As it was described in Chapter 1 our main aim was to investigate the possible application of a "milking" process using two-phase bioreactors in microalgal biotechnology. We choseb-carotene extraction from Dunaliella salina in a two-phase bioreactor with an aqueous phase and an organic solvent phase as our model system. The goal was to develop an alternative and more efficient process than the commercial production process ofb-carotene.
In chapter 2 biocompatibility of different solvents with values of log P octanol ranging from 3 to 9 for the cells of Dunaliella salina was investigated. Extraction ability of different solvents for bothb-carotene and chlorophyll was determined as well. Results showed that solvents having log P octanol & gt; 6 can be considered biocompatible for this alga. Moreover, pigment extraction ability of a solvent is inversely dependent on its log P octanol value. By increasing the degenerative hydrophobicity the extraction ability for both chlorophyll andb-carotene, decreased. However, this decrease was more pronouned for chlorophyll. Therefore, selective extraction ofb-carotene becomes feasible. Theb-carotene productivity per cell in a two-phase system with dodecane was the highest observed. Extraction ability of the biocompatible solvents dodecane, tetradecane and hexadecane was similar.
Effect of mixing rate which is supposed to lead to the facilitated release ofb-carotene from the cells of Dunaliella salina in two-phase bioreactors, was investigated in chapter 3 . Three pairs of bioreactors were inoculated at the same time, operated at 100, 150 and 170 rounds per minute, respectively and illuminated with a light intensity of 700mmol m -2s -1. Each pair consisted of one bioreactor containing only aqueous phase for the blank and one containing the water phase together with the biocompatible sovent (dodecane). Comparison of the viability and growth of the cells grown under different agitation rates showed that 170 rpm and 150 rpm were just as good as 100 rpm. Presence and absence of the organic phase had also no influence on the viability and growth of the cells. In contrast to the growth rate, the extraction rate ofb-carotene was influenced by the stirrer speed. The extraction rate increases at higher stirring rate. The effectiveness of extraction per amount of power in-put was comparable for all the applied mixing rates.
In chapter 4 the effect of light intensity on the extraction ofb-carotene from Dunaliella salina , in the fermentative extraction, was investigated. Three different average light exposures were applied: 1.5*10 -8(low), 2.7* 10 -8(intermediate) and 4.5* 10 -8(high)mmol s -1cell -1. Results showed thatb-carotene content of the cells increases by increasing the light exposure. Increase in theb-carotene content of the cells was not necessarily coupled with an increase in the volumetric production ofb-carotene.b-Carotene extraction rate was enhanced by the increase in the light exposure. The results suggest that extraction rate was related tob-carotene content of the cells and was not essentially related to the volumetric productivity ofb-carotene. Although the effectiveness of extraction with respect to the light input was comparable for all light intensities applied, increasing the light input per cell leaded to a higher volumetric extraction rate.
On the basis of the previous results the "milking" process forb-carotene production was developed and introduced in Chapter 5 . Growth of the cells was performed at low light intensity after which the cells were transferred to the production bioreactor, which was illuminated at a higher light intensity. The second bioreactor was a two-phase bioreactor consisting out of an aqueous and a biocompatible organic phase. In this bioreactor mixing and extraction were performed by re-circulation of the organic phase. The results showed that D. salina stayed viable for a long period (& gt;47 d) in the presence of a biocompatible organic phase at high light intensity. The cell growth, however, was very slow in this situation.b-Carotene could be continuously extracted to the organic phase. The cells kept producingb-carotene and the extracted molecules were substituted by the cells. As a resultb-carotene was continuously milked from the cells. Theb-carotene extraction efficiency in this system was more than 55%. The productivity of the system was 2.45 mg. m -2.d -1which is much higher than obtained in commercial plants forb-carotene production.
Several studies at macro-scale (bioreactors) and micro-scale (using microscopic techniques) were performed for better understanding the mechanism of the extraction process. The results are presented in Chapter 6 . Based on the results two hypothesis were made for the extraction: one of the mechanisms of extraction is transport of theb-carotene globules from the chloroplast to the space between the cell and the chloroplast membranes and subsequently from there to the outside by exo-cytosis. Another possible mode for the extraction could be release ofb-carotene from the globules as a result of alterations in the membrane of globules.b-Carotene molecules diffuse from the chloroplast to the space between the cell and the chloroplast membranes and from there to the medium either by diffusion or by exo-cytosis after accumulation in the vesicles.
In the last chapter, Chapter 7, we discuss the approaches which are helpful in answering two following fundamental questions: is it possible to milk all microalgae? And would this technique be suitable for mass production of high-value secondary metabolites? We think to answer these questions the mechanism of extraction and its relation with the production pathway of the target product should be exactly understood. Our previous results and the results of other researches suggest that chemical behavior and molecular structure of the solvent, chemical properties of the product and its location inside the cells and finally physiological behavior of the cell membrane and the cells by themselves are important parameters in a successful milking process.