Introduction

The compound 2,4-diacetylphloroglucinol (DAPG) is a broad-spectrum antibiotic that was originally isolated from a fluorescent Pseudomonas strain1. DAPG is synthesized by a type III polyketide synthase cluster comprising the phlABCDE operon2,3. The synthesis is initiated by the phlD gene, which produces phloroglucinol from malonyl-CoA2. Phloroglucinol is subsequentially acylated to produce monoacetylphloroglucinol (MAPG), and a second acylation leads to the production of DAPG4.

Subsequently, this molecule can be used to suppress disease causing organisms in soil and is an important component of “disease suppressive soils”5 and in biocontrol applications6. DAPG is highly effective in controlling the causal agent of “take-all decline” in wheat (Gaeumannomyces tritici)7 and “tobacco black root rot” (Thielaviopsis basicola)8. In addition to controlling these fungal ascomycetes, DAPG has been reported to inhibit the oomycete Pythium ultimum9, the Gram negative bacterium Erwinia carotovora subsp. atroseptica10 and the Gram-positive bacterium Clavibacter michiganensis subsp. michiganensis11. The ecology and use of DAPG in biocontrol applications remains under intensive study12,13,14.

Chromobacterium species belong to the Gram-negative class of bacteria, Betaproteobacteria. Strains of Chromobacterium have been studied for many reasons such as their ability to be an opportunistic pathogen15 and as a model system for quorum sensing in Gram-negative bacteria16. Chromobacterium have been a rich source of bioactive molecules17,18,19. Recently, they have been used and developed as biocontrol agents to control insect pests20,21 and to control plant pathogens22,23,24,25.

In this study, while exploring the diversity of Chromobacterium in a Brazilian river system, we identified strains of Chromobacterium vaccinii that possessed a polyketide synthase biosynthetic cluster with homology to the phlACBDE operon26. This provided motivation to determine the distribution of this cluster in Chromobacterium spp. and confirm the bioactivity of the cluster.

Materials and methods

Strains

The type strain of C. vaccinii DSM 25150T was obtained directly from the DSMZ-German Collection of Microorganisms and Cell Cultures. Chromobacterium vaccinii CR1 and CR5 were previously isolated from aquatic environments in Brazil26. Both strains were isolated from raw surface water, which was collected directly in sterile flasks from the body of water between 0 and 30 cm water depth, in which the collection and preservation of the samples occurred in accordance to the National Collection Guide and preservation of samples: water, sediment, aquatic communities and liquid effluents27. C. vaccinii CR1 was isolated from João Leite stream, Goiânia, Goiás, Brazil—Geographic location 16° 34′ 30.54" S; 49° 13′ 55.02" W. C. vaccinii CR5 was isolated from the Meia Ponte river, Goiânia, Goiás, Brazil—geographic location 16° 54′ 16.3'' S; 49° 07′ 37.8'' W. The isolation details were according to Gomes et al.28.

The aquatic environment where strains CR1 and CR5 were isolated from is a river and its tributary. These waters are also impacted by many anthropological pressures, such as industrial, agro-industrial activities (livestock and an intense production of horticultural products), in addition to being used for leisure. The climate of the basin is tropical with two well-defined seasons, dry and rainy, belonging to the Cerrado biome28.

Genome mining for secondary metabolites and DAPG cluster distribution

The website version of the antiSMASH 6.0 software was used to identify putative secondary metabolite clusters in the genomes of C. vaccinii CR1, CR5, and DSM 25150T29. Clusters that were not reliably assigned in the antiSMASH 6.0 software were manually searched using BLAST software to determine a potential function. To determine the distribution of the DAPG cluster, the cluster (identified as GenBank LQR3212800..LQR32_12830) was used as a reference. These loci were manually searched using BLAST software against all available genomes (as of June 1, 2022) in the Chromobacterium genus. In addition to the biosynthetic genes of the DAPG cluster (phlACBDE), we searched for the regulatory gene associated with the cluster (phlF, phlG, phlH) commonly found in Pseudomonas strains, accession: CM001558 was used as a reference.

Phylogenetic analysis

The core genome for the Chromobacterium genus was determined and aligned using the BIGSdb program30. Aquitalea magnusonii DSM 23134T was included as an outgroup. The phylogenetic tree was constructed using MEGA 11 software31. The neighbor-joining tree was determined using the Tamura-Nei model (0.40, gamma distributed with invariant sites). Measures of bootstrap support for internal branches were obtained from 1000 pseudoreplicates. For the PhlD phylogeny, the PhlD protein was identified in the genome of C. vaccinii CR1 and searched using BLAST software against all available non-redundant proteins in GenBank. All proteins available (9) in the Chromobacterium genus were included in the alignment. Due to the large number of Pseudomonas proteins available, only a representative sample (11) were included. A Streptomyces sp. protein (PhiD) was used as an outgroup. All alignments were performed using MUSCLE implemented under MEGA 1131. The alignment was model tested in MEGA 11 and the Maximum Likelihood method with WAG model was selected. Measures of bootstrap support for internal branches were obtained from 1000 pseudoreplicates.

Construction of phlACB expression vector

Genomic DNA was extracted from 1 ml of an overnight culture of C. vaccini DSM 25150T as previously described by32. The putative phlACB cluster (~ 2.8 KB) was amplified from C. vaccini DSM 25150T genomic DNA using the primers cluster F (ATG AAG AAG GCA GGC ATA GTG AGC TAT GGC AG) and cluster R (ATC TTC CAG CAC GAA CTT GTA GGC GTA TTG CCA) using Invitrogen Platinum SuperFi II DNA polymerase (Thermo Fisher Scientific, Waltham, MA) according to the manufacturer’s instructions. The gel purified PCR product was reamplified with new primers containing the restriction sites Nco I (5ʹ end) and Hind III (3ʹ end) for cloning into the pET28b expression vector. Following ligation, three independent expression clones with the phlACB cluster were identified from plasmids that had been transformed into E. coli DH 5α. The correct sequence of the amplified phlACB cluster in each clone was verified by comparison to the genome sequence of C. vaccini CR1 following sequencing of each clone using a MiSeq DNA sequencer. The plasmid clones phlACB.1, phlACB.3 and phlACB.5, as well as empty vector pET28b, were individually transformed into BL21 E. coli.

Metabolism of DAPG

Each E. coli BL21 phlACB cell line, as well as the control BL21 cells with pET28b, was grown overnight in LB broth with kanamycin (50 µg/ml). The overnight culture was then used to inoculate M9 minimal medium (200 µl of inoculum into 10 ml media) with kanamycin (50 µg/ml) and grown at 37 °C with shaking (180 RPM) until the OD 600 reached 0.6. Afterwards, each culture received 1 mg of DAPG (dissolved in ethanol, Cayman Chemical, Ann Arbor, Michigan) and Isopropyl-β-d-1-thiogalactopyranoside (IPTG) for a final concentration of 0.2 mM, and incubated at 30 °C Aliquots of the cultures were removed at various time points and stored at 4 °C.

LC–MS analysis of Chromobacterium metabolites

For each time point (1, 2, 3, 4, and 7 days) and each media condition (LB or KMB), 5 μL of Chromobacterium culture media was analyzed by LC–MS using a Vanquish HPLC equipped with an ODS-18 column (3 mm × 150 mm, 3 μm particle size, Inertsil, GL Sciences, Inc., Torrance, CA) elution of analytes was accomplished by a linear gradient form 5:95–95:5% 18 MΩ water + 0.1% formic acid: Optima grade methanol + 0.1%formic acid at a flow rate of 0.550 ml/min over 30 min at 50 °C. Column effluent was monitored by UV at wavelengths 254 nm, 280 nm, and 270 nm prior to introduction into an Orbitrap ID-X- tribrid mass spectrometer (ThermoElectron, West Palm Beach, CA) under Xcalibur 4.4 control. Mass spectral data from 200 to 2000 Da m/z were collected in the orbitrap at 120,000 resolution. Tandem mass spectral data of target metabolites was collected using higher-energy collision dissociation (HCD, CE = 25 and 35) in the Orbitrap analyzer.

Analysis of E. coli cultures expressing phlACB

Aliquots of E. coli media at times 8, 20.5, 28.5, 48, and 68 h from three transformants (pET28-phlACB-1, -3, -5) and one plasmid-only control (pET28) were centrifuged at 10,000 RPM for 10 min. A sample of the supernatant (250 μL) was removed and diluted with 250 μL 18 MΩ water. A portion of this dilution (5 μL) was analyzed by LC–MS under the same LC conditions listed above, with the only modification to the mass spectrometric analyses was to lower the mass of detection to 100–1000 Da m/z to allow for the detection of phloroglucinol and MAPG. Authentic standards for DAPG, and phloroglucinol (Acros Organics, Geel, Belgium) were analyzed at 100 μg/ml to establish retention times and MS2 fragmentation patterns of expected products. Based on the manufacturer's specifications the level of detection of these analytes is approximately 100 pg/ml under these conditions.

Results

Genome mining environmental strains of Chromobacterium vaccinii

Chromobacterium vaccinii strains CR1 and CR5 were originally isolated from aquatic environments in the state of Goiás, Brazil26. The type strain of C. vaccini DSM 25150T was also included in the analysis33. Table 1 shows five clusters were found across all three strains. The antiSMASH software was only able to assign the function of the violacein and the potential DAPG clusters. A large 68.4 kb trans-AT-PKS cluster was identified using BLAST software to be the biosynthetic cluster of cyclic depsipeptide FR90035934. The final two small clusters (6.9 and 8.6 kb) possess domains that suggests they are probable siderophores.

Table 1 Biosynthetic clusters identified in Chromobacterium vaccinii strains.
Full size table

Distribution of biosynthetic clusters in the Chromobacterium genus

The initial discovery of the potential DAPG biosynthetic cluster in genomes of select C. vaccinii strains prompted the exploration of the distribution of this cluster in the Chromobacterium genus, given the historical importance of this molecule that is normally found in Pseudomonas spp. strains12. After a broader investigation using all available Chromobacterium genomes, the cluster was identified in 11 genomes in the genus that appeared in three main clades (Fig. 1). The results show the cluster was found in all C. vaccinii genomes and a closely related Chromobacterium piscinae strain. A second clade containing the cluster was identified in Chromobacterium sphagni. The third clade was consists of Chromobacterium amazonense and a closely related, recently described new species of Chromobacterium sinusclupearum35. The cluster had a broad geographical distribution with seven of the strains from the United States36,37,38, two from Brazil26, one from China (CP022344), and one from Malayasia39.

Figure 1
figure 1

Distribution of DAPG and depsipeptide FR900359 biosynthetic clusters in Chromobacterium genus. The phylogeny is based on the core genome of the strains listed in the figure. Measures of bootstrap support for internal branches were obtained from 1000 pseudoreplicates. Aquitalea magnusonii DSM 23134T was used as an outgroup and not shown.

Full size image

All the biosynthetic genes of the DAPG cluster (phlACBDE) were found in all of the Chromobacterium strains reported to contain the cluster. Interestingly, the Chromobacterium strains do not contain all the regulatory genes associated with this cluster in Pseudomonas strains. Typically, the DAPG cluster in Pseudomonas strains contain the phlF, phlG and phlH genes, which are known to regulate the expression of the cluster12. The Chromobacterium strains only contain the phlF gene and lack the phlG and phlH genes, which suggests they utilize a different regulatory mechanism than in Pseudomonas strains.

Additionally, the biosynthetic cluster for the depsipeptide FR900359 was also confined to the C. vaccinii genomes and one very closely related strain, C. piscinae XC0014. The violacein biosynthetic cluster was found in all genomes in Fig. 1, except for the two Chromobacterium phragmitis genomes.

Phylogeny of DAPG cluster

A phylogeny of the polyketide synthase protein PhlD, which is fundamental to the DAPG cluster, was created to understand the possible origin on the cluster. Figure 2 shows the cluster was closely related to strains in the Pseudomonas fluorescens group40. The average amino acid identity between the proteomes of C. vaccinii CR1 and the closest Pseudomonas proteome was 47% while the average amino acid identity between two DAPG clusters was 80%. In all Chromobacterium genomes containing the cluster, the cluster was in the same physical region of the genome and shared the same flanking regions. A survey of Chromobacterium genomes lacking the cluster did not find any alternative genes or clusters at this insertion site.

Figure 2
figure 2

Maximum likelihood phylogeny of PhlD protein from the DAPG biosynthetic cluster. Measures of bootstrap support for internal branches were obtained from 1000 pseudoreplicates.

Full size image

Confirmation of biosynthetic cluster assignment and activity

To confirm the assignment and activity of the biosynthetic clusters LC–MS/MS was used to confirm the presence of the metabolites in culture. The three assigned biosynthetic clusters were violacein, depsipeptide FR900359, and DAPG. Figure 3 shows confirmation for the assignment of violacein and depsipeptide FR900359 using high accuracy mass spectrometry. For all three strains the violacein signal was strong and the depsipeptide FR900359 was much weaker. A very weak signal for DAPG was observed in the strains, which was consistent with the retention time and mass spectrometry of a DAPG reference standard. Based on the results, it was assumed the DAPG cluster was active but not highly expressed under these culture conditions. We evaluated DAPG production under a variety of culture conditions (71 different carbon sources using Biolog plates) to determine if we could improve expression of DAPG activity. Under all the growth conditions tested, none of the conditions significantly improved the production of DAPG.

Figure 3
figure 3

LC–MS data from Chromobacterium vaccinii CR1. (A) Extracted ion-chromatograms of violacein (black-m/z 344.1–344.2 Da); and depsipeptide FR900359 (brown-m/z 1002.5–1002.6 Da); (B) mass spectrum of violacein; (C) tandem mass spectrum of violacein; (D) mass spectrum of FR900359; (E) tandem mass spectrum of FR900359.

Full size image

Confirmation of DAPG cluster activity through heterologous expression in E. coli

To confirm the activity of the Chromobacterium phlACB gene cluster in the metabolism of DAPG, the cluster was transformed into E. coli. It should be noted that expression of the phlACB genes from Pseudomonas fluorescens in E. coli resulted in the conversion of phloroglucinol into 2-acetylphloroglucinol (MAPG, 20%) and 3% DAPG in the culture medium2. Initial testing of E. coli expressing phlACB from Chromobacterium did not produce DAPG from phloroglucinol. Achkar et al. also found that incubation of the E. coli transformant with DAPG led to complete conversion of DAPG to MAPG (9%) and phloroglucinol (22%) in 48 h2. This motivated us to establish the activity of phlACB in the direction of DAPG to phloroglucinol. Therefore, cultures were grown in 100 μg/ml DAPG to determine the activity of the transformation of DAPG to MAPG and phloroglucinol. DAPG biotransformation samples, from triplicate transformations, were analyzed by LC–UV–MS/MS to confirm the identity of the products (Fig. 4). Samples over the course of time during incubation were monitored by UV at 270 nm (top panel of Fig. 4A–D), the maximum absorbance for DAPG. The UV signal and extracted ion chromatogram of DAPG remained constant for the control reaction where plasmid-only E. coli was the biocatalyst over 48 h (Fig. 4A,C). E. coli expressing phlACB showed a decrease in UV and EIC signal for DAPG over the course of 48 h, while having an increased signal for MAPG (14.30 min as determined by UV and accurate mass, < 1 ppm error, due to the lack of authentic standard) and phloroglucinol (at 48 h, 2.92 min based on accurate mass and comparison with authentic standard). At longer reaction times with E. coli expressing phlACB, DAPG and MAPG signals decreased and additional UV peaks appeared.

Figure 4
figure 4

LC–UV–MS data of control and transformed E. coli incubated with 100 μg/ml DAPG. Culture media was diluted 1:1 with 18MΩ water prior to analysis. Black-UV at 270 nm (λmax for DAPG); brown-extracted ion chromatogram from 211.05 to 211.07 m/z; green-extracted ion chromatogram from 169.00 to 169.20 m/z; blue-extracted ion chromatogram from 127.03 to 127.05 m/z. DAPG at 24.3 min, MAPG at 14.3 min, phloroglucinol at 2.9 min. (A) 8 h control. (B) 8 h phlACB. (C) 48 h control. (D) 48 h phlACB. E. Structures, masses, and pathway of DAPG degradation.

Full size image

Discussion

DAPG has been an important molecule in the bioactivities of rhizosphere colonizing Pseudomonas species and the subject of extensive investigations for 40 years12. It is for this reason we found it interesting to discover strains of C. vaccinii isolated from aquatic environments have the DAPG cluster26. The type strain of C. vaccinii and other representative isolates of C. vaccinii were originally isolated from the rhizosphere and water environments associated with native and cultivated cranberries (Vaccinium macrocarpon Ait.)33. This coincidence suggests that DAPG could provide a similar ecological function in Chromobacterium as seen in Pseudomonas species. In Pseudomonas species, DAPG has long been known to suppress phytopathogens and other microbes in the rhizosphere41. In addition, it was shown that DAPG containing Pseudomonas strains were a key component of disease suppressive soils in controlling the wheat root disease “take-all decline”5. It is possible that DAPG in Chromobacterium may play a similar ecological role for plants in bogs or similar aquatic environments. It is also noteworthy that these genera are only distantly related, and the closest shared clade is at the phylum level (Proteobacteria) with Chromobacterium belonging to the class Betaproteobacteria and Pseudomonas belonging to the class Gammaproteobacteria.

The distribution of the DAPG biosynthetic cluster in the Chromobacterium genus is confined to a few different clades. It is interesting to find all members of C. vaccinii, C. sphagni and C. sinusclupearum have this cluster (Fig. 1). However, these species have limited genome representation with 5, 2, and 2 whole genomes sequenced, respectively. This suggests there is a selection pressure to maintain the cluster in the species, notably in C. vaccinii where all strains are from diverse environments. The analysis also shows the DAPG cluster has broad geographical distribution with genomes from North America, South America, and Asia. In addition to the DAPG cluster, we identified two other biosynthetic clusters in the genomes of interest, depsipeptide FR900359 and violacein. Most strikingly, depsipeptide FR900359 occurred in tandem with DAPG in the C. vaccinii strains while violacein appeared in all Chromobacterium genomes with the exception of the two C. phragmitis genomes. It is not clear if DAPG and depsipeptide FR900359 are an adaptation to a specific ecological need or if their correlation is a coincidence.

The depsipeptide FR900359 was found to be a powerful and selective inhibitor of G protein-coupled receptors (GPCR) in humans and other mammals42,43. GPCRs are important elements for intracellular signal transduction which respond to external ligand binding and initial response cascades through guanosine triphosphate and guanosine diphosphate exchange44. Now depsipeptide FR900359 and closely related compounds, collectively known as chromodepsins, are important pharmacological tools for use as therapeutics and tools to study GPCR function45. Interestingly, chromodespins were initially isolated from a plant (Ardisia crenata) used in traditional medicine46. Later, it was discovered the compound was actually produced by an endosymbiotic bacteria belonging to the genus Burkholderia47. The ecological role of this compound is not well understood. Chromodespins demonstrate a lack of inhibition toward plant GPCRs43. It has been suggested that they may provide a benefit to their plant host by acting as a herbivore deterrent47. Interestingly, chromodespins have been shown to have insecticidal activity against bean bug (Riptortus pedestris), as well as inhibit the GPCRs of the pest insects white fly (Bemisia tabaci) and silk moth (Bombyx mori)45. This suggests strains of C. vacinnii may also be useful in the control of insect pests.

All Chromobacterium genomes except the two C. phragmitis genomes contained the purple pigmented compound violacein. Violacein has been reported to have a variety of bioactive properties. It has been shown to have antifungal properties against a large variety of phytopathogenic and human fungal pathogens48. It has been reported to have antibacterial activities49 and antinematicidal activities50. The compound showed antifeedant, larvicidal and pupicidal activities against Asian armyworm (Spodoptera litura Fab.)51. These broad-based bioactivities suggest that violacein may also contribute to the biocontrol potential of these Chromobacterium strains.

The phylogeny of the PhlD protein (Fig. 2) suggests all DAPG clusters in Chromobacterium share a common ancestor and the closest non-Chromobacterium ancestor likely originated in the Pseudomonas flourescens group. From the phylogenetic tree it is difficult to definitively infer if the cluster had one transfer event into Chromobacterium and some descendent lineages lost the biosynthetic cluster or if other horizontal gene transfer (HGT) events occurred within the Chromobacterium genus (e.g. a HGT from C. vaccinii to C. sinusclupearum).

In wild type C. vaccinii isolates, the production of DAPG could only be detected in trace amounts. Efforts to increase the production of DAPG through screening a variety of media conditions and carbon sources were unsuccessful. It may be possible that molecules in aquatic environments induce the biosynthesis of DAPG in Chromobacterium. Nevertheless, we confirmed that the biosynthetic genes in Chromobacterium for DAPG production were intact and functional. Our results also suggest the regulatory mechanism of Chromobacterium strains is significantly different than Pseudomonas strains since they lack the known regulatory genes phlG and phlH12. In addition, they lack the Pseudomonas post translational regulatory genes rsmA and rsmE (data not shown)12. More experiments will need to be performed in the future to discover the regulatory mechanism of DAPG production in Chromobacterium. We assumed the lack of activity was due to the lack of expression of these genes. We transferred the phlACB cluster into E. coli to have direct control over the expression of these genes. Initial testing of the transformant failed to produce DAPG from phloroglucinol. This is consistent with previous observations that expression of the phlACB cluster in E. coli produced no or trace amounts of DAPG2,52. However, in past studies the reverse reaction, conversion of DAPG to MAPG and phloroglucinol was active2,52. Monitoring of the reverse reaction showed a time-dependent decrease in the amount of DAPG present in the media when phlACB was expressed. There was an increase in the UV270nm at retention times corresponding to the exact mass of MAPG and phloroglucinol (Fig. 4E) demonstrating the cluster could convert DAPG to MAPG and phloroglucinol. It should be noted in experiments where E. coli expressing phlACB transformants were incubated with phloroglucinol in an attempt to produce DAPG, the phloroglucinol used contained a small DAPG (< 1%) impurity. When incubated with the E. coli expressing phlACB transformants, this DAPG impurity was converted to phloroglucinol and no reaction was observed in the direction of DAPG formation. It is unknown why phlACB transformants favored the production of phloroglucinol and the biotransformation to DAPG did not occur. However, Liu et al.52 recently demonstrated enhanced production of DAPG in E. coli expressing phlACB transformants through co-expression of acc, marA and phlE genes.

A survey of literature identified a few examples of strains of Chromobacterium being evaluated as potential biocontrol agents to control plant pathogens22,23,24,25. Two of the studies use Chromobacterium haemolyticum C-61, which displayed strong antifungal activities against several plant pathogens23,24. These studies identified another novel cyclic lipopeptide which they named chromobactomycin, which was responsible for the observed activity24. The genome is available for this strain, and it does not contain the DAPG cluster. Han et al. identified Chromobacterium sp. JH7 during a screening assay to identify antagonists of Cylindrocarpon destructans, a root rot pathogen of ginseng22,23,24,25. The authors were not able to identify the strain to the species level. It is possible that DAPG or depsipeptide FR900359 may be associated with this activity, but there is currently no data to support this speculation. In a recent study, C. vaccinii was shown to inhibit several fungal plant pathogens through emission of bioactive volatiles22,23,24,25. We could not find any reported DAPG activity in Chromobacterium sp. in the literature.

Our results confirm the presence and activity of the DAPG biosynthetic cluster in the Chromobacterium genus. These findings and the limited investigations of the biocontrol activity against plant pathogens by the Chromobacterium genus suggests additional research in this area is warranted.