Simultaneous polyhydroxyalkanoates and rhamnolipids production by Thermus thermophilus HB8

A wide variety of microorganisms accumulate polyhydroxyalkanoic acids (PHAs), mainly polyhydroxybutyrate (PHB), as metabolic storage materials, which are deposited as intracellular water-insoluble inclusions (Griebel et al. 1968; Rehm and Steinbüchel 1999). Most fluorescent Pseudomonads strains belonging to rRNA homology group I, are able to synthesize and accumulate large amounts of PHAs consisting of various 3-hydroxy fatty acids with carbon chain lengths ranging from 6 to 14 carbon atoms (medium chain length-mcl) as carbon and energy storage compounds (Huisman et al. 1989). The composition of PHAs depends on the PHA synthases (polymerases) (PhaC), the key enzymes for PHA biosynthesis, the carbon source and the metabolic routes involved (Rehm 2003; Rehm and Steinbüchel 1999). Purified mcl-PHA synthases, from P. aeruginosa exhibit in vitro enzyme activity with (R)-3-hydroxydecanoyl-CoA as substrate (Amara and Rehm 2003; Qi et al. 2000; Ren et al. 2000). β-Oxidation is the main pathway for mcl-PHA biosynthesis when fatty acids are used as carbon sources, while de novo biosynthesis of fatty acids is the main route during growth on carbon sources which are metabolized to acetyl coenzyme A (acetyl-CoA), such as gluconate, acetate, or ethanol (Rehm et al. 2001). Besides the intracellular accumulation of PHAs, P. aeruginosa is capable to produce various exo-products, such as rhamnolipids (RLs), which are expressed in the onset of the stationary phase (Déziel et al. 1996).

RLs are glycolipidic bio-surfactants, which reduce water surface tension and emulsify oil. The RLs produced by P. aeruginosa in liquid cultures are mainly rhamnosyl-β-hydroxydecanoyl-β-hydroxydecanoate (mono-RL) and rhamnosyl-rhamnosyl-β-hydroxydecanoyl-β-hydroxydecanoate (di-RLs) (Ochsner and Reiser 1995). RL biosynthesis proceeds through transfer of two rhamnose moieties from TDP-L-rhamnose (Maier and Soberón-Chávez 2000). For the synthesis of mono-RL, the enzyme rhamnosyl-transferase 1 (Rt 1) catalyses the rhamnose transfer to β-hydroxydecanoyl-β-hydroxydecanoate, while rhamnosyl-transferase 2 (Rt 2) synthesizes di-RL from TDP-L-rhamnose and mono-RL. Genes for biosynthesis, regulation and induction of the Rt 1 enzyme are organized in tandem in the rhlABRI gene cluster (Ochsner and Reiser 1995). The gene rhlC encoding the Rt 2 enzyme has been described (Rahim et al. 2001), and is homologous to rhamnosyl-transferases involved in lipopolysaccharide biosynthesis. Some evidence was recently provided that RhlA is involved in synthesis of 3-(3-hydroxyalkanoyloxy)alkanoic acids (HAAs) (Déziel et al. 2003); however, it still remains unclear whether the PHA synthase is capable of catalyzing the synthesis of HAAs, which has been previously postulated (Campos-Garcia et al. 1998; Déziel et al. 2003; Rehm et al. 2001).

The expression of phaG gene, encoding transacylase from P. putida in mutants indicated that PhaG catalyzes diversion of intermediates of fatty acid de novo biosynthesis towards PHA biosynthesis, and in the transacylase-mediated PHA biosynthesis route from gluconate, PhaG is the only linking enzyme between fatty acid de novo biosynthesis and PHA biosynthesis. In addition, expression of the β-ketoacyl reductase gene rhlG from P. aeruginosa in mutants revealed that RhlG catalyzes diversion towards RL biosynthesis (Campos-Garcia et al. 1998). RhlG is thought to catalyze the NADPH-dependent reduction of β-ketodecanoyl-ACP, which is an intermediate of fatty acid de novo biosynthesis, resulting in β-hydroxydecanoyl-ACP, a putative precursor for RL biosynthesis. The proposed pathways for mcl-PHA and RLs biosynthesis suggested that both biosynthesis pathways are competitive (Rehm et al. 2001).

While the opportunistic pathogen P. aeruginosa has traditionally been considered the primary PHAs and RLs-producing microorganism, however it appears that the ability to produce PHAs and especially RLs (Abdel-Mawgoud et al. 2010) is in fact restricted to a limited number of bacterial species. Since there are a plethora of applications for both metabolites PHAs (Zinn et al. 2001; Valappil et al. 2006) and RLs (Abdel-Mawgoud et al. 2010) a low cost production would be very challenging. A novel approach for reducing their production costs was also reported for the simultaneous production of the intracellular PHAs and the extracellular RLs (Hori et al. 2002). PHAs production was reported using the remaining oil from RL production (Füchtenbusch et al. 2000). This approach was adopted to reduce PHA production cost by using the remaining carbon source for RL production. Palm oil, one of the typical plant oils, was also used as the sole carbon source for the simultaneous production of both products (Marsudi et al. 2008).

Bio-surfactants are produced by certain bacteria, yeasts, and filamentous fungi during cultivation on various carbon sources, in particular during growth on hydrophobic substances such as hydrocarbons leading to the assumption that biosurfactants serve to emulsify the hydrocarbons in the growth medium thus facilitating their uptake. A correlation between surfactant production and growth on water-insoluble substrates was shown (Itoh and Suzuki 1972). RLs are usually produced by P. aeruginosa also in media containing water-soluble carbon sources such as glucose, glycerol, mannitol, and ethanol as a carbon source, particularly, when the cells become limited for nitrogen (Guerra-Santos et al. 1986; Mulligan and Gibbs 1993; Robert et al. 1989; Sim et al. 1997; Wagner et al. 1983) indicating that they may serve other roles besides being involved in solubilizing hydrophobic substrates; nevertheless, the final surfactant concentration was lower than that obtained when n-alkanes and vegetable oils were the carbon substrates (Syldatk et al. 1985; Robert et al. 1989). The carbon source influences bio-surfactant synthesis by either induction or repression. An induction of glycolipid production by P. aeruginosa SB30, when n-alkanes were added to the medium was reported (Chakrabarty 1985). Catabolic repression of bio-surfactant synthesis by glucose, acetate, and tricarboxylic acids was also observed (Hauser and Karnovsky 1958, 1954). When P. aeruginosa UG2 was grown on hydrophobic substrates such as corn oil, lard, and long chain alcohols, production was around 100-165 mg of rhamnolipid/g substrate, whereas when hydrophilic substrates such as glucose and succinic acid were used, only 12-36 mg/g substrate was obtained (Mata-Sandoval et al. 1999).

Polyhydroxyalkanoates (PHAs) are intracellular storage compounds of carbon and energy that are produced by many bacteria in the form of inclusion bodies (Anderson and Dawes 1990). PHAs have attracted commercial biotechnological interest because of their biodegradability and biocompatibility (Reddy et al. 2003).

PHA biosynthesis by the thermophilic bacterium T. thermophilus HB8 has been previously reported (Pantazaki et al. 2003) while the soluble PHA synthase (Pantazaki et al. 2003), and a β-ketoacyl-CoA thiolase (Pantazaki et al. 2005), were purified from sodium gluconate-grown cells. In addition, PHAs production by T. thermophilus HB8 using whey as sole carbon source was also reported (Pantazaki et al. 2009). Recently, it was also demonstrated that T. thermophilus secreted RLs, when grown in the presence of insoluble substrates such as sunflower seed oil or oleic acid. The produced RLs have been identified by several methods e.g the orcinol method, Thin Layer Chromatography (TLC) and Attenuated Total Reflection Infrared (ATR-FTIR) spectroscopy and LC-(ESI)-MS (Pantazaki et al. 2010).

In this study we report the simultaneous production of both metabolites PHAs accumulated intra-cellularly and RLs secreted extra-cellularly by employing T. thermophilus HB8, as an approach to turn to advantage from the carbon source used extensively for the production of high added value bio-products from wastes aiming to obtaining the reduction of production cost. In this concept, this study specifically focused in finding the optimum growth condition by examining the combine effect of cultivation time and initial [PO₄ ^3-] on the simultaneous production of these metabolites to increase their production. The identification of the molecular composition of the produced PHAs and bio-surfactants RLs was also evaluated by GC chromatography and thin-layer chromatography (TLC)/LC-MS analysis respectively.

T. thermophilus was demonstrated to be capable to produce simultaneously two biotechnologically important bio-products; biodegradable biopolymers PHAs and RLs-type biosurfactants by utilizing either glucose or sodium gluconate as sole carbon source, although the ability to produce them, and especially RLs, is in fact restricted to a limited number of bacterial species (Abdel-Mawgoud et al. 2010). One of the long-term interests of the scientists is to produce RLs in strains other than the opportunistic known pathogen P. aeruginosa (Chwalek et al. 2006) due to the wide field of their applications. This paper contributes to this goal taken into account the advantages of T. thermophilus HB8 due to the exceptionality of the non-pathogenic and thermophilic nature of this strain, and its capability of naturally producing these metabolites. Apart from the advantages resulting from the thermophilicity of the microorganism that reflects positively on the solubility of substrates and the reduced likelihood of contamination should be taken into account the requirement of high temperature of 75°C and consequently the high cost of energy for bacterial growth, and production of both bio-products. A good compensation for overcoming this problem will be the use of inexpensive wastes substrates.

The usage of T. thermophilus for their production in low cost seems highly promising due to the high solubility of suitable carbon sources located in wastes, especially of water-immiscible substrates, where RLs production was high. The consumption of gluconate as sole carbon source by T. thermophilus resulted in higher production of both metabolites than that when glucose was used. Among the various initial [PO₄ ^3-] tested, the medium containing initially 25 mM of [PO₄ ^3-] resulted in the higher polymer production after 72 h of cultivation (Table 1), consistent with the report that the initial [PO₄ ^3-] in the culture medium influences both the cell density and the PHAs production (Lee et al. 2000). The recovery of PHAs requires the disruption of valuable biocatalysts of bacterial cells, while the production of secreted materials, such as RLs, in combination with PHAs production is expected to enhance the availability of biocatalysts because the cells can be used as "microbial cell factories" for the production of the exo-products before the eventual disruption to recover PHAs.

The investigation of the combine effect of both the cultivation time and the initial [PO₄ ^3-] on biomass, and on simultaneous PHAs and RLs production was evaluated from contour plots (Figure 3). As it was revealed, biomass was increased by increasing both initial [PO₄ ^3-] and cultivation time in the experimental ranges tested. However, a further increase of the initial [PO₄ ^3-] and cultivation time would lead the contours to convergence at a concrete point (Figure 3a). Moreover, PHAs production was increasing during cultivation time but there would not be further rise in PHAs production obtained by increasing [PO₄ ^3-] above a crucial concentration, i.e. above 30 mM (Figure 3b), as it can be deduced by the convergence of the curve toward the center point of initial phosphate concentration. In contrast, the increase of the initial [PO₄ ^3-] had a negative effect on RLs production, but the biosurfactant concentration increased during cultivation time (Figure 3c). These results revealed that PHAs synthesis occurred only during active cell growth, while substantial RLs production began early but is completed at the onset of the stationary phase. This strategy might be useful as the beginning to optimize other parameters affecting these processes such as nitrogen, carbon sources etc.

PHAs synthesized in both carbon sources are copolymers with a very similar composition of monomers possessing carbon chain lengths ranging from 6-12 carbon atoms. Since PHAs and RLs syntheses have a close metabolic relation, since (R)-3-hydroxyfatty acids are used as common precursors in both pathways, it has been proposed that both PHAs and RLs biosynthesis pathways may compete for (R)-3-hydroxyfatty acid ACP (Rehm et al. 2001; Soberón-Chávez et al. 2005). However, there is still no clear evidence for this probable competitive relation. A regulatory switch of the flow of beta-hydroxyalkanoic acids from PHAs polymerization to RLs synthesis is clearly indicated to occur when the culture reaches the stationary phase.

Recently, a proposed distributional pathways of (R)-3-hydroxyacyl groups for the synthesis of PHAs, HAAs ((R)-3-((R)-3-hydroxyalkanoyloxy) alkanoic acids), mono-RLs and di-RLs in Pseudomonas aeruginosa grown on medium-chain-length fatty acids was published. It was suggested that during the exponential cell growth, cells excessively produced free acids 3HAs (such as 3HO, 3HD, etc.), which are secreted into medium and retake-up later for RLs syntheses (Choi et al. 2011). An interrelationship between PHAs and RLs synthesis was found that in rhamnolipid-negative mutants, more (R)-3-hydroxyfatty acid precursors were fluxed toward PHA synthesis. However, the production of rhamnolipid in P. aeruginosa seemed to be very tightly regulated at the transcriptional level by the quorum-sensing response and by environmental conditions. Finally, investigation of metabolic pathway for PHAs and RLs synthesis using octanoic-1-¹³C acid suggested that the fatty acid substrate was converted to PHAs and the acyl groups in the RLs with a minimum number of rearrangements (Choi et al. 2011).

Concerning T. thermophilus HB8, there is no other information apart of our works for the biosynthetic pathway of PHAs and RLs neither for the corresponding genes in the genome, except that only a rhamnosyl-transferase (YP-144151; TTHA00885) was found, exhibiting a calculated molecular weight of 32,783, and consisting of 290 amino acids (Pantazaki et al. 2010). Recently it was also demonstrated the potential production of RLs by fatty acids by cultivating this bacterium in the presence of sunflower seed oil or oleic acid as carbon sources (Pantazaki et al. 2010) or by sugars such as glucose and its derivative sodium gluconate in this work. Taking into account that T. thermophilus produce mcl-PHAs, in which 3-hydroxydecanoate constituted the major constituent when grown in the presence of gluconate, and since 3-hydroxydecanoate can be formed via β-oxidation or via fatty acid de novo biosynthesis, we can suggest that both biosynthesis pathways, might be involved in RLs biosynthesis in this bacterium dependent on the carbon source used (Pantazaki et al. 2010). The β-oxidation leads to the desired R-configurated hydroxy acid and the de novo biosynthesis at first to the S-configurated compound (additional biochemical reactions are necessary to give also in this case the accurate configuration of the polymer precursor). These results confirmed that the type of RLs produced depends on multiple conditions, including the bacterial strain and the carbon source used as previously reported (Mulligan and Gibbs 1993; Robert et al. 1989).

The extracted mixture of RLs produced in the presence of sodium gluconate was subjected to identification and characterization using LC-(ESI)-MS. LC-MS analyses revealed that both mono- and di-RL moieties as well as di-rhamno-mono-lipidic congeners were present in the RLs mixture. Additionally, a plethora of diversity of mono- and di-RLs as well, was identified differing in the length of the lipidic chain, which additionally found to be saturated or unsaturated in some cases (Table 2). The combination of negative and positive ESI modes has been shown to be powerful tools in the characterization of RLs homologues. The wild type strain of T. thermophilus HB8 make all RL biosynthetic intermediates under the described growth conditions. This approach could be further applied successfully for the structural elucidation of biosurfactants producing from the substrate mixture of others sources of carbon.

The detailed and extensive RLs characterization of this work demonstrated that a plethora of diversity of mono- and di-RLs were identified differing in the length of the lipidic chain, which additionally found to be saturated or unsaturated in some cases. However, the diversity of β-hydroxyalkanoic acids that are participated in the PHAs synthesized by T. thermophilus was not as wide as this of β-hydroxyalkanoic acids that was involved in RLs synthesis and not at all so similar. Moreover, it is known that the composition of PHAs depends on substrate specificity of the PHA polymerases which uses (R)-3-hydroxyalkanoyl-CoA as substrates (Lee et al. 2004). This observation directs to the bold suggestion that RLs production might constitutes a mechanism of exhaustion of the accumulated β-hydroxyalkanoic acids unused "wastes" by-products from the cells to the exterior medium, since in this bacterium anyhow the RLs production began early but is completed at the onset of the stationary phase.

Secreted RLs are recorded as virulence factors of the producing microorganisms (Abdel-Mawgoud et al. 2010). Among the RLs constituents, there is evidence that attributes the implication of the lipidic chain to triggering of some biological effects such as cytotoxicity or heamolytic activity. RLs possessing the same glycoside chain can also have a different influence when the aglycone is only slightly modified. The hydrophobic aglycone backbone could probably intercalate into the hydrophobic membrane bi-layer, while the glycoside chain could interact with the polar group (Chwalek et al. 2006).

Surface-active compounds commonly used by industry are chemically synthesized, and their replacement with bio-surfactants could provide advantages. Bio-surfactants compared to chemically synthesized surfactants have been receiving increasing attention as a result of their unique properties, mild production conditions, lower environmental toxicity, and higher biodegradability (Mulligan 2005; Rosenberg and Ron 1999). However, the use of bio-surfactants has been limited currently due to their relatively high production cost since the main factor limiting commercialization of bio-surfactants is associated with non-economical large scale production. To surmount this obstacle and to compete with synthetic surfactants, inexpensive substrates and effective microorganisms must be developed for bio-surfactant production. Agro-industrial wastes are considered as the most promising substrates for low-cost bio-surfactant production and can alleviate many processing industrial waste management problems. This would lead to the greater possibility for economical bio-surfactant production and reduced pollution caused by those wastes (Maneerat 2005). Aforementioned, many microorganisms, especially Pseudomonads, were reported use various low-cost renewable resources as potential carbon sources, but they cannot produce bio-surfactants or only have very low yield. Screening of high RLs-producing microorganisms from the natural environment is a good choice, for RLs production.

These efficient cultivation processes and their mechanisms as they pertain to the simultaneous production of PHAs and RLs encourage us to the use of a range of inexpensive carbon sources containing sugars or fatty acids. The establishment of lab-scale bioreactor experiments that will design in a future project that will include more factors affecting both biosynthetic processes might significantly increase the production of both bio-products, PHAs and RLs.