Exploring 1-alkene biosynthesis in bacterial antagonists and Jeotgalicoccus sp. ATCC 8456

Abstract

Terminal olefins are important platform chemicals, drop-in compatible hydrocarbons and also play an important role as biocontrol agents of plant pathogens. Currently, 1-alkenes are derived from petroleum, although microbial biosynthetic routes are known. Jeotgalicoccus sp. ATCC 8456 produces 1-alkenes via the fatty acid decarboxylase OleT_JE. UndA and UndB are recently identified non-heme iron oxidases converting medium-chain fatty acids into terminal alkenes. Our knowledge about the diversity and natural function of OleT_JE, UndA, and UndB homologs is scarce. We applied a combined screening strategy—solid-phase microextraction coupled with gas chromatography–mass spectrometry (SPME GC–MS) and polymerase chain reaction (PCR)-based amplification—to survey an environmental strain collection for microbial 1-alkene producers and their corresponding enzymes. Our results reinforce the high level of conservation of UndA and UndB genes across the genus Pseudomonas. In vivo production of defined 1-alkenes (C₉–C₁₃; C₁₅; C₁₉) was directed by targeted feeding of fatty acids. Lauric acid feeding enabled 1-undecene production to a concentration of 3.05 mg l⁻¹ in Jeotgalicoccus sp. ATCC 8456 and enhanced its production by 105% in Pseudomonas putida 1T1 (1.10 mg l⁻¹). Besides, whole genome sequencing of Jeotgalicoccus sp. ATCC 8456 enabled reconstruction of the 1-alkene biosynthetic pathway. These results advance our understanding of microbial 1-alkene synthesis and the underlying genetic basis.

In this study, the repertoire of available UndA and UndB homologs and putative hosts for 1-alkene production was expanded and insights into conditions that enhance bacterial 1-alkene biosynthesis were obtained.

Introduction

1-Alkenes are important platform chemicals (Kourist 2015) that can be directly applied as drop-in compatible biofuels (Yan and Liao 2009) and also are interesting for biocontrol of plant pathogens. Especially 1-undecene got attention for its antagonistic properties (Hunziker et al. 2015). Although microbial biosynthetic routes are known (Robinson and Wackett 2019), 1-alkenes are now almost exclusively derived from fossil resources (Rui et al. 2015). The cytochrome P450 fatty acid (FA) decarboxylase OleT_JE, found in Jeotgalicoccus sp. ATCC 8456 (J8456) (Rude et al. 2011), catalyzes the decarboxylation of short- to long-chain FAs, converting them to 1-alkenes (C₃– C₂₁) (Belcher et al. 2014; Dennig et al. 2015; Fang et al. 2017). UndA and UndB were identified for their ability to convert medium-chain FAs into terminal alkenes. They belong to the diiron (Fe[II]) enzymes and catalyze the conversion of free FAs to 1-alkenes by oxidative decarboxylation. While UndA is a soluble protein, UndB belongs to the FA desaturases-like superfamily of integral membrane enzymes (Rui et al. 2014, 2015; Iqbal et al. 2023). Phylogenetic analysis showed that UndA is well conserved in the genera Pseudomonas, Acinetobacter, Burkholderia, and Myxococcus (Rui et al. 2014). UndB homologs were found only in a few Pseudomonas and related species (Rui et al. 2015). OleT_JE belongs to the cyp152 P450 enzyme family, so far identified in Kocuria rhizophila, Corynebacterium efficiens, Methylobacterium populi, and Bacillus subtilis (Rude et al. 2011). In this study, the diversity of microbial 1-alkene producers was investigated within an environmental strain collection derived from plant-associated antagonistic isolates, containing members of the families Bacillaceae, Enterobacteriaceae, Pseudomonaceae, and Staphylococcaceae, among others. Plant-associated antagonists are good producers of bioactive volatile compounds (Bruisson et al. 2020). Therefore, we expected them to be a good source for 1-alkenes. Identified 1-alkene producers were further investigated for undA, undB and oleT_JE homologous genes, using a proven degenerate primer-mediated PCR amplification screening approach (Karasev et al. 1994; Müller et al. 2015). Furthermore, it is important to understand the conditions under which 1-alkene production occurs in vivo. Previous analysis showed that the production of certain 1-alkenes by J8456 can be influenced by targeted feeding of FA precursors (Rude et al. 2011). Here, we performed further FA feeding experiments with Pseudomonas and Bacillus strains. To aid in the identification of further 1-alkene biosynthetic enzymes and accessory genes, we sequenced, assembled, and annotated the genome of J8456. The findings of this study advance our knowledge about 1-alkene production in vivo.

Materials and methods

All chemicals were of analytical grade and purchased from Sigma–Aldrich (St. Louis, Missouri, USA), Carl Roth (Karlsruhe, Germany), Merck (Darmstadt, Germany), or Thermo Fisher Scientific (Massachusetts, USA) if not stated otherwise.

Bacterial strains and cultivation conditions

To screen for 1-alkene producers, strains from the Strain Collection of Antagonistic Microorganisms (SCAM), Graz University of Technology, Austria (Gasser et al. 2009) were selected for their taxonomic relationship to bacteria capable of alkene biosynthesis (Supplementary Table 1) (Winters et al. 1969; Beller et al. 2010; Sukovich et al. 2010; Rude et al. 2011). To measure 1-alkene production using SPME GC–MS, 3 ml liquid media were inoculated with 1% (v/v) of prepared precultures and cultivated overnight at 30°C and 100 rpm in 20 ml GC–MS vials. All screenings were performed in duplicates in independent cultivation rounds. The following bacterial strains were selected for further analysis: Bacillus thuringiensis 3R2-29 (Berg et al. 2005), Pseudomonas aeruginosa QC14-3-8 (Lottmann 2000; Zachow et al. 2009), Pseudomonas brassicacearum L13-6-12 (Zachow et al. 2017), Pseudomonas poae RE*1-1-14 (Müller et al. 2013), Pseudomonas putida 1T1 (Opelt and Berg 2004) (characterized before), and two uncharacterized Pseudomonas sp. strains, denominated Baz30 and Baz53. To identify the latter two, 16S rRNA was sequenced and blasted against the reference RNA sequences (refseq_rna) database at NCBI. Pseudomonas brenneri CFML 97–391 and Pseudomonas canadensis 2–92 were the closest matches (blastn) with 99.50% and 99.72%, respectively (13 October 2020). They are referred as P. brenneri Baz30 and P. canadensis Baz53 from now on. All strains were cultured in nutrient broth (NB, 15.0 g l⁻¹ nutrient broth II; Sifin diagnostics GmbH, Berlin, Germany) medium, except J8456 (ATCC, Manassas, Virginia, USA; deposited as Micrococcus candicans ATCC® 8456) was cultured in tryptic soy broth yeast extract [TSBYE, 30.0 g l⁻¹ CASO broth and 0.5% (v/v) yeast extract] medium (Rude et al. 2011). For cultivation on solid media, NB-agar (NB and 15.0 g l⁻¹ agar) was used.

Identification and quantification of 1-alkenes

For quantification of 1-alkenes, headspace SPME GC–MS was applied. Air-tight sealed 20 ml GC–MS vials [ND18 Headspace Screw Vial (clear), 75.5 × 22.5 mm, rounded bottom; ND18 Magnetic Screw Cap (8 mm hole) with Silicone/PTFE Septa (white/blue), 1.5 mm 55° shore A; BGB Analytik Vertrieb GmbH, Rheinfelden, Germany] were used. The SPME fiber was conditioned once for 8 min at 40°C. All samples were extracted with a polydimethylsiloxane (PDMS)-coated fiber [30 µm, for non-polar semi-volatiles (MW 80–500 g mol⁻¹); Supelco, Pennsylvania, USA] for 15 min at 35°C. A splitless thermal desorption for 30 min at 270°C, 9.1473 psi, and a septum purge flow of 3 ml min⁻¹ in the heated injection port of the gas chromatograph followed. An Agilent HP-5 column (30 m × 320 µm, 0.25 μm film) was used with following temperature program: 40°C/hold 2 min; 5°C min⁻¹ to 110°C; 10°C min⁻¹ to 280°C; hold 3 min with a flow of 1.2 ml min⁻¹ under 9.1473 psi.

The samples were analyzed on a GC–MS system with He as carrier gas (Agilent 7890B/5977A Series Gas Chromatograph/Mass Selective Detector; and PAL RSI 85; CTC Analytics AG, Zwingen Switzerland) in the mass range of 50–350 g mol⁻¹ with a quadrupole mass spectrometer. Obtained mass spectra were investigated with the associated software (Agilent MSD Productivity) and the NIST EI Mass Spectral Library (NIST MS Search Version 2.2, 2014).

Analytically pure 1-alkenes were employed as standards. Measured retention times and distinctive molecular ions for each standard (Supplementary Table 2) were later used for identification of 1-alkenes. Linear calibration curves were generated using dilutions of 1-alkenes C₉–C₁₃; C₁₅; C₁₇–C₁₉ in the range of 0.39 and 28.9 µM. All measurements were done in duplicates or triplicates. The peaks were integrated using the RTE integrator.

Fatty acid feeding

The conditions leading to in vivo 1-alkene production were evaluated by feeding FAs as precursor for the reaction. FA stock solutions of 20 mM were prepared in dimethyl sulfoxide (DMSO). In GC–MS vials, 3 ml media were mixed with the FA stock to provide final concentrations of 200 µM before inoculation with 1% (v/v) preculture. J8456 was fed with C6:0, C8:0, C9:0, C10:0, C11:0, C12:0, C14:0, C16:0, and C20:0. B. thuringiensis 3R2-29, P. aeruginosa QC14-3-8, P. brassicacearum L13-6-12, P. brenneri Baz30, P. canadensis Baz53, P. poae RE*1-1-14, and P. putida 1T1 were fed with C10:0, C11:0, C12:0, C14:0, and C16:0. As control, 1% (v/v) of DMSO was used. J8456 cultures were incubated at 26°C for 24 h, all other strains at 30°C. Samples were run in triplicates.

PCR (-based screening)

A blast search was performed to get highly similar sequences for designing primers based on J8456 oleT_JE (GenBank accession no. HQ709266.1) (Rude et al. 2011), Pseudomonas fluorescens Pf-5 Chain A/undA [GenBank accession no. CP000076.1 (5024351–5025136)] (Rui et al. 2014), P. fluorescens Pf-5 undB [GenBank accession no. CP000076.1 (238682–239755)] (Rui et al. 2015). Obtained nucleotide sequences were aligned. Based on highly conserved regions, degenerate primers were designed (Supplementary Table 3a) which allowed partial amplification. Bacterial cells were denaturated at 98°C for 15 min, followed by centrifugation at 4000 rpm for 3 min. The supernatant was directly applied as template. The OneTaq® Quick-Load® 2X master mix with standard buffer (New England Biolabs, Ipswich, USA) was used for the PCR-based screening (Supplementary Tables 4a and 4b). Resulting partial sequences were analyzed by blastn. Per sequence, the four closest hits were aligned. In resulting consensus sequences, open reading frames (ORFs) were determined and used as basis for designing (degenerate) primers to obtain complete homologous sequences (Supplementary Table 3b). Therefore, a high-fidelity DNA Polymerase was used (20 U/ml Q5®; New England Biolabs, Ipswich, USA; Supplementary Tables 4c, 4d, and 4e). The strain identification was performed with the universal 27F-1492r primer pair for 16S rRNA amplification (Supplementary Tables 4f and 4g). All purified PCR products were sent for Sanger sequencing to Microsynth AG (Balgach, Switzerland). PCR purification was performed with the Wizard® SV Gel and PCR Clean-Up System (Promega, Wisconsin, USA) following the producer's manual.

Whole genome sequencing of J8456

The genome of J8456 was de novo sequenced with pacific biosciences single molecule real-time (SMRT) sequencing technology, short PacBio RS (Buermans and den Dunnen 2014), at GATC Biotech GmBH (Konstanz, Germany). The reads were assembled using the de novo Hierarchical Genome Assembly Process (RS_HGAP Assembly.2) implemented within the analysis pipeline SMRT Analysis 2.2 (Pacific Biosciences, Menlo Park, California, USA). The genes were annotated using BASys (Van Domselaar et al. 2005). Reconstruction of the FA metabolism and further genomic pathway information were gained using the Kyoto Encyclopedia of Genes and Genomes (KEGG) (Kanehisa and Goto 2000; Kanehisa 2019; Kanehisa et al. 2020).

Data accessibility

Identified undA and undB sequences (not identically available) can be found at GenBank under following accession numbers: MW390202–MW390206. The genome of J8456 was deposited at GenBank under the accession number CP066224.

Results

Identification of 1-alkene producers

To identify novel 1-alkene producing bacteria, we selected strains that are taxonomically related to those of known 1-alkene producers from the SCAM at Graz University of Technology, Austria (Gasser et al. 2009) (Supplementary Table 1). These selected strains were screened for 1-alkene biosynthesis by cultivating potential candidates in liquid media overnight in GC–MS vials, followed by analysis of the produced volatiles using SPME GC–MS. In total, one Bacillus strain (phylum of Firmicutes) and six Pseudomonas strains (phylum of Proteobacteria; class of Gammaproteobacteria) were identified as 1-alkene producers (Table 1). 1-Undecene was the major 1-alkene produced by Pseudomonas and Bacillus with concentrations ranging from 0.06 to 0.17 mg l⁻¹ (0.39–1.07 µM). In comparison, J8456 produced 0.05 mg l⁻¹ 1-pentadecene (0.24 µM) and 0.09 mg l⁻¹ 1-nonadecene (0.34 µM) without supplementation of FA to the media.

Table 1.

Bacterial 1-alkene producers identified in a preliminary screening in TSBYE without FA feeding.

Bacterial strain	Identified gene	1-C11 (µM)	1-C12 (nM)	1-C13 (nM)	1-C14 (nM)
Bacillus thuringiensis 3R2-29	n.d.	0.67 ± 0.02	n.d.	2.4 ± 2.4	n.d.
Pseudomonas aeruginosa QC14-3-8	n.d.	0.39 ± 0.06	0.6 ± 0.1	3.9 ± 0.6	n.d.
Pseudomonas brassicacearum L13-6-12	undA, undB	0.95 ± 0.19	n.d.	2.5 ± 1.7	n.d.
Pseudomonas brenneri Baz30	undA, undB	0.49 ± 0.05	n.d.	n.d.	n.d.
Pseudomonas canadensis Baz53	undA, undB	0.49 ± 0.10	n.d.	n.d.	1.6 ± 1.6
Pseudomonas poae RE*1-1-14	undA, undB	1.07 ± 0.34	n.d.	0.3 ± 0.3	n.d.
Pseudomonas putida 1T1	undA, undB	0.52 ± 0.37	n.d.	n.d.	n.d.

The 1-alkenes are 1-undecene (1-C11), 1-dodecene (1-C12), 1-tridecene (1-C13), and 1-tetradecene (1-C14). Identification and quantification by SPME GC–MS (sample size, n = 2); n.d.: not detected.

Fatty acid feeding

The effect of feeding FA precursors in vivo at a concentration of 200 µM was investigated and quantified for all strains using headspace SPME GC–MS. In J8456 cultures (Fig. 1 and Supplementary Table 5), the highest 1-alkene concentration was achieved with dodecanoic acid, which increased production of 1-undecene to a concentration of 3.05 ± 0.08 mg l⁻¹ (19.8 ± 0.5 µM). The production of corresponding 1-alkenes was also enhanced upon feeding decanoic acid (0.42 ± 0.01 mg l⁻¹; 3.31 ± 0.05 µM 1-C9), undecanoic acid (0.90 ± 0.10 mg l⁻¹; 6.41 ± 0.72 µM 1-C10), and tetradecanoic acid (0.71 ± 0.01 mg l⁻¹; 3.89 ± 0.08 µM 1-C13). Upon feeding hexadecanoic acid, a four-fold increase of 1-C15 compared to the control sample was measured. For B. thuringiensis 3R2-29, P. aeruginosa QC14-3-8, P. brassicacearum L13-6-12, P. brenneri Baz30, P. canadensis Baz53, P. poae RE*1-1-14, and P. putida 1T1, feeding 200 µM decanoic acid, tetradecanoic acid, and hexadecanoic acid did not result in the production of the expected 1-alkenes. Nevertheless, 1-undecene production increased in P. canadensis (up to 84.5%) and P. putida (up to 105.2%) (Fig. 2).

Figure 1.

Feeding saturated FAs in a concentration of 200 µM enhanced production of the targeted 1-alkenes in Jeotgalicoccus sp. ATCC 8456 cultures. Employed FAs: hexanoic (C6:0), octanoic (C8:0), nonanoic (C9:0), decanoic (C10:0), undecanoic (C11:0), dodecanoic (C12:0), tetradecanoic (C14:0), hexadecanoic (C16:0), eicosanoic (20:0) acid; control: DMSO; sample size, n = 3.

Figure 2.

Concentration of 1-undecene (light gray) upon feeding saturated FAs (200 µM) to B. thuringiensis 3R2-29, P. aeruginosa QC14-3-8, P. brassicacearum L13-6-12, P. brenneri Baz30, P. canadensis Baz53, P. poae RE*1-1-14, P. putida 1T1, and Jeotgalicoccus sp. ATCC 8456. The negative control (DMSO) is shown in dark gray. Sample size, n = 3.

Screening for homologous enzyme sequences

The identified 1-alkene-producing bacteria were screened for sequences homologous to oleT_JE  undA and undB using a degenerative PCR approach (Müller et al. 2015). No novel oleT_JE homologs could be found, but sequences of undA and undB were successfully obtained from P. brassicacearum L13-6-12, P. brenneri Baz30, P. canadensis Baz53, P. poae RE*1-1-14, and P. putida 1T1 (Supplementary Figure 1). Aligning against non-redundant protein sequences showcased the high level of conservation of UndA and UndB across the genus Pseudomonas (Supplementary Table 6). Compared to the best hits and the characterized UndA from P. fluorescens Pf-5 (PDB 4WWJ), all retrieved UndA sequences are missing 11 amino acid residues at the N-terminus. The identity varied between species, from 66.5 to 96.8% for undA, and from 74.7 to 96.6% for undB (Table 2). The partial UndA sequence found in P. brassicacearum was about 130 bp longer at the 5′-end, explaining the low identity. Based on the crystal structure of P. fluorescens Pf-5 UndA (PDB accession number 6P5Q; Zhang et al. 2019), homology modelling (Guex et al. 2009; Benkert et al. 2011; Bertoni et al. 2017; Bienert et al. 2017; Waterhouse et al. 2018) was performed, which showed high sequence identities between the P. fluorescens Pf-5 UndA and all partial UndA sequences identified, ranging from 89.66 to 90.72% (Supplementary Figure 2; sequence coverage compared to the template ranging from 0.87 to 0.98, QSQE from 0.70 to 0.75, and QMEAN from −0.88 to −0.45), supporting the assumption that the obtained sequences are homologous to UndA.

Table 2.

Percentage identity between obtained undA (a) and undB (b) nucleotide sequences of P. brassicacearum L13-6-12 (1), P. brenneri Baz30 (2), P. canadensis Baz53 (3), P. poae RE*1-1-14 (4), and P. putida 1T1 (5), determined by alignment pairwise comparison.

a	1	2	3	4	5
1	100.00	66.89	66.48	67.07	66.70
2	66.89	100.00	85.92	87.76	86.30
3	66.48	85.92	100.00	87.34	96.77
4	67.07	87.86	87.34	100.00	87.73
5	66.70	86.30	96.77	87.73	100.00
b	1	2	3	4	5
1	100.00	77.96	78.61	74.66	77.87
2	77.96	100.00	87.13	80.32	87.04
3	87.61	87.13	100.00	81.67	96.57
4	74.66	80.32	81.67	100.00	80.68
5	77.87	87.04	96.57	80.68	100.00

Recently, UndB was proven to be an integral transmembrane FA desaturase (Iqbal et al. 2023). The UndB sequence is characterized by the presence of three conserved histidine-rich motives: H₈₉DLIH₉₃, H₁₂₈(L/F)(N/H)H₁₃₁H₁₃₂, and H₂₉₇(G/S/A)IH₃₀₀H₃₀₁ (Rui et al. 2015). All three His-boxes were found in each of the obtained UndB sequences (Fig. 3), supporting the claim that the identified sequences are homologous to the known FA desaturase UndB.

Figure 3.

Alignment of translated protein sequences of UndB homologs from P. brassicacearum L13-6-12 (1; accession number WP_069 971 255), P. brenneri Baz30 (2; accession number MW390205), P. canadensis Baz53 (3; accession number MW390206), P. poae RE*1–1-14 (4; accession number WP_041 931 557), and P. putida 1T1 (5; accession number WP_179 676 382). The black rectangles indicate three conserved His-boxes: H₈₉DLIH₉₃, H₁₂₈(L/F)(N/H)H₁₃₁H₁₃₂ and H₂₉₇(G/S/A)IH₃₀₀H₃₀₁ (Rui et al. 2015).

Reconstructing the fatty acid synthesis pathway in J8456

In total, 2117 genes were annotated in the genome of J8456 using BASys (Van Domselaar et al. 2005). Mapping against the KEGG database revealed 1220 functional orthologues. A search for genes involved in the FA metabolism revealed 17 ORFs, showing a complete FA pathway (Rock and Cronan 1996; Freiberg et al. 2004) (Fig. 4 and Supplementary Table 7). About 3% of all presumed proteins can be assigned to the FA metabolism (Supplementary Figure 3). Compared to reported prokaryotic pathways (Rock and Cronan 1996; Freiberg et al. 2004), a possible FA synthesis route can be hypothesized: First, acetyl-CoA is carboxylated to malonyl-CoA by AccABC1D. Then, malonyl-CoA gets transferred to the ACP (AcpP) by FabD. Under emission of CO₂, malonyl-ACP condenses with another acetyl-CoA, catalyzed by FabH. Thereby, acetoacetyl-ACP (β-ketoacyl-ACP for further cycles) is generated. Acetoacetyl-ACP (β-ketoacyl-ACP) is reduced to β-hydroxyacyl-ACP by FabG. Dehydration with presumed FabZ leads to trans-2-enoyl-ACP and subsequent reduction to acyl-ACP catalyzed by FabI. For elongation, malonyl-ACP is added to acyl-ACP by FabF. The predicted thioesterase YneP was the only thioesterase found within the genomic sequence. It might be responsible for releasing free FAs from the acyl carrier protein (Torto-Alalibo et al. 2014), serving directly as precursor for conversion to terminal unsaturated olefins by OleT_JE or the newly discovered α-hydroxylase (P450_Jα) (Armbruster et al. 2020). Comparing annotated genes from J8456 with pathways of orthologue genes deposited in the KEGG database confirmed the capability to initialize FA synthesis and the availability of elongation cycles. Thereby, the second FabF gene (BASYS01401) was assigned as orthologue of FabB (KEGG, K00647), and either FabG (BASYS00708) or the uncharacterized oxidoreductase (Ymfl, BASYS00661) as orthologues of FabG (KEGG, K00059). Ymfl was referred to be responsible for the first reduction step in the FA biosynthesis with BASys but shared only 30% sequenced identity with FabG (BASYS00708). Additionally, two FabG homologs (BASYS01407 and BASYS01432) were not assigned to the corresponding KEGG orthology identifier, assuming that they might have been misclassified in a previous annotation, or due to a low sequence similarity were not found in the KEGG database. YneP as well as AcpP were found in the KEGG database and assigned to acyl-CoA thioester hydrolase and acyl carrier protein, respectively, supporting their postulated function. This gathered information allowed the reconstruction of a complete bacterial FA synthesis pathway type II (Rock and Cronan 1996; Freiberg et al. 2004).

Figure 4.

FA synthesis pathway of Jeotgalicoccus sp. ATCC 8456 reconstructed based on the sequenced genome.

Discussion

We obtained valuable insights into the conditions leading to increased bacterial 1-alkene biosynthesis in Jeotgalicoccus sp. ATCC 8456 by means of reaction engineering. Previous experiments showed that the production of certain 1-alkenes can be triggered by targeted FA feeding (Rude et al. 2011). In this study, feeding with lauric acid fully suppressed the production of other 1-alkenes, leading to exclusive synthesis of 3.05 mg l⁻¹ 1-undecene. A similar trend was observed for other tested FAs, though to a lower extent. The annotation of the whole genome sequence of J8456 opens possibilities for further research on biosynthetic prospects. For instance, analysis of the annotated genome from our study enabled the discovery of a new FA α-hydroxylase (P450_Jα) for 1-alkene biosynthesis (Armbruster et al. 2020). Until now, the synthesis pathway for free FAs, which are precursors for 1-alkene biosynthesis in J8456, was unknown. Based on genome annotation, it was possible to reconstruct the highly conserved bacterial fatty acid synthase (FAS) type II pathway (Rock and Cronan 1996; Freiberg et al. 2004) in J8456 and hypothesize a route for 1-alkene synthesis.

Several 1-alkene producers were identified from a large collection of bacterial antagonists of plant pathogens. P. canadensis Baz53 and P. putida 1T1 showed increased production of 1-undecene by 85% and 105% (1.34 and 1.10 mg l⁻¹, respectively) as a response to FA feeding. P. putida is known to cope well with metabolic burdens and the ability to catabolize a wide range of carbon sources. The genetic repertoire responsible for the fitness of P. putida KT2440 grown on FAs has been functionally analyzed before (Thompson et al. 2020). Based on the genetic pool of P. putida, we performed a blastn and blastp analysis to find possible explanations for the increased 1-undecene production. A second search for microbial FA transport proteins (Salvador López and Van Bogaert 2021) was performed as well. Both analyses did not identify any leads; there were no observable differences in the genetic repertoire associated with growth on FAs, nor in regard to microbial FA transport proteins between P. canadensis Baz53, P. putida 1T1, and the other bacterial strains examined. Further research is required to understand the observed difference between P. canadensis Baz53, P. putida 1T1 to the other strains.

Wang et al. demonstrated the biosynthetic potential of Pseudomonas and UndB with a product titer of 1102.6 mg l⁻¹ 1-undecene using lauric acid as sole carbon source (Wang, Yu and Zhu 2018). Acinetobacter baylyi ADP1, Saccharomyces cerevisiae, and Pichia pastoris were successfully engineered for 1-alkene production (Zhu et al. 2017; Zhou et al. 2018; Luo et al. 2019; Cai et al. 2022). Knoot and Pakrasi engineered the cyanobacterium Synechococcus sp. PCC 7002, enabling synthesis of non-natural hydrocarbon profiles by substituting the cyanobacterial native olefin synthase Ols by UndA variants (Knoot and Pakrasi 2019). Nevertheless, low catalytic efficiencies of native enzymes often impose a bottleneck to achieve higher product titers (Fu and Balskus 2020). The discovery of new enzymes, new production host, and feedstocks is therefore essential to advance the field of microbial hydrocarbon synthesis. Here, we expanded the repertoire of available UndA and UndB homologs and putative hosts for hydrocarbon production, which will support future metabolic engineering efforts.

The biological function of 1-alkenes in these bacteria is still uncertain. In J8456, longer chained 1-alkenes were detected, namely 1-pentadecene and 1-nonadecene, while 1-undecene, 1-tridecene, and 1-tetradecene were detected in Pseudomonas and Bacillus, when cultured without supplementing FAs. This is consistent with the reported specificity of UndA towards the in vivo synthesis of 1-undecene (Luo et al. 2019). J8456 was first isolated from the Korean fish paste Jeotgal (Yoon 2003; Robinson and Wackett 2019). The production of Jeotgal includes fermentation of fish at salt concentrations up to 40% (Koo et al. 2016). Halotolerant bacteria growing in salty environments accumulate medium- to long-chain hydrocarbons in the membrane (Kates 1986). The hydrocarbon metabolism of J8456, including the production of certain 1-alkenes, may result from the halotolerant properties of J8456. Shorter-chain 1-alkenes, such as 1-undecene, are volatile organic compounds, well-known for inter- and intraspecies communication (Schmidt et al. 2016), and commonly produced by antagonistic bacteria. Antifungal activity of 1-undecene was already shown in previous studies (Zhou et al. 2014; Hunziker et al. 2015). Robinson and Wackett (2019) hypothesized that UndA and UndB may have evolved for intra- and interspecies signaling and defense.

In this study, UndA was found in every strain except P. aeruginosa, J8456, and B. thuringiensis 3R2-29. The wide distribution of UndA in the genus Pseudomonas was consistent with literature reports (Rui et al. 2014; Robinson and Wackett 2019). Rui et al. mentioned that UndB occurs in some Pseudomonas species, but explicitly wrote that UndB is absent in P. aeruginosa and P. putida (Rui et al. 2015). In this project, a homologous UndB sequence was found in P. putida 1T1. Neither the degenerate primers for OleT_JE, nor the degenerate primers for UndA or UndB did work with B. thuringiensis. Using a homology model of OleT_JE, Rude et al. identified other 1-alkene producing organisms carrying cyp152 P450s. A FA hydroxylase from B. subtilis (GenBank accession number NP_388 092) with an identity of 41% to OleT_JE was identified and its 1-alkene producing activity demonstrated (Rude et al. 2011). The low percentage identity of 41% between OleT_JE and the FA hydroxylase found in B. subtilis could explain a possible limitation of the applied PCR-screening method for this class of hydroxylases.

In conclusion, these results aid in understanding the microbial 1-alkene biosynthesis and the underlying genetic diversity. A high conservation of UndA and UndB in the genus Pseudomonas was confirmed. OleT_JE, or homologs thereof, were only detected in its original organism, suggesting that OleT_JE is highly conserved in the genus Jeotgalicoccus.

Conflict of interest

None declared.

Funding

The research outlined in this publication was supported by the seed-funding program (Anschubfinanzierung) from TU Graz—Field of Expertise ‘Human and Biotechnology’ (Project: ‘Enzyme mining in 1-alkene producing microorganisms’)—is acknowledged for financial support. Part of the work was supported by the Austrian Center for Industrial Biotechnology (ACIB), which receives funding from the Federal Ministry of Science, Research and Economy (BMWFW), the Federal Ministry of Traffic, Innovation and Technology (bmvit), the Styrian Business Promotion Agency SFG, the Standortagentur Tirol, the Government of Lower Austria and ZIT—Technology Agency of the City of Vienna, through the COMET-Funding Program managed by the Austrian Research Promotion Agency FFG.