Currently, crude glycerol is generated in excess as an inevitable by-product from biodiesel industries ( da Silva et al., 2009). Crude glycerol is a very inexpensive (<US$200 per ton) and good carbon source from which many biochemicals and biofuels can be produced (Dharmadi et al., 2006). Accordingly, the use of glycerol as a carbon source can add profitability and sustainability to the biodiesel industry (da Silva et al., 2009; Jiang et al., 2009; Saxena et al., 2009). The production of 3-HP from glycerol has been extensively studied in recent years. The maximum theoretical yield (mol/mol) of 3-HP from glycerol is 1.0.
14.4.1 Glycerol Metabolism for the Production of 3-HP and Cell Growth
Glycerol is more reduced than traditional carbohydrates such as glucose, xylose, and so on. Due to its reduced nature, conversion of glycerol to pyruvate generates twice the amount of reducing equivalents than conversion of glucose or xylose. Therefore, the microorganisms that can grow on glycerol under anaerobic conditions, including Citrobacter,Clostridium,Enterobacter,Klebsiella, andLactobacillusspecies, require disposal of the two extra hydrogen atoms, which necessitates the formation of a by-product such as PDO to act as an electron sink. The production of PDO from glycerol proceeds via an aldehyde, 3-HPA, which can also be converted to 3-HP by proper aldehyde dehydrogenase(s) as observed in some strains ofKlebsiellaand Lactobacillus(Johnson and Lin, 1987; Nakamura and Whited, 2003; Dharmadi et al., 2006; Ashok et al., 2011). If these strains are genetically modified to overexpress a proper aldehyde dehydrogenase and block the enzymes leading to the production of PDO, more 3-HP can be produced.
Biosynthetic pathways for the production of 3-HP from glycerol can be classi- fied as either CoA-dependent or CoA-independent depending on the mechanism for the conversion of 3-HPA to 3-HP. Both the pathways generate NADH during the conversion of glycerol to 3-HP and considered “energy surplus” pathways (Dugar and Stephanopoulos, 2011). In both pathways, glycerol is first dehydrated to 3-HPA by either glycerol dehydratase encoded by dhaB or diol dehydratase encoded by pduCDE, both of which are vitamin B12-dependent (Bobik et al., 1999; Ashok et al., 2011; Luo et al., 2012). Once 3-HPA is produced, it is further converted to 3-HP in a CoA-dependent or CoA-independent manner. In the CoA-dependent pathway, 3-HPA is oxidized to 3-HP-CoA followed by phosphorylation to 3-HP–phosphate and subsequent dephosphorylation to 3-HP (Luo et al., 2011a, 2012). The CoA- dependent pathway is closely related to the degradation pathway of 1,2-propanediol inSalmonella enterica. In this pathway, 1,2-propanediol is converted to propionalde- hyde followed by disproportionation to propanol and propionic acid. The dispro- portionation occurs through a sequence of reactions catalyzed by CoA-dependent propionaldehyde dehydrogenase (PduP), phosphotransacylase (PduL), propionate kinase (PduW), and propanol dehydrogenase (PduQ) (Figure 14.3). The conversion of propionaldehyde to propionyl-CoA requires NAD+ as a cofactor, whereas that of propionyl phosphate to propanol requires NADH. When propionate is produced instead of propanol, ATP is generated. All of the genes necessary for consumption of 1,2-propanediol byS. entericare found in the propanediol utilization (pdu) locus as a single cluster (Leal et al., 2003; Liu et al., 2007; Xue et al., 2008). Unlike the CoA- dependent pathway, the CoA-independent pathway directly oxidizes 3-HPA to 3-HP by NAD+-dependent aldehyde dehydrogenase (Figure 14.3) and does not produce ATP.
When glycerol is used as a carbon and energy source, it should be oxidized to generate PEP and pyruvate via an oxidative pathway (Figure 14.4). In many microor- ganisms that can grow on glycerol under anaerobic conditions, the oxidative pathway always appears along with another reductive pathway leading to the production of
FIGURE 14.3 Biochemical reactions showing production of 3-HP from glycerol by CoA- dependent and CoA-independent pathways (Xue et al., 2008; Ashok et al., 2011; Luo et al., 2011a).
PDO. The reductive pathway from which 3-HP is also derived is highly depen- dent on the oxidative pathway for the generation of ATP and maintenance of redox balance. Under fermentative conditions, the oxidative metabolism starts with the dehydrogenation of glycerol to dihydroxyacetone (DHA) by the NAD+-dependent enzyme glycerol dehydrogenase. DHA is then phosphorylated by the PEP/ATP- dependent dihydroxyacetone kinase to dihydroxyacetone phosphate (DHAP), which is subsequently funneled to glycolysis (Figure 14.4). In few microorganisms and the species belonging to theKlebsiellagenus also have an alternative oxidative pathway that normally operates under respiratory conditions. This branch contains glycerol kinase(s), which catalyze the reaction of glycerol tosn-glycerol-3-phosphate and glycerol-3-phosphate dehydrogenases for the conversion ofsn-glycerol-3-phosphate to DHAP. Further conversion of DHAP results in the formation of various organic acids and reduced metabolites such as acetic acid, lactic acid, succinic acid, ethanol, 2,3-butanediol, and formic acid (Ashok et al., 2011).
14.4.2 Synthesis of 3-HP from Glycerol Through the CoA-Dependent Pathway
The production of 3-HP by variousLactobacillusspecies (L.strain 208-A,L. reuteri, andL. collinoides),S. enteric, andKlebsiella pneumoniaerelies on a CoA-dependent pathway. Yasuda et al. (2007) showed that resting cells of the mutant strain ofL.
reuterilacking glycerol dehydrogenase resulted in the production of 7.2 g/L of PDO
FIGURE14.4Metabolicpathwaydepictingoxidativemetabolismofglycerol(Ashoketal.,2011;Kumaretal., 391
and 8.5 g/L of 3-HP. Luo et al. (2011a) cloned and characterized the CoA-dependent propionaldehyde dehydrogenase (PduP) of L. reuteri. The PduP enzyme showed broad substrate specificity for aliphatic aldehydes including 3-HPA and could use both NAD+and NADP+as a cofactor. Among the substrates tested, the highest spe- cific activity was noticed with propionaldehyde. Specifically, theKmandVmaxvalues were 1.18 mM and 0.35 U/mg, respectively, at pH 7.8 and 37◦C.Klebsiella pnuemo- niaecan also naturally produce 3-HP from glycerol and has been shown to possess thepduoperon (Luo et al., 2012). To verify the role of PduP, a mutantK. pneumo- niaelacking aldehyde dehydrogenase (PuuC) was developed (Luo et al., 2011b). The mutant strain could produce 3-HP to the same extent as the wild type, indicating that the metabolic pathway for 3-HP synthesis was unaltered. Furthermore, overexpres- sion ofpduPenhanced 3-HP production, confirming that PduP is involved in 3-HP production inK. pneumoniae. In their next study, Luo et al. (2012) constructed apduP deletion mutant strain ofK. pneumoniae. The growth achieved by both the wild-type and mutant strains was almost the same until 15 hours, but the cell viability of the mutant strain declined remarkably after 15 hours. The mutant strain produced about 1.0 g/L 3-HP in 15 hours without any further production, while the wild-type strain continued to synthesize about 2.0 g/L 3-HP in 25 hours. The low 3-HP titer and cell death in the case of the mutant strain was attributed to the accumulation of cytotoxic intermediate metabolite 3-HPA (∼6 mM). Upon homologous overexpression ofpduP in the mutant strain, the level of 3-HPA was reduced and normal glycerol metabolism and cell growth were restored. These studies clearly suggest that, in natural strains,K.
pneumoniaeutilizes the CoA-dependent pathway to produce 3-HP from glycerol and PduP plays a vital role in 3-HP production. However, in other studies,K. pneumoniae was shown to efficiently produce 3-HP from glycerol via the CoA-independent path- way when a proper ALDH such as PuuC was overexpressed (see section 14.4.3). At present, 3-HP production by the CoA-dependent pathway is in the preliminary phase and far from commercialization. More extensive studies are needed to understand the detailed catalytic mechanism underlying 3-HP synthesis through PduP.
14.4.3 Synthesis of 3-HP From Glycerol Through the CoA-Independent Pathway
The CoA-independent production of 3-HP from glycerol can be accomplished via a two-step reaction catalyzed by glycerol dehydratase (GDHt) and aldehyde dehydro- genase (ALDH). Two types of strains, a nonnatural (E. coli) and a natural producer (K. pneumonia) of the target product have been investigated for CoA-independent 3-HP production.Klebsiella pneumoniaehas all of the genes for GDHt (e.g., DhaB) and ALDH, which are required for the conversion of glycerol to 3-HP. In addition, K. pneumoniaecan naturally produce coenzyme B12, which is an essential cofactor for DhaB. However, the level and/or mode of expression of the relevant genes for the CoA-independent pathway, especially ALDH, in wild-typeK. pneumoniaeare not appropriate for the optimal production of 3-HP. Conversely,E. colicannot produce coenzyme B12and does not have the genes for GDHt. Additionally, the expression level of 3-HPA-specific ALDH is low in wild-typeE. coli. However, E. coli has well-developed genetic tool boxes and is easy to culture.
14.4.3.1 Escherichia coli Suthers and Cameron (2005) reported the construc- tion of a recombinant E. coli by heterologously overexpressing the dhaB of K.
pneumoniaeand an ALDH ofSaccharomyces cerevisiae. The developed organism could produce 3-HP from glycerol at 0.1 g/L. Raj et al. (2008) screened a highly active ALDH enzyme (AldH) for the conversion of 3-HPA to 3-HP from E. coli K-12 MG1655. They developed the recombinantE. coliSH254 by introducing AldH (pCDF/aldH) along with DhaB (pBAD/dhaB) ofK. pneumoniaeDSM 2026 intoE.
coli BL21 DE3.When cultivated in M9 minimal medium supplemented with vita- min B12, strain SH254 could produce 3-HP at 0.58 g/L in flask culture and 2.5 g/L in a bioreactor experiment, respectively. However, the specific activities of the two enzymes in strain SH254 declined with time and the 3-HP production rate decreased in the later period. Furthermore, the imbalance between the activities of two enzymes (DhaB and AldH) caused intracellular accumulation of toxic 3-HPA. To address these problems, they developed a new recombinant strain,E. coliSH-BGA1, by cloning aldHin a high-copy plasmid and dhaBin a low-copy plasmid (Raj et al., 2009).
They also introduced glycerol dehydratase reactivase (GdrAB), which is known to improve the stability of DhaB by replacing damaged coenzyme B12molecules with new intact molecules during the dehydration reaction of glycerol (Mori et al., 1997).
Furthermore, they briefly optimized fermentation conditions such as pH, Isopropyl β-D-1-thiogalactopyranoside (IPTG) concentration, liquid-to-flask volume ratio, and substrate concentration. As a result, they could produce 4.4 g/L 3-HP at the flask level in 48 hours and 31 g/L 3-HP in a glycerol fed-batch bioreactor culture in 72 hours with a molar yield of 0.35 on glycerol. To further improve 3-HP production, they constructed another recombinant strain,E. coliSH-BGK1, by cloningα-ketoglutaric semialdehyde dehydrogenase (KGSADH) of Azospirillum brasilense in place of AldH of the recombinantE. coliSH-BGA1 (Rathnasingh et al., 2009). In its purified form, KGSADH (16.6±1.43 U/mg protein) was less active than AldH (38.10±0.76 U/mg protein) with 3-HPA as substrate. However, the AldH activity in crude cell extract of SH-BGK1 (1.94±0.15 U/mg protein) was higher than that of SH-BGA1 (1.56±0.08 U/mg protein). The higher ALDH activity in SH-BGK1 was expected to reduce the accumulation of toxic intermediate 3-HPA during the conversion of glycerol to 3-HP. In a glycerol fed-batch bioreactor experiment, strain SH-BGK1 produced 38.7 g/L 3-HP in 72 hours, which is the highest titer reported to date using glycerol as substrate. Recently, Kwak et al. (2012) developed a recombinantE. coliby expressing DhaB and DhaR (reactivation factor of glycerol dehydratase) fromLacto- bacillus brevisand AldH fromE. coli. The fed-batch cultivation of this recombinant strain with a two-step feeding strategy produced 14.3 g/L 3-HP under aerobic condi- tions (dissolved oxygen>10%). Acetate (4.2 g/L) and PDO (3.9 g/L) were the main by-products. Initially the cells were grown on glucose and once the culture optical density (OD) reached 100, the cells were fed with a mixture of glucose and glycerol.
14.4.3.2 Klebsiella pneumoniae Klebsiella pneumoniae has a de novo biosynthetic pathway for coenzyme B12 and well-developed metabolic pathways for glycerol assimilation (Ashok et al., 2011; Celi´nska, 2011). To divert 3-HPA to 3-HP in K. pneumoniae, oxidoreductases converting 3-HPA to PDO should be disrupted and an ALDH catalyzing the reaction of 3-HPA to 3-HP should be
overexpressed. However, the conversion of 3-HPA to 3-HP is always accompanied with NADH generation (Jo et al., 2008); therefore, continuous regeneration of NAD+ is critically important for uninterrupted 3-HP production. The regeneration of NAD+ from NADH can be accomplished most efficiently via the electron transport chain (ETC) under fully aerobic conditions (Richardson, 2000). However, there are several concerns about adopting aerobic conditions for 3-HP production withK. pneumoniae.
First, in the presence of oxygen, the production of coenzyme B12is significantly sup- pressed (Keuth and Bisping, 1994; Ye et al., 1996). In addition, the DhaB enzyme is inactivated by oxygen (Xu et al., 2009), and the entire reductive pathway for glycerol metabolism is downregulated. Therefore, the use ofK. pneumoniaeas a host strain for 3-HP production is challenging despite the important advantage that the strain can naturally synthesize vitamin B12.
In one study, Ashok et al. (2013a) developed recombinantK. pneumoniaeby (i) overexpressingγglutamyl-γ-aminobutyraldehyde dehydrogenase (PuuC) ofK. pneu- moniaeand (ii) disrupting two major oxidoreductases, DhaT and YqhD. PuuC, similar to AldH and KGSADH, was reported to have a high enzymatic activity for the conver- sion of 3-HPA to 3-HP, with NAD+as a cofactor (Raj et al., 2010). The recombinantK.
pneumoniaeΔdhaTΔyqhD(PuuC) could produce∼3.6 g/L 3-HP under microaerobic conditions in 12 hours of flask culture. However, when aeration was either increased or decreased, 3-HP production was seriously decreased due to insufficient production of coenzyme B12(under high aeration) or reduced regeneration of NAD+(under low aeration). These findings demonstrate that maintaining proper aeration is critical to successful production of 3-HP withK. pneumoniae. In glycerol fed-batch bioreactor cultivation under proper aeration conditions, the strain could produce 3-HP at>28 g/L in 48 hours with a glycerol carbon yield>40%. In another study, Ashok et al.
(2013b) attempted 3-HP production under anaerobic conditions in the presence of nitrate as an alternative electron acceptor to regenerate NAD+. The addition of nitrate facilitated NAD+regeneration and reduced the intracellular NADH level. However, the presence of this external electron acceptor decreased DhaB activity and 3-HP production. High carbon flux through the anaerobic respiratory pathway initiated by glycerol kinase (GlpK) was also observed in the presence of nitrate (Figure 14.4).
To reduce carbon flux through the oxidative pathway and improve 3-HP production, glpKanddhaTwere eliminated. The resulting strain showed 3-HP production of 22 g/L in 42 hours in an anaerobic bioreactor experiment in the presence of nitrate. Good cell growth and improved intracellular NAD+levels were also observed.
14.4.4 Coproduction of 3-HP and PDO From Glycerol
InK. pneumoniae, 3-HP and PDO are derived from the same intermediate, 3-HPA. The fate of 3-HPA depends on the activity of relevant enzymes (oxidoreductases and 3- HPA-specific ALDH) and the availability of redox cofactors (NAD+and NADH). If a single product, either 3-HP or PDO, should be produced at high yield, one of the other types of enzymes should be disrupted and one kind of cofactor should be regenerated efficiently. When compared to gene disruptions, cofactor regeneration is much more challenging. In K. pneumoniae, cofactor regeneration is performed by oxidative
metabolism of glycerol and, to this end, the carbon flow through oxidative pathway should become very high. One approach to deal with this redox balance problem is to produce both 3-HP and PDO together. If NADH or NAD+ can be recycled within the reductive pathway by producing 3-HP and PDO together, the carbon flow through the oxidative pathway is reduced and a higher carbon yield of coproduction can be achieved. Furthermore, since anaerobic or microaerobic conditions can be employed for this coproduction, the problems associated with the expression of genes for vitamin B12 (cob) and glycerol assimilation (dha) inK. pneumoniaecan be substantially alleviated. Several groups have adopted this approach and reported successful results withK. pneumoniae.
Zhu et al. (2009) developed a recombinant K. pneumoniae by introducing an ALDH fromE. coli. The recombinant strain was grown under three different aera- tion conditions and the highest amount of 3-HP (2.8 g/L) and PDO (9.8 g/L) was achieved under microaerobic conditions. Ashok et al. (2011) examined coproduc- tion using recombinantK. pneumoniaeDSM 2026 overexpressing PuuC and lacking DhaT. The recombinant strain showed greatly improved DhaB activity; thus, the car- bon flux through DhaB-mediated pathway was high. The removal of DhaT did not eliminate PDO production, but instead resulted in balanced production of 3-HP and PDO. These findings suggested that NADPH-dependent YqhD was responsible for PDO production in this recombinant strain. In one flask experiment conducted under microaerobic conditions with an initial glycerol concentration of 100 mM, equimolar 3-HP and PDO (40 mM each) were obtained in 24 hours. The overall yield of 3-HP and PDO was >80% and the production of by-products such as ethanol, acetate, and lactate was negligible. The fed-batch bioreactor cultivation of the same recombi- nant strain produced both compounds at a similar level (3-HP, 16.0 g/L; PDO, 16.8 g/L) within 24 hours. However, the cumulative yield of 3-HP and PDO on glycerol (mol/mol) was as low as 51% due to production of a significant amount of by-product acids and alcohols. In a coproduction study, Huang et al. (2012) investigated the effects of various AldH on the coproduction of 3-HP and PDO inK. pneumoniae and found thatγ-glutamyl-γ-aminobutyraldehydedehydrogenase ofE. coliwas most efficient. Bioreactor cultivation of the recombinantK. pneumoniaeoverexpressing this enzyme yielded 3-HP and PDO at 24.4 and 49.3 g/L, respectively, in 24 hours under anaerobic conditions. The cumulative molar yield of the two metabolites was 0.61 (0.18 on 3-HP and 0.43 on PDO). In their next study, the recombinant strain was cultivated under anaerobic, microaerobic, and aerobic conditions. They found that the concentrations and yields (mol/mol) of 3-HP improved continuously with increase in aeration rate under microaerobic conditions (Huang et al., 2013). The maximum titer of 3-HP (48.9 g/L) was achieved at an aeration rate of 1.5 vvm along with 25.3 g/L PDO and 24.9 g/L lactic acid in 28 hours. The molar yields of 3-HP, PDO, and lactate were 0.41, 0.25, and 0.22, respectively. The concentration (48.9 g/L) and volumetric productivity (1.75 g/L⋅h) of 3-HP achieved in this investigation were the highest reported to date.
Although Ashok et al. (2011) and Huang et al. (2012) reported successful copro- duction of 3-HP and PDO from glycerol by growing cells ofK. pneumoniae, a signif- icant amount of organic acids and alcohols were produced as by-products. To avoid
the formation of by-products and improve the yield, Kumar et al. (2012b) employed resting cells of recombinantK. pneumoniaeJ2B strain overexpressingKGSADHfor the coproduction of 3-HP and PDO. They grew cells under microaerobic conditions to maximize the whole-cell 3-HP production. The cells were then harvested and used for 3-HP production under nongrowing conditions. They found that fed-batch bioconversion in a 1.5-L bioreactor with 1.0 g-CDW/L under anaerobic conditions resulted in 11.3 g/L 3-HP and 15.2 g/L PDO in 12 hours with a cumulative yield of 0.71 (mol/mol). In addition to 3-HP and PDO, significant amount of lactic acid was also accumulated. In the next study, they constructed another recombinant strain ofK. pneumoniaeJ2B by deleting lactate dehydrogenase (ldhA) and overexpressing KGSADH. The lactic acid production was completely abolished in this new recom- binant. In addition, the final titer of 3-HP and PDO improved to 22.7 and 23.5 g/L, respectively, within 60 hours and the cumulative product yield was enhanced to 0.77 (Kumar et al., 2013).