Surfactants to Enhance Bioavailability of Recalcitrant Residual Petroleum Contamination

A rhamnolipid biosurfactant altered microbial community composition and hindered hydrocarbon biodegradation in fire-impacted contaminated soils from Lac Megantic, Quebec

 

 

Abstract

The use of surfactants to enhance the bioavailability of recalcitrant residual petroleum contamination has been proposed. However, surfactant addition often does not enhance biodegradation ofhydrocarbons, and has been attributed to toxicity. Herein, we show application of biosurfactant (rhamnolipid) in completely mixed, batch slurry bioreactors with soils from a petroleum-contaminated site led to a shift in the microbial community and a decrease in diversity, and the hydrocarbon-degrading bacteria likely used the rhamnolipid as the preferred carbon source. Addition of low and high concentrations of rhamnolipid showed no or an inhibitory effect, respectively, on biodegradation extents, compared to nutrient-amended systems. Dominant indigenous hydrocarbon-degrading bacteria isolated from the soil exhibited growth on rhamnolipid as a sole carbon source, suggesting similar taxa were potentially involved in the biodegradation of hydrocarbons and the uptake of the biosurfactant. This study provides new insights on the microbial community resposes of biosurfactant application in bioremediation.

1. Introduction

A catastrophic railway accident in July 2013 led to the death of 47 people, and the partial destruction of Lac-Mégantic,Quebec, Canada, due to spillage of oil being carried as freight and a subsequent fire from the oil spill. The train was carrying light hydrocarbons from the Bakken formations in North Dakota, U.S.A. It is estimated that over seven million liters of oil spilled, which  burned at the site or were released into the adjacent lake and river (Saint-Laurent et al., 2018). The contaminated soils were treated in full-scale biopiles, which led to a decrease in petroleum hydrocarbon contamination levels. However, some of the hydrocarbons remained un-degraded. An extremely slow rate of bioremediation following an initial phase of rapid decrease in PHCs has been observed in several studies (Akbari & Ghoshal, 2014; Gomez-Lahoz & Ortega-Calvo, 2005). It has been suggested that the limited bioavailability, due to low rates of dissolution of hydrocarbon solutes from the oil phase, low rates and extents of desorption, and sequestration of PHCs in pores inaccessible to bacteria in  soil aggregates, limit the extent of biodegradation of the potentially biodegradable fraction of petroleum resulting in a more persistent residual contamination (Akbari & Ghoshal, 2015a; Akbari et al., 2016; Ghoshal et al., 2004).  The residual contamination is often higher than regulatory clean-up limits. Even when not bioavailable to microbes, the residual contamination can affect humans and other ecological receptors  through specific exposure routes, and thus  may pose a risk to human health and the environment.

Several studies suggest the addition of surfactants enhances the biodegradation of high molecular weight hydrocarbons (Bueno-Montes et al., 2011; Posada-Baquero et al., 2019; Zhu & Aitken, 2010). Surfactants are amphiphilic compounds  with a hydrophobic tail, and a hydrophilic head, which partition at the oil-water interface and cause formation of oil-water emulsions. Formation of such emulsions increasesthe oil-water interfacial area for bacterial attachment and thus enhances the bioavailability and biodegradation of hydrocarbons. At surfactant doses where the  critical micelle concentration (CMC) is exceeded, micelles are formed in the aqueous phase which can partition significant amounts of hydrophobic, poorly soluble hydrocarbon compounds from the oil phase and effectively increase its solubility and thus bioavailability and biodegradation (Harvey et al., 1990; Mukherjee & Das, 2010). Some studies have also suggested enhanced biodegradation at sub-CMC concentrations (Zhu & Aitken, 2010),  and attributed it to “swelling of the soil−tar matrix” and the subsequent increase in diffusivity of hydrophobic compounds within the contaminant matrix (Yeom et al., 1996).

There are also several studies reporting no, or an inhibitory effect of surfactant addition on the biodegradation rates and endpoints (Jin et al., 2007; Shin et al., 2005). Mechanisms suggested for the negative effects of surfactant addition include increased hydrocarbon concentrations  which can be toxic to ecological receptors, and the antimicrobial activity of surfactants itself (Shin et al., 2005) or decreased microbial attachment at oil-water interfaces (Ortega-Calvo & Alexander, 1994). Alternatively, surfactants can potentially serve as a source of carbon for microorganisms. These hypotheses remain mainly untested in complex microbial communities from environmental samples. The majority of studies on surfactant addition for PHC bioremediation were conducted with a focus on physicochemical aspects of bioavailability, or in pure cultures whereas the microbial ecology and dynamics of microbial communities in response to surfactant addition have been largely overlooked.

Only two studies have investigated the effects of surfactant addition on microbial communities concurrently with assessments of biodegradation of hydrocarbons. Singleton et al. showed that application of Brij30, an anionic surfactant, and polyoxyethylene sorbitol hexaoleate, a nonionic surfactant, at sub-micellar concentrations enhanced the biodegradation rate of polycyclic aromatic hydrocarbons (PAHs) (Singleton et al., 2016). This was associated with an increase in the relative abundance of Alphaproteobacteria  affiliated operational taxonomic units (OTUs). The authors isolated strains which were able to grow on surfactants. However, those isolates were mostly unable to transform PAHs and were not associated with the most abundant OTUs or those enriched in the surfactant-amended systems with enhanced biodegradation. Therefore the behavior of those isolates do not provide insight on petroleum hydrocarbon degradation activity in the soil used in that study. In another study, Lladó et al. reported that application of Brij30 at above CMC concentrations resulted in a decrease in the size and diversity of hydrocarbon-degrading populations, and hindered biodegradation (Lladó et al., 2015). The authors attributed the observed effects to the toxicity of the applied surfactant to hydrocarbon degrading community.

Biosurfactants are predominantly produced by hydrocarbon-degrading and soil  microorganisms and generally have several advantages over synthetic surfactants: better biodegradability and lower toxicity, can be made from renewable sources or waste oils, and greater stability and performance at extreme temperatures, pH, and salinity (Bordoloi & Konwar, 2009; Mulligan et al., 2001; Souza et al., 2014). Biosurfactants act similarly to synthetic surfactants in terms of enhancing hydrocarbon solubilisation, oil emulsification and/or mobilization from pores, but also have well-known biological impacts that include inducing or influencing: bacterial toxicity/pathogenesis, microbial motility and adhesion, biofilm formation and displacement, cellular differentiation, cell signaling, chelation of toxic metals, and serving as a growth factor by facilitating uptake of substrates/nutrients or its uptake as a carbon source (Cameotra & Makkar, 2004; Christova & Stoineva, 2014; Kitamoto et al., 2002; Lang, 2002; Van Hamme et al., 2006; Warne Zoueki et al., 2010). Although the positive influence of biosurfactants on bioremediation is often reported, there have been numerous cases where no effects or negative effects were observed (Das & Mukherjee, 2007; Ławniczak et al., 2013). For example, Vipulanandan and Ren (2000) found that although a rhamnolipid biosurfactant enhanced naphthalene solubility by 30 times, it had little impact on its biodegradation by a Pseudomonas sp. The authors assumed, based on turbidity measurements that this was due to the biosurfactant being biodegraded as a competing substrate (Vipulanandan & Ren, 2000).

Despite the findings from several studies on the effects of surfactants and biosurfactants on the biodegradation performance and other microbial functions of bacterial isolates, the dynamics of the indigenous microbial community during bioremediation and in response to surfactant addition is not yet well understood. Studies with bacterial isolates may not explain microbial responses at the community level in contaminated soils because only a tiny fraction of bacteria are culturable and also that interactions and syntrophic relations in various groups of microbial communities can significantly impact their response.

Thus, we hypothesized that biosurfactant application has impacts on hydrocarbon bioremediation beyond the physio-chemical effects such as solubilization of hydrocarbon compounds and emulsification of oil, by altering the microbial community structure. Changes in the microbial community could be, for example, because biosurfactants are used as a preferred carbon source by hydrocarbon-degrading bacteria or due to toxicity.

We evaluated the effect of rhamnolipid addition at various doses on the biodegradation of petroleum hydrocarbons and the dynamics of microbial communities in a weathered oil-contaminated soil partially treated in industrial-scale biopiles, but with a significant residual contamination. Hydrocarbon-degrading  strains were isolated from the soil, and then their abilities to grow on biosurfactant as the sole carbon source in pure culture was examined. This approach determination of how the behavior of isolates in pure cultures correlated with their behaviour once present in the complex communities.

2. Material and Methods
   1. Microcosm experiments

The soils used in this study were from a biopile-treated, contaminated soil from Lac Mégantic and was provided by Englobe Corporation (Quebec). The physico-chemical properties of the soil are summarized in Table 1. The rhamnolipid (R90) were manufactured by AGAE Technologies and purchased from Sigma-Aldrich. Slurry experiments with 5 g soil and 35 mL Bushnell-Haas broth culture media (BH) were prepared in 250 mL amber glass jars with Teflon lined and sterilized screw cap, according to methods described elsewhere (Akbari & Ghoshal, 2014). The BH media was used because it was found to be effective in biostimulating hydrocarbon-degrading bacteria. (Bell et al., 2013; Chang et al., 2018)

Seven different slurry bioreactor systems representing controls and treatments containing BH and rhamnolipid at doses ranging from 1 to 100 CMC were set up. The controls included a killed control (autoclaved soil), unamended control (distilled water), BH (nutrient amendment) only, whereas the treatment systems BH along with rhamnolipid at  doses of 1 CMC (20 mg/L), 4 CMC (80 mg/L), 30 CMC (600 mg/L), 100 CMC (2000 mg/L). A total of 101 of microcosms were set-up. At various sampling time points, three bioreactors for each treatment were sacrificed for chemical and microbial analyses. The slurry bioreactors were mixed at 175 rpm and incubated at 17° C, the average summer temperature of the biopile site at Lac Mégantic.

2.2.  Total petroleum hydrocarbons analysis

At each sampling point, slurry microcosms pertaining to all treatments were sacrificed and sent to Maxxam Analytics (Montreal, Quebec) for Total Petroleum Hydrocarbon (TPH) analyses. Chemical extraction involded mixing slurries with a mixture of hexane-acetone (50%-50%) as solvent followed by a concentration step. Aliquots of concentrated-extractant were then analyzed by a gas chromatograph with a flame ionization detector (GC-FID, Agilent 6890) with a DB1 column. 1-Chlorooctadecane was used as a surrogate standard, and the extraction recovery was at least 75% in analyzed samples.

2.3.  Microbial community analysis

Prior to chemical analyses of bioreactors, 1.3 mL of the mixed slurry was sampled for DNA extraction with the Qiagen PowerSoil DNA isolation kit, according to the procedure provided by the manufacturer. Extracted DNA samples were submitted to the Genome Quebec Centre for next-generation sequencing using an Illumina MiSeq platform. PCR amplification and barcoding, and a quality check was performed before sequencing. The bacterial primers of 779 forward (5′- AACMGGATTAGATACCCKG -3′) and 1115 reverse (5′- AGGGTTGCGCTCGTTG -3′) were used for PCR amplification.

Analysis of sequencing reads was performed according to the procedure we have previously reported (Asadishad et al., 2018). After trimming, quality check, aligning, and chimera check of samples from days 0, 7 and 80, a total of 3138704 reads (about 90,000 reads on average for 35 samples) were obtained. OTU clustering was done at 97% similarity. Silva taxonomy reference was used for phylogenetic classification.

2.4.  Isolation of bacteria and pure culture experiments

Slurry microcosms with 7 g contaminated soil, 75 mL BH media and 750 µL sterile Bakken crude oil in 125 mL stoppered Erlenmeyer flasks were set up. The Bakken crude oil was provided by Valero Energy Inc. The microcosms were incubated at 25°C and mixed at 175 rpm for four days. Soil slurry samples were taken from the flasks which contained some oil droplets, and inoculated on R2A agar plates, and repeated culturing was used until distinct colonies were isolated on each plate. To identify the isolates, liquid cultures of the colonies were grown in Luria Bertani (LB) Lennox broth, centrifuged, and the DNA was extracted using the Zymo Research ZR Fungal/Bacterial DNA MiniPrep isolation kit. The PCR amplification was performed with 16S rRNA universal bacterial primers of 338 forward (5’-ACTCCTACGGGAGGCAGC-3’) and 1390 reverse (5’-GACGGGCGGTGTGTACAA-3’). The amplicons were submitted to the Genome Quebec Centre for Sanger Sequencing, and the 16S rRNA sequences obtained were compared to the online NCBI databases. To further verify if the isolates are able to use rhamnolipids as a carbon source, pure cultures of isolates were prepared with BH media and 600 mg/L (30 CMC) of sterilized rhamnolipid as sole carbon source. Cultures were incubated at 25°C with mixing at 175 rpm and monitored for growth.

3. Results and Discussion
   1. Biosurfactant addition did not enhance hydrocarbon biodegradation

The TPH time profiles for different bioreactors are presented in Fig. 1. A significant reduction in TPH (21-40%) was observed after 80 days in all microbially active systems. No reduction in killed control systems occurred over 80 days indicating that TPH removal in the treatment and control (except killed control) bioreactors were solely due to biodegradation, and no volatilization or other abiotic losses occurred. At day 80, the TPH of killed control systems was statistically significantly different than BH, BH+1CMC, BH+4CMC, and BH+30CMC (one-way ANOVA, Tukey HSD post-test, p value< 0.05), but was not significantly different than unamended control and BH+100CMC systems.

The systems with nutrient (BH) added but without the rhamnolipid biosurfactant showed the fastest decrease in TPH. The pseudo-first order degradation rate constant in these systems was 0.28 day^-1 as calculated using the method described in our prior studies (Akbari & Ghoshal, 2014). The calculated rate constant is comparable to our previous study of TPH biodegradation in soil slurries of a experiments with a hydrocarbon-contaminated clayey soil (Akbari & Ghoshal, 2014).

Generally, lower TPH reduction occurred in bioreactors with higher concentrations of biosurfactant added. For example, at day 14, the TPH was approximately 2500 mg/kg in the bioreactors with 30 CMC and 100 CMC and was higher than the other systems with nutrient (approximately 2000 and 2300 mg/kg). Similarly, at day 80, the TPH of the 100 CMC systems was higher than the rest of the nutrient-amended systems.

The limited efficacy of rhamnolipds on biodegradation is not surprising in the context of the literature on the effects of surfactants on PHC biodegradation. Among 218 microbial mixed cultures, the addition of rhamnolipids was found to have equally increased, decreased or had no effect on the diesel biodegradation extent (Owsianiak et al., 2009).  Zhang and Miller (1994, 1995) observed that the addition of rhamnolipids stimulated the uptake and biodegradation of octadecane for some species; but for other species inhibited octadecane biodegradation (Zhang & Miller, 1994; Zhang & Miller, 1995).  Inhibitory effects of higher concentrations of other non-ionic, anionic or biosurfactants on biodegradation activity in pure cultures have also been reported. For example, a decrease in the biodegradation of phenanthrene by a Mycobacterium genus in the presence of low doses (< 40 mg/L, ~ 0.1 CMC) of the anionic surfactant LAS (linear alkykane sulfonate) was observed. At higher doses, biodegradation and biomass growth completely stopped likely due to the toxicity of surfactant.(Jin et al., 2007). Furthermore, those authors found that the toxicity of LAS was less than cationic surfactants but greater than with nonionic surfactants. Signelton et al. reported no effect of Brij 30 on the removal of two- or three-ring PAHs in a contaminated soil, but enhanced removal of four- and five-ring PAHs. The enhancing effect was only at the lower dose of ~600 mg/L surfactant, whereas at a higher dose (~1800 mg/L) no significant effect was observed (Singleton et al., 2016). Lladó et al. reported that application of 4.5% Brij30 (w/w) generally had no effect on biodegradation of 5-ring PAHs but negatively impacted the removal of 4-ring PAHs and TPH (Lladó et al., 2015).

It could be perceived that the effect of surfactant addition in aged, weathered contaminated soils with limited-bioavailability of hydrocarbons, varies with the chemical profile of contamination and the dose and chemical structure of the surfactant. Furhter, the negative effect of higher doses of surfactants compared to lower doses as observed in soil studies suggests that the changes in microbial viability and/or communities compoistion likely contribute to the observed degradation profiles.

3.2.  The microbial community changed upon adding biosurfactant

The composition of microbial communities from different systems at the phyla level (classes of Proteobacteria) is presented in Fig. 2. Ten phyla had relative abundance of more than 1% in at least one of the samples analyzed. However, five major groups, namely, Actinobacteria, Bacteroidetes, Alphaproteobacteria, Betaproteobacteria, and Gammaproteobacteria constituted 80-96% of the total community in various samples. Statistical analysis showed that the three independent predictor variables (nutrient amendment, biosurfactant dose and time) significantly affected the total microbial community composition (multivariate analysis of variance, Pillai test, P < 10^-7 (Hand & Taylor, 1987)), and the relative abundance of Actinobacteria, Bacteroidetes, Alphaproteobacteria, Betaproteobacteria, and Gammaproteobacteria (P values: 0.03-0.09). Most significantly, the relative abundance of Alphaproteobacteria was particularly influenced by the rhamnolipid biosurfactant concentration (P values for rhamnolipid, nutrient, and time were 10^-10, 0.03, and 0.02, respectively). Higher biodegradation in microcosms by day 80 was associated with the higher relative abundance of Betaproteobacteria (Pearson correlation, r=0.69, FDR adjusted p value= 0.04), whereas in systems with lower biodegradation, higher relative abundance of Alphaproteobacteria were observed (Pearson correlation, r=-0.75, FDR adjusted p value= 0.02).

Non-metric multidimensional scaling (NMDS) plot of analyzed samples based on the OTU-level community compositions is presented in Fig. 3. In this three-dimensional graph, the closer the data points are, the more similar microbial communities’ structure. We presented the data in three dimensions because the two-dimensional analysis was not statistically representative based on the stress value. The graph shows a shift in the community at day 7 d in the bioreactors amended with various doses of rhamnolipid. In particular, the systems with the highest concentrations of surfactant at 30 and 100 CMC at day 7 were tightly clustered together, suggesting a similar community composition and shift from the BH only, or BH and lower doses of rhamnolipid (1 CM, 4CMC) systems. At day 80, the communities from BH+30CMC and BH+100CMC were relatively more distant from each other compared to day 7 (Figure 3). However, the shift in community with increasing surfactant concentration was still apparent at day 80 samples. Replicates of the samples from the same day-treatments bioreactors were relatively more clustered together, with the exception of un-amended, day 80 samples. It is noteworthy that one of the unamneded samples at day 80 which appears close to BH+100CMC at day 7  in fact was located in different planes, and the apparent proximity is only because of the projection angle of the graph. 

The most abundant OTUs and their correlations with the NMDS axes are presented in Fig. 3. The OTUs are labeled with the corresponding family. Given the short length of MiSeq sequencing reads, the phylogeny of OTUs could be reliably determined only at the family level. The most abundant OTU was affiliated with Sphingomonadaceae (Alphaproteobacteria), and as shown in the graph, its relative abundance was correlated with the rhamnolipid biosurfactant dose (indicated by biosurfactant arrow). Similarly, the relative abundance of Caulobacteraceae from the same class of Alphaproteobacteria was also correlated with the rhamnolipid dose. Moreover, as shown in the NMDS plot, OTUs affiliated with Beta- and Gammaproteobacteria (Chromatiaceae, Moraxellaceae, Comamonadaceae) were correlated with nutrient and were enriched in the presence of nutrients. Fig. 2 and 3 clearly illustrate that biosurfactant concentration impacted the microbial community composition at both phyla and OTU levels. Shifts in the microbial community with the addition of the Brij30 has been previously reported in two independent studies (Lladó et al., 2015; Singleton et al., 2016). The Pearson correlation analysis with NMDS axes (arrows in the NMDS plot) based on OTU level information confirmed that the biosurfactant concentration and the dominance of Sphingomonadaceae (Alphaproteobacteria) were important independent and dependent factors influencing the dynamics of microbial communities in the systems.

The inverse Shannon diversity of communities is presented in Fig. 4. A significant decrease (one-way ANOVA, Tukey post-test, P< 0.05) in diversity in slurries at day 7 compared to day 0 was observed. Because there was a decrease in diversity in the unamended bioreactor (only distilled water added) as well as with CMC and nutrient amended systems, it is likely that the diversity was impacted by the creation of a well-mixed soil slurry. The decrease in diversity of the slurry systems compared to the initial unsaturated soil was reported in our previous study (Asadishad et al., 2018). The significantly higher diversity of unsaturated soil compared to other environments is attributed to the existence of isolated niches in soil aggregates, which support greater diversity and reduces the competition in microbial communities. It is likely that only those taxa that could well adapt to new saturated, yet aerobic conditions and are competent in utilizing nutrients can thrive in mixed slurries where such isolated niches are destroyed. A greater decrease in diversity was observed in 30 CMC and 100 CMC systems compared to the rest of the systems at day 7, suggesting that higher concentrations of rhamnolipid also influenced microbial diversity.  The reduction in diversity in the presence of the high concentrations of surfactants is likely an indicator of selective enrichment. Signelton et al. also reported a very similar pattern of changes in the inverse Simpson indices with a decrease from 21.3±0.4 in systems without surfactant to 10.2±0.8 in systems with high doses of Brij30 amended (Singleton et al., 2016). In this study, significantly higher diversity in BH systems was observed in day 80 compared to the rest of the systems. One explanation could be that these systems have had the fastest reduction of TPH levels over 80 days, and the increased diversity reflects the removal of the contaminants. The BH systems reached their final residual level in the first two weeks, and the TPH level remained unchanged thereafter. If the residual contamination was not bioavailable, then the microbial community had enough time to recover by day 80. 

3.3. Isolates grew on biosurfactant as the sole carbon source

Several hydrocarbon-degrading bacteria from the soil were isolated. The growth response of isolates in pure cultures to the addition of 30 CMC rhamnolipid was compared to the changes in relative abundances of corresponding taxa in the microbial communities of various slurry bioreactors. As summarized in Table 1, seven distinct isolates were identified as members of five genera of Sphingomonas (Sphingomonadaceae), Pseudomonas (Pseudomonadaceae), Gordonia (Gordoniaceae), Bacillus (Bacillaceae) and Nocardia (Nocardiaceae). The full-length 16S rRNA sequences of the isolates matched to more than one strain in the database. Therefore, the strains could reliably be classified only at the genus level. All the identified genera include well-known hydrocarbon-degrading members. All isolates except Isolate 6 were able to grow in the presence of rhamnolipid as the sole carbon source.

Isolate 6 was affiliated with Bacillus genus. Consistent to pure culture experiments, the relative abundance of Bacillus in microbial communities of slurry bioreactors at day 80 was significantly lower in BH and surfactant amended systems than BH only systems (0.42±0.03% compared to 0.01±0.01%). The antimicrobial properties of rhamnolipid against Bacillus have been reported (Benincasa et al., 2004), although the mechanisms are not well-characterized. Some studies have suggested that an increase in the permeability of the cell membrane in the presence of rhamnolipid or its interactions with phospholipids are responsible for the observed antimicrobial activity (Magalhães & Nitschke, 2013). The relative abundance of Bacillus was less than 0.7% in all of the analyzed samples. Therefore, the vulnerability of this genus to rhamnolipid will have limited influence the structure of the whole community or on the biodegradation performance.

Among hydrocarbon-degrading isolates which grew on rhamnolipid as carbon source, Sphingomonas (Sphingomonadaceae) and Pseudomonas (Pseudomonadaceae) were abundant in microbial communities and had relative abundances of more than 1% on average in analyzed samples from various bioreactors. The relative abundances of these two families in samples of day 7 are presented in Fig. 5. The relative abundance of Sphingomonadaceae increased 3-5 times in the systems with the highest biosurfactant concentrations. Members of Sphingomonadaceae possess a wide range of metabolic activities and have been frequently detected in PAH-contaminated environments and are among the major PAH-degrading bacteria. For example, Sphingomonas was previously shown to be involved in biodegradation of high molecular weight PAHs with stable isotope probing and 16S rRNA gene sequencing (Jones et al., 2011). The Bakken oil added during enrichment culturing for isolation also consisted of high concentrations of PAHs (Yang et al., 2017). However, in our study with soil slurry systems with weathered contamination, this family was negatively correlated with TPH biodegradation (Pearson correlation, r=-0.68, FDR adjusted p value= 0.06).

The relative abundance of Pseudomonadaceae (Pseudomonas) was positively influenced by nutrient amendment. Nocardiaceae, however, was negatively influenced by the addition of nutrients (0.18±0.10% compared to 0.08±0.08% in systems without and with nutrients at day 7, respectively). Pseudomonas is known as an r-strategist (favoring high rates of reproduction and thrive under nutrient-rich conditions) whereas  Nocardiaceae is closely related to Corynebacteria  which are known as k-strategist bacteria (utilizing nutrients efficiently in limited resources environments).

In this study, biosurfactant addition, depending on the applied dose had no effect or an inhibitory effect on the biodegradation rate. The inhibitory effect was associated with a shift in the microbial community and reduced diversity. We were able to isolate dominant hydrocarbon degrading taxa (which were viable under our culturing conditions) and found that most of the isolates were able to grow in the presence of biosurfactant at the 30 CMC dose. This suggests that rhamnolipid toxicity is not relevant here and rhamnolipid likely served as a competing carbon source for hydrocarbon-degrading cells. Others have suggested that biosurfactant addition may hinder microbial activity because of the toxicity of solubilized contaminants (Shin et al., 2005). This is unlikely as an explanation for our observations due to low TPH concentration, as TPH content of the studied soils were generally less than  3000 mg-C/kg soil. Further, biodegradation at BH+30CMC was initially delayed but ultimately reached a similar endpoint as of the systems without surfactant.

Our results strongly suggest that biosurfactant was used as a carbon source by the indigenous microbial community. In BH+30CMC and BH+100CMC bioreactors, 15 mg-C of TPH and 12 or 41 mg-C of rhamnolipid were present initially. For a simplistic analysis and assuming a rate of 1 mg-C/bioreactor/day metabolism of rhamnolipid, and that rhamnolipid was the preferred substrate, TPH degradation in 30 CMC and 100 CMC could commence at day 12 and 41, respectively, which is qualitatively in agreement with TPH degradation pattern in Fig. 1. Also, taxa affiliated with major hydrocarbon-degrading isolates were enriched in the microbial communities of the systems with high doses of biosurfactant. Most isolates were able to use rhamnolipid as the sole carbon source in pure cultures. These suggest that similar or closely related taxa were involved in the uptake of biosurfactant and TPH degradation.

Soil contamination poses a serious risk to the environment, and in particular, groundwater resources (Akbari et al., 2012; Shayegan et al., 2015), and therefore, remediation actions are necessary. To this end, bioremediation and biostimulation of indigenous microbial communities along with the addition of biosurfactants to enhance the bioavailability of the less soluble fraction of hydrocarbons are widely used. The results of this study showed such remedial operations may not be universally effective. 

4.  Conclusion

Biosurfatcat addition influenced the biodegradation pattern by altering the microbial community composition and its metabolic activities. In particular, higher doses of biosurfactant caused a selective enrichment of Alphaproteobacteria affiliated taxa such as Sphingomonadaceae and a decrease in diversity. The metabolic activity of major hydrocarbon-degrading isolated from the same soil in pure cultures, dynamics of microbial communities in response to surfactant application, and observed biodegradation profiles suggested that rhamnolipid was likely used as the preferred carbon source. This metabolic behaviour at higher doses of biosurfactants can inhibit the biodegradation. Therefore microbial implications of biosurfactant applications should be taken into account in bioremediation operations.

Graphical abstract

Fig. 1

Fig. 1. TPH biodegradation in slurry bioreactors showed (i) a significant reduction in systems amended with nutrient, and (ii) similar biodegradation extents over 80 days in systems with nutrient, without or with lower doses of rhamnolipid added. BH: Bushnell Hass nutrient media, 1, 4, 30, 100 CMC indicates the dose of added rhamnolipid biosurfactant. Each data point is the average of three samples from three different sacrificial bioreactors. Error bars represent the standard deviation of the mean.

Fig. 2

Fig. 2. Overall phylogenetic diversity of bacterial communities in days 0, 7, and 80 in samples from slurry bioreactors. 16S rRNA gene sequences showed a shift in the community, in particular in the presence of higher doses of biosurfactant. BH refers to the nutrient amendment with Bushnell Hass media. 1, 4, 30, 100 C indicates 1, 4, 30, 100 CMC concentrations of rhamnolipid. d0, d7, and d80 refer to day 0, day 7, and day 80 samples. Phyla and Proteobacteria classes with >1% abundance in all samples are presented, and the rest are grouped as “other”.

Fig. 3

Figure 3. Nonmetric multidimensional scaling plot of the bacterial communities at days 0, 7 and 80 in different samples at OTU level (clustered at 97% similarity) and based on 16S rRNA reads (stress = 0.19, r² = 0.87) showed a shift in the community with different treatments and biosurfactant dose, in particular. Arrows show the effect of independent factors on the structure of bacterial communities based on Pearson correlation. Similarly, the correlation of most abundant OTUs with NMDS axes are shown, and OTUs are labeled with the corresponding classified family.

Fig. 4

Figure 4. OTU level diversity index (inverse Simpson) values for contaminated soil samples from slurry bioreactors shows the significant effects of higher doses of biosurfactants on day 7. BH: Bushnell Hass nutrient media and 1, 4, 30, 100 C refers to 1, 4, 30, 100 CMC doses of biosurfactant. d0, d7, and d80 refer to day 0, day 7, and day 80 samples.

Fig. 5

Figure 5. The relative abundance of the corresponding families of isolates from the contaminated soil (families with >1 % relative abundance) in microbial communities of slurry bioreactors at day 7, shows the significant enrichment of Sphingomonadaceae with increasing rhamnolipid concentration. BH, and 1C, 4C, 30C, 100C refer to Bushnell Haas nutrient media and rhamnolipid at 1, 4, 30, 100 CMC concentrations.

Table 1. Concentration of various elements and ions in background soil

pH	7.89±0.01	Total phosphorous (mg/ kg)	400
Iron (mg/g wet soil)	21.54 ± 4.11	Inorganic phosphorous (mg/ kg)	330
Calcium (mg/g wet soil)	5.72 ± 1.21	Total Kjeldahl nitrogen (mg/ kg)	760
Magnesium (mg/g wet soil)	6.49 ± 0.13	Nitrogen-NH–3 (mg/ kg)	ND
total organic carbon (g/g soil)	0.017	Nitrogen-NO2 & NO3 (mg/ kg)	5.9

Isolate	Closest relative by 16S rRNA gene (GenBank accession; % identity)	Appearance	affiliated family/genus	Growth on 30 CMC Rhamnolipid
Isolate_1	Pseudomonas mandelii strain SX1216 (MG576049 ,99%)	Small white	Pseudomonadaceae; Pseudomonas	+
Isolate_2	Gordonia amicalis strain NR_59 (KM113028.1, 98%)	Pink	Gordoniaceae; Gordonia	+
Isolate_6	Bacillus megaterium (HG975598, 99%)	Red	Bacillaceae; Bacillus	–
Isolate_7	Bacillus thuringiensis strain Y4-36 (GU143905, 97%)	White	Bacillaceae; Bacillus	+
Isolate_8	Sphingomonas sp. S1(2013) (KF544916, 99%)	Yellow	Sphingomonadaceae; Sphingomonas	+
Isolate_9	Pseudomonas silesiensis strain A3, (CP014870.1, 99%)	White smooth	Pseudomonadaceae; Pseudomonas	+
Isolate_10	Nocardia asteroides strain IIL-Asp25 (KX380900.1, 99%)	Filamentous	Nocardiaceae; Nocardia	+

Table 2. Identification of hydrocarbon-degrading isolates and their growth in the presence of rhamnolipid as the sole carbon source