Best practices in heterotrophic high-cell-density microalgal processes: achievements, potential and possible limitations.

Microalgae of numerous heterotrophic genera (obligate or facultative) exhibit considerable metabolic versatility and flexibility but are currently underexploited in the biotechnological manufacturing of known plant-derived compounds, novel high-value biomolecules or enriched biomass. Highly efficient production of microalgal biomass without the need for light is now feasible in inexpensive, well-defined mineral medium, typically supplemented with glucose. Cell densities of more than 100 g l(-1) cell dry weight have been achieved with Chlorella, Crypthecodinium and Galdieria species while controlling the addition of organic sources of carbon and energy in fedbatch mode. The ability of microalgae to adapt their metabolism to varying culture conditions provides opportunities to modify, control and thereby maximise the formation of targeted compounds with non-recombinant microalgae. This review outlines the critical aspects of cultivation technology and current best practices in the heterotrophic high-cell-density cultivation of microalgae. The primary topics include (1) the characteristics of microalgae that make them suitable for heterotrophic cultivation, (2) the appropriate chemical composition of mineral growth media, (3) the different strategies for fedbatch cultivations and (4) the principles behind the customisation of biomass composition. The review confirms that, although fundamental knowledge is now available, the development of efficient, economically feasible large-scale bioprocesses remains an obstacle to the commercialisation of this promising technology.

Introduction

Micropan> class="Species">algae, a large and heterogeneous group of microscopic algae, are an almost untapped pool of metabolic versatility. As many of the species occurring in nature have not yet been identified and/or physiologically characterised, their potential awaits exploitation in the biotechnological manufacturing of high-value biomolecules or deliberately enriched biomass (Guedes et al. 2011; Raja et al. 2008; Rosenberg et al. 2008; Wijffels 2008). The term ‘microalgae’ is typically used in its narrowest sense as a synonym for photoautotrophic, unicellular algae utilising CO2 and gaining energy from light. Although certain species are obligate photoautotrophs, numerous microorganisms currently classified as microalgae are in fact obligate heterotrophs (Droop 1974; Gladue and Maxey 1994), and others are capable of both heterotrophic and photoautotrophic metabolism either sequentially or simultaneously (Chojnacka and Marquez-Rocha 2004; Droop 1974; Gladue and Maxey 1994; Lee 2001).

Heterotrophic pan> class="Chemical">cultivation without light and with the controlled addition of an organic source of carbon and energy is similar to procedures established with bacteria or yeasts in multipurpose stirred closed tanks sterilised by heat. To date, only a small number of microalgal species have been cultured heterotrophically in conventional bioreactors (Chen 1996; Perez-Garcia et al. 2011). The few commercialised processes in which microalgae are grown under heterotrophic conditions are focussed on the manufacture of polyunsaturated fatty acids (PUFA) in 100-m3 scale (Behrens 2005). These biotechnological processes represent a sustainable alternative to the extraction of PUFA from fish oil (Apt and Behrens 1999; Barclay et al. 1994; Barclay 1992; Kyle and Gladue 1991; Kyle et al. 1991; Mendes et al. 2009; Wynn et al. 2005). Several other heterotrophic processes that utilise microalgae have been established at laboratory scale to deliberately enrich the biomass with compounds such as pigments and antioxidants (Pulz and Gross 2004; Raja et al. 2008; Spolaore et al. 2006). l-Ascorbic acid (Running et al. 1994) and polysaccharides (Ramus 1972) are examples of commercially valuable extracellular products obtained from microalgae. Classes of compounds that are found in microalgae and that exhibit desirable properties for treating inflammation, tumours and viral or microbial infections are attracting new interest (Guedes et al. 2011). Moreover, research in the rapidly expanding field of biofuels (Wijffels and Barbosa 2010) provides a valuable source of fundamental information on the physiology and biochemistry of microalgae, producing high-value compounds (e.g. Brányiková et al. 2010; Xiong et al. 2010b). The growing interest in microalgae, either non-recombinant or with appropriate genetic modification (Potvin and Zhang 2010; Specht et al. 2010), suggests that heterotrophic microalgal processes offer significant commercial opportunities (Rosenberg et al. 2008).

In contrast to planpan>ts or seaweeds, inpan> whipan> class="Chemical">ch biomass is fairly compact, the harvesting of unicellular microalgae dispersed in natural habitats of microbial consortia is not as straightforward. Low cell densities of several grams per litre are an important cost factor for established production processes with photoautotrophic microalgae in conventional open ponds or photobioreactors (Molina Grima et al. 2003). However, cell densities of more than 100 g l−1 cell dry weight, achieved with Chlorella, Crypthecodinium and Galdieria species, highlight the potential of heterotrophic microalgal processes (de Swaaf et al. 2003c; Doucha and Lívanský 2011; Graverholt and Eriksen 2007; Wu and Shi 2007). Moreover, systematic screening for new compounds is only feasible provided that sufficient quantities of concentrated biomass from axenic (pure) cultures are attainable (Olaizola 2003; Wijffels 2008).

From taxonomic studies, it is apan> class="Chemical">cknowledged that microalgae exhibit considerable metabolic plasticity (Trainor 2009). In response to their surroundings, particular species can occur in alternative phenotypes, and these can result in the altered formation of metabolites and/or products. Thus, the composition of biomass (or intracellular products) or the production of desired extracellular products is typically affected by culture conditions (Hu 2004; Illman et al. 2000; Jakobsen et al. 2008; Lv et al. 2010; Shi et al. 2006; Xiong et al. 2010b; Yongmanitchai and Ward 1991). In turn, this large environmental adaptability provides opportunities to modify the production of targeted natural compounds and to control their formation at high titres, yields, productivities and the required quality (purity). However, screening the various (natural) phenotypes under different conditions is a complex, time-consuming task involving a large number of culture variables. The basic principles of systematic screening were established during studies of the species suitable for use in aquaculture hatcheries (Gladue and Maxey 1994).

Although the opportunities for heterotrophic propan> class="Chemical">cesses with microalgae have been considered in several review papers (Apt and Behrens 1999; Borowitzka 1999; Lee 2001 as well as more recently by Eriksen 2008b and Perez-Garcia et al. 2011), few cover aspects of cultivation technology in depth (e.g. Chen 1996; Chen and Chen 2006). In an attempt to address the outstanding issues, this review paper outlines the current best practices in the heterotrophic high-cell-density cultivation of microalgae for the production of biomass or specific products for health and nutraceutical applications. The main topics dealt with include (1) the characteristics of microalgae suitable for heterotrophic cultivation, (2) the appropriate chemical composition of mineral growth media, (3) strategies for high-cell-density cultivation and (4) the principles of customising biomass composition. Thus, the potential and limitations of fedbatch technology are outlined. The generic process strategies described are based on experimental data collected for non-recombinant microalgae and are, in principle, also applicable to emerging strains improved by genetic engineering.

Cultivation of microalgae in conventional stirred bioreactors

The microalgal spepan> class="Chemical">cies which are currently attracting commercial interest grow under heterotrophic conditions and perform efficiently in conventional bioreactors in a similar manner to bacteria or yeast (Riesenberg and Guthke 1999). Such sophisticated, safe and controllable bioreactor systems are used to produce novel high-value compounds with microalgae. In contrast, established microalgal products are mostly manufactured by traditional outdoor photoautotrophic technologies (Lee 1997). If a product is unique or is not obtainable in the desired quality or quantity by other means (such as extraction from animal or plant material or chemical synthesis), the superior heterotrophic growth characteristics become less critical. Performance losses may also be acceptable in cases where patent infringements need to be prevented. Nevertheless, using microalgae instead of bacteria or yeasts provides the opportunity to establish a new intellectual property claim in the manufacture of competitive products (Borowitzka 1992).

General prerequisites and constraints

In order to be optimally suited for cultivation inpan> pan> class="Chemical">conventional stainless steel stirred bioreactors, a particular microalgal species should meet a number of desirable criteria (Table 1). The primary prerequisite is the ability to grow heterotrophically in an inexpensive, well-defined mineral medium with a high degree of resistance to mechanical and chemical stress.

Table 1

Prerequisites, benefits and constraints of heterotrophic cultivation in conventional stirred bioreactors

	Prerequisites/benefits	Constraints
Bioreactor cultivation	Performance independent of climate	High oxygen demand
	Reduced downstream costs	Sophisticated substrate feed control
	Enhanced productivity and/or titre	Rheological limitations (at high viscosity)
	Control of substrate concentrations	Critical/toxic levels of metabolites
	Scalable process strategies	High costs for (new) equipment
	Use of multi-purpose bioreactors
	Low land requirement
	Indoor and cGMP operation
Culture media	Energy of light not required	Enhanced risk of contamination (organic carbon substrate, temperature, pH)
	Defined (mineral) and inexpensive	Corrosion (high salinity, critical pH)
	Easy to sterilise	Expensive ingredients (vitamins, amino acids)
	Non-corrosive (low salinity, acidity)	Non-defined composition (e.g. yeast extract)
	Contamination protection (due to high salinity, extreme pH levels, high temperature > 40 °C)
Species	Available as axenic culture	Surface adhesion
	Reasonable specific growth rate	Aggregate formation
	Mechanical resistance	Secretion of viscous metabolites
	Temperature achievable with conventional cooling (25–40 °C)	Osmotic stress (at substrate over-dosing)
	Robust and resistant (to long periods of refrigeration, freezing, repeated cultivation, sudden condition changes)	Intracellular product harvest (hampered by rigid cell walls)

Compiled from: Borowitzka 1992; Chen 1996; Doucha and Lívanský 2008 and 2011; Doucha et al. 2009; Gladue and Maxey 1994; Perez-Garcia et al. 2011; Schmidt et al. 2005; Wu and Shi 2008

Prerequisites, benefits and constrainpan>ts of heterotrophic cultivation in conventional stirred bioreactors

Need for axenic cultures

An additional crupan> class="Chemical">cial prerequisite is the requirement for a monoculture in a long-term bioreactor operation. To date, this is still hampered by the dearth of axenic (pure) cultures of species isolated from the environment. In heterotrophic cultures, the advantage of preventing the growth of contaminants through selective photoautotrophic conditions is not a possibility. Thus, any (minor) contamination introduced with the inoculum could easily outgrow the desired microalgal species.

The particular obstacle can largely be overcome by modern methods of flow cytometry. For example, fluorescence-activated cell sorting permits the efficient differentiation and subsequent isolation of single cells based upon their morphology (size) and variation in autofluorescence (Cellamare et al. 2010; Sensen et al. 1993; Surek and Melkonian 2004). The rate of success for the isolation of axenic microalgal cultures collected from natural habitats was reported to be considerably higher than with a classical approach (Surek and Melkonian 2004). Furthermore, for selectivity reasons and the resulting ease of handling, microalgae such as Galdieria sulpuraria, which perform well at 42 °C and at a pH of 2, are desirable.

Restrictions due to high salinity

Should growth in the presence of pan> class="Chemical">sea salt (and thus at a very high salinity of about 35 g kg−1 of Na/KCl and high osmolarity) be essential for good performance of a particular species, additional investments in vessels coated with special materials like polyether ether ketone are necessary. However, high salinity has not always been linked to increased corrosion of metallic materials (Schmidt et al. 2005). The requirement of marine species for high salt concentrations is sometimes unwittingly overestimated and the salt concentration may be significantly reduced without productivity loss (Kiy et al. 2008). On the other hand, high salt conditions have been found to significantly enhance lipid formation. Upon changing the sodium chloride concentration from 10 to 20 g l−1 in a culture of N. laevis, the synthesis of total lipids, the production of eicosapentaenoic acid (EPA) and the accumulation of polar lipids increased while the synthesis of neutral lipids decreased (Chen et al. 2008).

Species used in heterotrophic processes

The ability of a number of microalgal spepan> class="Chemical">cies to grow with organic carbon substrates has been demonstrated previously (Droop 1974; Gladue and Maxey 1994). However, the number of current commercially important microalgae that are capable of growth on organic carbon substrates in the dark, and where experience of fedbatch cultivation has been gained, is very limited.

Growth characteristics

Both the (growth) kinetic anpan>d stoipan> class="Chemical">chiometric characteristics of these microalgae, along with values for other microalgal species obtained from batch cultures, are summarised in Table 2. This table encompasses fast-growing species with a specific growth rate higher than 0.09 h−1 (e.g. Chlorella, Crypthecodinium, Nitzia, Prototheca spp.) and species that grow at about half the rate, but where a lot of cultivation experience is available (e.g. Galdieria, Haematococcus, Nannochloropsis or Schizochytrium spp.). These specific growth rates correspond to doubling times of between 7 and 15 h. Interestingly, the Chlorella genus exhibits a wide range of growth rates with glucose, which vary with species and growth conditions, such as temperature, pH or dissolved oxygen concentrations (Shi et al. 2006). The latter condition, in particular, can be controlled in high-cell-density (heterotrophic) cultures as the specific growth rate can be deliberately reduced to assure sufficient oxygen supply (Doucha and Lívanský 2011). Furthermore, heterotrophic growth of Dunaliella sp. and Nannochloropsis sp. is possible but is not practicable due to its very slow growth (Gladue and Maxey 1994).

Table 2

Growth characteristics of microalgae in heterotrophic batch cultures

Species		μ max	Y x/s	T	pH	Carbon sourcesa	sinhib	Productsb	References (for growth parameters only)
		h−1	gCDW g−1	°C	−log [H+]		g l−1
Chlamydomonas reinhardtii	P/H	0.035	0.52	35	6.9	Acetate	>0.4	Biomass	Boyle and Morgan (2009)
									Chen and Johns (1996)
Chlorella protothecoides	P/H	0.090	0.47	28	6.6	Glucose (acetate, glycerol)	>24	Lutein, lipids, biodiesel violaxanthin	O'Grady and Morgan (2011)
									Xiong et al. (2008)
									Shi et al. (2000)
Chlorella pyrenoidosa	P/H	0.201	0.50	35	6.9	Glucose (acetate, glutamate, lactate)	>10	Ascorbic acid, lutein	Running et al. (1994), Wu et al. (2007)
Chlorella regularis	P/H	0.240	0.6	36	6.5	Glucose (acetate, ethanol)	>10	Biomass, intracellular phytochemicals	Endo et al. (1974), Sansawa and Endo (2004)
Chlorella vulgaris	P/H	0.180	0.55–0.69	36	6.0 – 7.5	Glucose (acetate, glutamate, lactate)	n.a.	Biomass	Doucha and Lívanský (2011)
Chlorella zofingiensis	P/H	0.031	0.44	30	5.5	Glucose (fructose, galactose, mannose, lactose, sucrose)	>20	Astaxanthin	Ip and Chen (2005b)
Crypthecodinium cohnii	H	0.089	0.56	25	7.2	Glucose (acetate)	>20	DHA	Jiang and Chen (2000)
Dunaliella sp.	P/H	< 0.010	n.a.	26	7.5–8.3	Acetate, lactate, glucose, glutamate, glycerol	n.a.	Biomass beta-carotene	Gladue and Maxey (1994)
Euglena gracilis	P/H	0.045	0.43	25	2.8–3.5	Glucose, (acetate, alanine, aspartate, asparagine, ethanol, glutamate)	n.a.	Alpha-tocopherol	Ogbonna et al. (1998), Fujita et al. (2008)
Galdieria sulphuraria	P/H	0.045–0.048	0.48–0.50	42	2	Glucose	>200	Phycocyanin	Schmidt et al. (2005)
						Sugar beet molasses (fructose, sucrose)	>350
Haematococcus pulvialis	P/H	0.009	n.a.	25	8	Acetate (glucose, asparagine)	>1.6	Astaxanthin, cantaxanthin, lutein	Hata et al. (2001)
Nannochloropsis oculata	P/H	< 0.007	n.a.	26	7.5–8.3	Glucose (ethanol)	n.a.	Biomass, EPA	Gladue and Maxey (1994)
Nitzschia alba	H	0.106	n.a.	30	n. a.	Lactate, succinate	n.a.	Biomass, EPA	Lewin and Lewin (1967)
						Glucose, glutamate
Nitzschia laevis	H	0.017	0.44	20	8.2	Acetate, glucose	n.a.	EPA	Wen et al. (2002),
Prototheca zopfii	H	0.330	0.81	21	7.2	Glucose (acetate)	n.a.	l-Ascorbic acid	Running et al. (2002)
Scenedesmus actus	P/H	0.040	n.a.	30	6	Glucose	>1	Biomass	Ogawa and Aiba (1981)
Schizochytrium sp.	H	0.071	0.42	27	7	Glucose	>200	PUFA, DHA, GLA	Ganuza et al. (2008)
Tetraselmis sueica	P/H	0.028	0.41	25	7.5	Acetate (glucose, glutamine, lactate)	n.a.	Lipids, PUFA n-3 HUFA	Day and Tsavalos (1996), Gladue and Maxey (1994)

μ max maximum specific growth rate, Y biomass yield determined in batch culture at given temperature (T), s inhib substrate concentration resulting in a decrease of the specific growth rate and/or biomass yield with the particular substrate, H heterotroph, P (facultatively) photoautotroph, n.a. not available, DHA docosahexaenoic acid, EPA eicosapentaenoic acid, GLA gamma-linolenic acid, HUFA highly unsaturated fatty, PUFA polyunsaturated fatty acids

aSubstrates known to support heterotrophic growth (adapted from Lee 2001)

bMainly adapted from Guedes et al. (2011), Raja et al. (2008) and Spolaore et al. (2006)

μ max maximum specifipan> class="Chemical">c growth rate, Y biomass yield determined in batch culture at given temperature (T), s inhib substrate concentration resulting in a decrease of the specific growth rate and/or biomass yield with the particular substrate, H heterotroph, P (facultatively) photoautotroph, n.a. not available, DHA docosahexaenoic acid, EPA eicosapentaenoic acid, GLA gamma-linolenic acid, HUFA highly unsaturated fatty, PUFA polyunsaturated fatty acids

aSubstrates known to support heterotrophin class="Chemical">c growth (adapted from Lee 2001)

The specifipan> class="Chemical">c microalgae’s tolerance to certain extracellular substrate concentrations of several grams per litre, as outlined in Table 2, is a feature that also allows particular species to be readily grown to high cell densities in batch culture. Galdieria sulphuraria reached the highest specific growth rates at glucose concentrations of between 2 and 166 g l−1, while a glucose concentration of 200 g l−1 was regarded as inhibiting its growth (Schmidt et al. 2005). Other highly tolerant species are Schizochytrium and Thraustochytrium, both known to accumulate large quantities of lipids within their biomass (Jain et al. 2007; Kiy et al. 2008). Although Chlorella sp. principally grow at glucose concentrations of more than 60 g l−1 (Ip and Chen 2005b), residual concentrations as low as 10 g l−1 significantly inhibit their growth (Sansawa and Endo 2004; Wu and Shi 2007; Xiong et al. 2008).

Instigating the photosynthetic apparatus in the dark

Most of the ‘classipan> class="Chemical">c’ microalgal species are recognised primarily as photoautotrophs. Studies on the effect of repeated (and long-term) propagation in the absence of light and using an organic carbon/energy source for the production of substances involved in the photosynthetic apparatus of microalgae are not yet conclusive (Graverholt and Eriksen 2007; Sansawa and Endo 2004; Shen et al. 2010; Xiong et al. 2010a). Some of these substances continue to be synthesised in the dark, for example, the light-harvesting pigment phycocyanin from G. sulphuraria (Eriksen 2008a).

Nevertheless, the inpan>ability of obligate photoautotrophs to grow anpan>d divide without photosynpan>thetipan> class="Chemical">cally derived energy is often the main obstacle preventing the efficient heterotrophic production of microalgal metabolites. This can be overcome through appropriate genetic engineering, for example, the introduction of a gene encoding a glucose transporter as demonstrated by Zaslavskaia et al. (2001). Another technically driven means involves exploiting mixotrophic cultures, where microalgal cultures are simultaneously exposed to an organic carbon source and light (Liang et al. 2009; Liu et al. 2009; Chen et al. 2006; Sloth et al. 2006; Feng et al. 2005; Garcia et al. 2005; Ma and Chen 2001). These do, however, require specially constructed illuminated bioreactors. It has also been suggested that light-dependent production can be deliberately induced by ‘oxidative stress’, triggered by substances other than light (e.g. H2O2 or Fe2+; Ip and Chen 2005a; Kobayashi et al. 1993).

Composition of culture media and microalgal biomass

Most of the culture media for growinpan>g mipan> class="Chemical">croalgae in vitro have been developed using, as a base, the stoichiometric composition of the microbial biomass grown under regular physiological conditions (Egli 2000; Egli and Fiechter 1981). Information on the composition of microbial biomass in relation to the formation of a particular product is very limited and may vary depending on species and culture conditions. For natural phytoplankton (representing a heterogeneous consortium of microalgae), the proportions of the elements are typically derived from the ‘Redfield ratio’ dating back to 1934, suggesting a molar ratio of C106N16P1 as described in Falkowski (2000). This has recently been further extended to include other important elements (Ho et al. 2003; Quigg et al. 2003). The stoichiometric composition of phytoplankton is comprehensively reviewed by Klausmeier et al. (2008). For heterotrophic cultures of Chlorella vulgaris, a molar stoichiometry of C3.96H7.9O1.875N0.685P0.0539K0.036Mg0.012 was determined (Sansawa and Endo 2004), and this has been reflected in optimised media compositions for biomass production in high-cell-density fedbatch processes (e.g. Doucha and Lívanský 2011; Xiong et al. 2008). All of the major molecules in microalgae (i.e. proteins, carbohydrates, lipids) contain carbon as the principal element, with oxygen, hydrogen and nitrogen at lower, or even zero, concentrations (Fig. 1). Typically, in media for heterotrophic cultures that support optimal growth, all of the constituents are supplied in stoichiometric excess to the organic carbon source. Applying stoichiometric principles to an established medium for photoautotrophic cultures of Chlorella spp. (Vonshak 1986), the medium was shown to be deficient in iron, magnesium, sulphur and nitrogen. When optimised, a fivefold increase in biomass concentration was achieved (Mandalam and Palsson 1998).

Fig. 1

Compositional variation in Chlorella sp. biomass. Left, molecular composition, lower and upper limits determined at low and high nitrogen availability in heterotrophic cultures of C. protothecoides (adapted from Xiong et al. 2010b). Right, proportions of macro and micro elements (adapted from Oh-Hama and Miyachi 1988)

Compositional variation inpan> Chlorella sp. biomass. Left, molecular composition, lower and upper limits determined at low and high nitrogen availability in heterotrophic cultures of C. protothecoides (adapted from Xiong et al. 2010b). Right, proportions of macro and micro elements (adapted from Oh-Hama and Miyachi 1988)

For most of the microalgal spepan> class="Chemical">cies capable of heterotrophic growth, glucose or acetate is an adequate source of energy and carbon (Table 2; Lee 2004; Lee 2001; Perez-Garcia et al. 2011). In addition, low-cost media formulations with molasses or carob pulp syrup, or the waste streams from sugar or milk processing industries, have been successfully used as alternatives to glucose (Mendes et al. 2007; Schmidt et al. 2005). Although microalgae grow with various carbonaceous compounds, glucose is the preferred carbon source because of its ease of handling, accessibility and safety (Lee 2004; Perez-Garcia et al. 2011; Sun et al. 2008). In particular, glucose is used for the production of high-value compounds where the processes need to be reproducible for prospective regulatory approval for pharmaceutical manufacture. Acetate and ethanol are possible alternatives but, because of their respective corrosive effects or high flammability, are only used when an exceptional productivity enhancement is achieved (de Swaaf et al. 2003b; de Swaaf et al. 2003c; Ogbonna et al. 1998). Although the cost of (pure) glucose for microalgal production of high-value compounds is less critical than in the biofuel field, opportunities for valorisation of the biomass after isolation of the target compound are desirable. Examples of such valorisation include the subsequent production of animal and fish feed or its use as an energy-rich biomass for the production of biofuels (Brennan and Owende 2010; Chisti 2007).

Nitrate, pan> class="Chemical">ammonia and/or urea are the preferred nitrogen sources at a bioreactor scale (Grobbelaar 2004). Tryptone, glycine and yeast extract have also been evaluated for their potential to enhance growth or product formation (Shen et al. 2010). Moreover, growth data suggest that nitrogen source preference might vary between the species (Shen et al. 2010; Xiong et al. 2008). Yeast extract, a complex component with a high carbon content, is not defined at the single-element level but is frequently used as a source of nitrogen, amino acids, vitamins and trace elements (Grant and Pramer 1962). As its composition varies from lot to lot, supplementation with yeast extract makes precise stoichiometric process control impossible. Moreover, the depletion of some of the components of yeast extract has been shown to result in a linear growth pattern. For example, in a heterotrophic culture with C. cohnii, a non-specific nutritional limitation was detected, which coincided with a change from exponential to linear growth while continuously feeding glucose to maintain a residual concentration of between 5 and 20 g l−1 (de Swaaf et al. 2003c). It was concluded that nitrogen, solely derived from the yeast extract, was most probably depleted, resulting in the onset of the linear phase, a cessation of cell division and the diversion of added glucose into the accumulation of lipids within the cell.

Furthermore, several micropan> class="Species">algae that are grown in pure culture with mineral medium require supplementation with the vitamins cobalamin and/or thiamine (Carvalho et al. 2006; Croft et al. 2005; Droop 2007). In turn, species capable of endogenously synthesising cobalamin need traces of cobalt (Grobbelaar 2004).

High-productivity processes

The development of appropriate strategies for enhancinpan>g biomass anpan>d/or produpan> class="Chemical">ct formation is based on understanding (Carvalho et al. 2006; Grobbelaar 2004; Hu 2004) and exploitation of the flexibility (adaptability) of biomass composition within its upper and lower limits as defined by different culture conditions and/or the altered supply of chemical elements in the culture medium. Achieving the desired (optimum) process performance, however, is far from straightforward, as some objectives are by their very nature contradictory (e.g. the highest product titre vs. the shortest process duration, the highest product purity vs. the highest biomass or product formation rates and yields).

High cell densities achieved

In the scientifipan> class="Chemical">c literature concerning the mass cultivation of microalgae, the term ‘high-cell-density culture’ (HCD) is ill defined but is generally applied to those values of biomass concentration that fall within the range of the highest values published for photoautotrophic or heterotrophic processes. For a photoautotrophic system, the highest biomass concentration achieved to date is 40 g l−1 of cell dry weight (CDW) with thin-layer cultures (Doucha and Lívanský 2006). However, in heterotrophic fedbatch cultures of Chlorella sp. concentrations ranging from approximately 100 g l−1 to greater than 150 g l−1 CDW are currently achievable (Table 3; e.g. de Swaaf et al. 2003b; Doucha and Lívanský 2011; Graverholt and Eriksen 2007; Schmidt et al. 2005; Hauser, unpublished data). Typically, such high cell densities are attainable in fedbatch operation where the increasing biomass is retained and thus accumulates in the bioreactor. A highly concentrated solution of the growth substrate is added in a controlled mode, which in turn determines the specific growth rate and limits the concentration of the residual substrate in the culture broth (e.g. glucose concentration in the feed solution ranging from 100 to 570 g l−1, where the highest concentration was used by de Swaaf et al. (2003c)). As the actual data relating to the time courses of the effective (working) culture volume are generally not provided in the publications (Table 3), it is not possible to calculate the amount of biomass produced (in grams) and, thus, the specific growth rates.

Table 3

Overview of fedbatch strategies for high-cell-density culture heterotrophic cultivations of microalgae

Species	x end		c p,end	Substrate	s 0,batch	s in	s fb	s res	V 0	t (t fb)	Control strategy	References
	(g l−1)		(g l−1)		(g l−1)	(g l−1)	(g l−1)	(g l−1)	(l)	(h)
Chlorella protothecoides	51.2	Lipids	25.8	Glucose	19.5	100	n.d.	2–24	5a	167 (133)	Pulsed	Xiong et al. (2008)
Chlorella protothecoides	48.0	Lipids	0.22	Glucose	40	240	n.d.	10–30	2.5	240 (147)	Pulsed	Shi et al. (2002)
Chlorella pyrenoidosa	116.2	Biomass	x end	Glucose	10	499	224	1.2–9.4	10	118.5 (95)	Stepwise	Wu and Shi (2007)
Chlorella regularis	84.0	Biomass	x end	Glucose	10	570	n.d.	0.5–10	30	30 (24)	Pulsed	Sansawa and Endo (2004)
Chlorella vulgaris	117.2	Biomass	x end	Glucose	65	500	130	5–45	35b	32 (16)	Pulsed	Doucha and Lívanský (2011)
Chlorella vulgaris	165.8	Biomass	x end	Glucose	20	500	253	0–17	6	60.5 (11)	Exponential	Hauser, unpublished data
Chlorella zofingiensis	51.8	Astaxanthin	0.03	Glucose	20	n.d.	n.d.	5–20	2	360 (264)	Pulsed	Sun et al. (2008)
Cryptecodinium cohnii	83.0	DHA	11.7	Ethanol	5.5	800	217	5–10	1	220 (191)	pO2	de Swaaf et al. (2003b)
Cryptecodinium cohnii	109.0	DHA	19.0	Acetic acid	5.75	1049	182	∼ 0	1	400 (380)	pH	de Swaaf et al. (2003c)
Cryptecodinium cohnii	26.0	DHA	1.7	Glucose	5	570	82	5–20	1c	120 (120)	Stepwise	de Swaaf et al. (2003c)
Cryptecodinium cohnii	42.0	DHA	1.9	Carob syrup	8.8d	n.d.	n.d.	5–10d	1.25	100.4	pO2	Mendes et al. (2007)
Euglena gracilis	48.0	α-Tocopherol	0.01	Glucose	15	696	n.d.	> 1	2.5a	182 (151)	Pulsed	Ogbonna et al. (1998)
Euglena gracilis	39.5	α-Tocopherol	0.04	Ethanol	15	800	n.d.	> 1	2.5a	455 (433)	Pulsed	Ogbonna et al. (1998)
Galdieria sulphuraria	109.0	Phycocyanin	2.9	Glucose	50	500	157	<0.5	1.2	336 (144)	Stepwise	Graverholt and Eriksen (2007)
Galdieria sulphuraria	72.0	Phycocyanin	0.28	Glucose	50	500	178	<0.3	1.5	250 (115)	pO2	Schmidt et al. (2005)
Galdieria sulphuraria	116.0	Phycocyanin	0.35	Molasses	52.5d	750	356d	2–5d	1.5d	338 (172)	pO2	Schmidt et al. (2005)
Nitzschia laevis	22.1	EPA	0.70	Glucose	20	400	22	<3	1.5	336 (192)	Stepwise	Wen et al. (2002)
Nitzschia laevis	40.0	EPA	1.1	Glucose	20	50	45	<6	2.2	336 (216)	Stepwise	Wen and Chen (2002)

x end concentration of biomass (determined as cell dry weight, CDW) at the process end point, c p,end concentration of particular products (docosahexaenoic acid, eicosapentaenoic acid) at the process end point, s 0,batch initial substrate concentration in batch phase, s in substrate concentration of the feed solution added, s fb accumulated amount of substrate added during the fedbatch phase per litre of the final volume, s res residual substrate concentration in bulk medium during the fedbatch phase, V initial culture volume, i.e. working volume or reactor volume, where V 0 is not available, scale-up was reported to be 200 l or 120 m3 in Apt and Behrens (1999), t total process duration with t fb as duration of the fedbatch phase, control strategy being either a predefined stepwise or exponentially increasing rate of substrate feed, pulsed addition of substrate or substrate addition resulting from a feedback control within defined limits of dissolved oxygen concentration (pO2) or pH, n.d. not determinable from the data provided in the paper, DHA docosahexaenoic acid, EPA eicosapentaenoic acid

aReactor volume

b200 l

c120 m3

dGiven in g glucose

x end conpan> class="Chemical">centration of biomass (determined as cell dry weight, CDW) at the process end point, c p,end concentration of particular products (docosahexaenoic acid, eicosapentaenoic acid) at the process end point, s 0,batch initial substrate concentration in batch phase, s in substrate concentration of the feed solution added, s fb accumulated amount of substrate added during the fedbatch phase per litre of the final volume, s res residual substrate concentration in bulk medium during the fedbatch phase, V initial culture volume, i.e. working volume or reactor volume, where V 0 is not available, scale-up was reported to be 200 l or 120 m3 in Apt and Behrens (1999), t total process duration with t fb as duration of the fedbatch phase, control strategy being either a predefined stepwise or exponentially increasing rate of substrate feed, pulsed addition of substrate or substrate addition resulting from a feedback control within defined limits of dissolved oxygen concentration (pO2) or pH, n.d. not determinable from the data provided in the paper, DHA docosahexaenoic acid, EPA eicosapentaenoic acid

aRean class="Chemical">ctor volume

n class="Chemical">c120 m3

dGiven in g n class="Chemical">glucose

Furthermore, biomass conpan> class="Chemical">centrations of more than 40 g l−1 CDW have been achieved in batch cultures with the few microalgal genera that tolerate exceptionally high substrate concentrations (Table 2; e.g. Galdieria, Schizochytrium and Thraustochytrium in Schmidt et al. 2005; Jain et al. 2007; Kiy et al. 2008, respectively). For instance, high-cell-density batch cultures with the obligate heterotrophic Prototheca sp. and Aurantiochytrium sp. were described by Running et al. (2002) and Jakobsen et al. (2008), respectively. In the latter, the cell dry weight increased from 40 to 90 g l−1 during a post-exponential growth phase while the lipid content of the biomass increased between circa 10% and 60%. In this context, the term ‘biomass growth’ needs to be differentiated as either the cells’ proliferation due to cell division or an increase in cell mass not directly linked to an increase in cell number.

Do high-cell-density culture and rapid growth imply high productivity?

Although microalgal biomass is the only target inpan> some propan> class="Chemical">cesses (Becker 2007; Brown et al. 1997; Doucha et al. 2009; Duerr et al. 1998; Tokusoglu and Unal 2003), the apparent prime focus on HCD cultures often obscures the ultimate objective for any microalgal process, which is to achieve the highest product concentration of the desired quality in the shortest possible time. From the basic data summarised in Table 3, the product to biomass yield (Y , in g g−1) and the volumetric productivities (r or r , in g l−1 h−1) can be calculated. In principle, however, interpreting productivities that have been computed on the basis of average values for the whole process duration can be misleading.

Maintaining growth at the maximum specifipan> class="Chemical">c growth rate (μ, in h−1) does not often correlate with the highest attainable rate of specific product formation (q p, in g g−1 h−1). Generally, the kinetic models accepted of microbial growth association also apply for microalgae. For instance, in continuous cultures, the maximum production of EPA by photoautotrophically grown Nannochloropsis sp. was achieved at dilution rates in the range of 0.004–0.013 h−1 while maximum biomass productivity was obtained at higher dilution rates ranging between 0.011 and 0.017 h−1 (Zou et al. 2000). Similarly, the highest DHA content as well as the highest degree of fatty acid unsaturation of Crypthecodinium cohnii was measured at low glucose concentrations (5 g l−1) and thus at a slow growth rate (Jiang and Chen 2000). The formation of EPA by Nitschia laevis in continuous heterotrophic culture was, on the contrary, reported to occur at a higher dilution rate and higher glucose concentration than maximum growth (Wen and Chen 2003). A positive growth association was however reported to apply to the biosynthesis of astaxanthin in Chlorella zofingiensis (Wang and Peng 2008). The above information illustrates that product formation kinetics can fall into either positive or non-/negative growth-associated product formation or a combination of both. Especially in those cases where product formation is not correlated with maximal growth, it is advantageous to have full control over the specific growth rate. In general, such fine-tuning of growth is achieved by the controlled addition of medium constituents in fedbatch cultivation.

Different carbon substrates lead to different biomass/substrate yields anpan>d also affepan> class="Chemical">ct the formation of the targeted product. As a consequence, maximum product formation does not always correlate with maximum biomass concentration. For instance, in a culture of C. cohnii grown with acetic acid, 77 g l−1 of biomass and 9.5 g l−1 of docosahexaenoic acid (DHA) were harvested after 210 h (de Swaaf et al. 2003c). In an alternative process strategy using ethanol as the carbon and energy source, less biomass (59 g l−1) generated more DHA (10.4 g l−1) in a shorter period of time (200 h; de Swaaf et al. 2003b). In another example, Euglena gracilis grown with glucose reached 48 g l−1 of biomass and 150 μg gCDW−1 of α-tocopherol in 182 h (Ogbonna et al. 1998). When fed with ethanol, the growth slowed significantly; however, the product yield was enhanced by a factor of approximately 10, reaching 1,200 μg gCDW−1 and 39.5 g l−1 of biomass in 455 h.

As biomass growth slows, for instance, due to the effepan> class="Chemical">ct of temperature or restricted substrate availability, cell division decelerates and the formation of storage products typically increases simultaneously. PUFA synthesis in certain species, such as eicosapentaenoic acid in Chlorella minutissima, has been reported to occur preferentially at temperatures lower than those required for optimal growth (Yongmanitchai and Ward 1991). However, in the diatom Phaeodactylum tricornutum, maximum product yield and maximum biomass concentration were achieved at the same cultivation temperature between 21.5 and 23 °C. In contrast, the lutein yield (Y ) of heterotrophically grown Chlorella protothecoides was increased from 4.25 to 4.59 mg g−1 by raising the cultivation temperature from 24 to 35 °C (Shi et al. 2006).

Strategies to enhance productivity in fedbatch cultures

Fedbatch pan> class="Chemical">cultivation is the most effective technique for reaching high biomass concentrations in a short time and controlled manner. Typically, this is achieved in cultures grown heterotrophically through controlling the rate of addition of the organic carbon and energy source (i.e. the substrate feed). Varying feed strategies can lead to different efficiencies of biomass and/or product formation.

In contrast to batpan> class="Chemical">ch mode, osmotic or toxic effects due to high substrate concentrations can be avoided in a fedbatch culture. Moreover, perfusion technology (with cell retention but exchange of culture medium) is appropriate when inhibitory metabolites, which would otherwise affect biomass growth or product formation, need to be removed (Wen and Chen 2002). Auto-inhibition effects have been described in the literature for the cultivation of different species or desired products, such as fatty acids (Bosma et al. 2008; Javanmardian and Palsson 1991; Zou et al. 2000).

The microalgal high-pan> class="Chemical">cell-density processes described in the literature (Table 3) employ several different strategies for substrate addition, which are optimised with respect to both the physiological requirements of the particular species and the technical restrictions of the available equipment. The majority of heterotrophic processes employing fedbatch mode, and using microalgae to produce high-value compounds that have been described systematically in recent publications, concentrate on some five to ten species and can be categorised by their target product as follows:

Polyunsaturated fatty acids, like DHA (Chi et al. 2009; de Swaaf et al. 2003c; Ganuza and Izquierdo 2007) and EPA (Wen and Chen 2002; Wen et al. 2002), using C. cohnii, Nitzschia laevis, Schizochytrium sp. or Ulkenia sp.

Carotenoids, like astaxanthin (Sun et al. 2008) and lutein (Shi et al. 2002), using Chlorella sp., and tocopherols with E. gracilis (Ogbonna et al. 1998)

Phycobiliproteinpan>s like phycocyanin using G. sulphuraria (Graverholt and Eriksen 2007; Schmidt et al. 2005)

The principal differenpan> class="Chemical">ces of the fedbatch processes listed in Table 2 are (1) the cells’ physiological state which is both affected by unrestricted or controlled substrate availability, (2) the limited (since controlled) availability of an element other than carbon and (3) the control strategy, which is either a predefined open-loop control or a feedback control (closed loop).

Pulsed addition of an organic carbon and energy source

The pulsed addition of a carbon sourpan> class="Chemical">ce is frequently employed in the high-cell-density cultivation of Chlorella sp. (Doucha and Lívanský 2011; Sansawa and Endo 2004; Shi et al. 2002; Sun et al. 2008; Xiong et al. 2008) and has also been applied in processes with E. gracilis (Ogbonna et al. 1998) and G. sulphuraria (Schmidt et al. 2005). Following the depletion of the substrate during an initial batch phase, the residual glucose concentration in the bioreactor is typically maintained within pre-determined upper and lower concentration limits. With C. zofingiensis and C. protothecoides, the upper limits were set at concentrations of approximately 24 or 30 g l−1, respectively. Higher levels would have inhibited biomass growth (Sansawa and Endo 2004; Shi et al. 2002; Sun et al. 2008; Xiong et al. 2008). The pulsed addition of a highly concentrated substrate solution was repeatedly triggered (about five to seven times during the entire process) whenever the glucose concentration dropped below a defined concentration of several grams per litre (Fig. 2). As described in a later section on feedback control strategies, overdosing with substrate can be prevented by implementing an automated dosing strategy based on monitoring the dissolved oxygen concentration.

Fig. 2

Principles of controlling growth and product formation through different modes of substrate addition. The three pictograms in each row show (from left to right) the time dependence of the following variables within the fedbatch phase of a bioreactor cultivation of C. vulgaris and their maximum (max), minimum (min), optimum (opt) or residual (res) values as indicated by the horizontal dashed lines: specific growth rate (μ), specific substrate utilisation rate (q ) and specific product formation rate (q ); concentrations of biomass (x) and the growth-limiting substrate (s) in the bioreactor, where the substrate concentration in the bioreactor was either determined by pulsed addition at predefined limits (s > s 1 and s < s 2) or reached a residual concentration (s res) below which the substrate cannot be utilised at the particular specific growth rate; mode of substrate addition. In the left-hand pictograms, growth-associated kinetics of the product formation was applied accordingly to the formula: q p = Y  · μ, where Y is a constant product yield per biomass coefficient. A constant specific growth rate near to its maximum (μ max) and correspondingly exponentially increasing biomass concentration is achieved with pulsed substrate addition (first row, derived from the data of Doucha and Lívanský 2011). With continuous substrate addition and an exponentially increasing feed rate (second row), a desired constant specific growth rate (μ opt < μ max) is controlled at the optimum for product formation. To reach near-optimum conditions, the exponential addition can also be approximated by a stepwise increasing feed rate (dashed step-like line)

Principles of pan> class="Chemical">controlling growth and product formation through different modes of substrate addition. The three pictograms in each row show (from left to right) the time dependence of the following variables within the fedbatch phase of a bioreactor cultivation of C. vulgaris and their maximum (max), minimum (min), optimum (opt) or residual (res) values as indicated by the horizontal dashed lines: specific growth rate (μ), specific substrate utilisation rate (q ) and specific product formation rate (q ); concentrations of biomass (x) and the growth-limiting substrate (s) in the bioreactor, where the substrate concentration in the bioreactor was either determined by pulsed addition at predefined limits (s > s 1 and s < s 2) or reached a residual concentration (s res) below which the substrate cannot be utilised at the particular specific growth rate; mode of substrate addition. In the left-hand pictograms, growth-associated kinetics of the product formation was applied accordingly to the formula: q p = Y  · μ, where Y is a constant product yield per biomass coefficient. A constant specific growth rate near to its maximum (μ max) and correspondingly exponentially increasing biomass concentration is achieved with pulsed substrate addition (first row, derived from the data of Doucha and Lívanský 2011). With continuous substrate addition and an exponentially increasing feed rate (second row), a desired constant specific growth rate (μ opt < μ max) is controlled at the optimum for product formation. To reach near-optimum conditions, the exponential addition can also be approximated by a stepwise increasing feed rate (dashed step-like line)

Pulsed fedbatch strategies are therefore appropriate to spepan> class="Chemical">cies where growth is inhibited by very high substrate concentrations but where residual substrate concentrations of several grams per litre can be tolerated. Biomass is developed at the highest specific growth rates when growth occurs in the continuous presence of an excess of substrate and in a balanced medium. The values for a specific growth rate achieved in such fedbatch processes are comparable to those rates attained during unrestricted, exponential growth in a batch culture (e.g. μ max > 0.18 h−1 for C. vulgaris; Doucha and Lívanský 2011). The pulsed fedbatch strategy is applicable to systems required to produce compounds where formation is tightly associated with fast biomass growth. In addition, the combined production strategies can be applied to systems where rapidly built biomass is essential during the first process phase, prior to a subsequent differently controlled phase that promotes product formation (e.g. Ganuza et al. 2008; Hata et al. 2001).

Continuous addition of an organic carbon and energy source

Potential strategies for continpan>uous substrate addition are usually pan> class="Chemical">categorised using mathematical functions that describe the time dependence of the rate of substrate addition (i.e. the feed profile). Such strategies encompass substrate addition at constant or predefined rates. Predefined rates can increase or decrease following linear or exponential functions or by equivalent, stepwise approximations (Fig. 2). However, the only feed profiles that are currently used to grow microalgae are those with stepwise increasing or decreasing feed rates (for example, Chlorella, Crypthecodinium, Galdieria and Nitzschia by Wu and Shi (2007), de Swaaf et al. (2003b), Graverholt and Eriksen (2007) and Wen and Chen (2002), respectively).

The optimum feed profile for product formation pan> class="Chemical">can be determined by using a rational approach provided that the effect of appropriate control variables (including μ) on the targeted productivity has been quantified and, preferably, captured in a model (Wu and Shi 2007; Zhang et al. 1999a; Zhang et al. 1999b; Zhang et al. 1999c). In addition, models based on hybrid neural networks can be used to predict optimum fedbatch strategies. A feed profile that included an exponentially increasing glucose addition over several steps, each at a constant feed rate, was successfully applied to a Chlorella pyrenoidosa culture to attain a biomass concentration of 116 g l−1 at an average productivity of 1.02 g l−1 h−1 (Wu and Shi 2007). Similarly, with a culture of G. sulphuraria, the feed rate was increased up to 3.7 times in one step per day, achieving 27.8 g l−1 of biomass and an average productivity of approximately 0.33 g l−1 h−1 (Graverholt and Eriksen 2007). Following an initial batch culture, the substrate was continuously added at a specific rate, which was lower than the specific glucose utilisation rate required to support the biomass, i.e. a μ max of 0.053 h−1. Residual glucose concentrations were maintained below 0.5 g l−1 and the specific growth rate was controlled at 0.046 h−1 during the fedbatch phase.

Specifipan> class="Chemical">c growth rate can be controlled at a particular defined rate, lower than its maximum, by continuous (preferably exponential) substrate addition. This allows product formation to be enhanced when its optimum is not linked to the fastest biomass growth (as illustrated in the pictograms in Fig. 2). However, the strategy of controlling the μ at a certain (optimum) value as demonstrated with bacteria and yeast has not yet been described in the literature on microalgae.

Substrate addition determined by feedback control

A feedback pan> class="Chemical">control strategy allows a continuous adaptation of the rate of substrate addition based on physiological criteria, such as the production of acids or bases (de Swaaf et al. 2003c), or oxygen utilisation (de Swaaf et al. 2003b; Schmidt et al. 2005). The changes in pH and pO2 values resulting from biomass growth (or product formation) are countered up to a predefined value.

In an auxostat-fedbatch system with Span> class="Chemical">chizochyrium sp., the pH was continuously controlled at a value of 7 through the addition of ammonium hydroxide solution (Ganuza et al. 2008). In this manner, both pH and an excess of nitrogen were controlled while the growth of the biomass reflected the utilisation of the carbon substrate. The substantial technical benefit of such a system is the potential to replace the ammonium hydroxide solution used for pH control with a potassium hydroxide solution. In this way, the desired onset of the production of lipids (DHA) is readily triggered through nitrogen deprivation (i.e. the nitrogen level decreases to zero from the concentration accumulated during the ammonium hydroxide addition phase). Effectively, this process strategy is fedbatch with respect to the nitrogen addition but is batch with respect to the (carbon) growth substrate. In a fedbatch process producing DHA with C. cohnii (de Swaaf et al. 2003b), the acetic acid used as the carbon/energy substrate was added via a pH control system to maintain a pH of 6.5. This resulted in the continuous addition of acetic acid at a rate which followed a time course determined by the metabolic requirements of the biomass.

Feedback pan> class="Chemical">control based on the dissolved oxygen concentration (pO2 or DO) is another process strategy that is technically feasible to prevent overdosing of the substrate. As determination of the residual concentration of the substrate typically involves a time delay, monitoring the decrease/increase in pO2 has the advantage such that the pulsed addition of substrate can be appropriately timed (Schmidt et al. 2005).

Strategies for customising biomass composition by adapting the culture medium

The following strategies, potentially combinpan>ed with fedbatch culture, allow product formation or biomass composition to be further controlled:

Controllinpan>g the availability of components in the growth medium other than carbon

Replacinpan>g a medium component with an alternative (e.g. using a different carbon or nitrogen source or exchanging sulphur with selenium)

Adapting the culture conditions (for example, T, pO2, pH) to conditions that would typically be outside of the optimal range for biomass growth

Affecting the proportions of macromolecules within the biomass

Biomass composition pan> class="Chemical">can be customised and/or product formation can be enhanced through tailoring the composition of the culture medium. The latter is achieved by controlling the consumption of a particular element through ensuring that its supply and, thus, availability are limited. An example of this principle is seen in the range of biomass composition achievable with C. protothecoides (Fig. 1).

Lipid pan> class="Chemical">content has been reported to increase under nutrient-deprived conditions such as low concentrations of nitrogen (Griffiths and Harrison 2009; Hsieh and Wu 2009; Illman et al. 2000; Jakobsen et al. 2008; Lv et al. 2010; Lynn et al. 2000; Rodolfi et al. 2009; Widjaja et al. 2009; Yongmanitchai and Ward 1991), phosphorus (Lynn et al. 2000; Reitan et al. 1994; Rodolfi et al. 2009) and silicon (Griffiths and Harrison 2009; Lynn et al. 2000). Low nitrogen concentrations were also reported to increase the cellular contents of lutein in C. protothecoides (Shi et al. 2002) and astaxanthin in C. zofingiensis (Ip and Chen 2005b).

N-deprivation regimes are most frequently expressed as the molar pan> class="Chemical">carbon-to-nitrogen (C-to-N) ratio (Ip and Chen 2005b; Shi et al. 2002; Sloth et al. 2006; Wen et al. 2002). Grown at a high C-to-N ratio, the cell dry mass of C. protothecoides contained up to 53.8% lipids compared to 25.2% obtained in low C-to-N medium. This lipid increase was accompanied by a drop in protein content from 25.8% at low C-to-N to 10.5% at a high C-to-N ratio (Xiong et al. 2010b).

Biomass enrichment by replacing its constituents

The enrichment of the mipan> class="Chemical">croalgal biomass with (readily bio-available) selenium at extraordinarily high volumetric productivities and final concentrations (> 0.4 mg Se per gram of biomass) is governed by a different mechanism than the enrichment by lipids. Microalgae, which are grown in medium containing a lower-than-stoichiometrically-required concentration of sulphur, are exposed to inorganic selenium in the form of selenite. Depending on the algae species, the cells are capable of incorporating the element at different levels, preferentially into intracellular protein, e.g. C. vulgaris or Scenedesmus quadricauda in Doucha et al. (2009) and Umysová et al. (2009), respectively.

Conclusions and future trends

Recent advanpan>pan> class="Chemical">ces in microalgal biotechnology have created opportunities for the efficient production of high-value (natural) compounds with the properties of plant-derived products that provide unique benefits (e.g. plant-like glycosylation) compared to their analogues resulting from chemical synthesis or recombinant microorganisms. The laboratory-scale bioreactor cultivations included within this review provide a first insight into the feasibility of carrying out heterotrophic processes with microalgae at an industrial scale. These processes have, in part, already been commercialised with the biotechnological production of PUFA. (To date, the authors are not aware of any literature on the heterotrophic large-scale fedbatch cultivation of microalgae, in contrast to literature on batch cultivation referred to by Apt and Behrens 1999; Behrens 2005; Wynn et al. 2005). This review confirms that the development of efficient, economically feasible large-scale bioprocesses remains an obstacle to the commercialisation of the promising microalgae technology.

The generic pan> class="Chemical">cultivation strategies outlined are based on the experimental data of natural microalgae but, in principle, could also be applied to emerging strains improved by genetic engineering. Fundamental knowledge enabling strain design may be derived from advanced metabolic flux analyses (Xiong et al. 2010b). A promising new avenue for transgenic microalgae is developing based on the knowledge gained over the past two decades, which includes the complete sequencing of the first microalgal genomes (Leon-Banares et al. 2004; Parker et al. 2008; Rosenberg et al. 2008; Walker et al. 2005a; Walker et al. 2005b).