Comparative genomics reveal unique markers to monitor by routine PCR assay bioinoculant of Sphingobium indicum B90A in hexachlorocyclohexane (HCH) contaminated soils

Abstract

Bioinoculants of Sphingobium indicum B90A have been used to decontaminate hexachlorocyclohexane (HCH)-contaminated soils in the past. There is no selective or convenient method available to track the added B90A in HCH-contaminated soils in the presence of several native sphingomonads. Here, we describe a method, BioMarkTrack, for tracking B90A bioinoculant by simple amplification of the B90A specific biomarker genes. Whole-genome sequence data of 120 different genera of sphingomonads (Sphingobium, Novosphingobium, Sphingomonas, Sphingopyxis, and Sphingosinicella) were retrieved from the NCBI database and annotated. Intra- and inter-genus similarity searches, including the genome of B90A as a reference was conducted. 122 unique gene sequences were identified in strain B90A, out of which 45 genes were selected that showed no similarity with the NCBI non-redundant (NR) database or gene sequences in the publicly available database. Primers were designed for amplification of 4 biomarkers. To validate the biomarkers B90A tracking efficacy in bioaugmented soils, a microcosm study was conducted in which sterile garden and HCH-contaminated dumpsite soils were amended with strain B90A. Amplification of the biomarker was observed both in sterile garden soil and HCH-contaminated dumpsite soil but not in control (lacking B90A) samples. Further, the primer set was used to track B90A in a bioremediation field trial soil, demonstrating the convenience and efficiency of the simple PCR-based method, which can be employed for tracking B90A in bioaugmented soils. The approach as presented here can be employed on different bioinoculants to identify unique biomarkers and then tracking these organisms during bioremediation.

Supplementary Information

The online version contains supplementary material available at 10.1007/s12088-024-01321-7.

Introduction

Globally, large number of organochlorine (OC) pesticides were used in past, in agriculture and public health care. Besides having benefits, these synthetic organic molecules are found to cause toxicity in non-targeted species due their prolonged persistence in the environment [1]. Among various methods (physiochemical and biological) of decontamination or remediation of polluted sites, microbe—mediated degradation is found to be effective. Diverse microbial species are used (either isolated from polluted sites or genetically modified) in the degradation of environmental pollutants [2]. The application of such potent microorganisms for the degradation of target pollutants into the environmental habitats (mostly soil) is defined as bioaugmentation. Several previous studies reported bioaugmentation as a good strategy for reducing the pollutant load to an appreciable level in a contaminated environment [3–6].

However, the relative progress of bioaugmentation depends upon several factors including the adaptability of the bioinoculant species at the non-native place, competitive interaction with indigenous microbial biota, predators, and various dynamic abiotic factors [7]. Therefore, monitoring or tracking the introduced bacterial species becomes imperative to ascertain the fate and persistence of the bioinoculant in the introduced environment. Standard plate count is one of the most practiced/adopted conventional methods of detection wherein microbial inoculant is identified based on morpho-physiological characteristics like cell morphology, colony color/pigment formation, growth on selective media, etc. Practicing these methods for tracking the bioinoculant in field samples becomes a strenuous exercise in the case of slow-growing bacteria and microbes sharing similar phenotypic and biochemical characteristics [8]. Further, advanced approaches such as quantitative PCR (q-PCR)- utilizing degradation pathway gene-specific primers [9, 10] and Carbon Stable Isotope Analysis (CSIA), a metabolite-based assay [11, 12], used to study the adaptation and biodegradation potential of model microorganisms were found to be promising methods, however, the high cost of nucleic acid stain/isotopes and dependency on GC–MS and GC-C- IRMS (for CSIA) make these techniques economically not feasible, especially while working on large sites in situ bioremediation. Some other culture-dependent methods for the detection of the target species include colony hybridization utilizing nucleic acids (DNA or RNA) probes that are radioactively or fluorescently labelled e.g. Fluorescent In Situ Hybridisation (FISH) [13, 14]. DNA-based techniques (such as dot/slot blot hybridization, Southern blot hybridization, and targeting for specific amplicon) facilitate the tracking of target microbe/s in complex environmental matrices and curtail culture dependence [14, 15]. The use of universal primers (sequences with conserved flanking regions) and probing either by biomarker gene sequence/s or catabolic genes in these methods could only help in the discrimination of microbial species up to genus or species level [16, 17]. Thus, identification/tracking of introduced bioinoculants in the field requires unique markers that should be species/strain-specific.

Hexachlorocyclohexane (HCH) was one of the most prevalent and extensively used organochlorine insecticide in the twentieth century whose application was eventually banned. Prolonged stability in the environment and bioaccumulation, causing toxicity in non-target organisms make HCH contamination a major global concern and priority [18]. Bioremediation attempts at many HCH contaminated soils have been executed using one or a consortium of HCH degraders [3, 5, 17, 19–24]. Conventionally, the viability of augmented (inoculated) strain/s and their HCH degradation potential have been monitored using the standard plate count method, pigment formation, and occurrence of halo or clearance zone of HCH around colonies [3, 21, 22]. Additionally, PCR amplification of HCH catabolic genes (lin genes) such as linA (encodes for dehydrochlorinase), and linB (encodes for halidohalide), products of which are involved in the HCH catabolic pathway has also been employed [3, 5]. However, the presence of lin genes in other indigenous microorganisms inhabiting the contaminated site restricts the lin-gene-based PCR method for tracking a bioinoculant. For example, microbial species such as Sphingomonas sp. NM05 [25], Xanthomonas sp. ICH12 [26], Microbacterium sp. ITRC1 [27], Sphingobium sp. strains (UM1, UM3, RL-3, F2, IP26, HDIP04) [17] isolated from an HCH dumpsite were reported to harbour lin gene/s for HCH isomer/s degradation rendering it difficult to track the bioinoculant, Sphingobium indicum B90A, a well-known HCH degrader.

Hence, an accurate method is required to track the added strain B90A in HCH-contaminated soils. Here we describe a method, termed BioMarkTrack (BMT) (Fig. S1, Supplementary material) that involves using a combination of genomics/ metagenomics approaches and routine PCR to identify bioinoculants in the bioaugmented soil using strain B90A as a model. This culture independent tracking method would serve as a useful tool in tracing B90A during bioaugmentation remediation of HCH contaminated soil habitats worldwide.

Materials and methods

Retrieval of genome sequences of Sphingomonadaceae family and genome annotation

Whole-genome sequences of S. indicum B90A and 30 other Sphingobium strains (31 genomes in total) were retrieved from the NCBI-genomes database (Table S1, Supplementary material). Genomes of different genera of Sphingomonadaceae family including strains belonging to genus Novosphingobium (26 genomes), Sphingomonas (43 genomes), Sphingopyxis (16 genomes) and Sphingosinicella (4 genomes) were also retrieved from the NCBI-genome database (Table S1, Supplementary material). The selected genome sequences (120 genomes) were annotated using Glimmer gene-finder [28] at Rapid Annotation Subsystem Technology (RAST) server [29]. The amino acid and nucleic acid sequences were retrieved from RAST server for further analysis. The functional annotation of S. indicum B90A genome was performed using RAST and Prokka [30].

Phylogenetic relatedness of Sphingobium strains

Phylogenetic clustering of 31 Sphingobium strains was performed on Pairwise Average Nucleotide Identity (ANI) values. ANI was performed using ANIb method from pyani master pipeline [31]. A two way-matrix obtained using ANI values was further processed and distance matrix and Hierarchical clustering were done using Pearson correlation algorithm on Multi experiment Viewer (MeV) [32]. The dendrogram so obtained was visualized on interactive Tree of Life (iTOL) server [33] (https://itol.embl.de/).

Identification of genes specific to S. indicum B90A

To determine the unique specific genes or markers of B90A, pan-genome was calculated using GET_HOMOLOGUE at 60_cov, 60_identity with 31 Sphingobium genomes included in the study. From pangenome matrix, genes accessory and specific to strain B90A were notified and extracted. Next, these gene sequences were then searched in the genomes of sphingomonads (other than Sphingobium strains) included in the study using all versus all blastn search [34]. All the genes that showed > 50% identity with any genes of sphingomonads were then excluded and genes with no or lower similarity were selected for further analysis. These finally selected genes were then searched against non-redundant database of sequences at NCBI BLASTn [34]. At the last, genes were sorted based on query coverage (QC), and genes showing QC < 30% were selected for further analysis. Thus, genes sharing very little or no significant similarity with any gene or nucleotide sequence at the publicly available database repository were included for further analysis.

Primer designing and specificity evaluation

Among the 45 unique genes identified in B90A, four genes were randomly selected showing least or no mapping to the metagenomic data (see next section for details). Using primer BLAST [35], four sets of primers (forward and reverse) designated as P1, P2, P3, and P4 were designed corresponding to the four unique genes of B90A. The primers sequence, GC content, and encoded protein functions have been listed in Table 1. Specificity of these primer sets for B90A were evaluated further. Firstly, genomic DNA of strain B90A was used as a template to amplify the unique regions using individual primer sets P1, P2, P3, and P4. Thermocycling was done using Peqlab PCR system (PEQSTAR, UK) under the following cycling conditions: an initial denaturation at 96 °C for 5 min followed by 30 cycles of denaturation at 96 °C for 45 s, annealing for 45 s at 55 °C and extension at 72 °C for 1 min with final extension for 8 min at 72°C. Strain specificity of primer pairs for strain B90A was analysed further through a second PCR amplification using genomic DNA of close monophyletic strains or subspecies of S. indicum that included S. lucknowense F2, Sphingobium sp. HDIPO4, S. chinhatense IP26, and S. japonicum UT26 [36].

Table 1.

List of four selected marker genes of S. indicum B90A

S. No	Gene ID	Forward primer sequence (5′–3′)	GC%	Reverse Primer sequence (3′–5′)	GC%	Product length (bp)	Function
1	B90A.3228	ATCGCCTTTCCATCTTCGG	53	AAGAAGGGACCAAGGGTTGC	55	947	Hypothetical protein
2	B90A.308	GCTGGTGACGACCTATGACT	55	GTTCCAGCCAGGTTCAATCG	55	1108	Polymyxin resistance protein ArnT, undecaprenyl phosphate-alpha-L-Ara4N transferase; Melittin resistance protein PqaB
3	B90A.3700	TGTCTTTCTTCGGCCTCTTCG	52.38	ATGAAATGAGATCGGCGGTTG	47.62	900	Hypothetical protein
4	B90A.285	CCGCTGGTCTACACGGAAAA	55	CCAGAGGTACGCCCATTCAC	60	801	Putative O-antigen synthesis protein, WbyH

Details like Gene IDs, primer sequences (forward and reverse), GC%, product length and encoded functions of genes have been tabulated

Mapping on hexachlorocyclohexane (HCH) dumpsite metagenome

Cultivation independent approach was employed to evaluate the uniqueness of four marker genes of B90A by checking them against the non-cultivable microbial community. For this, Hiseq Illumina paired end metagenome data (1.1 Gb) of HCH dumpsite, Lucknow, India (Accession number- ERP001726) was mapped over specific gene sequences of B90A using mummer module of MUMmer3.23 software [37] at -maxmatch -b -c -F parameters. For employing the strategy to other bioremediation systems, this step can be skipped if metagenomic data of the contaminated site is not available however, maximum number of genomes and metagenome together can refine the identification of unique genes. 16S rRNA gene sequence of strain B90A was also mapped with metagenome data of HCH dumpsite as control.

Further molecular validation of marker gene specificity for B90A in the genome pool present in HCH dumpsite soil was determined by PCR-amplification of soil metagenomic DNA using unique primer sets. For this exercise, metagenomic DNA was extracted from B90A inoculated HCH dumpsite (located at Ummari village, Lucknow, Uttar Pradesh, India; 26° 51′ 0.0000″ N & 80° 56′ 59.9892″) soil using Power Soil Kit (QIAGEN, Germany). Thermocycling was performed following similar cycling conditions as stated above.

Marker-based tracking of strain B90A in soil—microcosm study

Tracking of strain B90A in soil was carried out in different microenvironments employing our devised molecular method along with the phenotypic approach. Three soil types-autoclaved garden soil (a non-competitive habitat), autoclaved HCH contaminated soil (moderately conducive habitat), and non-autoclaved HCH soil (stressed habitat) were used in the microcosm experiment.

Soil collection and preparation

HCH contaminated soil was collected from the HCH dumpsite, located at Ummari village, Lucknow, Uttar Pradesh, India (26° 51′ 0.0000″ N & 80° 56′ 59.9892″ E). Garden soil was acquired from the campus garden at University of Delhi (28.6883° N, 77.2102° E). Soil samples were air-dried and finely sieved. 500 g of processed soil types were used for experimental set up. Soil types (garden soil and HCH polluted soil) were sterilized by autoclaving (121°C, 15 psi, 25 min).

Bacterial biomass for inoculation in soil

Primary microbial inoculum of strain B90A was raised in sterile LB-media (Difco-LB-Broth) from stock culture and incubating at 30 °C with continuous shaking at 200 rpm. After 36 h, 1% (v/v) of the preculture was transferred (1/100 volumetric ratio) into three 2L flasks containing 500 ml LB-media each. The flasks were incubated at 30 °C with continuous shaking at 200 rpm to achieve the OD_600nm between 1.0–1.5 (OD₆₀₀ 1.0 ⁓10⁸ cells (CFU/mL).

Establishment of soil model system/microcosms

500 g of each of the processed soil type was transferred in a separate tray (250 × 300 mm). B90A was harvested by centrifugation at 4000 rpm for 20 min at 4°C. Pelleted cell biomass was washed twice with 0.9% saline solution before mixing into the respective soil systems. Uninoculated soil samples were maintained as a control. All microcosm setups were incubated at room temperature. Parameters like aeration and moisture content were maintained in accordance with previous microcosm studies employing B90A [3, 17].

Soil sampling

Periodic soil sampling was done from 0 day to 12th day at an interval of 2 days, in pre-context of phenotypic detection of bacterial strain by plate counting and further for marker-based identification. Briefly, soil was mixed thoroughly, one gram of soil sample was taken randomly from six different points and pooled together to form 6 g of composite soil. From this composite sample, three replicates of soil sample, one gram each was taken for viability detection by culture plate method, and another 3 replicates (1 g each) were kept at -20°C for DNA extraction.

Detection of strain B90A in soil

Presence of strain B90A in soil was determined by standard plate count method. For this, 1 g of sampled soil suspended in 0.9% sterile saline solution was vortexed thoroughly for 1–2 min. Further, soil mix was serially diluted and 100 μl was plated onto the streptomycin (100 μg/ml) amended LB agar plates. Plates were incubated for 3–4 days at 30 °C for bacterial growth.

For marker-based detection of strain B90A in soil samples, microbial DNA was extracted from the stored (− 20 °C) soil samples using DNeasy Power Soil Kit (QIAGEN, Germany) according to manufacturer’s protocol. Concentration and purity of metagenomic DNA was assessed by Nano-Drop Spectrophotometer (Thermo-Scientific). PCR cycling was carried out as described above using one of the unique primer sets (P4—Forward primer—5′-CCGCTGGTCTACACGGAAAA-3′) and Reverse primer—5′-CCAGAGGTACGCCCATTCAC-3′), using metagenomic DNA as template. Equal amount of template DNA (100 ng) was taken for each soil sample for amplification. The amplified products were resolved by electrophoresis on 1% agarose gel and visualized using Amersham imager (GE).

Marker-based tracking of strain B90A in the bioremediation field

A bioremediation field trial is being conducted at an HCH dumpsite—total area 3852 m²; situated at Ummari village, Lucknow, Uttar Pradesh, India (26° 51′ 0.0000″ N & 80° 56′ 59.9892″ E). The total area is divided into 6 treatment plots meant for different bioremediation strategies including bioaugmentation (60 m × 19 m, 1140 m² area). Bioinoculum of strain B90A (10⁵–10⁶ cells g⁻¹ of soil) has been applied in this area. Tilling of bioaugmented soil was periodically performed for aeration, soil homogenization and mixing of the bacterial culture. Scheduled soil sampling and metagenomic DNA isolation (of the sample taken in March 2020) was carried out in a similar way done for microcosm study. Strain B90A was tracked via marker based as well as culture plate method as described above.

Results and discussion

Identification of unique/marker genes of S. indicum B90A

The amino acid sequences of all 31 Sphingobium genomes, including the target strain B90A were obtained by genome retrieval and annotation (refer Materials and Methods section). Pan genome analysis within the amino acid sequences obtained for all Sphingobium genomes was done. 155 protein sequences (Table S2, Supplementary material) were identified, which only belong to or found in strain B90A. Out of 155, only 122 protein gene sequences were selected (Table S3, Supplementary material), after finding out the percentage similarity of genes with other representative strains of 4 sphingomonad genera using all v/s all Blastn [34]. Excluded 33 genes were showing more than > 50% percentage identity with at least any one strain of sphingomonads. Next, sequence specificity of selected 122 genes with publicly available nucleotide or genome sequences was searched against the non-redundant (NR) database of NCBI using blastn. We report here, 45 of such gene sequences (Table S4, Supplementary material) which can be used as unique biomarkers of S. indicum B90A. These genes were sorted on the basis of coverage with the maximum identical sequence hit on NCBI blast (Table S4, Supplementary material). Gene sequences with QC < 30% were considered as the highly specific sequences of B90A. Finally, four genes (keeping sequence length near 1000 bp) were selected for further steps or narrowing down our search for marker/s for strain B90A. These were labelled as B90A.3228, B90A.3700, B90A.308 and B90A.285 (Table 1).

Priming of B90A genome with the designed primer sets

Four primer sets were designed, for each marker gene named P1, P2, P3, and P4 that specifically targeted strain B90A. PCR amplification of target genome using each primer set produced amplicon sized 947 bp, 1108 bp, 900 bp and 801 bp with P1, P2, P3 and P4, respectively (Fig. 1). No band other than expected/desired amplicon size was observed with any primer set that showed specificity and uniqueness of designed primer sets for a particular region in B90A genome. Hence, it proves that overall success of PCR based experiment, majorly depends upon the selection of specific and efficient primer set. Specificity (frequency of mispriming) and efficiency (ability to twofold increase in product for each PCR cycle) of primers primarily rely on the length and G + C content of the oligonucleotide [38]. Non-specificity of used primer sets for other than target gene sequences in B90A genome further emphasize it to be a good PCR based approach in tracking B90A in HCH contaminated site rather than using catabolic genes (lin genes) derived primer set as utilized in previous studies [3, 5]. The lin genes are reported to be present in varying copy numbers within specific species genome [39, 40] and shown to be acquired by microbial species present at the HCH contaminated site [17, 41]. For example, B90A genome itself contain two copies of linA gene [42]. Association of these lin genes with IS6100 elements results in high degree of genomic rearrangement which could result in either sharing or loss of gene [21, 43] within neighboring species and from species itself [44, 45]. Considering all such factors, it become imperative to use unique marker gene sequence/s for biomonitoring of microbial species during bioremediation approach. While comparing the four selected primer sets (Table 1) based on parameters like length, G-C content and melting temperature (Tm), it was observed that all sets shared similar characteristics. Since amplification produced by all the primer sets with B90A were clean with no non-specific binding and shared very similar characteristics, one primer set (P4) for B90A.285 was randomly chosen to employ in further tracking analysis. As mentioned earlier, an amplicon of size 801 bp was obtained on amplification with this primer set (P4) during strain specific PCR test.

Fig. 1.

PCR amplification of S. indicum B90A genomic DNA with four marker genes derived primer sets designated P1, P2, P3, and P4. Marker- 1 kb ladder (100 bp- 10kbp size fragments); lane 1–4 amplicon (sized 947 bp, 1108 bp, 900 bp and 801 bp) obtained with primer set P1, P2, P3 and P4, respectively

Metagenome mapping of marker genes of strain B90A

The main idea of the study was to detect the bioaugmented microorganism/s in a contaminated site. Thus, it is required to map the identified unique markers against cultivable and non-cultivable community as well of a contaminated site, the HCH dumpsite in the presented case. Prior metagenomic analysis of HCH contaminated site, reveals the abundance of genes coding for various enzymes or proteins which play significant role in adaptation/survival and degradation potential of native microbial communities, belongs to varied taxa [46–48]. Additionally, enormous transfer/exchange of catabolic (lin) and other specific genes (through conjugation and transposition), associated with HCH degradation within the established microbial species reported also. The coverage of four marker genes of B90A was determined in the metagenome of a HCH dumpsite (located in Lucknow, India) which was sequenced as a part of a previous study [49]. Less than 25% coverage was observed in case of B90A.3228, B90A.3700 and B90A.308 while no sufficient data was aligned over the sequence of B90A.285 while mapping the selected marker genes of strain B90A over the illumine Hiseq paired end data (1.1 GB) using mummer module of MUMmer3.23 (Fig. S2, Supplementary material). The finding supported our selection of primer set 4 for B90A.285, the unique marker gene for further analysis. 16S rRNA gene sequence of strain B90A, included as the control, showed coverage of ~ 90% when mapped with metagenome data, thus, projecting the uniqueness of the selected marker genes of B90A.

To validate the in-silico finding, PCR amplification using primer set 4 was performed using the DNA from the HCH dumpsite soil. Desired amplicon was not observed which emphasizes the uniqueness of the marker gene in the dumpsite soil metagenome. DNA isolated from strain B90A inoculated dumpsite soil gave an amplicon of size 801 bp revealing its presence only in strain B90A (Fig. S3, Supplementary material).

Selection of closest neighbours of S. indicum B90A using phylogenomic clustering and exclusive priming of B90A

The closest neighbours are likely to have similar genetic content and may provide similar results, thus to draw the uniqueness of B90A markers, identification of closest neighbours of strain B90A was required. The close relative bacterial species typically share conserved core genes and variable accessory genes that often move laterally between strains. A previous report has shown the close association of S. indicum B90A with three strains, S. japonicum UT26, S. chinhatense IP26 and Sphingobium sp. HDIPO4 [45]. Isolated from geographically distinct locations/places, these Sphingobium species shares only similar stressed condition/selection pressure, the HCH contamination. Another study has also proposed 11 Sphingobium strains as heterotypic synonyms among which S. japonicum UT26, S. chinhatense IP26, S. francense DSM26779 and S. lucknowense F2 were reclassified as heterotypic synonyms of S. indicum B90A. It also showed close monophyletic clustering between these strains on the basis of 16S rRNA sequences [36]. To figure out degree of genetic closeness or heterogeneity, genetic analysis should include all genetic elements (like primary chromosome, plasmids and secondary chromosome). As, in one of the study by Sangwan et al. [49], including S. indicum B90A and S. japonicum UT26 strains, phylogenomic analysis (ANI based) grouped these as genetic subspecies (ANI- 98.04%) whereas, S. sp. SYK-6 and S. chlorophenolicum L-1 were found distant related (ANI- 89%) to S. japonicum UT26, reported earlier closer using 16S rRNA gene sequence similarity [50]. Thus, in the present study, phylogeny of 31 Sphingobium strains obtained Average Nucleotide Identity (ANI) matrix was constructed. The dendrogram depicted close clustering between S. japonicum UT26, S. chinhatense IP26, S. francense DSM26779, Sphingobium sp. HDIPO4, S. lucknowense F2 and S. indicum B90A (Fig. S4, Supplementary material), like the previous reports. These strains were found to share > 97% of ANI values which was clearly observed in distance-based matrix of Sphingobium strains (Fig. S4, Supplementary material). Two species, Sphingobium algorifonticola TLA-22 and Sphingobium phenoxybenzoativorans SC3 were observed as most diverged and clustered separately with sharing least ANI values of 75.5% and 76.3% with other Sphingobium spp. (Fig. S4, Supplementary material).

Absence of marker gene corresponding amplicon (801 bp) in PCR amplification of genomes of neighbouring strains, S. lucknowense F2, Sphingobium sp. HDIPO4, S. chinhatense IP26, and S. japonicum UT26 validate the uniqueness of the selected gene marker for B90A (Fig. S5, Supplementary material).

Tracking B90A in bioaugmented soil microcosms and field trial soil

To fortify computer-simulated exercise for the identification of unique gene in B90A as a practiced wet-lab molecular approach, bioaugmented soil microcosms were set up and the bioinoculant S. indicum B90A was tracked by culture-based (plate count) and BMT methods. B90A was also tracked by both the methods in the bioremediation field trial soil.

Culturing based (plate count) detection of strain B90A

Respective soil microcosms were initially inoculated with strain B90A equivalent to 1 × 10⁸ cells per gram of soil which was determined as 1 × 10⁴ CFU per gram of soil via selective culture plating. Small, round yellow-colored bacterial colonies of strain B90A observed on plates along with production of water-soluble brown pigment during soil bacteria extraction from soil samples of microcosms. Brown pigment formation is the characteristic feature of strain B90A which differs it from other morphologically (yellow colored) identical sphingomonads [51]. In case of autoclaved and non-autoclaved HCH dumpsite soil microcosms, bacterial colonies of strain B90A were observed only in the soil samples on the day of inoculation (0-day) (Table 2). In the autoclaved garden microcosm soil, B90A was detected up to fourth day of inoculation (Table 2), with decrease in inoculant biomass in succeeding days along with growth of invader microbial species. A rapid fall in cell number of the inoculated B90A in HCH contaminated soil microcosms might occurred due to HCH toxicity as a very high HCH load (450 g kg⁻¹) in dumpsite soil was reported in a previous study [52]. Such high HCH load inhibited inoculated B90A to proliferate or survive in the soil matrix and rendered it unable to be detected phenotypically after the day of inoculation. Further, B90A biomass was not immobilized or encapsulated in any kind of matrix before added into the soil which would enhance the survival/longevity of bioinoculant as advocated in previous studies [3, 5, 53]. Monitoring of strain B90A via culture plating in case of autoclaved garden soil was possible due to absence of HCH. Phenotypic detection of B90A only up to fourth day of inoculation attributed to rapidly growing invader microbial species that took over, making B90A undistinguishable on the culture plates. Predatory and competitive interaction with such microbial occupiers either kill/inhibits or slowdowns the proliferation of bioinoculant biomass [54] which make hard to monitor inoculant bacterial species physiologically. No distinct bacterial colonies of strain B90A was observed in the soil samples from the field site after bioaugmentation. Strain B90A remained undetected even in the 0-day soil sample plates (Table S5, Supplementary material) due to confluent growth of other/non- specific antibiotic resistant native microbial community species, inhabiting HCH dumpsite soil.

Table 2.

Cell count of S. indicum B90A inoculated in soil microcosms for different time points

Sampling time (days)	Sterile garden soil	Sterile HCH contaminated dumpsite soil	Non-sterile HCH contaminated dumpsite soil	Control soil (Uninoculated Garden/HCH contaminated soil)
	Number of B90A bacterial cells (CFUg−1 of soil)
0	8.2 × 104	5 × 104	6 × 104	–
2	1.2 × 104	–	–	–
4	3 × 102	–	–	–
6	–	–	–	–
8	–	–	–	–
12	–	–	–	–

“–”No bacterial cell observed on agar-plate; Plate count performed in triplicates and average of triplicates given in table

BioMarkTrack-based detection of B90A

Tracking of strain B90A, inoculated in sterile and non-sterile soil types was done by conventional PCR using unique primer pair (P4). DNA were extracted from respective soil microcosms and amplified for unique gene amplicon. Distinct amplicon band (801 bp) confirms the presence of strain B90A in soil. The amplicon could be detected till day-6 soil samples from both autoclaved and non-autoclaved HCH soil microcosms (Figs. 2, 3). However, in autoclaved garden soil strain specific amplicon band was observed till day-4 soil samples (Fig. 4). Comparing the detection extent of B90A in HCH contaminated soil microcosms (both sterile and non-sterile) phenotypically (culture plating) or via molecular (marker-based PCR) approach, we were able to detect the strain B90A for more days after inoculation with the latter method. HCH toxicity, lack of growth supplements as no growth stimulants were added in the soil microcosms, could be the possible reasons for decline in B90A cell biomass that results in observed variation. Further, in autoclaved garden soil B90A was tracked phenotypically and molecularly (DNA) up to 4th day of inoculation. The detection could be possible for more days in garden soil but growth of non-specific microbes might predate B90A and degrade the DNA content present in the soil as compared to HCH soil microcosms in which due to high toxicity of HCH, growth of invader bacterial species would have been less. Next, in soil samples from B90A-augmented bioremediation field, B90A was detected from day-0 to day-4 (Fig. 5). At field scale, factors like uneven distribution of bacterial cells in the field soil, predation by indigenous microbial population and insufficient growth supplements could be the reason for the decrease in bioaugmented biomass and non-amplification after 4 days of inoculation [3, 5, 20]. Variation in bacterial cell count in on-site and pot-scale level bioremediation approach was reported in a study [5], due to proper soil homogenization and controlled experimental conditions in pot scale.

Fig. 2.

Marker gene-based detection of S. indicum B90A in field soil (autoclaved) microcosms at different time periods. Marker 10kbp ladder (100 bp- 10kbp); Lane-1 0 day; 2- 2 days; 3- 4 days; 4- 6 days; 5- positive control; 6- negative control

Fig. 3.

Marker gene-based detection of S. indicum B90A in field soil (non-autoclaved) microcosms at different time periods. Marker 10 kbp ladder (100 bp-10kbp); Lane-1 0 day; 2- 2 days; 3- 4 days; 4- 6 days; 5- positive control; 6- Negative control

Fig. 4.

Marker gene-based detection of S. indicum B90A in garden soil (autoclaved) microcosms at different time periods. Marker 100 bp ladder (100 bp-3kbp); Lane-1 0 day; 2- 2 days; 3- 4 days; 4- 6 days; 5- negative control; 6- positive control

Fig. 5.

Marker gene-based detection of S. indicum B90A mixed in HCH dumpsite field soil at different time periods. Marker 100 bp ladder (100 bp-3kbp) Lane-1 0 day; 2- 1 days; 3- 2 days; 4- 4 days; 5- 6 days; 6- positive control; 7- negative control

The detection of B90A both in microcosms and in field trial soil could be carried out with high efficiency during initial days however the specificity significantly decreased later possibly since the bioinoculum released into the soil often result in transient loads of the microbial strain/s that generally fade away with time [8]. As stated earlier, this could be due to the sum of multiple variables such as HCH toxicity, lack of growth supplements, predation by native microbes, etc. making the survival of the bioinoculant difficult to understand or predict. For detection, however, BioMarkTrack method came out as a useful tool as B90A could be detected in bioaugmented soil even for lesser cell number/biomass that remain undetected phenotypically by plate count method.

Conclusion

Biomonitoring of augmented microbial biomass in bioremediation under natural conditions is a challenging task and often a major impediment in assessing the success of bioremediation. We have devised a biomarker-based tracking method, BioMarkTrack (BMT), which can be employed to successfully detect a bioaugmented microorganism. Unique marker genes of the bioinoculant are identified by in silico analysis and used to design primer sets for PCR amplifying them in bioaugmented soils. We demonstrated the usefulness of the BMT method by detecting a pesticide (HCH) degrading bacterium, Sphingobium indicum B90A, in bioaugmented microcosms and remediation field soils. Our results reveal that bioaugmented B90A could be detected till later time points (06 days after inoculation in HCH soil microcosms and 04 days after inoculation in the HCH bioremediation field site soil) employing the BMT detection method as against the plate count method, where B90A could be detected at 0 day of inoculation only in the HCH soil microcosm. This clearly reflects the utility of the BMT in tracking a bioaugmented organism. Although there are limitations of PCR such as requirement of specialized equipment (which is routinely present in all molecular biology labs), careful handling and storage of samples before extraction of soil DNA and presence of DNase, RNase and organic compounds like humic acid and fulvic acid in the soil that inhibits or results in poor DNA amplification [8, 55] and require standardization, BMT method still holds ground for better detection over conventional morphophysiological methods such as, plate count method. Also, this culture independent method involving in silico identification of bioinoculant specific biomarkers and their amplification directly from the soil DNA intends to overcome the limitations of culturing, time constraints (for slow growing bioinoculant, for example), higher cost and complexities associated with the advanced machines/techniques usage in other methods. Once the set of biomarker genes have been identified and validated for detecting the presence of an organism, the BMT protocol can be repeatedly used with an ease of simple PCR amplification for biotracking a bioinoculant.

Supplementary Information

Below is the link to the electronic supplementary material.

Data availability

The accession number of genome sequences used in the analysis have been listed in Table S1 (Supplementary material). Metagenomic reads of HCH dumpsite were retrieved from NCBI GenBank database with accession number: ERP001726.

Declarations

Conflict of interest

The authors declare that they have no conflict of interests.