Isolation and identification of novel, native geosmin and 2-methylisoborneol degrading bacteria

Several studies have recorded that biological processes could be a viable treatment option for removing T&O compounds (Saito et al. 1999; Ho et al. 2007) and cyanotoxins (Lawton et al. 2011), as an eco-friendly green solution. In general, biological processes are of low cost, require little maintenance, and do not rely on the addition of chemicals, which often results in unwanted by-products (McDowall et al. 2009; Hsieh et al. 2010; Tian 2013). Previous studies have described the genus Arthrobacter (Saadoun & El- Migdadi 1998), genus Bacillus (Lauderdale et al. 2004), and genus Rhodococcus (Xue et al. 2012) as potential geosmin and 2-MIB degraders. However, studies on batch culture biodegradation of geosmin and 2-MIB are limited. The present study used enrichment to isolate bacteria from seventeen Sri Lankan water bodies where T&O issues were prevalent. The Biolog MT2 assay was employed to screen the isolated bacteria to metabolize geosmin and 2-MIB, since this assay had previously been shown to be an effective method of demonstrating the metabolism of microcystin variants (Manage et al. 2009).

Surface water and lake sediment samples were collected in sterile Pyrex glass bottles and black polypropylene bags from June 2016 to June 2019 across 17 water bodies covering seven provinces in Sri Lanka (Table 1). Sampling was carried out in both dry and wet seasons. Surface water samples were collected from the upper 10–20 cm of the water column using a pre-sterilized glass container, and sediment samples were collected from the top 5–10 cm of the lake sediment using a sterilized core sampler. The collected samples were immediately placed in sterile containers; surface water in the glass bottles and sediment in sterile black polypropylene bags to prevent contamination. Samples were stored at 4 °C overnight, and a metal sieve with 150 μm mesh was used to remove zooplankton and vegetation, as previously described for degradation studies (Edwards et al. 2008). Both geosmin and 2-MIB standards were purchased from Sigma Aldrich, USA, and aliquots from each water sample (10 mL) were processed to identify and quantify the presence of naturally occurring geosmin and 2-MIB by high-performance gas chromatography/mass spectrometry (GC/MS) coupled with solid phase micro extraction (SPME) (Ganegoda et al. 2020).

To enrich bacteria capable of degrading geosmin and 2-MIB, both compounds were added to each 100 mL water sample to reach a final concentration of 20 ng/L for each compound. Similarly, composite 5 g of sediment was added to 100 mL of filter-sterilized reservoir water, which geosmin, and 2-MIB were spiked at a final concentration of 20 ng/L (ppt). All flasks were incubated at 28 °C ± 1 °C with shaking at 34 × g for 14 days. After 14 days of enrichment, 1 mL of sample was removed aseptically from each flask, and serial 10-fold dilutions (to 10⁻⁵) were made using 0.01 M phosphate buffered saline (PBS) solution, and 1 mL of sub-sample from each dilution was removed, mixed with 20–25 mL of molten Luria-Bertani (LB) agar, poured onto sterile petri dishes, and incubated in the dark at 25 °C ± 1 °C for 3 days (Manage et al. 2009). Colonies with differing morphological features were picked and re-suspended in liquid LB medium, and pure cultures were obtained by repeated streaking onto LB agar plates. For the Biolog MT2 assay, a loop of each bacterial isolate was transferred into 5 mL of liquid LB medium and incubated overnight in the dark at 25 °C ± 1 °C. The cultures in the exponential growth phase were then washed twice by centrifugation at 1,000 × g for 15 min, the resultant bacterial pellets were re-suspended in sterile 0.01 M PBS and incubated at 25 °C ± 1 °C for 24 h to deplete residual carbon. The turbidity of all bacterial suspensions was equalized at A 590 = 0.35 and used as a reference for bacterial concentration. Geosmin and 2-MIB were added to Biolog MT2 plates in triplicate to obtain final concentrations of 10, 40, and 100 ng/L. Wells were inoculated with an equalized bacterial suspension of 150 μL, and then the plates were incubated at 25 °C ± 1 °C. Absorbance at 595 nm was recorded by using an enzyme-linked immunosorbent assay (ELISA) microplate reader (MULTISKAN EX, Thermo Scientific, USA) immediately after inoculation (0 h), and at 3, 6, 15, 18, 24, and 48 h intervals. Metabolism of geosmin and 2-MIB reduces tetrazolium violet dye, giving a color reaction that can be quantified spectroscopically (Garland & Mills 1991). Bacterial isolates found using the Biolog MT screen to metabolize geosmin and 2-MIB were evaluated for their ability to degrade both odorants. Out of 73 bacterial isolates, 10 were capable of metabolizing both compounds, these 10 selected bacterial isolates were subjected to flask die-away kinetics experiments, conducted in triplicate. Selected bacteria strains were grown overnight in LB liquid medium (25 °C ± 1 °C), then washed following carbon depletion (0.5 mL) and added to universal glass bottles containing 9 mL of 0.2 μm filter-sterilized raw water from their original locations. Since geosmin and 2-MIB are semi-volatile compounds, flasks cannot be used for degradation kinetic study. Thus, 14 equal sets of samples were prepared each day in triplicate (3 × 14 samples). Both geosmin and 2-MIB standards were spiked to each bottle under aseptic conditions at a final concentration of 20 ng/L. Triplicate samples were prepared for each isolate and incubated (at 25 °C ± 1 °C with shaking at 34 × g). Three bottles were removed at 24-h intervals under sterile conditions, filtered using 0.2-μm syringe filters (Whatman) to remove bacteria and analyzed for geosmin and 2-MIB according to the methods given by Manage et al. (2009). Experiments with sterile controls were performed for each sample, treated the same way except for not being inoculated.

In the 3.2 degradation kinetics of geosmin and 2-MIB, Bacillus cereus, Bacillus subtilis, and Acinetobacter guillouisae were identified as the most potent bacterial degraders of geosmin and 2-MIB among the 10 bacterial isolates, following the degradation kinetics study. Consequently, optical density at 600 nm was measured every 3 h to construct the growth curves for these bacteria. This allowed for identifying the growth phases of these potential degraders of geosmin and 2-MIB (Manage et al. 2009).

The geosmin and 2-MIB degrading bacteria; B. cereus, B. subtilis, and A. guillousae were selected for further degradation studies. The bacterial strains were grown in 5 mL liquid LB medium as described previously for 24 h to achieve exponential growth, following a wash with an equal volume of 0.01 M PBS by centrifugation at 1,000 × g for 15 min. The resulting supernatant was discarded, and the pellet was suspended in sterile 0.01 M PBS and kept overnight to exhaust residual carbon content, if any. Then, the suspension was centrifuged at 334 × g for 5 min, and the pellet was washed three times using PBS by centrifugation (334 × g). After that, the optical density was equalized at A 590 = 0.35. A total of 1 mL of each equalized bacterial suspension was inoculated into 100 mL filter-sterilized (0.2 μm) lake water containing geosmin and 2-MIB at a final concentration of 20 ng L⁻¹. Ten of these experimental setups were prepared for each bacterium in triplicates and incubated at 28 °C following a shake at 34 × g. Then, 10 mL subsamples of aliquots were removed at every 6-h interval. Once the 10 mL aliquot was removed, geosmin and 2-MIB concentration in the sample were measured. Controls were maintained for each experiment setup.

The bacterial degradation rate (h) of geosmin and 2-MIB was calculated using the equation; h = ln(C/C₀)/(t₂ − t₁), where C₀ and C are the concentrations of odorants at the beginning and at the end of the time interval t, respectively, (Manage et al. 2000).

Since the degradation of both geosmin and 2-MIB was not constant, different degradation rates were calculated for lag, exponential, and stationary growth phases.

To extract genomic DNA and to perform gene-specific polymerase chain reaction (PCR), pure cultures of selected bacterial isolates were sent to Genetech, Sri Lanka, and the molecular identification of bacteria by performing DNA sequencing of the resultant PCR products was conducted in Macrogen, Korea. The quality of the sequence was checked by using the sequence analysis software (ABI), and the products of the universal forward and reverse primers (785 F and 907 R) (contigs) were aligned using Kodon (Applied Maths, Saint-Martens-Latem, Belgium). The method described by Barghouthi (2011) was employed with some modifications. The analyzed sequences were compared to DNA sequences in public databases using the BLASTn function of NCBI (http://www.ncbi.nlm.nih.gov). Individual isolates were classified according to the identities of the sequenced loci, which were revealed through BLASTn. A > 97% level of sequence identity was specified as a species, while >94.5% sequence identity was demarcated as a genus-level match. DNA sequences of all isolates, along with related bacteria and some known geosmin and 2-MIB degrading bacteria, were used to construct a maximum likelihood (ML) phylogenetic tree (Figure 7) using MEGA 4 (Ventura et al. 2007; Hoefel et al. 2009; Xue et al. 2012).

Furthermore, the 10 selected bacterial isolates named B1: Myroides xuanwuensis, B2: Providencia vermicola, B3: Providencia rettgeri, B4: Myroides odoratitimus, B5: Proteus mirabilis, B6: Bacillus cereus, B7: Bacillus subtilis, B8: Acinetobacter guillouisae, B9: Acinetobacter indicus, and B10: Pseudomonas stutzeri demonstrated significant metabolism across three different concentrations of 2-MIB (10, 40, and 100 ng/L). However, the bacterial responses to 2-MIB varied compared to geosmin (Figure 2(a) and 2(b)). B. cereus (B6) and A. guillouisae (B8) demonstrated rapid utilization of the low concentration of 2-MIB (10 ng/L). The other isolates followed a similar pattern in terms of their degradation of geosmin. B. subtilis (B7) displayed the highest utilization across all tested concentrations of 2-MIB, with an absorption of over 3.4. In contrast, P. stutzeri (B10) and P. vermicola (B2) exhibited moderate metabolism, with absorption levels above 2.0. P. mirabilis (B5) and A. indicus (B9) showed no significant variation in utilizing all three concentration gradients of 2-MIB compared to the other isolates. However, the response of P. rettgeri (B3) to 100 ng/L of 2-MIB was notably stronger.

Notably, B. subtilis (B7) demonstrated the ability to metabolize both geosmin and 2-MIB at a faster approach than the other isolates (Figure 2(a) and 2(b)). These characteristics suggest that these bacteria could be valuable in future experiments to explore further potential for degrading geosmin and 2-MIB in contaminated water.

Geosmin and 2-MIB degrading bacteria were identified at the molecular level using 16S rRNA gene analysis and presented in Table 2.

The ‘Lag phase’ of A. guillousae remained for 6 h, similar to the other two bacteria, followed by a longer ‘Exponential phase’ (from 6 to 15 h), and then to a ‘Stationary phase’ from 15 to 24 h of incubation (Figure 5(c)).

B. cereus showed rapid degradation of geosmin with its complete removal at 24 h of incubation, whereas B. subtilis exhibited a faster degradation pattern than A. guillousae, and both species showed complete degradation of geosmin at 48 h of incubation (Figure 6).

The results showed different degradation patterns of 2-MIB by the three bacterial species. B. subtilis showed rapid and complete degradation of 2-MIB at 48 h following A. guillousae at 120 h and B. cereus at 192 h, respectively. Thus, it was found that the degradation pattern and time required for the complete degradation of geosmin and 2-MIB differed by isolated bacteria species, which may be due to the different chemical structures.

Bacterial degradation rates of geosmin and 2-MIB were calculated using the equation; h = ln(C/C₀)/(t₂ − t₁), described in the ‘Methodology’ section. The calculated degradation rates of geosmin and 2-MIB in different growth phases are given in Table 3.

Out of all three phases, the highest degradation rate of geosmin was recorded during the stationary phase in all three bacteria. B. cereus (0.1733 ng L⁻¹h⁻¹) showed the highest rate, and the other two bacteria species showed more or less similar degradation rates between 0.0822 to 0.0866 ng L⁻¹h⁻¹ (Table 2). The degradation rate of geosmin was remarkable in the exponential growth phase of bacteria. The highest degradation rate was achieved by B. subtilis (0.0821 ng L⁻¹h⁻¹) followed by B. cereus (0.0581 ng L⁻¹h⁻¹) and A. guillousae (0.0511 ng L⁻¹h⁻¹). During the lag phase, the highest degradation rate of geosmin was recorded by B. cereus as 0.0271 ng L⁻¹h⁻¹ followed by B. subtilis (0.0175 ng L⁻¹h⁻¹) and A. guillousae (0.0085 ng L⁻¹ h⁻¹) in descending order.

2-MIB degradation rates of the three bacteria species were lower than the recorded geosmin degradation rates (Table 2). The lowest degradation rate was found in the lag phase, and the highest was synonymous with the stationary phase for all three bacteria species. Amongst the three species, B. subtilis showed the maximum degradation rate (0.0529 ng L⁻¹h⁻¹) during the stationary phase. All the other bacteria showed a similar degradation rate of 2-MIB during the exponential phase (Table 3).

The nucleotide sequences for the 16S rRNA genes have been deposited in the NCBI database under accession numbers MK606113, MK601700, MK968362, MK601701, MK601699, MK968363, MK982381, MK968347, MK972672, and MK968348 accordingly.

The evolutionary relationships of geosmin and 2-MIB degrading bacteria are illustrated in Figure 7.

Geosmin and 2-MIB are two challenging compounds to oxidize using conventional water purification methods due to their structural stability (Cook & Newcombe 2004; Hung & Lin 2006). Although microbial treatment methods used by the water industry are few (Persson et al. 2007), microbial removal offers a highly effective means to treat geosmin and 2-MIB in source water (Juttner & Watson 2007). Several studies have implicated a variety of microorganisms involved in the removal of geosmin and 2-MIB from water (Saadoun & El-Migdadi 1998; Ho et al. 2007; McDowall et al. 2009). The present study agrees with these authors by demonstrating that a greater diversity of bacterial genera is capable of degrading geosmin and 2-MIB. The susceptibility of both geosmin and 2-MIB to biodegradation can be attributed to their structural similarities to biodegradable alicyclic alcohols and ketones (Rittmann et al. 1995).

The most crucial finding in this study was the diversity of bacteria capable of utilizing varying concentrations of geosmin and 2-MIB. The results revealed that B. cereus and B. subtilis showed pronounced utilization of geosmin while, P. vermicola, B. subtilis, and P. stutzeri exhibited significant metabolism of 2-MIB. This suggests the possibility of diverse biochemical pathways in these bacteria for degrading both compounds. Interestingly, 2-MIB degradation was slower compared to geosmin degradation, likely due to the more complex structure of the 2-MIB. (Juttner & Watson 2007) (Figure 1).

Xue et al. (2012) highlighted that bacterial degradation of geosmin requires an adaptation period before the effective degradation. The results from the present study support this finding. As an example, the P. mirabilis strain was found to degrade geosmin in 5 days of incubation time. Geosmin concentration did not show a significant reduction during the first 60 h; however, in the subsequent 60 h, the concentration decreased from 13 ng/L to an undetectable level (less than 1.5 ng/L). This observation suggests that geosmin degradation may involve a lag period and/or that some enzymes responsible for degradation may not be completely activated at the beginning (Xue et al. 2012). Alternatively, this could expand our understanding of the potential role of bacterial by-products or extracellular enzymes in the degradation process. This highlights a pathway for further investigation, as it could provide valuable insights into the mechanisms underlying geosmin degradation. Future studies could focus on analyzing the metabolic activity of the bacteria over time and characterizing any extracellular enzymes or by-products that may contribute to degradation. The remaining nine bacterial strains in the study also showed a similar degradation pattern for geosmin. In contrast, this pattern was not observed with 2-MIB. For instance, P. vermicola was capable of degrading 2-MIB in 4 days. The degradation kinetics demonstrated that more than 50% (62.5%) of 2-MIB was degraded within the first 48 h, and no adaptation time was required (Figure 3).

The Biolog MT2 assay has recently been employed in various microbiological studies to screen the ability of isolated bacteria to metabolize different cyanobacterial toxins (Manage et al. 2009). While several other cyanobacterial secondary metabolites have previously been analyzed using this technique, this is the first recorded study, to analyze the metabolism of geosmin and 2-MIB using the Biolog MT2 assay.

The present study revealed that out of 120 bacterial isolates tested in the Biolog MT2 assay, a total of 73 isolates (45 for geosmin and 28 for 2-MIB) showed positive results for the degradation of geosmin and 2-MIB. Of these, 25 bacteria were selected for batch degradation kinetics, and all 25 were able to degrade geosmin and/or 2-MIB in 14 days of incubation, as evidenced by GC-MS analysis. A similar study conducted by Manage et al. (2009), utilized the Biolog MT2 assay to screen microcystin-LR (MC-LR) metabolism and found that all positive bacteria in the assay were capable of degrading MC-LR in batch studies, as evidenced by liquid chromatography-mass spectrometry analysis. Employing the Biolog MT plate assay facilitated rapid (approximately 24 h), cost-effective screening of bacteria in a high-throughput format (96-well plates), using significantly lower amounts of geosmin and 2-MIB as standards. For example, geosmin and 2-MIB were used at 5 μg per isolate in the Biolog MT plate, compared to 300 μg to follow degradation in die-away kinetics as described herein. Additionally, after degradation by the latter method, sample processing and GC analysis were necessary to identify the degradation products.

The present study found that 80% of bacterial strains showed the fastest utilization of geosmin and 2-MIB at the highest concentrations in the concentration gradient, indicating that these bacteria can effectively degrade higher concentrations of both compounds. Previously, only a limited number of bacterial species have been reported to degrade geosmin and 2-MIB, specifically three genera: Arthrobacter (Saadoun & El-Migdadi 1998), Bacillus (Lauderdale et al. 2004), and Rhodococcus (Xue et al. 2012). This is the first report to our knowledge, demonstrating a wide range of bacterial isolates capable of utilizing two chemically distinct compounds that commonly cause taste and odor issues in water. The present study demonstrated that the isolated bacteria could utilize three different concentrations of geosmin and 2-MIB (Figure 2). Given the widespread taste and odor problems worldwide (Juttner & Watson 2007; Sorial & Sirinivasan 2011; Xue et al. 2012; Tian 2013; Ganegoda et al. 2018), this approach shows promise for developing reliable methods to degrade these compounds in the future water treatment strategies. Bacteria with ‘universal’ geosmin and 2-MIB degradation capabilities have better potential in such applications than single-variant degraders.

In this study, two Gram-positive bacteria, B. cereus and B. subtilis, were isolated from freshwater sources and identified as capable of degrading both geosmin and 2-MIB. B. cereus degraded geosmin within 24 h, whereas B. subtilis required 48 h. Moreover, B. subtilis degraded 2-MIB in 48 h, while B. cereus only achieved degradation of 2-MIB after a 7-day incubation period. This observation is intriguing and may suggest distinct biochemical pathways for the degradation of geosmin and 2-MIB, despite the structural similarities between these compounds. Additionally, eight Gram-negative bacteria were identified as capable of degrading both geosmin and 2-MIB. Among them, Pseudomonas stutzeri was the most effective Gram-negative degrader of 2-MIB. Notably, P. stutzeri has been previously reported as a naphthalene-degrading bacterium, capable of breaking down various contaminants and interacting with hazardous pollutants. This suggests that Pseudomonas sp. may have the genetic capacity to degrade complex organic compounds, as reported by Lalucat et al. (2006). Further molecular studies on this are currently underway.

The top three bacterial isolates for degrading geosmin and 2-MIB were identified as B. cereus, B. subtilis, and A. guillouisae, with B. subtilis degrading both odorant compounds within 48 h. This is the first report highlighting the potential of A. guillouisae and A. indicus to degrade both geosmin and 2-MIB. The 10 bacterial strains identified in this study belong to the phyla Firmicutes (genus Bacillus), Bacteroidetes (genus Myroides), Actinobacteria (genus Acinetobacter), and Proteobacteria (genera Providencia, Proteus, and the family Pseudomonas). These phyla are well-known for their metabolic diversity and ability to degrade various chemical compounds (Ventura et al. 2007). Furthermore, members of these groups have been isolated from a wide range of environments, including freshwater, saltwater, soil, and sludge. This study reports the ability of several members of these phyla to degrade geosmin and 2-MIB.

Individual bacteria capable of degrading geosmin and 2-MIB were isolated from water and sediment samples collected from 17 different freshwater sources, including the Nallachchiya reservoir. In the Nallachchiya reservoir, no degradation of geosmin or 2-MIB by the indigenous microbial flora was observed during the enrichment and die-away experiments. This lack of degradation could be attributed to the relatively low bacterial count in the samples compared to the much higher bacterial abundance found in other reservoirs.

The intermediate products identified in the degradation of geosmin and 2-MIB include 3,6-dimethylbicyclo, 2-methyl-2-bornene, and 1-ethoxy-4-ethoxybenzene. (Figure 4). Moreover, recent researchers, Saito et al. (1999), showed that 3,6-dimethylbicyclo is a breakdown product of geosmin, and Schumann & Pendleton (1997) as well as Song & O'Shea (2007) stated that 2 methyl-2-bornene is a degradation by-product of 2-MIB. The findings of the previous studies agree with the recorded by-products in the present study, and the by-product chemicals are considered to be odorless and non-toxic (Schumann & Pendleton 1997; Saito et al. 1999). Furthermore, 1-ethoxy-4-ethylbenzene is a safe substitute for the flavoring component ‘Anethole’ (Brunke et al. 2016). Further studies on by-products are in progress. Moreover, these identified bacteria could be used on a biological filter to remove geosmin and 2-MIB from water as an eco-friendly approach, with a low cost and high environmental-friendly avenue.

B. cereus, B. subtilis, and A. guillousae were identified as the best bacterial strains to degrade geosmin and 2-MIB, with B. subtilis degrading both compounds within 48 h. This is the first report on A. guillouisae's potential to degrade both geosmin and 2-MIB. This research shows that a more comprehensive array of bacterial genera can degrade geosmin and 2-MIB, yet uncharacterized degradation mechanisms. We also found geosmin and 2-MIB degraders in an aquatic environment where the levels of geosmin and 2-MIB are incredibly low. More importantly, the current study showed that geosmin and 2-MIB could be removed from water via biodegradation, a low-cost process that would not need the addition of chemicals.

The National Science Foundation (NSF) is acknowledged for the financial support provided (Grant number RG/2016/EB04). Center for Water Quality and Algae Research, Department of Zoology, Central Instrumentation Facility, Faculty of Applied Sciences, University of Sri Jayewardenepura, National Water Supply and Drainage Board (NWSDB) are acknowledged for the instrument facilities, technical advice and other logistic support provided.

The research study was funded by the NSF, Sri Lanka (Grant number RG/2016/EB04).

Ethics: Human participants

No human participants, human studies nor potentially identifiable human images were used/presented in this study

Ethics: Animal testing

No animal testings/studies were used in this study

All relevant data are available from the online repository at https://www.ncbi.nlm.nih.gov/.

The authors declare there is no conflict.