Accelerating the discovery of alkyl halide-derived natural products using halide depletion

The study addresses the challenge of accelerating discovery of microbial natural products, especially those bearing or derived from halogens, which are valuable in pharmacology and often feature complex biosynthetic logic. While genome mining has revealed many biosynthetic gene clusters (BGCs), discovery remains hindered by redundancy, re-discovery, and the need for genetic manipulation or heterologous expression. Halogenated natural products are widespread and halogenases can alter small-molecule properties; however, products derived from cryptic halogenation are often unhalogenated and evade standard MS-based halogen detection methods. The authors hypothesized that limiting the availability of halide anions—essential substrates for halogenases—would perturb pathways requiring halogenation, enabling targeted discovery of halogenated metabolites and products of cryptic halogenation via comparative metabolomics. They developed and validated a halide depletion workflow combining controlled cultivation with and without halides, LC-MS/MS, comparative metabolomics, and molecular networking to identify such metabolites first in Cylindrospermum licheniforme and then in Nostoc punctiforme ATCC 29133.

The paper situates its approach within several lines of prior work: natural products as key sources of bioactive scaffolds and enzymatic novelty; halogen atoms as modulators of physicochemical properties in drugs; and halogenases as biosynthetic enzymes of growing interest, including as selective biocatalysts. Existing mass spectrometry strategies for finding halogenated metabolites often rely on characteristic isotopic patterns of Cl/Br, fragment ions, or elemental analysis, which miss unhalogenated end products of cryptic halogenation. Cryptic halogenation is a transient activation strategy that yields diverse motifs (cyclopropanes, biaryls, terminal alkynes, paracyclophanes). Bioinformatic genome-mining using halogenase sequences can help but is complicated by poorly characterized alkyl halide-derivatizing enzymes compared with halogenases. Isolated reports suggested medium halide levels can affect halogenated product titers, motivating a systematic depletion approach for discovery.

• Concept: Manipulate substrate availability by removing halide anions (e.g., chloride) from microbial media to depress production of halogen-dependent metabolites. Compare metabolomes of cultures grown in halide-replete versus halide-free conditions.
• Validation organism: Cylindrospermum licheniforme ATCC 29412, known to make cylindrocyclophanes via cryptic halogenation. Cultured in minimal media with and without chloride while keeping osmolarity and other nutrients comparable. LC-MS profiling, statistical analysis (Student’s t-test), and molecular networking identified differentially abundant features and clustered related compounds.
• Discovery organism: Nostoc punctiforme ATCC 29133. Bioinformatic search for BGCs encoding homologues of the halogenase CylC and the alkyl halide-derivatizing enzyme CylK, but lacking the resorcinol-forming enzyme CylI, identified a distinct BGC (pks3) reannotated as ngl (NgIA–NgIR). Halide depletion experiments with LC-MS general and lipidomics methods highlighted chlorinated metabolites most impacted by halide removal. Molecular networking connected features into a metabolite family.
• Genetics: Constructed markerless deletion mutants Δngl (deleting core biosynthetic genes nglA–nglO) and ΔnglO (deleting the CylK homologue nglO) via Gibson assembly of pRL278-based shuttle vectors and biparental conjugation. LC-MS compared mutant and wild type metabolomes to connect features to ngl.
• Isolation/structure elucidation: Grew ΔnglO at scale (10–20 L) to overproduce targets, extracted with chloroform/methanol, used flash silica chromatography and preparative HPLC (and saponification when needed) to purify nostochloroside A (5) and an esterified analogue (7). Structure elucidation by 1D/2D NMR (COSY, H2BC, HMBC), derivatizations, and GC-MS of permethylated sugars after acid hydrolysis identified the headgroup as 6-O-methylgulose. Chemical degradation located the internal hydroxyl at C6; stable isotope feeding with deuterated fatty acids localized the chlorine to C12 of the lipid tail.
• Enzymology: Heterologously expressed and purified a soluble truncated NgIO (NgIO′, lacking a C-terminal RTX domain). In vitro assays incubating NgIO′ with 5 detected oligomeric products up to octamers by LC-MS; controls without enzyme showed no oligomerization. No activity on 7 or mixed 5/7 for 7-derived oligomers. AlphaFold modeling compared to CylK structures to infer active site conservation and a hydrophobic binding pocket consistent with lipidic substrate recognition, supporting an etherification mechanism using O6 as nucleophile to displace the alkyl chloride at C12.
• Transcriptomics/phenotyping: RNA-seq of N. punctiforme under halide-replete versus depleted conditions at 2 and 4 weeks revealed stress responses and increased ngl transcription under depletion, despite greatly reduced product titers. Assessed growth, aggregation, motility, and nitrogen fixation phenotypes of Δngl versus wild type.
• LC-MS/MS and molecular networking: Agilent 1260/6530 Q-TOF with defined gradients for general and lipidomics separations; MS/MS-driven networking (MetGem, Cytoscape) to cluster features and relate adducts/derivatives. Statistics on triplicate biological replicates per condition.

• Validation in C. licheniforme: Halide depletion markedly reduced features in the cylindrocyclophane network. A putative biosynthetic intermediate (3) showed the largest decrease (fold change 1,525; P = 2.4 × 10^-2). Cylindrocyclophane F (4) and the hydroxymonoalkylresorcinol monomer (2) also decreased (fold changes 13.5, P = 8.9 × 10^-3; and 88.2, P = 4.3 × 10^-3, respectively). Molecular networking clustered 22 related features, most responding to halide removal, many lacking halogen isotopic signatures, consistent with products of cryptic halogenation.
• Discovery in N. punctiforme: Identified a distinct CylC/CylK-homologous BGC (ngl). LC-MS under halide depletion uncovered a family of chlorinated metabolites most depleted when chloride was absent. Representative: C28H37OCl (5) [M+H]+ m/z 555.3679, fold change 87.5 (P = 5.5 × 10^-9); C28H35OCl (6), fold change 421.1 (P = 6.9 × 10^-10). Lipidomics revealed additional high-responding features; C26H49OCl (7) [M+NH4]+ m/z 838.6566, fold change 607.0 (P = 1.8 × 10^-12). Network analysis linked five features (7–11) as fatty-acid ester adducts of 5; a pseudo-dimer (12) of 5 appeared in some halide-replete samples and was absent under depletion, suggesting cryptic halogenation involvement. One additional chlorinated feature (C30H59O2Cl) appeared unrelated to the main family.
• Genetic validation: Metabolites 5–11 were undetectable in Δngl, confirming ngl involvement; ΔnglO overproduced 5–11, consistent with NgIO processing 5–11 into downstream products, while the pseudo-dimer 12 was not consistently observed in vivo. An unrelated small chlorinated metabolite persisted in both mutants, indicating production outside ngl.
• Structural elucidation: Compounds 5 and 7 are chlorinated glycolipids (nostochlorosides) with a 6-O-methylgulose headgroup and an unusual C17 lipid tail chlorinated at C12. Esterified variants 7–11 differ by fatty acid chain length and unsaturation.
• Enzymatic mechanism: Purified NgIO′ catalyzed oligomerization (not macrocyclization) of 5 to chlorinated oligomers (dimer through octamer) via etherification, consistent with displacement of the C12 chloride by a hydroxyl nucleophile. MS/MS suggested linkage through the lipid moiety, not sugars. Structural modeling indicated conservation of halide-binding and nucleophile-activating residues akin to CylK and a hydrophobic cavity compatible with lipid substrates; data support O6–C12 ether linkage formation. Wild-type extracts contained species matching in vitro dimer/trimer.
• Physiology and regulation: Δngl showed increased aggregation and decreased motility on soft agar but no growth defects, including under nitrogen-fixing conditions; 5–11 titers were lower during nitrogen fixation in wild type despite reported ngl upregulation in prior studies. RNA-seq under halide depletion showed upregulation of stress responses and a 2–3 fold increase in ngl transcription by week 4, yet metabolite production dropped up to ~600-fold, indicating halide availability, not transcription, limits product formation.
• Overall: Halide depletion is a powerful, function-agnostic strategy that illuminates halogen-dependent pathways and cryptic halogenation products, enabling rapid linkage of metabolites to BGCs and discovery of new enzymatic reactivity (NgIO-mediated etherification).

The findings confirm that limiting a key biosynthetic substrate (halide anions) effectively suppresses pathways requiring halogenation, thereby contrasting metabolomes to prioritize relevant features. In C. licheniforme, depletion of known cylindrocyclophane intermediates validated the approach. Applying the workflow to N. punctiforme revealed the nostochlorosides, linked them genetically to ngl, and uncovered an unprecedented alkyl halide-utilizing etherification catalyzed by the CylK-family homologue NgIO. The strategy circumvents the need for activity-based screens, complex genetic manipulations, or reliance on halogen isotopic signatures (which miss cryptic products). It also provides a practical route to connect metabolites to halogenase-containing BGCs in native producers, especially when full pathway reconstitution in heterologous hosts is impractical. The observed decoupling between ngl transcriptional upregulation and product titers under halide depletion underscores substrate-level control over pathway output, validating the mechanistic basis of the workflow. The discovery expands known reactivity of CylK homologues to include hydroxy nucleophile etherification (in addition to resorcinol and carboxylate O-alkylation), and reveals unusual structural motifs (gulose-containing headgroups, odd-chain C17 tails) with potential biological significance. Broadly, manipulating precursor availability can be generalized to other building blocks (metals, amino acids, sugars) to target diverse biosynthetic chemistries.

This work introduces halide depletion as a simple, generalizable, and function-agnostic discovery strategy for halogenated natural products and cryptic halogenation derivatives. The method rapidly identified known pathway perturbations in C. licheniforme and led to the discovery and structural elucidation of the nostochlorosides (chlorinated gulose-containing glycolipids) from the ngl BGC in N. punctiforme. Genetic deletions confirmed pathway linkage, and in vitro biochemistry revealed that the CylK-family enzyme NgIO catalyzes an unprecedented etherification-driven oligomerization of an alkyl chloride-bearing glycolipid, expanding the enzymology of alkyl halide-utilizing enzymes. Future directions include full structural determination of oligomers like the pseudo-dimer 12 by isolating sufficient material or using synthetic substrates, biochemical dissection of the ngl PKS assembly line to explain the odd-chain C17 tail biosynthesis, identification of the acyltransferase for esterified nostochlorosides, and broader application of substrate-depletion strategies (halides and beyond) across microbial taxa to accelerate BGC-to-metabolite connections and discover new enzyme chemistries.

• Environmental scope: Halide depletion may be unsuitable for marine organisms or others requiring halides for growth; its physiological impacts are less studied than high-salt tolerance and may cause non-specific effects in some taxa.
• Material constraints: Low titers and oligomerization hindered isolation of sufficient quantities of the pseudo-dimer 12 and higher oligomers for full NMR characterization; estimated culture volumes (>1,000 L) would be required under current conditions.
• Genetic/biochemical completeness: The ngl BGC does not obviously encode the acyltransferase forming esterified nostochlorosides (7–11), complicating heterologous reconstitution; parts of the PKS logic remain to be biochemically validated.
• Detection variability: The pseudo-dimer 12 was inconsistently observed in vivo, possibly due to extraction/ionization challenges of higher oligomers.
• Generalizability: While powerful for halogen-dependent pathways, the approach targets a subset of natural products and relies on growth in defined, low-salt minimal media.