Use of Glycerol in Biotechnological Applications

Since decades the limited access to petroleum oil is a major concern and substitutions for fossil fuels are needed. One promising substitute is biodiesel, which is widely produced from vegetable oils, e.g. from rape seeds, soybeans, sunflower seeds or animal fats. In the synthesis of biodiesel, oils and fats are transesterified to fatty acid methyl ester in the presence of sodium hydroxide or potassium hydroxide. In this process, glycerol is generated as stoichiometric byproduct with a ratio of 10% (w/w) with respect to biodiesel produced. In 2009, the biodiesel production of the world reached 16 million tons (Licht, 2010), with the lion's share produced by the European Union with 9 million tons (EBB, European Biodiesel Board 2010), followed by the United States with a production of 2.7 million tons (Licht, 2010). Hence, 1.6 million tons of glycerol were produced as obligatory by-product. Glycerol finds applications as an ingredient of various products, such as creams, food, feed, and pharmaceuticals, but the demand for glycerol in these processes is limited. Integrated conversion of raw glycerol from the biodiesel process to value-added products is a driver towards higher cost efficiency of biodiesel production. Glycerol is a good source of carbon and energy for growth of several microorganisms and may be suitable for the biotechnological production of a number of chemicals in fermentative processes. To date several microbiological productions have been adjusted to glycerol as carbon and energy source or, if glycerol is close to the desired product, are based on glycerol as substrate anyway. For instance, the biotechnological production of 1,3propanediol and dihydroxyacetone has predominantly been carried out from glycerol, since these processes are catalyzed in a two and one step reaction, respectively. Bacterial 1,3propanediol production from glycerol is involving two enzymes. First glycerol is dehydrated by glycerol dehydratase to 3-hydroxypropionaldehyde, which is subsequently converted to 1,3-propanediol by 1,3-propanediol dehydrogenase. Predominantly, 1,3propanediol production has been approved with Clostridium strains, Klebsiella pneumoniae, or Escherichia coli (Zhu et al. 2002, Biebl et al. 1998, Forsberg, 1987). 1,3-propanediol finds application in the production of the polyester polytrimethylene terephthalate. Strains of Gluconobacter oxydans are used for producing dihydroxyacetone (glycerone) from glycerol (Flickinger & Perlman, 1977). Dihydroxyacetone is used as a building block in organic chemistry and as a skin tanning agent in cosmetics. Besides products that can be derived from glycerol in one or two reactions further products requiring more complex conversions have been investigated, e.g. succinic acid production


Introduction
Since decades the limited access to petroleum oil is a major concern and substitutions for fossil fuels are needed.One promising substitute is biodiesel, which is widely produced from vegetable oils, e.g. from rape seeds, soybeans, sunflower seeds or animal fats.In the synthesis of biodiesel, oils and fats are transesterified to fatty acid methyl ester in the presence of sodium hydroxide or potassium hydroxide.In this process, glycerol is generated as stoichiometric byproduct with a ratio of 10% (w/w) with respect to biodiesel produced.In 2009, the biodiesel production of the world reached 16 million tons (Licht, 2010), with the lion's share produced by the European Union with 9 million tons (EBB, European Biodiesel Board 2010), followed by the United States with a production of 2.7 million tons (Licht, 2010).Hence, 1.6 million tons of glycerol were produced as obligatory by-product.Glycerol finds applications as an ingredient of various products, such as creams, food, feed, and pharmaceuticals, but the demand for glycerol in these processes is limited.Integrated conversion of raw glycerol from the biodiesel process to value-added products is a driver towards higher cost efficiency of biodiesel production.Glycerol is a good source of carbon and energy for growth of several microorganisms and may be suitable for the biotechnological production of a number of chemicals in fermentative processes.To date several microbiological productions have been adjusted to glycerol as carbon and energy source or, if glycerol is close to the desired product, are based on glycerol as substrate anyway.For instance, the biotechnological production of 1,3propanediol and dihydroxyacetone has predominantly been carried out from glycerol, since these processes are catalyzed in a two and one step reaction, respectively.Bacterial 1,3propanediol production from glycerol is involving two enzymes.First glycerol is dehydrated by glycerol dehydratase to 3-hydroxypropionaldehyde, which is subsequently converted to 1,3-propanediol by 1,3-propanediol dehydrogenase.Predominantly, 1,3propanediol production has been approved with Clostridium strains, Klebsiella pneumoniae, or Escherichia coli (Zhu et al. 2002, Biebl et al. 1998, Forsberg, 1987).1,3-propanediol finds application in the production of the polyester polytrimethylene terephthalate.Strains of Gluconobacter oxydans are used for producing dihydroxyacetone (glycerone) from glycerol (Flickinger & Perlman, 1977).Dihydroxyacetone is used as a building block in organic chemistry and as a skin tanning agent in cosmetics.Besides products that can be derived from glycerol in one or two reactions further products requiring more complex conversions have been investigated, e.g.succinic acid production with Anaerobiospirillum succiniciproducens (Lee et al. 2001) or Escherichia coli (Blankschien et al., 2010, Zhang et al., 2010), citrate production with Yarrowia lipolytica (Rywinska & Rymowicz, 2010), ethanol production with Escherichia coli, Saccharomyces cerevisiae, or Hansenula polymorpha (Hong et al., 2010, Shams Yazdani & Gonzalez, 2008, Yu et al., 2010), production of amino acids with Corynebacterium glutamicum (Rittmann et al., 2008), or propionate production with Propionibacteria (Himmi et al., 2000).Crude glycerol preparations from biodiesel factories differ considerably from the pure chemical glycerol, e.g. by their salt content.The applicability of crude glycerol from biodiesel production plants has already been demonstrated for production strains of Clostridium and Klebsiella in 1,3-propanediol production (Gonzalez-Pajuelo et al., 2004, Mu et al., 2006), Yarrowia lipolytica in citrate production (Papanikolaou et al., 2002), Kluyvera cryocrescens and Klebsiella pneumoniae in ethanol production (Choi et al., 2011, Oh et al., 2011), as well as Basfia succiniciproducens in succinic acid production (Scholten et al., 2009).This chapter will summarize state-of-the-art of glycerol-based biotechnological processes and will discuss future developments.

Use of glycerol as a carbon source in biotechnological applications
Glycerol has many applications, it is used for the production of food, cosmetics, paints, pharmaceutics, paper, textiles, leather and for the production of various chemicals (Wang et al., 2001).It can be used as a stabilizing agent for storage of cells and proteins.Physiologically, glycerol is essential for the biosynthesis of membranes, since it is the backbone of glycerolipids.And for its function as a component of lipids and fats it is an abundant source of carbon and energy in nature.Formerly, glycerol was a valuable product derived from glucose via dihydroxyacetonephosphate and glycerol-3-phosphate by glycerol-3-phosphate dehydrogenase (EC 1. 1.1.94)and glycerol-3-phosphatase (EC 3. 1.3.21).The yeast Saccharomyces cerevisiae which uses glycerol as an osmolyte under osmotic stress conditions was engineered for efficient glycerol production from glucose.S. cerevisiae possesses two isozymes of each, glycerol-3-phosphate forming glycerol-3-phosphate dehydrogenases and glycerol-3-phosphatases (Larsson et al., 1993, Pahlman et al., 2001).Sulphite treatment of yeasts enabled glycerol production (Petrovska et al., 1999), as did metabolic engineering on the glycerol production pathway, e.g.deletion of the triosephosphate isomerase gene (Overkamp et al., 2002), deletion of the alcohol dehydrogenase gene, overexpression of a glycerol-3-phosphate dehydrogenase gene (Navarro-Avino et al., 1999), or overexpression of a glycerol exporter gene (Cordier et al., 2007).Nowadays, glycerol arises from the biodiesel production.In 2010 glycerol formed as byproduct in the biodiesel process amounted to 1.6 million tons (Licht, 2010), which is extending the glycerol demand by far, thus, making microbial glycerol production unprofitable.
described in yeasts, such as sodium dependent symport for Debaryomyces hansenii (Lucas et al., 1990) and proton dependent symport for Pichia sorbitophila (Lages & Lucas, 1995) and Saccharomyces cerevisiae (Lages & Lucas, 1997, Ferreira et al., 2005).In yeasts, active glycerol transport is mostly regarded of importance with respect to the use of glycerol as an osmolyte under osmotic stress conditions.The imported glycerol can enter the metabolism at the level of the glycolytic intermediate dihydroxyacetone-phosphate.Two routes for the formation of dihydroxyacetone-phosphate from glycerol have been reported.In the first, ATP-dependent glycerol kinase (EC 2.7.1.30) phosphorylates glycerol to glycerol-3-phosphate, which is subsequently oxidized to dihydroxyacetone-phosphate by glycerol-3-phosphate dehydrogenase, which are either quinone or FAD-dependent (EC 1.1.5.3) or NAD-dependent (EC 1. 1.1.8).In the second pathway, glycerol is oxidized to dihydroxyacetone by glycerol dehydrogenase (EC 1. 1.1.6)before being phosphorylated to dihydroxyacetone-phosphate by ATP-or phosphoenolpyruvate-dependent dihydroxyacetone kinase (EC 2.7.1.29).

Dihydroxyacetone
Dihydroxyacetone (DHA; 1,3-Dihydroxypropan-2-one, Glycerone) is produced biotechnologically with a global market of 2,000 tons per year (Pagliaro et al., 2007).DHA is produced biotechnologically, since its chemical synthesis is expensive and requires safety measures to cope with hazardous reactants (Hekmat et al., 2003).The most popular use of DHA is as coloring agent in sunless tanning products, such as creams and lotions (Levy, 1992).The tanning effect depends on Maillard-like reactions of DHA with the amino acids of the outer skin layer.Historically, also applications for medical treatments of glycogenesis, a glycogen storage disease, and diabetes mellitus have been described.Currently, the use of DHA as building block for chemical synthesis appears to have the highest potential for a production process based on biodiesel-derived glycerol (Enders et al., 2005, Zheng et al., 2008, Hekmat et al., 2003).Gluconobacter oxydans is used for the production of DHA from glycerol.This bacterium belongs to the family of Acetobacteraceae (acetic acid bacteria), which are able to oxidize many carbohydrates and alcohols incompletely.To this end, G. oxydans possesses a variety of membrane-bound dehydrogenases.The membrane-bound pyrroloquinoline quinone (PQQ)-dependent glycerol dehydrogenase (EC 1. 1.99.22)catalyzes oxidation of the secondary hydroxy group of glycerol to DHA (Matsushita et al., 1994).The enzyme is the protein product of sldA and sldB (Prust et al., 2005) and its localization allows extracellular oxidization of glycerol to DHA in the periplasma with a concurrent reduction of the membrane-localized PQQ.Besides membrane-bound glycerol dehydrogenase (Fig. 2), G. oxydans also possesses an intracellular catabolic pathway for the use of glycerol as a carbon source for growth (Fig. 1), in which glycerol is phosphorylated by ATP-dependent glycerol kinase to yield glycerol-3-phosphate which in turn is oxidized to dihydroxyacetonephosphate by NAD-dependent glycerol-3-phosphate dehydrogenase (Claret et al., 1994).Since a functional glycolysis pathway is missing in G. oxydans and since its tricarboxylic acid cycle is incomplete, dihydroxyacetone-phosphate is metabolized via the pentose phosphate pathway (Greenfield & Claus, 1972).Problems have occurred in the process of DHA production by G. oxydans, most importantly inhibition of the biotransformation process by the substrate glycerol and the product DHA as both inhibit growth and DHA production (Claret et al., 1992, Claret et al., 1993, Bauer et al., 2005).These problems have been met by optimizing production conditions including immobilization of G. oxydans cells to a polyvinyl alcohol matrix, which resulted in cells active for 14 days while maintaining glycerol dehydrogenase activity above 90 % (Wei et al., 2007).Also fed-batch cultivations were shown to be supportive for DHA production compared to batch fermentations, as higher total glycerol amounts could be converted into DHA by avoiding inhibitory concentrations of glycerol, however, in this case production yields are limited with respect to DHA accumulation (Bories et al., 1991).Further enhancements of DHA production were achieved when using repeated fed-batch cultivations, in which DHA concentrations were kept below the inhibitory concentration.By this method immobilized cells were reused for up to 100 fed-batch cycles reducing costs and time for cleaning, sterilization, and inoculation (Hekmat et al., 2003).Other studies focused on optimizing culturing conditions for DHA production focused on cultivation media, aeration, or pH (Wethmar & Deckwer, 1999, Svitel & Sturdik, 1994, Tkac et al., 2001, Holst et al., 1985).After the genome sequence of G. oxydans became public (Prust et al., 2005), metabolic engineering of G. oxydans has been reported for optimizing DHA production.Overexpression of the genes encoding the glycerol dehydrogenase, sldAB, led to increased biomass formation and higher DHA yields from glycerol (Herrmann et al., 2007).As the formation of glyceric acid as byproduct interferes with DHA production, disruption of adhA, the gene for the PQQ-dependent alcohol dehydrogenase, which is involved in the oxidation of glycerol to glyceric acid was shown to improve the product yield by abolishing glyceraldehyde and glyceric acid formation (Habe et al., 2010a).In addition, the mutant lacking adhA was less inhibited by high initial glycerol concentrations than the parent strain.In line with the notion that glyceric acid inhibits growth and DHA production by G. oxydans, it was found that addition of glyceric acid to the medium reduced growth and DHA production (Habe et al., 2010a) and that AdhA activity strongly increased when very high glycerol concentrations were used (Habe et al., 2009d).In the next step, the combination of adhA disruption and sldAB overexpression resulted in a strain with very high productivity and strongly increased tolerance towards glycerol (Li et al., 2010b).Since the production of DHA by the obligate aerobic G. oxydans is characterized by a very high demand for oxygen to oxidize reduced PQQ (Hekmat et al., 2003, Svitel & Sturdik, 1994), the gene encoding Vitreoscilla hemoglobin, an oxygen transporting protein, was shown to be beneficial (Li et al., 2010a).Besides G. oxydans strains other acetic acid bacteria have also been reported for DHA production from glycerol, e.g.Gluconobacter melanogenus (Flickinger & Perlman, 1977) and Acetobacter xylinum (Nabe et al., 1979).Until today, production of DHA from biodieselderived raw glycerol has not yet been reported.

1,2-PDO production by E. coli
First approaches towards metabolic engineering of E. coli strains for production of 1,2-PDO from glucose by Cameron et al. involved overexpression of genes for either aldose reductase or glycerol dehydrogenase (Cameron et al., 1998).Later on, the same group coexpressed glycerol dehydrogenase genes from E. coli or K. pneumoniae together with the E. coli methylglyoxal synthase gene in E. coli and reached up to 0.7 g l -1 , an almost three-fold increase when compared to overexpression of E. coli or K. pneumoniae glycerol dehydrogenase gene alone (0.25 g l -1 ) (Altaras & Cameron, 1999).Additional overexpression of yeast alcohol dehydrogenase or E. coli 1,2-PDO reductase further improved production performance and 1,2-PDO concentrations of 4.5 g l -1 and a yield of 0.19 g (g glucose) -1 were achieved in a fed-batch fermentation process (Altaras & Cameron, 2000).Elimination of lactate dehydrogenase by gene deletion improved 1,2-PDO production by an E. coli strain overexpressing genes coding for methylglyoxal synthase from Clostridium acetobutylicum and glycerol dehydrogenase from E. coli (Berrios-Rivera et al., 2003).Glycerol has a higher degree of reduction than glucose, thus, higher 1,2-PDO yields are theoretically possible using glycerol (0.72 g g -1 ) as compared to glucose (0.63 g g -1 ) (Clomburg & Gonzalez, 2011).Moreover, 1,2-PDO has been reported to be a natural product of anaerobic fermentation of glycerol in E. coli by Gonzalez et al. (Gonzalez et al., 2008).Glycerol is converted to 1,2-PDO in a pathway consisting of glycerol dehydrogenase for oxidation of glycerol to dihydroxyacetone and phosphorylation of the latter to dihydroxyacetone-phosphate by phosphoenolypyruvate-dependent dihydroxyacetone kinase.Subsequently, dihydroxyacetone-phosphate is reduced to 1,2-PDO which requires two NADH.Thus, to generate NADH and ATP required in these reactions a portion of dihydroxyacetone-phosphate needs to be catabolized in glycolysis and onwards to ethanol (Gonzalez et al., 2008).Recently, Clomburg et al. rationally engineered E. coli for effective 1,2-PDO production from glycerol.They introduced a 1,2-PDO production pathway via methylglyoxal synthase, glycerol dehydrogenase, and aldehyde oxidoreductase.Replacement of the phosphoenolpyruvate-dependent dihydroxyacetone kinase by ATP-dependent dihydroxyacetone kinase from Citrobacter freudii elevated dihydroxyacetone-phosphate availability and enhanced 1,2-PDO production from 0.02 to 0.15 g g -1 .Moreover, the formation of byproducts succinate, acetate, ethanol, and formate increased but lactate was reduced.To eliminate byproduct formation several gene deletions, e.g. of genes coding for lactate dehydrogenase, acetate kinase, phosphate acetyltransferase, formate hydrogen lyase, fumarate reductase, alcohol dehydrogenase, and pyruvate dehydrogenase, were tested alone or in combination.Ethanol formation could not be abolished as deletion of the alcohol dehydrogenase gene strongly decreased 1,2-PDO production (0.02 compared to 0.12 g g -1 ) and glycerol consumption, but increased formation of acetate and succinate as byproducts.However, deletion of the genes for acetate kinase, phosphate acetyltransferase, and lactate dehydrogenase resulted in an increased product yield of 0.21 g g -1 , but entailed increased ethanol, formate, and pyruvate formation.The use of raw glycerol by this strain reduced formate formation and increased the 1,2-PDO yield (0.24 g g -1 ) (Clomburg & Gonzalez, 2011).Taken together, engineered E. coli strains allow for 1,2-PDO yields from glycerol that are comparable to yields from glucose, making glycerol a feasible substrate for microbial 1,2-PDO production.

1,2-PDO production by other microorganisms
Recombinant strains of S. cerevisiae and of C. glutamicum have been developed for production of 1,2-PDO, as well.S. cerevisiae carrying multiple genome-integrated copies of E. coli methylglyoxal synthase and glycerol dehydrogenase genes was able to produce 1,2-PDO (Lee & da Silva, 2006).Plasmid-borne expression of the latter genes combined with the deletion of the endogenous triosephosphate isomerase gene resulted in the production of 1.1 g l -1 1,2-PDO (Jung et al., 2008).Expression of E. coli methylglyoxal synthase gene and a Corynebacterium glutamicum aldo-keto reductase gene enabled C. glutamicum for the production of 1,2-PDO from glucose (1.8 g l -1 ) (Niimi et al., 2011).

1,3-Propanediol
1,3-propanediol (1,3-PDO) can be used in several chemical applications.It is a substrate in the formulation reactions for polyesters, polyethers, polyurethanes, adhesives, composites, laminates, coatings, and moldings.In addition, 1,3-PDO itself is used e.g. as solvent or antifreeze agent.Most importantly, the production of polytrimethylene terephthalate (PTT) from 1,3-PDO and terephthalic acid is the driver of the growing market for 1,3-PDO.The polymer PTT is a promising polyester with numerous applications, e.g. as compound in textile, carpet, upholstery, or specialty resins.Due to its properties, PTT might be favored over polymers such as nylon, polyethylene terephthalate and polybutylene terephthalate.PTT has environmental benefits since it is biodegradable (for recent reviews see (da Silva et al., 2009, Liu et al., 2010, Carole et al., 2004).The chemical producers Shell and DuPont produce PTT from 1,3-PDO which is commercialized under the names Sorona, Hytrel (DuPont) and Corterra (Shell).1,3-PDO is a success story of a glycerol-based biotechnological process.When 1,3-PDO was first discovered to be a fermentative product of glycerol in 1881 (Werkman & Gillen, 1932), it received little interest until the development of PTT in 1941 (Whinfield & Dickinson, 1946).However, while terephthalic acid was readily available the expense of 1,3-PDO production limited efficient PTT production.Nowadays, chemical production of 1,3-PDO from ethylene oxide or acrolein as well as its biotechnological production enable supply of 1,3-PDO at low cost.It has to be noted that chemical synthesis of 1,3-PDO suffers from high energy consumption, toxic intermediates, and expensive catalysts as major drawbacks.Currently, the market for 1,3-PDO is estimated to be about 50,000 tons per year (Liu et al., 2010), but due to a growing production of PTT a market volume of 230,000 tons is foreseen for 2020 (Carole et al., 2004).An indication of the growing interests and the potential of biotechnological 1,3-PDO production are the current decisions of the joint venture of DuPont and Tate & Lyle to extend their 1,3-PDO biotech plant in Louden, TN, USA, by 35 %.The actual capacity of the plant is 45,000 tons per year (Greenwood, 2010).Also the French company Metabolic Explorer (Clermont-Ferrand, France) decided to build a plant for biotechnological 1,3-PDO production in Malaysia in partnership with Malaysian Bio-XCell.The plant with a capacity to produce 50,000 tons 1,3-PDO annually is expected to start with a production of 8,000 tons per year (Degalard, 2011).While the process of DuPont and Tate & Lyle is based on sugars from corn hydrolysates, the Metabolic Explorer process will be based on crude glycerol.

Metabolic engineering of K. pneumoniae, C. acetobutylicum and E. coli for 1,3-PDO production
Several metabolic engineering approaches were made for the optimization of 1,3-PDO production and reduction of byproduct formation with K. pneumoniae and C. acetobutylicum.E. coli, not a natural producer of 1,3-PDO, was engineered for 1,3-PDO production.

K. pneumoniae
K. pneumoniae was genetically manipulated to gain higher 1,3-PDO titers and production rates and to reduce process competing byproducts.When overexpressing the gene encoding the first enzyme of 1,3-PDO production glycerol dehydratase (dhaB) Zhao et al. (2009) found no effect on 1,3-PDO yields but reported decreased formation of the byproducts ethanol and 2,3-butanediol and an increase of acetic acid production (Ma et al., 2009).The overexpression of the 1,3-PDO dehydrogenase gene dhaT was shown to positively affect 1,3-PDO production in many studies and to reduce the formation of the toxic intermediate and substrate of 1,3-PDO dehydrogenase 3-HPA (Rao et al., 2010, Ma et al., 2009, Zhuge et al., 2010, Hao et al., 2008b).While aiming to reduce concentrations of the inhibitory intermediate 3-HPA Hao et al. (2008) overexpressed the 1,3-PDO dehydrogenase gene (dhaT) in K. pneumoniae TUAC01.During fermentation with 30 or 50 g l -1 glycerol 3-HPA accumulation was significantly reduced to 1.49 and 2.02 mM, respectively, compared to the parental strain which produced 7.55 and 12.57 mM, respectively (Hao et al., 2008b).Accordingly, Ma et al. (2010) overexpressed the 1,3-PDO dehydrogenase gene in K. pneumoniae and found a 3-fold decrease of 3-HPA production and an increase in 1,3-PDO production by 16.5% (Ma et al., 2010b). Similarly, Zhao et al. (2009) reported a strong decrease of the toxic intermediate 3-HPA by overexpression of the 1,3-PDO dehydrogenase gene leading to an increased molar yield of 0.64 mol mol -1 compared to 0.51 mol mol -1 of the parental strain and decreased lactic acid, ethanol and succinic acid concentrations in fed-batch fermentations (Ma et al., 2009).Zhuge et al. (2010) combined overexpression of dhaT and yqhD (encoding NADH-and NADPHdependent 1,3-PDO dehydrogenases from K. pneumoniae and E. coli, respectively) in K. pneumoniae.The recombinant strain had a slightly elevated product titer (18.3 g l -1 compared to 17.1 g l -1 ) and increased molar 1,3-PDO yield (0.51 to 0.57 mol mol -1 ) in batch fermentations.Furthermore, the byproducts 3-HPA, succinic acid, lactic acid, acetic acid and ethanol were significantly reduced (Zhuge et al., 2010).When analyzing a K. pneumoniae mutant strain defective in the genes for NADH-dependent 1,3-PDO dehydrogenase and the oxydative glycerol utilization pathway Seo et al. (2009) reported that the mutant surprisingly retained the ability of 1,3-PDO production.1,3-PDO yields were low but a strongly reduced byproduct formation was reported (Seo et al., 2009).Later, Seo et al. (2010) published the identification of a second 1,3-PDO dehydrogenase, which possesses high homology to E. coli YqhD and is, in contrast to the dhaT encoded enzyme, NADPH-dependent.Overproduction of the NADPH-dependent 1,3-PDO dehydrogenase from K. pneumoniae resulted in restoration of 1,3-PDO production in the mutant defective in the genes for NADH-dependent 1,3-PDO dehydrogenase and the oxydative glycerol utilization pathway and moreover byproduct formation remained low in the recombinant strain, as the oxidative pathway of glycerol utilization was absent.Although the 1,3-PDO concentrations were lower as compared to the parental strain (4.7 compared to 7.9 g l -1 ), the molar yield of 0.54 mol mol -1 of the recombinant strain was higher compared to 0.48 mol mol -1 (Seo et al., 2010).To reduce byproduct formation Horng et al. (2010) constructed a K. pneumoniae mutant lacking the genes for glycerol dehydrogenase and dihydroxyacetone kinase.This mutant was reported to have ceased lactate, 2,3-butanediol, and ethanol byproduct formation and furthermore showed higher 1,3-PDO productivity when glycerol dehydratase and 1,3-PDO dehydrogenase were overexpressed (Horng et al., 2010).A lactate dehydrogenase mutant was constructed by Xu et al. (2009) in K. pneumoniae HR526.The accumulation of lactate was reduced from 40 g l -1 in the parental strain to less than 3 g l -1 in the lactate dehydrogenase mutant, which showed very low lactate dehydrogenase activity.The mutant furthermore produced higher 1,3-PDO concentrations (95 g l -1 as compared to 102 g l -1 ) with a higher yield (0.48 mol mol -1 to 0.52 mol mol -1 ) and higher production rate (1.98 g l -1 h -1 to 2.13 g l -1 h -1 ).Reduced lactate production increased NADH availability for 1,2-PDO production in K. pneumoniae (Xu et al., 2009) and K. oxytoca (Yang et al., 2007).Similarly, inactivation of the NADH-dependent aldehyde dehydrogenase gene in K. pneumoniae abolished ethanol formation, improved NADH availability and resulted in a 1,3-PDO production rate of 1.07 g l -1 h -1 and a yield of 0.70 mol mol -1 (Zhang et al., 2006).

E. coli
Wild-type E. coli is unable to convert glycerol to 1,3-PDO (Tong et al., 1991), but already in 1991 Tong et al. constructed a recombinant E. coli strain for the production of 1,3-PDO in an early application of metabolic engineering.A K. pneumoniae ATCC 25955 genomic library in E. coli was screened for anaerobic growth on glycerol and dihydroxyacetone and 1,3-PDO production.The selected recombinant possessed glycerol dehydratase, 1,3-PDO dehydrogenase, glycerol dehydrogenase, and dihydroxyacetone kinase from K. pneumoniae and produced 1 g l -1 1,3-PDO with a yield of 0.46 mol mol -1 (Tong et al., 1991).Cofermentation of glycerol with glucose or xylose increased yields from 0.46 mol mol -1 to 0.63 mol mol -1 in the presence of glucose and to 0.55 mol mol -1 when cofermented with xylose (Tong & Cameron, 1992).Skraly et al. (1998) further optimized 1,3-PDO production with E. coli as they constructed an artificial operon containing the K. pneumoniae genes of 1,3-PDO production.They could show that the recombinant E. coli strain yielded 0.82 mol mol -1 (glycerol only) in a fed-batch process cofermenting glycerol and glucose, but only 9.3 g l -1 of glycerol where converted (Skraly et al., 1998).The additional expression of the glycerol dehydratase reactivating factor genes gdrAB together with the expression of K. pneumoniae genes encoding glycerol dehydratase and 1,3-PDO dehydrogenase yielded 8.6 g l -1 1,3-PDO in a fed-batch fermentation (Wang et al., 2007).A higher final 1,3-PDO concentration of 13.2 g l -1 was obtained when they substituted K. pneumoniae 1,3-PDO dehydrogenase with the E. coli YqhD, which possesses NADPH-dependent 1,3-PDO dehydrogenase activity (Tan et al., 2007).YqhD from E. coli was also used in addition to the vitamin B12-independent glycerol dehydratase from C. butyricum in a two-stage fermentation process (aerobic biomass production from glucose followed by anaerobic 1,3-PDO production from glycerol).By this process 1,3-PDO concentrations of 104.4 g l -1 were reached, with a production rate of 2.62 g l - 1 h -1 and a molar yield of 1.09 mol mol -1 (90.2% g g -1 ) (referred to glycerol only) (Tang et al., 2009).Some efforts have also been made in the abolishment of toxic intermediate metabolites in E. coli during 1,3-PDO production.In E. coli high glycerol concentrations inhibit growth and 1,3-PDO production due to intracellular accumulation of glycerol-3phosphate (Cozzarelli et al., 1965), the product of glycerol kinase, first enzyme in the oxidative pathway of glycerol utilization in E. coli.Zhu et al. (2002) could show that the glycerol-3-phosphate concentration increased when glycerol concentrations were elevated.The glycerol-3-phosphate accumulation was due to inefficient expression of the glycerol-3phosphate dehydrogenase gene, and was overcome by usage of a glycerol kinase mutant, which showed 2.5-fold increased 1,3-PDO production (Zhu et al., 2002).The reduction of methylglyoxal formation by expression of the Pseudomonas putida glyoxalase I resulted in increased 1,3-PDO production by 50% (Zhu et al., 2001).

Raw glycerol
Investigations of raw glycerol use for the production of 1,3-PDO have been made with Clostridium and Klebsiella strains.K. pneumoniae produced 1,3-PDO concentrations from raw glycerol from soybean oil biodiesel production close to that from pure glycerol (51. 3 g l -1 ) with a productivity of 1.7 g l -1 h -1 (compared to 2 g l -1 h -1 ) (Mu et al., 2006).Similar concentrations (56 g l -1 ) were reported by Hiremath et al. who used glycerol obtained from Jatropha seed biodiesel production (Hiremath et al., 2011).C. beijerinckii was also reported to produce 1,3-PDO from raw glycerol (Gungormusler et al., 2011).Notably, productivity was increased to 1.51 g l -1 h -1 when using raw glycerol as compared to 0.84 g l -1 h -1 when using pure glycerol (Um et al., 2010).Moon et al. compared 1,3-PDO production with Klebsiella and Clostridium strains from glycerol derived from biodiesel production from waste vegetable oil and soybean oil (Moon et al., 2010).Possibly due to inhibitory methanol concentrations, the use of soybean derived glycerol was better than use of raw glycerol derived from waste vegetable oil from different suppliers.In this study, a higher tolerance of Klebsiella strains to the different raw glycerols used compared to the Clostridium strains was observed (Moon et al., 2010).C. butyricum was also shown to utilize raw glycerol for 1,3-PDO production (Papanikolaou et al., 2000, Gonzalez-Pajuelo et al., 2004).Inhibitory effects of raw glycerol components on 1,3-PDO production with C. butyricum were found not to be due to NaCl or methanol, but rather due to oleic acid (Chatzifragkou et al., 2010).

Ethanol
Ethanol as a bio-fuel is mainly gained from sugarcane in Brazil, from corn in the USA and from sugar beets in the EU (da Silva et al., 2009).Since ethanol production is already done in a tens of billions scale, production deriving from crude glycerol may contribute only a small fraction.Nevertheless there has been considerable research on this topic in order to use crude glycerol efficiently to produce ethanol (Licht, 2010).Unfortunately, the well known ethanol producer Saccharomyces cerevisiae grows very slow on glycerol and, thus, the growth had to be considerably improved, e.g. by selecting S. cerevisiae strain CBS8066-FL20 which grows much faster (0.2 h -1 rather than at 0.01 h -1 ) (Ochoa-Estopier et al., 2011).Ethanol accumulation of the yeast Hansenula polymorpha was improved from 0.83 g l -1 to 2.74 g l -1 by expression of genes encoding pyruvate decarboxylase (pdc) and aldehyde dehydrogenase II (adhB) from Zymomonas mobilis.Combined with the expression of glycerol dehydrogenase (dhaD) and dehydroxyacetone kinase (dhaKLM) genes from Klebsiella pneumonia even more ethanol (3.1 g l -1 ) was produced (Hong et al., 2010).Elementary mode analysis and metabolic evolution of E. coli mutants led to conversion of 40 g l -1 glycerol to ethanol reaching 90% of the theoretical yield (Trinh & Srienc, 2009).E. coli strains have been engineered to produce ethanol and H 2 or ethanol and formate from crude gycerol.Due to overexpression of genes for glycerol dehydrogenase (gldA) and dihydroxyacetonekinase (dhaKLM) 95% of the theoretical maximum yield and specific production rates of 15-30 mmol (g cell) -1 h -1 could be obtained (Shams Yazdani & Gonzalez, 2008).A newly isolated bacterium, Kluyvera cryocrescens S26, was able to produce 27 g l -1 ethanol with yield of 0.8 mol mol -1 and a productivity of 0.61 g l -1 h -1 (Choi et al., 2011).

Succinate
Succinic acid (succinate) is a so called platform chemical based on which a variety of other chemicals are produced, e.g.tetrahydrofuran, γ-butyrolactone, adipic acid, 1,4-butanediol and n-methyl-pyrrolidone. Based on this spectrum of products there are several markets succinic acid is involved in, such as pharmaceuticals, chemistry of biodegradable polymers, surfactants and detergents (Zeikus et al., 1999).Various microorganisms have been engineered for succinate production (Wendisch et al., 2006).Succinic acid is an intermediate of the TCA cycle with four carbon atoms and two carboxylic groups.Today most of the produced succinic acid derives from the petrochemical industry and only a small part already comes from biotechnological processes.For the chemical synthesis the nonrenewable fossil fuel butane is the starting point leading through maleic anhydride to succinic acid (Zeikus et al., 1999).Production of succinate from glycerol is interesting because both share the same level of reduction, thus, when produced from glycerol and CO 2 no further electron source is necessary.Various attempts have been made to use bacteria to efficiently produce succinate using natural succinic acid producers as well as metabolically engineered strains (Zeikus et al., 2007, Ingram et al., 2008, Lin et al., 2005, Samuelov et al., 1991, Okino et al., 2005, Singh et al., 2009, Van der Werf et al., 1997, Zhang et al., 2009).Among the natural producers, Anaerobiospirillum succiniciproducens was shown to use glycerol as sole or combined carbon source to efficiently produce succinic acid (Lee et al., 2004).They found a high succinic acid yield of 133% (or 160% when yeast extract was additionally fed) for glycerol concentrations of 6.5 g l -1 .Glycerol entailed less formation of acetic acid as byproduct and, thus, an easier downstream processing (Lee et al., 2001).Succinate production by the related bacterium Basfia succiniciproducens from crude glycerol in a continuous cultivation process allowed for product yields of about 1 g g -1 (Scholten et al., 2009).Succinate production from glycerol by E. coli appeared to be favored under aerobic conditions as indicated by elementary mode analysis (Chen et al., 2010).Using microaerobic conditions, a recombinant producing pyruvate carboxylase from Lactococcus lactis and lacking pathways to byproducts showed succinic acid yields on glycerol comparable to those on glucose (Blankschien et al., 2010).About 80% of the maximum theoretical yield could be achieved by inserting three key mutations affecting phosphoenolpyruvate carboxykinase (pck), part of the phosphotransferase system (ptsI) and the pyruvate formate lyase (pflB) (Zhang et al., 2010).

Citrate
Citric acid is produced by fermentation at a scale of about 1,600,000 t/a (Papanikolaou et al., 2002, Berovic & Legisa, 2007), and it is sold for about 0.8 €/kg (Weusthuis et al., 2011).The main markets are the food and the pharmaceutical industries as well as applications in cosmetics, detergents and cleaning products.Citric acid is produced almost exclusively by Aspergillus niger in a submerged fermentation process using starch-or sucrose-based media like molasses (Soccol et al., 2006).
Different strains of the yeast Yarrowia lipolytica have been investigated for citric acid production using glycerol as a carbon source.For Yarrowia lipolytica Wratislavia AWG7 a maximal yield of 0.67 g g -1 was reported in continuous culture using glycerol as carbon source with only low contamination by the common byproduct isocitric acid (Rywinska et al., 2011, Rywinska & Rymowicz, 2010). In 2002, Papanikolaou et al. reported about the Y. lipolytica strain LGAM S(7)1 being capable of growing on crude glycerol as carbon source and producing up to 35 g l -1 citric acid (Papanikolaou et al., 2002) .In 2010 and 2011, much higher citric acid concentrations of 112 g l -1 and a yield of 0.6 g g -1 using crude glycerol and the acetate-negative mutant strain Y. lipolytica A-101-1.22 (Rymowicz et al., 2010) or the acetate-negative strain Y. lipolytica N15 could be achieved (Kamzolova et al., 2011).

Amino acids
Amino acids are a multi-billion dollar business (Wendisch, 2007).They are used as flavor enhancers (L-glutamate), feed additives (L-lysine, L-methionine, L-threonine, Ltryptophane), to produce sweeteners such as aspartam (L-aspartate, L-phenylalanine), and in various pharmaceutical applications.The biggest products are L-glutamate (2,160,000 tons per year) and L-lysine (1,330,000 tons per year) (Ajinomoto, 2010a, Ajinomoto, 2010b).Pathways for amino acid synthesis start from intermediates of glycolysis (e.g.L-serine from 3-phosphoglycerate, L-valine from pyruvate), glycolysis and pentose phosphate pathway (aromatic amino acids L-tyrosine, L-phenylalanine, and Ltryptophane, from phosphoenolpyruvate (PEP) and erythrose-4-phosphate), and tricarboxylic acid cycle intermediates (L-glutamate, L-glutamine from 2-oxoglutarate, L-lysine, L-aspartate from oxaloacetate) (Schneider & Wendisch, 2001, Wendisch, 2007, Gopinath et al., 2011).In principle, production of amino acids from glycerol should be possible.In the following section the biosynthesis of L-glutamate, L-lysine, and L-phenylalanine are described, since strains for production of these from glycerol have been described.Corynebacterium glutamicum is a natural L-glutamate producer (Eggeling & Bott, 2005), but the excretion of L-glutamate needs to be "triggered", e.g. by limitation of biotin (Shiio et al., 1962).Biotin is essential for the activity of acetyl-CoA carboxylase, necessary for fatty acid synthesis, and hence for membrane precursors, thus effect of biotin limitation on Lglutamate production is thought to be due to a higher permeability of the cell membrane (Shimizu & Hirasawa, 2007).Also addition of detergents like Tween 40 (Takinami et al., 1965), antibiotics like penicillin (Nara et al., 1964), and ethambutol (Radmacher et al., 2005, Stansen et al., 2005) trigger L-glutamate production.In C. glutamicum L-glutamate is mainly synthesized by NADPH dependent glutamate dehydrogenase from the tricarboxylic acid cycle intermediate 2-oxoglutarate (Bormann et al., 1992).This holds true for high ammonia concentrations, when ammonia is low L-glutamate is synthesized via L-glutamine by glutamine synthetase and glutamine-2-oxoglutarate aminotransferase.Crucial for Lglutamate production is the anaplerosis of the tricarboxylic acid cycle by either pyruvate carboxylase or PEP carboxylase.Pyruvate carboxylase has been shown to be indispensable under detergent triggered production conditions (Peters-Wendisch et al., 2001) and vice versa under biotin limiting conditions PEPcarboxylase is responsible for anaplerosis (Sato et al., 2008, Delaunay et al., 2004, Lapujade et al., 1999).C. glutamicum was engineered for glycerol utilization by expression of the genes for glycerol facilitator, glycerol kinase, and glycerol-3-phosphate dehydrogenase from E. coli (Rittmann et al., 2008).Under ethambutol triggered L-glutamate production conditions recombinant C. glutamicum showed reduced Lglutamate yields from glycerol compared to glucose, 0.11 g g -1 compared to 0.20 g g -1 , respectively (Rittmann et al., 2008).Production of L-lysine, which is used as a feed additive, is also carried out with C. glutamicum (Wendisch, 2007, Eggeling & Bott, 2005).The precursors of L-lysine production are the tricarboxylic acid cycle intermediate oxaloacetate and the glycolytic intermediate pyruvate.Deregulation of the L-lysine production pathway by introduction of feedback resistant variants of the key enzyme aspartate kinase, which usually is inhibited by L-lysine and L-threonine (Kalinowski et al., 1991) enables C. glutamicum for L-lysine production.Further increases were made by overexpression of the gene for pyruvate carboxylase (Peters-Wendisch et al., 2001), which provides L-lysine precursor oxaloacetate by the anaplerotic reaction from pyruvate (Peters- Wendisch et al., 1998).The anaplerotic reaction from PEP was shown to be dispendable for L-lysine production from glucose, however, it might play an important role if glucose is phosphorylated by use of ATP or polyphosphate and not PEP, which was shown to enhance L-lysine production and might elevate PEP availability (Lindner et al., 2011).Vice versa also an inactivation of the gene for PEP carboxykinase, catalyzing decarboxylation of oxaloacetate to PEP, entailed increased L-lysine production (Riedel et al., 2001).NADPH supply is very important for L-lysine production, as four molecules of NADPH are needed for one molecule L-lysine.The main path of NADPH generation is the oxidative branch of the pentose phosphate pathway (Marx et al., 1996), where NADP is reduced to NADPH by glucose-6-phosphate dehydrogenase and 6-phosphogluconate dehydrogenase, thus to enhance L-lysine production numerous attempts have been made towards increasing the pentose phosphate pathway flux, hence NADPH availability, hence L-lysine production.A deletion of the phosphoglucose isomerase gene drives the complete flux from glucose-6-phosphate into the pentose phosphate pathway and was shown to increase L-lysine production but to the cost of reduced growth (Marx et al., 2003).Redirection of the glycolytic flux towards the entry of the pentose phosphate pathway was furthermore achieved by overexpression of the fructose-bisphosphatase gene (Becker et al., 2005, Georgi et al., 2005) as well as use of feedback resistant variants of glucose-6-phosphate dehydrogenase and 6-phosphogluconate dehydrogenase (Becker et al., 2007, Ohnishi et al., 2005).Also the increase of NADP availability by overexpression of a NAD kinase gene resulted in increased L-lysine production (Lindner et al., 2010).To establish L-lysine production from glycerol Rittmann et al. introduced the Escherichia coli glycerol utilization genes in a metabolic engineered C. glutamicum L-lysine producing strain (deregulated Llysine pathway and higher anaplerotic from pyruvate to oxaloacetate).L-lysine yields were slightly lower from glycerol as glucose, 0.19 g g -1 compared to 0.26 g g -1 , respectively (Rittmann et al., 2008).Glycerol has also been used as a source of carbon for the production of the polymer of L-lysine -Poly-L-lysine with Streptomyces sp.(Chen et al., 2011b, Chen et al., 2011a).-Poly-L-lysine is an antimicrobial agent against bacteria, yeasts, and viruses (Shima et al., 1984) and therefore interesting for the pharmaceutical industry (Shih et al., 2004) and it is used as food preservative.The main use of the aromatic amino acid phenylalanine is in production of the sweetener aspartam.Biosynthesis of aromatic amino acids from PEP and erythrose-4-phosphate involves the shikimic acid pathway and dedicated terminal biosynthesis pathways for tryptophan, tyrosine and phenylalanine (Sprenger, 2007).Biosynthesis of aromatic amino acids e.g. in Escherichia coli and C. glutamicum was engineered e.g. by gene deregulation (Berry, 1996, Herry & Dunican, 1993), by gene copy number increase (Chan et al., 1993), and by the use of feedback-resistant enzyme variants, e.g.variants of 3-deoxy-D-arabinoheptulosonate 7-phosphate synthase, the first enzyme of the shikimic acid pathway, variants of anthranilate synthase of the tryptophan pathway in E. coli (Tribe & Pittard, 1979) or anthranilate phosphoribosyltransferase of the tryptophan pathway in C. glutamicum (O'Gara & Dunican, 1995).In addition, strains were engineered for increased supply of the precursors PEP and erythrose-4-phosphate.In C. glutamicum, PEP availability was increased in PEP carboxylase mutants and erythrose-4-phosphate concentrations were elevated by overexpression of the transketolase gene (Ikeda & Katsumata, 1999, Ikeda et al., 1999, Katsumata & Kino, 1989).Similar approaches were made in E. coli, where PEP carboxylase or pyruvate kinase gene knock outs and overexpression of PEP carboxykinase increased PEP supply (Miller et al., 1987, Gosset et al., 1996, Chao & Liao, 1993, Backman, 1992).Furthermore, overexpression of genes encoding PEP synthase, PEP carboxykinase, and the use of an ATP-dependent glucose phosphorylation system instead of the PEP-dependent phosphotransferase system had positive effects on the availability of shikimic acid pathway precursor PEP (Patnaik et al., 1995, Liao, 1996, Gulevich et al., 2004).In E. coli, the availability of erythrose-4-phosphate could be increased by overproduction of transketolase and transaldolase or by phosphoglucose isomerase gene disruption (Draths & Frost, 1990, Draths et al., 1992, Lu & Liao, 1997, Mascarenhas et al., 1991, Frost, 1992).Up to now, only Lphenylalanine production from glycerol has been shown, but results might be transferable to the other aromatic amino acids.Similar final concentrations of L-phenylalanine were reported for an engineered E. coli strain regardless of the use of glycerol, glucose or sucrose as carbon source.Notably, a higher yield was reported when glycerol was used (0.58 g g -1 ) as compared to the use of sucrose (0.25 g g -1 ) (Khamduang et al., 2009).Polyamines may be derived from amino acids (Schneider & Wendisch, 2010).While strain development for sugar-based production of polyamines such as the diamine 1,4diaminobutane, which is used e.g. in the polyamide market, has been successful (Schneider & Wendisch, 2010), glycerol-based production of poylamines has not yet been reported.

Hydrogen
Hydrogen production is highly desirable as a source of clean energy to be used, e.g. in fuel cells.Processes for the use of glycerol or crude glycerol respectively are under investigation.Besides microbial strategies to generate H 2 from crude glycerol there are also promising chemicals techniques such as steam reforming, partial oxidation, auto thermal reforming, aqueous-phase reforming, and supercritical water reforming (Xiaohu Fan, 2010).Currently, only low concentrations of glycerol can be used in microbial H 2 production process to avoid that other products like 1,3-propanediol or ethanol are produced along with H 2 .Enterobacter aerogenes HU-101 showed hydrogen yields of 1.12 mol mol -1 using crude glycerol, but at relatively low glycerol concentrations of 1.7 g l -1 (Ito et al., 2005).Mixed cultures isolated from soil or wastewater converted crude glycerol to H 2 with a yield 0.31 mol mol -1 and to 1,3-Propanediol with a yield of 0.59 mol mol -1 .These values are lower than the ones on glucose but similar to the ones with pure glycerol, suggesting that inhibiting substances in crude glycerol may not be a problem in this process (Selembo et al., 2009).Production rates of 0.68 ± 0.16 mmol H 2 l -1 h -1 could be achieved by an evolved E. coli BW25113 frdC negative strain along with some ethanol production (Hu & Wood, 2010).

Glyceric acid
Glyceric acid is a known byproduct of dihydroxyacetone production from glycerol with Gluconobacter oxydans.The path from glycerol to glyceric aid, which might be suitable for chemical applications (Habe et al., 2009a), occurs via two dehydrogenases.First, alcohol dehydrogenase oxidizes glycerol to glyceraldehyde which is subsequently oxidized further to glyceric acid by a so far unidentified enzyme (Habe et al., 2009d).In a screen of various acetic acid bacteria Acetobacter tropicalis was the best glyceric acid producing strain (Habe et al., 2009b).A. tropicalis produced 101.8 g l -1 glyceric acid while Gluconobacter frateurii accumulated 136.5 g l -1 (Habe et al., 2009d).The involvement of a membrane-bound alcohol dehydrogenase in glyceric acid production was investigated with G. oxydans IFO12528.Gene disruption of the alcohol dehydrogenase entailed severely reduced glyceric acid concentrations, indicating a role of the alcohol dehydrogenase in glyceric acid production (Habe et al., 2009d).G. frateurii was engineered for glyceric acid production by disruption of the glycerol dehydrogenase gene sldA, thus, eliminating dihydroxyacetone production.The growth retardation of this strain on glycerol alone was overcome by addition of sorbitol to the medium.A higher glyceric acid concentration of 89.1 g l -1 was reached with the sldA mutant compared to the parental strain (54.7 g l -1 ) as production dihydroxyacetone as byproduct was avoided (Habe et al., 2010b).Glyceric acid production from raw glycerol pretreated with activated charcoal by Gluconobacter sp.NBRC3259 reached comparable concentration of glyceric acid (45.9 g l -1 and 54.7 g l -1 ) and of dihydroxyacetone (28.2 g l -1 and 33.7 g l -1 ) as production from pure glycerol (Habe et al., 2009c).

Biosurfactants
Surfactants are used in numerous applications such as cleaners, emulsifiers, in coatings, laundry detergents, or in paints.The global surfactant market is predicted to reach 17.9 billion dollars by 2015 (Global Industry Analysts, 2010).Biosurfactants consist of a hydrophilic part and a hydrophobic/lipophilic part, making them amphiphiles/tensides.The hydrophilic part can consist of a sugar, peptide or protein, while the hydrophobic part contains fatty acids or fatty alcohols.The great advantage of biosurfactants over chemically produced tensides is that they are biodegradable, hence environmentally friendly, less toxic, and they can be produced from renewable resources.Natural biosurfactant producers are bacteria, yeast, and fungi (Mulligan, 2005).Biosurfactants are already used in many applications, e.g. in the food, agriculture, chemical, pharmaceutical, and cosmetic industries (for recent review see (Pacwa-Plociniczak et al., 2011)).Glycerol is the backbone of the lipid component of biosurfactants and use of pure and raw glycerol in biosurfactant production has been investigated with a variety of organisms.The yeast Pseudozyma antarctica produced 16.3 g l -1 glycolipid biosurfactants from glycerol (Morita et al., 2007).Glycolipid surfactants production was also shown with Candida bombicola and product concentrations of 12.7 g l -1 of sophorolipids could be obtained from glycerol and oleic acid (Ashby & Solaiman, 2010).When a biodiesel co-product stream consisting of 40% glycerol, 34% hexane-solubles, and 26% water was used production of sophorolipids by C. bombicola was strongly increased (from 9 g l -1 to 60 g l -1 ) (Ashby et al., 2005).Production of glucosylmannosyl-glycerolipid with Microbacterium spec.was reported to be 1.5-fold higher when glycerol instead of glucose is used in media containing the complex medium compounds peptone and yeast extract (Lang et al., 2004, Wicke et al., 2000).When several carbon sources were analyzed for rhamnolipid production by Pseudomonas aeruginosa EM1, glycerol was the best carbon source besides glucose, yielding 4.9 and 7.5 g l -1 respectively (Wu et al., 2008).Glycerolbased biosurfactant production with Pseudomonas aeruginosa UCP0992 yielded 8.0 g l -1 biosurfactants (Silva et al., 2010), processes with Rhodococcus erythropolis yielded 1.7 g l -1 biosurfactants (Ciapina et al., 2006) and with Ustilago maydis 32.1 g l -1 glycolipid biosurfactants from 50 g l -1 glycerol could be produced (Liu et al., 2011).

Conclusions
Glycerol availability has increased tremendously as it arises as byproduct of the biodiesel process.Besides using glycerol as a chemical in creams and other small-scale applications, glycerol may be used as starting material for large-scale biotechnological processes.Several microbiological process are based on glycerol as substrate anyway, e.g. the biotechnological production of 1,3-propanediol and dihydroxyacetone.In addition, as glycerol is a good source of carbon and energy for growth of several microorganisms it may be suitable for the biotechnological production of a number of chemicals in fermentative processes.Microbial catalysts have been developed for the production of succinic acid, citric acid, glyceric acid, propionic acid, ethanol, 1,2-propanediol, 2,3-butanediol, biosurfactants and amino acids.Successful implementation of these processes critically depends on strain optimization, recalcitrance to inhibitors present in crude glycerol preparations from biodiesel factories and measures to reduce glycerol price volatility.If successful, coupling biodiesel production to the production of value-added products from the side-stream glycerol "on the spot" would represent an excellent example of applying the biorefinery concept to a large-scale process.