This study (Lansford et al. 1990) specifically addressed the saline groundwater water resources in New Mexico. The objective was to identify suitable areas where large (1,000-ha) microalgae facilities could be established based on land and water availability. This report did not consider CO2 availability. Selection criteria developed by SERI, such as water quality, land slope, and climatic conditions, were used in this analysis. The groundwater resource information was reviewed for potential suitability for microalgae culture. Total gross water reserves of some 20 billion acre-feet were projected, of which about one-quarter was fresh water and the remainder of varying degrees of salinity. Freshwater sources would likely not be available in large quantities, as they would have higher value uses, would already be appropriated, or otherwise restricted. Thus, this report focused on saline groundwater sources. A first cut was by depth, likely well yields, size of reserves, and chemical composition. A detailed analysis of six groundwater basins of varying sizes and quality was then carried out. A qualitative analysis was carried out first, based on available data (Table III. C.2). Of these six, only two met all the criteria established by SERI for a microalgae facility, with the Tularosa Basin judged the best choice and Crows Flats next. For these regions, about 2.3 to 5 million acre-feet of useable water resources were identified as available. This report is an excellent example of the challenges of developing such a detailed resource analysis, to arrive at even an initial estimate for a single U. S. state, even after relaxing a key resource constraint (e. g., CO2).

I Publications:

Lansford, R.; Hernandez, J.; Enis, P. J. (1987) “Evaluation of available saline water resources in New Mexico for the production of microalgae.” FY 1986 Aquatic Species Program Annual Report, Solar Energy Research Institute, Golden, Colorado, SERI/SP-231-3071, pp. 227-248.

Lansford, R.; Hernandez, J.; Enis, P. J.; Truby, D.; Mapel, C. (1990) “Evaluation of available saline water resources in New Mexico for the production of microalgae.” Report, Solar Energy Research Institute, Golden, Colorado, SERI/TP 232-3597, 83 pp.

Table III. C.2. Suitable water resources in New Mexico.

(Source: Landsford et al. 1990.)

‘Unappropriated water is available, but competition from agriculture is Ikely because water quality is suitable for agriculture. "Unappropriated water is available, but competition from existing uses may exclude microalgae production.

50aia on depth to groundwater was available only far the Pecos Valley not the Pecos Basin.

"Ownership was not described lor the Pecos Basin. Ownership in the Pecos Valley was predominantly private.

‘Further study is recommended for the area around Roswell if less than a 1000 ha facility is considered.

III. C.7. Conclusions

The various ASP resources analyses indicated significant potential land, water, and CO2 resources, even within the limited geographic area (the southwestern United States) that was the focus of the ASP. Several quads (1015 Btu) of fuels were projected for the various available resources. Other areas, from Florida to California, could also be considered. Microalgae systems actually use fairly little water, compared to irrigated crop plants. In addition, many waste and saline water resources may be available and suitable for microalgae production. Many CO2 sources are available, and algal ponds could be purposefully co-located with CO2 sources, or even vice versa. This is already being done at a commercial microalgae facility in Hawaii. Finally, land is hardly a major limitation: two hundred thousand hectares, less than 0.1% of climatically suitable land areas in the United States, could, with maximal productivities, produce about 1 quad of fuels. Thus, although there are many practical limitations, which may make some earlier predictions optimistic, resource limitations should not be an argument against microalgae biodiesel systems.

Research during FY 1985-86 (Laws 1987a) elaborated on the two key findings mentioned earlier: effects of a 3-day dilution interval and of the foil arrays. The effects of foil arrays were tested over a 12-month period in the 48-m2 flume with Cyclotella sp., a diatom, which, like Chaetoceros, is a good lipid producer. The experiment involved alternatingly operating the pond with and without the foils for 2-week periods. The presence of foils increased productivity by almost a third, similar to the prior experiments.

The dilution effect was investigated with T. suecica, also in the 48-m2 flume, with similar results as before, in terms of both overall and maximal 3rd day productivity. However, solar conversion efficiencies were lower than observed in previous years, perhaps due to the approximately 3°C higher temperature during this year, compared to the previous one. The author speculated that this could have been close to the maximal permissible temperature for growth of T. suecica, and thus resulted in lower productivities.

However, the effect of dilution interval on production in the 48.4-m2 flume was somewhat puzzling (Laws 1986). These findings were a subject of considerable discussion and controversy. One possible explanation was the measurement of actual biomass density, which varied from about 27-28 g/m2 after dilution, to 80, 140, and 160 g/m2 for the 2-, 3-, and 4-day dilutions periods, respectively. However, this was considered an “unlikely” explanation. Indeed, the highest productivity was observed on day 3, with a steep decline on day 4. However, 4-day cycle cells still had lower productivity on day 3. Some “lingering effect of exposure to supraoptimal density conditions” was speculated to account for this phenomenon. The classical technique for studying such phenomena is the P versus I curve. Such studies were carried out with T. suecica cultures grown in the smaller 9.2-m2 flumes. However, as the author noted, the results were “somewhat discouraging” as there was no difference as a function of dilution intervals, and productivities were only about 24 g/m2/d, much lower than reported with the larger flumes. Thus, this issue remained as a major focus of this project.

IV. A.2.a. General Considerations

The conclusions presented elsewhere in this report focus attention on the fundamental issue of how to maximize the overall productivities of microalgae systems and suggest that there still is considerable scope for improvements. The issue of productivity, in its various guises and aspects, from species control to lipid (oil) yield and harvesting, is therefore recommended as a central subject for any future U. S. R&D program in microalgae biodiesel production. Essentially, the focus would be on developing the microbial catalysts that can convert solar energy to a liquid fuel at high overall efficiency. This effort will require a relatively long-term R&D effort, which would, at least initially, be focused on the fundamental and early-stage applied research required for such a biocatalyst development effort.

This recommendation implies that engineering design, cost and resource analyses, and even outdoor pond operations, discussed in Section III, would be relatively minor parts of such a projected R&D program, at least initially. The argument is that most of the variables of large-scale microalgal culture can be scaled-down to very small-scale, even laboratory systems. This allows detailed investigation of the key parameters in maximizing productivities. At present the central issues in scale-up are those of algal species dominance and grazer control (and other biological invasions). However, at this point these are secondary to the necessity and priority of establishing a high benchmark for productivity and lipid induction, under physicochemical and other environmental conditions that would allow extrapolation to large-scale outdoor systems at typical locations.

Again, this will require a relatively long-term R&D effort to accomplish, although guideposts, such as efficiency goals, to the needed advances can be provided. One issue is how to select strains for genetic improvements. This is still a difficult choice, as a relatively large investment is required to develop any novel genetic system. On the other hand, selection of the best strains for such a targeted genetics development effort is still some time off. Thus, a parallel track is recommended: strain selection (screening and improvements) would be carried out alongside with genetic engineering studies to demonstrate productivity enhancements using microalgal strains with already well-developed genetic systems. Such improvements would be in the efficiency of photosynthesis, described next, and lipid productivities, extending the ASP research reviewed in Section II.

A. (Top) — Histogram showing luciferase expression in protoplasts of C. ellipsoidea. Expression of the luciferase gene is expressed in relative light units (RLU), which are the net photons counted during a 5-min period. See text for explanation.

B. (Bottom) — Kinetics of luciferase expression in C. ellipsoidea protoplasts. Each symbol represents the result of a single assay. Control cultures were grown in the dark (▲) or light (A). Duplicate cultures of plasmid-treated protoplasts were also grown in either the dark (■) or light (□).

(Source: Jarvis and Brown 1991).

Development of Homologous Selectable Markers for Monoraphidium and Cyclotella:

Transient expression assays can be useful for the rapid assessment of DNA uptake and expression by cells as demonstrated by the expression of luciferase in Chlorella protoplasts, described earlier. However, attempts to produce similar results in other algal strains were unsuccessful. The problem with an experiment that produces no signal is that it is impossible to know if this is because the DNA did not get into the cell, or if the DNA entered the cell but was not expressed at detectable levels. In the latter case, poor expression could result from degradation of the foreign DNA, inappropriate regulatory signals, or differences in the codon usage.

One of the most promising organisms with regard to high lipid production and tolerance to environmental fluxes was the green algaM. minutum (strain MONOR2). However, MONOR2 DNA was shown to be highly unusual in GC content and degree of methylation. As mentioned elsewhere in this report, successful transformation of the green alga C. reinhardtii, which also has an elevated GC content, required the use of homologous selectable markers. The literature suggested that this unusual GC content would inhibit the expression of foreign genes, such as bacterial antibiotic resistance genes that had been used successfully as transformation markers in plant and mammalian systems. Based on this information, it was decided to attempt to develop homologous selectable markers for transforming MONOR2 and other strains with programmatic importance. Use of a selectable marker, in contrast to a transient expression assay, would allow the identification of very rare transformation events. Under the appropriate selection conditions, one transformed cell can be detected in a very large population of nontransformed cells, whereas in transient assays, a significant number of cells in a population must be expressing the foreign gene in order to detect the new enzymatic activity. The use of a homologous gene as a marker would greatly increase the chance for successful expression of the introduced gene, as there would be no problems associated with codon bias or foreign regulatory sequences. Although some success was achieved toward the development of a homologous selectable marker system, the emphasis of the research at NREL was shifted after the successful development of a transformation system for diatoms that used a chimeric selectable marker. A significant effort was put into the development of homologous markers, particularly for non-diatom species, from 1989 to 1994, so it is relevant here to summarize the progress made in this area.

The general protocol for developing a homologous selectable transformation system involves several steps. First, a mutation is created or identified in a specific gene. The gene should be essential for growth under “normal” conditions; however, the mutated strains will grow under modified growth conditions. This will allow for positive selection of transformed cells. Then the corresponding wild-type gene is isolated and inserted into a plasmid vector. The wild-type gene is introduced into the mutant cells, and transformants are detected by the ability to grow under the normal, defined growth conditions. In contrast to the transient assay described earlier, use of a selectable marker involves not only DNA entry and expression, but also stabilization of the new DNA in the cell and viability and growth of the newly transformed cells. Genes with good potential for use as selectable markers should not only code for a protein essential for growth

under defined conditions, but should also produce a protein that can be detected by a simple enzymatic assay. In addition, the use of a gene that has been well characterized in other systems will help isolate the gene from the species of interest and simplify the development of enzyme assays and growth conditions for isolating mutants and transformed cells.

Two genes that meet these criteria were targeted for the development of homologous selectable markers for MONOR2 and for C. cryptica T13L. One codes for the enzyme nitrate reductase (NR). NR had been used successfully to transform Chlamydomonas (Kindle et al. 1989) and several species of fungi (Daboussi et al. 1989) and methods were available to isolate NR mutants and selection of transformed strains. In addition, there was some interest at NREL in the role of nitrogen uptake and utilization in lipid accumulation, and isolating the wild-type NR gene would permit further investigation of these questions.

NR mutants can be isolated based on their resistance to chlorate. Cells with functional NR will take up chlorate along with nitrate and reduce the chlorate to the toxic compound chlorite. Therefore, cells with a mutation in the NR gene will be unable to grow using nitrate as the sole N source, but will be able to grow in the presence of chlorate, as long as urea or ammonium is added as an alternative N source. Using this scheme, several putative NR mutants grew from non-mutagenized cells of MONOR2 and C. cryptica T13L. Biochemical assays suggested that at least two of the MONOR2 mutants contained defects within the NR structural gene.

The next step was to isolate the wild-type gene from MONOR2 for complementation of the NR- minus mutants. A partial cDNA clone of NR from Chlorella vulgaris was obtained from Dr. Andrew Cannons (University of Southern Florida). Southern blot analysis indicated that the Chlorella DNA sequence showed significant homology to a sequence in MONOR2 genomic DNA. Degenerate primers for use in the PCR were designed based on conserved regions in the NR genes from three green algal species and several higher plants. A 700-bp PCR product was generated using MONOR2 genomic DNA as a template and confirmed to represent a fragment of the NR gene by sequence analysis. A MONOR2 genomic DNA library was constructed in a lambda phage vector. Although the library appeared to be representative of the algal genome in that it contained approximately 300,000 separate clones of about 20,000 bp each, repeated screening of the library with the NR gene fragment failed to produce any positive results. Two additional libraries were constructed, but again, screening with the MONOR2 NR sequence did not result in the isolation of a genomic NR sequence. It was concluded that the libraries were probably incomplete; i. e., they did not contain DNA representative of the total algal genome, possibly because of problems associated with the unusual composition of the MONOR2 DNA. This project was put on hold when successful transformation was achieved in C. cryptica, and had not been pursued further when the project was terminated in 1996.

A gene that encodes the enzyme orotidine-5′-phosphate decarboxylase (OPDase) was also targeted for use as a selectable transformation marker. OPDase is a key enzyme in the synthesis of pyrimidines. Organisms with defects in the OPDase gene will only grow if pyrimidines such as uracil are added to the growth medium. OPDase mutants can be selected by growing cells in the presence of the drug 5-fluoroorotic acid (FOA); OPDase converts FOA into a compound that

is toxic to the cells. Therefore, OPDase mutants would grow in the presence of FOA and require uracil; wild-type cells (or mutants transformed with the wild-type OPDase gene) would be susceptible to FOA and would require added uracil in the growth media. NREL researcher Eric Jarvis attempted to develop the OPDase system as a selectable marker for MONOR2. Cells were mutagenized by exposure to UV light, then grown in the presence of uracil and FOA. Putative OPDase mutants were identified as FOA-resistant colonies. Based on growth studies and spectrophotometric measurements of OPDase activity, one isolate of MONOR2 (3180a-1) was identified as a probable OPDase mutant for use as a host strain in the transformation system.

The next step, as for NR, was isolate the wild-type OPDase gene from MONOR2. OPDase had previously been isolated from several species and demonstrated significant sequence conservation between genes from different organisms. Dr. Jarvis made a number of attempts to isolate the OPDase gene from MONOR2 via PCR, using degenerate primers based on conserved OPDase gene sequences. Several PCR products were generated using this approach, but sequence analysis of the cloned DNA fragments resulted in no clones with homology to the OPDase gene. Why this approach did not work for OPDase is unclear, as this same PCR technique had been used to isolate a fragment of NR. A second approach, in which a MONOR2 genomic DNA library was screened for OPDase sequences using heterologous probes, was also unsuccessful.

By 1994, a transformation system had been developed for the diatoms using a chimeric gene as a selectable marker (discussed in the following section); however, there was still interest in producing a selectable marker system that would work for high lipid (although genetically recalcitrant) green algal strains, such as MONOR2. Work began on developing a new selectable marker system that used a mutated version of the acetolactate synthase (ALS) gene as a selectable marker. ALS is an enzyme involved in the synthesis of branched-chain amino acid such as leucine and valine. In plants, this enzyme is inhibited by sulfonylurea and imidazolinone herbicides. Previous work at NREL by Galloway (1990) showed that many microalgae are also sensitive to these herbicides. Eric Jarvis repeated these experiments for MONOR2 and demonstrated that these cells are sensitive to low levels of the sulfonylurea herbicides chlorsulfuron and sulfometural methyl. The approach was to isolate the wild-type gene for ALS from MONOR2, and then to produce a gene that encodes a herbicide-resistant form of the enzyme by site-directed mutagenesis. Degenerate primers were produced based on known ALS sequences and used, this time successfully, to isolate an ALS gene fragment from MONOR2 DNA. This sequence was used to screen the MONOR2 DNA libraries for a full-length ALS sequence, but once again, the screening efforts were unsuccessful.

The feeling among the NREL researchers was that the use of a homologous selectable marker system would still be the best approach for developing genetic transformation systems for some organisms, in particular, those with unusual DNA compositions, and for haploid organisms for which generation of mutants should be relatively straightforward. Despite the promise of M. minutum as a high lipid producer, it may have not been the best organism for these studies because of its highly unusual DNA properties and “tough” cell wall that complicated biochemical extractions and assays. Some of the cloning problems seen with this organism might have been

solved if time had permitted the generation of a cDNA library, or a new genomic DNA library using bacterial host strains optimized for use with highly modified or high GC DNA.

In the time frame between the end of the California projects and before initiation of the Roswell project, John R. Benemann and Dr. David Tillett carried out a research project in support of these outdoor pond projects (Benemann and Tillett 1987, 1990). This project included studies of the effect of nutrient limitations, specifically N and Si, on lipid induction, the effects of fluctuating conditions (temperatures, O2, light) on culture dominance and productivity, and the development of a mathematical model of the algal pond environment. That “Algal Pond Model,” (APM) could be used in designing experimental protocols and predicting culture performance.

A basic premise of this project was that the algal pond environment experienced by microalgal cells is characterized by relatively consistent and predictable fluctuations in a rather limited set of variables, specifically light intensity, temperature, pH/pCO2, and pO2. The first two are essentially uncontrolled, although somewhat predictable, variables. The second two factors are consequences of light and temperature, as well as the pond chemistry and hydraulics (outgassing), and algal productivity. Further, it was argued that some of these variables, such as the last three, could be well modeled, based on mass and heat balances and algal growth models. One uncertainty is whether short-term (<0.1 h) light fluctuations caused by pond mixing, can be averaged over the time scales of interest (>1 h). With this assumption it would be possible to predict, and reproduce on a small-scale, the key environmental parameters of outdoor ponds to which microalgae likely would respond, determining productivity and culture stability.

Such “down-scaling” of parameters, as summarized in Table III. B.4., would allow more realistic modeling at the laboratory, or very small outdoor scale, of the conditions encountered by microalgae in large-scale ponds. And it would allow, in turn, more controlled and easily interpretable experiments on species productivity and even dominance. As one of its objectives, and in collaboration with the prior studies (Section III. B.5.; Weissman and Tillett 1989, 1992), this project resulted in the development of an APM. Many experiments were also conducted on species growth responses and competition under fluctuating environmental conditions. Finally, lipid productivity was investigated under conditions of nutrient limitation.

This schematic for an advanced wastewater treatment process uses a multi-stage pond system for complete organic waste degradation and nutrient removal. The initial wastewater treatment ponds are shown, followed by a smaller intermediate “green algae” pond for N depletion and a final pond for cultivating N-fixing blue-green algae and removing residual phosphates. CO2 supplementation would be required in the last two ponds, and could increase productivity in the initial pond. (Source: Benemann et al. 1978.)

Figure III. A.4. Growth of microalgae on pond effluents.

Graph showing typical changes in density, chloropyll a content and ammonia concentrations during batch growth of green microalgae on effluents from wastewater treatment ponds after settling of the primary culture. As the culture grew from the residual concentrations of about 20 mg to about 170 mg/L, ammonia concentrations and chlorophyll a levels decreased. A 4- or 5- day culture period appears to be optimal. (Source: Benemann et al. 1978.)

III. D.1. Introduction

One of the major accomplishments of the ASP was the development of detailed engineering/cost projections for large-scale microalgae biofuels production. These analyses generally supported the view that microalgae biomass production could be performed at sufficiently low cost as to plausibly become a renewable energy source, assuming however, that the rather ambitious R&D goals of the ASP could be met. A major conclusion from reviewing these studies is that most R&D goals for this technology are related to the algal cultures themselves (productivity, species control, and harvestability), rather than the engineering aspects, such as the ponds, CO2 transfer, or biomass processing.

Historically, the first engineering and cost analysis for large-scale microalgae production of fuels was that of Oswald and Golueke (1960). These authors projected the costs of electricity generated from biogas (methane) obtained from the anaerobic fermentation of algal biomass. The algae were to be cultivated in very large (40-ha) raceway type ponds, mixed with pumps, and supplied with CO2 from a power plant. Other nutrients would come from the digesters. Municipal wastewaters would be used as make up for water and nutrients (C, N, P, etc.). The ponds were to be of earthen construction, with a depth of about 30-cm. Harvesting was assumed to be by simple settling. Electricity costs were projected to be competitive with nuclear power. Although few details were provided, the general concept outlined in this early publication has remained essentially unchanged. Perhaps the greatest change is that biomass productivities thought to be achievable at that time were less than 50 mt/ha/yr of biomass, while current projections are roughly two to five times higher.

With the initiation of the ERDA/DOE funded projects at the University of California-Berkeley during the mid-1970s (Section III. A.), additional engineering and cost analyses were conducted

(Benemann et al. 1977). The early studies were based on large (8-20-ha) ponds, with multiple channels and mixing by recirculation pumps (the required deep concrete sumps and splash pads were a major cost factor). Both the settling pond for harvesting algae by sedimentation and a covered anaerobic lagoon were part of this initial design. Total systems costs were only about $10,000/ha (somewhat over twice that in current dollars). Based on a projected yield of about 500 GJ/ha/y (10 GJ/t of algal biomass) of biogas, costs were projected at about $3/GJ. Although optimistic, this study served as a starting point for more detailed later studies.

I Publications:

Oswald, W. J.; Golueke, C. G. (1960) “Biological transformation of solar energy.” Adv. Appl. Microbiol. 11:223-242.

Benemann, J. R.; Baker, D.; Koopman, B. L.; Oswald, W. J. (1977) “A systems analysis of bioconversion with microalgae.” Proc. Symposium Clean Fuels from Biomass and Wastes, (Klass, D., ed.) Institute of Gas Technology, Chicago, pp. 101-126.

During the final year of the Hawaii ARPS project (Laws 1987), the goal was to screen for additional algal species in the smaller flumes and to further study the effect of dilution intervals. Four species were tested in the 9.2-m2 flumes: Navicula sp., C. cryptica, C. gracilis, and Synechococcus sp. From prior work (see earlier Section, and also Laws et al. 1986, 1987a), photosynthetic efficiencies of 9.1% were reported with T. suecica, during a 78-day period, and 9.6% for 122 days with C. cryptica. With the three other organisms listed above, somewhat lower efficiencies were noted during shorter time periods: 7.8 % for Navicula sp., 8.5% for C. gracilis, and 8.6% for Synechococcus. Somewhat “surprisingly” (their characterization), they observed that in a 2-day batch growth mode, initial cell concentrations ranging from about 50 to 400 mg/L (AFDW) had no major effect on productivity. For C. cryptica, at an initial concentration of 40 mg/L at a depth of 12 cm, this would give an areal cell density of about 5 g/m2. For an equal daily productivity of 30 g/m2/d, averaged over 2 days, this would require the cells to divide 2.5 times the first day, and once the second day. Not impossible, certainly, but somewhat problematic. There is indeed some likelihood that some systematic measurement error influenced their productivity measurements (John Ryther, private communications, circa. 1986).

This report also described lipid induction by Si limitation by C. gracilis and C. cryptica. In both microalgae Si limitation greatly reduced overall productivities, and lipid productivities, even though lipid contents increased. Laws (1987) concluded that lipid productivities would be maximized by maximizing total biomass production.

In the final paper, Laws et al. (1988), reported on long-term (13-month) production of C. cryptica in the large flume, with a 9.6% solar conversion efficiency reported with the foils and 7.5% without the foils, similar to earlier results with T. suecica. For 122 days, at optimal dilution (2- day batch cycle) productivity of about 30 g/m2/d was measured. This is, indeed, a high sustained productivity.

Many environmental factors affect the performance of the complex photosynthetic machinery in microalgae, reducing its efficiency to well below the maximum at which photosynthesis can perform. That maximum is dictated by the underlying mechanisms, biophysical constraints, and physiological adaptations. One objective of applied microalgal R&D would be to develop strains and techniques that achieve productivities as close as possible to the maximum.

However, somewhat surprisingly, there is still argument about the maximum limit for photosynthetic efficiencies. The arguments boil down to the mechanisms assumed and the many possible loss factors that may or may not be considered. Most researchers agree that an absolute minimum of eight quanta (photons) of light absorbed are required by the two-photosystem mechanism (Z-scheme) of photosynthesis to reduce one molecule of CO2 (and closer to 10 to 12 quanta if the energy needs for CO2 fixation and cell metabolism are considered). However, there have been many reports of higher efficiencies. For example, recently Greenbaum et al. (1995) reported that some algal mutants lacking one photosystem still fixed CO2 (and produced H2), suggesting less than 8 (and as few as 4) quanta per CO2 reduced. However, recent reports cast doubts on this interpretation, and the two-photosystem mechanism appears robust.

The maximum efficiency can be estimated at about 10% of total solar (Bolton 1996). Such efficiencies have been used in the projections for microalgae biodiesel production (see Section III. D.). However, high sunlight conversions are observed ony at low light intensities. Under full sunlight, typically one-third or less of this maximal efficiency, biomass productivity is obtained, because of the light saturation effect.

Light saturation is simply the fact that algae, like many plants, can use efficiently rather low levels of light, typically only 10% of full sunlight (and often even less). Above this level, light is wasted. In fact, full sunlight intensities can damage the photosynthetic apparatus, a phenomenon known as photoinhibition. Light saturation and photoinhibition result from several hundred chlorophyll molecules collaborating in light trapping, an arrangement ideally suited for dense algal cultures, where on average a cell receives little light. However, exposed to full sunlight, the photosynthetic apparatus cannot keep up with the high photon flux and most of the photons are wasted, as heat and fluorescence, and can damage the photosynthetic apparatus in the process. One possibility, suggested by Neidhardt et al. (1998), is that photosynthetic productivity and light utilization could be maximized in microalgae by reducing the size of the light-harvesting antenna through mutation or genetic engineering. This is an interesting idea that will be discussed further in the next section.

I Publications:

Bolton, J. R. (1996) “Solar photoproduction of hydrogen.” Report to the Int. Energy Agency, under Agreement on the Production and Utilization of Hydrogen, IEA/H2/TR-96.

Greenbaum, E.; Lee, J. W.; Tevault, C. V.; Blankinship, S. L.; Metz, L. J. (1995) “Carbon dioxide fixation and photoevolution of hydrogen and oxygen in a mutant of Chlamydomonas lacking photosystem I.” Nature, August 3rd, (1995).

Kok, B. (1953) “Experiments in photosynthesis by Chlorella in flashing light.” In Algal Culture: From Laboratory to Pilot Plant (Burlew, J. B., ed.), Carnegie Inst. of Washington, Publ. 600, pp. 63-75.

Kok, B. (1973) “Photosynthesis.” Proceedings of the Workshop on Bio Solar Hydrogen Conversion (Gibbs, M., et al., eds.), September 5-6, Bethesda, Maryland, pp. 22-30.

Melis, A.; Neidhardt, J.; Bartoli, I.; Benemann, J. R. (1998) Proc. Biohydrogen ’97.

Neidhardt, J.; Benemann, J. R.; Baroli, I.; Melis, A. (1998) “Maximizing photosynthetic productivity and light utilization in microalgae by minimizing the light-harvesting chlorophyll antenna size of the photosystems.” Photosynthesis Res., in press.

Based on the frustrating efforts to produce viable protoplasts from microalgae discussed earlier, efforts were initated to develop other methods for introducing DNA into microalgal cells through the intact algal cell walls. At the time this research was going on, the only microalga for which there was a reproducible transformation system was C. reinhardtii. Early efforts to transform this organism were facilitated by the availability of wall-less cells, either genetic mutants (cw — 15), or cells whose walls were degraded using autolysin, a species-specific cell wall-degrading enzyme produced during mating by C. reinhardtii gametes. High-frequency nuclear transformation was accomplished by agitating these wall-less cells in the presence of plasmid DNA, glass beads, and polyethylene glycol (Kindle 1990). This method was reported to work for walled cells, but at a very low frequency. DNA could also be introduced into walled cells of Chlamydomonas and into higher plant cells using microprojectile bombardment, or biolistics; however, this technique requires very expensive, specialized equipment. (This technique will be described in detail “Development of a Genetic Transformation System for the Diatoms Cyclotella and Navicula.”)

During the early 1990s, several reports demonstrated the feasibility of using silicon carbide whiskers (SiC) to mediate the entry of DNA into intact plant cells (Kaeppler et al. 1990; Asano et al. 1991). NREL researcher Terri Dunahay decided to try this approach to introduce DNA into intact algal cells. As reliable selectable markers were not yet available for any oleaginous microalgal strain, she decided to use Chlamydomonas as a model system. A strain of C reinhardtii that contains a defect in the gene for nitrate reductase (CC2453 nit1-305 mt-) was obtained from the Chlamydomonas Genetics Center at Duke University, Durham, North Carolina. These cells cannot use nitrate as a N source, but grow well in the presence of ammonia or urea. Kindle (1990) had shown previously that NR-deficient cells could be transformed with the Chlamydomonas wild-type gene for NR; transformed cells expressing the added DNA could be detected by their ability to grow on nitrate as the sole N source. A plasmid containing the wild-type NR gene from Chlamydomonas was obtained from Dr. P. Lefebvre at the University of Minnesota, St. Paul, Minnesota. A protocol for SiC-mediated transformation was developed based on the glass bead transformation protocol of Kindle (1990). Exponentially growing cells were washed once in NH4+-free medium, then suspended in the same medium with plasmid DNA, sterilzed SiC whiskers, and polyethylene glycol (mw 8,000) to a final concentration of 4.5%-5.0% w/v. The samples were agitated using a vortex mixer for periods from 30 seconds to 10 minutes, then diluted into NH4+-free medium containing 0.6% agar (top agar) and plated onto agar plates that contained the same medium. Transformed colonies (containing a funtional NR gene) appeared in 1-2 weeks.

Attempts to transform walled cells of Chlamydomonas using SiC were made in parallel with glass bead-mediated transformation to compare the two procedures. The results of a typical experiment are shown in Figure II. B.7. The number of transformants obtained using SiC varied

between experiments, but generally were in the range of 10-100 per 107 cells, comparable to transformation efficiency obtained with glass beads. Probably the most significant finding was the difference in cell viability after being agitated with either glass beads or SiC fibers. The viability of the cultures was greater than 80% even after agitation with SiC fibers for 10 minutes; only 10% of the cells survived agitation with glass beads for 60 seconds. The fact that SiC — mediated transformation appears to be a more “gentle” protocol than glass bead treatment may be important when adapting the transformation procedure to other species that may have different wall properties. This work resulted in two publications (Dunahay 1993; Dunahay et al. 1997). The second paper was a collaboration with Dr. Jonathan Jarvik at Carnegie Mellon University. Dr. Jarvik’s laboratory adopted and refined the SiC protocol and now uses it routinely to generate transformants in Chlamydomonas strains with intact walls. After the initial development of the SiC protocol, there was some work at NREL to adapt this procedure for other algal strains of interest to the biodiesel project. Initially, no genetic markers for these strains were available; however, the viability of Monoraphidium and Cyclotella were tested following agitation with SiC; both strains showed high survival rates after extended agitation with SiC. However, the successful development of a transformation system for Cyclotella using biolistics (discussed later) precluded further work on SiC-mediated algal transformation. A few attempts were made to generate transformants of Cyclotella or Navicula using SiC once a selectable marker system was developed. Only one transformant was generated in one experiment. The silica frustule of the diatoms likely acts as a significant barrier to penetration by SiC fibers. SiC would probably work better for introducting DNA into non-diatom cells such as Monoraphidium; these cells are very small and may not be a good target for biolistics, but might be readily pierced by SiC fibers.

. Cell survival and transformation efficiency of intact C. reinhardtii following vortex mixing with SiC fibers or glass beads. (Source: Dunahay 1993.)