Genomic insights into the metabolic complementarity of a synthetic marine consortium for hydrocarbon degradation in coastal sand microcosms

Abstract

Microbial consortia are essential for the degradation of complex hydrocarbon mixtures in coastal environments, yet the functional basis of consortium performance rarely resolved at strain level. In this study, we employed a Plackett–Burman experimental design to rationally assemble and optimize a synthetic consortium for the degradation of medium-grade crude oil in a sand–seawater microcosm. Six strains were selected: Stutzerimonas stutzeri CBC1 and CBC2, S. balearica GOM11, Pseudomonas sp. CBC3, Halomonas litopenaei GOM23 and Enterobacter kobei GOM24. After 45 days, the consortium removed 15.72% of total hydrocarbons relative to the abiotic control, with markedly higher fraction-specific removal determined by GC–MS: 94.7% of phenanthrene, 63.1% and 41.0% of the C8-C27 and C18-C27 aliphatic fractions respectively. Comparative genomic analysis revealed a non-uniform distribution of catabolic functions consistent with metabolic complementary. Stutzerimonas and Pseudomonas strains carried ring-hydroxylating dioxygenases (nah/nag-like clusters) together alkane-degrading alcohol/aldehyde dehydrogenases (alkJ, alkH), whereas H. litopenaei GOM23 and E. kobei GOM24 encoded alternative long-chain alkane-hydroxylating systems (ladA). GOM24 further carried an expanded repertoire of ABC transporters and oxidative-stress genes (ahpC/ahpF), consistent with a role in substrate uptake and stress tolerance rather than in primary ring cleavage. Enzymatic assays supported this partitioning, catechol 1,2-dioxygenase activity was confined to the same four strain that encoded the gene and was highest in CBC1 and GOM11 (0.0472 and 0.0443 U·mg⁻¹ protein, respectively). No genome carried the classical alkB alkane-monooxygenase, pointing to alternative and partially redundant alkane-oxidation routes. By combining statistically guided assembly with comparative genomics and targeted enzyme assays, this work identifies a six-strain combination that removes more crude oil than the full eleven-isolate community and resolves how its catabolic functions are partitioned among strains.

Supplementary Information

The online version contains supplementary material available at https://doi.org/10.1007/s10532-026-10379-7.

Introduction

Oil spills are recurrent disturbances in marine ecosystems, particularly in coastal zones where sand can retain petroleum for extended periods (Zhang et al., 2018). Sandy beaches, which account for more than 31% of the world’s ice-free coastlines, are highly susceptible to contamination because of their high permeability (Pontee 2013; Luijendijk et al. 2018).

Once released into marine ecosystems, spilled oil is transported by currents and waves, increasing the impacted area (Weisberg et al. 2017). Following the Deepwater Horizon (DWH) spill, weathered oil residues accumulated along 1000 km of the Gulf of Mexico shoreline, with up to 2 kg of polycyclic aromatic hydrocarbons (PAHs) per meter of beach and their long-term burial within sediment layers (Huettel et al. 2018). Oil reaching coastal sediments undergoes weathering, yielding residues with altered physicochemical properties that influence oil bioavailability (Bacosa et al. 2015; Huettel 2022).

Both sand grain size and oil viscosity correlate with oil infiltration, as coarser sands and lower viscosity oil favor deeper penetration. Environmental factors such as salinity further modulate PAH solubility and adsorption, thereby affecting biodegradation rates (Wang et al. 2021). Interactions with co-occurring pollutants including microplastics can additionally modify hydrocarbon persistence and distribution (Padervand et al. 2020).

On beaches, weathered oil can form sand-associated aggregates that are resistant to degradation and may even protect the oil from further weathering by limiting exposure to direct sunlight, thereby hindering photodegradation, which has been identified as an essential component of oil breakdown (John et al. 2016).

Hydrocarbon degradation in coastal sands is primarily mediated by specialized bacterial taxa, including Alcanivorax, Halomonas, Marinobacter, Pseudomonas and Acinetobacter (Shin et al. 2019). Petroleum biodegradation relies on the enzymatic capacity of bacteria to activate hydrocarbons via oxygenation and generate intermediates that subsequently enter the central metabolic pathway (Aziz et al. 2024). The process involves multiple stages including hydrocarbon contact, transmembrane transport, and intracellular transformation (Xu et al. 2018).

Petroleum is a complex mixture of alkanes, and aromatic compounds such as polycyclic aromatic hydrocarbons (PAHs), many of which are recalcitrant to degradation by individual microorganisms. Consequently, no single organism can fully mineralize all hydrocarbon fractions. Instead, microbial communities outperform individual isolates due to metabolic complementarity, as different species possess distinct pathways and mechanisms for hydrocarbon degradation. This synergistic metabolic activity enhances the overall efficiency of petroleum biodegradation while also contributing to greater ecological stability (Mekonnen et al. 2024; Chunyan et al. 2023; Parhamfar et al. 2020).

Rationally designed synthetic consortia, optimized using statistical approaches such as the Plackett–Burman (PB) fractional factorial model, provide a powerful framework for identifying strain combinations that synergistically enhance pollutant biodegradation (Magallanes and Olivieri 2010; Mazzucotelli et al. 2014; Rojas-Vargas et al. 2022). These approaches reduce experimental complexity while identifying critical interactions within microbial systems.

Although previous studies have demonstrated the potential of microbial consortia for petroleum hydrocarbon biodegradation, several limitations remain. Many consortia are assembled empirically or based on taxonomic diversity, without systematically evaluating the contribution of individual strains to overall degradation performance. In addition, degradation efficiency is frequently assessed without linking hydrocarbon removal patterns to the functional genomic potential of the consortium members. Consequently, the relationship between statistically selected strain composition, degradation of specific hydrocarbon fractions, and the distribution of catabolic functions among consortium members remains insufficiently resolved, particularly in coastal sand–seawater systems.

We hypothesized that a statistically selected marine bacterial consortium would exhibit greater crude-oil degradation performance than non-optimized strain combinations and that this performance would be associated with a non-uniform distribution of hydrocarbon-degradation and stress-response functions among its constituent strains. In this study we aimed to identify an effective combination of marine bacterial isolates for medium-grade crude-oil degradation using a Plackett–Burman screening design; characterize the aliphatic and aromatic hydrocarbon fractions removed by the selected consortium using gas chromatography–mass spectrometry (GC–MS); and determine whether the genomes of the selected strains contain complementary catabolic and stress-response functions potentially associated with the observed degradation profile. Functional genomic predictions for aromatic ring cleavage were further supported by measuring catechol 1,2-dioxygenase activity in individual consortium members.

Materials and methods

Bacterial isolation and culture conditions

Eleven marine isolates were obtained from Gulf of Mexico and Pacific Ocean water, sediment, and tar samples. Bacteria were cultured in modified Burk’s mineral medium (MM) adapted for the cultivation of marine isolates (Rojas-Vargas et al. 2022) containing 0.183 g·L^-1 Na₂SO₄, 0.073 g·L^-1 CaCl₂, 0.267 g·L^-1 NH₄Cl, 0.16 g·L^-1 MgCl₂·6H₂O, 0.0002 g·L^-1 Na₂MoO₄·2H₂O, 0.005 g·L¹ FeSO₄, 11.68 g·L^-1 NaCl, 0.8 g·L^-1 K₂HPO₄, and 0.2 g·L^-1 KH₂PO₄ supplemented with 0.05% peptone. The NaCl concentration together with the other inorganic salts in the medium, provides the necessary osmotic pressure and ionic strength required for the optimal growth of marine and estuarine bacteria used in this study. All isolates were able to grow in this saline medium. Strains were prescreened in seawater supplemented with petroleum as the sole carbon source. The halotolerance of the six consortium isolates was confirmed across a NaCl gradient (0–50 g·L⁻¹; Supplementary Fig. S3), demonstrating optimal growth at marine salinities (11.6–30 g·L⁻¹ NaCl).

Sand and seawater preparation

Approximately 10 kg of sand was collected from a beach with no reported history of hydrocarbon contamination in Copala, Guerrero, México (N 16° 28′ 30.468''W, 98° 51′ 46.645''). The sand was sieved through 2 mm mesh, and sterilized by five autoclaving cycles (121 °C, 15 min each). Additionally, ~ 10 L of seawater was collected from the same site, filtered through a 0.22 µm membrane, and subsequently heat-sterilized.

Pathogenicity assays with Galleria mellonella

The pathogenicity of the eleven marine isolates evaluated in this study were analyzed using Galleria mellonella larvae as a model, following the protocol proposed by Loh et al. (2013). Cream-colored larvae (n = 10 per treatment; three biological replicates) approximately 3 cm in size with a healthy appearance, as indicated by swarm formation and mobility, were injected in the last proleg with 10 μL of inoculum containing ~ 100 colony-forming unit (CFU) of each isolate suspended in phosphate buffer. The negative controls included noninjected larvae, buffer-injected larvae and larvae injected with E. coli DH5α; the positive control strain was P. aeruginosa PAO1. Survival was monitored for 6 days at 30 °C, considering the following parameters: activity, cocoon formation, melanization, and survival.

Crude oil biodegradation in microcosms

The microcosm consisted of 40 g of sterilized sand and 10 ml of sterilized seawater in 125 ml Erlenmeyer flasks. Medium-grade crude oil was added to a final content of 0.1% (w/v) as the sole carbon source, along with nitrogen (NH₄NO₃) and phosphorus (KH₂PO₄ and K₂HPO₄) at a C:N:P ratio of 100:5:1.

For inoculum preparation, each isolate was grown individually in modified Burk’s mineral medium (MM) supplemented with 0.05% peptone at 30 °C for 24 h at 180 rpm. Bacterial cells were harvested by centrifugation (4000×g for 10 min), washed twice with sterile 10 mM MgSO₄ solution to remove residual nutrients, and resuspended in the same solution. Cell suspensions were standardized to an optical density at 600 nm (OD600) corresponding to 1×10⁸ CFU·mL⁻¹.

The washed cell suspensions were inoculated into the seawater–sand microcosm according to the strain combination defined by the PB design, at final concentration of 1×10⁸ CFU·mL⁻¹ per isolate. The microcosms were incubated at 30 °C for 45 days, with weekly aeration performed by manual shaking. Biomass was estimated on the basis of protein content using the method of Lowry et al. (1951). No hydrocarbon interference or protein precipitation was observed during the assay.

Plackett–Burman experimental design

A 12-run Plackett–Burman (PB) matrix was constructed to evaluate the individual effect of each of the 11 marine bacterial isolates on crude oil degradation. The 11 independent variables were assigned alphabetical codes (A–K), each corresponding to the presence or absence of a specific bacterial isolate listed in Table 1 (A: Dietzia sp. GOM26, B: S. stutzeri CBC1, C: Halomonas sp. GOM23, D: A. xenomutans GOM8, E: Enterobacter sp. GOM24, F: Stutzerimonas balearica GOM11, G: S. stutzeri CBC2, H: H. aestusnigri GOM5, I: Paenarthrobacter sp. GOM3, J: Stutzerimonas sp. GOM25, K: Pseudomonas sp. CBC3).

Table 1.

Plackett–Burman experimental design matrix

Codes	Bacterial isolate
A	Dietzia sp. GOM26
B	S. stutzeri CBC1
C	Halomonas sp. GOM23
D	A. xenomutans GOM8
E	Enterobacter sp. GOM24
F	Stutzerimonas sp. GOM11
G	S. stutzeri CBC2
H	H. aestusnigri GOM5
I	Paenarthrobacter sp. GOM3
J	Stutzerimonas sp. GOM25
K	Pseudomonas sp. CBC3

Run	A	B	C	D	E	F	G	H	I	J	K	TPH removal (%± SD)
1	+1	+1	− 1	+1	+1	+1	− 1	− 1	− 1	+1	− 1	3.44 ± 0.33
2	− 1	+1	+1	− 1	+1	+1	+1	− 1	− 1	− 1	+1	15.72 ± 0.78
3	+1	− 1	+1	+1	− 1	+1	+1	+1	− 1	− 1	− 1	12.35 ± 0.37
4	− 1	+1	− 1	+1	+1	− 1	+1	+1	+1	− 1	− 1	8.33 ± 0.76
5	− 1	− 1	+1	− 1	+1	+1	− 1	+1	+1	+1	− 1	13.20 ± 0.62
6	− 1	− 1	− 1	+1	− 1	+1	+1	− 1	+1	+1	+1	0.12 ± 0.13
7	+1	− 1	− 1	− 1	+1	− 1	+1	+1	− 1	+1	+1	3.65 ± 0.22
8	+1	+1	− 1	− 1	− 1	+1	− 1	+1	+1	− 1	+1	2.29 ± 0.36
9	+1	+1	+1	− 1	− 1	− 1	+1	− 1	+1	+1	− 1	0.79 ± 0.11
10	− 1	+1	+1	+1	− 1	− 1	− 1	+1	− 1	+1	+1	5.20 ± 0.16
11	+1	− 1	+1	+1	+1	− 1	− 1	− 1	+1	− 1	+1	0.24 ± 0.47
12	− 1	− 1	− 1	− 1	− 1	− 1	− 1	− 1	− 1	− 1	− 1	1.31 ± 0.24
13	+1	+1	+1	+1	+1	+1	+1	+1	+1	+1	+1	7.63 ± 0.13

Eleven independent variables (A–K) representing different bacterial isolates were tested across 12 experimental runs, including a 13th run with the full consortium

Presence (+ 1) and absence (− 1) of each strain are indicated

The response variable is expressed as TPH Removal (%± SD), determined by gravimetric analysis after 45 days of incubation

The presence of an isolate in each experimental run was coded as + 1 and its absence as − 1. When present, each isolate was inoculated at a final concentration of 1 × 10⁸ CFU·mL⁻¹ suspended in sterile 10 mM MgSO₄). An abiotic control without bacterial inoculum and a 13th run containing all 11 isolates were also included (Table 1). Each experimental condition was performed in quintuplicate (n = 5). Crude oil removal efficiency (%) determined gravimetrically after 45 days of incubation, was used as the response variable. No time-course degradation measurements or kinetic model fitting were performed; therefore, the statistical parameters derived from the PB analysis describe the fit of the screening model and do not represent biodegradation rate constants or kinetic parameters.

Screening for significant factors using Plackett–Burman design

The Plackett–Burman design was used to find the most significant parameters for oil degradation, and the percentage removal was used as the response variable. Statistical analysis of the data was conducted using a Pareto chart, and analysis of variance (ANOVA) was performed using Design-Expert software v7.0.0 (Stat-Ease Inc., MN, USA). A total of 13 experimental runs were generated; the gravimetric removal percentages determined from five replicates were averaged and used for statistical analysis.

Hydrocarbon extraction

Hydrocarbon extraction for gravimetric determination was carried out using a mechanical agitation method with modifications. For GC‒MS analysis, after 45 days of incubation, the liquid phase of the microcosms was evaporated at 30 °C. Solid‒liquid extraction of total HCs was performed by adding 10 mL of HPLC-grade dichloromethane (DCM) (CH₂Cl₂; Sigma‒Aldrich Inc., MO, USA) to each microcosm composed of sand, seawater and oil, which was recovered in 50 mL PYREX round-bottom tubes that had been previously washed three times with DCM. The sand was subsequently washed five more times with 3 mL of DCM, and each wash solution was added to the recovered solvent. The tube was left overnight at 4 °C to achieve phase separation.

The organic phase was recovered manually using a Pasteur pipette and evaporated to a volume of approximately 1 mL using a Rotavapor R-300 (Büchi, Flawil, Switzerland). The recovered solution was then transferred with a Pasteur pipette to a glass column packed with anhydrous sodium sulfate and fiberglass. The extract obtained was recovered and evaporated in Wheaton glass vials, and three more washes were performed to ensure full oil recovery. The percent medium oil removal was determined by gravimetry and calculated using the following formula:

$$Oilremoval\left(\%\right)=\left(\text{X}-\text{Y}\right)/\text{X}\times 100$$

where X is the original mass of the oil added to the microcosm and Y is the mass of the residual oil.

Extraction of total hydrocarbon oil via accelerated solvent extraction

After 45 days of incubation for GC‒MS analysis, the liquid phase of the microcosms was evaporated at 30 °C. Solid‒liquid extraction was performed using 3 g of solid from each flask via accelerated solvent extraction (ASE) with a DIONEX ASE 350 and DCM as the solvent. Afterward, the internal standards were added, and the samples were heated to 100 °C for 5 min for a total of 5 heating cycles.

Gas chromatography‒mass spectrometry

The extracts were concentrated to approximately 2 ml using a SyncorePlus concentrator (Büchi) at 37 °C and 60 rpm. The samples were then subjected to low-flow nitrogen in Wheaton amber glass vials to achieve a final volume of 1 ml. The extracts were diluted and injected onto an Agilent 6890 N gas chromatograph equipped with a 30 m HP-5MS column using helium as the carrier gas. Mass spectra were analyzed using built-in Mass Hunter software. For GC–MS chemical profiling, four independent biological microcosms (n = 4) were processed per treatment. The GC‒MS results were used to determine crude oil biodegradation levels. The target compounds identified and quantified included aliphatic hydrocarbons (from n-octane to n-tetracontane, including pristane and phytane) and aromatic hydrocarbons (including naphthalene, phenanthrene, anthracene, and their alkylated derivatives). Additionally, the Bacab alfa oil fractions used in this study were characterized by GC‒MS (Supplementary Table 1).

GC–MS qualitative screening

Qualitative metabolite profiling of organic extracts was performed using a GC–MS system operated with Agilent MassHunter software. Data acquisition and qualitative processing were conducted using internal standardized protocols for organic compound screening. Chemical identification of the chromatographic peaks was tentatively assigned by mass spectral matching against qualitative mass spectral libraries. The analysis was conducted as an untargeted qualitative screening without chemical derivatization.

Catechol 1,2-dioxygenase activity assay

Catechol 1,2-dioxygenase (C12O, Enzyme Commission number (EC) 1.13.11.1) activity was determined spectrophotometrically in cell-free total protein extracts from cultures grown in Burk’s mineral medium supplemented with benzoate (1 g L⁻¹), tracking the cleavage of catechol to cis,cis-muconate at 260 nm, as described by Rodríguez-Salazar et al. (2020). The reaction mixture (0.2 mL) contained 50 mM phosphate buffer (pH 7.5), 20 μg of total protein extract, and 50 μM catechol as substrate. Assays were conducted at 30 °C. Protein extract from the reference strain GOM2 was included as a positive control, while uninduced Escherichia coli DH5α served as a negative control. One unit (U) of enzyme activity was defined as the amount of enzyme converting 1 μmol of catechol to cis,cis-muconate per minute (ε260​ = 16,800 M⁻¹ cm⁻¹). Specific activity was expressed as units per milligram of total protein (U⋅mg⁻¹ protein).

Bacterial genotyping by 16S rRNA analysis

Each isolate, which had been preserved in glycerol at − 80 °C, was cultured overnight in 5 mL of LB medium in a test tube at 30 °C and 180 RPM. Cells were harvested by centrifugation, and the pellet was used for DNA extraction. Bacterial DNA was extracted using the Quick-DNA™ Miniprep Kit (Zymo Research, Irvine, CA, United States). Polymerase chain reaction (PCR) was carried out with Taq DNA polymerase (NZYTech, Lisbon, Portugal) using the following primers to amplify the complete 16S rRNA gene: 27F (AGAGTTTGATCMTGGCTCAG) and 1492R(I) (GGTTACCTTGTTACGACTT).

The PCR products were band-purified with a Zymoclean Gel DNA Recovery Kit (Zymo Research) and subsequently sequenced using the Sanger method at the DNA Synthesis and Sequencing Unit of the Institute of Biotechnology (IBt), UNAM, Mexico.

Whole-genome sequencing

The genomes of previously unsequenced strains that significantly positively affected oil degradation were sequenced. DNA was extracted from each isolate using the Quick-gDNA Miniprep Kit (Zymo Research). Genome sequencing and assembly were performed according to the protocols of Millán-López et al. (2024). Briefly, the genomes were sequenced using an Illumina MiSeq platform with a 600-cycle kit and a 2 × 300 bp configuration. All bioinformatic tools and software were executed using default parameters unless explicitly stated otherwise. Genome assembly was performed using DISCOVAR de novo (v52488) (Weisenfeld et al. 2014), and quality was analyzed with Quast (v5.2.0) (Gurevich et al. 2013). Gene prediction and annotation were performed using Bakta (v1.8.2) (Schwengers et al. 2021) with its full database. Completeness and contamination tests were performed with CheckM (v1.2.2) (Parks et al. 2015). The average nucleotide identity (ANI) test was conducted with PyAni software (v0.2.12) (Pritchard et al. 2016), using MUMmer.

Bioinformatic search for petroleum degradation-related genes

Genes associated with hydrocarbon degradation were identified using Basic Local Alignment Search Tool (BLAST)-KOALA (Kanehisa et al. 2016), and metabolic pathways were mapped with Kyoto Encyclopedia of Genes and Genomes (KEGG) Mapper (Kanehisa and Sato 2019). RAST (Aziz et al. 2008) and Hydrocarbon Aerobic Degradation Enzymes and Genes (HADEG) (Rojas-Vargas et al. 2023) were used with a 30% threshold to validate gene annotations related to hydrocarbon degradation.

To provide all the data about the annotations, we additionally performed BLASTP (v2.17.0) (Camacho et al. 2009) searches of each genome against the HADEG protein database and against a set of genes not represented in HADEG (stress-response, membrane-transport, and other aromatic metabolism genes) obtained from searching the functional annotation in OrthoDB (Zdobnov et al. 2020), with a 30% identity threshold, a minimum 50% query coverage and an E-value ≤1e−5. Hits are reported in Supplementary Table 2. The reference set was assembled to cover representative functional annotations reported in the analysis that are not represented in the HADEG database like oxidative stress, chaperones, heat-shock/osmotic stress, and membrane transport.

Statistical analyses and visualization

The 16S rRNA sequences were aligned using MUSCLE5 (Edgar 2022). Phylogenetic reconstruction was performed using the maximum likelihood method implemented in IQ-TREE (v2.0.7) (Minh et al. 2020). Branch support was assessed using 1000 ultrafast bootstrap replicates. The genome representation plots were constructed with the Circular Plot tool from Artemis (18.0.2) (Carver et al. 2011). Statistical analyses were performed in R version 4.2.3, and the ggplot2 package version 3.3.2 (Wickham 2016) was used to create the figures.

Results

Origin and selection of marine bacterial isolates

Stutzerimonas stutzeri CBC1, S. stutzeri CBC2 and Pseudomonas sp. CBC3 were previously identified as hydrocarbonoclastic bacteria that can metabolize oil and degrade monoaromatic hydrocarbons (Millán-López et al. 2024). The selected strains, Paenarthrobacter GOM3, Halopseudomonas aestusnigri GOM5, and Alloalcanivorax xenomutans GOM8, were identified during the formation of a consortium for the removal of light oil from seawater (Rojas-Vargas et al. 2022). To broaden the metabolic versatility of the consortium, five additional marine isolates (GOM11, GOM23, GOM24, GOM25, and GOM26) were included based on preliminary assays showing their specialized catabolic potential. Specifically, GOM11, GOM23, and GOM26 were able to grow on phenol (5 mM) and light crude oil (0.1%), whereas GOM23 and GOM25 were able to grow on phenanthrene. Furthermore, GOM25 was identified as a specialist in the degradation of high-molecular-weight PAHs, such as pyrene, and GOM24 was confirmed to grow on light crude oil (0.1%). The geographical origins and environmental sources of all the isolates are summarized in Table 2.

Table 2.

Environmental origin and geographical coordinates of the eleven marine bacterial strains evaluated in this study

Source	Sampling
	Isolate	Marine coordinates (GMS)	Depth (m)
Tar	Dietzia sp.	8°31′12.0"N 93°23′24.0"W	246
Water	S. stutzeri CBC1	32°21′36.0"N 117°04′48.0"W	20
Sediment	Halomonas sp. GOM23	19°00′49.8"N 95°17′51.0"W	308
Sediment	A. xenomutans GOM8	19°30′00.0"N 96°00′00.0"W	307
Sediment	Enterobacter sp. GOM24	18°58′47.3"N 93°29′05.6"W	448
Water	Pseudomonas sp. GOM11	24°00′00.0"N 86°48′00.0"W	1233
Water	S. stutzeri CBC2	32°21′53.4"N 117°04′42.4"W	20
Tar	H. aestusnigri GOM5	18°45′36.0"N 94°15′36.0"W	350
Sediment	Paenarthrobacter sp. GOM3	18°26′24.0"N 94°18′00.0"W	275
Water	Stutzerimonas sp. GOM25	24°00′00.0"N 87°42′00.0"W	1233
Water	Pseudomonas sp. CBC3	32°21′53.4"N 117°04′42.4"W	20

The halotolerance profile of each strain was systematically evaluated across 0 to 50 g⋅L⁻¹ NaCl (Supplementary Fig. 3). All isolates sustained robust growth under marine-equivalent salinities (11.6–30 g⋅L⁻¹ NaCl), with Pseudomonas sp. CBC3 and Halomonas litopenaei GOM23 displaying halotolerant kinetics (OD600​ >0.8). All strains maintained active biomass accumulation at seawater salinity (~ 30 g⋅L−1 NaCl), therefore, osmotic inhibition is therefore unlikely to explain the undergraduate fraction in the crude oil microcosms, reflecting instead the chemical persistence of heavy branched and complex aromatic oil residues.

Pathogenicity screening and safety assessment of the consortium isolates in the Galleria mellonella model

Survival analysis of G. mellonella larvae revealed that none of the eleven isolates exhibited pathogenic effects during the 6-day assay. The survival patterns of the larvae were comparable to those observed in the nonpathogenic control group (E. coli DH5α), confirming that the eleven isolates lack virulence traits under the conditions evaluated. In contrast, the positive control, Pseudomonas aeruginosa PAO1, resulted in 100% larval mortality within 18 h (Supplementary Fig. 1).

Selection and design of the synthetic consortium using Plackett–Burman

The 11 marine isolates were randomly distributed according to a 12-run Plackett–Burman (PB) screening design (Table 1), where the presence (+ 1) and absence (− 1) of each isolate were systematically varied to evaluate their individual contributions to crude oil degradation after 45 days of incubation. An additional run (Run 13), containing all 11 isolates, was evaluated independently of the main PB matrix to assess the baseline performance of the unselected community.

Gravimetric analysis revealed that endpoint crude oil removal after 45 days varied markedly across the 13 experimental runs, ranging from 0% to 15.72% ± 0.78%. Notably, Runs 6, 9, and 11 exhibited no detectable oil degradation, indicating strong heterogeneity in inter-strain interactions. Furthermore, Run 13 (full community) resulted in only 7.63% removal. The highest degradation efficiency (15.72%) was achieved in Run 2, exceeding both the abiotic control and the full consortium run.

Statistical analysis of the PB screening model (Fig. 1) revealed positive (blue) and negative (red) main effects on final oil removal. The mathematical model was validated by analysis of variance (ANOVA), which was highly significant (F = 8.60, p = 0.0160). The coefficient of determination (R² = 0.9117) indicated that PB model accounted for 91.17% of the observed variability in endpoint oil removal, demonstrating a strong correlation between inoculum composition and degradation efficiency. The strains with positive effects were E (Enterobacter sp. GOM24; F = 12.70, p = 0.0161), F (Stutzerimonas balearica GOM11; F = 11.37, p = 0.0181), and C (Halomonas sp. GOM23; F = 10.22, p = 0.0241); thus, these strains can be considered key variables for process optimization, as they had the greatest effects and exceeded the critical t-value of 2.57. Run 2, which achieved the greatest degradation performance and included all three strains with significant contributions, was selected as the optimized configuration for consortium construction. The remaining factors (B, D, G, H, J, A, I, and K) did not significantly affect the oil degradation system.

Fig. 1.

Effects of bacterial strains on medium crude oil degradation. The statistical evaluation of each consortium member was performed through an Plackett–Burman (PB) experimental design. Blue bars indicate a positive synergistic effect on degradation, while red bars represent a negative effect or competition under the tested conditions. The standardized effect values prioritize the strains with the most significant impact on the overall metabolic performance of the consortium

Furthermore, all six key strains exhibited broad halotolerance with optimal growth at 11.6–30 g·L⁻¹ NaCl (Supplementary Fig. S3), suggesting that salinity was unlikely to be main constraint on degradation performance.

Taxonomic identification and phylogenetic diversity

16S rRNA gene analysis revealed that the selected isolates belonged primarily to the Gammaproteobacteria class (72.7%), with a relatively small representation of Actinomycetia (Fig. 2). Among the characterized isolates, the Stutzerimonas and Pseudomonas genera were the most frequent, along with members of Dietzia, Halomonas, Alloalcanivorax, Paenarthrobacter, and Enterobacter.

Fig. 2.

Maximum likelihood phylogenetic tree based on the 16S rRNA gene sequences of the 11 marine isolates used in the Plackett–Burman design. Colored blocks indicate bacterial classes, with Gammaproteobacteria representing the most abundant group (72.7%). Isolates selected for the final optimized consortium (Trial 2) are highlighted with the "Consortium" label. Branch support was assessed using 1000 bootstrap replicates

Chemical profiling of crude oil degradation

After 45 days of incubation, the metabolic activity of the synthetic consortium (conformed by 6 strains) significantly altered the hydrocarbon profile of the Bacab-Alpha crude oil. GC‒MS analysis revealed distinct degradation patterns in the aliphatic fractions (Fig. 3A).

Fig. 3.

Relative hydrocarbon degradation profile after 45 days of incubation. Bars represent the mean degradation values (n = 4) calculated as the reduction in concentration relative to the abiotic control (Control—Treated)/Control, as determined by GC–MS. A Aliphatic degradation and B aromatic degradation. For aliphatic fractions, individual data points for each biological replicate are overlaid on the bars. Statistical significance was assessed using Student's t-test: double asterisks (**) indicate p < 0.01 (highly significant), while a single asterisk (*) indicates p < 0.05 (significant)

The light-chain fraction (C8-C17) exhibited the most significant reduction, achieving a mean relative removal of 63.1% ± 10.4% compared with that of the abiotic control. The degradation values of individual replicates within this group were as high as 84.1%, highlighting the catabolic potential of the consortium.

The medium-chain fraction (C18–C27) followed a similar trend, with a reduction of 41.0% ± 10.1%, while the heavy aliphatic fraction (C28-C40) was the most recalcitrant, with a lower mean removal of 16.9% ± 11.2%. Despite the inherent heterogeneity of the sand matrix, reflected in the dispersion of individual data, the inoculated microcosms consistently had lower residual concentrations than the abiotic control did.

Statistical analysis (Student's t test, n = 4) revealed significant removal of light and medium aliphatic fractions: n−C18​ to C27​ range (p = 0.0075) and n−C8​ to C17​ fraction (p = 0.0397). The heavy fraction (n−C28​ to C40​) showed no significant difference (p = 0.1322).

Analysis of the aromatic fraction revealed that compounds such as naphthalenes, methylnaphthalenes, biphenyls, acenaphthylene, and high-molecular-weight benzo compounds were below the detection limit in the inoculated microcosms. Total PAH removal, calculated relative to the abiotic control, reached 86.4%. This significant reduction was primarily driven by the efficient degradation of phenanthrene, which reached 94.7% relative to that in the abiotic control group, with a significant p value of 0.0413, indicating efficient degradation of tricyclic PAHs (Fig. 3B).

Qualitative GC–MS analysis of the organic extracts revealed a highly reproducible chemical profile across all analyzed experimental replicates. Spectral library matching allowed the tentative identification of parent polycyclic aromatic hydrocarbons and sulfur heterocycles (such as dibenzothiophene and substituted naphthalenes), alongside a characteristic series of aliphatic carboxylic acids, including hexanoic, octanoic, nonanoic, n-decanoic, and dodecanoic acids. Highly polar transient intermediates, such as catechol or related diols, were not detected within the qualitative library-search profiles under these non-derivatized analytical conditions (Supplementary Table 3).

Genomic features and taxonomic confirmation via ANI

Whole-genome sequencing (WGS) of the E. kobei GOM24, H. litopenaei GOM23, and S. balearica GOM11 isolates revealed high-quality assemblies ranging from 4.46 to 4.98 Mb, with > 99% completeness and < 1% contamination. Average nucleotide identity (ANI) analyses confirmed species-level identification (Table 3) (Enterobacter kobei 99.25%, Halomonas litopenaei 98.56%, Stutzerimonas balearica 99.25%), all of which were well above the 95% species delimitation threshold. All the raw data are available within BioProject ID PRJNA1370045.

Table 3.

Genomic features and taxonomic identification of the key consortium isolates based on whole-genome sequencing

Identification	Integrity (%)	Contamination (%)	Assembly size (bp)	Contigs	N50 (bp)	L50	GC (%)	ANI valsue (%)	Reference strain
Enterobacter kobei GOM24	99.70	0.68	4,982,984	115	257,585	6	54.9	99.25	E. kobei MNCRE12
Halomonas litopenaei GOM23	100	0.47	4,875,208	69	199,526	7	64.2	98.56	H. litopenaei SYSU ZJ2214
Stutzerimonas balearica GOM11	100	0.13	4,459,115	48	248,694	7	64.7	99.25	S. balearica DSM 6083

Genes and pathways associated with hydrocarbon degradation

Genomic mining across the six isolated genomes revealed a robust machinery comprising 1615 genes dedicated to hydrocarbon degradation (obtained from HADEG), membrane transport, and physiological stress responses [obtained from OrthoDB (Zdobnov et al. 2020)]. Specifically, these included pathways for aliphatic alkane degradation, aromatic hydrocarbon degradation, central aromatic catabolism, membrane transport systems, and stress response mechanisms, including chaperones/heat shock, oxidative stress, and general/osmotic stress tolerance (Supplementary Table 4).

Functional annotation through Rapid Annotations using Subsystems Technology (RAST), HADEG, and BLAST Koala revealed a specialized metabolic repertoire for hydrocarbon degradation across the genomes. Analysis of the fatty acid, lipid, and isoprenoid subsystems by RAST revealed 93 nonredundant genes, with a core of 29 genes shared across all the consortium members.

With respect to the metabolism of aromatic compounds, we identified 72 nonredundant genes, with a core of 5 genes shared among all the consortium members. These core genes include fumarylacetoacetate hydrolase family proteins, succinyl-CoA:3-ketoacid-coenzyme A transferase subunits A and B (EC 2.8.3.5), 3-dehydroquinate dehydratase II (EC 4.2.1.10), and benzoate transport proteins. The greatest abundance of genes involved in aromatic degradation was identified in the Stutzerimonas (52.65%) and Pseudomonas (19.78%) genera (Fig. 4).

Fig. 4.

UpSet plot of aromatic compound metabolism across the consortium. The vertical bars represent the number of non-redundant genes in each intersection, while the horizontal bars indicate the total gene count for each strain. A functional core of 5 genes is shared by all members, while the large number of strain-specific genes highlights the metabolic specialization and functional redundancy that drives the degradation of aromatic fractions

Genes related to stress response and membrane transport were enriched in E. kobei GOM24 relative to the other sequenced genomes (22.90% and 25.85%, respectively), indicating unique potential for cellular homeostasis, motility (chemotaxis), and nutrient acquistion (Supplementary Fig. 2).

Gene analysis performed with BlastKOALA and HADEG allowed the identification of several putative genes associated with the degradation of alkanes (Fig. 5A). While the classical alkB gene was not detected, gene analysis revealed a diverse set of alternative alkane-hydroxylating systems. These include long-chain alkane monooxygenase (ladA), which is supported by the auxiliary oxidoreductases alcohol dehydrogenase (alkJ) and aldehyde dehydrogenase (alkH).

Fig. 5.

Genomic potential and terminal oxidation pathway for n-alkanes. A Heatmap of gene copy numbers for aliphatic degradation based on the HADEG database; the bottom bar categorizes enzymes by metabolic subsystem. B Proposed genomic reconstruction of aliphatic hydrocarbon metabolism. Distribution of key catabolic genes (ladA, alkJ, aldH) among the six strains of the consortium. The interaction between specialized monooxygenases and redundant dehydrogenases ensures efficient alkane processing and carbon flow into central metabolism. The layout illustrates the functional redundancy and metabolic synergy driving aliphatic hydrocarbon conversion.

Modified from Pandolfo et al. (2023)

The genomic configuration of the consortium includes a complete enzymatic set for the terminal oxidation of alkanes. Genes encoding long-chain alkane monooxygenase (ladA), alcohol dehydrogenase (alkJ), and aldehyde dehydrogenase (aldH) were identified, along with a complete repertoire of genes for the β-oxidation pathway (Fig. 5B). Additionally, the sequences of the ahpC and ahpF genes, which encode alkyl hydroperoxide reductases, were detected across the analyzed genomes.

Analysis of aromatic degradation-related sequences revealed genes encoding monooxygenases and dioxygenases associated with benzoate, fluorobenzoate, aminobenzoate, toluene, and xylene pathways. Notably, the enrichment of the ring-hydroxylating dioxygenase (RHD) system is prominent in the strains S. stutzeri (CBC1, CBC2), S. balearica (GOM11), and Pseudomonas sp. (CBC3). These genomes contain conserved gene clusters with high synteny to nah/nag-like clusters, including the key components nagAa (ferredoxin reductase) and nagAb (Rieske-type [2Fe−2S] ferredoxin) (Fig. 6A). The metabolic reconstruction (Fig. 6B) illustrates the enzymatic sequence starting from phenanthrene to the central metabolism. Downstream of the initial oxidation, the genes nagK (associated with lactone formation) and nagL (involved in aldehyde processing) were identified across the consortium, although their distribution varied among the strains. Finally, sequences belonging to the cat operon were detected in all genomes, providing the genetic basis for the conversion of catechol and protocatechuate into the β-ketoadipate pathway or the tricarboxylic acid (TCA) cycle.

Fig. 6.

Genomic potential and metabolic pathway for aromatic hydrocarbon degradation. A Heatmap of aromatic degradation genes. Distribution and copy number of genes identified via the HADEG database for the breakdown of diverse aromatic compounds (e.g., naphthalene, benzoate, toluene). The bottom color bar classifies the enzymes by substrate specificity and metabolic intermediate. B Proposed phenanthrene degradation pathway. Schematic representation of the upper and lower pathways leading to the TCA cycle. The diagram highlights the enzymatic steps—from initial dioxygenation (nagAa/Ab) to the Cat operon—and identifies the specific consortium members contributing to each catalytic stage. The results demonstrate high functional redundancy in the initial steps of PAH (Polycyclic Aromatic Hydrocarbon) degradation

Genes associated with the central β-ketoadipate pathway, including the catABC cluster, were detected in multiple genomes. Specifically, the enzyme catechol 1,2-dioxygenase (EC 1.13.11.1) was identified in four of the six strains: CBC1, CBC3, GOM11 and GOM23.

Catechol 1,2-dioxygenase activity

Catechol 1,2-dioxygenase activity was evaluated in cell-free protein extracts from the six consortium members after growth with benzoate as the inducer (Table 4). The highest initial specific activities were detected in S. balearica GOM11 (0.0472 ± 0.0031 U mg⁻¹ protein) and S. stutzeri CBC1 (0.0443 ± 0.0029 U mg⁻¹ protein). Intermediate activities were observed for H. litopenaei GOM23 (0.0165 ± 0.0078 U mg⁻¹ protein) and Pseudomonas sp. CBC3 (0.0161 ± 0.0028 U mg⁻¹ protein), whereas S. stutzeri CBC2 and E. kobei GOM24 exhibited comparatively low activities (0.0045 ± 0.0004 and 0.0030 ± 0.0006 U mg⁻¹ protein, respectively) Table 4.

Table 4.

Enzymatic activity of catechol 1,2-dioxygenase (EC 1.13.11.1) in cell-free protein extracts of bacterial consortium members and reference strains grown on benzoate (1 g·L⁻¹)

Bacterial strain	Substrate / inducer	Initial specific activity (0–30 min) (U·mg⁻1 protein)	Kinetic profile/behavior
Stutzerimonas balearica GOM11	Benzoate	0.0472 ± 0.0031	High-turnover initial response
Stutzerimonas stutzeri CBC1	Benzoate	0.0443 ± 0.0029	High-turnover initial response
Halomonas litopenaei GOM23	Benzoate	0.0165 ± 0.0078	Moderate linear activity
Pseudomonas sp. CBC3	Benzoate	0.0161 ± 0.0028	Moderate linear activity
Stutzerimonas stutzeri CBC2	Benzoate	0.0045 ± 0.0004	Basal/background level
Enterobacter kobei GOM24	Benzoate	0.0030 ± 0.0006	Basal/background level
Stutzerimonas stutzeri GOM2 (Control +)	Benzoate	0.0442 ± 0.0028	High sustained linear activity
Escherichia coli DH5 α (Control -)	Uninduced	0.0045 ± 0.0002	Basal/background level

Values represent mean ± standard deviation (n = 3). Enzyme activity was assayed at 30 °C in 50 mM phosphate buffer (pH 7.5) with 50 µM catechol by monitoring cis,cis-muconate formation at 260 nm (ε260​=16,800 M⁻¹ cm⁻¹). One unit (U) corresponds to 1 µmol of catechol converted per minute. Cells were grown in Burk’s mineral medium supplemented with 1 g·L⁻¹ benzoate

For GOM23 and CBC3, activity remained detectable during the 30–90 min interval, although at lower levels (0.0085 ± 0.0005 and 0.0071 ± 0.0004 U mg⁻¹ protein, respectively). In contrast, substrate depletion during the initial assay period precluded calculation of sustained activity for GOM11 and CBC1. Overall, these results demonstrate catechol 1,2-dioxygenase activity in multiple members of the consortium and provide experimental support for the genomic prediction of aromatic ring-cleavage capacity.

Despite the presence of these catabolic clusters, catechol and protocatechuate intermediates were not detected during the GC–MS analysis of the sand–seawater microcosms. This absence aligns with the high-turnover catechol 1,2-dioxygenase activity observed in vitro using cell-free extracts from liquid media, which demonstrates a high intrinsic capacity for immediate ring cleavage that prevents intermediate accumulation under controlled conditions.

Discussion

Rational design and biosafety

Generating synthetic bacterial consortia requires careful consideration of the functional capabilities of individual strains and their ability to coexist in a shared environment. In this study, most of the isolates belonged to Gammaproteobacteria (72.7%), a taxon frequently enriched in environments containing hydrocarbons and widely recognized for its metabolic versatility and rapid response to petroleum inputs (Kostka et al. 2011). However, the effectiveness of the synthetic consortium in degrading medium crude oil cannot be explained solely by the taxonomic composition but rather by the redundancy and complementarity of its metabolic capabilities.

Before the consortium assembly, the isolates were evaluated using a Galleria mellonella infection model, a widely used alternative for studying microbial pathogenicity. Although insects lack an adaptive immune response, their innate immune responses share functional similarities with those of vertebrates (Loh et al. 2013; Tsai et al. 2016). None of the strains caused significant mortality over six days, in contrast to the rapid lethality of Pseudomonas aeruginosa PAO1. The lack of pathogenicity in our isolates aligns with the genomic heterogeneity of environmental bacteria, where a large variable genome allows individual strains to tailor their functional repertoire to specific lifestyles and ecological niches (Loper et al. 2012).

Plackett–Burman (PB) analysis revealed high heterogeneity among experimental runs, indicating that specific strain combinations promote efficient removal, whereas others exhibit antagonistic effects (Hegazy et al. 2015; Rojas-Vargas et al. 2022). Several trials revealed no detectable degradation, whereas Run 2 achieved the highest oil removal efficiency (15.72%). These results highlight the importance of rational consortium design and that increasing species richness does not guarantee better degradation but rather implies greater complexity of the interactions within the system (Head et al. 2006). This was evident in Run 13 (all 11 isolates), where degradation decreased to 7.63%. This reduction suggests that taxonomic richness does not inherently correlate with functional performance, likely because of metabolic interference or ecological competition for limiting nutrients such as N and P, within the unoptimized community. Furthermore, the presence of antagonistic interactions such as the production of secondary metabolites that inhibit primary degraders underscores the necessity of a statistically guided assembly to filter and identify the most synergistic microbial modules.

Metabolic strategies for hydrocarbon degradation

The crude oil used in this study (Bacab Alpha Crude) is inherently recalcitrant because of its predominance of medium- and long-chain hydrocarbons (C8 to C40) as well as a diverse aromatic fraction (Supplementary Table 1). Compared with the light crude oils commonly reported in biodegradation studies, this type of medium crude is inherently more recalcitrant because of its high content of long-chain n-alkanes and polycyclic aromatic hydrocarbons (PAHs), which limit microbial accessibility and degradation, such as benzo[ghi]perylene and various methylated derivatives of naphthalene and phenanthrene (Kanaly and Harayama 2000).

The inclusion of strains affiliated with genera such as Stutzerimonas and Pseudomonas is consistent with their well-documented role as specialists in PAH degradation, such as phenanthrene degradation. In contrast, genera such as Enterobacter and Halomonas may contribute to the stability and resilience of consortia (Mnif et al. 2009). These taxa are known to participate in the degradation of medium-chain alkanes or confer tolerance to environmental stressors such as elevated salinity (McGenity et al. 2012; Corti Monzon et al. 2018). Such metabolic complementarity promotes a division of labor within the microbial community, a key factor for overcoming the challenges associated with the degradation of chemically complex matrices such as crude oil in sandy sediments.

Furthermore, the incorporation of taxa not traditionally considered primary hydrocarbon degraders may nevertheless play essential ecological roles, including the detoxification of intermediate metabolites, biosurfactant production, and the transformation of partially degraded compounds. These processes increase substrate availability and improve the efficiency of specialist degraders (Abbasian et al. 2015). Consequently, the functional integration of both specialist and supportive taxa may enhance the overall metabolic performance of the consortium.

When consortia are constructed to degrade complex substrates, it is essential to consider their coexistence with other strains as well as the rational partitioning of metabolic pathways among strains, as different organisms may use distinct degradation routes (Jawed et al. 2019; Lu et al. 2019; Liu et al. 2024).

Functional reconstruction through BlastKOALA revealed a highly redundant beta-oxidation pathway, providing genomic support for the observed capacity of the consortium to degrade aliphatic fractions. Furthermore, the presence of genes associated with flagellar assembly and chemotaxis suggests that several consortium members possess the potential for motility and chemotactic responses, traits that may facilitate access to poorly soluble hydrocarbons in heterogeneous matrices.

Despite the absence of the canonical alkB gene, the consortium demonstrated efficient degradation of medium-chain aliphatics, suggesting the involvement of an alternative enzymatic repertoire. The metabolic complementarity for this process is illustrated in Fig. 5B, where genomic reconstruction reveals a specialized division of labor during terminal oxidation. The initial monooxygenation step is potentially localized in H. litopenaei GOM23 and E. kobei GOM24, both of which harbor long-chain alkane monooxygenase (ladA).

In contrast, genes encoding the subsequent oxidative steps, such as alcohol dehydrogenase (alkJ) and aldehyde dehydrogenase (aldH), were identified more ubiquitously across the consortium’s genomes, including members of Stutzerimonas (CBC1, CBC2, and GOM11) and Pseudomonas (CBC3). This distribution is consistent with a potential division of metabolic functions, in which GOM23 and GOM24 may contribute to initial alkane oxidation, whereas other consortium members harbor enzymes involved in subsequent oxidation and β-oxidation steps. The differential distribution of hydrocarbon-degradation functions among consortium members is consistent with observations in naturally occurring hydrocarbon-degrading communities, in which distinct bacterial populations can contribute different metabolic capabilities to hydrocarbon transformation (Yun et al. 2024). In our consortium, the presence of ladA in GOM23 and GOM24, together with the broader distribution of alkJ and aldH among the remaining members, suggests a potential partitioning of functions during alkane oxidation.

While the degradation of aliphatics relies on the ladA system, the remarkably high efficiency of phenanthrene removal (94.7%) highlights a distinct, noncanonical metabolic strategy for aromatic compounds. The results of the genomic analysis suggest that this performance is mediated by nagAa (ferredoxin reductase) and nagAb (Rieske-type ferredoxin) gene clusters (Fig. 6B). Owing to the known substrate promiscuity of the naphthalene 1,2-dioxygenase system, the consortium likely bypasses traditional phenanthrene-specific pathways, channeling intermediates through the salicylate and gentisate routes toward the TCA cycle (Cao et al. 2021). This enzymatic flexibility is a key feature of the strains, which effectively repurpose naphthalene-degrading machinery to overcome the structural recalcitrance of three-ring polycyclic aromatic hydrocarbons (PAHs). Although the complete degradation of phenanthrene was not directly measured or confirmed in this study, the genomic potential and high depletion rates strongly indicate active cleavage and downstream processing of these aromatic intermediates.

The strong genomic presence of the catABC cluster and benzoate transport proteins across multiple consortium members suggests that aromatic degradation is channeled through these central pathways (Figs. 6A, B). This genomic architecture indicates that the consortium is equipped to process aromatic rings through central pathways at rates that prevent the buildup of detectable metabolic byproducts. Such rapid flux into the TCA cycle is an essential trait for a high-performance consortium, as it ensures that potentially autooxidative or toxic intermediates, such as catechol, are rapidly processed and converted, thereby maintaining the overall stability of the microbial community.

The genomic evidence for aromatic hydrocarbon degradation was further supported by the detection of catechol 1,2-dioxygenase activity in cell-free extracts of several consortium members. This enzyme catalyzes the ortho-cleavage of catechol, a central intermediate in the degradation of several aromatic compounds, connecting upstream aromatic transformations with the β-ketoadipate pathway. Consistent with the genomic reconstruction, catechol 1,2-dioxygenase activity was not uniformly distributed among the consortium members. The highest initial specific activities were observed in S. balearica GOM11 and S. stutzeri CBC1 (0.0472 ± 0.0031 and 0.0443 ± 0.0029 U mg⁻¹ protein, respectively), whereas H. litopenaei GOM23 and Pseudomonas sp. CBC3 showed intermediate activities. In contrast, S. stutzeri CBC2 and E. kobei GOM24 exhibited activities comparable to background levels. Although catechol was not detected in the untargeted GC–MS profiles, its absence is consistent with its role as a short-lived, highly reactive, and potentially cytotoxic transient intermediate. Because catechol undergoes rapid intracellular turnover by active ring-cleavage enzymes such as the catechol 1,2-dioxygenase activity demonstrated herein it is efficiently channeled into the β-ketoadipate pathway without accumulating in detectable amounts in the extracellular matrix. Nevertheless, the combination of genomic predictions and experimentally measured catechol 1,2-dioxygenase activity provides stronger evidence for the capacity of the consortium to process aromatic intermediates. These findings are consistent with functional differentiation among consortium members during aromatic hydrocarbon biotransformation, although direct metabolic exchange between strains would require further experimental validation.

Synergistic interactions and metabolic cooperation in the consortium

Metabolic cooperation and functional complementarity are increasingly recognized as key determinants of the performance of synthetic hydrocarbon-degrading consortia, because different members can specialize in distinct catabolic or supportive functions (Liu et al. 2024; Wang et al. 2026).

Our genomic data are consistent with such a division of metabolic functions in the degradation and metabolic processing of phenanthrene and other complex hydrocarbons. Stutzerimonas and Pseudomonas strains (CBC1, CBC2, CBC3, and GOM11) act as metabolic engines, initiating this process through the ring-hydroxylating dioxygenase (RHD) system. Their enzymatic flexibility allows the initial hydroxylation of aromatic rings into simpler intermediates such as salicylate.

At this critical juncture, a metabolic relay occurs where Enterobacter kobei GOM24 serves as a metabolic stabilizer. Although it possesses fewer primary catabolic genes, our analysis (Fig. 4) highlights that GOM24 contributes the greatest number of unique functional roles (14 roles) within the consortium. The high catechol 1,2-dioxygenase specific activities recorded for Stutzerimonas balearica GOM11 (0.0472 ± 0.0031 U⋅mg⁻¹ protein) and Stutzerimonas stutzeri CBC1 (0.0443 ± 0.0029 U⋅mg⁻¹ protein) align with previous findings for marine isolates from the Gulf of Mexico, such as Pseudomonas stutzeri GOM2, which possesses highly functional dioxygenase machinery adapted to marine conditions (Rodríguez-Salazar et al. 2020).

The qualitative GC–MS profiling consistently revealed residual parent aromatic hydrocarbons and aliphatic carboxylic acids across replicates, whereas transient intermediates such as catechol and cis,cis-muconic acid were not detected. Rather than indicating pathway inactivity, the absence of detectable catechol aligns with a high intracellular metabolic flux, as supported by the catechol 1,2-dioxygenase activity determined in vitro using cell-free liquid culture extracts. In active biodegradation systems, catechol and its immediate downstream metabolites represent critical yet short-lived central intermediates (Kamaraj et al. 2026) that undergo rapid enzymatic turnover to prevent intracellular toxicity, thus preventing their accumulation above analytical detection limits. The specific activity of catechol 1,2-dioxygenase measured in cell-free extracts confirms that once catechol is formed, it is rapidly cleaved via ortho-ring fission and channeled into the β-ketoadipate pathway. This high enzymatic turnover prevents the accumulation of toxic diols in the extracellular medium, maintaining their steady-state concentration below the detection limits of non-derivatized GC–MS screening.

Furthermore, GOM24 stands out for its extensive investment in membrane transport (ABC transporters) and stress response genes (ahpC/ahpF), which are critical for maintaining community stability and mitigating the oxidative stress generated during active hydrocarbon oxidation.

The intensive oxidation of crude oil fractions generates reactive oxygen species (ROS), which can compromise membrane integrity and enzymatic activity. Alkyl hydroperoxide reductases encoded by ahpC/ahpF constitute an important bacterial defense system against organic peroxides and oxidative stress (Si et al. 2023). The occurrence of these genes across all consortium members, and their increased copy number in Stutzerimonas and Pseudomonas, therefore, suggests an enhanced genomic capacity to cope with oxidative stress during hydrocarbon transformation. This genetic redundancy suggests an enhanced capacity for these 'catabolic engines' to mitigate the oxidative stress generated during the active cleavage of aromatic rings. Furthermore, the specialized functional profile of E. kobei GOM24, characterized by its unique repertoire of ABC transporters and its own antioxidant machinery, complements this resilient architecture. This collective investment in cellular homeostasis ensures the overall stability and metabolic longevity of the consortium under the high oxidative demands of hydrocarbon oxidation.

Additionally, the presence of Halomonas GOM23 likely facilitates this process by enhancing the ecological resilience of the consortium. As a halotolerant taxon, Halomonas is recognized for its ability to maintain metabolic activity under the osmotic fluctuations typical of marine sandy matrices (McGenity et al. 2012). This synergistic interaction where CBC strains initiate degradation, GOM24 stabilizes the intermediates, and GOM23 provides environmental robustness allows the consortium to function as a single metabolic unit, ensuring that the energy derived from hydrocarbon breakdown sustains the growth of all members.

The sand-seawater microcosms used in this study reproduce relevant components of a coastal sandy environment, but they represent controlled laboratory conditions. The crude-oil concentration used in the microcosms (0.1% w/v; 1 g L⁻¹ nominal concentration) was selected to allow reproducible assessment of hydrocarbon removal and consortium performance. Oil-contaminated beaches can exhibit highly variable hydrocarbon loads depending on spill magnitude, weathering, and time since contamination; for example, total petroleum hydrocarbon concentrations of approximately 1900–4500 mg kg⁻¹ have been reported in moderately to heavily contaminated beach sands following the Deepwater Horizon spill (Kostka et al. 2011).

Under field conditions, hydrocarbon biodegradation is influenced by multiple physicochemical and biological factors, including temperature, pH, salinity, oxygen and nutrient availability, organic matter, sediment sorption, co-occurring contaminants, oil weathering, and interactions with indigenous microbial communities. Therefore, the degradation efficiencies reported here under controlled sand–seawater microcosm conditions should be considered preliminary. Further validation across environmentally relevant conditions, using non-sterile coastal sediments and larger-scale mesocosms, will be required to assess the consortium’s robustness and potential for field application.

Conclusion

The statistically optimized synthetic consortium developed in this study demonstrated a robust capacity for the degradation of both aliphatic and aromatic hydrocarbons in a sand–seawater model system. Although total oil removal was constrained by the physicochemical complexity and thermodynamic limitations of the sandy matrix, the consortium exhibited highly specialized catabolic efficiency, achieving near-complete degradation of phenanthrene (94.7%).

This performance can be mechanistically attributed to the synergistic interaction of complementary metabolic pathways and versatile enzymatic systems. Specifically, the recruitment of nagAa and nagAb components from the naphthalene 1,2-dioxygenase complex enabled the consortium to overcome the structural recalcitrance of three-ring PAHs through a noncanonical metabolic route. Furthermore, the genomic insights provided here elucidate a sophisticated division of labor: while Stutzerimonas and Pseudomonas strains act as the primary metabolic engine, Enterobacter kobei GOM24 and Halomonas litopenaei GOM23 ensure community resilience and stability by maintaining cellular homeostasis and removing metabolic bottlenecks.

Overall, these findings highlight the functional robustness and biotechnological potential of rationally designed microbial consortia under controlled experimental conditions. While further field-scale validation is required to evaluate performance in natural dynamics, this study provides a preliminary framework for targeted bioremediation applications, demonstrating that metabolic complementarity is key to enhancing pollutant removal in coastal sand-seawater systems.

Supplementary Information

Below is the link to the electronic supplementary material.

Author contributions

ML-KS and LP-L contributed to the design and conceptualization of the research. ML-KS and JM-ZA performed the computational and bioinformatic analyses. ML-KS, AG, performed the GC–MS experiments and analyzed the data. LP-L coordinated the group, managed the resources, and guaranteed their availability to perform the experiments and analysis. All authors contributed to the writing of the manuscript and approved the final version.

Funding

Millán-López is a student in the the “Programa de Maestría y Doctorado en Ciencias Bioquímicas”, Universidad Nacional Autónoma de México (UNAM), and received fellowship 957581 from the Consejo Nacional de Humanidades, Ciencias y Tecnologías (CONAHCYT). Research was funded by the PAPIIT DGAPA UNAM program (IG200223), the CONAHCYT National Laboratory of Oceanography (LANCO) and SECIHTI MADTEC-2025-M-60.

Data availability

The data that support the findings of this study are openly available in the NCBI Sequence Read Archive (SRA) under BioProject accession number ID PRJNA1370045.

Declarations

Conflict of interest

The authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.