Biodegradation of DDT using multi‐species mixtures: From genome‐mining prediction to practical assessment

Abstract

DDT (dichlorodiphenyltrichloroethane) is a commonly used insecticide that is recalcitrant and highly stable in the environment. Currently, DDT residue contamination, especially in agricultural soil, is still a concern in many countries, threatening human health and the environment. Among the approaches to resolve such an issue, novel biodegradation‐based methods are now preferred to physicochemical methods, due to the sustainability and the effectiveness of the former. In this study, we explored the possibility of building mixed microbial cultures that can offer improved DDT‐degrading efficiencies and be more environmentally transilient, based on genome annotation using the KEGG database and prediction of interactions between single strains using the obtained metabolic maps. We then proposed 10 potential DDT‐degrading mixed cultures of different strain combinations and evaluated their DDT degradation performances in liquid, semi‐solid and solid media. The results demonstrated the superiority of the mixtures over the single strains in terms of degrading DDT, particularly in a semi‐solid medium, with up to 40–50% more efficiency. Not only did the mixed cultures degrade DDT more efficiently, but they also adapted to broader spectra of environmental conditions. The three best DDT‐degrading and transilient mixtures were selected, and it turned out that their component strains seemed to have more metabolic interactions than those in the other mixtures. Thus, our study demonstrates the effectiveness of exploiting genome‐mining techniques and the use of constructed mixed cultures in improving biodegradation.

Genome‐mining can be the base for rational construction of mixed cultures containing microbial strains that metabolically complement each other and thus offering improved biodegradation of pollutants, such as DDT. The constructed mixed cultures performed better than the single member strains in terms of both DDT degradation and environmental transilience.

INTRODUCTION

DDT [1,1,1‐trichloro‐2,2‐bis(p‐chlorophenyl) ethane], one of the most used insecticides in the past few decades, is a serious environmental pollutant. Due to its low solubility in water and the ability to accumulate through the food chain, DDT is regarded as a persistent organic pollutant. The half‐life of DDT is up to 30 years (Mansouri et al., 2017). The toxicity of DDT primarily rests on its ability to prevent Na⁺ channels on nerve cell membranes from closing after the depolarization phase of action potential (Silver et al., 2014). As a consequence of this sustained opening, the affected nerve cells become overexcited, which can lead to convulsions and death (due to convulsions of muscles related to respiratory activity) (Dong, 2007). Moreover, DDT also causes unwanted side effects in many creatures, including humans and other animal groups such as fishes, amphibians, birds and mammals (WHO, 1979). Besides directly affecting human health, DDT also has major impacts on the ecosystem. With its lipophilic properties, as well as low solubility in water, DDT is stored for a long time in fatty tissues of animal bodies. This leads to the accumulation of DDT in living bodies and the harmful effects of DDT on humans and organisms at higher trophic levels are even more severe due to the biomagnification effect (Deribe et al., 2013; Di Landa et al., 2009).

Although DDT has been banned for a long time, due to its recalcitrant properties, up to now, DDT residue is still found in soil in many different areas around the world. In a recent study analysing 317 soil samples from 11 European countries, p,p′‐DDE, o‐p′‐DDD and p‐p′‐DDT were detected in 72, 23 and 23 samples, respectively, with the respective concentrations up to là 310, 40 and 10 μg kg⁻¹ (Silva et al., 2019). The situation is more serious in many Asian countries such as India, China, Vietnam, etc. (Sharma et al., 2019; Silva et al., 2019; Tzanetou & Karasali, 2022). A review of about 120 studies on more than 2000 soil samples in China revealed that DDT could be detected with concentrations up to 488 μg kg⁻¹ (Yu et al., 2020). In India, residual DDT was found in many sites at concernable levels, for example, up to 5.794 mg L⁻¹ in the Bhopal River (Mamta Rao & Wani, 2015). In Vietnam, DDT was found in higher concentrations than in other countries in the region in both environmental and biological samples. According to survey results and statistics of Vietnam Environment Administration on chemical residues as of June 2015, there were 1562 residues due to plant protection chemicals, such as DDT, of which concentrations are still at threatening levels in many locations (https://www.gef.monre.gov.vn/wp‐content/uploads/2016/01/POP_bao‐caohien‐trang_final_print‐1.pdf). According to Tham et al. (2019), the DDT concentrations in surface sediment samples collected from estuaries on Vietnam's central coastline were around 0.462 to 26.7 ng g⁻¹ dry wt (Tham et al., 2019). Notably, a survey carried out in five Vietnamese provinces found that DDT was present in more than 90% of soil samples collected (Ding et al., 2024). Such a situation urges us to seek sustainable solutions to degrade DDT in contaminated sites.

Physical and chemical methods for degrading DDT exist and some have been in use in practice, including burial and isolation, catalysed incineration, using UV rays, etc. (Chattopadhyay & Chattopadhyay, 2015; Xia et al., 2012). However, these methods exhibit two major disadvantages: namely, causing secondary pollution and high costs. Therefore, addressing DDT pollution by physical and chemical means is now undesirable, and instead, people are turning more towards bioremediation. Biodegradation helps to thoroughly decompose pollutants and make use of microbial sources that are available in nature, so it is environmentally friendly and offers complete remediation as well as a reasonable cost (Cao et al., 2024; Chen et al., 2023; Lin et al., 2022).

Bioremediation of DDT is solely based on the use of microorganisms (Mansouri et al., 2017). There are many mechanisms by which microorganisms remove organic substances containing chlorine, such as: (i) bioaccumulation: the ability to retain or precipitate these substances by organisms, or (ii) biosorption: the ability to adsorb these substances onto biological structures, or (iii) biotransformation: the ability to convert these substances into other, less toxic products (Camacho‐Pérez et al., 2012). In particular, biological transformation is the main path for the biodegradation of the mentioned substances. The biodegradation of halogen‐containing substances is generally done usually through aerobic pathways (oxidation reactions) or anaerobic pathways (dehalogenation reactions) (Mansouri et al., 2017). According to Mansouri et al. (2017), DDT metabolism is catalysed by enzymes such as DDT 2,3‐dioxygenase and DDT dehydrochlorinase produced by different bacteria. Currently, there are a variety of microbial strains that are discovered to be capable of degrading DDT, among which some display notable degrading performance. For example, a Pseudomonas sp. strain was reported capable of degrading more than 60% of DDT within 5 days (Kamanavalli & Ninnekar, 2004). In another study, strain Stenotrophomonas sp. DDT‐1 was reported to be able to degrade totally >90% of DDT after 82 days (Pan et al., 2016). Such strains offer the possibility of using microbial preparations for bioremediation of DDT in contaminated sites.

The use of preparations based on single strains of microorganisms has been studied to a certain extent. However, the effectiveness of these preparations in practice is not yet researched thoroughly. In addition, as DDT is a substance with a complex structure, a single strain will not likely be able to effectively degrade this substance. Therefore, the approach of using DDT‐degrading mixed cultures has also gained research attention because such cultures could degrade DDT more effectively and completely than single ones (Mansouri et al., 2017). Mixtures are also better able to adapt to adverse environments (Aparicio et al., 2018), due to their high diversity of species, which constitutes their high flexibility in their habitat, allowing them to remain functional in many different conditions. However, thus far most studies have been carried out on natural mixed cultures. For example, it was reported that the existing microbial community available in soil, after enrichment, could be better able to degrade DDT than single strains (Wang et al., 2017). Recently, there has been one study on using artificial combinations consisting of eight strains with good DDT‐degrading ability isolated from industrial sediment containing DDT in Mexico. The results showed that there was no difference in DDT‐degrading efficiency between the mix cultures and single strains, but new metabolites, which were not found in single strains' sample, were detected in the mixed cultures (Garcia Lara et al., 2021).

Given the current situation, it is apparent that there have not been many studies on the degradation of DDT by artificial microbial combinations or complexes, especially those composed of bacterial strains, while this has a potential to be an effective approach because nowadays strain combinations can be rationally designed by genomic and metabolic analysis tools. Genomic analysis may help evaluate and predict interactions between different microorganisms when participating in the process of decomposing a pollutant, thereby helping develop an effective approach (Bhatt, 2019). With the recent development of new generation sequencing methods, more and more microbial studies are using the whole‐genome‐sequencing (WGS)‐based approach, through genome mining analysis (Carrillo & Blais, 2021), which mainly focuses on predicting biologically active compounds and synthesis pathways of these substances (Ziemert et al., 2016). Some databases such as CAZy (Carbohydrate‐Active enZYmes), MetaCyc, Reactome, and KEGG are often used to discover some degrading enzymes, metabolic pathways, etc. (Dilokpimol et al., 2016; Ma et al., 2022). Among them, the KEGG database (Kyoto Encyclopedia of Genes and Genomes) is the oldest and considerably most complete database, with tools to represent the network of metabolic pathways that is intuitive and comprehensible. Most importantly, only the KEGG database contains data on degrading pathways of xenobiotics, including DDT (Altman et al., 2013). Given these considerations, in this study, we used the KEGG database for analysing genome data to establish metabolic maps of selected bacterial strains capable of degrading DDT and predict their interactions to construct potential mixtures that can effectively degrade DDT and adapt better to varied environmental conditions. Such ‘in silico’ work was also validated by benchwork experiments. This approach has not been addressed in previous studies and is one of the most important features of this study.

EXPERIMENTAL PROCEDURES

Material

In this study, we used 10 DDT‐degrading microbial strains that were previously isolated, selected and identified by our research group. These strains were selected among more than 40 DDT‐degrading strains we isolated from different soil samples in Vietnam. They include Pseudomonas sp. T006, Pseudomonas putida T087, Ralstonia sp. TN030, Pseudomonas sp. PAM64, Pseudomonas sp. PAM67, Pseudomonas sp. Y077, Stenotrophomonas maltophila Y042, Stenotrophomonas maltophila Y050, Streptomyces sp. DDT21 and Streptomyces sp. DDT23. They were deposited in the public culture collection of Phenikaa University under the accession numbers from PU50001 to PU500010, respectively.

The following media were used for culturing microorganisms in this study: (i) mineral salts medium (MSM) (per litre): 0.4 g MgSO₄·7H₂O, 0.002 g FeSO₄·7H₂O, 0.2 g K₂HPO₄, 0.2 g (NH₄)₂SO₄, 0.08 g CaSO₄, 1000 mL H₂O, pH 7.0–7.2; (ii) trace elements (TE) solution (per litre): 1 g FeSO₄·7H₂O, 0.07 g ZnCl₂, 0.1 g MnCl₂·4H₂O, 0.006 g H₃BO₃, 0.13 g CaCl₂·6H₂O, 0.002 g CuCl₂·6H₂O, 0.024 g NiCl₂·6H₂O, 0.036 g NaMo₄·2H₂O, 0.238 g CoCl₂·6H₂O, pH 7.0–7.2; (iii) Luria‐Bertani (LB) medium (per litre): 10 g peptone, 5 g yeast extract, 10 g NaCl, 1000 mL H₂O, pH 7.0–7.2; (iv) yeast – starch (YS) medium (per litre): 2 g yeast extract, 10 g starch, 1000 mL H₂O. When used in semi‐solid or solid forms, these media were supplemented with agar (or agarose) at the rate of 2.5 or 16 g L⁻¹, respectively.

All chemicals used were purchased from reliable suppliers (such as Xilong (China), Biobase (Canada), Sigma‐Merck (USA), etc.). 4,4′‐DDT (99.8% purity) was supplied by Dr. Ehrenstorfer (Bavaria, Germany).

Methods

DNA extraction and whole genome sequencing (WGS)

DNA samples used for whole genome sequencing were extracted using Qiagen DNA extraction kits: genomic DNAs of actinomycetous isolates were extracted using DNeasy® PowerSoil Pro Kit (Qiagen, USA), genomic DNAs of bacterial isolates were extracted using the DNeasy® Blood & Tissue Kit (Qiagen, USA). The DNA extraction procedure was performed as in the manufacturers' manual. DNA extract solutions meeting the quality requirements were sent for sequencing by LOBI (Vietnam). The whole genome sequencing platform used by the sequencing service provider was BGI sequencing DNBseq 150PE or Illumina HiSeqXten sequencing 150PE. Each DNA fragment was sequenced in both directions (paired‐end).

WGS analysis and genome annotation

The raw data was further processed as follows: The raw data was purified using the fastq tool, in which bases with quality scores less than 30 and read segments shorter than 70 bp in length were removed. To assess the possibility of data being contaminated, the raw data were initially identified using the Kraken2 with miniKraken2 database. The purified data were used for de novo genome assembly using SPAdes with the k‐mer with default parameters. The quality of de novo assembled genome was evaluated based on the following criteria using QUAST software: total genome size, largest contig length and N50 index.

Genes of the de novo assembled genomes were predicted and functionally annotated using prokka (Seemann, 2014). Protein sequences after being annotated by prokka were annotated according to the KEGG database using the KEGG Automatic Annotation Server (KASS) and KEGG Mapper—Reconstruct Pathway. Annotated proteins were sorted into COG (Cluster of Orthologous Groups) groups using EggNOG (evolutionary genealogy of genes: Non‐supervised Orthologous Groups) application. When annotating genome according to the KEGG database, we focused mainly on two groups of pathways: ‘Global and overview maps’ pathways (used to annotate basic pathways) and ‘Xenobiotics biodegradation and metabolism’ (used to annotate possible pathways involved in DDT degradation). Therefore, these two pathway groups were of primary use in the analysis of the metabolic capacity and DDT metabolism potential of each isolate. The number of enzymes in these two pathway groups of single strains was enumerated and compared.

Identification results for the studied strains based on de novo assembled genome were performed by uploading the respective assembled genome data to the Type Strain (Genome) Server system (TYGS), a platform belonging to DSMZ (Deutsche Sammlung von Mikroorganismen und Zellkulturen).

Constructing strain combinations based on metabolic interaction prediction

Based on the results of genomic annotation with KEGG database, KO (KEGG Orthology) identifiers of proteins of each strain of interest were collected and assembled. The assembled identifiers data file was then uploaded to the KEGG mapper – reconstruction page to acquire the overall metabolic pathways of the strain. Metabolic maps of different strains were observed and compared with each other to identify possible complementary metabolic interactions between the strains. For a strain mixture, a general metabolic map could be generated by combining the maps of the member strains, while the specific pathways of each strain were highlighted after being filtered using Microsoft Excel. In addition to the metabolic maps, information about the general physiological and biochemical characteristics of the strains was also considered when constructing the potentially advantageous strain combinations.

Preparing mixtures containing the constructed strain combinations for practical tests

To empirically evaluate the actual advantages of the strain combinations constructed based on metabolic interaction predictions, the mixture suspensions containing those combinations were prepared.

Preparation of the biomass of each bacterial strain: A fresh liquid culture of the strain was prepared by shaking incubation in LB broth at 200 rpm, at 30°C and usually for 12 h. Such a bacterial culture was centrifuged at ca. 4000 × g for 15 min. to collect biomass. The cell pellet was then washed in saline solution twice by vortexing and then centrifuged at 6000 rpm for 15 min. The cell pellet after washing will be mixed with a suitable volume of MSM to obtain a cell suspension having a density of approximately 1 × 10⁹ CFU mL⁻¹.

Preparation of the biomass of each actinomycetous strain: This was prepared in the form of solely spores. The strain was cultured on YS agar plates until it matured and produced spores, usually after 5–7 days. The spores were subsequently collected by pouring liquid MSM medium onto the agar surface and scraped off by using a sterile microstreaker. The density (in CFU mL⁻¹) of such a suspension was then determined by serial dilution and the spread plate method. When used in experiments, such a stock suspension will be diluted with an appropriate volume of MSM solution to obtain a suspension having a density of approximately 1 × 10⁹ CFU mL⁻¹.

Preparation of the mixtures: A mixture suspension was created by mixing equal volumes of suspensions of single strains having the same biomass density (10⁹ CFU mL⁻¹) according to the determined formula of the respective strain combination it contains. Thus, in such a mixture, the total microbial density was approximately 10⁹ CFU mL⁻¹. The mixing ratio of a mixture could be 1:1:1 (in a 3‐strain mixture), 1:1:1:1 (in a 4‐strain mixture) or 1:1:1:1:1 (in a 5‐strain mixture), depending on the mixture formula.

Determination of DDT degradation capabilities of the mixtures

Different mixtures and single strains were tested for their DDT‐degrading efficiencies in liquid, semi‐solid (0.25% agar) and solid (1.6% agar) MSM media containing 30 mg L⁻¹ DDT. For inoculation, at the beginning 100 μL of each microbial suspension (prepared as described above) was dropped into 1000 μL of such MSM liquid medium or onto the agar surface of 1000 μL of such MSM semi‐solid medium or MSM solid medium in a test tube. The test tube, if containing semi‐solid or solid medium, was gently spinned by hand to spread the solution evenly on the agar surface. For each experimental case, three tubes (triplicates) were prepared. The cultures were incubated at pH 7.0 and at 30°C (and shaken at 200 rpm only for the liquid cultures). The DDT degradation efficiency of each culture was inferred by measuring the residual DDT concentration in the medium after 7 and 14 days using gas chromatography with flame ionization detector (GC‐FID) (see the description of chemical analyses below).

Prior to the GC‐FID analysis at each checkpoint, for DDT extraction in a liquid medium test, 1000 μL of chloroform is added to each test tube. This tube will be vortexed for about 20 s to extract DDT. Similarly, for DDT extraction in a semi‐solid or solid medium test, 1000 μL of chloroform is added to each test tube and the solidified agar mass is cut with a sharp stick into four pieces of equal volume. The test tube was then left to stand for 16 hours. Approximately 300 μL of each extract solution (chloroform phase) containing DDT was sampled for GC‐FID analysis.

Testing the effects of different external factors on DDT‐degrading performances of the mixtures

Effect of temperature

Ten per cent of each mixture of interest was added to a test tube containing liquid MSM containing 30 mg L⁻¹ DDT to obtain a total volume of 2 mL. The culture was incubated in a shaking incubator (at 150 rpm) at different temperatures of 10, 15, 25, 30 and 40°C.

Effect of substrate concentration

Ten per cent of each mixture of interest was added to a test tube containing liquid MSM containing a tested DDT concentration (0, 2, 5, 10, 20 or 30 mg L⁻¹). The culture was incubated in a shaking incubator (at 150 rpm) at 30°C.

Effect of additional carbon sources

Ten per cent of each mixture of interest was added to a test tube containing liquid MSM containing 30 mg L⁻¹ DDT to obtain a total volume of 2 mL. Simultaneously, a tested carbon source was added to the culture at a concentration that ensures that the added quantity of moles of C is equal to a fixed level: that is, 1 g L⁻¹ for fructose or galactose (M = 180.16), 0.95 g L⁻¹ for maltose, lactose or sucrose (M = 242.3), and 0.52 g L⁻¹ for phenol (M = 94.11). The culture was incubated in a shaking incubator (at 150 rpm) at 30°C.

Effect of additional nitrogen sources

Ten per cent of each mixture of interest was added to a test tube containing liquid MSM containing 30 mg L⁻¹ DDT to obtain a total volume of 2 mL. Simultaneously, a tested nitrogen source was added to the culture at a concentration that ensures that the added quantity of moles of N is equal to a fixed level: that is, 1 g L⁻¹ for (NH₄)₂SO₄ (M = 132.14), 1.23 g L⁻¹ for ammonium citrate (M = 243.22), 1.53 g L⁻¹ for KNO₃ (M = 101.1), 1.51 g L⁻¹ for peptone (total N ≥ 14%) and 1.14 g L⁻¹ for glycine (M = 75.07). Peptone represents a nitrogen‐rich organic source, and glycine represents an easy‐to‐use amino acid, while the others represent sources containing ${\mathrm{NH}}_4^{+}$ or ${\mathrm{NO}}_3^{-}$. The culture was incubated in a shaking incubator (at 150 rpm) at 30°C.

Effect of trace element proportion

Ten per cent of each mixture of interest was added to a test tube containing liquid MSM containing 30 mg L⁻¹ DDT to obtain a total volume of 2 mL. Trace element solution was also added to the culture at a tested ratio (0.1%, 0.5%, 1%, 2% and 5%). The culture was incubated in a shaking incubator (at 150 rpm) at 30°C.

Effect of seed culture amount

Each mixture of interest was added at a tested ratio (5%, 10%, 15% or 20% v/v) to a test tube containing liquid MSM containing 30 mg L⁻¹ DDT to obtain a total volume of 2 mL. The culture was incubated in a shaking incubator (at 150 rpm) at 30°C.

Measurement of dechlorination rates of the cultures: Each culture in the above‐mentioned experiments was incubated for 7 days and the concentrations of free chloride ions in the medium at day 0 and day 7 were measured (as described below in the next section) to calculate the dechloration rate of the culture. Based on that, the effects of different factors or different levels of a factor could be evaluated.

Effect of moisture

Ten per cent of each mixture of interest was added to a test tube containing liquid MSM and quartz sand (with a total volume of 1000 μL), mixed in a ratio such that the medium had a tested moisture of 20% (10% MSM, 80% quartz sand), 30% (10% MSM, 10% H₂O, 70% quartz sand), 40% (10% MSM, 20% H₂O, 50% quartz sand) or 50% (10% MSM, 30% H₂O, 40% quartz sand). DDT was added to each tube to the final DDT concentration of 20 mg L⁻¹. The culture was incubated at 150 rpm, 30°C, for 7 days and its DDT degradation efficiency was inferred by measuring the residual DDT concentrations in the medium at day 0 and day 7 using GC‐FID analysis (as described below). For DDT extraction, at each checkpoint, 1000 μL of chloroform is added to each test tube. Then, it will be vortexed for about 20 seconds to extract DDT. Approximately 300 μL of each extract solution (chloroform phase) containing DDT was sampled for gas chromatography analysis.

Chemical analyses

Determination of free chloride ions for dechlorination activity measurements

This method is built on the method of Bergmann and Sanik (1957). First, 200 μL of a sample of interest was supplemented with 20 μL of a solution containing 0.25 M ferric ammonium sulphate (Fe(NH₄)(SO₄)₂·12H₂O) in 9 M nitric acid and then with 20 μL of a solution containing mercuric thiocyanate saturated in ethanol. The resulting solution was mixed thoroughly to observe the colour change (gradually turning yellow if free Cl‐ is present). After 10 min., the absorbance of the Cl‐ complex was measured at 460 nm using a spectrophotometer (Labomed, USA). The blank sample and the calibration samples containing different concentrations of NaCl (5, 10, 30 and 50 mg L⁻¹) were processed similarly. The actual concentration of free chloride ions in the sample is determined based on the calibration curve established.

Determination of DDT concentration in an extract solution using GC‐FID

Each extract solution (with chloroform as solvent) was analysed by gas chromatography using an Agilent 7890 system, with HP‐5 glass capillary column (ID: 0.32 mm, length: 30 m; film thickness: 0.25 μm) and N₂ as the carrier gas at the speed of 1.2 mL min⁻¹. Injector mode was: split of 1:10. Injection volume was 1 μL. Heat program was as follows: injector temperature: 250°C; oven temperature (oven): starting at 170°C (hold for 5 min), then increased at the speed of 35°C min⁻¹, to 275°C (hold at 5 min); detector temperature (flame ionization detector): 275°C. The peak of DDT appeared at about 10 min after the initiation of the heat program. The concentration of DDT in each sample was determined based on a calibration curve established by measuring standard samples during the experiments.

Data processing and analysis

The true DDT‐degrading performance of a mixture is evaluated through the percentage difference between the actual DDT‐degrading efficiency (calculated from actual measurement results) and the average DDT‐degrading efficiency of the single strains in the mixture. Theoretically, the latter would be the degradation efficiency of the mixture if no (metabolic) interactions between the strains happened.

Unless otherwise stated, the experiments were done in triplicates and the reported data were average values and standard deviations calculated by Microsoft Excel using standard algorithms. The significance of data differences between different experimental cases was evaluated using the basic t‐test.

RESULTS AND DISCUSSION

DNA extraction and whole‐genome sequencing (WGS)

Whole‐genome DNA extracts of the strains were obtained successfully (Figure S1) as they all met the requirements of samples for sequencing. They were submitted for WGS and the quality control results showed that they were of good quality, with an average quality score above 30 for all sequencing segments (Chattopadhyay & Chattopadhyay, 2015). This data was then cleaned, and the post‐purification sequences retained more than 90% of the original data. Furthermore, preliminary Kraken classification results showed that all the samples were of high purity. Their BUSCO's genome integrity scores reached over 99.0%, which indicate that all the genomes have high integrity.

The WGS‐based identification results by Kraken (Table S1) were consistent with our previous identification results and thus the taxonomic classifications of the strains were confirmed.

Genome annotation and KEGG metabolic maps

From the genome data, the protein data of each strain could be predicted by using Prokka and KASS for gene function annotation, and these results revealed more about the metabolic potentials of the strains and their relationships (Figure 1). It is apparent that the protein profiles of the two Streptomyces strains (DDT21 and DDT23) were closer to each other than those of other strains. Similar observations can be made for the two Stenotrophomonas maltophila strains (Y042 and Y050) and for the five Pseudomonas strains (PAM64, PAM67, T006, T087 and Y077). As for Ralstonia metallidurans TN030, it is interesting that its protein profile is closer to those of the Pseudomonas strains.

FIGURE 1.

Comparison of COG and eggNOG annotations of single strains. C: Energy production and conversion; D: Cell cycle control, cell division, chromosome partitioning; E: Amino acid transport and metabolism; F: Nucleotide transport and metabolism; G: Carbohydrate transport and metabolism; H: Carbohydrate transport and metabolism; I: Lipid transport and metabolism; J: Translation, ribosomal structure and biogenesis; K: Transcription; L: Replication, recombination and repair; M: Cell wall, membrane, envelope, biogenesis; N: Cell motility; O: Post‐translational modification, protein turnover and chaperones; P: Inorganic ion transport and metabolism; Q: Secondary metabolites biosynthesis, transport, and catabolism; S: Function unknown; T: Function unknown; U: Intracellular trafficking, secretion and vesicular transport; V: Intracellular trafficking, secretion and vesicular transport; Unclassified.

By using ‘Global and overview maps’ and ‘Xenobiotics biodegradation and metabolism’ pathways of KEGG database for gene annotation, we can clearly see the differences between the strains in terms of the abundance of the enzymes in global metabolic pathways as well as that of the specific enzymes involved in DDT metabolism (Table 1 and Figure S2, Table S2). Furthermore, the pathways of the strains can be seen and compared illustratively through their respective metabolic maps (Figure 2 and Figures S3–S9).

TABLE 1.

Quantities of enzymes in major metabolic pathways of the strains as annotated from their genomes according to the KEGG database.

ID	Pathway	DDT21	DDT23	PAM64	PAM67	T006	T087	TN030	Y042	Y050	Y077
Global and overview maps
01100	Metabolic pathways	848	711	793	770	792	758	769	573	583	792
01110	Biosynthesis of secondary metabolites	341	329	285	291	279	287	298	247	253	291
01120	Microbial metabolism in diverse environments	257	206	245	216	244	216	249	136	137	237
01200	Carbon metabolism	101	81	96	88	89	90	102	80	81	91
01210	2‐Oxocarboxylic acid metabolism	24	26	26	24	23	23	24	20	18	23
01212	Fatty acid metabolism	22	18	28	21	26	21	23	21	22	27
01230	Biosynthesis of amino acids	106	104	106	104	101	104	103	95	93	102
01232	Nucleotide metabolism	46	46	40	37	40	36	37	35	40	39
01250	Biosynthesis of nucleotide sugars	36	31	41	41	39	36	34	30	32	35
01240	Biosynthesis of cofactors	144	133	141	140	141	141	131	111	105	142
01220	Degradation of aromatic compounds	26	12	28	20	28	22	23	3	2	20
Xenobiotics biodegradation and metabolism
00362	Benzoate degradation	17	9	27	22	27	23	23	5	5	24
00627	Aminobenzoate degradation	13	8	9	4	7	3	13	1	1	7
00364	Fluorobenzoate degradation	4	1	6	6	7	6	7	1	1	6
00625	Chloroalkane and chloroalkene degradation	8	4	7	5	5	6	5	3	3	6
00361	Chlorocyclohexane and chlorobenzene degradation	5	1	3	2	3	2	6	1	2	3
00623	Toluene degradation	2	2	2	2	3	2	3	1	1	3
00622	Xylene degradation	3	1	8	5	9	6	8	0	0	5
00633	Nitrotoluene degradation	1	3	1	2	1	2	4	1	1	3
00642	Ethylbenzene degradation	1	1	1	1	1	1	1	1	1	1
00643	Styrene degradation	5	4	8	8	8	7	4	4	3	8
00791	Atrazine degradation	7	4	6	3	6	4	3	0	0	6
00930	Caprolactam degradation	5	5	5	2	6	2	4	1	1	6
00621	Dioxin degradation	3	0	4	1	4	1	4	0	0	2
00626	Naphthalene degradation	4	3	3	2	2	3	2	2	2	3
00624	Polycyclic aromatic hydrocarbon degradation	2	0	2	2	2	2	2	0	0	2
00984	Steroid degradation	7	4	7	1	1	1	3	0	1	2
00980	Metabolism of xenobiotics by cytochrome P450	6	2	4	3	3	3	3	3	3	3
00982	Drug metabolism − cytochrome P450	6	3	5	4	4	4	3	3	3	4
00983	Drug metabolism – other enzymes	13	14	9	10	9	10	9	8	10	10

FIGURE 2.

Theoretical metabolic maps of some strains of interest. (A) Ralstonia sp. TN030; (B) Streptomyces sp. DDT21; (C) Pseudomonas sp. PAM67; (D) Stenotrophomonas maltophila Y042. Lines in different colours indicate different pathways: orange: Chloroalkane and chloroalkene degradation; red: Chlorocyclohexane and chlorobenzene degradation; green: Dioxin degradation; black: General pathways.

Among the strains studied, the actinomycetous strain DDT21 has the most diverse enzymes in considered metabolic pathways, as demonstrated by the high number of enzymes in global pathway groups such as ‘Metabolic pathways’, ‘Biosynthesis of secondary metabolites’, ‘Microbial metabolism in diverse environments’ and some groups of pathways for degrading xenobiotics (Table 1). Strain TN030 is also noteworthy because it has the highest number of enzymes belonging to two pathway groups: ‘Chlorocyclohexane and chlorobenzene degradation’ and ‘Dioxin degradation’ (Table 1). These two groups of pathways are involved in the degradation of chlorinated aromatic compounds, very similar in structure to DDT. Therefore, TN030 has the great potential to effectively degrade DDT.

Regarding the Pseudomonas strains (PAM64, PAM67, T006, T087 and Y077), although in general their global metabolic pathways seem to be less diverse than those of DDT21, their xenobiotics degradation pathways are more diverse than those of DDT21 and can be considered similar to those of TN030 (Table 1, Figure 2A,C and Figures S4–S6, S9). We can also see that those Pseudomonas strains are very similar to each other in metabolic pathways (Table 1, Figure 2C and Figures S4–S6, S9), with almost the same numbers of pathways related to the degradation of xenobiotics, especially those for benzoate compounds, chloroalkane, chloroalkene, chlorocyclohexane, chlorobenzene and toluene, which are related to DDT degradation. Notably, PAM64 and T006 appear to harbour more dioxin degradation pathways (Table 1), as well as more pathways of xylene degradation. These observations suggest that those two strains can probably degrade DDT more flexibly and effectively than the other Pseudomonas strains.

Among the strains, the two Stenotrophomonas strains (Y042 and Y050) appear to harbour the least number of global metabolic pathways as well as xenobiotics degradation pathways (Table 1, Figure 2D and Figure S8). However, they possess some genes encoding some enzymes that those Pseudomonas strains do not have, such as carboxymethylenebutenolidase and haloalkane dehalogenase (only in Y050) (Table S2). Probably, these enzymes somehow make the two Stenotrophomonas strains unique in terms of dealing with DDT in practice.

Constructing strain combinations for mixed cultures based on prediction of metabolic interactions between the single strains

As mentioned above, among the strains, DDT21 and TN030 appear to have more diverse metabolic capabilities, including those of general pathways that may help them deal better with diverse substrates and environmental conditions and those of specific pathways related to DDT metabolism. Therefore, these two strains were chosen as the core strains, around which we constructed the strain combinations that can efficiently degrade DDT and adapt well to diverse practical environments.

Of the five Pseudomonas spp. strains, Pseudomonas sp. T087 strain is suspected to be a new species and its metabolic map appears to be the simplest. Therefore, we chose the other four strains (PAM64, PAM67, T006 and Y077) for strain combination construction. Among the two Stenotrophomonas strains, Y042 was selected because of its superior oxidative enzyme activity (Table S3) and greater diversity in some metabolic pathways.

To predict the interactions between the studied strains, the metabolic maps of any two strains were matched to figure out their metabolic complementarities to each other. The results show that among the possible strain combinations, the most potential one, in which the strains well complement each other in terms of metabolism, is that of DDT21 and TN030, the core strains (Figure 3). Apparently, DDT21 and TN030 complement each other to complete some common metabolic pathways such as the pentose phosphate pathway or some branches of the central catabolism pathways. Such favourable complementation can be expected, as DDT21 and TN030 have the most diverse metabolic pathways, as mentioned above. Thus, we can confirm that those two strains are the right choice as core strains of the mixed cultures to be constructed.

FIGURE 3.

The ‘combined’ metabolic map of strains Streptomyces sp. DDT21 and Ralstonia sp. TN030. Lines in green and red indicate the pathways unique for DDT21 and TN030, respectively, which thus complement each other. The red node represents DDT to provide some direction guides when navigating through each metabolic map.

Additionally, the Pseudomonas strains offer the potential to support DDT21 and TN030 in the combinations due to the metabolic flexibility of Pseudomonas species. Our previous research results showed that T006 and Y077 were two single strains that were robust in degrading DDT, producing biosurfactants, and in oxidative enzyme and dechlorinating enzyme activities (Table S3). Therefore, two three‐strain mixtures were designed by combining only one of these two single strains with the core strains DDT21 and TN030. Indeed, the addition of one of such strains can complement several metabolic pathways (such as pinene and camphor degradation pathways) to those of the core strains (Figure 4, notice the blue and light navy blue lines).

FIGURE 4.

Metabolic pathways of different strain combinations. (A) (DDT21, TN030, Y077); (B) (DDT21, TN030, Y077, PAM64); (C) (DDT21, TN030, T006, PAM64, Y042). Red – TN030; Green – DDT21; Light navy blue – Y077; Blue – T006; Yellow – PAM64; Pink – Y042; Black – common pathways.

In addition to the three‐strain mixtures, it is also essential to investigate whether using more Pseudomonas strains in the combinations can offer more potential for mixed cultures. T006 and Y077 are actually very similar in their metabolic pathways (see their respective maps, Figures S5 and S9). The same observation is with PAM64 and PAM67, the other strains of the 4 above‐selected Pseudomonas strains (Figure S4 and Figure 2C). Therefore, PAM64 and PAM67 will not be present in the same combination; and the same for T006 and Y077. Thus, the two strains of each pair PAM64/PAM67 or T006/Y077 should not be present together but the strain combinations between the two pairs (PAM64 + T006; PAM64 + Y077; PAM67 + T006; or PAM67 + Y077) should be tested to evaluate their complementarity. Indeed, such a combination can offer a little more complementarity among the strains (Figure 4B), compared to the three‐strain combinations. For example, strain PAM64 can contribute some additional catabolic pathways, such as the lysine degradation pathway, to those of the three‐strain combination of DDT21, TN030 and Y77 (Figure 4B).

Belonging to a special genus distantly different from those of the strains mentioned above, strain Stenotrophomonas maltophila Y042 is predicted to increase metabolic diversity if combined in a mixed culture. Stenotrophomonas spp. are also known to be good at degrading DDT (Pan et al., 2016). In fact, the combination of Y042 with a four‐strain mixture (DDT21, TN030, T006 and PAM64) (Figure 4C) offers the supplementation of some metabolic pathways, such as the pinene, camphor and geraniol degradation pathways (left corner, bottom of the map), several amino acid metabolic pathways and pyruvate metabolic pathways, in association with the central pathways, such as a step in the glucolysis/gluconeogenesis carried by glucose‐1‐phosphatase (centre, top of the map) and a step in fructose and mannose metabolism carried by 2‐dehydro‐3‐deoxy‐L‐rhamnonate dehydrogenase (centre).

Based on the metabolic interaction analysis and predictions stated above, we proposed the 10 best mixtures that can have hypothetically enhanced performances in degrading DDT and be possibly more resilient due to their metabolic flexibility (Table 2).

TABLE 2.

The 10 best hypothetical mixtures and their strain compositions proposed on the basis of metabolic interaction predictions.

Mixture name	Strain composition
HD1	DDT21, TN030, T006, PAM64, Y042
HD2	DDT21, TN030, T006, PAM67, Y042
HD3	DDT21, TN030, Y077, PAM64, Y042
HD4	DDT21, TN030, Y077, PAM67, Y042
HD5	DDT21, TN030, T006, PAM64
HD6	DDT21, TN030, T006, PAM67
HD7	DDT21, TN030, Y077, PAM64
HD8	DDT21, TN030, Y077, PAM67
HD9	DDT21, TN030, T006
HD10	DDT21, TN030, Y077

DDT degradation: the selected mixed cultures versus the single cultures

The best 10 microbial mixtures that we predicted were tested for their capabilities of degrading DDT in liquid, semi‐solid and solid media, in comparison with those of the single strains. The difference between the DDT removal efficiency of each mixture and the average DDT removal efficiency of its respective single strains was calculated (Figure 5, detailed data in Table S4).

FIGURE 5.

True DDT‐degrading performances of the mixtures versus their respective single strains over time, as presented by the percentage difference between the actual degradation efficiency of each strain combination and the average degradation efficiency of its respective single strains. (A) In liquid media after 1 week; (B) In semi‐solid media after 1 week; (C) In solid media after 1 week; (D) In liquid media after 2 weeks; (E) In semi‐solid media after 2 weeks; (F) In solid media after 2 weeks. Detailed degradation efficiency data are presented in Supporting Information.

In liquid medium, the DDT removal efficiencies of the mixtures were slightly higher than those of the single strains, by around 10% in general (Figure 5A,D) (p < 0.05 in most of the cases). The combinations with better DDT‐degrading capabilities in liquid media after one week were HD2, HD4, HD5, HD6 and HD10 (Figure 5A), and those after 2 weeks were HD1, HD2, HD3, HD8 and HD10 (Figure 5D). The results suggest that the strains in the mixtures could have complementary interactions in DDT degradation as predicted from the genome mining, but they did not exert those very strongly in liquid media, which is probably due to their lack of close contacts in such environments.

In semi‐solid medium, the DDT removal efficiencies of the mixtures were largely higher than those of the single strains, by up to around 40% (Figure 5B,E) (p < 0.05 in all the cases). The mixtures also degraded DDT well in semi‐solid medium after 2 weeks. Compared with their DDT‐degrading efficiencies in liquid medium, those in semi‐solid medium were significant and clearly different between different mixtures. Prominent combinations, in terms of DDT‐degrading performance, include HD1, HD3, HD4 and HD7 (Figure 5B,E).

For solid medium, the DDT‐degrading capabilities of the mixtures were not as good as those of the single strains after 1 week (Figure 5C) but became much higher than the latter after 2 weeks, by more than 20% (Figure 5F) (p < 0.05 for all the cases except HD09). Overall, after 2 weeks, the combinations showing better removal efficiencies (vs. the single strains) were HD1, HD4, HD7 and HD10. These results are interesting and in fact difficult to explain. In solid media, it may take more time for bacteria to interact with each other as they need time to spread and produce extracellular chemicals, through which interactions can be executed. As the total input cell density was roughly identical for all single cultures and mixed cultures, the density of each individual strain in a mixed culture was lower than that in a single culture. It is obvious that the DDT‐degrading performances of the individual strains are not the same. Therefore, in the first week, when the interactions between the strains in mixed cultures might still be little, the single cultures performed more efficiently in degrading DDT due to the high cell density of the only strain in each culture. However, in the second week, probably due to more growth and interactions of the strains in the mixtures, their DDT‐degrading performances improved and exceeded those of the single cultures. Indeed, there is no previous reference about mixed culture biodegradation to compare, but there have been some reports that biodegradation in a solid medium could be slow at the beginning but completed faster in the end, compared to that in a liquid medium (Deroiné et al., 2015; Mbarki et al., 2019).

From all the results above, we can see that in general the mixed cultures were more efficient than the single cultures in degrading DDT. This performance difference is more apparent in solid medium, and particularly in semi‐solid medium. There cannot be a simple explanation for such an observation but we believe that the sole reason may lie in the hydrophobicity of DDT, which leads to its adsorption onto solid carried, that is, agarose in this study. In liquid medium, even with shaking incubation, DDT might still separate from the water phase, thus it was difficult for the microorganism to access and their complementary interactions might not help much, either. In solid medium, DDT might adsorb and be trapped onto the agarose surface and the microorganisms might have more access to DDT and degrade it by their extracellular enzymes, which can be improved through their metabolic interactions. However, in solid medium, it might require more time for the microorganisms to grow, spread and produce enzymes, in comparison to when they were in semi‐solid medium. That may be why the degradation of DDT by the mixtures was the most efficient in semi‐solid medium. This is actually good in terms of practical application as semi‐solid medium conditions are probably the closest to those of soil. We should not exclude another possibility that some strains could use agarose as the carbon source to grow better, which might also lead to the more efficient biodegradation in solid and semi‐solid media.

The different mixtures containing the different strain combinations also had clear differences in DDT‐degrading performance. Among them, the combinations showing better and stable performances in all medium conditions are: HD1, HD4, HD5, HD7 and HD10. Looking at the compositions of the predicted combinations (Table 2), we see that besides the two core strains, DDT21 and TN030, all these combinations also contain either Pseudomonas sp. Y077 or Pseudomonas sp. T006 plus Pseudomonas sp. PAM64, which are metabolically very similar as discussed above. Therefore, it is possible that the interactions between the Pseudomonas sp. strain with the two core strains DDT21 and TN030 may contribute to enhancing the DDT‐degrading capabilities of the combinations. Some combinations (HD1, HD4 and HD7) also contain another Pseudomonas strain, which is one of the other two similar strains PAM64 or PAM67 and may have a little additive metabolic effect. HD1 and HD4 contain another strain, Stenotrophomonas maltophila Y042, which is an auxiliary strain as discussed above. It is interesting to note that these two five‐strain mixtures are in general the best in degrading DDT (in all media). It is a question why HD2 and HD3 were not that good. It is probable that the interactions between T006 and PAM67 (in HD2) or between Y077 and PAM64 (in HD3) were not similar to those of the other pairs and could not bring about the same effects.

It is also interesting to investigate the metabolites of the degradation processes by the mixtures and the single strains. It is highly possible that the mixtures and the single strains employ different biodegradation pathways (Ding et al., 2021; Zhu et al., 2019). This should be investigated in further studies.

DDT‐degrading performances of the mixed cultures in response to different environmental conditions

The mixed cultures were grown under different conditions and their DDT‐degrading performances were evaluated through their DDT removal efficiencies (as for the test of the moisture effect) or their dechlorination activities (as for the tests of the effects of other conditions).

Considering the effect of temperature, the mixtures were capable of degrading DDT at a wide temperature spectrum from 10 to 40°C (Figure 6A). They grew and had the highest dechlorination activities at 25–30°C. In our previous study on the single strains, DDT21 and Y042 grew poorly and displayed a weak de‐chlorination activity at extreme temperatures of 10 and 40°C (Figure S10). However, the mixtures HD1, HD4 and HD8 containing DDT21 and Y042 still displayed quite good dechlorination activities at those suboptimal temperatures. This result shows that the strains might interact with each other and thus enable the mixtures to better adapt to changing temperatures. Particularly, the strains that are adapted to a wider range of temperatures, such as TN030 and Y077 (as shown in our previous results, Figure S10) might play some roles. These results support our hypothesis about the enhanced versatility of the mixed cultures due to the complementary interactions between their member strains, especially in the cases of HD1, HD4 and HD8. These mixtures will have practical advantages when used in areas with severe climatic conditions, where the temperature is below 10°C in winter or over 40°C in summer.

FIGURE 6.

Effects of environmental conditions on the de‐chlorination activities of the different constructed DDT‐degrading mixed cultures. (A) Temperature; (B) PH; (C) Carbon source; (D) Nitrogen source; (E) Concentration of pollutant; (F) Trace element solution ratio; (G) Seeding rate; (H) Moisture.

Regarding the effect of pH, the results (Figure 6B) show that the studied mixtures were able to dechlorinate well at pH 5–9, and the optimum is mostly pH 7. However, there are some exceptions, such as HD3 and HD5 dechlorinating the most at pH 9 or HD4 performing the best at pH 5. It is uncertain whether this result could be due to the microbial interactions in those combinations that allow some more‐enduring strains to play dominant roles. For example, P. lalkuanensis PAM64, a strain that grows well at pH 9 (Figure S11) could dominate in HD3 and HD5. Similarly, HD4 could be dominated by S. maltophila Y042, which is known to prefer pH 5 (Pan et al., 2016). Noticeably, HD10 had good de‐chlorination activities in a wide pH spectrum from pH 3 to 11. Interestingly, this combination contains only three strains (Streptomyces sp. DDT21, Ralstonia metalidurans TN030 and Pseudomonas sp. Y077). Probably the combination of only these 3 strains is sufficient to ensure a good adaptation to pH changes. The genera Streptomyces, Ralstonia and Pseudomonas are known to be able to grow and function at wide pH ranges (Kaur et al., 2022; Kulig et al., 2013; Zeng et al., 2020).

Another factor affecting the DDT‐degrading performances of the mixtures is nutrients, including the carbon source and the nitrogen source. In response to different supplemented carbon sources, the mixtures performed differently (Figure 6C). They tent to remove Cl⁻ better when supplemented with phenol or simple sugars, such as galactose and fructose, than when supplemented with the other complex sugars. When growing with phenol as an additional carbon source, the strong activities of the mixtures suggest that they have good capabilities of degrading aromatic compounds, which is probably related to the good DDT‐degradation potentials of the single strains. Regarding the capability of growing with other different carbon sources, we can see that the mixtures HD7, HD8 and HD10 displayed the most active and flexible dechlorination activities with the diverse supplemented carbon sources (Figure 6C). These combinations all contain the three strains DDT21, TN030 and Y77, and probably the interactions between those strains may be the reason for the better adaptation of the mixtures to different carbon sources. Interestingly, these three mixtures (HD7, HD8 and HD10) also performed more versatilely to different nitrogen sources than the others (Figure 6D), although it seemed that the latter all could use the different supplemented nitrogen sources (peptone, glycine (an easy‐to‐use amino acid), ammonium citrate, nitrate and ammonium sulphate – the nitrogen sources that are commonly accessible to microorganisms). In addition, HD2, HD6 and HD9 are also the mixtures that functioned well with various nitrogen sources (Figure 6C).

We also tested the effect of DDT concentration on the dechlorination activities of the mixtures as this is an important factor in deciding their application potential in practice (Pan et al., 2016). The results (Figure 6E) show that, in general, the chlorine‐removing rates of the mixtures decreased as the initial DDT concentration increased. This is actually understandable because when DDT concentration increases, the metabolism load for the microorganisms increases while too much DDT may otherwise inhibit them (Pan et al., 2016). Among the tested mixtures, HD10 displayed good dechlorination capabilities at different concentrations of DDT, meaning it still removed Cl⁻ at remarkable rates when the DDT concentration was high (at 20–30 mg L⁻¹). Surprisingly, this mixture was also the best adapted to diverse pHs, as discussed above, which possibly can be an explanation. Indeed, dechlorination is associated with pH change (usually lowering the pH as H⁺ is also released during dechlorination) (Brovelli et al., 2012), and thus the strain combination that can adapt better to pH changes may as well adapt better to various DDT concentrations.

The next factor that we investigated is the ratio of trace elements supplemented to the cultures as microbial growth and activities can be limited by these substances (Brock et al., 2003). The results (Figure 6F) showed that the combinations HD2, HD4, HD6, HD7, HD8, HD9 and HD10 maintained stable de‐chlorination activities with different levels of trace minerals. Furthermore, the addition of about 0.1%–0.5% trace elements to the culture media appeared to be generally suitable for all the mixtures, while adding too much might inhibit some of them.

A microbial preparation is considered more effective for practical application when it is used with a low seeding rate. Therefore, we also evaluated the performances of the mixtures of interest at different seeding rates. The results (Figure 6G) showed that, in general, when increasing the seeding rate from 5% to 20%, the growth and the dechlorination activity of every mixture increased gradually. Among the mixtures, HD6, HD9 and HD10 are those that had higher de‐chlorination activities than the others even at low seeding rates (i.e., 5%–10%). HD2, HD4 and HD7 can also be considerable.

The last factor that we investigated was moisture as its effect on microorganisms is as important as those of temperature and pH. The results (Figure 6H) showed that the mixtures generally performed de‐chlorination well at the moisture contents ranging from 20% to 30% and worse at those higher than 40%. Among them, the combinations HD7, HD8 and HD10 had stable DDT‐degrading performances under different moisture conditions. Relatively stable to be considered are HD1, HD3, HD4 and HD5.

Combining all the results together, we can see that HD4, HD7 and HD10 are the mixtures both capable of effectively resolving DDT and adaptable to a wide range of practical conditions. Such advantages may be due to the interactions between the strains in those mixtures. Noteworthily, in all those 3 mixtures, in addition to the two central strains DDT21 and TN030, there is also the appearance of Pseudomonas strain Y077. We propose a hypothesis to explain this observation as follows. First, Pseudomonas is a genus comprising bacteria that can function robustly and flexibly in different environmental conditions (Silby et al., 2011). However, not only the three above‐mentioned mixtures contain Pseudomonas strains. The uniqueness of HD4, HD7 and HD10 is the strain Pseudomonas sp. Y077, with its distinctive metabolic contribution to the mixtures. Digging into its KEGG map, we can see that adding Y077 to a mixture helps complete a metabolic pathway located near the central group of metabolic pathways in the middle of the map (Figure 4), due to an enzyme that Y077 may uniquely contain: D‐lactate dehydratase (K05523). This enzyme catalyses the reaction that converts 2‐oxopropanal to lactate. 2‐oxopropanal or methylglyoxal is a reduced form of pyruvic acid and is a by‐product of glycolysis and several other central carbon metabolism pathways (Inoue & Kimura, 1995). Methylglyoxal is a highly cytotoxic compound and can be secreted externally by bacteria (Chakraborty et al., 2014; Zhang et al., 2014). Two strains TN030 and DDT21 do not have this enzyme (Figure 3). Therefore, the supplementation of this enzyme by Y077 may help support the growth and performance of partner bacteria. In addition, Y077 also provides several supplementary pathways involved in the metabolism of several different amino acids such as glycine, serine, threonine and lysine. It is uncertain that the above‐mentioned features of Y077 are the reason for the advantages of HD4, HD7 and HD10 but this can be a possibility, which obviously requires further studies to clarify.

CONCLUSIONS

In summary, in this study we have been successful in rationally designing and constructing some DDT‐degrading mixed cultures based on genome‐mining and prediction of metabolic interactions among the single strains. The constructed mixed cultures displayed significantly improved performances versus the single cultures, in terms of both DDT degradation and adaptation to environmental conditions. The three best‐performing mixtures contained the strains that appear to have the most metabolic complementation. Furthermore, the best performance of the mixtures in semi‐solid medium suggests their application potential in practical soil bioremediation. Altogether, our results demonstrate the effectiveness of the genome‐mining approach and the use of constructed mixed cultures in improving biodegradation, especially for xenobiotics such as DDT.

AUTHOR CONTRIBUTIONS

Hai The Pham: Conceptualization. Hai The Pham and Phuong Ha Vu: Experimental design. Phuong Ha Vu, Dang Huy Nguyen, Tung Son Vu, Anh Hien Le, Trang Quynh Thi Tran, Yen Thi Nguyen and Bao Gia Truong: Experimental conducting and data analysis. Thuy Thu Thi Nguyen, Hanh My Tran and Bao Gia Truong: Technical assistance. Thuy Thu Thi Nguyen, Hanh My Tran, Linh Dam Thi Mai, Ha Viet Thi Bui, Huy Quang Nguyen, Thao Kim Nu Nguyen and Huyen Thanh Thi Tran: Discussion and Manuscript revising. Hai The Pham, Phuong Ha Vu and Dang Huy Nguyen: Manuscript drafting and finalizing. Hai The Pham, Phuong Ha Vu and Huyen Thanh Thi Tran: Project administration.

FUNDING INFORMATION

This research was funded by the Ministry of Science and Technology under grant number NVQG‐2021/DT.02.

CONFLICT OF INTEREST STATEMENT

The authors declare no conflicts of interest.

Supporting information

Appendix S1.

Vu, P.H. , Nguyen, D.H. , Vu, T.S. , Le, A.H. , Tran, T.Q.T. , Nguyen, Y.T. et al. (2024) Biodegradation of DDT using multi‐species mixtures: From genome‐mining prediction to practical assessment. Microbial Biotechnology, 17, e70021. Available from: 10.1111/1751-7915.70021

DATA AVAILABILITY STATEMENT

The whole‐genome sequences of the strains used in this study were deposited in GenBank with the accession numbers as follows: SRX21629081 for Pseudomonas sp. T006, SRX21629080 for Pseudomonas putida T087, SRX21629073 for Ralstonia sp. TN30, SRX21629083 for Pseudomonas sp. PAM64, SRX21629082 for Pseudomonas sp. PAM67, SRX21629079 for Pseudomonas sp. Y077, SRX21629087 for Stenotrophomonas maltophila Y042, SRX21629086 for Stenotrophomonas maltophila Y050, SRX21629091 for Streptomyces sp. DDT21 and SRX21629090 for Streptomyces sp. DDT23. They can also be found at the following URL: https://drive.google.com/drive/folders/1LKQaV8hqqrDVvbYAbIFHy5JUGu2xp3dW?usp=sharing.