The screening of important cultural factors by Taguchi method shows that initial pH had the least effect on the production of ethanol from glycerol by E. aerogenes S012. The reason could be that the range of pH values tested was very close to the optimum value. Ito et al. (
2005) found pH 6.8 to be optimum for hydrogen and ethanol production by E. aerogenes HU-101, while Nakashimada et al. (
2002) found pH 7.0 and 6.3 best for growth and hydrogen production, respectively, by E. aerogenes AY-2 strain. Therefore, best pH for fermentation by E. aerogenes S012 is between 6.3 and 6.8.
The best cultural conditions were 38°C, 200 rpm and 72 h, as deduced from Eq.3. However, Figure
1 shows that the fermentation virtually ceased after 48 h. This confirms another deduction from Eq.3 which found that running the fermentation beyond 48 h may not increase product formation significantly. However, we believe that this trend could also be an indication that the presence or depletion of some factor(s) adversely interfered with the activity of the biocatalyst. Among the factors suspected were lowered medium pH below the organism’s tolerable limit by metabolites, accumulation of NADH, product (ethanol) inhibition, and depletion of nutrients. It is important to keep the range of tested cultural conditions within the tolerance limit of the organism. For instance, temperature should be between 20-45°C and agitation between 120–200 rpm. Too high agitation speed might rupture the cell wall of the bacteria leading to death. Also temperatures outside the stated range may not support the growth of the organism. It is important to note that the response values and deductions of Eq.3 were based on experiments with feedstock concentration of 30 g/l, and may be considered qualitative values.
It could be seen that trace amounts of succinate, lactate, and acetate were produced as byproducts of the fermentation. These probably lowered the pH of the medium to a value at which the organism could not function anymore. For each set of fermentation, the pH was measured after every 24 h. The pH was about 5.9 after 24 h; then about 5.5 after 48 h, and the value stayed almost unchanged after 72 h. Meanwhile the initial pH was 7.0 ± 0.2.
Most enteric bacteria possess type 1 NAD+-dependent glycerol dehydrogenase (GDH-1), and ferment glycerol by oxidative and reductive pathways. In the oxidative pathway, glycerol is converted to dihydroxyacetone phosphate (DHAP) by two successive enzymes, namely GDH-1 and DHA kinase (DHAK), and NAD+ is reduced to NADH. DHAP then enters glycolysis to produce pyruvate, which can be broken down to various products depending on the organism.
In the reductive pathway, however, glycerol is dehydrated into 3-hydroxypropionaldehyde, 3-HPA, by glycerol dehydratase (GD). Then a NADH-linked 1,3-Pdo dehydrogenase (1,3-PdoDH) reduces 3-HPA to 1,3-Pdo and NADH is reoxidized to NAD+. Biomass yield during glycerol dissimilation also requires NAD+ for oxidation reactions, giving rise to NADH. In other words, microbial cells can only grow in a redox-balanced process in which NAD+ is regenerated from NADH. Although E. coli does not produce 1,3-Pdo like other Enterobacteriaceae do because it possesses type 2 GDH (GDH-2), it produces 1,2-pdo, which also does the work of regenerating NAD+ (Gonzalez et al.
2008). It follows that during biosynthesis of ethanol from glycerol by cells that don’t produce 1,2- or 1,3-pdo, cell growth releases a surplus of NADH, which inhibits further growth and cell function if not reoxidized. Thus for cell growth or glycerol metabolism, NAD+ concentration needs to be higher than NADH.
It follows that the key function of 1,3-pdo is to regenerate NAD+ reduced in the oxidative pathway or during cell growth. Therefore, species which do not synthesize 1,3-pdo need an alternative source of electron sink to regenerate NAD+. This could be achieved either by utilizing oxygen or by using complex compounds in the medium in which case they don’t synthesize biomass, thereby avoiding the necessity of NAD+ regeneration (Choi et al.
2011). Gonzalez et al. (
2008) corroborated this when they reported that metabolism of glycerol in microbial species unable to synthesize 1,3-pdo takes place through a respiratory pathway that requires an external electron acceptor.
1 shows that E. aerogenes S012 produced neither 1,2-Pdo nor 1,3-Pdo at the experimental conditions, but gave small amounts of succinic, lactic, and acetic acids as the only organic coproducts. The screening experiments gave similar results (data not shown). Therefore, we assumed that there was an accumulation of NADH in the cells, which halted the assimilation of glycerol. Hence, there was need for an electron acceptor such as oxygen for regeneration of NAD+. Choi et al. (
2011) discovered that Kluyvera cryocrescens S26 produced neither 1,2- nor 1,3-Pdo. This corroborates our report that it is possible for bacterial cells to produce ethanol from glycerol without co-producing propanediol (Pdo). K. cryocrescens S26 also coproduced H2 and small amounts of formic, lactic and succinic acids as the only organic acids.
The result shown in Figure
2 confirmed that low amount of oxygen was required by E. aerogenes S012 for optimum yield of ethanol from glycerol. This was also supported by Choi et al. (
2011), Oh et al. (
2011), and Jitrwung and Yargeau (
2011). In their optimization research, Jitrwung and Yargeau (
2011) found that E. aerogenes ATCC 35029 required a low oxygen concentration, rather than oxygen free conditions, to optimally produce hydrogen and ethanol from glycerol. They reported that E. aerogenes requires and consumes externally-supplied gaseous oxygen before using oxygen from decomposable organic or inorganic compounds/salts in the medium.
Choi et al. (
2011) found that K. cryocrescens S26 was at its best in microaerobic condition. The bacterium produced 27 g/l ethanol with a productivity of 0.61 g/l/h at a low oxygen level. They reported that limited oxygen served as electron acceptor that consumed excess reducing equivalents (NADH) generated during biomass synthesis. However they noted that high oxygen supply would result in most of the carbon being integrated into cellular mass and converted to CO2, leading to poor metabolite formation. Choi et al. (
2011) also confirmed that high oxygen supply decreased ethanol production by K. cryocrescens S26, but increased acetic and lactic acid products as well as encouraged excessive cell growth. They concluded that oxygen plays a key role in a switch between biomass formation and ethanol production pathways. Excess oxygen favors the former while low oxygen the later. Oh et al. (
2011) arrived at similar conclusion: high aeration rate increased biomass accumulation but lowered ethanol production by Klebsiella pneumoniae GEM 167, while low aeration rate of 0.5 vvm (vessel volume/minute) gave highest ethanol production of 21.5 g/l and productivity of 0.93 g/l/h at 200 rpm and 37°C. In contrast to all these, however, Chen et al. (
2003) reported that microaerobic condition (air at 0.4vvm) favored more cell growth and less ethanol production by Klebsiella pneumoniae DSM 2026 rather than anaerobic condition (nitrogen at 0.4 vvm). They noted that complete absence of air was more desirable for ethanol production by K. pneumoniae DSM 2026.
From the foregoing, presence of oxygen can enhance or disrupt ethanol production by many glycerol-metabolizing facultative anaerobes depending on the amount of air supplied. High oxygen supply enhances the generation of ATP by reducing NADH. Then ATP is used for biomass synthesis. Thus high oxygen level drives more carbon flux towards biomass production. But in microaerobic conditions, the limited oxygen converts NADH produced during cell growth into NAD+ while maintaining carbon flux into ethanol production. Obviously, K. cryocrescens S26 (Choi et al.
2011), K. pneumoniae GEM 167 (Oh et al.
2011), E. aerogenes ATCC 35029 (Jitrwung & Yargeau
2011), and E. aerogenes S012 (current work) follow the microaerobic pathway.
Ethanol is generally known to inhibit the growth of microorganisms. Therefore, ethanol exerts a product inhibition on the microbial agents of its biosynthesis. Although Kluyvera cryocrescens S26 produced high amounts of ethanol from glycerol, Choi et al. (
2011) reported that 50 g/l ethanol in the medium reduced the growth of the bacterium by 80%, supporting the proposition. But our finding that E. aerogenes S012 survived (but did not grow) in a growth medium containing over 60 g/l ethanol suggests that ethanol product inhibition may not be the repressing factor to further metabolism of glycerol beyond 48 h.
The result shown in Figure
3 attested to the fact that excess air represses ethanol production. Just as Choi et al. (
2011) reported higher levels of acetic and lactic acids, this report also found predominance of acetic, lactic and succinic acids in high supply of oxygen. Similarly, Hong et al. (
2009) found that E. coli AC-521 produced 85.8 g/l lactic acid in 88 h under aerated fermentation.
The optimized fermentation (Figure
4) showed that best ethanol yield of 1.12 mol/mol-glycerol was produced after 48 h, although maximum productivity (0.74 g/l/h) was recorded after 24 h. We reported in a previous study (Nwachukwu et al.
2012) that the wild strain, E. aerogenes ATCC 13048, seemed to be at its best for utilizing glycerol within 48 h. The result in the current study showed that the mutant strain, E. aerogenes S012, retained this high ethanol yield (or glycerol-to-ethanol conversion efficiency) of the wild strain.
The better ethanol production observed in Figure
4 than the production observed in Figure
1 showed that low oxygen was required for optimum ethanol production by E. aerogenes S012. Ethanol yield of 1.12 mol/mol-glycerol was in excess of the theoretical maximum (1.0 mol/mol-glycerol). The reasons for this have been stated previously to be probably due to additional carbon source provided to the bacterium by the amino acids of the complex medium, or some unknown carbon/electron sources in the medium (Nwachukwu et al.
2012). Dharmadi et al. (
2006) and Murarka et al. (
2008) reported that E. coli needed to be supplemented with rich nutrients as tryptone or yeast extract in order to dissimilate glycerol. Yeast extract contains nitrogen and carbohydrates. Therefore, as Choi et al. (
2011) also noted, those carbohydrates may be utilized by the microbial cells as carbon source for ethanol production and biomass yield.
Several researchers have reported using recombinant species to ferment glycerol. Yang et al. (
2007) found that engineered Klebsiella oxitoca M5al produced 19.5 g/l ethanol at the rate of 0.56 g/l/h; Durnin et al. (
2009) reported a production of 20.7 g/l with a productivity of 0.22 g/l/h by a recombinant E. coli; and Oh et al. (
2011) used a genetically modified K. pneumoniae to produce 25 g/l ethanol from glycerol. Cheng et al. (
2007) reported that a wild K. pneumoniae M5al produced 18 g/l ethanol from pure glycerol, with a productivity of 0.28 g/l/h in a N2 gas-created anoxia. This work found that non engineered mutant, E. aerogenes S012, produced 25.4 g/l ethanol from pure glycerol in 48 h, with productivity of 0.53 g/l/h and yield of 1.12 mol/mol-glycerol. This means that 56.23% of the used glycerol was converted to ethanol.
It is well known that manipulation of culture parameters is a process-based improvement strategy to optimize product formation (Yazdani and Gonzalez,
2007), while detection of mutants, genetic operations, and metabolic pathway control are strain-based improvement strategies. The use of process-based protocols to optimize product formation is preferable to the use of recombinant DNA mechanisms. The reason is that commercial/industrial use of genetically modified (GM) organisms is restricted by government regulations through EPA and USDA (D. Glass Associates Inc
2012). D. Glass Associates Inc (
2012)) reported that biofuels companies may experience significant interference and restrictions due to the government regulations on the testing and commercial use of GM microorganisms, algae, or transgenic plants being developed for biofuels applications. Although the GM products could be more efficient for commercial applications, Bell and Attfield (
2009) noted that these regulatory constraints can increase the capital and running costs of biofuels production. Therefore, improving biological catalysts of bioethanol production by measures that do not include genetic engineering is very desirable.
This work used TSB for all fermentations. According to Gullapalli et al. (
2007), the type of medium used in fermentation greatly influences the type and quantity of products formed. Choi et al. (
2011) found that yeast extract was best for biomass accumulation and ethanol yield by K. cryocrescens S26. However, Gullapalli et al. (
2007), in their bioproduction of D-psicose by E. aerogenes, discovered that the highest transformation rate was obtained when tryptic soy broth (TSB) was the fermentation broth. In a preliminary study, Banna and Ouro (2008, unpublished) used Luria Bertani (LB) broth supplemented with tryptone and yeast extract as fermentation medium for E. aerogenes. They found that less than 10 g/l glycerol was utilized after 48 h, producing about 5 g/l ethanol. Although this represented a yield equivalent to theoretical maximum of 1 mol/mol-glycerol, the ethanol production, ethanol productivity, and glycerol utilization were very poor. The poor fermentation was probably due to the type of medium used or excess air in the system.
In conclusion, Enterobacter aerogenes S012 promises to be good and efficient biocatalyst for conversion of glycerol to ethanol. It produced 25.4 g/l ethanol within 48 h at optimum temperature and agitation speed of 38°C and 200 rpm, respectively, under a low amount of oxygen. It also demonstrated an ability to produce fewer co-products at trace concentrations and complete absence of 1,3-Pdo. This will make ethanol extraction/purification easier and more economically efficient. Previously, the wild strain, E. aerogenes ATCC 13048, was reported to be effective at converting low concentrations of recovered glycerol into ethanol (Nwachukwu et al.
2012). Therefore, we believe that the mutant strain, E. aerogenes S012, will be even more effective in converting higher concentrations of recovered glycerol into ethanol. Further work is necessary to verify this assumption. Moreover, the ability of this E. aerogenes S012 to convert glycerol to ethanol at effectiveness of 1 mol ethanol/mol glycerol in high glycerol-containing medium is potentially of great importance to biofuels industry. This ability makes the invention a potentially excellent biocatalyst for industrial conversion of glycerol to ethanol. Since the technology involves using biocatalyst developed by adaptive mutation, it is not subject to government regulations of GM products.
It was also observed in this work that the pH of the fermentation medium dropped to a value below 5.6 from the initial value of 7.0 ± 0.2 after 48 h. Additional work is necessary to confirm that E. aerogenes S012 is at its best for ethanol production at pH 6.3-6.8. Finally, it has been reported that hydrogen production is associated with ethanol production during microbial glycerol metabolism (Varrone et al.
2012). Choi et al. (
2011) and Ito et al. (
2005) also corroborated this during the fermentation of crude glycerol by the bacteria K. cryocrescens S26 and E. aerogenes HU-101, respectively. Therefore, further research is necessary to determine the amount of hydrogen produced in the glycerol metabolism by E. aerogenes S012.