type. Both wild type and the ∆lrhA strain had significantly lower production of xenematide and nematophin upon supplementation with leucine, although the fold difference exhibited by ∆lrhA was relatively modest. However, leucine supplementation of the ∆lrhA strain failed to significantly affect the production of the leupeptins, xenocoumacin 1, and xenortides, indicating that LrhA may be involved in a leucine-dependent signal transduction cascade influencing the production of these antibiotics.

X. nematophila produces γ-keto-N-acyl amides that are modulated by LrhA

Through the GNPS platform, we were able to compare the MS/MS profiles of the differentially regulated features against public MS² databases. Although this enabled us to identify known X. nematophila metabolites, like the leupeptins and xenematides, it also suggested the presence of two seemingly unreported N-acyl amide-type compounds that were differentially regulated between the wild type and ∆lrhA strains. Notably, one of these N-acyl amides (1, [M + H]⁺ experimental m/z 399.2977, [M + H]⁺ calc’d m/z 399.2966, C₂₀H₃₈N₄O₄) was more abundant in the ∆lrhA strain, whereas the other (2, [M + H]⁺ experimental m/z 340.2474, [M + H]⁺ calc’d m/z 340.2482, C₁₉H₃₃NO₄) was more abundant in the wild type strain (Fig. 2A and 3A). This suggests that LrhA negatively impacts the production of N-acyl amide 1 and positively impacts the production of N-acyl amide 2 under the conditions tested. Interestingly, one of the most conserved gene cluster families across Xenorhabdus spp. encodes for N-acyl amide tripeptides termed the bovienimides; however, these compounds were not differentially regulated between the wild type and ∆lrhA strains (9, 36).

Fig 2.

(A) Production of m/z 399.2966 (1) in X. nematophila strains. (B) LC-MS/MS analysis of 1 and its fragment ions. (C) LC-MS/MS analyses of unlabeled (m/z 399.2966) and labeled (m/z 405.3167) 1 in the ∆lrhA strain supplemented with ¹³C-L-Arginine. (D) m/z values corresponding to labeled and unlabeled 1 and its fragments. (E) ¹H-¹H COSY NMR correlations of natural 1. HMBC NMR correlations are shown for the synthetic 1. (F) Synthetic route toward 1. (G) Retention time analyses of natural 1, synthetic 1, and a co-injection of natural and synthetic 1. Left y-axis represents counts for synthetic and co-injection traces, whereas the right y-axis represents counts for the natural trace.

Fig 3.

(A) Production of m/z 340.2482 (2) in X. nematophila strains. (B) LC-MS/MS analysis of 2 and its fragment ions. (C) m/z values corresponding to compound 2 and its fragments. (D) Synthetic route toward compound 2. (E) Retention time analyses of natural 2, synthetic 2, and a co-injection of natural and synthetic 2. (F) Chiral functionalization analysis of natural γ-keto N-acyl proline and synthetic L-(2) and D-(2′) γ-keto N-acyl proline.

Given the differential production of 1 and 2 between the ∆lrhA and wild type strains, and the diverse biological functions of animal and microbial N-acyl amides (37–45), we proceeded to structurally characterize these metabolites. Tandem MS/MS analysis, ¹³C-L-arginine isotope feeding studies, and Marfey’s analysis confirmed that 1 contained an L-arginine motif (Fig. 2B through D; Fig. S1). We isolated 1 (~0.1 mg) from a large-scale cultivation of the ∆lrhA strain (18 L, lysogeny broth, LB). Structural analysis of 1 by 1D-(¹H) and 2D-NMR (gCOSY, gHSQCAD, and gHMBCAD) supported the presence of a rare γ-keto-amide (Fig. S2 to S5; Table S1). Taken together, the MS and NMR data supported 1 to be an L-arginine N-acylated with a C₁₄ acyl chain comprising a γ-keto-acyl functional group (Fig. 2E). To verify the proposed structure, we synthesized 1 (Fig. 2F). Briefly, we performed a Grignard-based alkylation of succinic anhydride followed by a BOP-coupling to H-L-Arg(Pbf)-OMe to afford a protected analog of 1; sequential acid- and base-mediated deprotection of the analog yielded 1. The synthetic material shared the same retention time (Fig. 2G), tandem MS spectra (Fig. S1), and NMR spectra (Fig. S2) with the natural material.

Similar to 1, tandem MS analysis of 2 also exhibited diagnostic signs of a C₁₄ acyl chain with a γ-keto-amide functional group but contained a feature consistent with proline, rather than arginine (Fig. 3B and C). We hypothesized that 2 was the proline derivative of 1; thus, we also synthesized this analog (Fig. 3D) and confirmed its structure via LC-MS and NMR spectroscopy (Fig. 3E; Fig. S6 and S7). To determine whether X. nematophila utilized L- or D-proline in the biosynthesis of 2, we synthesized both L-(2) and D-(2′) γ-keto N-acyl proline and performed chiral derivatization with the chiral auxiliary agent (R)−1-(9-anthryl)−2,2,2-trifluoroethanol, as previously described (46) (Fig. S8). We compared these functionalized standards with the metabolite in wild type X. nematophila extracts—similarly functionalized—through LC-MS and determined that the naturally occurring compound has an L amino acid (2) (Fig. 3F).

γ-Keto N-acyl amino acids are derived from γ-ketomyristic acid

To our knowledge, γ-keto N-acyl amino acid metabolites have not been described to date. We hypothesized that compounds 1 and 2 could be derived from three candidate fatty acid substrates: γ-ketomyristic acid (GKA), α,β-unsaturated myristic acid (ABUA), or β,γ-unsaturated myristic acid (BGUA). GKA could be directly condensed with L-arginine and L-proline to form 1 and 2, respectively, whereas ABUA and BGUA would require hydration and oxidation at the 3′ position of the acyl chain before or after condensation with L-arginine or L-proline. Metabolomic profiling and chemical synthesis of a standard revealed that only GKA was detected in culture extracts of the X. nematophila strains (Fig. 4A and B; Fig. S11). Notably, although the presence of GKA has been proposed via gas chromatography-MS in cheese (47) and in a mummy (48), this study unambiguously supports its structure and existence.

Fig 4.

(A) Production of biosynthetic precursor GKA in X. nematophila strains. (B) Retention time analysis of natural GKA from ∆lrhA and synthetic GKA from ∆lrhA supplemented with synthetic GKA. Left y-axis represents the counts for the co-injection of synthetic GKA in ∆lrhA extracts and the right y-axis represents the counts for natural GKA. (C) Production of 1 in X. nematophila wild type strain upon acid supplementation. (D) Production of 1 in X. nematophila ∆lrhA strain upon acid supplementation. (E) Chemical structures of compounds GKA, ABUA, and BGUA.

To study the effect of GKA, ABUA, and BGUA fatty acids on the production of the major γ-keto N-acyl amino acid metabolite 1, we individually supplemented ΔlrhA and wild type cultures with 100 µM synthetic standards of GKA, ABUA, and BGUA. Although GKA supplementation modestly upregulated the biosynthesis of 1 in wild type X. nematophila, it drastically increased the levels of 1 by over 100-fold in the ΔlrhA strain (Fig. 4C and D). Interestingly, we also observed a GKA-lysine analog in all strains by LC-MS; however, the production of this compound was completely dependent on GKA supplementation (Fig. S12). In contrast to GKA supplementation, ABUA and BGUA were unable to significantly increase the production of 1 in any strain, suggesting that they are not preferred substrates for the biosynthesis of 1. In contrast, we observed that ABUA supplementation produced the proposed alternative acyl-arginine derivative 1A in wild type and ΔlrhA strains (Fig. S11). BGUA supplementation led to the production of 1B in wild type and ΔlrhA strains, but it also led to the minor production of 1A, presumably due to stereoisomerization from β,γ-unsaturation to α,β-unsaturation (Fig. S11). These results support that the wild type and ΔlrhA Xenorhabdus strains accept GKA as a natural substrate toward γ-keto N-acyl arginine biosynthesis, that the endogenous ligation enzymes are promiscuous, producing unnatural unsaturated acyl-arginine derivatives 1A and 1B upon ABUA and BGUA supplementation, and that their formation is elevated in the ΔlrhA mutant.

γ-Keto N-acyl arginine 1 and γ-keto N-acyl proline 2 are GPCR ligands and differentially affect nematode physiology

The assorted bioactivities of N-acyl amides are well-documented in the literature and include phenomena such as antibiotic activity, immune regulation, quorum sensing, and GPCR-mediated signaling (39–42, 45). To begin probing the bioactivities of 1 and 2, we performed PRESTO-Tango screening against 314 distinct human GPCRs with initial 100 µM doses of these compounds (Fig. 5A and B) (49). This screen revealed that 1 (GKA-Arg) and 2 (GKA-Pro) exhibited activation of the GPCRs C3AR1 and CHRM2, respectively. Subsequent dose-response analysis of 1 and 2 on C3AR1 and CHRM2, respectively, revealed that 1 began to activate C3AR1 at doses as low as 100 nM and above while 2 could only significantly activate CHRM2 at the highest tested dose of 100 µM (Fig. S13). Notably, C3AR1 binds anaphylatoxin C3a, which initiates a cascade that is involved in various pro-inflammatory responses including vasodilation, chemotaxis, histamine release from mast cells, and bacterial opsonization (50, 51), whereas CHRM2 is involved with the regulation of acetylcholine release and has been implicated in various functions including heart rate modulation and higher order cognitive processing of humans (52–54). In the nematode Caenorhabditis elegans, acetylcholine release is associated with inhibition of egg-laying (55). Furthermore, studies have shown an association between the CHRM2 gene and schistosomiasis: an acute parasitic nematode disease that can lead to liver damage, kidney failure, and cancer (56).

Fig 5.

(A) Heatmap depicting the activation of 314 GPCRs by compounds 1 and 2 (100 µM) as analyzed by PRESTO-Tango (49, 57). A subset of labels are shown. Active hits correspond to C3AR1 (compound 1) and CHRM2 (compound 2). Color intensity represents the log₂ (activation fold-change) over a DMSO vehicle control. (B) Zoomed-in image of the PRESTO-Tango heatmap highlighting GPCRs that are activated by either compound 1 or 2. (C) S. carpocapsae egg counts on day 0 to ensure consistency between untreated and compound 2-treated wells. (D) Ratio of hatched S. carpocapsae eggs to total eggs on day 4. (E) Length of hatched S. carpocapsae nematodes on day 4. Statistical analyses performed through 1-way ANOVA with multiple comparisons; n.s. indicates a non-significant difference, * indicates a difference with a P-value less than 0.05, ** indicates a difference with a P-value less than 0.01, and *** indicates a difference with a P-value less than 0.001.

Because 1 and 2 could modulate animal signal transduction, we further tested the biological effects of these compounds on the growth of various nematodes. Treatment of S. carpocapsae, the mutualistic nematode host of X. nematophila, with compound 2 significantly reduced both the hatching fidelity (Fig. 5D) and overall length of the nematodes (Fig. 5E). Supplementary studies indicate that compound 2 may affect the hatching fidelity of C. elegans and length of Steinernema anatoliense nematodes as well (Fig. S14). Notably, treating nematode eggs with compound 1 did not affect their hatching fidelity or nematode length, regardless of species.

DISCUSSION

It is widely accepted that the vast biosynthetic potential encoded in bacterial genomes is incommensurate with the number of known bacterial small molecules. As such, tools and strategies that manipulate bacterial secondary metabolism to activate the production of low-abundance metabolites are becoming vital to the continued discovery of new natural products (58–62). Here, through studying the role of the LrhA transcriptional regulator on X. nematophila secondary metabolism, we were able to identify a rare γ-ketoacid and two previously unreported N-acyl amides, GKA-Arg and GKA-Pro. These N-acyl amides are derived from arginine or proline acylated with the unusual γ-keto myristic acid moiety, activate distinct human GPCRs, and reduce the hatching fidelity and length of S. carpocapsae nematodes.

We show that the transcription factor LrhA regulates the production of various X. nematophila metabolites. Notably, the relative insensitivity of the ∆lrhA mutant, relative to wild type, to the impact of leucine on production of secondary metabolites suggests that LrhA may help transduce leucine as a signal toward regulation of secondary metabolite production. In X. nematophila, Lrp is a global regulator necessary for mutualistic colonization and supporting fitness of the nematode host, as well as killing and degradation of the insect host. An Lrp-leucine complex may activate transcription of LrhA, which in turn negatively impacts production of secondary metabolites. We propose that the effect of LrhA on X. nematophila secondary metabolism is likely indirect, since existing microarray data do not indicate a direct influence on biosynthetic gene clusters for the metabolites observed in this study (5). Instead, it may be due to altered metabolite availability caused by the global impact of LrhA on amino acid and sugar metabolism.

Given the differential production of compounds 1 and 2 between ∆lrhA and wild type X. nematophila strains, these findings may indicate that under conditions in which LrhA is inactive, X. nematophila negatively impacts the physiology of its mutualistic nematode host by production of a specific γ-keto N-acyl amide signaling molecule. This signaling could possibly be through interaction with a GPCR receptor, since PRESTO-TANGO analysis of the arginine and proline N-acyl amides found that the compounds are agonists for C3AR1 and CHRM2, respectively. Interestingly, C. elegans nematodes, predominantly males, secrete N-acyl glutamine, which can accelerate development and shorten the lifespan of the animal (37). Our findings raise the intriguing possibility that the bacterial mutualist of a nematode has co-opted a signaling pathway to modulate egg hatching and development of its nematode host.

MATERIALS AND METHODS

General methods

Commercially available cloning strains DH5α and DH10B were purchased from New England Biolabs (NEB) and Thermo Fisher Scientific for routine cloning experiments. Primers and oligonucleotides were ordered from the Keck Oligonucleotide Synthesis facility at the Yale School of Medicine. Sequencing experiments were performed at the Keck DNA Sequencing facility at the Yale School of Medicine. All PCR reactions were performed using the Q5 High-Fidelity DNA Polymerase (NEB) according to the manufacturer’s instructions with the inclusion of 5% (vol/vol) DMSO. Thermal cycling was performed on a C1000 Touch thermal cycler. Traditional molecular biology cloning, Gibson assembly, and QuickChange site-directed mutagenesis techniques were employed for routine cloning and mutagenesis. Agarose gel electrophoresis was performed using 0.7% agarose gels stained with GelGreen Nucleic Acid Gel Stain (Biotium) in tris-acetate-EDTA buffer. Gel purifications and PCR cleanups were performed using the NucleoSpin Gel and PCR Clean-up kits (Macherey-Nagel) according to the manufacturer’s protocols. Cloned plasmids and constructs were purified using the QIAprep Spin Miniprep Kit according to the manufacturer’s protocol. Ligations were performed at a 1:3 vector-to-insert ratio using T4 DNA ligase (NEB) according to the manufacturer’s protocol. A BTX Gemini X2 HT was used for electroporation with the electroporation parameters set at 1800 V, 25 µF, and 200 Ω. Electroporation cuvettes had a gap width of 1 mm. Sodium dodecyl sulfate-polyacrylamide gel electrophoresis (SDS-PAGE) analyses were performed using 4–15% Mini-PROTEAN TGX Precast Protein Gels (Bio-Rad) in Tris-glycine running buffer. Liquid chromatography/high-resolution quadrupole time-of-flight electrospray ionization mass spectrometry (LC-HR-ESI-QTOF-MS) analyses were performed on an Agilent iFunnel 6550 QTOF MS fitted with a Dual Agilent Jet Stream (AJS) ESI source coupled to an Agilent 1290 Infinity HPLC system. A Kinetex C₁₈ 1.7 µm 100 Å 100 × 2.1 mm (Part Number 00D-4475-AN) ultra-high-performance liquid chromatography column was used. General HR-ESI-QTOF-MS experiments employed a method comprising a 10 µL sample injection at 25°C, 0.7 mL/min flow rate with a H₂O:MeCN gradient solvent system containing 0.1% formic acid: 0–30 min, 5%-98% acetonitrile (MeCN); 30–35 min, 98% MeCN; 35–37 min, 98%-5% MeCN; and 37–42 min equilibration, 5% MeCN. Mass spectra were acquired in the range of 25–3200 m/z in positive ion mode. Collected data were analyzed using Agilent MassHunter Qualitative Analysis Software (Version B.06.00, Agilent Technologies). Lower resolution LC-MS analyses were performed on an Agilent 1260 Infinity system using a Phenomenex Luna 5 µm C₁₈ (2) 100 Å 4.6 × 250 mm (Phenomenex, CA, USA) column and a PDA detector, coupled with a quadrupole electrospray ionization mass spectrometry instrument (Agilent 6120). General LC-MS experiments employed a method comprising a 10 µL sample injection at 25°C, 0.7 mL/min flow rate with a H₂O:MeCN gradient solvent system containing 0.1% formic acid: 0–30 min, 10%–100% MeCN; 30–35 min, 100% MeCN; 35–37 min, 100%–10% MeCN; and 37–42 min equilibration, 10% MeCN.

Bacterial strains and growth conditions

E. coli and Xenorhabdus nematophila strains were maintained on lysogeny broth (LB) agar plates [BD; 1% (wt/vol) tryptone, 0.5% (wt/vol) yeast extract, 1% (wt/vol) NaCl, and 1.5% (wt/vol) agar] at 30°C (for X. nematophila) or 37°C (for E. coli) in a standing incubator. Liquid cultures were maintained in LB medium [BD; 1% (wt/vol) tryptone, 0.5% (wt/vol) yeast extract, and 1% (wt/vol) NaCl (for metabolomic experiments, proline and leucine were also supplemented; see “Comparative Metabolomics Analysis” section for more details)] at 250 rpm in a shaking incubator. Antibiotics were purchased from American Bioanalytical. Ampicillin, streptomycin, spectinomycin, chloramphenicol, and kanamycin antibiotics were used at concentrations of 100 µg/mL, 100 µg/mL, 100 µg/mL, 34 µg/mL, and 100 µg/mL, respectively, unless otherwise noted. X. nematophila wild type (HGB800, ATCC 19061) and ΔlrhA (HGB1320) (16) strains are from the Goodrich-Blair lab stocks.

Comparative metabolomics analysis

We performed comparative metabolomics, according to a published metabolomics protocol (63), to identify metabolites that are regulated by the ΔlrhA strain. Wild type (HGB800) and ΔlrhA (HGB1320) Xenorhabdus nematophila strains were grown on LB agar plates at 30°C for 48 h. Colonies were picked and grown overnight under aerobic conditions in 5 mL of LB at 30°C and 250 rpm. Overnight cultures were centrifuged and resuspended in an equal volume of M9 medium (without glucose). Five milliliters of fresh M9 minimal medium (MP-Bio), without glucose, supplemented with specifically 10 g/L casamino acids and 10 g/L L-proline (M9PCA-G medium), to mimic the hemolymph environment, were inoculated at 1:200 dilution with the overnight culture and grown at 30°C and 250 rpm for 48 h (for L-leucine feeding experiments, 10 g/L of L-proline and 10 g/L of L-leucine were used as the carbon source). After growth, formic acid was added to a net concentration of 0.1%, the culture was centrifuged, and the cleared supernatant was dried under reduced pressure. The supernatant concentrates were resuspended in 200 µL methanol (MeOH) and centrifuged, and the soluble MeOH extracts were analyzed by HR-ESI-QTOF-MS. Experiments were performed in triplicate. Comparative metabolomics analysis was performed using the XCMS-online platform (xcmsonline.scripps.edu) (26). Compounds that significantly differed between strains (fold-change >2, and P-value < 0.01) were used to generate an inclusion list for pathway-targeted MS² analysis. MS² library searching and molecular networking were accomplished through the GNPS web platform (gnps.ucsd.edu).

¹³C isotope-labeling experiments

We performed ¹³C isotope-labeling experiments by supplementing the compound 1 overproducing strain ΔlrhA with ¹³C-L-arginine. ΔlrhA colonies were picked and grown overnight in 2 mL of LB at 30°C and 250 rpm. Two milliliters of LB, with or without 1 g/L ¹³C-L-arginine, were inoculated at 1:100 dilution with the overnight culture and grown at 30°C and 250 rpm for 48 h. After growth, formic acid was added to a net concentration of 0.1%, the culture was centrifuged, and the cleared supernatant was dried under reduced pressure. The supernatant concentrates were resuspended in 160 µL MeOH and centrifuged, and the soluble MeOH extracts were analyzed by HR-ESI-QTOF-MS. Experiments were performed in triplicate.

Purification, isolation, and structural characterization of metabolite 1

Large-scale cultures of ΔlrhA were employed to isolate and structurally characterize 1. ΔlrhA colonies were picked and grown overnight in 90 mL of LB at 30°C and 250 rpm. Overnight cultures were centrifuged and resuspended in an equal volume of M9PCA-G medium. In total, 18 × 1 L of M9PCA-G medium was inoculated at 1:200 dilution with the overnight culture and grown at 30°C and 250 rpm for 48 h. After growth, formic acid was added to a net concentration of 0.1%, the culture was centrifuged, and the cleared supernatant was dried under reduced pressure. The dried supernatant was extracted with 2 × 1 L of MeOH, the salts were filtered off with a frit funnel, and the MeOH extracts were dried under reduced pressure. The crude MeOH extract was subjected to manual flash C₁₈ column chromatography (200 g) with a step gradient elution (0%, 20%, 40%, 60%, 80%, and 100% MeOH in water, supplemented with 0.1% formic acid). The 80% and 100% MeOH fractions were dried under reduced pressure and further fractionated on an Agilent Prepstar HPLC system using an Agilent Polaris 5 µm C₁₈-A 250 × 21.2 mm column with linear gradient elution (20%–80% MeCN in water supplemented with 0.1% formic acid over 60 min, 8 mL/min, 1 min fraction collection window). Fractions 37–39, containing compound 1, were combined and dried under reduced pressure and further purified on a Phenomenex Luna 10 µm C₈(2) 100 Å 10 × 250 mm column with linear gradient elution (30%–60% MeCN in water supplemented with 0.1% formic acid over 60 min, 2 mL/min, 1 min fraction collection window). Fraction 23 containing compound 1 underwent a final purification using a Kinetex 5 µm Phenyl-Hexyl 100 Å 250 × 4.6 mm column with linear gradient elution (30%–60% MeCN in water supplemented with 0.1% formic acid over 60 min, 2 mL/min, 1 min fraction collection window) to yield 0.2 mg of natural 1 with minor impurities.

Acid precursor feeding in Xenorhabdus nematophila strains

The utility of GKA, ABUA, and BGUA as substrates for compound 1 production was studied in the wild type and ΔlrhA strains. Strains were picked and grown overnight in 1 mL of LB at 30°C and 250 rpm. One milliliter of fresh LB, with or without 0.5 mM synthetic GKA, ABUA, or BGUA substrates, was inoculated at 1:100 dilution with the overnight cultures and grown at 30°C and 250 rpm for 48 h. Controls of media alone, without cells, and with and without acid substrates, were included. After growth, formic acid was added to a net concentration of 0.1%, the culture was centrifuged, and the cleared supernatant was dried under reduced pressure. The supernatant concentrates were resuspended in 80 µL MeOH, sonicated for 3 min and centrifuged, and the soluble MeOH extracts were analyzed by HR-ESI-QTOF-MS. The media control samples were subtracted from the experimental samples. Experiments were performed in triplicate.

PRESTO-Tango GPCR assay

Presto-Tango (49) analysis was performed as previously published (57). Cells were dosed with 100 µM of either compound 1 or 2 to generate Fig. 5A and B, and a dose-response curve was generated in a similar manner with varying concentrations of either compound 1 or 2 to construct Fig. S13.

Statistics

Data were analyzed with either an unpaired t-test or one-way analysis of variance (ANOVA) with Tukey’s post-test at a 95% CI using GraphPad Prism version 3.0 a for Macintosh (GraphPad Software, San Diego, CA).

General synthetic methods

All reactions were conducted in oven-dried glassware under an N₂ atmosphere unless otherwise stated. Commercial reagents and materials were used as received. In addition, 1 M decylmagnesium bromide solution in Et₂O was purchased from Santa Cruz Biotechnology, Inc., succinic anhydride was purchased from Alfa Aesar, H-L-Arg(Pbf)-OMe was purchased from Bachem, PyBOP was purchased from Sigma-Aldrich, L-Pro-OMe was purchased from Alfa Aesar, D-Pro-OMe was purchased from Alfa Aesar, dodecanal was purchased from Alfa Aesar, malonic acid was purchased from Alfa Aesar, (carbethoxymethylene)triphenylphosphorane (C₆H₅)₃P=CHCO₂CH₂CH₃ was purchased from Alfa Aesar, and t-BuOH was purchased from Alfa Aesar. Solvents and reagents used are abbreviated as trifluoroacetic acid (TFA), dimethyl sulfoxide (DMSO), dichloromethane (DCM), tetrahydrofuran (THF), carbon tetrachloride (CCl₄), dimethylformamide (DMF), methanol (MeOH), chloroform (CHCl₃), triethylamine (Et₃N), acetonitrile (MeCN), diethyl ether (Et₂O), hexanes (Hex), and ethyl acetate (EtOAc). Room temperature is abbreviated as rt. Reactions were monitored for completion by thin-layer chromatography (TLC) or LC-MS. Purification of the synthetic molecules was performed on a Biotage-Isolera One instrument (Biotage, Charlotte, NC, USA). Various purification cartridges were employed, including Biotage SNAP KP-Sil-HS 10 g, KP-Sil-HS 25 g, KP-Sil-HS 50 g, KP-C₁₈-HS 12 g, and KP-C₁₈-HS 30 g cartridges. Purification also was performed on an Agilent Prepstar HPLC system using an Agilent Polaris 5 µm C₁₈-A 250 × 21.2 mm column when necessary. Low-resolution or high-resolution LC-MS experiments were performed on instruments as described above in the General Molecular Biology Methods. NMR spectra were measured on an Agilent 400 MHz spectrometer equipped with a Broadband Probe or an Agilent DD2 600 MHz spectrometer equipped with a 3 mm cold probe in deuterated solvents. Multiplicities are abbreviated as singlet (s), doublet (d), triplet (t), quartet (q), quintet (qu), doublet of doublets (dd), multiplet (m), doublet of triplets (dt), and doublet of multiplets (dm).

SUPPLEMENTAL MATERIAL

The following material is available online at https://doi.org/10.1128/aem.00528-24.

Supplemental material. aem.00528-24-s0001.pdf.

Table S1; Figures S1 to S14.

ASM does not own the copyrights to Supplemental Material that may be linked to, or accessed through, an article. The authors have granted ASM a non-exclusive, world-wide license to publish the Supplemental Material files. Please contact the corresponding author directly for reuse.