|Home | About | Journals | Submit | Contact Us | Français|
Fatty acid ethyl esters (FAEEs) were produced in this study by the use of an engineered Escherichia coli p(Microdiesel) strain. Four fed-batch pilot scale cultivations were carried out by first using glycerol as sole carbon source for biomass production before glucose and oleic acid were added as carbon sources. Cultivations yielded a cell density of up to 61 ± 3.1 g of cell dry mass (CDM) per liter and a maximal FAEE content of 25.4% ± 1.1% (wt/wt) of CDM.
Fuel consisting of methyl esters of fatty acids (FAMEs) and ethyl esters of fatty acids (FAEEs), also referred to as “biodiesel,” is made from renewable farm products such as vegetable or animal oils, fats, or recycled cooking greases that are transesterified using methanol derived from fossil fuel or bioethanol from renewable resources (22, 27). Fatty acids from animal and plant origin (typically C8 to C22) are not good fuels, but transestrification of the fatty acids from the glycerol triesters to methanol or ethanol yields molecules that are excellent fuels for diesel engines (15). Biodiesel possesses several environment-friendly features: it is biodegradable and nontoxic, it exhibits low sulfur content and an absence of aromatic compounds, and it can be produced mostly from renewable resources. FAMEs and FAEEs have similar characteristics with respect to chemical and physical fuel properties and engine performance (32, 34).
In 2006, Kalscheuer et al. (20) introduced the term microdiesel, which could be defined as the ethyl esters of a mixture of fatty acids (FAEE) that are synthesized in an engineered Escherichia coli strain harboring the plasmid p(Microdiesel), which encodes two functions: the generation of ethanol from carbohydrates like glucose and the catalysis of the transesterification reaction yielding ethyl esters.
E. coli does not produce FAEEs by its natural metabolism, and it synthesizes ethanol only anaerobically, among several other fermentation products from acetyl coenzyme A (acetyl-CoA), via two sequential NADH-dependent reductions catalyzed by AdhE, a multifunctional alcohol dehydrogenase (13, 21). Thus, genes encoding the NADH-oxidizing system from Zymomonas mobilis have been heterologously expressed in E. coli (1, 16, 20). This NADH oxidizing pathway consists of the pyruvate decarboxylase (PDC) (EC 220.127.116.11) catalyzing the nonoxidative decarboxylation of pyruvate to acetaldehyde plus carbon dioxide and of two alcohol dehydrogenase isozymes (ADH) (EC 18.104.22.168) catalyzing the NADH-dependent reduction of acetaldehyde to ethanol (28, 47). Ethanol is produced under aerobic conditions in substantial amounts that are sufficient for the transesterification process to yield enough FAEEs. An acyltransferase (encoded by atfA) is the key enzyme of wax esters and of triacyl glycerol biosynthesis in Acinetobacter baylyi strain ADP1 and other bacteria (17). It exhibits an extremely low specificity for the acyl acceptor molecule and also for the alcohol donor (18, 19, 39, 44).
Fermentation economics depend mainly on investment costs for the fermentation plant, raw materials, and maintenance plus costs for upstream and downstream processes as well as the process yield and throughput (31). Since FAEEs are intracellularly synthesized, the amount of biomass produced is an important factor that limits the amounts of FAEEs produced during cultivations. Therefore, suitable substrates and economically feasible downstream processes are of great importance in microdiesel production. The objectives of this study were therefore to improve the process previously described (20) and to investigate the application of glycerol as a cheap carbon source for pilot-scale cultivation of E. coli harboring p(Microdiesel).
Cells of the recombinant E. coli p(Microdiesel) strain were cultivated in LB medium for preculture and subsequently in modified M9 mineral salts medium (35) containing 0.1 M IPTG (isopropyl-β-d-thiogalactopyranoside) and 100 μg/ml ampicillin. High-cell-density cultivations were done in a Biostat UD30 stainless steel reactor (B. Braun Biotech International, Melsungen, Germany). The bioreactor was filled with 20 or 16 liters of mineral salts medium, and the pH value was set to 7.0 before in situ sterilization and during cultivation by the addition of 4 M HCl or 3.2 M oleic acid and 4 M NaOH or 4 M NH4OH. Fermentations were carried out at 30°C with a pO2 between 20 and 30% saturation for batch and fed-batch cultivations, respectively. Agitation and aeration rates were adjusted between 100 and 800 rpm or 0.8 and 1.0 volume per volume and min (vvm), respectively. Foam was controlled by a mechanical foam destroyer and by the addition of a Wacker Silikon Antischaum emulsion SLE antifoam agent (Darwin Vertriebs GmbH, Ottobrunn, Germany), when necessary. Acetate, glycerol, ethanol, and other metabolites were analyzed in the cell-free aqueous culture supernatants by the use of a LaChrom Elite high-performance liquid chromatography (HPLC) apparatus (VWR-Hitachi International GmbH, Darmstadt, Germany). At the end of cultivation, cells were harvested by centrifugation at 4°C in a CEPA type Z41 or type Z61 continuous centrifuge (Carl Padberg Zentrifugenbau GmbH, Lahr, Germany). Harvested cells were frozen at −30°C and then lyophilized (Beta 1-16; Martin Christ GmbH, Osterode, Germany).
E. coli Top10 harboring p(Microdiesel) was cultivated in mineral salts medium containing different concentrations of glycerol, glucose, and sodium oleate in Erlenmeyer flasks. The microscopic examination of cells showed the accumulation of refractile bodies inside the cells; the presence of lipids was confirmed by thin-layer chromatography (TLC) analysis (17). Glycerol was used in all subsequent cultivations, partly because of its low price; glycerol of sufficient quality is obtained in some industries such as biodiesel and soap production as a residual product (27).
The first fed-batch cultivation experiment resulted in a cell density of 32 ± 2.6 g/liter after 33 h of cultivation (Table (Table1).1). The growth rate was 0.37 cells h−1 during the batch phase of cultivation. Feeding of glucose and oleic acid started after 14 h of incubation. After 15 h, the formation of the proteins encoded by p(Microdiesel) was induced by the addition of IPTG. The maximum concentration of ethanol of 1.0% (vol/vol) in the medium was reached at the end of the cultivation; cells grew at a rate of 0.17 cells h−1 in the fed-batch phase of growth. However, most glycerol was consumed at the end of this cultivation. FAEEs were analyzed by TLC, gas chromatography (GC), and coupled GC-mass spectrometry (GC-MS) using the NIST mass spectral search program (Windows software, version 1.6d) as described elsewhere (20). The highest FAEE content obtained in this cultivation was 25.4% ± 1.1% (wt/wt) of cell dry mass (CDM) after 29 h of cultivation. Microscopic examination of cells incubated with Nil red showed the absence of microdiesel bodies before induction (Fig. (Fig.11 A) but the presence of fluorescent bodies after 19 h (Fig. (Fig.1B)1B) and at the end of the cultivation period (Fig. (Fig.1C1C).
The second fed-batch cultivation started with a filling of 16 liters of mineral salts medium in the bioreactor, and the cultivation period was extended to 54 h. Instead of oleic acid and sodium oleate, which were used in the previous fed-batch cultivation, now only sodium oleate was fed as source of fatty acids. Due to the toxicity of sodium oleate at higher concentrations, a 10% (wt/vol) stock solution was used as feed (Table (Table1).1). The growth rate was 0.39 cells h−1 during the batch phase and 0.18 cells h−1 during the fed-batch phase. A total of 45 ± 1.2 g of CDM/liter was obtained after 54 h. In this cultivation, a cell density about 41% higher than that seen with the previous cultivation was obtained. The amounts of microdiesel reached 14.4% ± 0.4% (wt/wt) of CDM after 38 h and did not significantly change until the end of the cultivation period.
To obtain cell density and FAEE content volumes higher than those seen in previous cultivations, the cultivation period was reduced to 47 h, and the total amounts of glucose and glycerol used for feeding were increased about 3-fold (Table (Table1).1). A cell density of 61 ± 3.1 g of CDM/liter was obtained, which corresponds to 1.4- and 2-fold increases in CDM in comparison to the second cultivation (45 g of CDM/liter) and the first cultivation (32 g of CDM/liter), respectively. The growth rate during the batch phase of this cultivation was 0.34 cells h−1; it decreased to 0.20 cells h−1 during the fed-batch phase. The acetate concentration in the medium was 1.2% (wt/vol) after 35 h and might have limited cell growth. Later, acetate was consumed, and its concentration decreased to only 0.2% (wt/vol) at the end of the cultivation period. Glucose and glycerol were not detected in the medium at the end of the cultivation period, whereas ethanol was detected at a concentration of 0.26% (vol/vol). Lipid analysis revealed a FAEE concentration of 18.5% ± 0.9% (wt/wt) at the end of cultivation, which is 22% higher than the level seen in the previous cultivation.
In the fourth cultivation, only sodium oleate was fed as a fatty acid source as in the second cultivation; however, its amount had been increased by 25% to obtain enhanced FAEE contents of cells (Table (Table1).1). After 66 h of cultivation, 54.2 ± 2.0 g of CDM/liter was obtained (Fig. (Fig.2),2), which was 20% (wt/wt) higher than the amount obtained in the second cultivation (45 ± 1.2 g of CDM/liter) which was similar to the fourth cultivation in having been fed with sodium oleate as a source of fatty acids. Glycerol and ethanol were detected at concentrations of 4% and 2% (vol/vol), respectively, at the end of the cultivation period (Fig. (Fig.2).2). Cells grew at a rate of 0.38 h−1 during the batch phase and of 0.11 h−1 during the fed-batch phase of cultivation. The cells accumulated FAEE to 23.4% ± 0.4% (wt/wt) as the maximum after 45 h. However, the accumulation level dropped to 20.4% ± 0.9% before cell harvest, as shown in Fig. Fig.2.2. Increases in FAEE content of 30% and 9% were obtained in this cultivation in comparison to the second and third fed-batch cultivations, respectively.
In the first cultivation, about 0.7 kg of CDM was obtained after 33 h, containing 8.9 g/liter FAEEs (0.27 g/liter/h), when cells were fed with oleic acid and sodium oleate. Furthermore, the second cultivation resulted in 1.2 kg of CDM in only 54 h, containing FAEE at up to 10.8 g/liter (0.2 g/liter/h). The maximum amount of CDM obtained in this study was 1.6 kg after 47 h of cultivation, with about 14.8 g/liter FAEE (0.32 g/liter/h), when cells were fed with sodium oleate. The last fed-batch cultivation resulted in 1.5 kg of CDM after 66 h, containing about 19 g/liter FAEE (0.29 g/liter/h). Oleate was converted to FAEE with efficiencies of 48.6, 54.0, 39.1, and 74.6% on a weight basis during cultivations 1, 2, 3, and 4, respectively.
During this study, acetone and chloroform were used for extraction of microdiesel from the cells in both the small-scale and the pilot-scale experiments. Acetone was as effective as chloroform in extraction and is preferred because it is cheaper and relatively safe and is even available from renewable sources as a byproduct of microbial fermentations (8). TLC was used for routine analysis to investigate the components of the microdiesel produced in each of the cultivation experiments. Ethyl oleate and ethyl palmitate were used as standard ethyl esters of fatty acids in addition to oleic acid as a reference for nonesterified fatty acids (Fig. (Fig.3,3, lanes EO, EP, and A). A sample from the first cultivation (Fig. (Fig.3,3, lane B) showed that two spots seemed to represent ethyl esters of oleate and palmitate, in addition, spots that appeared to represent oleic acid and triolein were seen. Samples from cultivations 2, 3, and 4 (lanes C, D, and E in Fig. Fig.3)3) exhibited the same constituents, except for the sample separated in lane E, which was withdrawn after a prolonged incubation period of the cells in cultivation 4. Here, triolein was obviously partially consumed and more FAEEs could be detected, as indicated by the intensity of the spot, taking into account the fact that all samples were treated identically.
Furthermore, although only at low concentrations, FAEEs were also detected extracellularly in the cultivation broth, as shown in Fig. Fig.33 (lane E). Spots illustrated in lanes G and H of Fig. Fig.33 were obtained from samples extracted with acetone and a sample of commercial rapeseed biodiesel, respectively. Our samples and biodiesel exhibited almost the same retention time in TLC analyses. A vial containing a larger microdiesel sample is shown in lane I of Fig. Fig.3.3. This sample was prepared by stirring 25 g of freeze-dried cells of E. coli p(Microdiesel) into 400 ml of chloroform. Cells were then removed by filtration, and chloroform was removed by evaporation. The residual lipid constituted 18.3% ± 1.2% (wt/wt) of the dry matter of the cells, as analyzed by gas chromatography (data not shown). Analyses by gas chromatography (GC) and by GC coupled to mass spectrometry revealed the presence of C10:0, C11:0, C12:0, and C14:0 ethyl esters in addition to C16:0, C16:1, C18:0, C18:1, and C20:0 ethyl esters (data not shown).
Microdiesel could be an attractive alternative bioenergy source and could contribute to the energy market as a substitute for diesel in addition to conventional biodiesel. However, a broader use of biodiesel and a more significant substitution of petroleum-based fuels in the future will be possible only if production processes are developed that are based not solely on oilseed crops but also on bulk plant materials such as cellulose. One advantage of microdiesel is that production could be based on the establishment of engineered strains which are able to synthesize large quantities of lipids as well as catalyzing the transesterification process. Another strategy published recently aimed at growing engineered E. coli strain on renewable substrates such as hemicelluloses (38). This would save edible lipid sources from being used in the bioenergy sector (23). An interesting aspect of this study was the use of glycerol as a carbon source, given that glycerol has previously been used for large-scale cultivation of recombinant E. coli strains to produce a variety of products (24, 42, 43, 46). The upsurge in biodiesel production has resulted in a market surplus of glycerol, a byproduct of the chemical transesterification process (27).
Bioethanol is of great importance in that it is itself a biofuel and is required for the transesterification process in microdiesel biosynthesis; many studies have reported on the optimization of its production (11, 41, 49). However, it has been also reported that ethanol causes multiple effects on E. coli K-12 cells, including reduction of growth rates and inhibition of cell division at concentrations higher than 0.6 M and loss of viability at a concentration of 1.6 M (12) and damage to the cell membrane and degradation of ribonucleic acids (36). The maximum ethanol concentration obtained in this study was 2% (vol/vol). This concentration corresponds to 15 g/liter and is three to five times higher than the concentrations obtained during aerobic and anaerobic cultivations in LB medium supplemented by 2% (wt/vol) glucose (20). In the third cultivation, the maximum ethanol concentration was 0.39% (vol/vol), which was the lowest among all cultivations, albeit the rate of growth (μ = 0.2) during the fed-batch phase of growth was the highest seen among all of the cultivations. This could explain why the third cultivation yielded a higher cell density and why less glucose was used.
Plasmid loss occurring during the fermentation process is a critical drawback in efforts to reach a high production level of the desired products. Two major reasons for plasmid loss are known: plasmid instability and depression of the growth rate of plasmid-bearing cells (6, 37). Plasmid instability is mainly due to the segregational plasmid loss during cell division that was observed in the second, third, and fourth cultivations, which had longer incubation periods than the first cultivation. Strategies to improve the stability of recombinant plasmids to achieve high density of cells without losing the ability of product formation are being investigated in our laboratory (25).
The dependence on feeding a source of lipid to cells synthesizing these FAEEs is considered the bottleneck of microdiesel production. This could be overcome by establishing the plasmid used in this study or a similar plasmid for production of microdiesel used in an oleaginous microorganism. Many wild-type bacterial strains produce neutral lipids, including Rhodococcus opacus (accumulating lipids at up to 87% [wt/wt] of CDM) (2, 3, 5), R. ruber (26% [wt/wt]) (5), Streptomyces lividans (125 mg/ml) (29), Pseudomonas aeruginosa strain 44T1 (38% [wt/wt]) (10), Acinetobacter sp. strain 211 (25% [wt/wt]) (4), and Gordonia sp. (up to 80% [wt/wt]) (14). A strain of E. coli modified to overproduce fatty acids by engineering the fatty acid de novo biosynthesis pathway and fatty acid degradation was also previously reported (26, 38). Moreover, expression of a plant thioesterase could increase the abundance of shorter-chain fatty acids to improve fuel quality (9, 26, 38, 45, 48). On the other hand, certain fungi, such as Mortierella ramanniana and species of the genera Lipomyces and Rhodotorula, accumulate substantial amounts of lipids (above 50% of CDM) (33), Debaryomyces hansenii accumulates up to 75% (wt/wt) lipids (7), and Yarrowia lipolytica produces 43% (wt/wt) lipids of CDM when cultivated on industrial glycerol (30). An endophytic fungus, Gliocladium roseum (NRRL 50072), produced a series of volatile hydrocarbons and hydrocarbon derivatives under microaerophilic conditions (40). Genes for mycodiesel production could also be expressed in some of the oleaginous fungi to establish processes independent from fatty acid or lipid feeding.
We are indebted to the Egyptian Ministry of Higher Education and Scientific Research (MHESR) and the German Academic Exchange Service (DAAD) for the GERSS (German-Egyptian Research Short-Term Scholarship) grant provided to Yasser Elbahloul.
Technical assistance by Herbert Ahlers, Ahmed Sallam, Mohamed Ibrahim, and Yasser Hafez is gratefully acknowledged.
Published ahead of print on 7 May 2010.