Improving fucoxanthin and docosahexaenoic acid production in Tisochrysis lutea

In this thesis, two strategies (process optimization and strain improvement) were applied to improve fucoxanthin (Fx) and docosahexaenoic acid (DHA) productivities with Tisochrysis lutea. Experiments were conducted with different temperatures, light intensities, dilution rates and biomass concentrations, to obtain the optimal parameters for Fx and DHA productivities, at different scales, indoors and outdoors. Continuous cultivation at selected conditions (red light, low light and low temperature) combined with cell sorting was applied as an approach for strain improvement, resulting in 3 improved phenotypes. We achieved the highest ever reported Fx and DHA productivities, using the optimal cultivation parameters and improved strains obtained in this thesis.In chapter 2, we optimized the temperature, light intensity and dilution rate to improve Fx content and productivity using batch and continuous chemostat experiments. The highest Fx content (16.39 mg g-1) was found at 50 μmol m-2 s-1, dilution rate 0.47 d-1, 30 °C; and the maximum Fx productivity (9.81 mg L-1 d-1) was found at 300 μmol m-2 s-1, dilution rate 0.80 d-1, 30 °C. Continuous chemostat cultivation resulted in 3-9× higher Fx productivity than obtained in batch with the same light intensity, showing a clear advantage in Fx production. We found that the Fx content was negatively correlated to the light absorbed and Fx productivity was positively affected by dilution rate. In this chapter, we achieved the highest ever reported Fx content by process optimization. Overall, process optimization can improve Fx production significantly.In chapter 3, T. lutea was cultivated outdoors using a pilot flat-panel photobioreactors (40 L) with high (1.1 g L-1) and low (0.4 g L-1) biomass concentrations at semi-continuous mode. The light can hardly be controlled in outdoor conditions; however, one can change the biomass concentration inside the reactors to control the light perceived by cells (light/cell). To make a good comparison with existing commercial Fx producers, the microalga Phaeodactylum tricornutum was cultivated in parallel under identical conditions. In addition, a monitoring tool based on fluorescence spectroscopy was developed to predict biomass and Fx in a fast and non-invasive manner. We found that both biomass and Fx productivities were higher with low biomass concentration in both microalgae. However, Fx content was higher with a high biomass concentration. The Fx productivity reached 2.09 mg L-1 d-1 in T. lutea, which was higher than that (1.73 mg L-1 d-1) in P. tricornutum. The developed prediction models showed a high correlation between fluorescence and cell concentration and fucoxanthin content; the same model works for both species. This is the first time for a model to be used for different microalgal species. In this chapter, we reported how biomass and Fx productivities can be manipulated in outdoor conditions and estimated using fluorescence monitoring.In chapter 4, T. lutea was grown in Erlenmeyer flasks and Algaemists under batch and continuous chemostat modes. The Fx and lipids (polar and neutral) were measured daily. Plate reader and fluorescence-activated cell sorting (FACS) were used to measure the fluorescence of the culture in batch and the fluorescence of single cells in continuous cultivation. Nitrogen plays an important role in Fx and polar lipids production. Fx content decreased 52.94% from nitrogen-sufficient to starvation conditions in batch cultivation; Fx increased 40.53% as nitrogen supplementation increased from 21.76 to 74.79 mg d-1 in continuous cultivation. The polar lipids followed the same trends as Fx. Both polar and neutral lipids in living T. lutea cells can be stained with Nile red and the ratio between them can be monitored using FACS at 617 nm and 585 nm. We found that Fx, chlorophyll a and polar lipids were always positively correlated to each other in both batch and continuous cultivations with different nitrogen levels. Therefore, these compounds can be measured simultaneously. Chlorophyll a was selected as a proxy of Fx and polar lipids due to the positively correlated emission signals at 720 nm. The cell viability analysis showed that ≈80% cells can survive after staining + FACS sorting. In this chapter we analysed the dynamics and correlations of Fx and lipids at different cultivation conditions, and developed a high throughput method to monitor and sort Fx and lipids based on single-cell fluorescence using FACS.In chapter 5, six light spectra were applied as illumination in continuous chemostat mode at 50 μmol m-2 s-1. This light intensity was used because it leads to the highest Fx content as shown in chapter 2. FACS was also used to monitor the single-cell fluorescence. In this chapter blue + green light resulted in the highest Fx content (16.8 mg g-1 DW) while red + green + blue light resulted in the highest Fx productivity (2.77 mg L-1 d-1). Both lowest Fx content (13.3 mg g-1 DW) and productivity (1.16 mg L-1 d-1) were found with the red light. It was hypothesized that high Fx producing cells under adverse conditions (red light) could lead to improved strains. This hypothesis was confirmed by a sorted strain from a culture cultivated under red light, which showed 16% higher Fx productivity. Although different light spectra did not affect total lipids (17.31-19.46% DW) and DHA (2.08-3.21% DW) contents, the neutral lipids increased significantly with light containing green light. Single-cell Fx content was positively correlated to emission signals at 720 nm, indicating the method developed in chapter 4 is also valid at different light spectra. Different light spectra could be applied as selection pressure to result in populations with different Fx content that can be selected using FACS for T. lutea strain improvement.In chapter 6, continuous chemostat cultivation at a low light intensity of 50 μmol m-2 s-1 was used to obtain T. lutea cells with high Fx content. FACS was applied to select and sort cells with high Fx fluorescence (top 0.8%) using the method developed in chapter 4. A novel phenotype resulted from 2 rounds of continuous cultivation and 2 rounds of FACS selection. This phenotype forms cell aggregate, has no flagella, and showed to be a stable phenotype after 15 months. The outdoor Fx productivity increased 3.1× compared to the original strain, up to 2.1 mg L-1 d-1, which is the highest ever reported Fx productivity outdoors. The novel phenotype has three main advantages: [1] high biomass and valuable compounds productivities (4-9× higher), [2] low contamination risk (self-aggregating cells with mucus layer), [3] low harvesting cost, as the culture settles spontaneously in less than 60 min. The findings of this chapter show that productivities can be improved by the selection of improved phenotypes using FACS. The resulting phenotype shows high potential for improving the industrial production of T. lutea.In chapter 7, the effect of decreasing temperature on Fx and lipids production was studied using continuous turbidostat experiments. Following, long-term continuous low temperature adaptation combined with FACS selection was applied to develop winter strains. Both Fx content and productivity decreased from 30 to 15 °C with wildtype T. lutea. Although the DHA content did not decrease at 15 °C, its productivity decreased 85%. These results indicated that 15 °C is unsuitable for T. lutea production. To obtain an improved strain that can grow at 15 °C, we used a combined strategy: firstly, a gradual temperature decrease from 20 to 15 °C was applied for adaptation; secondly, long-term continuous turbidostat cultivation was performed at 15 °C; following, FACS was applied to select cells with high fluorescence at 720 nm. A winter strain was obtained after two rounds of FACS selection with a growth rate of 0.37 d-1 at 15 °C. The Fx and DHA productivities of the winter strain reached 1.65 and 8.95 mg L-1 d-1, respectively, whereas the original strain showed productivities of 0.01 for Fx and 0.15 mg L-1 d-1 for DHA, under the same cultivation conditions. The winter strain obtained was delivered to our project partner NECTON, S.A. (Olhão, Portugal) for industrial outdoor production.In chapter 8, we summarize Fx and DHA content and productivities reported in microalgae and show that T. lutea is competitive in the production of these compounds with relatively high content and productivity. Following, we discussed how to use the strategies applied in this thesis, i.e. process optimization and strain improvement, to achieve higher Fx and DHA productivities than currently. Scale up of improved strains is discussed for commercial exploitation. Using our improved strains and optimal parameters, we estimate that lower (70-80%) biomass and Fx production costs could be achieved in industrial scale. More selection rounds are recommended for further improvement of strains with high Fx content and adaptability to fluctuating temperatures.Overall, this thesis obtained the optimal process parameters for Fx and DHA contents and productivities with T. lutea. High throughput method and prediction model were developed for Fx and lipids quantification and cell selection. Three improved strains were obtained using continuous cultivation at selected conditions combined with cell sorting, which filled the gap of low productivities and seasonality of T. lutea, and opened new possibilities for Fx and DHA cost-effective production.