A gene deletion experiment and its effect on cellular behavior can be simulated in a manner similar to the linear optimization of growth (or for any other objective function). The results can be used to guide the design of metabolic engineering strategies. Gene-reaction associations in the model describe the relationship between genes and their corresponding reactions; therefore, reactions can be removed from the model on the basis of individual gene deletions. The possible results from a simulation of a single gene deletion are unchanged maximal growth (nonessential), reduced maximal growth, or no growth (“sick” or lethal effect). The gene deletion analyses can be carried out genome-wide with the outcome tabulated to provide a comprehensive overview of gene essentiality for the system.

In certain cases, mutations in two genes produce a phenotype that is surprising in the light of each mutation’s individual effect. This phenomenon, which defines genetic interaction, can reveal functional relationships between genes and path­ways. For example, double mutants with surprisingly slow growth define syner­gistic interactions that can identify compensatory pathways or protein complexes (Harcombe et al. 2013). Like single gene deletions, double gene deletions can be simulated to encompass all possible double gene deletions in the network. This provides a powerful tool to simulate otherwise prohibitively difficult wet-bench genetic interaction experiments.

The addition of lime aims to produce two effects:

1. Raising pH to pHc. This is species dependent and varies from 7.97 for Chla — mydomonas reinhardtii (wild type) to 10.2 for Nannochloris (Schlesinger et al. 2012). The amount of lime can be reduced by ceasing CO2 addition and allowing the available inorganic carbon to be photosynthesized to produce an increase of almost two pH units in the microalgal culture.

2. Inducing autoflocculation. This is also species dependent and, according to Schlesinger et al. (2012), varies from 6.07 pM (0.4 mg/L of Ca(OH)2) for Chlamydomonas reinhardtii (wild type) to 16.2 pM (1.20 mg/L of Ca(OH)2) for Nannochloris.

The amount of lime to be added to attain pHc (and avoiding a starvation step prior to autoflocculation) is as follows:

{(i) . [(> X l°-pHc) — (1 X 10—)] . (gCa(OH’).)

As Ca(OH)2 is a strong base, it completely dissociates in water, and for each mole of H+ to be neutralised, half a mole of Ca(OH)2 is needed, with 74 being the molecular weight of Ca(OH)2 and 1 + nCa(OH)2 the proportion of Ca(OH)2 in lime. The appropriate molarity of Ca(OH)2 (Mautoflocc) should be added to induce auto­flocculation, and experimental data show that it is at most 16.2 pM (1.20 mg/L Ca (OH)2) for Nannochloris (Schlesinger et al. 2012). Following this, the expression becomes

Sukenik and Shelef (1984) determined that at least 0.1-0.2 mM phosphate and 1.5-2 mM Ca+2 must be present in the culture in order for autoflocculation to proceed. These concentrations are already present in culture mediums of sea water (Ca in sea water 411 ppm = 10 mM), particularly with the addition of N/P
fertilisers. Using this approximation, we can compare the amount of lime in various Ca(OH)2 purities to be added to different microalgae cultures. Assuming the most unfavourable conditions (no starvation to raise pH, a low cell density, a high lime consuming species, a low purity (60 %) Ca(OH)2 lime), we would have to add a theoretically small 2.004 mg/L (or only around 2 tonnes per GL) of lime to increase from pHculture = 7 to pHc =11, including the 1.2 mg/L (Nannochloris sp.) to induce autoflocculation. This calculation ignores the losses associated with mixing for uniform lime distribution. Table 13.1 presents a list of nine microalgal cultures and required amounts of lime to initialise autoflocculation for a range of species. The authors note that the required lime per unit volume of around 2 tonnes per GL is a relatively small amount with correspondingly low associated product and transport costs, particularly when compared to the tonnes of lime applied per ha over years to mitigate acidic soils in conventional terrestrial agriculture.

13.7 Conclusion

Schlesinger et al. (2012) demonstrated the economic potential of low-cost micro­algal dewatering by autoflocculation. Because the underlying mechanisms are reasonably well understood, there is the potential to optimise autocoagulation strategies for a given microalgae and culture composition. Critical to success is an upfront analysis of the species-specific characteristics of the microalgae in question (Brady et al. 2014; Folkman and Wachs 1973). Note lastly that there is the potential for autocoagulation to increase the ash content of the harvested microalgae which might complicate subsequent processing. The ash content will be the lowest when autocoagulation occurs by surface charge reversal and higher when autocoagulation occurs by mineral precipitation. Mild acid rinses of harvested microalgae should lessen the ash content in both cases.

Acknowledgements Funding from the Sandia National Laboratories LDRD Office is gratefully acknowledged by PVB. Sandia is a multiprogram laboratory operated by Sandia Corporation, a Lockheed Martin Company, for the United States Department of Energy’s National Nuclear Security Administration under contract DE-AC04-94AL85000.

16.2.1 Historical and Current Microalgae Digestion Perspective

The first authors to report anaerobic digestion of microalgae biomass was Golueke et al. (1957) where authors investigated the anaerobic digestion of Chlorella vulgaris and Scenedesmus grown as part of a wastewater treatment process. Anaerobic digestion of microalgae was then continued in a series of studies on advanced integrated large-scale wastewater pond production, with resultant mic­roalgae biomass used for anaerobic digestion as a method of converting solar energy to chemical energy (Benemann et al. 1977; Chen and Oswald 1998; Golueke and Oswald 1963; Golueke et al. 1964; Green et al. 1995a, b; Oswald 1976; Oswald et al. 1994). In general, they identified several key challenges that could hinder the optimal digestion of microalgae biomass. Table 16.1 provides a summary of the research on several different microalgae species to date, and highlights the biogas potential from microalgae. The variation in biogas volumes produced reported in literature illustrates the variation in biogas production between both different species of microalgae and variability within the same species (Table 16.1). The production variation for a single species is highlighted for

C. vulgaris, with the anaerobic digestion in five separate experiments yielding five different biogas volumes per gram of volatile solids (VS) for each experiment, ranging from a low of 189 mL/g/VS up to 450 mL/g/VS. It also highlights how the pre-treatment, digester configuration and resultant microbial communities can impact on the final biogas production.

When investigating biogas production from microalgae biomass, several dif­ferent terminologies are utilised to report the biogas production from microalgae substrates. Units range from biogas production for grams of COD destroyed, biogas produced for each gram of VS added to digester and biogas produced for each gram of total solids added to digester. The standard methods used to determine VS are also the same as standard methods used to determine the organic weight (ash free dry weight—AFDW) of microalgae (Clesceri et al. 1998). Ash free dry weight is used extensively by phycologists to report quantities of microalgae biomass. When reporting microalgae biomass, the organic weight or the VS (digestible component) of the microalgae biomass is only a percentage of the total solids and ash content varies between species. The variation in AFDW and VS can differ by up to 50 % between species and can significantly affect calculating or modelling the biogas production from different microalgae species (Ward et al. 2014). Several key issues have been identified in scientific literature that may be detrimental to the anaerobic digestion of microalgae biomass and result in reduced biogas production. The following sections highlight the major factors that can influence biogas production from microalgae.

Ela Eroglu, Steven M. Smith and Colin L. Raston

Abstract Immobilized cells entrapped within a polymer matrix or attached onto the surface of a solid support have advantages over their free-cell counterpart, with easier harvesting of the biomass, enhanced wastewater treatment, and enriched bioproduct generation. Immobilized microalgae have been used for a diverse number of bio­processes including gaining access to high-value products (biohydrogen, biodiesel, and photopigments), removal of nutrients (nitrate, phosphate, and ammonium ions), heavy metal ion removal, biosensors, and stock culture management. Wastewater treatment processes appear to be one of the most promising applications for immo­bilized microalgae, which mostly involve heavy metal and nutrient removal from liquid effluents. This chapter outlines the current applications of immobilized mic­roalgae with an emphasis on alternative immobilization approaches. Advances in immobilization processes and possible research directions are also highlighted.

2.1 Introduction

Algal bioprocesses are advantageous in integrating wastewater treatment processes with valuable biomass production. Algal biomass can be further exploited for various purposes such as biofuel generation in the form of biodiesel, biohydrogen, or biogas;

E. Eroglu (H) • S. M. Smith

School of Chemistry and Biochemistry, The University of Western Australia, Crawley,

WA 6009, Australia e-mail: elaeroglu@gmail. com

E. Eroglu • S. M. Smith

ARC Centre of Excellence in Plant Energy Biology, The University of Western Australia, Crawley, WA 6009, Australia

C. L. Raston

Centre for NanoScale Science and Technology, School of Chemical and Physical Sciences, Flinders University, Bedford Park, SA 5042, Australia

© Springer International Publishing Switzerland 2015

N. R. Moheimani et al. (eds.), Biomass and Biofuels from Microalgae,

Biofuel and Biorefinery Technologies 2, DOI 10.1007/978-3-319-16640-7_2 food additives; slow-release fertilizers or soil conditioners; cosmetics; pharmaceu­ticals; and several other valuable chemicals (Johnson and An 1991; Mallick 2002; Mulbry et al. 2005). Microalgal cells have several other advantages in not requiring many resources to generate their biomass, providing an economical operation at lower costs, with the dissolved oxygen released by the algae being useful to elevate the oxygen levels of water effluents, and can be utilized for the reduction of carbon dioxide emissions using CO2 for their biomass and/or energy production.

Harvesting and dewatering of algal biomass from its liquid environment is one of the major challenges of algal bioprocesses. Several studies have focused on harvesting of microalgae using a wide range of technologies from sand filtration to high-speed centrifugation (Mallick 2002; Oswald 1988). Some of the most recent technologies for algal dewatering are further discussed in Chaps. 12, 13, and 14 of this book. Immobilization of algal cells has been proposed mainly to overcome the burdens of difficult harvesting and dewatering stages, in addition to providing the retention of the high-value-added algal biomass for further processes (de la Noue and de Pauw 1988; Mallick 2002).

Immobilization of various cells in either polymeric or biopolymeric matrices has several advantages over their free-cell counterparts, since immobilized cells occupy less space, are easier to handle, and can be used repeatedly for product generation (Mallick 2002). Immobilization of cells has also been proposed to increase the biosorption capacity and bioactivity of the biomass (Akhtar et al. 2004; de-Bashan and Bashan 2010). It allows bioprocesses with higher cell densities and also easy harvesting of biomass from its liquid environment (Mallick 2002). Immobilization processes have several other advantages as being resistant to harsh environments such as salinity, metal toxicity, and pH; protecting the aging cultures against the harmful effects of photoinhibition; yielding higher biomass concentrations; recov­ering the cells in a less-destructive way; and enhancing the cost-effectiveness of the process by reusing the regenerated biomass (Bailliez et al. 1986; Hall-Stoodley et al. 2004; Liu et al. 2009). Given the use of large-scale bioreactors represents a significant challenge associated with algal biomass recovery, immobilization sys­tems are becoming attractive alternatives for scale-up processing (Christenson and Sims 2011; Hoffmann 1998).

Various immobilization processes are in use, such as adsorption, confinement in liquid-liquid emulsions, capturing with semipermeable membranes, covalent cou­pling, and entrapment within polymers (de-Bashan and Bashan 2010; Mallick

2002) . Among others, the most common immobilization processes are the entrap­ment of the cells within polymeric matrices and self-adhesive attachment of cells onto the surfaces of solid-supports (Godlewska-Zylkiewicz 2003). Both synthetic and natural polymers can be applied as the immobilization matrix (de-Bashan and Bashan 2010).

Important criteria for successful entrapment are to set the algal cells free within their partition, while pores inside the gel matrix allow the diffusion of substrates and the metabolic products toward and from the cells (Mallick 2002). Nevertheless, entrapment systems still hold some drawbacks in reducing the mass transfer kinetics of the uptake of metal ions (Aksu et al. 2002). However, these can be avoided by careful choice of the immobilization method and the nature of the matrix, which will be further discussed in detail.

Key applications suggested for immobilized algal cells include removal of nutrients from aqueous solutions (Chevalier and de la Notie 1985), biodiesel pro­duction (Bailliez et al. 1985; Li et al. 2007), biosorption of heavy metals from wastewaters (de-Bashan and Bashan 2010), photoproduction of hydrogen and photopigments (Bailliez et al. 1986; Laurinavichene et al. 2008), providing an alternative technique to the common cryopreservation processes (Chen 2001; Faafeng et al. 1994; Hertzberg and Jensen 1989), and also toxicity testing (Boze­man et al. 1989). These processes will also be discussed in detail in the following sections of this review article.

When nutrient concentrations are reported for wastewater, total concentrations of nutrients are often reported. Dissolved inorganic nutrients, such nitrate and ammo­nium N and phosphate P are directly available to microalgae. Part of the total N and P in wastewater, however, may be associated with organic molecules (dissolved organic nutrients) or with particulate matter (either organic or inorganic). These nutrients are not necessarily available to microalgae. Relatively few studies have investigated the bioavailability of nutrients to microalgae in different types of wastewater.

The bioavailability of P is highly variable and may vary between 3.4 and 81 % in different types of wastewaters or surface waters (Ekholm and Krogerus 2003). The bioavailability of total P depends on the dominant P-forms that are present (Van Moorleghem et al. 2013; e. g., Li and Brett 2013). Polyphosphates and organic phosphate monoesters, for instance, have a high bioavailability. Microalgae are capable of producing phosphatase enzymes to dissociate phosphate ester bonds (e. g., Huang and Hong 1999). Phytic acid, on the contrary, has a very low bio­availability and is a major phosphate reserve in plant seeds. As it cannot be digested by livestock, it is present in high concentrations in manure from grain-fed livestock (Jongbloed and Lenis 1998). To increase the bioavailability of P from phytic acid to livestock, phytase enzymes are sometimes added to the animal feed, and this may result in a higher bioavailability of P in the manure. If the wastewater contains a lot of sediments, phosphate may be chemically bound to iron (Fe) or aluminum (Al), and this phosphate has a low bioavailability. Phosphate that is precipitated with calcium (Ca) as apatite minerals also has a low bioavailability to microalgae, despite being often detected by colorimetric phosphate analyses. Wastewater from animal manure has a high content of humic substances and these can also bind phosphate. It is assumed that phosphate is bound to oxidized Al or Fe ions that are stabilized by humic acid complexes (Gerke 2010). These humic acid-metal phos­phate complexes also have a low bioavailability to microalgae.

Less information is available about the bioavailability of organic N forms. About half of the N in animal manure, wastewater may be present as organic N (Cai et al.

2013) . Free amino acids, nucleotides, as well as urea are organic N forms that are highly bioavailable to microalgae. Peptides or proteins have a slightly lower bio­availability, and humic substances contain N and exhibit even less bioavailability (Bronk et al. 2007). Humic substances have a C:N ratio of 18-30:1 for humic acids and 45-55:1 for fulvic acids (See and Bronk 2005). Humic substances can absorb amino acids and ammonium ions, which can represent about half of N in humic substances. Part of the organic N present in wastewater is slowly made available by bacteria that live in symbiosis with microalgae (Pehlivanoglu and Sedlak 2004).

MAGE is a genetic engineering method that relies on recombineering to produce frequent and large scale genetic changes to cells. Simply, it is a cyclical and scalable recombineering system that allows multiple genetic changes in a high throughput manner. A given cycle in MAGE requires cell growth, the introduction of recombineering substrate, and provision of providing synthesized DNA con­structs a continuous and interlinked process. Target cells are continuously grown and drawn out of a cell growth chamber into an exchange chamber where the substrates required for recombineering are added under the right conditions. The induced cells are then mixed with the recombinant and synthetic DNA constructs in another chamber and then moved to an electroporator to induce the uptake of the synthetic DNA constructs. Those modified cells are then reintroduced into the growth chamber and grown. This whole process occurs continuously; and move­ment to each stage is facilitated by microfluidics, or any suitable pumping assembly (Isaacs et al. 2011).

Just like recombineering, MAGE can be used to create a mismatch, deletion or insertional genetic changes. The efficiency of each genetic change is dependent on the size of the homology flanking each introduced DNA construct, the size of the desired change, and the number of cycles. This efficiency is also correlated with the Gibbs free energy from the hybridization that occurs between the DNA strands. Knowledge of the efficiency of each genetic change and the parameters that affect it can be used to tune the MAGE process to produce colonies or strains with desired genetic traits (Isaacs et al. 2011). The MAGE system was initially demonstrated by modifying a 1-Deoxy-D-xylulose 5-phosphate (DXP) biosynthesis pathway in Escherichia Coli. The DXP pathway was modified to increase production of isoprenoid lycopene. New genetic modifications were introduced in greater than 30 % of the cell population every 2-2.5 h under optimum conditions. After just five cycles of MAGE, the average genetic change across the entire cell population was

3.1 bp, and increased to 5.6 bp afters 15 cycles. Ultimately, 24 genes were opti­mized simultaneously and about 15 billion genetic variants were produced at an average rate of 430 million bp changes per MAGE cycle in a total of 35 cycles. This translated to an up to 390 % increase in isoprenoid lycopene production, clearly demonstrating MAGE’s potential (Isaacs et al. 2011). Development of a system similar to MAGE in algae is dependent on the design of a suitable genetic engi­neering method; recombineering strategies are not established in algae.

The downstream metabolic pathways which utilise the ATP and NADPH generated by photosynthesis offer a second area of opportunity for genetic improvement. Essentially, biocrude is reduced carbon dioxide and photosynthesis is a process of reduction. Metabolic engineering based, for example, on flux analysis (Dal’Molin et al. 2011) offers the opportunity to increase the capture of the energy derived from solar irradiation into desired ‘chemical energy-rich’ product (e. g. triacylglycerides, secondary metabolites or recombinant proteins) in production models utilising extraction processes or for sale of raw biomass.

To achieve the maximum PCE for a particular product, it is important to achieve four key things:

1. The optimised channelling of the captured solar energy into the biochemical pathway producing the target product

2. The optimisation of the efficiency of the biochemical pathway producing the product

3. To minimise energetic losses through cell division and metabolism

4. To export the product to avoid inhibitory feedback loops

It could be argued that while the engineering of pathways for production models utilising extraction processes or for the sale of biomass is likely to benefit most from genetic optimisation, the benefits of metabolic engineering are attenuated when whole biomass processing strategies such as HTL to biocrude are intended. However, even in this scenario, the elemental composition of microalgal biomass is still relevant to the quality of biocrude output and biomass. For example, oil-rich biomass can be considered to be more reduced than its more oxygenated carbo­hydrate-rich counterpart and as such tends to have a higher energy content or calorific value in terms of MJ kg-1. Thus, for example, while TAG accumulation has traditionally been examined as a feedstock for biodiesel production, for which it is particularly suited, its accumulation can also be beneficial for the production of biocrude. Similarly, protein accumulation results in high nitrogen and sulphur levels which incurs additional nutrient requirements and poses a problem for bio­crude quality. Generally, algae are considered to be * 50 % carbon; however, this is not an inflexible rule. The Redfield ratio purports that the stoichiometric ratio of main elements is expected to be 106C:16N:1P; however, this is a molar ratio, and when masses are calculated, the corresponding mass ratios would be approximately 41C:7N:1P. Thus, algae can vary significantly in their carbon content, and where algae can accumulate large amounts of lipid stores, then a reasonable carbon range could be anticipated from *40-60 % carbon. In this respect, where carbon assimilation can be increased without compromising total productivity, metabolic engineering can play an important role in strain development for whole biomass processing.

Balancing Carbon Storage in Microalgae (Case Studies) Studies have shown that the content and biochemical composition of microalgae can vary significantly as their metabolic pathways shift in response to environmental stimuli and changes in nutritional conditions—protein (6-52 %), carbohydrate (5-23 %) and lipid (7-23 %) (Brown et al. 1997; Chen et al. 2011; Guschina and Harwood 2006; Johnson and Alric 2013; Poerschmann et al. 2004; Roessler 1990). While some algae use lipids as energy storage, e. g. diatoms, for many species cultivated under nutrient replete normal growth conditions, carbon reserves are preferentially stored as starch (as in higher plants) which are readily catabolised to glucose for ATP (Johnson and Alric 2013). Though in nutrient deplete unfavourable growth con­ditions where temperature, light, salinity, pH, etc., may be suboptimal, the carbon can be stored in the form of triacylglycerols as a reserve for future use (Roessler 1990; Sheehan et al. 1998). By taking a nutritional and/or an environmental approach, the regulatory processes involved in these carbon storage schemes can be manipulated, although the effects may be species specific. While some of these approaches may not be practical in a commercial context, it is interesting to understand genetic activities related to such phenomena.

Nitrogen depletion was first demonstrated by Spoehr et al. in Chlorella pyre- noidosa and has since been applied to a number of other strains as the standard for lipid induction (Ben-Amotz et al. 1985; Spoehr and Milner 1949). Studies have shown significant increases in lipid content (from 1 to 85 % of cellular mass) in microalgal cultures at the expense of cellular growth (Rodolfi et al. 2009; Roessler 1990; Siaut et al. 2011). As the synthesis of proteins and nucleic acids rely on nitrogen, these pathways inevitably cease to function and the flow of carbon diverts to storage compounds (Berges et al. 1996; Roessler 1990). Similarly, silicon depletion in diatoms such as Cyclotella cryptica has been observed to induce a multifold increase in ACCase activity leading to increased lipid accumulation, while a reported 50 % reduction in carbohydrate storage was noted. Pulse chase experiments suggested in the first 12 h after silicon depletion, *55-68 % of the lipids produced via de novo synthesis (Roessler 1988a, b and 1990).

The effects of salinity on metabolism have been mainly studied in halotolerant species such as Dunaliella which is found to draw upon starch reserves to physi­ologically respond to osmotic stress (Craigie and McLachlan 1964). As salinity within the medium is increased, Dunaliella cells contract rapidly, and the cells then metabolise glucose and fructose into a glycerol pool within the cytoplasm in order for the cell to regain its volume (Baird and DeLorenzo 2010). On the other hand, when salinity in the medium is decreased, the available glycerol pool is metabolised back into starch reserves (Ben-Amotz and Avron 1973; Borowitzka and Brown 1974). Other growth factors such as temperature stress, pH variation and light intensity have also been reported to influence the lipid composition in a range of species (Guckert and Cooksey 1990; Roessler 1990).

Knowledge in how different species respond to different nutrients and envi­ronmental conditions will greatly aid researchers in developing a greater under­standing of mass culture and biological response. The disadvantage of these simple strategies for increasing lipid content is that they are fundamental responses to stress, represent a loss of net productivity and can be subject to operational limi­tations. However, elucidating the genomic changes elicited by these responses can enable engineering strategies which offers the prospect of more rapid, direct and controllable ways to siphon off the biological gains of photosynthesis into a desirable form.

Engineering of Lipid Pathways While much is known about lipid metabolism in higher plants from research models such as Arabidopsis (Beisson et al. 2003), lipid metabolism in microalgae is substantially different relative to higher plants and also between microalgal genera. The neutral and polar lipids, and the enzymes and metabolic pathways involved in their biosynthesis and catabolism have been recently described with the current focus being upon gene identification to enable proper metabolic engineering (Dal’Molin et al. 2011; Guschina and Harwood 2013; Khozin-Goldberg and Cohen 2011; Liu and Benning 2013; Rismani-Yazdi et al.

2011) . C. reinhardtii remains the most extensively studied model for microalgal lipid metabolism (Liu and Benning 2013; Merchant et al. 2012), but it does not appear to use phosphatidylcholine as a substrate in TAG synthesis or to accumulate TAG unless under stress conditions (unlike Nannochloropsis which can synthesise TAG under normal cultivation conditions) or in starch accumulation (sta) mutants (Li et al. 2010; Work et al. 2010; Zabawinski et al. 2001). Recently, the metabolic pathways of Dunaliella tertiolecta (Rismani-Yazdi et al. 2011) and Monoraphidium neglectum (Bogen et al. 2013) have been reported, and there is some knowledge of other microalgae (Guschina and Harwood 2013), but knowledge of specific enzy­matic processes and the genes involved requires further advancement before effective metabolic engineering strategies become commonplace and metabolic flux models will also assist with this (Dal’Molin et al. 2011).

In early work on acetyl-CoA carboxylase (ACCase) in which Dunahay et al. (1996) transformed the diatom Cyclotella cryptica with additional copies of the ACCase gene within the TCA cycle to increase the flux of carbon towards lipid biosynthesis, the resultant increase in enzyme activity did not increase lipid accu­mulation. Similarly, the recent engineering of increased expression DGAT strains in C. reinhardtii (La Russa et al. 2012) saw enhanced DGAT mRNA levels but failed to increase intracellular TAG accumulation. Thus, claims that this field of engi­neering lipid metabolism in microalgae is still in its infancy are valid (Merchant et al. 2012), and much more work needs to be completed before we are likely to see the potential gains in industrial production candidates that have been anticipated. Possible reasons for this are that microalgae have some form of feedback regulation and that they have multiple and divergent DGAT2 isoforms (Chen and Smith

2012) , and while all algal species have at least one DGAT2 from the animal clade, it currently appears that only green algae have DGAT2s similar to higher plants. Thus, single gene engineering strategies may be of limited application when dealing with a gene network which we do not currently comprehend sufficiently. Chen and Smith (2012) call for further investigation of DGAT2 enzymatic characteristics as functionality and substrate preferences are currently not fully understood. PDAT is the other of the two enzymes involved in the final step of TAG production and it has recently been examined (Boyle et al. 2012; Yoon et al. 2012). Yoon et al. (2012) demonstrated that PDAT is indeed involved in TAG biosynthesis in C. reinhardtii through RNAi-induced PDAT knock-down mutants. Thus, both DGAT and PDAT represent valid targets for metabolic engineering, but more must be understood about the other metabolic processes acting in the TAG ‘neighbour­hood’. By simultaneously manipulating all of the critical genes that influence the metabolic flux, success is far more likely. RNAi and new CRISPR/Cas and TALEN technologies offer the potential to dissect these pathways and indeed optimise individual catalytic steps through genetic editing and the amino acid level.

Another promising genetic approach has focused upon the engineering of lipid catabolism rather than biosynthesis. Using this approach, Trentacoste et al. (2013) incorporated antisense and RNAi into the diatom Thalassiosira pseudonana tar­geting a newly identified gene Thaps3_264297, which was reported to be a mul­tifunctional lipase-phospholipase-acyltransferase, which showed consistent decrease in microarray transcript abundance throughout the lipid accumulation phase of silicon withdrawal. Thaps3_264297 is homologous to human CGI-58 whose mutation in humans can lead to excessive accumulation of neutral lipid droplets in various tissues. Trentacoste et al. (2013) found that these knock-down mutants had increased accumulation of TAG droplets and total lipid production without negatively affecting cell division and biomass growth. Examples of other targetable enzymes may include malate dehydrogenase (mme gene), pyruvate for­mate-lyase (pfl gene) or the fatty acid synthase complex (FAS) to drive carbon towards fatty acid synthesis (Perez-Garcia et al. 2011; Yu et al. 2011), and these targets will be further refined with the improvement of metabolic flux models.

Metabolic engineering is not just about increasing the flux towards fatty acid synthesis and TAG accumulation, but also qualitatively concerning the types of lipids that are produced. For example, metabolic engineering has been successful in altering the fatty acid profile of Phaeodactylum tricornutum (Radakovits et al.

2011) to yield shorter acyl chains. The capacity to manipulate both chain length and the degree of saturation has significant potential for adjusting fuel properties.

Engineering of Carbohydrate Pathways Similar to the metabolic models for fatty acid biosynthesis and catabolism, models for carbohydrate metabolism are also under-development, for example, in Phaeodactylum (Kroth et al. 2008). For bio­crude production, increasing the carbohydrate concentration in cells can also increase the total carbon content although oxygen content increases and this strategy seems secondary relative to strategies maximising lipid content. In the C. reinhardtii Stm6 mutant, the deletion of the Moc1 gene via random gene insertion resulted in modified respiration metabolism with the downstream effect of accu­mulating large starch reserves within the chloroplast (Schonfeld et al. 2004). There are also ambitions for microalgae, in particular cyanobacteria, to produce carbo­hydrates at the industrial scale (Ducat et al. 2012; Wijffels et al. 2013), and the company Algenol that utilises GM algae to produce and secrete ethanol is a good example of this.

Challenges for Effective Deployment of these Technologies in Commercial Sys­tems Where microalgae are accumulating energy storage compounds, they become better candidates for production; however, within the ecology of mass cultivations, they also become better candidates for predation, increasing the energy return for micro-organisms grazing upon them.

Electric field-driven water hydrolysis generates H2 which forms bubbles that adhere to the microalgae and carries them to the surface to accelerate harvesting (Mollah et al. 2004). This process has been run in both batch and continuous mode and found that higher power positively correlated with increased algal cell separation in eutrophic lake water (Alafara et al. 2002). The advantage of this method is there is no need for the addition of chemicals to induce separation. However, metal ions are released from the electrodes and will accumulate in the biomass. Scrapers similar to those used in the other flotation techniques can be used to remove the biomass from the surface of the water and provide concentrations in the range of 20-70 g/L. The need for electrical power and the positive correlation with increased power combine to make commercial application a challenge.

Once the microalgae biomass has been removed from the water (typically to 20-30 % dw biomass), a number of options are available for the conversion of the biomass into fuel or valuable co-products. These approaches can be broadly cate­gorised into two system types. The traditional approach is to extract the metabolite of interest (typically lipids or vegetable oil) and then convert the oil into a fuel via transesterification (biodiesel) or hydrotreating. The extraction process requires the use of a solvent such as hexane that requires the biomass to be completely dried; as a result, further thermal energy and capital expenditure are required to dry the biomass to below 10 % moisture to allow solvent extraction (Xu et al. 2011). Recently, wet solvent extraction methods have been proposed, which negate the need for drying; however, these have yet to be proved on a large scale (Kanda and Li 2011).

The further cost of drying has led researchers to investigate in situ conversion methods that convert the oil contained in the biomass into biodiesel directly with the biodiesel subsequently separated from the biomass (Ehimen et al. 2010; Johnson and Wen 2009; Levine et al. 2010; Wahlen et al. 2011). Again, these methods have limitations including high solvent consumption, higher processing temperatures and low yields (de Boer et al. 2012).

The challenges associated with extraction have led to further investigation into conversion methods that use the whole algal cell (Biller and Ross 2011; Liu et al.

2013) . Again, approaches can be split into wet and dry, with dry methods requiring the whole biomass to be dried and then converted to biofuels via traditional thermal methods, e. g. pyrolysis and gasification (Babich et al. 2011). In this case, the dried microalgae biomass simply becomes one of many potential feedstocks that can be used in existing biomass to liquid processes. Again, the need to dry the biomass and the fact that these processes are uneconomical with existing low-cost traditional biomass sources has resulted in this approach being all but abandoned for microalgae.

Another method, which is currently being investigated by research institutions and implemented by companies throughout the world, is hydrothermal liquefaction. In this process, originally developed by Shell, wet biomass is converted directly into a hydrocarbon liquid (bio-oil) under relatively mild temperature and pressure conditions, with the oil to be subsequently upgraded to produce biofuel (Goudriaan et al. 2005, 2008). This process uses wet biomass and produces oil at approximately 35-40 % yield (dry biomass basis). The oil can be used as a blend stock in traditional refining equipment, and it is very attractive from an economic viability viewpoint. The major weaknesses in this approach are the complete loss of any valuable co-products and the challenges associated with recycling nutrients (Liu et al. 2013).

An overview of the different extraction and conversion processes discussed above is provided in Fig. 17.2.

In some studies, more than one culture was immobilized to achieve a multifunctional immobilization matrix. For example, Adlercreutz et al. (1982) co-immobilized mixed cultures of algae (Chlorella pyrenoidosa) and bacteria (Gluconobacter oxy — dans) inside calcium alginate beads for the continuous production of dihydroxyac — etone. They did not observe any significant loss of activity within the first six days of this bioprocess. They used the algal cells as an in situ oxygen supplier, which was directly used by the bacteria during the conversion of glycerol to dihydroxyacetone (Adlercreutz et al. 1982). Co-immobilization of microalga S. obliquus with Bacillus subtilis bacteria in carrageenan beads was studied inside air-lift reactors, for enhancing the production of alpha-amylase enzyme (Chevalier and de la Node 1988). Microalgal cells were again used as an in situ oxygen generator for the bacterial cells, which were mainly responsible for the synthesis of alpha-amylase enzyme. Co-immobilization overcame the existing oxygen diffusion problems and yielded higher alpha-amylase activity by a factor of around 20 %. They also observed higher growth rates for the algal cells when co-immobilized with bacteria, compared to the immobilization with algal cells alone (Chevalier and de la Node 1988).

Immobilization of Dunaliella tertiolecta in alginate (Grizeau and Navarro 1986) and Dunaliella salina in agar-agar (Thakur and Kumar 1999) increased the amount of glycerol production. Immobilized algae were also used for the generation of keto acids from amino acids (Wikstrom et al. 1982).

Luan et al. (2006) achieved successful removal (90 %) of a highly toxic tribu — tyltin using alginate-immobilized C. vulgaris cells. They observed that less than 10 % of the tributyltin was accumulated inside the cells, while the remainder was adsorbed by both the immobilization matrix and the cell walls.

He et al. (2014) recently constructed an algal fuel cell with immobilized C. vulgaris cells in sodium alginate placed inside a cathode chamber of the fuel cell. The aim was to achieve a complete process that combines biomass production, electricity generation, and wastewater treatment all at the same time. They observed a significant chemical oxygen demand (COD) removal efficiency of 92.1 %.