Oak Ridge National Laboratory, Oak Ridge, Tennessee Jean A. Solomon 12/86 — 11/87 DK-4-04142-01

The purpose of this project was to determine if flow cytometry could be used to select for subpopulations of high lipid-producing algae within an algal culture. Flow cytometry is a method that measures the light scattered or emitted by particles as they pass through a laser beam. Scattered light is believed to reflect the size, shape, and refractive properties of cells. Dr. Solomon initially used exponentially growing and nutrient-stressed cells of the chrysophycean alga Boekelovia to demonstrate that the extent of right-angle scatter, which indicates changes in internal cell morphology, can be correlated to the lipid content of microalgal cells.

In subsequent studies, a lipid-specific fluorescent dye, Nile Red (see work by Dr. Cooksey, described in Section II. A.l. f.), was used to stain intracellular lipids. Nile Red is excited at a wavelength of 488 nm, and emits yellow-green light at 520-580 nm. In contrast, chlorophyll autofluorescence can be measured at wavelengths greater than 630 nm. Therefore, in contrast to the scattered light data mentioned above, flow cytometric analyses of cells stained with Nile Red would be more specific for changes in lipid content. Preliminary experiments in which cells of Boekelovia were stained with Nile Red demonstrated that increased yellow green fluorescence could be correlated with increased numbers of lipid droplets in the cells, suggesting that this method could work to screen for cells with high lipid contents.

Of the three microalgal species analyzed by TEM (Section II. B.l. d.), only Isochrysis was found to be appropriate for flow cytometric analysis. The cells of this strain are small and spherical, the optimal shape for flow cytometric analysis, and take up Nile Red well. In contrast, Ankistrodesmus cells are long and thin (40 nm x 4 nm). Nannochloropsis did not take up the Nile Red dye, possibly because of cell wall properties that also prevented good chemical fixation for microsopy.

In the initial experiments, cells were screened for lipid content based on Nile Red fluorescence alone. Several improvements to this procedure were implemented during the course of the study. First, efforts were made to optimize the Nile Red staining protocol. The best staining was achieved using a concentration of 1 mg Nile Red in 1 mL of cell suspension. The solvent for the Nile Red stain was changed from heptane to acetone, due to interfering fluorescence from undissolved heptane droplets. Finally, the researchers found that the fluorescence signal from Nile Red is unstable and decays rapidly. However, the fluorescence level stabilizes after about 45 minutes, so all readings were taken at least 45 minutes after staining the cells with Nile Red.

Another important change was to measure the chlorophyll autofluorescence as well as Nile Red fluorescence. This ensured that only viable cells containing lipid and intact chloroplasts would

be analyzed. In addition, the amount of chlorophyll is an indication of cell size. Cell sorting based on the ratio of chlorophyll fluorescence to Nile Red fluorescence would normalize the results to account for differences in cell size and age and allow detection of individual cells with unusually high lipid levels resulting from natural genetic variation. A decrease in the ratio of chlorophyll to Nile Red fluorescence would indicate lipid accumulation.

In one set of experiments, Isochrysis cultures were stressed by transferring the cells into N — deficient media, then screened for lipid content using flow cytometry, either by lipid content alone (Nile Red fluorescence) or by monitoring the chlorophyll to Nile Red fluorescence ratios. The daughter cells containing high or low levels of lipid were recultured in N-replete medium for

1 week or 1 month, then subjected again to N deprivation, and resorted. The lipid content of the daughter population was compared to that of the parent cells. These experiments produced inconsistent results. In some cases, the population of daughter cells selected for their high lipid content showed a wider range of lipid contents than the parent cells; other sorts produced daughters without significant differences in lipid content from the parents. One interesting observation was the bimodal distribution of cells in all populations subjected to N stress. Cells fell into two classes with low or high chlorophyll-to-lipid ratios. This again supports the author’s theory, discussed in the section on ultrastructural analysis of lipid accumulation, that cells respond as individuals to a lipid trigger, rather than gradually increasing the lipid content of the entire culture.

These results suggested that flow cytometry might be used to select for populations of high lipid algae if more was understood about the relationship of the physiological state of parent cells to lipid accumulation. Analysis of the growth of Isochrysis in N-replete media showed several phases, including a period of exponential growth that declined to a stationary phase. After about

2 weeks, the nutrients were depleted and the cells entered a stressed phase. A series of experiments was performed in which cells in various growth phases were stained with Nile Red and sorted based on lipid content alone (no chlorophyll measurements). The sorts on cells in exponential phase were usually not successful, but if the parent cells were in stationary phase or from very old cultures (stressed), the mean lipid content of the daughter population was about 20% higher than that of the parental cells. These results suggested that successful screening for high-lipid cells using flow cytometry was related to the cell cycle. The exponential cultures contained cells in all stages of cell growth and division. Cells that were preparing to divide would be larger and would be selected as high-lipid, so that the set of high lipid cells selected would actually contain only the largest and oldest cells, rather than high lipid genetic variants. Stationary or stressed populations are not actively dividing, so the cells are more uniform in size and sorting of the cells for high lipid should be more likely to identify true high lipid variants within the population.

The experiments using Isochrysis cultures in various growth stages as the parental population were repeated to test these assumptions. Cells were sorted based either on lipid levels alone (Nile Red fluorescence only) or using the ratio of chlorophyll to Nile Red fluorescence to control for cell size. The results presented in the available reports support the hypotheses. Sorting cells based solely on lipid content produced high-lipid daughter populations if the parent population

was in a stationary or stressed growth phase. Exponential cultures produced variable results. If the parental cells were sorted based on ratios of chlorophyll-to-lipid fluorescence, high-lipid populations could be produced from exponentially growing parent cultures.

These conclusions were based on very limited data. Only a few experiments were performed. In addition, several daughter cultures did not grow up after the sorting process, and failure of a growth chamber resulted in the loss of some cultures. The data were written up in only two technical reports to SERI, and it was difficult to determine the exact protocol used for each sorting experiment. However, the results of these studies are intriquing, and suggest that flow cytometry might be a viable method for screening for high lipid genetic variants within (or between) strains of oleaginous microalgae. The procedure would be limited to strains in which the cells are small and spherical. Dr. Solomon’s data suggest the best results would be achieved by using stationary or stressed cells as the parent population. In addition, cells should be selected based on a low chlorophyll-to-Nile Red fluorescence ratio, which would indicate high-lipid levels with respect to cell size. However, it is interesting that high-lipid daughter strains were not produced in the experiments in which exponentially growing cells were transferred directly to N — deficient media; yet, cells allowed to gradually deplete their N supply to induce the stressed condition could be used successfully in a flow cytometric screen. This suggests either that lipid accumulation occurs by different mechanisms under these two conditions, or, more likely, that the stressed cells from very old cultures had all entered a similar metabolic state so that size and lipid contents would be more indicative of genetic differences.

I Publications:

Solomon, J. A (1985) “Ultrastructure evaluation of lipid accumulation in microalgae.” Aquatic Species Program Review: Proceedings of the March 1985 Principal Investigators ’ Meeting, Solar Energy Research Institute, Golden, Colorado, SERI/CP-231-2700, pp. 71-82.

Solomon, J. A. (1987) “Flow cytometry techniques for species improvement.” FY1986Aquatic Species Program: Annual Report (abstr.), Solar Energy Research Institute, Golden, Colorado, SERI/SP-231-3071, p. 252.

Solomon, J. A.; Hand, R. E.; Mann, R. C. (1986a) “Ultrastructural and flow cytometric analyses of lipid accumulation in microalgae.” Annual Report, Oak Ridge National Laboratory, Oak Ridge, Tennessee, ORNL/M-258, 50 pp.

Solomon, J. A.; Hand, R. E.; Mann, R. C. (1986b) “Ultrastructural and Flow Cytometric Analyses of Lipid Accumulation in Microalgae: A Subcontract Report.” Solar Energy Research Institute, Golden, Colorado, SERI/STR-231-3089.

Solomon, J. A.; Palumbo, A. V. (1987) “Improvement of microalgal strains for lipid production.” FY 1987 Aquatic Species Program: Annual report, Solar Energy Research Institute, Golden, Colorado, SERI/SP-231-3206.