Rhamnolipid production was monitored for a period of 216 h using different substrates in Pseudomonas aeruginosa PAO1 and Burkholderia thailandensis E264 which showed comparable crude yields attained by both after 216 h. The crude yield for P. aeruginosa, however, was significantly higher at the early stages of fermentation (72 or 144 h). Additionally, P. aeruginosa produced rhamnolipid with odd and even carbon chain lipid moieties using odd carbon chain fatty acid substrates (up to 45.97 and 67.57%, respectively). In contrast, B. thailandensis produced rhamnolipid with predominantly even carbon chain lipid moieties (up to 99.26). These results indicate the use of the fatty acid synthesis (FAS II) pathway as the main source of lipid precursors in rhamnolipid biosynthesis by B. thailandensis. Isotope tracing using 0.25% stearic acid – d 35 + 1% glycerol as carbon substrate showed a single pattern of deuterium incorporation: with predominantly less than 15 deuterium atoms incorporated into a single Di-C14-C14 rhamnolipid molecule. This further indicates that the FAS II pathway is the main source of the lipid precursor in rhamnolipid biosynthesis by B. thailandensis. The pathogenicity of these strains was also assessed, and results showed that B. thailandensis is significantly less pathogenic than P. aeruginosa with an LC50 at 24 h > 2500, approximately three logs higher than P. aeruginosa using the Galleria mellonella larva model.
The first report of rhamnolipid production was made in the mid twentieth century, with Pseudomonas aeruginosa identified as the producer organism. Since then, production of rhamnolipid from strains of P. aeruginosa has been extensively investigated to understand the structure, cellular function and mechanism of rhamnolipid production by this organism (Irorere et al. 2017). However, heightened industrial interest has led to an increased demand for rhamnolipids and concerns about the safety of the product and its industrial production processes, since P. aeruginosa is an opportunistic human pathogen. These facts have stimulated the search for alternative, safer rhamnolipid producing microorganisms.
In the last decade, a number of Burkholderia species have being shown to be capable of rhamnolipid production, including Burkholderia thailandensis E264, Burkholderia pseudomallei, Burkholderia glumae BGR1, Burkholderia plantarii and Burkholderia kururiensis (Costa et al. 2011; Dubeau et al. 2009; Tavares et al. 2013). The orthologues of the rhamnolipid genes (rhlA, rhlB and rhlC) have been identified in the first three strains (Irorere et al. 2017). Among these strains, B. thailandensis E264 is of particular interest mainly due to its significantly lower levels of pathogenicity (Brett et al. 1998; Haraga et al. 2008; Koh et al. 2012).
The production of rhamnolipid by B. thailandensis was first reported less than a decade ago. The di-rhamnolipid C14-C14 was reported as the most abundant congeners produced by this strain, compared to P. aeruginosa strains in which di-rhamnolipid C10-C10 is the most abundant rhamnolipid congener (Dubeau et al. 2009). Also, a genome wide study of B. thailandensis E264 showed that the rhamnolipid genes rhlA, rhlB and rhlC are located in a single gene cluster. This cluster is duplicated in the genome of the organism, with both clusters reported to be functioning in rhamnolipid production (Dubeau et al. 2009). This is also different to P. aeruginosa in which the rhlA and rhlB genes are contained in a single operon, while rhlC is located in a remote location. Also, these genes are not duplicated in the genome of P. aeruginosa (Irorere et al. 2017).
While the production of rhamnolipid has been established in strains of Burkholderia, little is known about the mechanism of rhamnolipid biosynthesis in these species. As earlier stated, the majority of what we know about the function and mechanism of biosynthesis of rhamnolipid are obtained from studies using P. aeruginosa strains. However, the rhamnolipid congener distribution and the genomic arrangement of the rhamnolipid genes are different in Burkholderia compared to Pseudomonas. These differences suggest that the metabolism of substrates for rhamnolipid biosynthesis might follow different patterns. To fully exploit these strains as alternatives for industrial rhamnolipid production, it is important to understand their mechanism of rhamnolipid biosynthesis and factors affecting production and composition of the rhamnolipid congeners.
This study was therefore designed to understand the differences in growth and rhamnolipid biosynthesis by B. thailandensis in different nutrient limiting media and carbon substrates. Results indicate that the fatty acid synthesis pathway is the main source of the lipid precursor in rhamnolipid biosynthesis in B. thailandensis. This differs from P. aeruginosa which uses β-oxidation intermediates as the main suppliers of lipid precursors in rhamnolipid biosynthesis (Abdel-Mawgoud et al. 2014).
Materials and methods
Microbial strains, media and shake flask fermentation
Burkholderia thailandensis E264 (ATCC 700388) and P. aeruginosa PAO1 were the strains used in this study (Brett et al. 1998; Stover et al. 2000). Cells were maintained in 50% glycerol at − 80 °C prior to use. Cultures were prepared from freezer stocks in nutrient agar (Sigma Aldrich) plates grown for 24 h at 30 °C. Single colonies were then used to make starter cultures in nutrient broth (Sigma Aldrich) grown overnight at 30 °C with shaking at 200 rpm.
The phosphate-limited media, the proteose-peptone/glucose acid salt media (PPGAS) (Caiazza et al. 2005; Mulligan et al. 1989) and a mineral salt media (MSM) (Moya Ramírez et al. 2015) were used as fermentation media for the production of rhamnolipid. PPGAS comprises Tris-HCl 15 g/l, MgSO4·7H2O 0.4 g/l, NaCl 1.4 g/l, NH4Cl 1 g/l, proteose peptone 10 g/l and carbon source. While MSM comprises NaNO3 2 g/l, Na2HPO4 0.9 g/l, KH2PO4 0.7 g/l, MgSO4·7H2O 0.4 g/l, CaCl2·2H2O 0.1 g/l, trace element solution 0.1% (v/v) and carbon source. Trace element solution comprises ZnSO4·7H2O 0.7 g/l, CuSO4·5H2O 0.5 g/l, MnSO4·H2O 0.5 g/l, H3BO3 0.26 g/l, MoNa2O4·2H2O 0.06 g/l and FeSO4·7H2O 0.001 g/l. To prevent precipitation during autoclaving, initial preparation of MSM included the carbon source, NaNO3, Na2HPO4 and KH2PO4. Stock solutions of the other components including the trace elements were prepared, filter sterilised and calculated volumes were added to the autoclaved media. The pH of PPGAS and MSM were adjusted to seven before autoclaving. The addition of the other media components to MSM after autoclaving did not have any significant effect on the final pH of the medium.
Initial shake flask experiments were set up using both PPGAS and MSM prepared as described above with 2% of either oleic acid—C18 (BDH chemicals), heptadecanoic acid—C17 (Sigma Aldrich, UK) or glycerol (BDH chemical) as carbon source. As heptadecanoic acid and stearic acid are solid powders, after autoclaving but before inoculation, the media were kept in a water bath (Davidson and Harvey Ltd., UK) at 60 °C with shaking at 50 rpm until the white lump formed following autoclaving was dispersed within the medium. The media were then allowed to cool to room temperature before inoculation.
Ten millilitres of overnight starter culture were used to inoculate flasks containing 90 ml fermentation broth to give a 10% inoculum concentration. A sample was taken immediately to determine the initial microbial concentration. The flasks were then incubated at 30 °C for 216 h and samples taken regularly to determine cell growth. All shake flask experiments were performed in triplicate.
Shake flasks experiments were also set up for deuterium tracing studies using MSM as the rhamnolipid production medium supplemented with either 1% glycerol (BDH chemicals) and 0.25% stearic acid (Sigma Aldrich) or 1% glycerol and 0.25% stearic acid – d 35 (Sigma Aldrich). Stearic acid – d 35 was used in the deuterium tracing studies, as it is readily available compared to oleic acid – d 34 , while MSM was used due to higher crude rhamnolipid yield compared to PPGAS.
Determination of microbial growth and biomass concentration
One millilitre of fermentation broth was taken at intervals for determination of viable cell count as described by Miles et al. (1938). Samples were serially diluted, and 20 μl of the diluted samples were inoculated into nutrient agar and incubated for 24 h at 30 °C. Viable cell count was calculated in log10 cfu/ml. Analyses were carried out in triplicate for each flask.
To determine biomass concentration, 1 ml of broth was transferred to a preweighed Eppendorf tube and centrifuged for 15 min at 13,000×g. The supernatant was discarded, and the pellet washed with PBS, re-centrifuged, and the PBS discarded. Pellets were dried at 80 °C.
Rhamnolipid extraction and purification
Rhamnolipid extraction was carried out using an acid precipitation and solvent extraction method (Smyth et al. 2014). First, cultures were centrifuged at 13,000×g for 15 min. Supernatants were collected and the pH adjusted to 2.0 using 1 M HCl. An equal volume of ethyl acetate (Sigma Aldrich) was then used to extract rhamnolipid from the supernatant three times with the aqueous phase discarded after each extraction. Anhydrous MgSO4 (Sigma Aldrich) was added to the organic phase containing the rhamnolipid at a concentration of 0.01 g/ml to remove any residual aqueous phase left after extraction. The ethyl acetate (organic phase) containing the extracted rhamnolipid was filtered using a grade 1 filter paper (Whatman® qualitative filter paper, grade 1) into a round bottom flask and dried under vacuum in a rotary evaporator (Buchi, Flawil, Switzerland) leaving the thick rhamnolipid in the flask. This was then re-dissolved in a small volume of chloroform/methanol (1:1) and transferred into preweighed glass scintillation vials. The samples within the vials were dried using nitrogen gas and the weight of the vials measured until a constant weight was reached.
To remove unwanted impurities and excess carbon source, solid-phase extraction (SPE) of the crude extracts was carried out. This was done using Strata SI-1 Silica (55 μm, 70 A) Giga tubes (Phenomenex). The tubes were conditioned using HPLC-grade chloroform, and the rhamnolipid samples were dissolved in a small amount of chloroform and introduced into the conditioned column. The column was eluted with chloroform until the eluent from the column became clear, indicative of the removal of residual fatty acids and other contaminants. A chloroform/methanol solution in the ratio of 5:0.3 and 5:0.5 was then used to elute the mono-rhamnolipid followed by a 1:1 solution of chloroform/methanol to elute the di-rhamnolipids. Both mono- and di-rhamnolipid were collected together in a preweighed scintillation vial and dried under nitrogen gas.
Surface tension measurements and determination of the critical micelle concentration
Surface tension was measured using a KRUSS KI0ST Tensiometer with a platinum ring. The surface tension values of a series of rhamnolipid concentrations in water were measured to estimate the critical micelle concentration (CMC). The pH of rhamnolipid samples from B. thailandensis E264 was adjusted to seven, to allow the rhamnolipid to completely dissolve in water. Surface tension of each concentration was taken in triplicate and a graph of surface tension against the corresponding concentration was plotted. CMC was determined by extrapolating the intercept of the two sections of the graph.
Rhamnolipid analysis by LC-QToF-MS
SPE-purified rhamnolipid extracts were characterised by liquid chromatography-hybrid quadrupole time-of-flight mass spectrometry (LC-QToF-MS) using methods described previously (Funston et al. 2016; Smyth et al. 2010b). The Agilent Poroshell SB-C3, 2.1 × 100 mm, particle size 2.7 μm column was used as the static phase to analyse rhamnolipid from B. thailandensis E264 while the Poroshell 120, EC-C18, 2.1 × 100 mm, 2.7 μm was used as static phase to analyse rhamnolipids from P. aeruginosa. Mobile phase 1 consisted of HPLC grade water (4 mM ammonium acetate) while mobile phase 2 was acetonitrile.
Pathogenicity assay using Galleria mellonella
For Galleria mellonella analysis, single colonies of respective organisms were incubated overnight at 30 °C and 200 rpm in nutrient broth (NB). Fifty millilitres of the overnight cultures were centrifuged, and cell pellets were washed in PBS, centrifuged and re-suspended in PBS. Cells were diluted to OD600 of 0.46 (7.5 × 108 cfu/ml) and 0.28 (1.35 × 109 cfu/ml) for B. thailandensis E264 and P. aeruginosa PAO1, respectively. These were further diluted by serial dilution to make working stocks for each individual organism. The cell concentrations of the working stocks were confirmed by plate count. Both the initial stocks and the working stocks were prepared daily and used immediately after preparation.
Rearing and infection of G. mellonella
Galleria mellonella were purchased from Pets at Home (Coleraine, Northern Ireland, UK) and maintained in the dark on woodchips at 15 °C prior to use. Bacteria suspensions were diluted to known concentration in PBS, and 20-μl samples (100 cfu) were injected into G. mellonella via the foremost proleg. Ten larvae were injected for each strain of bacterium, and larvae were incubated for 48 h at 37 °C in perforated Petri dishes to provide adequate ventilation. Larvae injected with 20 μl of sterile PBS were used as a control, and the numbers of dead larvae were scored periodically. Larvae were considered dead when they turn grey/black and did not move in response to a gentle physical stimulation.
To determine if the death of G. mellonella is due to the pathogenic effect of the infecting cells or toxicity of the cell components, larvae were also injected with heat-killed cells at similar concentrations (i.e. 100 cfu in 20-μl samples). Cells were heat killed at 80 °C for 2 h in a block heater (Stuart SBH200D, Staffordshire, UK). Cells were confirmed as dead if no growth was observed in an overnight culture of 100-μl sample spread plated on nutrient agar.
To determine the lethal cell number (LC50) of each strain at 24-h post infection, different cell numbers of respective strains were used to inject G. mellonella larvae and the number of dead cells was scored at 24 h post infection. The LC50 is the concentration that corresponded to 50% of the larvae death after 24 h, and this was determined by plotting a graph of cell concentration vs percentage death.
All statistical analyses were carried out using GraphPad Prism version 6. Difference in rhamnolipid crude yield from various carbon sources was analysed using two-way ANOVA with Tukey’s multiple comparison (P ≥ 0.05). While differences in biomass and rhamnolipid concentration in the deuterium studies were analysed by Mann-Whitney unpaired t test.
Microbial growth and rhamnolipid production by B. thailandensis and P. aeruginosa
For rhamnolipid production, both B. thailandensis and P. aeruginosa were grown for a period of 216 h at 30 °C in MSM or PPGAS media supplemented with different carbon sources. Figure 1 shows the growth curve of both organisms in media supplemented with glycerol, oleic acid or heptadecanoic acid. No significant difference was observed in the initial average cell concentration across and within bacteria strains used in this experiment, with an average initial cell concentration of approximately log10 8.8 cfu/ml. In both media, B. thailandensis reached an early stationary phase at 24 h and a late stationary phase at 72 h in media supplemented with fatty acids, or 120 h in media supplemented with glycerol. It maintained this phase until 216 h when fermentation was stopped (Fig. 1a, b). These results correlated with the biomass concentrations which either showed no difference between 72 and 216 h or increased until 216 h, depending on the carbon substrate (Figure S1). On the other hand, P. aeruginosa reached early stationary phase around 12 h and went into the late stationary phase between 24 and 48 h. In PPGAS, this phase lasted for a maximum of 72 h with a decline phase observed at 120 h in all carbon substrates (Fig. 1c). In MSM, cells supplemented with fatty acids maintained the late stationary phase until the end of fermentation while cells supplemented with glycerol went into lag phase at 168 h (Fig. 1d).
Triplicate extractions were carried out for each fermentation medium and carbon source after 216 h of fermentation to assess the crude yield of rhamnolipid. Results showed a rhamnolipid crude yield of 1.306 and 0.537 g/l by B. thailandensis in PPGAS using either heptadecanoic acid or glycerol as carbon source respectively (Fig. 2a). These were comparable to results obtained with P. aeruginosa under similar conditions, although the crude yield by P. aeruginosa was significantly higher (P ≥ 0.05) than B. thailandensis using oleic acid as a carbon source in PPGAS (Fig. 2a). Rhamnolipid crude yield in MSM was observed to be considerably higher (P ≥ 0.05) compared to PPGAS from all three carbon sources used in the study. The highest yields by P. aeruginosa and B. thailandensis in MSM were obtained using heptadecanoic acid as a source of carbon (7.543 and 5.588 g/l, respectively). With oleic acid as carbon substrate, the rhamnolipid yield by B. thailandensis was 4.99 g/l comparable to a yield of 4.67 g/l in P. aeruginosa at the end of the fermentation (Fig. 2b).
From the growth analysis, it was observed that the growth pattern of B. thailandensis is different from P. aeruginosa. It had a steady stationary phase until the end of fermentation in all media and carbon sources compared to P. aeruginosa which went into decline prior to the end of fermentation. To see if rhamnolipid yield followed a similar pattern to microbial growth, RL extraction was carried out at 72, 144 and 216 h from MSM supplemented with glycerol. Results showed significant increases in rhamnolipid production by B. thailandensis with crude yields of 0.15, 0.38 and 0.79 g/l at 72, 144 and 216 h (Fig. 3). In contrast, no significant increase in rhamnolipid production by P. aeruginosa was observed at all three extraction times (Fig. 3).
The rhamnolipid obtained from P. aeruginosa using either heptadecanoic acid or oleic acid as substrate both had CMCs of 60 mg/l with surface tensions of 31.63 ± 0.06mN/m and 30.30 ± 0.17mN/m respectively at the CMC concentrations. While rhamnolipids from B. thailandensis E264 using either heptadecanoic acid or oleic acid as substrate had CMCs of 250 mg/l with surface tensions of 39.16 ± 0.47and 41.90 ± 0.40, respectively.
Congener distribution of rhamnolipid from different carbon sources in B. thailandensis and P. aeruginosa
To compare the differences in rhamnolipid congener distribution from the different carbon sources and media, LC-QToF-MS was carried out. Crude extracts were purified by solid-phase extraction prior to analysis. As expected, the results showed that the di-rhamnolipid C14-C14 with a pseudomolecular ion of 761 m/z was the most abundant rhamnolipid congener produced by B. thailandensis (Table S1). The pseudomolecular ion of 649 m/z corresponding to di-rhamnolipid C10-C10 was observed as the most abundant congener produced by P. aeruginosa across all carbon sources (Table S2). As expected, the chromatogram of extracts from P. aeruginosa cultures supplemented with heptadecanoic acid, showed additional peaks at pseudomolecular ion values of 461, 489, 517, 607, 635 and 663 m/z. These are indicative of the following odd carbon chain rhamnolipid congeners: mono-C8-C9/C9-C8, C9-C10/C10-C9, C10-C11/C11-C10 and their corresponding di-rhamnolipid congeners, respectively (Table S2). Interestingly, corresponding odd carbon chain rhamnolipid congeners were less obvious in B. thailandensis cultures supplemented with heptadecanoic acid (Table S1).
Fragmentation of the parent ion obtained from LC-QToF-MS was carried out for each sample using tandem mass spectrometry (MS/MS) to differentiate isomeric congeners that cannot be resolved chromatographically such as mono or di-C10-C12/C12-C10/C11-C11. These results were used to compare the relative proportion of odd and even carbon chain length rhamnolipids produced with heptadecanoic acid as carbon substrate summarised in Table 1. With heptadecanoic acid as a carbon source, P. aeruginosa produced rhamnolipid with odd carbon chain lipid moieties with a cumulative relative abundance of 32.43 and 45.97% in MSM and PPGAS cultures, respectively. In contrast, B. thailandensis cultures supplemented with heptadecanoic acid were observed to produce significantly lower quantities of rhamnolipids with odd carbon chain lipid moieties. The cumulative relative abundance of odd carbon chain length rhamnolipids produced by B. thailandensis are 0.76 and 2.95% in MSM and PPGAS cultures, respectively (Table 1).
The most abundant odd carbon chain rhamnolipid congeners produced by P. aeruginosa in both MSM and PPGAS were di-C9-C10/C10-C9 and di-C10-C11/C11-C10. Di-C14-C15/C15-C14 was the only odd carbon chain rhamnolipid congener produced by B. thailandensis from MSM while di-C13-C14/C14-C13 was observed in addition in PPGAS (result not shown).
Growth of B. thailandensis in deuterated stearic acid supplemented media
The results observed from the biosynthesis of rhamnolipid using different fatty acid substrates suggest that the lipid component in rhamnolipid biosynthesis is supplied by the fatty acid synthesis pathway (FAS II) and not β-oxidation in the presence of a fatty acid substrate. To investigate this further, we designed an experiment following the procedure described by Zhang et al. (2012) but replacing glucose with glycerol. We supplemented MSM with 1% glycerol and 0.25% stearic acid – d 35 . A control experiment supplementing MSM with 1% glycerol and 0.25% undeuterated stearic acid was also set up. The hypothesis was that if β-oxidation supplies the lipid precursor in rhamnolipid biosynthesis, then rhamnolipid with at least seven fully deuterated carbons will be observed in the mass spectrometry. This is as described previously in Pseudomonas cultures using stearic acid – d 35 as co-substrate (Zhang et al. 2012). However, if FAS II supplies the lipid precursor then there will be random incorporation of deuterium within the lipid moieties of the synthesised rhamnolipid.
First, we assessed the impact of using deuterated fatty acid as co-substrate in rhamnolipid biosynthesis on the growth and rhamnolipid production of B. thailandensis E264. As seen in Fig. 4a, deuterated and undeuterated stearic acid supplemented media had no significant differences in microbial concentration at each time point during the growth of B. thailandensis. This indicates that deuterated stearic acid does not have any effect on the growth of B. thailandensis. The biomass concentration was also assessed after 216 h of fermentation. No significant difference in biomass concentration was observed at 216 h of fermentation using either deuterated or undeuterated stearic acid as co-substrate (Fig. 4b). Additionally, no significant difference was observed in the yield of rhamnolipid from deuterated or undeuterated stearic acid as co-substrate (Fig. 4c). Put together, these results suggest that the use of deuterated stearic acid as co-substrate did not have any significant effect on the growth and rhamnolipid biosynthesis of B. thailandensis.
Characterisation of rhamnolipid from deuterated and undeuterated stearic acid
The composition of purified rhamnolipid extracts from deuterated and undeuterated fatty acid is presented in Table 2. The relative abundance is expressed in percentage, and the natural contributions of 13C have being corrected in the relative abundance values are presented in Table 2.
In cultures treated with undeuterated stearic acid as co-substrate, six major rhamnolipid congeners were identified: mono-C14-C14, di-C14, di-C12-C12, di-C12-C14/C14-C12, di-C14-C14 and di-C14-C16/C16-C14 (Table 2). In media containing deuterated stearic acid, these six congeners were also identified; however, additional peaks corresponding to various rates of deuterium incorporation were also found. In all the six major congeners, there is a progressive decline in the number of deuterium atom incorporation, with single deuterium incorporation having the highest percentage abundance across all congeners. Additionally, the maximum amount of deuterium incorporation observed was 15, seen only in the di-C14-C14 at a very low abundance of 0.06% (Table 2).
Pathogenicity of B. thailandensis E264 and P. aeruginosa PAO1
Using the G. mellonella model, we compared the pathogenicity of B. thailandensis and P. aeruginosa. Results showed 100% mortality in larvae infected with 100 cfu of P. aeruginosa within 24 h of infection (Fig. 5). However, infecting larvae with 100 cfu of B. thailandensis showed 100% survival at 24 h. Although, mortality rate increases after 24 h and a 100% mortality can be observed at 48 h post infection. The LC50 of both organisms at 24 h was determined and results showed that P. aeruginosa has LC50 of < 3.5 cfu while B. thailandensis has an LC50 of approximately 2500 cfu. This indicates that the LC50 of B. thailandensis at 24 h in G. mellonella is approximately three logs higher than that of P. aeruginosa. No mortality was observed using heat-killed cells of either organism (Fig. 5).
The two major studies on rhamnolipid production by B. thailandensis E264 were carried out in nutrient rich broth, supplemented with either glycerol or canola oil as carbon sources (Dubeau et al. 2009; Funston et al. 2016). In the latter study, it was shown that the organism had a sustained stationary phase until the end of fermentation (Funston et al. 2016). Our study confirms that B. thailandensis has a sustained stationary phase even when grown in minimal or nutrient limiting conditions, with no decline phase observed in its growth curve during the period of this study. However, P. aeruginosa went into decline phase at all carbon substrate conditions in PPGAS and with glycerol as substrate in MSM (Fig. 1). We chose P. aeruginosa PAO1 for this comparative study, as it has been shown to be a model for the study of rhamnolipid production (Müller et al. 2010).
We have earlier shown that the rhamnolipid genes (rhlA, rhlB and rhlC) are expressed at elevated levels throughout the stationary phase of growth in B. thailandensis (Funston et al. 2016). This is not the case in P. aeruginosa where rhlA and rhlB are downregulated as the cells enter into stationary growth phase with rhlC expression elevated at the onset of stationary phase and then declining subsequently (Perfumo et al. 2013). An extended stationary phase coupled with continuous expression of rhamnolipid synthesis genes will potentially result in continuous rhamnolipid production. This is because the rate of rhamnolipid production is known to increase as the cells enter their stationary phase (Benincasa et al. 2002). This may therefore account for the continuous increase in rhamnolipid yield, observed throughout the period of fermentation in B. thailandensis, compared to P. aeruginosa where the rhamnolipid yield at 72 h was comparable to that at 216 h (Fig. 3).
Additionally, no significant difference in rhamnolipid crude yield from B. thailandensis was observed at 216 h compared to P. aeruginosa in either PPGAS (with either heptadecanoic acid or glycerol as carbon source) or MSM (with oleic acid as carbon source). Given that these media and growth conditions were designed for rhamnolipid production in P. aeruginosa, it is possible that optimising production conditions for B. thailandensis can give yields with values higher than those observed in P. aeruginosa. However, it is important to note that while rhamnolipid crude yield in B. thailandensis can be comparable to that from P. aeruginosa at 216 h, crude yield in P. aeruginosa peaked at 72 h in this study or at 48 h as reported in previous comparative studies (De-Rienzo et al. 2016). At this time (72 h), the rhamnolipid crude yield from P. aeruginosa was significantly higher compared to B. thailandensis under all conditions studied. As the industrial cost of running a fermenter for 216 h is substantially higher than that needed to run the same fermentation for 72 h, industries would prefer to use P. aeruginosa in their rhamnolipid production; considering cost savings only. However, the development of methods to enhance the early initiation of rhamnolipid production in B. thailandensis will further boost the industrial appeal of this organism as an alternative rhamnolipid producer. For example, a study in 2010 with P. aeruginosa showed that iron limitation resulted in the expression of rhlA, seven- to tenfold earlier compared to iron rich media and this simultaneously led to early initiation of rhamnolipid production (Glick et al. 2010). Also, the introduction of exogenous quorum sensing molecules is known to induce and increase the production of rhamnolipid in P. aeruginosa (Dusane et al. 2010). Similar techniques as those previously can be applied to induce the synthesis of rhamnolipid early in the growth of B. thailandensis. Thus, improving the overall yield of rhamnolipid from these strains and subsequently make them more industrially relevant.
The industrial appeal of B. thailandensis is further boosted by its significantly lower pathogenicity confirmed in this study. B. thailandensis E264 is a biosafety level 1 organism and is generally regarded to be non-pathogenic. However, its physiological similarity with the pathogenic B. pseudomallei, the causative agent of melioidosis, coupled with a recent report of disease caused by infection with strains of B. thailandensis (not E264) (Glass et al. 2006) led us to compare the pathogenicity of P. aeruginosa PAO1 and B. thailandensis E264. As expected, results using the G. mellonella virulence model showed that B. thailandensis E264 is a less pathogenic alternative for rhamnolipid production compared to P. aeruginosa PAO1 (Fig. 5). The LC50 at 24 h of B. thailandensis was also shown to be three logs higher than that of P. aeruginosa. Furthermore, no lethal effect of heat-killed cells was observed in both strains ruling out the possibility of cell toxicity rather than pathogenicity.
One of the main objective of this study was to analyse the differences is congener distribution produced by B. thailandensis and P. aeruginosa when grown under different substrate conditions. Results showed that B. thailandensis produces rhamnolipid with longer chain length lipid moieties (predominantly C14-C14) and fewer congeners (< 15 including isomeric congeners). While P. aeruginosa produces shorter chain length rhamnolipid lipid moieties (predominantly C10-C10), with more diverse congener distribution (> 36 different congeners identified in this study). These results corroborate previous studies (De-Rienzo et al. 2016; Dubeau et al. 2009; Elshikh et al. 2017; Funston et al. 2016).
However, no report on the effect of fatty acid substrate chain length on the rhamnolipid lipid moieties of B. thailandensis is available in the literature. It has previously been reported that when P. aeruginosa is fed with odd carbon chain fatty acid as a carbon substrate, the cells produce rhamnolipids with odd and even carbon chain lipid components. While cells fed with even chain fatty acids produce rhamnolipid with purely even carbon chains (Hori et al. 2011). This led to the assumption that the lipid component of rhamnolipid is either supplied by β-oxidation or that β-oxidation intermediates are diverted into the fatty acid synthesis pathway (FAS II) and subsequently used as lipid precursors for rhamnolipid biosynthesis. The latter assumption was later proved by Zhang et al. using isotope tracing and gene expression studies of both β-oxidation and FAS II synthesis genes (Zhang et al. 2012). This is also backed by previous studies that have shown that P. aeruginosa is able to shunt intermediates of β-oxidation into the FAS II pathway (Yuan et al. 2012). However, a more recent report has shown that β-oxidation directly supplies the lipid precursor in rhamnolipid biosynthesis without the need for FAS II elongation (Abdel-Mawgoud et al. 2014).
To see if either of these is true for B. thailandensis, this study assessed the rhamnolipid congener distribution of B. thailandensis. Odd or even carbon chain length fatty acids or glycerol were used as carbon source; P. aeruginosa was used as a control. As expected, it was observed that when grown with either even chain fatty acid or glycerol, P. aeruginosa produces rhamnolipid with solely even carbon chain lipid moieties. While when grown in odd carbon chain fatty acid (C17), it produced rhamnolipid with both odd and even carbon chain lipid moieties (Table S2). Similarly, when B. thailandensis was grown with even carbon chain fatty acid or glycerol as carbon source, only rhamnolipids with even carbon chain lipid moieties were observed. Interestingly, when B. thailandensis was grown with odd carbon chain fatty acid as the sole carbon source, it produces rhamnolipid with predominantly even carbon chain lipid moieties (Table S1). This result was unexpected and suggested that the rhamnolipid lipid moieties of B. thailandensis are obtained predominantly from the FAS II pathway. Different from P. aeruginosa, known to obtain its lipid precursor predominantly from β-oxidation (Abdel-Mawgoud et al. 2014).
To further investigate this, we carried out isotope tracing studies using 1% glycerol + 0.25% stearic acid – d 35 as carbon sources, 1% glycerol + 0.25% stearic acid was used as a control. We first analysed the effect of using deuterated co-substrate in bacteria growth and rhamnolipid production. This was done as previous report has shown that P. aeruginosa is inhibited by deuterated substrates (Smyth et al. 2010a) and the organism needed to be ‘trained’ to grow on them. Results indicate that the use of deuterated stearic acid as co-substrate in rhamnolipid biosynthesis did not have any effect on either the growth or rhamnolipid production of B. thailandensis (Fig. 4a–c). This could be because deuterated stearic acid is only a co-substrate in this study; therefore, cells can easily adapt to it and grow using glycerol supplied as an additional substrate. This is further evidenced in a previous study that used deuterated stearic acid as co-substrate in rhamnolipid production by P. aeruginosa. The report showed that deuterated stearic acid did not have any effect on either microbial growth or rhamnolipid yield (Zhang et al. 2012).
The synthesised rhamnolipid were then characterised to see the pattern of deuterium incorporation within the various congeners. Isotopologous rhamnolipid congeners with varied levels of deuterium incorporation in all congeners were observed. Deuterium incorporation from 0 to 15 can be seen in the most abundant Di-C14-C14 rhamnolipid congener. To see if FAS II synthesis can explain these various levels of deuterium incorporation, we carried out isotope tracing analysis for the synthesis of (R)-β-hydroxytetradecanoyl-ACP, the lipid precursor for the synthesis of Di-C14-C14. The analysis showed that in a single C14, isotopes ranging from C14-d 0 to C14-d 10 are possible (Table S3 and Figure S2). This therefore suggests that a maximum of 20 deuterium atoms can be incorporated into the lipid structure of the Di-C14-C14 using FAS II lipid precursors.
If β-oxidation of stearic acid – d 35 directly supplies the lipid precursor to synthesise Di-C14-C14, as previously suggested for P. aeruginosa (Abdel-Mawgoud et al. 2014), then a minimum of 27 deuterium atoms will be present in a single C14 chain, as it will be fully deuterated. This will result in 54 deuterium atoms in a Di-C14-C14 rhamnolipid congener. Also, if β-oxidation intermediates (C8-d 15 ) are diverted to FAS II and elongated to give the C14 lipid precursor, as was also suggested for P. aeruginosa (Zhang et al. 2012), then a minimum of 15 deuterium atoms will be present in a single C14 lipid chain. This will give a minimum number of 30 deuterium atoms in a Di-C14-C14 rhamnolipid congener.
Based on the explanations mentioned and results from isotope tracing, we suggest that the lipid precursor for rhamnolipid biosynthesis in B. thailandensis is obtained from FAS II or de novo fatty acid synthesis. Combining results obtained from studies using different fatty acid substrate and that of isotope tracing, we concluded that FAS II is the main supplier of the lipid precursor in B. thailandensis. However, the observation of small quantities of single odd chain lipid moiety in the Di-C14-C14 using odd chain fatty acid as carbon source suggests the use of either direct β-oxidation or elongated β-oxidation products in rhamnolipid biosynthesis. These, however, contribute less than 3% of the total rhamnolipid yield based on our study. These results suggest that the rhlA of B. thailandensis has a preference for β-hydroxylacyl-ACP from FAS II compared to those from β-oxidation. The fact that B. thailandensis produces rhamnolipid with long chain lipid moieties (majorly C12, C14 and C16) further supports the supply of lipid precursors by the FAS II pathway. This is because fatty acid biosynthesis in most biological systems is known to be terminated with the C16 saturated straight chain palmitic acid (Smith 1994).
The use of FAS II as the main supplier of the lipid precursor in rhamnolipid biosynthesis is industrially relevant. It implies that B. thailandensis will be able to use a wide range of substrates with yields comparable to those of fatty acid substrates, compared to P. aeruginosa in which yield from sugars are significantly lower than those from lipids. Although in this study, fatty acid substrates produced the highest yields compared to glycerol. This might be because hydroxyfatty acids from β-oxidation are recruited for other cell processes, thus more β-hydroxyfatty acids from FAS II are available for rhamnolipid biosynthesis. This creates an opportunity to genetically modify B. thailandensis for increased rhamnolipid production, by creating mutants unable to divert FAS II products to other metabolic cell processes. A clear example is the enhanced rhamnolipid production in polyhydroxyalkanoate (PHA)-deficient mutants of B. thailandensis (Funston et al. 2017). Contrarily, PHA-deficient mutants of P. aeruginosa did not produce significantly higher rhamnolipid yields compared to wild type; however, rhamnolipid-deficient mutants produced significantly higher yields of PHA (Choi et al. 2011).
In conclusion, these results show that B. thailandensis is a potentially less pathogenic substitute for industrial production of rhamnolipid compared to P. aeruginosa. However, to fully exploit these strains, it is important that more research be carried out to understand factors that can improve rhamnolipid yield, especially in the initial stages of cell growth as explained earlier. Furthermore, studies of rhamnolipid biosynthesis using different fatty acid substrates coupled with isotope tracing suggest that rhamnolipid biosynthesis in B. thailandensis differs significantly from P. aeruginosa. This further strengthens the need for studies in rhamnolipid biosynthesis in the Burkholderia species, to further understand their natural function and mechanism of biosynthesis. These studies will help improve the industrial appeal of these less virulent strains in rhamnolipid biosynthesis compared to P. aeruginosa.
Abdel-Mawgoud AM, Lépine F, Déziel E (2014) A stereospecific pathway diverts β-oxidation intermediates to the biosynthesis of rhamnolipid biosurfactants. Chem Biol 21(1):156–164. https://doi.org/10.1016/j.chembiol.2013.11.010
Benincasa M, Contiero J, Manresa MA, Moraes IO (2002) Rhamnolipid production by Pseudomonas aeruginosa LBI growing on soapstock as the sole carbon source. J Food Eng 54(4):283–288. https://doi.org/10.1016/S0260-8774(01)00214-X
Brett PJ, DeShazer D, Woods DE (1998) Burkholderia thailandensis sp. nov., a Burkholderia pseudomallei-like species. Int J Sys Evol Microbiol 48(1):317–320
Caiazza NC, Shanks RMQ, O'Toole GA (2005) Rhamnolipids modulate swarming motility patterns of Pseudomonas aeruginosa. J Bacteriol 187(21):7351–7361. https://doi.org/10.1128/jb.187.21.7351-7361.2005
Choi MH, Xu J, Gutierrez M, Yoo T, Cho YH, Yoon SC (2011) Metabolic relationship between polyhydroxyalkanoic acid and rhamnolipid synthesis in Pseudomonas aeruginosa: comparative 13C NMR analysis of the products in wild-type and mutants. J Biotechnol 151(1):30–42
Costa SG, Déziel E, Lépine F (2011) Characterization of rhamnolipid production by Burkholderia glumae. Lett Appl Microbiol 53(6):620–627. https://doi.org/10.1111/j.1472-765X.2011.03154.x
De-Rienzo MAD, Kamalanathan ID, Martin PJ (2016) Comparative study of the production of rhamnolipid biosurfactants by B. thailandensis E264 and P. aeruginosa ATCC 9027 using foam fractionation. Process Biochem 51(7):820–827. https://doi.org/10.1016/j.procbio.2016.04.007
Dubeau D, Deziel E, Woods D, Lepine F (2009) Burkholderia thailandensis harbors two identical rhl gene clusters responsible for the biosynthesis of rhamnolipids. BMC Microbiol 9(1):263
Dusane DH, Zinjarde SS, Venugopalan VP, McLean RJ, Weber MM, Rahman PK (2010) Quorum sensing: implications on rhamnolipid biosurfactant production. Biotechnol Genet Eng Rev 27:159–184
Elshikh M, Funston S, Chebbi A, Ahmed S, Marchant R, Banat IM (2017) Rhamnolipids from non-pathogenic Burkholderia thailandensis E264: physicochemical characterization, antimicrobial and antibiofilm efficacy against oral hygiene related pathogens. New Biotechnol 36:26–36
Funston SJ, Tsaousi K, Rudden M, Smyth TJ, Stevenson PS, Marchant R, Banat IM (2016) Characterising rhamnolipid production in Burkholderia thailandensis E264, a non-pathogenic producer. Appl Microbiol Biotechnol 100(18):7945–7956. https://doi.org/10.1007/s00253-016-7564-y
Funston SJ, Tsaousi K, Smyth TJ, Twigg MS, Marchant R, Banat IM (2017) Enhanced rhamnolipid production in Burkholderia thailandensis transposon knockout strains deficient in polyhydroxyalkanoate (PHA) synthesis. Appl Microbiol Biotechnol 101(23):8443–8454. https://doi.org/10.1007/s00253-017-8540-x
Glass MB, Gee JE, Steigerwalt AG, Cavuoti D, Barton T, Hardy RD, Godoy D, Spratt BG, Clark TA, Wilkins PP (2006) Pneumonia and septicemia caused by Burkholderia thailandensis in the United States. J Clin Microbiol 44(12):4601–4604. https://doi.org/10.1128/jcm.01585-06
Glick R, Gilmour C, Tremblay J, Satanower S, Avidan O, Déziel E, Greenberg EP, Poole K, Banin E (2010) Increase in rhamnolipid synthesis under iron-limiting conditions influences surface motility and biofilm formation in Pseudomonas aeruginosa. J Bacteriol 192(12):2973–2980. https://doi.org/10.1128/jb.01601-09
Haraga A, West TE, Brittnacher MJ, Skerrett SJ, Miller SI (2008) Burkholderia thailandensis as a model system for the study of the virulence-associated type III secretion system of Burkholderia pseudomallei. Infect Immun 76(11):5402–5411. https://doi.org/10.1128/iai.00626-08
Hori K, Ichinohe R, Unno H, Marsudi S (2011) Simultaneous syntheses of polyhydroxyalkanoates and rhamnolipids by Pseudomonas aeruginosa IFO3924 at various temperatures and from various fatty acids. Biochem Eng J 53(2):196–202. https://doi.org/10.1016/j.bej.2010.10.011
Irorere VU, Tripathi L, Marchant R, McClean S, Banat IM (2017) Microbial rhamnolipid production: a critical re-evaluation of published data and suggested future publication criteria. Appl Microbiol Biotechnol 101(10):3941–3951. https://doi.org/10.1007/s00253-017-8262-0
Koh SF, Tay ST, Sermswan R, Wongratanacheewin S, Chua KH, Puthucheary SD (2012) Development of a multiplex PCR assay for rapid identification of Burkholderia pseudomallei, Burkholderia thailandensis, Burkholderia mallei and Burkholderia cepacia complex. J Microbiol Methods 90(3):305–308. https://doi.org/10.1016/j.mimet.2012.06.002
Miles AA, Misra SS, Irwin JO (1938) The estimation of the bactericidal power of the blood. Epidemiol Infect 38(06):732–749. https://doi.org/10.1017/S002217240001158X
Moya Ramírez I, Tsaousi K, Rudden M, Marchant R, Jurado Alameda E, García Román M, Banat IM (2015) Rhamnolipid and surfactin production from olive oil mill waste as sole carbon source. Bioresour Technol 198:231–236. https://doi.org/10.1016/j.biortech.2015.09.012
Müller MM, Hörmann B, Syldatk C, Hausmann R (2010) Pseudomonas aeruginosa PAO1 as a model for rhamnolipid production in bioreactor systems. Appl Microbiol Biotechnol 87(1):167–174. https://doi.org/10.1007/s00253-010-2513-7
Mulligan CN, Mahmourides G, Gibbs BF (1989) The influence of phosphate metabolism on biosurfactant production by Pseudomonas aeruginosa. J Biotechnol 12(3):199–209. https://doi.org/10.1016/0168-1656(89)90041-2
Perfumo A, Rudden M, Smyth TJ, Marchant R, Stevenson PS, Parry NJ, Banat IM (2013) Rhamnolipids are conserved biosurfactants molecules: implications for their biotechnological potential. Appl Microbiol Biotechnol 97(16):7297–7306
Smith S (1994) The animal fatty acid synthase: one gene, one polypeptide, seven enzymes. FASEB J 8(15):1248–1259
Smyth TJ, Perfumo A, Marchant R, Banat IM, Chen M, Thomas RK, Penfold J, Stevenson PS, Parry NJ (2010a) Directed microbial biosynthesis of deuterated biosurfactants and potential future application to other bioactive molecules. Appl Microbiol Biotechnol 87(4):1347–1354
Smyth TJ, Perfumo A, Marchant R, Banat IM (2010b) Isolation and analysis of low molecular weight microbial glycolipids. In: Timmis KN (ed) Handbook of hydrocarbon and lipid microbiology. Springer Berlin Heidelberg, Berlin, pp 3705–3723
Smyth TJ, Rudden M, Tsaousi K, Marchant R, Banat IM (2014) Protocols for the detection and chemical characterisation of microbial glycolipids. In: McGenity T, Timmis K, Nogales B (eds) Hydrocarbon and lipid microbiology protocols. Springer Protocols Handbooks. Springer, Berlin, pp 29–60
Stover CK, Pham XQ, Erwin AL, Mizoguchi SD, Warrener P, Hickey MJ, Brinkman FSL, Hufnagle WO, Kowalik DJ, Lagrou M, Garber RL, Goltry L, Tolentino E, Westbrock-Wadman S, Yuan Y, Brody LL, Coulter SN, Folger KR, Kas A, Larbig K, Lim R, Smith K, Spencer D, Wong GKS, Wu Z, Paulsen IT, Reizer J, Saier MH, Hancock REW, Lory S, Olson MV (2000) Complete genome sequence of Pseudomonas aeruginosa PAO1, an opportunistic pathogen. Nature 406(6799):959–964
Tavares LFD, Silva PM, Junqueira M, Mariano DCO, Nogueira FCS, Domont GB, Freire DMG, Neves BC (2013) Characterization of rhamnolipids produced by wild-type and engineered Burkholderia kururiensis. Appl Microbiol Biotechnol 97(5):1909–1921. https://doi.org/10.1007/s00253-012-4454-9
Yuan Y, Leeds JA, Meredith TC (2012) Pseudomonas aeruginosa directly shunts β-oxidation degradation intermediates into de novo fatty acid biosynthesis. J Bacteriol 194(19):5185–5196. https://doi.org/10.1128/jb.00860-12
Zhang L, Veres-Schalnat TA, Somogyi A, Pemberton JE, Maier RM (2012) Fatty acid cosubstrates provide β-oxidation precursors for rhamnolipid biosynthesis in Pseudomonas aeruginosa, as evidenced by isotope tracing and gene expression assays. Appl Environ Microbiol 78(24):8611–8622. https://doi.org/10.1128/aem.02111-12
The authors will like to appreciate Sarah Dobbin for her technical support during the analysis of rhamnolipid samples. VUI acknowledges the support provided by an Ulster University Vice Chancellor’s Research Scholarship.
This article does not contain any studies with human participants or animals performed by any of the authors.
Conflict of interest
The authors declare that they have no conflict of interest.
Electronic supplementary material
About this article
Cite this article
Irorere, V.U., Smyth, T.J., Cobice, D. et al. Fatty acid synthesis pathway provides lipid precursors for rhamnolipid biosynthesis in Burkholderia thailandensis E264. Appl Microbiol Biotechnol 102, 6163–6174 (2018). https://doi.org/10.1007/s00253-018-9059-5
- FAS II
- B. thailandensis
- P. aeruginosa
- Deuterium labelling