The APM (Benemann and Tillett 1987, 1990; Tillett 1989) describes a shallow, paddle wheel — mixed pond system. It incorporated climatic, design, operational, physicochemical, and biological parameters and submodels to predict the pond, and algal culture, behavior. Simplying assumptions in the model include the following:

• the pond exhibited no nutrient, biomass or temperature inhomogeneities (e. g., gradients),

• mixing was essentially plug flow for large systems and mixed tank behavoir for smaller ones,

• CO2 and O2 outgassing can be estimated from measured gas transfer coefficients, and

• evaporation was 1.5 times pan A evaporation data, for example.

Climate input parameters were obtained from local U. S. climate stations in machine-readable form, including daily data on air temperatures (diurnal), relative humidity, rainfall, wind speed and total solar radiation. These data were then averaged over several years to provide daily and monthly average data sets. The use of such climate data is critical for predicting pond conditions at a particular site and location. Design variables included pond area, depth, mixing velocity, and length-to-width ratios. Physical inputs were the alkalinity of the media, the starting and ending pH for the carbonation station, and the outgassing coefficient. Finally, for a biological parameter a productivity assumption (from 20-40 g/m2/d, or a % maximum sunlight conversion efficiency) was used, along with C content, O2 yield, heat of combustion, and saturating light intensity for photosynthesis (assuming simple light saturation, and application of the Bush equation).

The model included an energy balance (input sunlight, radiation, evaporation, air temperatures, etc.), which continuously predicted the pond temperature, based on pond depth, and ambient conditions. From the productivity assumption (on a diurnal basis) and a CO2 outgassing coefficient for the ponds, the total inorganic C balance can be calculated based on alkalinity and pH as the major determinants of inorganic C in ponds. Like CO2 (and pH, etc.), O2 is also dependent on productivity, and outgassing. The model was written in Fortran (Tillett 1989).

The model was validated at the Roswell test site, with the 3-m2 ponds, which were instrumented (for pH, DO, wind, air temperature, etc.) and a data acquisition system developed to obtain short — (< 1h) and long — (>1 d) term data. The model was also validated with the larger ponds. Measured and predicted pond temperatures agreed well, as seen in Figure III. B. 10a for a diurnal data set for a heated and unheated small pond, and Figure III. B. 10b for a single unheated pond for 2 weeks.

Simulations were also run for a larger, 1,000-m2 earthen pond as built at Roswell (an arid and cool climate), which were then compared with a site in West Palm Beach, Florida (a humid,
warm climate), using monthly average climate data. The model was exercised for 4 representative months of the year for both locations, and assuming ponds of 10, 20, 30 and 100­cm depth. (This last point was to demonstrate the limited effect of managing pond temperature extremes by depth.) Minimum water temperatures do not rise above 10°C for shallow ponds for most of the year in Roswell; they never drop below this level in West Palm Beach. Maximum summer temperatures seem to be only modestly higher in Florida than in New Mexico. These results point to low temperatures as a major factor in Roswell operations.

Only a limited attempt was made to verify the model in relation to productivity, by using a fitted Ik (saturating light intensity) parameter, as well as an assumed heat of combustion (5.7 Kcal/g) and biochemical conversion efficiency (from prior work by Weissman and Goebel 1987). Although the agreement between calculated and measured productivities was excellent (both gave about 15.3 g/m2/d), this was probably fortuitous, as the Ik actually used in the model was well below what had been previously measured with the same organism (Monoraphidium sp.). This requires further investigation.

One interesting use of the model was to predict CO2 utilization and outgassing from various assumed pH, pCO2 and alkalinity regimes. This is a central issue in the operation of algal pond systems, as these parameters must be used to optimize for productivity and overall CO2 utilization efficiencies. The higher the pH and the lower the alkalinity, the greater the utilization efficiency. However, for CO2 supply, pH must be decreased transiently. With moderate alkalinity (2.5-5 mM), and CO2 requirements, the pH transients can be relatively small, allowing minimal outgassing even for seawater systems, resulting in predicted CO2 use efficiencies of over 85%.

A final important parameter is dissolved O2, which was predicted to accumulate within 1 hour to a level of 210% of air saturation with a 30-cm deep culture and a productivity of 30 g/m2/d. Much higher concentrations would build up during the day, and are actually observed in ponds. This could be a major factor in reduced productivities in ponds (see Section III. B.5.).

This model is of sufficient detail and predictive value to minimally direct the laboratory and small-scale outdoor research in making these more representative of the outdoor pond environments. Laboratory applications are discussed in the following section.

Figure III. B.10. Comparison of measures and predicted pond temperatures in Roswell, New Mexico.

a. ). Top: Comparison of measured and predicted diurnal temperature profiles for heated and unheated ponds. October 4, 1987.

b. ). Bottom: A comparison of predicted (line) versus measured (symbols) maximum-minimum temperatures in miniponds using simplified inputs for Roswell, N. M.

This 2-year project (Benemann et al. 1979; Eisenberg et al. 1980) extended from the end of 1977 to Winter 1979. Initially the approach to establish microstrainable cultures using the 12-m2 ponds, described in Section III. A.3., continued to be investigated. Essentially the same results as before were obtained: detention time was found to be the key environmental variable determining algal colony size (but not necessarily species composition) and a negative correlation was found between numbers of algal grazers and the large colonial algal types easy to harvest with microstrainers. Apparently the grazers preferentially consumed the smaller algae. Overall, the harvestability results with the microstrainers continued to be poor, so this line of research was abandoned during the initial period of this project.

Simultaneously with these studies, another project was being carried out under EPA sponsorship to study the settling of algae in the City of Woodland waste treatment ponds. This project used a “phase isolation” process, in which the algal cells were allowed to spontaneously settle when sewage inflow was stopped (Koopman et al. 1978, 1980). Although generally long times were required for this settling process (2-3 weeks), it was decided to investigate this general phenomenon of “bioflocculation” in high rate ponds. The process involved removing the algae from the paddle wheel-mixed ponds and placing them in a quiescent container, where they would spontaneously flocculate and rapidly settle. There are several apparently distinct mechanisms by which algae flocculate and then settle, including “autoflocculation”, which is induced by high pH in the presence of phosphate and divalent cations (Mg2+ and Ca2+), and flocculation induced by N limitation. Bioflocculation refers to the tendency of normally repulsive microalgae to aggregate in large flocs, that then exhibit a rather high sedimentation velocity. The mechanisms of bioflocculation involve extracellular polymers excreted by the algae, but the details remain to be investigated.

Settling tests were carried out with the cultures from the 12-m2 ponds. As with microstrainer harvesting, detention time and mixing velocity were the most important variables in promoting a bioflocculating culture. The rather rapid settling of many of the cultures was very encouraging. Also, the initial experiments with the 0.25-ha pond demonstrated a fairly rapid (<24 h) bioflocculation process.

At this time (mid 1978), the 0.25-ha pond had been divided into two 0.1-ha high rate ponds mixed with paddle wheels, and the bioflocculation-settling process using these pilot ponds became the focus of further research. Two 32,000 L settling ponds, with concave bottoms as deep as 2.5 m, were constructed to test bioflocculation-settling with the effluents of these ponds at the pilot scale. In initial experiments the two ponds were operated in parallel with two smaller ponds, at similar dilution rates (and with the only difference that one set of ponds was screened to remove larger grazers). The results (summarized in T able III. A.1.) were reasonably reproducible between the ponds. Algal solids removal through bioflocculation, as measured by 24-h settling in an “Imhoff Cone,” was high, about 90%.

Additional experiments, with the ponds operated in parallel at 2.5- and 5-day retention times (during September 1978) again resulted in removal efficiencies of over 90% for both ponds, including the 32,000 L settling tanks, based on measurements of suspended solids and chlorophyll. However, the shorter retention-time pond had more than twice the productivity of the longer-retention time pond (15 versus 7 g/m2/d). Bioflocculation was established as the method of choice for algal harvesting, as it seemed to be achievable even with high productivity cultures. Culture settleability was routinely determined during all the experiments with the high rate ponds. Table III. A.2. summarizes productivity and settleability for more than 1 year for the two 0.1-ha ponds. The 24-h laboratory settling tests correlated well with settling in the large (32­m3) settling ponds. Overall, a settling efficiency of greater than 85% was achieved, if the better of the two ponds was selected, without compromise in productivity. Annual gross productivity averaged almost 20 g/m2/d, estimated at about 90% algal solids. In many cases bioflocculation removals were higher than 95% on a chlorophyll a basis, indicating almost complete sedimentation of the algal biomass.

One factor that had a major effect on bioflocculation was mixing speed. At very low mixing velocities, the cultures settled very poorly. This is not unexpected, as any settling algae would have dropped out of the photic zone and been replaced by suspended cells in the absence of continuous mixing. However, a more fundamental investigation of the bioflocculation phenomenon (well reported in the ecological and some physiological literature) remains to be carried out.

Another activity carried out during this project was the cultivation of N-fixing microalgae using the supernatant of secondary growth ponds (see Figure III. A.3.). One problem was that culture collection strains of N-fixing cyanobacteria inoculated into the N-deficient ponds quickly succumbed. However, when the N-deficient wastewater samples were incubated in sunlight for a few weeks, N-fixing strains, which were indeed culturable, appeared spontaneously. One strain, an Anabaenopsis sp., was successfully cultured over long periods. However, the productivity of N-fixing cultures was only about half of what was observed with green algae under similar conditions. That is understandable because of the high energy requirement of N-fixation, which is not desirable in biomass fuel production. However, use of such algae for phosphate removal from wastewaters is of interest.

This project also investigated the anaerobic digestion of algal biomass using both a large (1,400- L) and several smaller (20-L) digesters. Algae provided a good substrate for anaerobic digestion, although not as good as conventional sewage sludge based on the conversion of organic C to methane. However, attempts to regrow algae on the effluent of the digesters resulted in poor productivities, as some unidentified factor appeared to limit growth on the regenerated nutrients. This was not investigated further, but clearly requires some attention in the future.

As part of this project, the energy requirements for high-rate pond operations, in particular mixing, were studied. The shaft power input required for mixing followed closely the predicted cube power law, demonstrating the need to keep mixing velocities below 30 cm/s to avoid excessive power inputs. At 15 cm/s, power inputs were only about 1 kWh/d (for the 0.1-ha

pond), increasing to about 10 kWh/d at 30 cm/s. Motor efficiency at the higher velocity (67%) was twice that at the lowest velocities, not a deciding factor. Clearly, mixing speeds higher than 30 cm/s are impractical, at least for energy production systems.

An energy analysis of such a microalgae wastewater treatment process was quite promising, with the wastewater system being at least energy self-sufficient based on the requirements for sewage and algal pond effluent lift requirements, as well as algal pond mixing and other requirements. Such systems should be able to generate a net positive energy output if operated to maximize algal biomass through the addition of CO2 (Figure III. A.3.).

Overall this project marked a major advance in this technology, as it demonstrated at the pilot plant scale a relatively low-cost and reasonably reliable microalgae harvesting process that does not unduly interfere with other objectives, in particular microalgae productivity. (See Benemann et al. [1980] and Eisenberg et al. [1980].)

A relatively detailed analysis of an algal wastewater treatment-energy production process was carried out by Benemann et al. (1977) as part of a larger study that examined a system integrating wastewater algal ponds with tree biomass production. The so-called “Photosynthetic Energy Factory” (InterTechnology Solar Corporation 1978) was to use the effluents of a waste treatment pond system to fertilize short-rotation trees for fuel farming. In turn, the power plant burning the woody biomass would provide CO2 for the algal ponds.

A design of the algal pond subsystem was carried out by Benemann et al. (1978) for a typical municipal community of 50,000 people, generating approximately 18,000 m3 of municipal wastewater per day. The assumption was that algal biomass would be grown up to the N growth potential of the wastewater, containing 65 mg/L of useable N (as organic N and ammonia). This required recycling about 5 to 7 tons of CO2 per day from the power plant to the algal ponds. A temperate site with an average insolation of about 15 GJ/m2/d was assumed, with a solar conversion efficiency averaging only 2.7% of visible light (about 1.35% of total solar), somewhat higher in winter than summer. This is considerably lower than current assumptions.

This study, for the first time, took into consideration monthly variations in temperature, insolation and other parameters. Algal harvesting was assumed to be with microstrainers (this analysis was carried out while this option was still being investigated, see Section III. A.3.). This report also carried out the first, though preliminary, analysis of the mixing power required for such large algal ponds and of the transfer requirements for CO2 to the algal culture. A 160-ha algal pond system was required to treat this wastewater flow year-round. This was about three times larger than a conventional oxidation pond system. Costs were projected to be competitive with conventional wastewater treatment systems.

Energy outputs were twice the energy inputs, based on digester gas production and requirements for pumping the wastewater, mixing the ponds, etc. The overall economics were very favorable because of the wastewater treatment credits.

Although this concept appeared favorable, in practice the relatively small scale of the locally available municipal wastes could supply only a small fraction of fertilizer needs for the very large (> 10,000 ha) energy plantations being projected. It does, however, point to the potential of this technology in wastewater treatment.

I Publications:

Benemann, J. R.; Koopman, B. L.; Baker, D.; Goebel, R.; Oswald, W. J. (1977) “Preliminary design of the algae pond subsystem of the photosynthesis energy factory.” Final Report to Inter-Technology Solar Corp., Sanitary Eng. Res. Lab., Univ. of Calif.-Berkeley.

InterTechnology/Solar Corp. (1978) “The photosynthesis energy factory: analysis, synthesis and demonstration.” U. S. DOE HCP/T3548-01.

This project evolved from one that focused on a demonstration of the ARPS concept using a single flume, to the investigation of fundamental issues in algal mass culture, using several smaller ponds and a simplified system design. In particular, this project reported very high productivities achieved by two methods: organized mixing in ponds (e. g., the foils), and optimal batch dilution (2- or 3-day intervals, depending on species). However, the basis for these productivity enhancements was speculative, and it proved difficult to demonstrate the reproducibility of these effects. The effects of foils could be better ascribed to degassing of oxygen from the ponds with foils (e. g., higher mixing power inputs) and the results from the 3- day dilution experiments to some uncontrolled factors, in addition to possible methodological problems (Laws et al. 1985; 1986a, b; 1987).

Laws (1989; see also Laws and Berning 1990) continued this research with Electric Power Research Institute funding for 1 year, moving the system to Kona, Hawaii. No significantly different information was produced. However, Laws concluded that lack of land area, and high costs, would make such a process impractical for fuel production in Hawaii.

The problem of light saturation has been a subject of research in photosynthesis for almost 5 decades, with the report by Kok (1953) that microalgae cultures exposed to short (milliseconds) flashes of bright light, followed by longer periods of darkness, exhibited the same light conversion efficiencies as cultures exposed to the same total photon flux averaged for the entire period. The interpretation was straightforward: only a limited number of photons can be used per unit time, and the millisecond light/dark periods allow averaging high photon fluxes. A large body of literature has developed on this subject, including laboratory work by the ASP (Terry 1984, 1986; see also Section II). The mass culture work in Hawaii (Section III. B.2.), among many others, attempted to use this phenomenon to increase algal productivities. However, practical applications are not plausible because of the very short time periods involved. Another approach, central to the Japanese microalgae program (Section IV. B.1.c.), has been to diffuse light throughout the depth of the culture, using optical fibers, thus avoiding high a surface irradiance. But this approach is also not practical for biodiesel production because of the very high cost of the system.

A potential practical solution to the light saturation problem, and also probably to photoinhibition, has been recognized for many years (e. g., Kok 1973): reduce the number of chlorophyll molecules cooperating in photosynthesis (the so-called “antenna” chlorophylls) from a few hundred to a few dozen. This would allow the photosynthetic apparatus to absorb only as much light as it can use. The benefits of reduced absorption are that it would:

• reduce waste,

• limit photooxidative damage to the photosynthetic reaction center, and

• increase the overall productivity of an algal culture, by a factor of at least 3 (see Benemann and Oswald 1996 for a recent discussion).

However, it has only recently become possible to consider achieving this objective, through the detailed understanding of photosynthesis at the molecular level, and the development of genetic engineering tools that could now allow us to redesign the photosynthetic apparatus. Recent work by Melis et al. (1998) and Neidhardt et al. (1998) demonstrated, at the physiological level, the feasibility of obtaining high efficiencies and high light saturation levels with algal cultures. Much more research is required, but the molecular and genetic tools are available to achieve the

desired high photosynthetic efficiencies by algal mass cultures. Such tools can also be used to direct the flow of photosynthate to desired metabolic products, such as lipids (see Section II).

Future R&D should demonstrate the feasibility of genetically engineering an improved photosynthesis system using algae for which such genetic systems are already well established. Once proven, these techniques can then be transferred to strains suitable for mass culture.

Successful genetic transformation of microalgal strains with demonstrated potential for biodiesel fuel production was finally accomplished in 1994. Two factors that were critical in the development of the transformation system were:

• the cloning of the acetyl-CoA carboxylase gene from C. cryptica, and thus the availability of ACCase regulatory sequences to drive expression of a foreign gene in the diatoms, and

• the purchase by NREL of a microprojectile accelerator (also known as a particle gun) that can efficiently deliver DNA-coated gold or tungsten beads into walled cells.

Except for the transient expression of luciferase in Chlorella protoplasts, all previous attempts at NREL to transform microalgae had been unsuccessful. Whether the problem was the inability to deliver foreign DNA into the cells through the algal cell wall, or inefficient expression of the foreign gene, is not clear.

As discussed in the previous section of this report, a significant amount of work went into developing homologous selectable markers for microalgae, primarily for Monoraphidium. However, there were some attempts, mainly with diatoms, to use a heterologous antibiotic resistance gene as a selectable marker. The GC content of bacteria and diatoms are relatively similar; thus, codon bias should not prevent expression of a bacterial gene in the diatoms. The antibiotic kanamycin and its more potent analog G418, have been used extensively for genetic transformation in higher plants. These antibiotics function by binding to 30S ribosomes and inhibiting protein synthesis. Resistance to kanamycin or G418 can be induced in cells by expressing the bacterial gene neomycin phosphotransferase (nptII). This enzyme phosphorylates the antibiotic, preventing binding to the ribosome. Previous work at NREL (Galloway 1990) demonstrated that some algal strains are sensitive to kanamycin, suggesting that the kanamycin — G418/nptII system might be the basis of a successful transformation system for microalgae.

Further testing showed that most of the algal strains were sensitive to low concentrations of G418; however, the conditions for complete inhibition of cell growth had to be determined empirically for each strain. The required concentration of the antibiotic depended both on the osmoticum of the plating medium and on the plating density of the cells. For example, C. cryptica T13L grows well on both 10% and 50% ASW. When 2 x 106 cells of T13L were plated on 50% ASW agar plates, the cells were resistant to 50 pg-mL-1 G418. The same number of cells plated onto 10% ASW plus 50 pg-mL-1 G418 showed no growth, yet 3 x 107 cells produced a confluent lawn of colonies under the same conditions.

Early attempts to use the nptII gene as a selectable marker used a plasmid construct that had been used successfully for transformation in higher plants. This plasmid, pCaMVNeo, was obtained from Dr. Michael Fromm at the USDA Plant Gene Expression Center, Albany California.

pCaMVNeo contains the nptII gene driven by the cauliflower mosaic virus 35S ribosomal gene promoter (CaMV35S). Attempts were made to introduce pCaMVNeo into C. crypticaCYCLO1 by electroporation, and later into C. ellipsoidea or CYCLO1 by agitating the cells with glass beads or SiC fibers. No G418-resistant colonies were generated by these methods.

After the acetyl-CoA carboxylase (acc1) gene was cloned from C. cryptica T13L, NREL researcher Paul Roessler decided to try to use the 5′- and 3′-regulatory regions from this gene to drive expression of nptII in T13L. A plasmid (pACCNPT10) was constructed that contained a chimeric gene consisting of the coding region of the nptII gene flanked by 445 bp of the acc1 5′ region (the putative promoter) and 275 bp of acc1 coding region following the nptII stop codon, followed by the acc1 3′ noncoding regions (the putative transcriptional terminator). To increase the chance of encompassing the entire acc1 promoter, a second plasmid, pACCNPT5.1, was constructed that contained 819 bp of upstream sequence. In addition, all but 13 bp of the acc1 coding region was removed from the 5′ end of chimeric gene. Details of the plasmid constructions can be found in Dunahay et al. (1995), and plasmid maps are shown in Figure II. B.8.

DNA entry into the algal cells was accomplished using the DuPont/Bio-Rad PDS/1000He microprojectile accelerator. The process, called biolistics, had been used successfully for introducting DNA into walled cells of higher plants, fungi, bacteria, and Chlamydomonas. In this procedure, plasmid DNA is precipitated onto small tungsten or gold particles and accelerated into cells using a burst of helium pressure. Early versions of this device used a gun powder charge to accelerate the particles. Because of prohibitive costs and restrictive licensing agreements, a homemade version of the particle gun was designed and built at NREL. No transformants were generated using this device, but as these experiments were performed before the acc1-nptII chimeric plasmids where constructed, whether the device actually functioned as planned is unclear. Ultimately, a commerical microprojectile accelerator was purchased. This device was optimized for very simple operation and used helium pressure to propel the DNA-coated particles. These properties resulted in greater reproducibility between shots and decreased toxicity caused by gases generated during the explosive charge.

There was some initial skepticism on the part of at least one NREL researcher as to whether microprojectile bombardment would work to introduce DNA into diatoms through the Si frustule. However, the diatoms were transformed using the particle gun and the chimeric vectors in the first try. This turned out to be a simple and reproducible procedure (Figure II. B.9.). For each transformation, algal cells were harvested and spread in an approximate monolayer in the center of an agar plate containing growth medium and 50 pg-mL-1 ampicillin to inhibit bacterial growth. The plates were allowed to dry for 2 hours before bombardment. Just before bombardment, plasmid DNA was precipitated onto 0.5-1.0 pm tungsten particles, which were then propelled into the cells using the microprojectile accelerator. The exact parameters used are described in Dunahay et al. (1995). The cells were incubated for 2 days under nonselective conditions to allow the cells to recover and express the nptII gene. The cells were then washed from the orignal plates and replated onto agar that contained the appropriate concentration of G418. G418-resistant colonies appeared in 7-10 days. These putative transformants were picked

from the plates and tested for continued resistance to G418. The presence of the foreign gene was tested by hybridizing the algal DNA with an nptII gene probe (Southern analysis). The cells were tested for the presence of the and for the NPTII protein by probing with an NPTII-specific antibody (Western blotting), Figures II. B. 10 and II. B. 11.

Both the pACCNPT 10 and pACCNPT 5. 1 plasmids worked well to generate transformants in two strains of C. cryptica (T13L and CYCLO1), as well as in the diatom N. saprophila (NAVIC1). These two species belong to different orders (C. cryptica is a centric diatom, Order Centrales; N. saprophila is a pennate diatom, Order Pennales). Southern analysis indicated that the plasmid DNA was not replicating independently in the cells but had integrated into the host genome, presumably into the nuclear DNA. The chimeric gene integrated into one or more independent sites, often in form of tandem repeats. The nptII DNA remained stably integrated into the host genome for more than 1 year, even when the cells were grown under nonselective conditions.

The successful development of a genetic transformation system for the diatoms was a major achievement for the ASP. This was the first report of genetic transformation of any diatom species, and one of the few reports in which a heterologous gene was used as a selectable marker for stable nuclear transformation of an alga. The use of algal regulatory sequences to drive expression of the bacterial gene in diatoms apparently was a key factor in the successful development of a transformation protocol for these organisms. When the pCaMVNeo plasmid was introduced into diatoms via particle bombardment, no G418-resistant transformants were generated. However, when another plasmid that contains the CaMV35S promoter and the firefly luciferase gene were introduced into the diatoms by cotransformation with pACCNPT5.1, a number of transformants selected based on their resistance to G418 also expressed significant luciferase activity. This result suggests that even though microalgae can in some cases recognize and use foreign promoter sequences, homologous promoters may be necessary to drive expression of foreign selectable markers at levels high enough to overcome the selective pressure. The research that resulted in the development of a genetic tranformation system for diatoms resulted in a publication (Dunahay et al. 1995) that was a finalist for the Provasoli Award for best publication in the Journal of Phycology for that year. In addition, a patent describing this technology was applied for and issued in August 1997. Diatoms represent a very large proportion of the world’s biomass, and are responsible for nearly one-fourth of the net primary production. However, little is known about the biochemistry and molecular biology of these organisms. The availability of a genetic transformation system for diatoms could have a major impact on increasing the understanding of the basic biology of these organisms and should promote their use in biotechnological applications in addition to the intended goal of lipid production. The following section will describe the initial attempts to use the genetic transformation protocol to manipulate levels of storage lipids in C. cryptica.

A major issue in microalgae mass cultures is the understanding of the factors that determine species dominance. A considerable theoretical background can be gleaned from the ecological and ecophysiological literature, where this problem, as it applies to lotic systems (ponds, lakes, rivers, oceans) is the subject of an enormous literature (briefly reviewed in Benemann and Tillett [1987]). Fundamental to this is Hutchins’ so-called “Paradox of the Plankton.” Hutchins pointed out the fact that only a limited number of nutrients could limit algal growth, and thus, applying Liebing’s Law of the Minimum, only a limited number (one per nutrient) of algal species that should be able to compete in any environment. Rather, we find literally hundreds, if not thousands of species and uncountable strains in even the smallest and most uniform of environments. Indeed, even Hutchins did not realize the greatness of microbial biodiversity in nature, hinted at by earlier work in clonal variations (See Section II. B.), but only recently revealed in ever-increasing detail using modern tools of molecular phylogenetics. The solution to the Paradox is that natural environments are not steady-state systems, but are exposed to periodic and random fluctuations in physical-chemical (let alone biotic) environmental parameters. These fluctuations allow for additional niches, allowing for organisms specialized in the exploitation of particular temporal combination of limiting factors. Also, non-steady-state conditions would select for different species and strains. For example, in a continuous cultivation using natural samples for inoculum into enriched media, the first algal species to appear and dominate are soon replaced by other species, which are slower growing but better at exploiting a light-limited (dense culture) environment.

Continuous algal cultures in 1-L vessels were set up in the laboratory, which allowed operations under fluctuating conditions of temperature, O2 concentrations, and pH. In these experiments, several algal strains were inoculated together, then productivity and species dominance were observed for as long as 3 weeks (over 10 dilution times). In initial experiments, Chlorella and Chaetoceros co-dominated; Cyclotella was lost or greatly diminished, possibly because of different light spectral use of the two types of algae (greens and diatoms). Many other experiments were carried out, with fluctuating pH, dilution rates, and light intensities and even gas sparger types on the dominance of these and one additional algal strain (Ankistrodesmus). All these factors tended to affect species dominance, even the gas sparger, and the results were not clearly interpretable in terms of major dominance factors. One conclusion from these initial experiments was that several factors, alone and in combination, can determine species dominance.

The lipid contents of several strains were determined for cultures in exponential growth phase and for cultures that were N-limited for 7 days or Si-limited for 2 days. In general, nutrient deficiency led to an increase in the lipid content of the cells, but this was not always the case. The highest lipid content occurred with NAVIC1, which increased from 22% in exponential phase cells to 49% in Si-deficient cells and to 58% in N-deficient cells. For the green alga MONOR2, the lipid content increased from 22% in exponentially growing cells to 52% for cells that had been N-starved for 7 days. CHAET14 also exhibited a large increase in lipid content in response to Si and N deficiency, increasing from 19% to 39% and 38%, respectively. A more modest increase occurred for nutrient-deficient AMPHO1 cells, whereas the lipid content of CY CLO2 was similar in exponential phase and nutrient-deficient cells, and actually decreased in AMPHO2 as a result of nutrient deficiency.

These results suggested that high lipid content was indeed achievable in many strains by manipulating the nutrient levels in the growth media. However, these experiments did not provide information on actual lipid productivity in the cultures, which is the more important factor for developing a commercially viable biodiesel production process. This lack of lipid productivity data also occurred with most of the ASP subcontractors involved in strain screening and characterization, but was understandable because the process for maximizing lipid yields from microalgae grown in mass culture never was optimized. Therefore, there was no basis for designing experiments to estimate lipid productivity potential.

I Publications:

Barclay, B.; Nagle, N.; Terry, K. (1986) “Screening microalgae for biomass production potential: protocol modification and evaluation.” FY1986Aquatic Species Program Annual Report, Solar Energy Research Institute, Golden, Colorado, SERI/SP-231-3071, pp. 22-40.

Barclay, W. R.; Terry, K. L.; Nagle, N. J.; Weissman, J. C.; Goebel, R. P. (1987) “Potential of new strains of marine and inland saline-adapted microalgae for aquaculture.” J. World. Aquaculture Soc. 18:218-228.

In the early 1980s, Dr. Meints and Dr. Van Etten were studying Chlorella-like green algae that live in a symbiotic relationship within cells of the protozoan Hydra viridis. They found that the algal cells could be excised from the hydra, which could exist free of the symbiont if given proper nutrients. However, it was not possible to culture the algae free of the hydra host. Further study demonstrated that when the algal cells were isolated from the host, virus particles rapidly began to multiply within the algae, resulting in lysis of the algal population within 24 hours. Ultrastructual and biochemical studies on this algal virus system produced the following results:

• The virus consisted of a large (approximately 190-nm), polygonal particle, containing 30 to 40 polypeptides, the most abundant of which was a 46 kDa glycoprotein, presumably associated with the viral capsid.

• The virus genome consists of about 130 kbp of double-stranded DNA.

• New virus particles were assembled in the cytoplasm of the algal cells and released upon lysis of the algal cell wall.

This virus, called HVCV (for Hydra viridis Chlorella virus), was one of the few viruses described in eukaryotic algae. HV CV might play a role in initiating or maintaining the symbiotic

relationship between the alga and its hydra host, possibly by altering the algal cell wall. Subsequent studies identified viruses in four other strains of Hydra obtained from commercial sources. The viruses fell into two classes, based on particle size, bouyant density, and DNA restriction patterns (HV CV -1, HV CV -2). In addition, a similar virus was isolated from symbiotic Chlorella from Paramecium bursaria (PBCV-1).

To facilitate the study of these viruses, it was desirable to identify an algal strain that could be cultured free of the hydra or Paramecium host, which was susceptible to infection by the virus. This would allow production of large quantities of the virus and the study of viral replication and development. Sixteen strains of culturable Chlorella, which had been isolated from invertebrates such as Paramecium, Hydra, Stentor, and sponges, plus two free-living strains, were obtained. Attempts were made to infect these Chlorella strains with all the virus strains described earlier. None of the HV CV viruses (from Chlorella-Hydra hosts) were able to infect any Chlorella strain tested. However, two culturable Chlorella strains originally isolated from Paramecium (Chlorella strains N1a and NC64) were infected with PBCV-1 (the P. bursaia Chlorella virus). The infection led to lysis of the algal cells and production of large amounts of infectious viral progeny. This result led to the development of a plaque assay system for the algal viruses, similar to a bacteriophage assay on bacterial lawns. The availability of this system, which caused synchronous infection of the algal cells and the production of large quantities of viral particles, allowed the researchers to characterize the virus biochemically. It also allowed researchers to study the regulation of viral gene expression and the effects of viral replication on algal physiology and gene expression. A large number of publications resulted from this research (see below). Several of the most interesting and possibly relevant findings are summarized here.

• The virus particles attach at one vertex of their polygonal capsid to receptor sites on the algal cell wall. A lytic enzyme produced by the virus degrades the wall at this site, and the viral DNA is released into the cell. Living algal cells are not required for virus attachment and wall degradation (viruses can attach to and degrade isolated wall fragments), but living cells are necessary for release of the viral DNA. Complete viral capsids are assembled in viral assembly sites within the cytoplasm and subsequently filled with DNA. Virus particles are released through holes produced at discrete locations in the algal cell wall.

• The plaque assay system was used to screen for other virues that infect algae.

Viruses that infect Chlorella strains N1a or NC64A were found to be very common in nature. The viruses all had similar features, including a large, polygonal capsid and dsDNA; however, some viruses were distinguishable based on plaque size, reactivity to anti-PBCV -1 antisera, variations in the DNA restriction patterns and the extent of nucleotide modification. [7]

expression of overlapping genes or transcription of genes from both DNA strands. Analysis of DNA from some viral isolates showed that the viral DNA is modified to varying extents, primarily in the methylation of adenine and cytosine residues. The data suggested that the virus produces a unique restriction enzyme that is specific for non-methylated sequences for degradation of the host DNA. The virus also produces a corresponding methyltransferase, which recognizes the same sequence as the restriction endonuclease. The methyltransferase methylates newly synthesized viral DNA, protecting it from degradation in the next round of infection.

Dr. Meints received funding from SERI from 1986 through 1989. The overall goal of the SERI — funded research was to use the algal virus system to develop methods for genetically manipulating microalgae with potential for liquid fuel production. The research from Dr. Meints’ laboratory is reviewed below with respect to the specific goals of the project.

Executive Summary

From 1978 to 1996, the U. S. Department of Energy’s Office of Fuels Development funded a program to develop renewable transportation fuels from algae. The main focus of the program, know as the Aquatic Species Program (or ASP) was the production of biodiesel from high lipid-content algae grown in ponds, utilizing waste CO2 from coal fired power plants. Over the almost two decades of this program, tremendous advances were made in the science of manipulating the metabolism of algae and the engineering of microalgae algae production systems. Technical highlights of the program are summarized below:

Applied Biology

і A unique collection of oil-producing microalgae.

The ASP studied a fairly specific aspect of algae—their ability to produce natural oils. Researchers not only concerned themselves with finding algae that produced a lot of oil, but also with algae that grow under severe conditions—extremes of temperature, pH and salinity. At the outset of the program, no collections existed that either emphasized or characterized algae in terms of these constraints. Early on, researchers set out to build such a collection. Algae were collected from sites in the west, the northwest and the southeastern regions of the continental U. S., as well as Hawaii. At its peak, the collection contained over 3,000 strains of organisms. After screening, isolation and characterization efforts, the collection was eventually winnowed down to around 300 species, mostly green algae and diatoms. The collection, now housed at the University of Hawaii, is still available to researchers. This collection is an untapped resource, both in terms of the unique organisms available and the mostly untapped genetic resource they represent. It is our sincere hope that future researchers will make use of the collection not only as a source of new products for energy production, but for many as yet undiscovered new products and genes for industry and medicine.

і Shedding light on the physiology and biochemistry of algae.

Prior to this program, little work had been done to improve oil production in algal organisms. Much of the program’s research focused attention on the elusive “lipid trigger.” (Lipids are another generic name for TAGs, the primary storage form of natural oils.) This “trigger” refers to the observation that, under environmental stress, many microalgae appeared to flip a switch to turn on production of TAGs. Nutrient deficiency was the major factor studied. Our work with nitrogen-deficiency in algae and silicon deficiency in diatoms did not turn up any overwhelming evidence in support of this trigger theory. The common thread among the studies showing increased oil production under stress seems to be the observed cessation of cell division. While the rate of production of all cell components is lower under nutrient starvation, oil production seems to remain higher, leading to an accumulation of oil in the cells. The increased oil content of the algae does not to lead to increased overall productivity of oil. In fact, overall rates of oil production are lower during periods of nutrient deficiency. Higher levels of oil in the cells are more than offset by lower rates of cell growth.

Breakthroughs in molecular biology and genetic engineering.

Plant biotechnology is a field that is only now coming into its own. Within the field of plant biotechnology, algae research is one of the least trodden territories. The slower rate of advance in this field makes each step forward in our research all the more remarkable. Our work on the molecular biology and genetics of algae is thus marked with significant scientific discoveries. The program was the first to isolate the enzyme Acetyl CoA Carboxylase (ACCase) from a diatom. This enzyme was found to catalyze a key metabolic step in the synthesis of oils in algae. The gene that encodes for the production of ACCase was eventually isolated and cloned. This was the first report of the cloning of the full sequence of the ACCase gene in any photosynthetic organism. With this gene in hand, researchers went on to develop the first successful transformation system for diatoms—the tools and genetic components for expressing a foreign gene. The ACCase gene and the transformation system for diatoms have both been patented. In the closing days of the program, researchers initiated the first experiments in metabolic engineering as a means of increasing oil production. Researchers demonstrated an ability to make algae over-express the ACCase gene, a major milestone for the research, with the hope that increasing the level of ACCase activity in the cells would lead to higher oil production. These early experiments did not, however, demonstrate increased oil production in the cells.