Of the species examined, P. tricornutum and T. sueica had the highest overall productivities. These species also had the highest lipid productivities, which were 4.34 and 4.47 g lipid^m-2^d-1, respectively. For both species, the maximal productivities were obtained in batch cultures, as opposed to semi-continuous or continuous cultures. Although the lipid contents of cells were often higher in response to N deficiency, the lipid productivities of all species tested were invariably lower under N deficiency because of an overall reduction in the culture growth rates. For the species tested under continuous or semi-continuous growth conditions, lipid productivities were reduced from 14% to 45% of the values measured for N-sufficient cultures.

The results also pointed to the importance of identifying strains that are not photoinhibited at light intensities that would occur in outdoor ponds. Finally, this work highlighted the fact that some microalgae accumulate carbohydrates during nutrient-deficient growth; such strains are clearly not acceptable for use as a feedstock for lipid-based fuel production.

(Also see references listed in the following section.)

II. A.2.c. Selection of High-Yielding Microalgae from Desert Saline Environments

University of California, San Diego William H. Thomas 1983 — 1986 XK-2-02170-0-01

The work carried out under this subcontract represented one of the first efforts to collect microalgae from inland saline habitats and to screen those strains for rapid growth rates and lipid content. Collecting trips were made to eastern California and western Nevada, and initial culturing efforts were conducted at the Sierra Nevada Aquatic Research Laboratory near Mammoth Lakes, California, and the Desert Studies Center (Zzyzx Springs), which is near Baker, California. Various saline waters and soils were sampled during these collecting trips. The collection sites included Pyramid Lake, Black Lake, Owens Lake, Walker Lake, Saline Valley, Zzyzx Springs, Armagosa River, Sperry River, Harper Lake, and Salt Creek. The water samples were enriched with N, P, Si, and trace metals, then incubated under natural conditions. The algae that grew up were isolated by the use of micropipettes. Soil samples were placed in “Zzyzx medium” before algal isolation (see Thomas et al. [1986] for media compositions).

Diatoms, green algae, and cyanobacteria were the dominant types of algae isolated using these procedures. Of the 100 strains isolated, 42 were grown under standardized conditions in various

artificial media that were designed to mimic the water from which the strains were originally obtained. The pH of these various media formulations was typically high because of the presence of high levels of carbonate and bicarbonate, and the total dissolved solids ranged from approximately 1.5 g^L-1 to over 260 g^L-1. The growth of the cultures was visually scored, and nine of the fastest growing strains were further analyzed with respect to growth under scaled-up conditions.

For larger scale cultures, 6 L of medium that was enriched with N (nitrate, urea, or ammonium), phosphate, trace metals, and vitamins were placed in 9-L serum bottles, and the cultures were illuminated with fluorescent bulbs at a light intensity of 18% full sunlight. To enhance growth, the cultures were bubbled with 1% CO2 in air, and more nutrients were added as the cell density of the cultures increased. The strains tested in this manner included Nitzschia, Ankistrodesmus, Nannochloris, Oocystis (two strains), Chlorella (three strains), and Selenastrum. The estimated productivities ranged from 8.8 g dry weighHm-2M-1 for for Nitzschia S-16 (NITZS1[5]) grown in the presence of urea to 45.8 g dry weighHm-2M-1 for Oocystis pusilla 32-1, which was also grown with urea as the N source. (These productivity values were considered overestimates in that there was incidental side lighting of the flasks under the incubation conditions.) Also the productivity values did not always correlate with final biomass yield values, indicating that growth saturation was reached at different culture densities for the various strains. The maximum biomass yield was obtained with O. pusilla 32-1 (2.29 g dry weighHL-1). The results of these experiments indicated that certain strains had a clear preference for either urea or nitrate as the N source. Because urea is significantly less expensive than nitrate, these results have economic implications with respect to algal mass culture. However, the results regarding a preference for a particular N source were not always reproducible.

Additional experiments were carried out to assess the combined effect of temperature and salinity on the growth of several of the isolates. A thermal gradient table was used for these experiments in which the incubation temperature at six stations on the table varied between 11 °C and 35°C. Salinities of the media were varied in five increments along the other axis (as much as twice that of the natural waters), leading to a total of 30 different combinations of temperature and salinity. Growth was determined via optical density measurements, and contour lines were drawn upon a matrix chart of the various temperature/salinity combinations. This approach was later used by other subcontractors and SERI in-house researchers to determine the optimal growth characteristics of numerous promising algal strains. Results from this analysis were reported for eight different strains. In general, the strains grew better at higher salinities, indicating their halophilic nature, and had temperature optima for growth between 20°C and 30°C. Of the strains tested, the Mono Lake isolate NITZS1 had the highest temperature optimum (between 30 and 36°C).

The effect of light intensity on the growth of Ankistrodesmus falcatus 91-1 (ANKIS1, from Pyramid Lake) and O. pusilla 32-1 (from Walker Lake) was determined in 0.83-L cultures that were placed at varying distances from a tungsten lamp source. Neutral density filters were

placed between the light source and the cultures. This arrangement provided between 30% and 70% of full sunlight. For ANKIS1, maximum productivity (21.9 g dry weight^m-2^d-1) was attained when the cells were subjected to 50% full sunlight. At 30% full sunlight, the productivity fell off to 14.7 g dry weight^m-2^d-1 (because of light limitation), and at 70% full sunlight, the productivity was reduced to 19.0 g dry weight^m-2^d-1 (most likely because of photoinhibition). For O. pusilla, a maximum productivity of 25.8 g dry weight^m-2^d-1 was also attained at 50% full sunlight, whereas productivity at 30% and 70% full sunlight was 18.7 and 23.2 g dry weight^m-2^d-1, respectively. These productivity values are believed to be more accurate than those reported in the preceding section because light was able to enter the culture vessels only from one side. The relative high densities of these cultures (>1 g dry weights-1) permitted the cells to tolerate higher light intensities than would be possible in less dense cultures, because of the self-shading of the cells.

Experiments were also conducted to determine the effects of varying the culture vessel width (i. e., culture depth) on overall productivities. ANKIS1 was grown in containers that were 5, 10, and 15 cm wide. The containers were illuminated with a tungsten lamp at 50% full sunlight, and were bubbled vigorously with 1% CO2 in air. The results of these experiments indicated that growth rate, when expressed as g dry weight^L-1^d-1, was highest in the 5-cm thick culture and lowest in the 15-cm thick culture. However, when productivities were expressed as g dry weight^m-2^d-1, which takes into account the actual surface area that is illuminated, the thicker cultures were more productive. For example, the volumetric productivities over 10 days were 0.72, 0.35 and 0.31 g dry weight^L-1^d-1) for the 5-, 10-, and 15-cm thick cultures, respectively, whereas the corresponding areal productivities for these cultures were 41.1, 40.2, and 52.7 g dry weight^m-2^d-1. Because the economic constraints regarding an actual algal biodiesel production facility dictate the use of open pond systems, the areal productivity values are the more important consideration, although less water (and consequently less water handling) is required when dealing with more dense cultures. The productivities reported for these experiments may be overestimates of what these strains could achieve in outdoor mass culture because of the optimized mixing and aeration regime.

In the final year of this subcontract, additional collecting trips were taken to gather more microalgal strains. These strains, along with some that had been collected during earlier trips, were screened more rigorously than before. In this revised selection process, the strains were subjected to higher light intensities and higher temperatures, and the abilities of the strains to grow in “SERI standard media[6]“ were investigated. This selection procedure resulted in the isolation of 41 additional strains. Initial screening of strains involved incubating the isolates at 25 °C and 30°C under 40% full sunlight (provided by a 2000-W tungsten-halide lamp) in the SERI standard medium that most closely resembled the water from which they were originally isolated. Twelve of the strains that grew best under these conditions were then tested under the same temperature and light conditions in an early version of standard SERI media (Type I and Type II at low, medium, and high salinities; see media compositions in Thomas et al. [1985]).

The results indicated that most of the strains had a definite preference for a particular medium type and level of salinity. The results also indicated some inconsistencies in the growth rates of cells grown in the two experiments. For example, Chlorella BL-6 (CHLOR2) grew very well in the preliminary experiments in Type II/low salinity medium (2.48 doublings^-1), but grew much more poorly when grown in all five SERI media (including Type II/low salinity) in the second set of experiments. Conversely, Chlamydomonas HL-9 grew much more quickly in the second set of experiments than in the first set. The reasons for these discrepancies are unclear, as the culture conditions were essentially the same, and underscore the need to perform replicate experiments. Several marine microalgae were also tested for the ability to grow in SERI standard media. Phaeodactylum tricornutum, Chaetoceros gracilis, and Platymonas all grew well (>1.25 doublings^-1) in at least one SERI medium. Isochrysis T-ISO was unable to grow in any SERI medium, however.

The combined effects of temperature and salinity on the growth rates of eight of these newly collected strains were determined by the use of a temperature-salinity gradient table. In general, the strains grew best in the range of salinity that was similar to that of the water from which they were originally isolated. The optimal temperature for growth was generally in the 25°C to 35°C range, although one Chlorella strain from Salt Creek grew well at 40°C. Based on the results of these experiments, two strains were selected for analysis of growth characteristics in larger scale (12 L) cultures at 50%-70% full sunlight. CHLOR2 achieved a productivity of 55.5 g dry weight^m-2^d-1 under these conditions, and Nannochloris MO-2A had a productivity of 31.9 g dry weight^m-2^d-1.

This subcontract represented one of the first efforts in the ASP to collect and screen microalgal strains to identify suitable biofuel production strains. As a consequence, many of the screening and characterization protocols were still being developed; therefore, there is a substantial lack of uniformity in the testing of the various strains isolated. Nonetheless, a number of promising strains were isolated during the course of this research, and several methods were developed that helped establish standard screening protocols used by other ASP researchers.

Publications:

Thomas, W. H.; Gaines, S. R. (1982) “Algae from the arid southwestern United States: an annotated bibliography.” Report for Subcontract XK-2-0270-01. Solar Energy Research Institute, Golden, Colorado, October 1982.

Thomas, W. H.; Seibert, D. L.R.; Alden, M.; Eldridge, P.; Neori, A.; Gaines, S. (1983b) “Selection of high-yielding microalgae from desert saline environments.” Aquatic Species Program Review: Proceedings of the March 1983 Principal Investigators ’ Meeting, Solar Energy Research Institute, Golden, Colorado, SERI/CP-231-1946, pp. 97-122.

Thomas, W. H. (1983b) “Microalgae from desert saline waters as potential biomass producers.”

Progress in Solar Energy 6:143-145.

Thomas, W. H.; Seibert, D. L.R.; Alden, M.; Eldridge, P. (1984c) “Cultural requirements, yields, and light utilization efficiencies of some desert saline microalgae.” Aquatic Species Program Review: Proceedings of the April 1984 Principal Investigators ’ Meeting, Solar Energy Research Institute, Golden, Colorado, SERI/CP-231-2341, pp. 7-63.

Thomas, W. H.; Tornabene, T. G.; Weissman, J. (1984d) “Screening for lipid yielding microalgae: Activities for 1983.” Final Subcontract Report. Solar Energy Research Institute, Golden, Colorado, SERI/STR-231 -2207.

Thomas, W. H.; Seibert, D. L.R.; Alden, M.; Eldridge, P. (1985) “Selection of desert saline microalgae for high yields at elevated temperatures and light intensities and in SERI Standard artificial media.” Aquatic Species Program Review: Proceedings of the March 1985 Principal Investigators ’Meeting, Solar Energy Research Institute, Golden, Colorado SERI/CP-231-2700, pp. 5-27.

Thomas, W. H.; Seibert, D. L.R.; Alden, M.; Eldridge, P. (1986) “Cultural requirements, yields and light utilization efficiencies of some desert saline microalgae.” Nova Hedwigia 83:60-69.

From 1992 to 1995, Dr. Jane Schneider worked at NREL with Dr. Roessler on a project funded by the United States-Israel Binational Agricultural Research and Development Fund. The research was performed in collaboration with Dr. Assaf Sukenik and other scientists at the Israel Oceanographic and Limnological Institute in Haifa. The goal of the research was to understand the biochemistry of lipid synthesis in the eustigmatophyte Nannochloropsis sp., particularly with respect to fatty acid desaturation pathways. There has been a significant amount of research on lipid synthesis pathways in higher plants, and the pathways hve been assumed to be similar in lipogenic algae. However, unlike plants, nutrient deprivation produces major effects on the quantity and quality of lipids in algae; so there are likely to be significant differences in the biochemical pathways. In addition, like many algae, Nannochloropsis contains a high proportion of long fatty acids (i. e., C-20, C-22) with a high degree of unsaturation (20:5). These very long chain-polyunsaturated fatty acids (VLC-PUFAs) are important in aquaculture applications as they improve the nutritional quality of feed for fish and shellfish, and have nutritional and pharmaceutical applications for humans. Understanding the details of the biochemistry of lipid accumulation in microalgae could help researchers develop strategies for genetic manipulation of lipid synthesis pathways to affect not only the quantity but also the quality (chain length, degree of desaturation) of lipids produced for optimal biodiesel performance.

In one set of experiments, pulse-chase radiolabeling was used to study de novo synthesis of lipids in Nannochloropsis. Exponentially growing cells under low light were fed 14C-bicarbonate or acetate for 1 hour. The cells were then washed and allowed to grow in unlabeled medium. At various time points, cells were removed and lipids extracted. The substrates resulted in a different distribution of labeled carbon in the lipids and fatty acids. The work demonstrated the probable existence of two pools of malonyl-CoA used as substrates for fatty acid synthesis, and resulted in a new model for the sites of desaturation of fatty acids and the identification of a new acyltransferase activity in this organism. In another set of experiments, Nannochloropsis cells were mutagenized using UV light and screened for unusual fatty acid profiles using gas chromatography. This work resulted in the isolation of a mutant deficient in 20:5 fatty acids, probably due to a mutation affecting a desaturase enzyme that utilizes 20:4 fatty acids as substrate.

These experiments will not be described in detail here, primarily because the funding for this research did not come from DOE. In addition, it would require a lengthy discussion of the details of fatty acid synthesis and processing for the reader to understand the relevance of the findings. Readers interested in the details of this research are referred to the three publications that resulted from this research (Schneider and Roessler 1994; Schneider et al. 1995; Schneider and Roessler 1995).

Part II of this report details the specific research accomplishments of the program on a year-to-year basis. In order to provide a consistent context and framework for understanding this detail, we have attempted to outline the major activities of the program as they unfolded over the course of the past two decades. The timeline on the following page shows the major activities broken down into two main categories—laboratory studies and outdoor testing/systems analysis. For the sake of clarity and brevity, many of the research projects and findings from the program are not presented here. Instead, only those findings that form a thread throughout the work are highlighted. There were many other studies and findings of value in the program. The reader is encouraged to review Part II of this report for a more comprehensive discussion of the research.

The collection and screening effort resulted in a large number of strains that had many characteristics deemed important for a biodiesel production organism. In reviewing the many procedures used by ASP researchers, however, it is clear that a more consistent screening protocol might have yielded results that could be compared more meaningfully. Although these types of standard protocols were being developed near the end of the collection and screening effort, they were not consistently used. Furthermore, because an optimized microalgal-based biofuel production process was never fully developed, the screening protocols could not be based on an actual process. Therefore, whether the screening criteria used in the ASP were accurate predictors of good performance in a biodiesel production facility is not really known. For example, lipid productivity over a given amount of time is one of the most important factors in a production process. There were no clear guidelines as to whether lipid productivity in an outdoor pond was better in a continuous, steady-state process or in a multistage process involving substantial culture manipulation (e. g., nutrient level control, “ripening” tanks). This information is critical, and has a profound impact on the type of screening that should be conducted.

In addition to the need for a better understanding of the most economically feasible commercial biodiesel production process, additional information about the true constraints with regard to lipid properties (e. g., fatty acid chain length, degree of unsaturation, polar lipid constituents, etc.) there is a need for better information on the impact of lipid composition on fuel quality. The lipophilic dye Nile Red was used as a screening tool to rapidly assess the lipid contents of isolates, but in retrospect this technique probably does not provide the level of detail regarding lipid quality that may be necessary. Indeed, the variability in the ability of various strains to take up this dye is a major problem that must be recognized. With the rapid advances that have been made in recent years in automated high performance liquid chromatography and detection, this technique seems readily adaptable for use as a powerful screening tool.

For future screening endeavors, we recommend that an effort be made to naturally select strains at the locations that would likely be commercial microalgal production sites. In this manner, the algae would be exposed to the prevailing environmental conditions, particularly the indigenous waters. In small open-air vessels, the medium would be “inoculated” with a variety of indigenous strains, and a process of natural selection would occur such that the most competitive strains would dominate the cultures after a short while. Of course, the disadvantage of this method is that the dominant strains may not be good lipid producers. For this reason, genetically manipulating the dominant strains by classical or recombinant means may be necessary, such that they remain competitive and yet make acceptable amounts of lipid. Whether such manipulations can be made awaits further experimentation.

One thing that was clear from the collection and screening effort was that diatoms and green algae would most likely be well represented in a “natural selection” screen as described in the preceding paragraph. Therefore, future efforts should probably focus on developing sophisticated genetic engineering tools focused on these groups. Such tools could be transferable to many different species within these groups; such transfer would be facilitated by the fact that powerful methods for generating genetic sequence information are becoming routine.

SERI in-house algal strain collection and screening efforts during 1986-1987 were focused in three separate areas. First, detailed characterization of previously collected strains continued. Second, because the desert southwest sites targeted for biodiesel production facilities can be quite cool during the winter, a new effort to collect strains from cold-water sites was initiated. Finally, a strategy was developed and implemented to reduce the number of strains that had accumulated as a result of in-house and subcontracted research efforts, which allowed researchers to focus on the most promising strains.

Water samples isolated from various sites were screened for the presence of algal viruses using the plaque assay. If viruses were not detected initially, an enrichment protocol was used in which a few algal cells were added to the water sample; after 48-72 hours the algal and cell debris were removed by centrifugation and the sample was reassayed for the presence of virus. Using these procedures, more than 50 individual algal viral isolates were identified. Although the viruses were all large polygonal particles with dsDNA genomes, analysis of the viral DNA by digestion with restriction endonucleases showed there were at least 15 different types of virus found. Sites that contained virus were further analyzed for the natural algal hosts; however, none were identified. It is unclear why the natural hosts could not be found. Dr. Meints proposed either that the viruses were propagated or maintained by some unknown mechanism, or, more likely, that the natural host was present in the sites tested at very low concentrations. This is possible in that each virus that infects an algal cell could produce 350 new virus particles, and that up to 100 algal cells can exist within a single Paramecium. Based on the density of viral particles isolated from the various sites, a single Paramecium with algal symbionts could theoretically sustain a virus population in 350 liters of water. Because a natural alga host for the viruses was not found, the goal of characterizing the viral host was dropped, and further screening efforts were discontinued.

In a separate, but related, series of experiments, Dr. Meints received 250 water samples collected by Dr. Ralph Lewin of the Scripps Institute in La Jolla, California. These were also screened for the presence of algal viruses, but viral particles were not found in any of Dr. Lewin’s samples.

і Demonstration of Open Pond Systems for Mass Production of Microalgae.

Over the course of the program, efforts were made to establish the feasibility of large-scale algae production in open ponds. In studies conducted in California, Hawaii and New Mexico, the ASP proved the concept of long term, reliable production of algae. California and Hawaii served as early test bed sites. Based on results from six years of tests run in parallel in California and Hawaii, 1,000 m2 pond systems were built and tested in Roswell, New Mexico. The Roswell, New Mexico tests proved that outdoor ponds could be run with extremely high efficiency of CO2 utilization. Careful control of pH and other physical conditions for introducing CO2 into the ponds allowed greater than 90% utilization of injected CO2. The Roswell test site successfully completed a full year of operation with reasonable control of the algal species grown. Single day productivities reported over the course of one year were as high as 50 grams of algae per square meter per day, a long-term target for the program. Attempts to achieve consistently high productivities were hampered by low temperature conditions encountered at the site. The desert conditions of New Mexico provided ample sunlight, but temperatures regularly reached low levels (especially at night). If such locations are to be used in the future, some form of temperature control with enclosure of the ponds may well be required.

і The high cost of algae production remains an obstacle.

The cost analyses for large-scale microalgae production evolved from rather superficial analyses in the 1970s to the much more detailed and sophisticated studies conducted during the 1980s. A major conclusion from these analyses is that there is little prospect for any alternatives to the open pond designs, given the low cost requirements associated with fuel production. The factors that most influence cost are biological, and not engineering-related. These analyses point to the need for highly productive organisms capable of near-theoretical levels of conversion of sunlight to biomass. Even with aggressive assumptions about biological productivity, we project costs for biodiesel which are two times higher than current petroleum diesel fuel costs.

The goal of this subcontract was to isolate and characterize strains of microalgae from the southeastern United States that have attributes desirable for a biodiesel production strain. During the first year of this work, field trips were made to several sites in Alabama to collect microalgal strains from a variety of habitats. Freshwater and brackish water strains were collected from rivers, lakes, estuaries, and ponds, and marine strains were collected from the waters surrounding Dauphin Island in the Gulf of Mexico. Collected samples were inoculated into various artificial media, including Bold’s Basal Medium, Chu no. 10, and “f/2” (Barclay et al. 1986). Artificial sea salts were used in place of seawater for the saltwater media. For initial strain selection, the cultures were incubated at 29-30°C with shaking at a light intensity of 100 to 125 qE^m-2^s-1 provided by cool white fluorescent bulbs with a 14 h:10 h light:dark cycle. The fastest growing strains were isolated via micropipetting or by spreading samples on agar plates. In these preliminary experiments, the marine strains exhibiting the fastest growth were Cyclotella DI-35 (CYCLO1), Hantzschia DI-160 (NITZS2), and Chlorococcum DI-34. The freshwater strains exhibiting the fastest growth rates were Chlorella MB-31, Scenedesmus TR-84, Ankistrodesmus TR-87, and Nitzschia TR-114.

CYCLO1, Nitzschia TR-114, and Scenedesmus TR-84 were selected for more detailed growth analyses under various combinations of temperature, salinity, and light intensity. A temperature gradient table was employed for these experiments that was similar in design to the tables used by Dr. William Thomas (discussed earlier) and SERI researchers for screening purposes. Growth of standing cultures was determined by measuring final cell densities after 12 days of incubation. CYCLO1 achieved maximum cell density at a temperature of 30°C, a light intensity of 100 qE^m-2^s-1, and a salinity of 15 ppt (parts per thousand). Growth was nearly as good at a light intensity of 200 qE^m-2^s-1 and a temperature of 35°C, and substantial growth occurred at a salinity of 32 ppt. Growth did not occur at 15° to 20°C. Nitzschia TR-114 achieved maximal cell density at 30°C, 15 ppt salinity, and 100 qE^m-2^s-1; growth was severely inhibited at 0 and 45 ppt salinity. Growth was similar at 100 and 200 qE^m-2^s-1 for this strain, although the higher light intensity seemed to increase the thermal tolerance of the cells. The freshwater strain Scenedesmus TR-84 grew best at 25°C, and grew increasingly slower as the salt concentration of the medium was increased.

The lipid contents of several strains isolated during the initial collecting trips were determined. For 14-day-old cultures that were reportedly N-limited (although no evidence is provided to support this), the lipid contents (as a percentage of the organic mass) were as follows: CYCLO1, 42.1%; Nitzschia TR-114, 28.1%; Chlorella MB-31, 28.6%-32.4%; Scenedesmus TR-84, 44.7%; Ankistrodesmus TR-87, 28.1%; and Hantzschia DI-160 (NITZS2), 66%.

Additional strains were collected the next year from intertidal waters near Biloxi, Mississippi and St. Joseph Bay, Florida. Preliminary screening experiments indicated that five strains (all of which were diatoms) had the best growth rates and lipid accumulation potential: Navicula acceptata (two strains, NAVIC6 and NAVIC8), N. saprophila (NAVIC7), Nitzschia dissipata (NITZS13), and Amphiprora hyalina (ENTOM3). These strains and CYCLO1 were grown semi­continuously in media with six different salinities at 25°, 30°, and 35°C. Cells were grown at light intensities of 80 qE^m-2^s-1 and 160 qE^m-2^s-1 (approximately 4% and 8% of full sunlight, respectively). The media were produced by adding various quantities of artificial sea salts to “f/2” medium; the resulting conductivities were <1, 10, 20, 35, 45, and 60 mmho^cm-1. (Note: seawater is typically 35-45 mmho*cm-1.) All strains exhibited more rapid growth under 160 qE^m-2^s-1 illumination than at 80 qE^m-2^s-1; and even higher growth rates might well have been obtained at light intensities greater than 160 qE^m-2^s-1.

ENTOM3 grew best (2.0 to 2.3 doublings^-1) at 30°C in media with conductivities of 20-60 mmho’cm’1. Growth was better with urea or nitrate as the N source rather than with ammonium. The lipid content of nutrient-sufficient cells was 22.1% of the organic mass, and increased to 37.1% and 30.2% under Si-deficient and N-deficient conditions, respectively.

CYCLO1 achieved the highest growth rates (2.8 to 3.0 doublings^-1) at 35°C between 10 and 35 mmho-cm-1. Cells grew best with nitrate as a N source, followed by ammonium and then urea. The highest lipid content was observed in N-deficient cells (42.1%), but was also elevated in Si — deficient cells (38.6%) relative to nutrient-sufficient cells (13.2%).

NAVIC8 grew most rapidly (3.8 doublings^"1) at 35°C and 45 mmho^cm"1. Nitrate and ammonium were more suitable N sources than urea. Lipid contents of 21.8%, 48.5%, and 32.4% were observed for cells grown under nutrient-sufficient, Si-deficient, and N-deficient conditions, respectively.

During the final year of this subcontract, additional promising strains were isolated. Included in this group was Navicula BB-324 (NAVIC9), which had a growth rate exceeding 2.5 doublings^-1 at 30°C in artificial seawater and SERI Types I/10, I/25, I/40, II/10, II/25, II/40, and II/55 media. Navicula SB-304 (NAVIC8) also exhibited excellent growth (1.5-3.0 doublings^-1) in each medium. These two strains had Si starvation-induced lipid contents of 42.5% and 47.2%, respectively. Other notable strains were Nitzschia SB-307 (NITZS13), which had a maximal growth rate of 2.5 doublings^-1 and a lipid content of 45%-47% under nutrient — stressed conditions. Amphiprora BB-333 (ENTOM3), Chaetoceros BB-330 (CHAET66), and Cylindrotheca AB-204 also grew rapidly (2.3-6.0 doublings^-1), with stress-induced lipid contents ranging from 16.5%-37.1%.

In conclusion, many promising strains were isolated as a result of this subcontract. The nutrient status of the cells again was played an important role in lipid accumulation. Furthermore, the nature of the N source included in the medium had a substantial impact on growth of the cultures. Several of these strains were further tested in outdoor mass culture, as described in Section III.

Another strategy that has been proposed to increase the proportion of lipid in algal cells is to limit the flow of newly assimilated carbon into other cellular pathways. Many diatoms, including

C. cryptica, can produce a significant amount of a storage carbohydrate called chrysolaminarin, a P-(1 ^3)-linked glucan. Although some data were available on the chemistry of this compound, the biochemical pathways involved in the synthesis of chrysolaminarin were not known. The synthesis of most storage polysaccharides involves the condensation of nucleoside diphosphate sugars; for example, starch is formed in plants from ADPglucose, and UDPglucose is used to form sucrose in plants and glycogen in mammalian cells. These reactions are catalyzed by nucleoside diphosphate sugar pyrophosphorylases, such as UDPglucose pyrophosphorylase (UGPase), which catalyzes the following reaction:

glucose-1-phosphate + UTP ^ UDPglucose + PPi

Roessler first looked for nucleoside diphosphate sugars pyrophosphorylases in cell-free extracts of C. cryptica, and identified significant amounts of UGPase activity. The enzyme activity was characterized to optimize in vitro assay conditions. The enzyme was activated in the presence of Mn2+ and Mg2+ but was not affected by 3-phosphoglycerate or inorganic phosphate; these chemicals are known to affect the activity of ADPglucose pyrophosphorylase in higher plants. Incubation of cell-free extracts with UDP [14C]glucose resulted in the incorporation of the labeled carbon into a P-(1^3)-glucan polymer, presumably chrysolaminarin, supporting the role of UGPase in chrysolaminarin synthesis in diatoms. Subsequent studies identified a second enzyme, UDPglucose:P-(1^3)-glucan-p-glucosyltransferase (also known as chrysolaminarin synthase), which catalyzes the synthesis of glucan using UDPglucose as substrate. The specific activity of both enzymes was examined in C. cryptica cells under Si-replete and Si-depleted conditions. The activity of UDPglucose pyrophosphorylase was similar under both conditions; however, the activity of chrysolaminarin synthase decreased by 31% in Si-deficient cells, suggesting that the partitioning of newly assimilated carbon into lipid may be partly due to decreased synthesis or inhibition of the chrysolaminarin synthase enzyme (Roessler 1987; 1988a).

Further research on UGPase in C. cryptica was put on hiatus for several years while the emphasis was on ACCase (discussed earlier) and on the development of genetic engineering protocols for microalgae (discussed in Section II. B.3.). However, the development of a successful genetic transformation system for C. cryptica, as well as advances in techniques that allow the down — regulation of particular genes (i. e., antisense RNA, ribozymes) generated a renewed interest in UGPase. NREL researcher Eric Jarvis spent 6 months working at Ribozyme Pharmaceuticals, Inc., a biotechnology company in Boulder, Colorado, learning about these new methods. Antisense RNA is a method in which a cell is transformed with a synthetic gene that produces an RNA sequence complimentary to a specific messenger RNA (mRNA). Although the exact mechanism is not clear, the antisense RNA prevents translation from its complimentary mRNA, effectively lowering the level of that particular protein in the cell. Ribozymes are also RNA molecules produced by synthetic genes that can bind to, and cleave, very specific RNA sequences. Ribozymes can be designed to degrade specific mRNA molecules, effectively decreasing expression of a specific gene.

In C. cryptica, chrysolaminarin can make up 20%-30% of the cell dry weight, and thus chrysolaminarin synthesis pathways presumably compete for newly fixed carbon with the pathways for lipid biosynthesis. Dr. Jarvis and Dr. Roessler proposed that inhibiting chrysolaminarin production by inhibiting one or more genes in the carbohydrate synthesis pathway could result in the flow of more carbon into lipid production. Based on the earlier studies on chrysolaminarin synthesis, Dr. Jarvis initiated an effort to isolate the UGPase gene from C. cryptica DNA. A fragment of the C. cryptica UGPase gene was first produced by the PCR using degenerate oligonucleotide primers based on conserved sequences from known UGPase genes from potato, human, yeast, and Dictyostelium. This fragment was cloned and sequenced; the derived amino acid sequence showed 37% identity with the corresponding sequence from potato UGPase, confirming that a C. cryptica UGPase gene fragment had been cloned. The cloned PCR product was then used as a probe to isolate a genomic DNA clone containing the entire C. cryptica UGPase gene from a lambda library. One clone contained a DNA segment with a single long open reading frame, the 5′ end of which showed homology to known UGPase genes. Surprisingly, the 3′ end of this DNA showed homology to known genes coding for the enzyme phosphoglucomutase (PGMase). In chrysolaminarin synthesis, PGMase catalyzes the following reaction:

glucose-6-phosphate ^ glucose-1-phosphate

The glucose-1-P produced in this reaction is the substrate for UGPase in the production of UDPglucose, an immediate precursor of chrysolaminarin, as described earlier. Although PGMase and UGPase are thought to catalyze successive steps in the chrysolaminarin biosynthesis pathway, this was the first report of a naturally occurring fusion of these two genes in any organism. The C. cryptica UGPase/PGMase gene, designated uppl, contained 3,640 bps, including 3 introns, and coded for a protein composed of 1,056 amino acids, with a molecular weight of 114.4 kd.

To confirm that the protein coded for by uppl actually catalyzes both the UGPase and PGMase reactions, the protein was isolated from extracts of C. cryptica by sequential column chromatography (ion exchange, hydroxylapatite, and gel filtration). The two enzyme activities co-eluted throughout the purification procedure, and all fractions containing UGPase/PGMase activity contained a 114 kd protein as determined by SDS-PAGE. These results supported the presence of both enzyme activities in C. cryptica on a single multifunctional protein. A patent submitted by NREL on this unique gene was allowed in October 1996. The research at NREL involving attempts to manipulate uppl gene expression to affect carbon partitioning in C. cryptica will be discussed in Section II. B.3. of this report.

The research pathway in the lab can be broken down into three types of activities:

• Collection, screening and characterization of algae.

• Biochemical and physiological studies of lipid production

• Molecular biology and genetic engineering studies

There is a logic to the sequence of these activities. Researchers first identified a need to collect and identify algae that met minimal requirements for this technology. Collection and screening occurred over a seven-year period from 1980 to 1987. Once a substantial amount of information was available on the types of oil-producing algae and their capabilities, the program began to switch its emphasis to understanding the biochemistry and physiology of oil production in algae. A natural next step was to use this information to identify approaches to genetically manipulate the metabolism of algae to enhance oil production.

Algae collection efforts initially focused on shallow, inland saline habitats, particularly in western Colorado, New Mexico and Utah. The reasoning behind collecting strains from these habitats was that the strains would be adapted to at least some of the environmental conditions expected in mass culture facilities located in the southwestern U. S. (a region identified early on as a target for deployment of the technology). Organisms isolated from shallow habitats were also expected to be more tolerant to wide swings in temperature and salinity. In the meantime, subcontractors were collecting organisms from the southeastern region of the U. S. (Florida, Mississippi, and Alabama). By 1984, researchers in the program had developed improved tools and techniques for collecting and screening organisms. These included a modified rotary screening apparatus and statistically designed saline media formulations that mimicked typical brackish water conditions in the southwest. In 1985, a rapid screening test was in place for identifying high oil-producing algae. In the last years of the collection effort, the focus switched to finding algae that were tolerant to low temperature. This expanded the reach of the collection activities into the northwest. By 1987, the algae collection contained over 3,000 species.

As the collection efforts began to wind down, it became apparent that no one single species was going to be found that met all of the needs of the technology. As a result, about midway through the collection efforts, the program began studies on the biochemistry and physiology of oil production in algae in hopes of learning how to improve the performance of existing organisms. A number of ASP subcontractors struggled to identify the so-called “lipid trigger.” These studies confirmed observations that deficiencies in nitrogen could lead to an increase in the level of oil present in many species of algae. Observations of cellular structure also supported the notion of a trigger that caused rapid build up of oil droplets in the cells during periods of nitrogen depletion.

Pre-1980	1980		1994 1995 1996
Lab Studies
By 1987, over 3,000 strains of algae had been collected.
Isolation and characterization of ACCase enzyme
Genetic Engineering of Algae
expression of foreign gene in algae using protoplasts
1st successful genetic transformation of diatom
Outdoor Culture Studies and Systems Analysis
Algae Production in Wastewater Treatment
<100 sq. m. Pond Studies (CA, HI)
1000 sq. m. Pond Study (NM)
I Systems Analysis and Resource Assessment

In the end, however, the studies conducted both by NREL researchers and program subcontractors concluded that no simple trigger for lipid production exists. Instead, we found that environmental stresses like nitrogen depletion lead to inhibition of cell division, without immediately slowing down oil production. It appeared that no simple means existed for increasing oil production, without a penalty in overall productivity due to a slowing down of cell growth. The use of nutrient depletion as a means of inducing oil production may still have merit. Some experiments conducted at NREL suggested that the kinetics of cell growth and lipid accumulation are very subtle. With a better understanding of these kinetics, it may be possible to provide a net increase in total oil productivity by carefully controlling the timing of nutrient depletion and cell harvesting.

In 1986, researchers at NREL reported on the use of Si depletion as a way to increase oil levels in diatoms. They found that when Si was used up, cell division slowed down since Si is a component of the diatoms’ cell walls. In the diatom C. cryptica, the rate of oil production remained constant once Si depletion occurred, while growth rate of the cells dropped. Further studies identified two factors that seemed to be at play in this species:

1. Si-depleted cells direct newly assimilated carbon more toward lipid production and less toward carbohydrate production.

2. Si-depleted cells slowly convert non-lipid cell components to lipids.

Diatoms store carbon in lipid form or in carbohydrate form. The results of these experiments suggested that it might be possible to alter which route the cells used for storage (see schematic below):

Through the process of photosynthesis, algae cells assimilate carbon. There are numerous metabolic pathways through which the carbon can go, resulting in synthesis of whatever compounds are needed by the cell. These pathways consist of sequences of enzymes, each of which catalyzes a specific reaction. Two possible pathways for carbon are shown on the previous page. They represent the two storage forms that carbon can take.

Researchers at NREL began to look for key enzymes in the lipid synthesis pathway. These would be enzymes whose level of activity in the cell influences the rate at which oils are formed. Think of these enzymes as valves or spigots controlling the flow of carbon down the pathway. Higher enzyme activity leads to higher rates of oil production. When algae cells increase the activity of active enzymes, they are opening up the spigot to allow greater flow of carbon to oil production. Finding such critical enzymes was key to understanding the mechanisms for controlling oil production.

By 1988, researchers had shown that increases in the levels of the enzyme Acetyl CoA Carboxylase (ACCase) correlated well with lipid accumulation during Si depletion. They also showed that the increased levels correlated with increased expression of the gene encoding for this enzyme. These findings led to a focus on isolating the enzyme and cloning the gene responsible for its expression. By the end of the program, not only had researchers successfully cloned the ACCase gene, but they had also succeeded in developing the tools for expressing foreign genes in diatoms.

In the 1990s, genetic engineering had become the main focus of the program. While we have highlighted the successes of over-expressing ACCase in diatoms, other approaches were also developed for foreign gene expression—in green algae as well as in diatoms. Another interesting sideline in the research involved studies aimed at identifying key enzymes involved in the synthesis of storage carbohydrates. Instead of over-expressing these enzymes, researchers hoped to inactivate them. Returning to our “spigot” analogy, this approach was like shutting off the flow of carbon to carbohydrates, in the hopes that it would force carbon to flow down the lipid synthesis pathway (again, see the schematic on the previous page). This work led to the discovery of a unique multifunctional enzyme in the carbohydrate synthesis pathway. This enzyme and its gene were both patented by NREL in 1996.

Outdoor Testing and Systems Analysis

The first work done in earnest by DOE on demonstration of algae technology for energy production predates the Aquatic Species Program. In 1976, the Energy Research and Development Administration (before it was folded into DOE) funded a project at the University of California Berkeley’s Richmond Field Station to evaluate a combined wastewater treatment/fuel production system based on microalgae. Over the course of several years, the Richmond Field Station demonstrated techniques for algae harvesting and for control of species growing in open ponds.

By the time the Aquatic Species Program took on microalgae research, emphasis had already moved from wastewater treatment based systems to dedicated algae farm operations. From 1980 to 1987, the program funded two parallel efforts to develop large scale mass culture systems for microalgae. One effort was at the University of California, and it was based on a so-called “High Rate Pond” (HRP) design. The other effort was carried out at the University of Hawaii, where a patented “Algae

Raceway Production System” (ARPS). Both designs utilized open raceway designs. The HRP design was based on a shallow, mixed raceway concept developed at Berkeley in 1963 and successfully applied in wastewater treatment operations in California. The ARPS was really a variation on the same concept. Both efforts carried out their test work in ponds of 100 square meters or less. They studied a variety of fundamental operational issues, such as the effects of fluid flow patterns, light intensity, dissolved oxygen levels, pH and algae harvesting methods.

At the conclusion of the smaller scale tests conducted in California and Hawaii, the program engaged in a competitive bidding process to select a system design for scale up of algae mass culture. The HRP design evaluated at UC Berkeley was selected for scale-up. The “Outdoor Test Facility” (OTF) was designed and built at the site of an abandoned water treatment plant in Roswell, New Mexico. From 1988 to 1990, 1,000 square meter ponds were successfully operated at Roswell. This project demonstrated how to achieve very efficient (>90%) utilization of CO2 in large ponds. The best results were obtained using native species of algae that naturally took over in the ponds (as opposed to using laboratory cultures). The OTF also demonstrated production of high levels of oil in algae using both nitrogen and silica depletion strategies. While daily productivities did reach program target levels of 50 grams per square per day, overall productivity was much lower (around 10 grams per square meter per day) due to the number of cold temperature days encountered at this site. Nevertheless, the project established the proof-of-concept for large scale open pond operations. The facility was shut down in 1990, and has not been operated since.

A variety of other outdoor projects were funded over the course of the program, including a three-year project on algal biodiesel production conducted in Israel. In addition, research at the Georgia Institute of Technology was carried out in the late 1980s. This work consisted of a combination of experimental and computer modeling work. This project resulted in the development of the APM (Algal Pond Model), a computer modeling tool for predicting performance of outdoor pond systems.

Two types of systems analysis were conducted frequently over the course of the program—resource assessments and engineering design/cost analyses. The former addresses the following important question: how much impact can algae technology have on petroleum use within the limits of available resources? Engineering designs provide some input to this question as well, since such designs tell us something about the resource demands of the technology. These designs also tell us how much the technology will cost.

As early as 1982, the program began to study the question of resource availability for algae technology. Initial studies scoped out criteria and methodology that should be used in the assessment. In 1985, a study done for Argonne National Lab produced maps of the southwestern U. S. which showed suitable zones for algae production based on climate, land and water availability. In 1990, estimates of available CO2 supplies were completed for the first time. These estimates suggested that that there was enough waste CO2 available in the states where climate conditions were suitable to support 2 to 7 quads of fuel production annually. The cost of the CO2 was estimated to range anywhere from $9 to $90 per ton of CO2. This study did not consider any regionally specific data, but drew its conclusions from overall data on CO2 availability across a broad region. Also in 1990, a study was funded to assess land and water availability for algae technology in New Mexico. This study took a more regionally specific look at the resource question, but did so by sacrificing any

consideration of available CO2 supplies. This last study sums up the difficulties faced in these types of studies. The results obtained on resource availability are either able to provide a complete, but general, perspective on resources or they are more detailed in approach, but incomplete in the analysis of all resources.

Engineering design and cost studies have been done throughout the course of the ASP, with ever increasing realism in the design assumptions and cost estimates. The last set of cost estimates for the program was developed in 1995. These estimates showed that algal biodiesel cost would range from $1.40 to $4.40 per gallon based on current and long-term projections for the performance of the technology. Even with assumptions of $50 per ton of CO2 as a carbon credit, the cost of biodiesel never competes with the projected cost of petroleum diesel.