The place of diatoms in the biofuels industry

Abstract

In spite of attractive attributes, diatoms are underrepresented in research and literature related to the development of microalgal biofuels. Diatoms are highly diverse and have substantial evolutionarily-based differences in cellular organization and metabolic processes relative to chlorophytes. Diatoms have tremendous ecological success, with typically higher productivity than other algal classes, which may relate to cellular factors discussed in this review. Diatoms can accumulate lipid equivalently or to a greater extent than other algal classes, and can rapidly induce triacylglycerol under Si limitation, avoiding the detrimental effects on photosynthesis, gene expression and protein content associated with N limitation. Diatoms have been grown on production scales for aquaculture for decades, produce value-added products and are amenable to omic and genetic manipulation approaches. In this article, we highlight beneficial attributes and address potential concerns of diatoms as biofuels research and production organisms, and encourage a greater emphasis on their development in the biofuels arena.

Figure 1.  Scanning electron micrographs of different diatom species.

(A) Thalassiosira pseudonana, (B)Cocconeis sp., (C)Lampriscus sp., (D)Gyrosigma balticum, (E)Cyclotella cryptica, (F)Nitzschia sp., (G)Thalassiosira weissflogii and (H)Achnanthes sp.

Figure 2.  Fluorescence and light micrographs of lipid droplets in diatoms.

(A) Nitzschia curvilineata, (B)Cyclotella cryptica, (C)C. cryptica, (D)Nitzschia alba and (E)Nitzschia filiformis. Green fluorescence is BODIPY staining of neutral lipids, red is chlorophyll autofluorescence. (B) is a differential interference contrast image with no fluorescence.

Figure 3.  Unrooted consensus phylogeny of photosynthetic eukaryotes and derived taxa.

Branch lengths do not correspond to phylogenetic distances. Red arrow locates diatoms and green arrow locates chlorophytes.

Reproduced with permission from [176].

Figure 4.  Neutral lipid accumulation using BODIPY fluorescence in Thalassiosira pseudonana comparing treatment with nocodazole, or limitation for Si or N.

Cultures were harvested from exponential phase and inoculated into fresh medium for the various treatments. The nocodazole treatment was done in a separate experiment from Si and N.

Figure 5.  Control of biomass accumulation in cultures of Cyclotella cryptica by altering the starting concentration of silicon.

Incremental amounts of silicic acid were added to the cultures at day 1 and growth followed thereafter. On day 7 when all cultures were limited, the ratio of cell numbers to starting concentration were in exact correspondence.

A frequently stated reason for choosing unicellular microalgae as biofuel-production organisms is that they are more efficient than vascular plants at turning light and nutrients into valuable products. The economics of algal biofuels is critically dependent on productivity, and one can apply the same criterion and ask the question: among the microalgae, are certain classes more productive? Before answering that question, we need to acknowledge that within a given class of algae, there will be differences in measured productivity values when comparing individual species. These differences can stem from intrinsic factors related to their cell biology or from growth conditions at the time of sampling, which can involve both environmental and ecological variables. Comparing productivity in controlled culture conditions has value, because it reflects what is possible, but it also encompasses biases due to the fact that conditions are never exactly the same in two culture systems. Indications are that algal biofuels will be developed on a regional basis in a wide variety of environments [1]. Another consideration is the utility of comparing maximum productivity values. In a production system operating for a good portion of the year, it would be far more desirable to have strains that are consistently productive under a variety of environmental conditions than strains that have higher productivity under optimal conditions for a shorter period of time. A final consideration in productivity comparisons is that if the number of studies is relatively small, which is usually the case, poor statistical comparisons result. An alternative to measuring productivity values in culture is to evaluate an algal class’ productivity on the basis of its abundance in nature in a variety of environments. This approach may directly relate to biofuel production systems that would be deployed in different environments and provide robust statistics, since the sample size is large.

Diatoms (Figure 1) are among the most productive and environmentally flexible eukaryotic microalgae on the planet. They are estimated to be responsible for 20% of global carbon fixation and have become the dominant primary producers in the ocean [2,3]. In addition to being highly abundant, they are also highly diverse, with estimates of over 100,000 species [4]. Their most distinctive cellular feature is their cell wall made of nanostructured silica (Figure 1), which is reproduced with fidelity through generations by genetically controlled assembly processes. Typical diatoms range in size from a few to a few hundred microns, with the majority of species between 10 and 50 µm.

Diatoms have characteristics, both conceptual and proven, which make them amenable to large-scale biofuel cultivation. In 1998, the Aquatic Species Program recommended a list of 50 microalgal strains, selected from a pool of over 3000 candidates, which held the most promise as biofuel-production organisms [201]. Sixty percent of the selected strains were diatoms, which were chosen based on criteria such as high growth rates and lipid yields, tolerance of harsh environmental conditions, and performance in large-scale cultures [201]. Diatoms are excellent lipid accumulators, and we frequently see cells similar to the ones shown in Figure 2 where a substantial portion of the cells’ volume is occupied by lipid droplets that accumulate rapidly, especially under Si limitation.

There is a puzzling underrepresentation of diatoms in the microalgal biofuels arena. Based on our survey of the literature and research experience, this is not because they are inferior to other classes of algae. The purpose of this article is to highlight attributes of diatoms that make them highly desirable as organisms for biofuel research and production, and to address potential concerns and misconceptions about diatoms that we have encountered in the literature and during dialog with colleagues over the past several years.

Evolutionary aspects of diatoms

Diatoms are members of the Stramenopile or heterokont class of algae (Figure 3), which are highly different from, and have a more complex evolutionary history than, green algae and vascular plants [5]. Common to all photosynthetic eukaryotes was a primary endosymbiosis event where a heterotrophic eukaryote engulfed or was invaded by a cyanobacterium. Most of the cyanobacterial genes were transferred to the nucleus as the cyanobacterium evolved into the chloroplast. The progenitor plant cell resulting from primary endosymbiosis gave rise to the glaucophytes, red algae, green algae and plants. A secondary endosymbiosis later occurred between a different heterotrophic eukaryote and a red alga. The red algal endosymbiont was the progenitor of the plastids found in Stramenopiles, a group that includes diatoms, brown macroalgae and plant parasites (Figure 3). In addition to diatoms, prospective biofuel production organisms such as Nannochloropsis are in the Stramenopiles.

The evolutionary story is a bit more complex for diatoms; evidence suggests that at some stage in the development of the progenitor plant cell, a Chlamydial invasion occurred [6]. There is also evidence that green algal nuclear genes are present in the secondary endosymbiont [7], suggesting an endosymbiotic episode with a green alga. Diatom genomes also contain abundant numbers (up to 5%) of genes derived from bacteria of various classes, with more than half of these genes being shared between two evolutionarily diverse diatoms, Thalassiosira pseudonana and Phaeodactylum tricornutum[8]. Thus diatom genomes, and their resultant metabolomes, are a complex mixture of components derived from highly divergent sources [8,9]. This is reflected in a unique combination of metabolic processes in diatoms, which combine both plant- and animal-like characteristics [10]. Fundamental aspects of diatom metabolism differ substantially from other classes of algae [11], and caution should be taken when making generalizations about algae based on studies on one organism.

The four major groups of diatoms (based on the features of their silica cell walls) are the radial centrics, the bipolar and multipolar centrics, the araphid pennates, and the raphid pennates; each group arose and diversified sequentially under decreasing CO₂ levels during the Mesozoic era [5]. Thus, their carbon concentrating and fixation systems evolved under substantially different conditions; for example, CO₂ levels were, at one point, eight-times higher than present [12]. It is reasonable to expect that changing environmental CO₂ levels were drivers (at least in part) of diatom diversification and that differences in the physiology of the major groups may reflect this adaptation. Diatoms also adapted to grow and dominate in low-nutrient open ocean waters resulting from changing global geochemical conditions [5]. Conversely, diatoms may have affected environmental conditions themselves; a rapid rise in diatom abundance during the Eocene may have resulted in a massive drawdown of atmospheric CO₂ that facilitated global cooling [13].

The fossil record indicates that over the Mesozoic era larger eukaryotic phytoplankton of the red algal lineage, which includes dinoflagellates, coccolithophorids and diatoms, displaced a large portion of other algae in the ocean, which were predominantly cyanobacteria and very small green algae [14]. This suggests fundamental productivity advantages to the red algal line [14], which will be discussed later. Enhanced sinking rates of the larger phytoplankton increased the burial of organic carbon in the continental margins and shallow seas, which generated most of our known petroleum reserves [15,16], and reduced atmospheric CO₂ levels. While some early reports suggested that diatoms were the major source of carbon for fossil fuels, dinoflagellates and coccolithophorids were the dominant types of phytoplankton during the major carbon export period [14]. Currently, diatoms are responsible for a large fraction of the organic carbon buried on continental margins [17] and are major contributors to nascent petroleum reserves.

Diatom productivity

Currently, in the oceans, diatoms are one of the most productive classes of microalgae. They are responsible for a substantial portion (40%) of marine primary productivity [3] and most of the CO₂ export (drawdown and sinking) from the atmosphere, of the order of 50% [18]. The intrinsically high productivity of diatoms not only makes them major players in the global carbon cycle, but also suggests their utility as biofuel production organisms.

▪ Natural productivity

Diatoms are efficient at nutrient utilization and are entirely responsible for the initial nitrate uptake in the equatorial upwelling zone of the eastern Pacific Ocean [18]. Other eukaryotic phytoplankton obtain their N through the action of grazers, which allows N from diatoms to be recycled [18]. One of the contributing factors to diatom dominance is that they possess a large nutrient storage vacuole, which dinoflagellates and coccolithophorids lack [19–21]. Vacuolization has important consequences in terms of productivity. Nutrient utilization efficiency is generally related to surface to volume ratio, and smaller cells tend to be more efficient for this reason. One way to compensate for a larger surface area is to store nutrients inside the cell in a vacuole, which in essence reduces the cytoplasmic volume and increases the surface area from which nutrients can be utilized by the cytoplasm [19]. As long as extracellular nutrients are present at levels that minimize diffusive limitation to the cell surface, the gains in productivity in vacuolated cells are quite substantial [19]. Thus, in nutrient-replete conditions, diatoms will hoard excess nutrients, storing enough nitrate to enable several cell divisions, and preventing the growth of other phytoplankton [15]. In a survey of in situ growth rates of marine phytoplankton, maximum doubling rates between 2 and 4 per day were repeatedly measured for both pennate and centric diatoms, which were higher than those for other algae with corresponding sizes [22]. Diatoms can even dominate under nutrient-limiting conditions; in one example a diatom species was shown to outcompete non-N-fixing cyanobacteria under low nitrate concentrations in a eutrophic lake [23]. Diatoms consistently outcompete other eukaryotic phytoplankton under conditions of mixing and high turbulence, in which pulses of nutrients occur between short periods of quiescence [24,25]. Furthermore, the relative dominance of diatoms in the oceans relates to geologic time periods of generally greater ocean turbulence [26–29].

Diatoms also have a greater carbon fixing ability than other co-existing groups of microalgae, as shown from calculations of productivity per unit of crop carbon [30]. This has not only been repeatedly demonstrated in laboratory cultures under optimal conditions [31–33], but in environmental samples from different locations [30]. In a comparative study, diatoms grew faster under low light conditions than competing algae [34]. Higher photosynthetic efficiency for diatoms not only relates to other Stramenopiles, but also in comparison to green algae (molecular mechanisms will be discussed below). A comparison between P. tricornutum and Chlorella vulgaris indicated that under nonfluctuating light conditions, efficiencies of converting light into biomass were equivalent, but under fluctuating light conditions, the diatom was nearly twice as efficient as C. vulgaris[35]. Efficient carbon fixation ability along with nutrient utilization could increase the appeal of diatoms for co-processes such as CO₂ abatement and waste water remediation.

▪ The relevance of natural productivity to biofuels production

Several of the concepts presented in the previous paragraphs are of relevance to biofuels production. Diatoms naturally bloom under nutrient-replete conditions, meaning maximal growth rate to highest density is obtained, which are ideal characteristics for a biofuel production system. Diatom blooms also tend to crash precipitously, which is a tendency of all bloom species. Crashes can result from a number of factors including nutrient exhaustion, grazing, sedimentation, or viral lysis and remains a fundamentally poorly understood process in all microalgae [36,37]. Environmental studies have shown that once chlorophyll levels begin to decrease (signaling the beginning of the end of the bloom), triacylglycerol (TAG) content increases [38,39]. Translating that into a biofuels production scenario, TAG accumulation precedes cell death and thus crashing would not necessarily be detrimental to production. Diatoms also grow best under highly mixed conditions, which are desirable for large-scale cultivation. Their large contribution to carbon export in the oceans relates to their relatively rapid sinking rates, which could translate into enhanced harvesting ability in a production system (see below). Their ability to sequester nutrients could be advantageous for crop management; supplying a nutrient pulse in which most of a particular nutrient is rapidly taken up and removed from the culture medium could reduce nutrient availability for contaminating organisms. Vacuole size and nutrient storage capacity should affect different aspects of production schemes. Chlorophytes, including Chlorella species, Chlamydomonas reinhardtii and Dunaliella salina, have small vacuoles (3–14% of their protoplast volume) compared with diatoms (45–90% of their protoplast volume), and should have a greater dependence on external nutrient conditions including consistency of supply [19]. In terms of lipid productivity, cells with larger vacuoles may take longer than cells with smaller ones to become nutrient limited, due to the enhanced nutrient storage capability.

▪ Diatom productivity in large-scale culture systems

Diatoms have been cultivated in large-scale outdoor systems for purposes other than biofuel production, such as aquaculture, for decades [40]. Widely ranging values for algal biomass productivity are reported in the literature and are due in large part to differences in cultivation parameters. Such parameters include, but are not limited to, light source and intensity, depth of culture, mixing rate, temperature, nutrients and supplementary carbon additions in the form of sugars or CO₂. In an attempt to limit some of the variability for the purpose of comparison, only values obtained from microalgae grown in outdoor ponds, raceways or open photobioreactors illuminated with natural light are included here. Diatom biomass productivity values range from 5 to 25 g m^-2 day^-1[30,41–43], and are in line with values obtained for other types of microalgae including the Eustigmatophyte Nannochloropsis salina[44], the cyanobacterium Spirulinaplatensis[45] and the Chlorophyte Chlorella sp. [46], whose biomass productivity values range from 10–25 g m^-2 day^-1. Cyclotella cryptica, a model diatom species from the Aquatic Species Program, was shown to have consistent productivity levels of 20 g m^-2 day^-1 (erroneously listed as 12 g m^-2 day^-1 in the text of this paper [41]).

What cellular factors contribute to diatom productivity?

As discussed previously, the evolution of diatoms is more complex and unique than that of green algae and terrestrial plants. As a result, there are fundamental differences in the biology (i.e. metabolism, physiology and subcellular organization) of the different algal classes [11]. Acknowledging diversity within a class, it is still reasonable to expect there will be advantages or disadvantages to certain classes in a large-scale biomass production setting. It is useful to outline some of the variation between classes so as to better understand the biology of the organisms that we propose to cultivate for fuel. Since chlorophytes are commonly considered as algal biofuels production strains and are fundamentally different than diatoms, we will focus on comparing these two classes. We acknowledge that there are other classes of photosynthetic organisms (Nannochloropsis, cyanobacteria) being considered for large-scale cultivation that are not included in this discussion.

▪ Differences in photosynthesis & carbon fixation comparing diatoms & green algae

Dissipation of excess light in photosynthesis either occurs at the level of photosystem II (PSII) and especially light-harvesting complex II (LHCII) via thermal dissipation or adjustment of the absorption cross section; or by redirecting electrons into pathways other than CO₂ fixation. Diatoms possess an additional xanthophyll-based cycle involved in energy dissipation relative to green algae and higher plants [47]. In a comparative study, the diatom P. tricornutum favored modulation at PSII/LHCII, with a substantial increase in de-epoxidation of diadinoxanthin to diatoxanthin, which reduced the flow of electrons into secondary processes. This was more efficient than C. vulgaris, where thermal dissipation at LHCII was less [48]. Green algae undergo state transitions where reorganization of antenna proteins shifts primary activity between PSII and PSI, especially under fluctuating light conditions [49]. State transitions increase photosynthetic efficiency, as has been demonstrated by mutants in this process, which have reduced capacity to make ATP via cyclic-photophosphorylation and have impaired growth [49]. No evidence for state transitions or a reaction center-based quenching mechanism has been found in diatoms [50], yet they are inherently more efficient than chlorophytes in balancing photosynthetic electron flow [48]. Perhaps the lower efficiency of chlorophytes involves the need for physical movements of antenna proteins over 500 nm [49], as compared with localized quenching by additional xanthophyll as occurs in diatoms. Diatom thylakoid membranes are arranged in groups of three and are not differentiated into grana and stroma lamellae [51,52], which affects redox signaling. Diatoms lack the α-carotene biosynthetic pathway, which means that both photoprotective and light harvesting pigments are synthesized from the same precursors, in contrast to the chlorophytes [48]. In addition, in contrast to the chlorophytes, diatom light-harvesting proteins are not differentiated into major and minor complexes [53,54], the Calvin–Benson cycle is not controlled by light via the redox state of thiol groups [55] and the oxidative pentose phosphate cycle is apparently not present in the chloroplast [55]. It is reasonable to assume that these substantial differences in photosynthesis, and carbon fixation and flux, contribute to diatoms’ environmental success.

▪ Carbon fixation efficiency & concentrating mechanisms

Both diatoms and chlorophytes possess carbon concentrating mechanisms (CCMs), which effectively increase CO₂ concentration in the proximity of RuBisCO. It is not clear whether there are advantages in CCM efficiency intrinsic to either class; however, there are significant differences in the ability of RuBisCO to discriminate between CO₂ and O₂. The chlorophytes have Form1B RuBisCO, with selectivity (Srel) values ranging from 54 to 83, whereas diatoms have Form1D RuBisCO, with a much higher Srel of 106–114 [56]. CCMs can involve two different mechanisms: biochemical (e.g., C4 metabolism), in which carbon is fixed by an enzyme other than RuBisCO and the fixation product or a derivative is decarboxylated in the vicinity of RuBisCO; or biophysical, which involves localized enhancement of CO₂ via equilibrium shifts involving pH changes or active transport of inorganic carbon across membranes [57]. The latter involves carbonic anhydrase (CA) and bicarbonate transporters. In diatoms, evidence for both C₃ and C₄ biochemical fixation exists, and the preference can apparently differ in relatively closely related species [58], which has led to confusion in attempts to make generalizations. P. tricornutum and T. pseudonana encode 9 and 13 CAs, respectively, but only P. tricornutum has β-CAs that are targeted to the chloroplast [59]. It is as yet unclear whether differences in CCMs relate to the different evolutionary time periods under which different classes of diatoms arose in relation to CO₂ levels. The documented differences may either reflect an organization that maximizes efficiency under a given CO₂ condition or, alternatively, that a variety of organizational schemes provide acceptable efficiency under current conditions. More research, especially localization studies, is required to clarify the situation. Variation in CCM mechanisms may not be unexpected; according to a modeling study, the advantage of a CCM would be best suited for an intermediate-sized algal cell, and there should be a high degree of co-adaptation in different uptake systems [60].

▪ Compartmentation & carbon metabolism in diatoms

Subcellular compartmentation enables incompatible metabolic processes to occur by physically separating them, and many biochemical processes (lipid metabolism, oxidative phosphorylation, photosynthesis) take place on the membranes of subcellular compartments, which also provide a greater area for these processes to occur [61]. Stramenopiles have additional intracellular compartmentation relative to chlorophytes as a result of their secondary endosymbiotic origin. In particular, the chloroplast is surrounded by two additional membranes, the chloroplast endoplasmic reticulum and the periplastidic membrane [62,63]. These two membranes are closely associated, but there is a noticeable space between the periplastidic membrane and the outer chloroplast membrane [64]. This periplastid compartment appears to be functionally significant, as several enzymes involved in myriad of cellular processes, including lipid transport and metabolism, have been shown to be targeted there [65]. Though the specific roles of this compartment are far from resolved, it is clear that it plays a part in regulating metabolism and overall cellular function.

Similar to green algae and terrestrial plants, diatom genomes contain genes for the essential enzymes of the conserved eukaryotic central carbon metabolic pathways of carbon fixation, glycolysis/gluconeogenesis and the citric acid cycle [8,10]. One difference in the glycolysis pathway is that neither T. pseudonana nor P. tricornutum encode a gene for a glucokinase [66]. Perhaps more surprising is the localization of the complete lower half of the glycolysis pathway from triose phosphate isomerase to pyruvate kinase to the mitochondria [66]. Diatoms apparently have the unique ability to convert glyceraldehyde-3-phosphate into pyruvate, along with the attendant intermediates, entirely in the mitochondria. The role of this partial pathway in the mitochondrion is unclear, though it could be a strategy to channel substrates towards the citric acid cycle. In terrestrial plants, an increased association between cytosolic glycolytic enzymes and the mitochondrial surface occurs during higher respiration, with attendant-increased efficiency through substrate channeling [67]. A mitochondrial-localized glycolysis pathway could also avoid the need for the malate-aspartate shuttle to import the reducing equivalents needed to generate ATP through oxidative phosphorylation.

Diatoms contain a complete urea cycle [10], which is not found in chlorophytes, and several metazoan-like urea cycle components are localized to the diatom mitochondria. In metazoans, the ornithine-urea cycle is only known for its essential role in the removal of fixed N; in diatoms, however, it apparently serves as a distribution and repackaging hub for inorganic carbon and N [68]. Experimental manipulation of the cycle indicates that it contributes significantly to the metabolic response of diatoms to episodic N availability [68].

Diatoms also differ from other classes of algae in terms of carbohydrate storage. Instead of starch, diatoms store fixed carbon as a complex carbohydrate called chrysolaminarin, which contains β(1→3) and β(1→6) linked glucose units in a ratio of 11:1 [69]. Chrysolaminarin is soluble and is stored in the chrysolaminarin vacuole (CV). The CV is quite large [70], and recent imaging work in our laboratory in collaboration with Robyn Roth at Washington University (MO, USA) indicates that it can comprise as much as 50% of the cell volume. Diatoms are not necessarily better at storing carbohydrates than green algae. For example, C. reinhardtii has quite an impressive ability to accumulate starch [71]. However, it may be important to consider the relative energetics involved in accessing a soluble carbohydrate (chrysolaminarin) stored in a vacuole versus an insoluble form (starch) stored in the chloroplast.

▪ The silica cell wall

The most conspicuous feature of diatoms is their silica-based cell wall, also called the frustule (Figure 1). The frustule consists of two halves that overlap like a petri dish, the epi- and hypo-theca. The ‘top’ and ‘bottom’ of the dish are made of structures called valves, and the sides are most typically composed of a series of thin siliceous bands that encircle the cell, overlap each other, and provide the overlap between the theca, called the girdle bands. During vegetative division, diatoms make new valves in an intracellular compartment called the silica deposition vesicle (SDV), and once complete, the entire structure is exocytosed [72]. After the two daughter cells separate, individual cells can sequentially synthesize new girdle bands, which enable elongation of the cell. The source of Si for diatoms is dissolved silicic acid, Si(OH)₄[73], which is taken up at low concentrations by specific silicic acid transport proteins, and at high concentration by diffusion across the plasma membrane [74].

The silica-based cell wall not only serves a protective function, but can contribute to diatoms’ overall productivity. Silica is polymerized from soluble silicic acid precursors in an energetically favorable autocatalytic process when silicic acid concentrations are sufficiently high (above 2 mM) and/or the pH is lowered [75]. In diatoms there are organic components that speed up this process; however, they are present in, at most, a few percent relative to the silica that is formed. The energetic needs for silica polymerization are thus much lower than an organically based cell wall, and calculations done for silicified plant cell walls indicated that silicification required only 3.7% of the energy compared with a lignin-based wall, and 6.7% for a polysaccharide wall [76]. In addition to the energetic savings, there should be savings in carbon, since a large percentage of the carbonaceous cell wall components are replaced with silica and the carbon can be used for other cellular purposes. Another feature of the cell wall that may relate to biofuels and specifically TAG accumulation, is that very different cell shapes can result, which may impose constraints on the size and number of lipid droplets that can be accommodated in the cell (Figure 2). There may actually be species-specific geometrical features that relate to optimal productivity.

The silica cell wall also apparently plays an important role in carbon concentrating mechanisms. Data suggest that diatom biosilica is an effective pH buffer, which increases carbonic anhydrase activity near the cell surface to enable the conversion of bicarbonate to CO₂[77].

Attributes of diatoms for biofuel production

In addition to efficient photosynthesis and nutrient utilization, diatoms possess several other attributes that relate directly or indirectly to useful characteristics for a biofuels production system.

▪ Environmental & trophic flexibility of diatoms

It is envisioned that algal biofuels development will have a regional character, and there will be a need to develop strains optimally adapted to different environments [1]. Further, the development of different types of culture strategies, ranging from outdoor open systems to enclosed photobioreactors to hybrid systems, and different trophic strategies, from auto- to hetero- to mixo-trophic, must be considered [78]. Thus algae with diverse culturing characteristics are desirable.

On this note, diatoms are found in a wide variety of environments, including general environments such as marine, freshwater and soil, but also specific extreme or variable environments of salinity, temperature (ice or hot springs) or pH. The most intensively studied diatoms are marine, and all of the species with sequenced genomes fall into that category. This does not preclude freshwater diatoms from being developed for biofuels production, but considering the issue of water supply, saline-tolerant species may be the most appropriate for development. Diatoms isolated from thalassic hypersaline marine environments include both centric and pennate species that grow well at salinities ranging from 0.5 to 15%, and three strains of Pleurosigma strigosum cannot grow in salinities less than 5% total salts, representing truly halophilic diatoms [79]. Different strains of P. tricornutum were shown to have near-linear growth rates (divisions per day) in salinities ranging from 5 to 70 psu [80]. Numerous species grow under acidic conditions (pH 3.3–4.4 [81]) and diatoms have been found to grow at pHs consistently maintained between 1 and 2.5 [82]. Conversely, a wide array of both pennate and centric species have been found in alkaline lakes, such as those in East Africa, where the pH can reach 10.6 [83]. Diatoms can grow at 60–65°C in hot springs [84] and at 40°C in thermal muds, which are a 5 g l^-1 colloidal aluminum silicate suspension mixed with thermomineral water, rich in sodium chloride and sulfates [85]. Psychrophilic diatoms grow on polar ice and in hypersaline brine channels at -1–2°C [86], with photosynthetic activities optimized between 4 and 7°C [87]. Diatoms can form biomasses in ice with chlorophyll concentrations that are matched only in hypertrophic waters or in extremely eutrophicated sediments [88]. The Fragilariopsis genus may be the most prevalent biomass producer among polar plankton [89,90], serving as the base of the entire polar food web. The range of different environments where diatoms can dominate sheds light on their diversity; in terms of biofuels production, growth in extreme conditions, such as salinity or pH can be utilized to keep pond contaminants and predators at low levels.

Diatoms exhibit trophic flexibility. The majority of species are autotrophic, but mixotrophy and obligate heterotrophy occur. C. cryptica, a candidate biofuels production strain, is capable of slow growth (10% of autotrophic) on glucose only in the dark, likely due to its glucose transport mechanism being induced in the absence of light [91]. The mixotrophs and obligate heterotrophs are generally able to use simple sugars as carbon sources, but some species are able to break down complex carbohydrates. There are two forms of mixotrophs: those that can grow strictly auto- and hetero-trophically, or those whose growth is only enhanced in the presence of an extracellular carbon source [92]. Several Nitzschia species can switch between strict auto- and hetero-trophic growth [93] and have been observed to burrow into solid medium containing agar, presumably breaking it down as a carbon source. In the natural environment, heterotrophic diatoms can be found on macroalgae (kelp), from which they obtain their carbon [94]. Nitzschia alba is an obligate heterotroph that has lost its chloroplasts, and can grow on both simple and complex carbohydrates [94–96], demonstrates growth rates in the field appreciably higher than other heterotrophic protists, and doubles in 6 h when cultured in the presence of glucose [95]. The ability of diatoms to breakdown complex carbohydrates suggests possible applications with different carbon feedstocks, such as cellulose.

▪ Lipid amount & composition

A primary target for biofuels production is the accumulation of lipid, either in the form of total lipid or TAG, which provide carbon for conversion to a variety of types of fuel [1]. Diatoms are excellent lipid accumulators, which partly may be due to extra buoyancy needed to counteract the effect of their relatively heavy silica cell walls. It is vitally important when comparing lipid content numbers to appreciate the variability in conditions in different studies; strict comparisons cannot generally be made. However, by comparing data from a sufficiently large number of studies some general conclusions can be formed.

In terms of total lipids, Shifrin and Chisholm [97] reported values of 12–45% of dry cell weight (DCW) for diatoms under conditions of N limitation. A comparative analysis of several species in different algal classes by Hu et al.[98] showed that on a DCW basis, chlorophytes had an average total lipid content of 25.5% under normal growth and 45.7% under nutrient stress. Regarding diatoms, lipid content was 22.7% DCW during normal growth and 44.6% during stress as indicated in the text, but this value was mistakenly listed as 37.8% in their Figure 1, as the line depicting the average is clearly above 40% [98]. A literature survey by Griffiths and Harrison [99] of 55 microalgal species in various classes showed that diatoms as a class performed above average at 27.4% (19% better than average) for total lipid content under nutrient-replete conditions and 36.1% (13% better than average) under N-deficient conditions. Diatoms also accumulate lipid as a result of Si limitation, and the average lipid content for diatoms limited for Si relative to N deficient conditions was 41%, which was 28% higher compared with all species and 13% higher when compared within the diatoms [99].

In terms of neutral lipid (TAG), Si-deficient cultures of C. cryptica were reported to contain 64% on an ash-free dry weight (AFDW) basis, compared with 32% in Si-replete cultures, and Cylindrotheca fusiformis contained 57% in Si-deficient cultures and 17–20% in Si-replete cultures [201]. A wide variety of TAG composition has been identified in diatoms, with fatty acid (FA) side-chains ranging from 14:0/16:0/16:1 to 16:1/20:5/22:6; however, there were ten predominant species, which were enriched in C14 and C16 [100].

Diatom FAs are among the most highly enriched in C14 chain lengths compared with other classes of algae [98], especially chlorophytes, which appear to lack this particular length [101–103]. Reports range from 4 to 32% C14:0 of the total FA composition for diatoms [98,103,104]. C16 FAs are highly represented in diatoms and C16:0 and C16:1 combined can represent 65–70% of the total FA, with C16:1 being predominant. C18 lengths are poorly represented in diatoms, but can be a predominant form in chlorophytes, especially polyunsaturated forms [98,102]. In general, shorter chain lengths are desirable to improve cold-flow properties of biofuels [105] and saturated FAs are more desirable because they increase the ignition quality (cetane number) of the fuel [106].

A critical point to make regarding measurement of total lipid or TAG in diatoms is that DCW measurements will underestimate the actual amount of lipid in terms of percent cellular carbon when compared with a similar measurement in other classes of algae because of the contribution of the noncarbonaceous silica cell wall. Comparison of eight different diatom species indicated ash contents from 49 to 59% [107]; further, diatoms had an average of twice the ash content of a variety of nonsilicified algae and four times the content of two Chlamydomonas species examined [97]. The silica content of three diatom species was measured at 22–47% of the dry weight [108]. Factoring this into DCW measurements, this suggests that lipid content in diatoms on a per cell carbon basis is substantially higher than usually reported. Illustrating this point is a comparison of Si-induced lipid accumulation in C. cryptica, in which total lipid based on DCW was 42% [97] compared with 54.1% when calculated on an AFDW basis [109]. The data suggest that diatoms are at least on par with, if not superior to, other classes of algae in terms of their ability to accumulate lipids.

▪ Other valuable hydrocarbons, EPA, DHA & isoprenoids

Diatoms are unusually highly enriched for eicosapentaenoic acid (EPA; C20:5, ω-3), which is not prevalent in chlorophytes [98,103,104]. Fish do not naturally produce EPA, but obtain it from algae they ingest, to which diatoms should contribute significantly. Docosahexaenoic acid (DHA) levels in diatoms are less than EPA, on the order of 5% of total lipid [98,103,104]. Diatoms contain a rich variety of carotenoids in the form of β-carotene, as well as a number of oxygenated derivatives called xanthophylls. These include fucoxanthin, neofucoxanthin, diadinoxanthin and diatoxanthin, some of which play important roles as accessory pigments and photoprotective agents in the light-harvesting complex of the photosystems [110]. Reports of diatom carotenoid content ranges between 0.07 and 0.6% of the dry weight (0.7–6.3 mg pigment/g dry wt), with pennates exhibiting a higher content than the centrics that were examined [108]. Some diatom xanthophylls are used commercially for their color and antioxidant properties [111], and it should be possible through recombinant techniques to dramatically increase production of specific desired xanthophylls in microalgae that naturally produce these pigments. Diatoms possess both the 2-C-methyl-D-erythritol 4-phosphate (MEP) or nonmevalonate pathway for carotenoid synthesis, and the mevalonate pathway (MEV) for synthesis of sterols, sesquiterpenes and triterpenoids [112,113].

▪ Antipathogenic or medical applications of extracted diatom components

Although not a major field of research, added-value chemical components of diatoms have been described. These include antibacterial and antifungal compounds [114–116], which are present in FA extracts [117,118]. Antibacterial activity extracted from Skeletonema costatum was shown to inhibit the growth of Vibrio, a pathogen of fish and shellfish [118]. Other extracts of S. costatum and Haslea ostrearia demonstrated anti-tumoral activities against human lung cancer and also had anti-HIV effects [119–121]. Highly branched C25 isoprenoidpolyenes (polyunsaturated sesterpenes oils orhaslenes) were identified as being responsible for these activities [122]. A recent report documented that marine benthic diatoms contain compounds able to induce leukemia cell death and modulate blood platelet activity [123] suggesting that diatoms may represent a richer source of interesting bioactive compounds than previously recognized.

▪ The advantages of Si limitation-based lipid accumulation induction

As far as has been documented, diatoms accumulate TAG only under limitation conditions. Published literature has shown that TAG accumulates under either Si or N limitation [98,109], and in Figure 4 we present a comparison of these conditions with a treatment where the cells were exposed to the microtubule-based cell cycle inhibitor nocodazole. There are several important points to consider from this. First is that only cell cycle arrest, and not nutrient limitation, is required for induction of TAG accumulation in diatoms (Figure 4). This has been substantiated by other experiments in our laboratory in which cell cycle stage is directly measured, and the results demonstrate that TAG does not begin to accumulate until arrest has occurred [Smith SR et al. Unpublished Data]. Often, it is stated in the literature that TAG accumulation occurs only under nutrient stress conditions; however, experiments in our laboratory indicate that, for diatoms stopping the cell cycle, even under nutrient-replete conditions is sufficient (Figure 4). A second point is that TAG accumulation under Si limitation lags slightly behind nocodazole treatment, but reaches and eventually exceeds the maximum level (Figure 4). The difference in induction start points may be related to a longer time required to halt the cell cycle under Si limitation. The third point is that TAG accumulation under N limitation lags substantially in timing and level of induction compared with nocodazole and Si. After a very long limitation period, abundant lipid droplets will form under N limitation, but only several days after similar sized droplets are seen under Si limitation [Traller JC, Unpublished Data]. The reason for the differential response comparing Si with N limitation is likely due to the differences in extent of intracellular storage of the two elements. N is stored intracellularly in a variety of forms and can be accessed over time, resulting in continued growth even after cells are placed in a N-free medium. Si is not appreciably stored in diatom cells, and hence placement of cells into a Si-free medium results in relatively rapid arrest and, therefore, a faster onset of TAG accumulation.

In terms of a production scenario, a lag time for induction of TAG accumulation would be unfavorable – increased time means decreased overall productivity. Thus, N limitation would not be desirable for diatoms, especially when Si limitation is an option. For this reason, it is a valid question whether diatoms without a Si requirement, such as P. tricornutum, would be competitive in a production system. The rate of TAG accumulation in chlorophytes resulting from N limitation varies among species. In C. reinhardtii, experiments on wild-type, starchless and complemented mutants showed that TAG levels increased on the order of twofold within 24 hours [124,125], suggesting a more rapid response under N limitation than with diatoms. Consistent with this, growth arrest in C. reinhardtii under N limitation appears to be fairly immediate [126]. In C. vulgaris, growth arrest was not immediate, and increased cell number was evident 2–4 days after placement in a N-free medium [126]. To our knowledge, no correlation between TAG accumulation and cell cycle arrest has been reported in chlorophytes, although correlation has been made between life cycle stage and TAG accumulation [101].

▪ The detrimental effects of N limitation

In addition to the time consideration, there are other drawbacks of having to rely on N limitation to induce TAG accumulation. There are numerous detrimental effects of N limitation on photosynthesis [127,128], which include decreased chlorophyll, alteration of thylakoid stacking, reduction in PSII cross-section and decreased energy transfer from the LHC to reaction centers. N limitation reduced photosynthetic rates and acetate uptake in Chlamydomonas[125] and LHC protein content decreased 10–40-fold, contributing to chlorosis [129]. Chlorophyll levels typically decrease under N limitation, which does not occur under Si limitation in diatoms [127,130]. N limitation also dramatically reduces ribosome content in C. reinhardtii[131]. N-deprived microalgae typically have reduced protein content; for P. tricornutum, N starvation led to a decrease in abundance of RuBisCO, especially the small subunit [132]. De novo protein synthesis is necessary for an increase in acetyl CoA carboxylase activity, the first committed step in FA biosynthesis [100,133]. General protease activity was increased in a diatom and Dunaliella tertiolecta under N limitation [99,134]. In terms of transgenic approaches to improve productivity, because of the overall decrease in protein synthesis under N limitation [127], any expressed proteins are likely to be less effective than desired. The combined effects suggest that N limitation is truly stressful to the cells, which raises the question of how cell ‘health’ is related to its ability to synthesize and accumulate TAG. In diatoms, Si limitation has little direct effect on other aspects of cellular metabolism than the cell cycle [135]; for example, protein and RNA synthesis are not appreciably affected [136].

▪ The ease of extraction of lipid from diatoms

On occasion we have heard comments that it must be more difficult to extract lipid from diatoms because of their solid cell walls. This is not true. Although cell wall components are solid, the overall structure is composed of discrete siliceous substructures that are joined together by organic molecules. The individual parts are relatively easily disrupted, and organic solvents readily extract lipid and pigments [137,138]. The diatom cell wall is actually easier to break open than the wall of some green algae; for example, RNA isolation procedures for diatoms involve simply vortexing cells in the presence of Tri Reagent [139], whereas for some green algae frozen cells have to be ground in a mortar and pestle prior to extraction. Furthermore, pigments are easily extracted from diatoms and blue-green algae using 90% acetone, whereas harsher solvents are required for the extraction of pigments from coccoid green algae [140]. In a biofuels production setting, where the energetic cost of extracting lipid is high, diatoms and other more easily extractible algae may provide a significant economic benefit.

▪ Settling rates, flocculation, harvesting & shear stress

One of the most expensive aspects of microalgal biofuels production is harvesting and dewatering; the need to concentrate a small percentage of solids from a much larger volume of water [141]. Factors that would reduce these costs include increasing settling rates or the ability of microalgal cells to flocculate. As stated previously, diatoms are major exporters of organic carbon from the ocean surface due to their increased settling rates relative to other phytoplankton [142]. The efficiency of this settling rate is likely to be at least partly dependent on the shape of the diatom. One report documented that elongated pennate diatoms were problematic in a water treatment facility because they did not settle well in the facility’s sedimentation tanks, causing clogs in the downstream sand filters [143]. However, diatom cell walls are covered with organic material having properties similar to other algal cell walls, and respond well to both cationic and anionic chemical flocculation treatments [144]. Further, diatoms are known to flocculate naturally and chain-forming marine diatoms can aggregate into centimeter-sized flocs in as little as 24 h [145]. The ability of diatoms to naturally flocculate is species-specific [146], facilitated by a variety of mechanisms ranging from direct adhesion to the cell surface production of exopolymers [147]. Some species may even produce anti-flocculation compounds [147]. Increased exopolymer secretion in conjunction with cellular protuberances that trap other particles is a common sinking mechanism in marine species, which is often triggered near the end of a bloom and generates sinking rates of 100 m day^-1 or more [148]. These data indicate that understanding diatom self-flocculation strategies may be of value in the context of biofuels production.

Diatoms are well adapted to turbulence, and in fact are more productive in high turbulence conditions than other classes of marine microalgae [149]. However, there is some apparent contradiction in the literature concerning this point, in that some studies indicate that diatoms are especially sensitive to shear during cultivation [150,151]. In one report, the authors concluded that micro-eddies on the scale of the cell size were responsible for shear stress [150]. A suggestion was made that diatoms may be especially sensitive to shear stress because of their rigid silica walls [152]; however, the particular species used in this study was predominantly unsilicified. Another study examined the passage of two diatom species through pumps and valves and concluded that the extent of cellular damage was related to operational parameters, such as the pressure drop and the rotational speed of the pump [153]. Cultivation in our laboratory has revealed that some species are impervious to highly turbulent mixing, whereas others are sensitive to modest agitation, suggesting that species-specific differences need to be taken into account with regard to shear stress; such differences have also been reported in the literature [150,154]. The data suggests that the issue is not with diatoms as a class, but particular diatom species subjected to particular turbulence conditions.

▪ Susceptibility to predators or pathogens & resistance to contamination

In spite of the protective cell wall, diatoms are not immune to predators and pathogens. A major class of diatom predator in the ocean is copepods, which are small crustaceans [155]. Interestingly, diatoms have apparently developed both physical and chemical defenses against copepods. A study demonstrated that diatom cell wall architectures were optimized to resist crushing [156], in this context by copepod jaws, and diatoms also produce aldehyde derivatives that have been implicated in reducing copepod reproductive viability [155]. Apofucoxanthinoid compounds (derived from fucoxanthin) produced by diatoms act as copepod feeding deterrents at concentrations of 2.2–20.2 ppm [157]. Small flagellates also feed selectively on diatoms [158]. In general, diatoms represent a bulky food for ingestors and their cell walls must be compromised during ingestion or else the intracellular contents are not completely accessible. Chytrid fungi infect diverse prokaryotic and eukaryotic algae, including diatoms [159]. Phytomyxea (plasmodiophorids) parasites have also been shown to parasitize diatoms [160]. Diatom viruses were first described in 2004 [161] and are either ssRNA or ssDNA viruses that infect the genera Rhizosolenia or Chaetoceros[162]. It is likely that the measures to deal with predation and pathogens that are being considered for other classes of algae will need to be applied to diatoms as well for biofuels production. In terms of cross-contamination of cultivation systems or environmental contaminants, one advantage of the defined shapes and sizes of diatoms is that it is generally easy to use microscopy to identify differences between diatom species, as well as to distinguish diatoms from other microalgae. The generally higher nutrient uptake rate of diatoms compared with other classes of algae (as mentioned previously) could also contribute to reduced contamination.

▪ Genomic & transgenic aspects of diatoms

Genetic modification is likely to play an important role in the development of algal biofuels, both as a research tool and in possible field applications once appropriate regulations are in place. Diatoms are well established in terms of genomic and transgenic capabilities. There are currently four available diatom genome sequences, for T. pseudonana, P. tricornutum, Fragilariopsis cylindrus and Pseudo-nitzschia multiseries. Of the four, F. cylindrus and Pseudo-nitzschia multiseries may not have biofuels relevance, because neither grows at a fast enough rate for a production system. Sequenced genome sizes in diatoms vary from compact at 27 Mbp to larger on the order of 300 Mbp, and GC content is close to 50% [8,10]. The latter point is important because many genes already developed for mammalian or yeast purposes (e.g., eGFP) can be used directly in diatoms with no modification, such as codon optimization [163]. Another important point is that there is no evidence for gene silencing in diatoms, as occurs in some chlorophytes [164]. In one study, a transgenic diatom was cultured in the absence of antibiotic selection for 1 year and retained resistance [165].

Currently diatom transformation is limited to the biolistic approach [165,166], which is effective, but inefficient. Higher efficiency methods would be desirable; for example, to enable saturation screening of insertional mutants or to complement mutations by introducing a library of genes. Integration of introduced DNA occurs randomly in diatom genomes [162], which can result in different levels of expression depending on the site of integration. This is not necessarily negative; on the one hand, more involved screening is required to identify useful transformants, but the variation in expression can be useful in obtaining a graded response. Homologous recombination has not been reported in diatoms, so gene replacement is not yet an option. Promoters that drive overexpression of genes have been identified [163,165,166], as well as a regulated promoter that can be turned on or off by changing the N source in the growth medium [167]. Gene knockdowns have also been accomplished using RNAi approaches [168]. Although there is room for improvement, a complete set of basic genetic manipulation tools are available for diatoms.

▪ Diatom metabolic diversity could provide abundant targets for bio-engineering

The evolutionary divergence of centric and pennate diatoms occurred at least 90 million years ago (based on the fossil record), yet their genomes differ from one another to the same extent as those of fish and mammals, which diverged approximately 550 million years ago [8]. The relatively rapid rate of diatom evolution has been at least partially attributed to the acquisition of horizontally acquired genes, which are thought to confer novel cellular and metabolic functions, and are drivers of diversification [8]. However, there are also substantial differences between diatoms in the targeting and homology of core genes involved in highly conserved eukaryotic pathways and processes. These include differences in the homology and distribution in enzymes of carbohydrate metabolism, the pentose phosphate pathway and carbonic anhydrases [59,68,169–171]. Differences in these core pathways are likely to affect metabolism and/or cellular energetics. These differences indicate that the trophic and environmental flexibility found within the diatoms has a genetic basis. Diversity in the core metabolism of diatoms means that there are multiple possibilities for manipulation of enzymes and pathways to improve growth and lipid accumulation characteristics.

Perceived disadvantages of diatoms in large-scale cultivation

The vast majority of diatoms require Si for growth, and thus a major difference between cultivating diatoms and other classes of algae is the requirement to add soluble Si to the medium to support growth. The presence of an additional nutrient can be perceived as a drawback in a culture system because of increased cost and, as in the case of Si, ancillary problems that could arise because of its aqueous chemistry. Complete details of the aqueous chemistry of Si can be found in Iler [75]. As a brief summary, soluble Si is present in aqueous solution as silicic acid, Si(OH)₄, which is the predominant form at pH 8.0, with a small fraction (3%) of the monoanion Si(OH)₃O^-. The solubility of silicic acid is limited at neutral pH to approximately 2 mM, above that concentration it begins to autopolymerize to first form multimers, which eventually coalesce into solid silica, SiO₂, which can precipitate from solution [75]. Silicic acid can form complexes with metal ions, and magnesium silicates are especially prone to precipitation [75]. Another perceived problem with silica relates to the important distinction between the amorphous silica found in diatoms and crystalline silica found in asbestos. Although chemically identical, their toxic effects are completely different. Needle-shaped crystalline silica is a causative agent in silicosis, in which scarring occurs in the upper lobes of the lungs. Amorphous silica is harmless to all organisms other than insects (see below) when breathed, ingested or contacted; thus, large-scale diatom cultivation will not produce a hazardous silica product.

There are advantages to the Si requirement for diatoms in cultivation. As discussed previously (Figure 4), Si limitation results in faster TAG accumulation than N limitation, with attendant reduction in production time. An interesting concept is whether a Si limitation lipid strategy actually requires more N than a species that accumulates under N limitation. There are two considerations regarding this: increased neutral lipid productivity under Si limitation (Figure 4) may outweigh the extra cost of N, and N could be included during the Si-limitation phase at levels much lower than used for growth to prevent the detrimental effects of N limitation as discussed previously, but not in excess of that, thus minimizing cost. Another benefit of the diatom’s Si requirement is that biomass accumulation can be precisely regulated by the amount of Si added to batch cultures. Many diatom species have well-defined silica content in their cell walls; thus, there is a strict dependency on how many cell divisions can occur with a given starting amount of Si in the medium. This is illustrated in Figure 5, where batch cultures of C. cryptica were grown with different starting concentrations of Si. Growth arrested in each culture on day 7, and the resulting culture density was precisely in the ratio of the starting amount of Si in the media. Such control could be highly valuable in a large-scale culture system because biomass attainment would both be predictable and controllable. Similar control may not be likely with other nutrients, such as N, given the issue of intracellular storage and the abundant presence of N sources in the environment (including N fixers). There are scenarios where Si limitation may be problematic in a production setting; for example, Si-containing dust could blow into an open system and perturb the control. However, the effects of this type of contamination should also be predictable once the rate of dissolution of airborne Si into the medium is determined. It should be noted that it is not likely that all diatom species will exhibit such precise control of biomass relative to Si concentration. Heavily silicified diatoms are known to reduce in size with succeeding generations (eventually triggering the sexual cycle), which leads to a mixed size population of cells with varying amounts of silica in their walls. An important point is that size reduction is not universal in diatoms, and in most laboratory species there is no evidence of size reduction, and have documented reasons why it does not have to occur [172].

Given its limited solubility (0.07% in water), the method of delivery of Si to a culture system deserves some consideration. Most production systems are maintained at or above pH 8, which is favorable for silicic acid solubility; however, there are still limits. A maximum of 2 mM Si could be added to a batch culture, but in addition to pushing the solubility limit, such high concentrations could affect the growth of diatoms. Environmental concentrations of silicic acid are rarely above 100 µM, and we have observed detrimental effects on some species in laboratory cultures at 400 µM. Fortunately, the amount of free silicic acid in a culture can be readily measured, either by estimating the amount based on biomass accumulated (Figure 5), or by direct chemical measurement using a colorimetric assay [173]. By regular monitoring, one could add Si when needed to promote high biomass accumulation without detrimental side effects.

A valid concern is what would be the added cost of including Si as a component of the growth medium. There is a lack of publically available information on Si’s projected costs; in nine cost analysis publications over the past 2 years, only one found one that mentioned the words Si or silica, and there were no cost figures for Si included [141]. Our discussion is therefore limited to general concepts. First is the availability of precursor raw material; Si is the second most abundant element on the earth’s surface, and so supplies are not limited. The true cost will be in the conversion of precursors to silicic acid. A variety of types of silica ranging from crystalline quartz (e.g., sand) to amorphous diatomaceous earth will dissolve in water, with amorphous forms dissolving faster [75]. Second is the cost related to the recycling of silica from spent medium or biomass. Because diatom silica readily dissolves at high pH [75], recovery of silica from lipid-extracted diatoms should be feasible, without detrimental effects on other cellular components, such as protein. Therefore, recovered Si could be recycled repeatedly, since the chemical form of dissolved Si will not be altered, as might happen with other nutrients such as N. A third factor is the possibility of using diatom silica as a value-added product, which is discussed below. The overall cost of adding Si to a production system will depend on these factors and has to be balanced against potential increases in productivity using diatoms in a Si limitation approach. To date, no studies have addressed these issues in a quantitative way, which obviously needs to be done to evaluate the possibilities and to provide impetus for potentially improving processing methods to reduce costs.

Diatom silica generated in a production system can be considered as a marketable product. Diatomaceous earth is already widely used in commercial processes and products. Due to the porous nature of diatom frustules, they are excellent filtering materials for very fine particles, and are primarily used in pool filters and for filtration of beer and wine. Diatom silica is also highly hygroscopic, and it is used as an insecticide because its contact with the cuticle stimulates desiccation of the insect in conjunction with perforating the cuticle with sharp-edged small particles. It is also used as a neutral anthelmintic (dewormer) in mammals. The abrasive nature of microscopic diatom particles is taken advantage of in metal polishes and toothpastes. Diatom silica, which is essentially glass, has optical properties, and when added to paint, makes it flat. Diatomaceous earth is added to plastic wrap to separate the layers slightly to prevent them sticking tightly to each other. Dynamite is a stabilized form of nitroglycerin, which is absorbed into diatomaceous earth.

Mined diatomaceous earth is usually a mixture of diatom species, and so can be relatively heterogeneous. It is possible that diatom silica isolated from a monoculture pond may have more consistent properties. One possible future application of diatom silica is in nanotechnology. Diatoms make nanostructured silica in a variety of shapes that cannot be achieved by current synthetic materials approaches. Silica has limited properties as a material but several approaches have been developed by which diatom silica can either be directly converted or used as a template to generate a completely different chemistry while maintaining the nanoscale structure [174].

Diatoms are likely to have other added-value applications. They are currently used as small percentage supplements to animal feed, and we are aware of studies evaluating the use of diatoms as a larger component of animal feed, ranging from fish to cattle. EPA from cultivated diatoms is less expensive than from fish sources [175]. Other applications, such as amino acids for cosmetics, antibiotics and antiproliferative agents, are at early stages of development [175].

Conclusion

A question commonly asked is, “If diatoms have such desirable properties and potential advantages, then why aren’t they more highly represented in biofuels research and production projects?” After considering their attributes as described in this article, it is puzzling. Two possible explanations come to mind. One is the long history of the study of chlorophytes due to their intimate relationship with terrestrial plants (especially related to photosynthesis) and our food supply. Another is related to the fact that diatoms dominate marine environments, whereas chlorophytes and plants dominate terrestrial environments. The ocean is a difficult system to study, and there are fewer research institutions with a marine focus than those dedicated to terrestrial environments; this imbalance may be an additional reason why the virtues of diatoms have been under appreciated. Generalized statements about the properties of ‘algae’ are also often heard. It is hoped that this review has highlighted the substantial differences not only between different classes of microalgae (specifically between diatoms and chlorophytes), but between different species within a class. Generalizations are not only inaccurate, but obscure the rich diversity of microalgae, which is a ‘value-added product’ in itself. At this early stage in the development of algal biofuels, little is known about the fundamental mechanisms involved in carbon fixation, flux and partitioning in microalgal cells, nor are scientists well versed in how to grow microalgae at scale. The deployment of algal biofuels on a production scale will involve largely uncontrollable and only partially predictable environmental conditions; thus, development of strains with diverse environmentally related productivity attributes will be necessary. For these reasons, it is valuable to investigate diverse classes of algae as research models and as potential production organisms. In a practical sense, this will define what is possible in terms of environmental adaptability and productivity, and allow a survey of the gene pool, which could be used to improve production species. As a class of algae, diatoms have all of the requisite attributes to be included as important contributors to the development of algal biofuels technology.

Future perspective

There has been substantial progress in the development of algal biofuels in the past few years. This may not yet be reflected in the publication record, because of the lag time in getting data published and corporate policies to protect data; however, at algal biofuels meetings it is clear that there is an undercurrent of excitement about things moving forward. The two major challenges in the development of this technology are the development of the “biology” to understand and improve factors that result in consistently high productivity of microalgal strains and development of large-scale culture systems that can grow algae and produce fuel precursors economically and at predictable production levels. Of the two challenges, understanding the biology is likely to be more straightforward, given the powerful tools already available to investigate and manipulate cellular metabolism. After an initial phase, over the next several years, of understanding and manipulating metabolic factors involved in fuel precursor synthesis in model species and initial production strains, a transition into applying these tools to a wider variety of production strains, which will be adapted to specific environmental conditions, is anticipated. This may also include multispecies cultivation, in which one or the other species is the primary producer under given conditions. This approach will require greater understanding of the ecological aspects of microalgal growth. Throughout the development of the organismal basis for biofuels production, an intimate interaction with the engineering and production aspects of large-scale cultivation will be needed.

Productivity

The ability of a species to grow rapidly to high density and accumulate fuel precursor molecules.

Diatoms

Unicellular microalgae of the heterokont class, characterized by silica-based cell walls and high productivity.

Lipid accumulation

The process of triacylglycerol accumulation in cells leading to the formation of lipid droplets.

Si requirement

The need by diatoms for an additional nutrient (silicon for cell wall synthesis) relative to other classes of algae. Silicon limitation will induce lipid accumulation in most diatoms.

Environmental flexibility

The ability of a microalgal species to adapt and grow in different environmental conditions, which is relevant to the development of biofuels production systems.

Executive summary

Evolutionary aspects of diatoms

	▪ Diatoms have a different evolutionary history than chlorophytes, which resulted in fundamental differences in cellular organization and metabolic processes.
	▪ On geologic timescales diatoms have become the predominant carbon fixers in the ocean.

Diatom productivity

	▪ Diatoms out-compete other classes of algae for growth, which may result from superior nutrient acquisition and storage and photosynthetic efficiency.
	▪ Diatoms are environmentally adaptable and trophically flexible.

What cellular factors contribute to diatom productivity?

	▪ Diatoms exhibit more efficient photosystem electron flux and increased selectivity for CO2 compared with other classes of algae.
	▪ Diatoms have substantial differences in intracellular compartmentation, the presence and localization of metabolic pathways and forms of carbohydrate stored compared with chlorophytes.
	▪ The rate of lipid accumulation is substantially improved under silicon limitation relative to nitrogen limitation, and numerous detrimental effects of nitrogen limitation are avoided.
	▪ The silica cell wall may reduce carbon requirements and energy for cell wall synthesis.

Attributes of diatoms for biofuels production

	▪ Diatoms have been cultivated for decades in outdoor open systems for aquaculture.
	▪ Diatoms naturally grow under bloom conditions – maximum growth to high density.
	▪ Diatoms accumulate abundant lipids, and lipid content may be underestimated unless based on an ash-free dry weight basis.
	▪ Diatoms produce added-value products.
	▪ The diatom silica cell wall may contribute to improved production through precise control of biomass accumulation and increased settling rates.

Perceived disadvantages of diatoms in large-scale cultivation

	▪ Diatoms require silicon to grow, but the costs of supply, recycling and value-added products relative to productivity have not been determined.
	▪ Abundant silicon supply exists, and recycling should be feasible.

Financial & competing interests disclosure

Diatom biofuels research in the Hildebrand laboratory is supported by AFOSR grant FA9550-08-1-0178, DOE grant DE-EE0001222 and NSF grant CBET-0903712. SRS was supported by the Department of Defense through the National Defense Science and Engineering Graduate Fellowship Program. The authors have no other relevant affiliations or financial involvement with any organization or entity with a financial interest in or financial conflict with the subject matter or materials discussed in the manuscript apart from those disclosed.

No writing assistance was utilized in the production of this manuscript.