Sunlight, the ultimate energy source of the earth, is of great importance for growth and development of all living creatures. With the alternation of days and nights and the changing of the four seasons, light intensity in every corner of the earth varies and changes permanently. Thus, the ability for living creatures to detect the ambient light change becomes necessary. Not like animals, which use opsin-based visual system to receive information about the change of light, plants managed to develop a set of different nonopsin photoreceptors, which are able to perceive a wide range of light qualities and intensities.

In the range of 400-850nm of the light spectrum, plants have three major classes of photoreceptors, the cryptochromes (cry), the phototropins (phot), and the phytochromes (phy). Among those photoreceptors, cryptochromes and phototropins are in charge of monitoring the blue/ultraviolet (B/UV-A) region of the spectrum, while the phytochromes monitor mainly the red (R) and far red (FR) wavelength.

Figure 2.1.1 Domain features of plant phototropin

The phototropins could be divided into two major regions: the LOV domain region and a Ser/Thr kinase region. Two LOV domains are sensor domains which are in charge of receiving blue light signals and the Ser/Thr kinase domain is the signal output domain. All the phototropins discovered up to now contain two LOV domains and one kinase domain, each LOV domain binds one FMN as a chromophore.

As early as in the late 19 century, Charles Darwin carried the first delicate experiment to study the phototropic behavior of grass coleoptiles (Darwin, 1880). In the book Power of Movement in Plants, he first described the bending of grass seedling coleoptiles toward a light source. For about hundred years, this phototropic response still remains a mystery in the biochemical level. The hypothesis, that blue light-dependent protein phosphorylation was an important component in the signal perception and transduction pathway of phototropism, got strong support from the results obtained from Briggs group (Reymond et al., 1992).

Later, a phototropic mutant of Arabidopsis thaliana, strain JK224 was found (Khurana and Poff, 1989), which showed only little blue light-induced phosphorylation. Complete lack of blue light-dependent phosphorylation was observed in null mutant of NPH1 (Non-Phototropic Hypocotyl 1) locus of Arabidopsis (Liscum and Briggs, 1995). Subsequent map-based cloning of nph1 demonstrated that it encodes a 120-kD plasmamembrane-associated Ser/Thr protein kinase (Huala et al., 1997). This protein was first named as NPH1 since it is the product of nph1 gene and later changed to Phototropin because its involvement in phototropism.

Phototropin is composed of two LOV domains in the N-terminal part and one Ser/Thr kinase domain in the C-terminal part (Figure 2.1.1). The LOV domains belong to the PAS (Per-Arnt-Sim) domain superfamily. Proteins which contain LOV domains are sensitive to the change of light, oxygen and voltage (Zulin et al., 1997; Taylor and Zhulin, 1999) and are named LOV for that reason. Each LOV domain binds a flavin mononucleotide (FMN) as chromophore. One oxidized FMN is non-covalently bound to one LOV domain, the complex is called LOV^D ₄₇₇ (Crosson and Moffat, 2001; Swartz et al., 2001).

Figure 2.1.2 Photocycle of LOV1 domain from C. reinhardtii phototropin (Kottke et al 2003)

When activated by blue light, a photon is absorbed and an excited singlet state is generated, then the singlet state converts to a triplet state LOV1₇₁₅ by intersystem crossing (Swartz et al., 2001; Kennis et al., 2003; Kottke et al., 2003). There are two subspecies of LOV1₇₁₅ , namely LOV1_715a and LOV1_715b, both resembling triplet states (Kottke et al., 2003). The N5 atom of the isoalloxazine ring of the excited triplet state accept a hydrogen atom from the cysteine which is present in the highly conserved sequence NCRFLQ (Asn-Cys-Arg-Phe-Leu-Gln) in all LOV domains (Kennis et al., 2003; Corchnoy et al., 2003; Salomon et al., 2000). The protonation of FMN not only stablizes the triplet state but also increases the electrophilicity of the C(4a) atom which triggers the attack from the thiol anion. Thus, a cysteinyl-FMN adduct is formed with a absorption maximum of 390nm (LOV^S ₃₉₀) (Swartz et al., 2001; Kennis et al., 2003; Kottke et al., 2003; Salomon et al., 2000). The LOV^S ₃₉₀ cysteinyl-FMN adduct is thought to be the active signaling form of phototropin (Crosson and Moffat, 2002; Salomon et al 2001). Replacement of cysteine in NCRFLQ with either serine or alanine banishes the ability of light-dependent formation of LOV^S ₃₉₀ (Swartz et al., 2001; Kottke et al., 2003).

Arabidopsis phot1 with two cysteine to alanine mutations in NCRFLQ of both LOV1 and LOV2 lost the ability to complement the phot1 null mutant. The two LOV domains seem to play different roles. Arabidopsis phot1 with cysteine to alanine mutation in only LOV1 manages to complement the phot1 null mutant while Arabidopsis phot1 with cysteine to alanine mutation in LOV2 fails to complement the phot1 null mutant, which suggest the LOV2 domain is the more important sensor domain in phototropin (Christie et al., 2002). Similar studies about the phot2-dependent chloroplast avoidance response in fern also lead to the same conclusion (Kagawa et al., 2004).

Phototropins, although the name comes from its relationship with phototropism, are not only the sensors for phototropism. It has been shown in Arabidopsis, that they are also involved in chloroplast relocation and stomatal opening (Briggs and Christie et al., 2002). In Arabidopsis, there are two phototropins (phot1 and phot2) and both of them are involved in the functions mentioned above. As reported originally by Liscum and Briggs (1995), the phot1- mutant lost both hypocotyl and root phototropism in response to low-intensity of blue light. Phot1 is in charge of phototropism in high-intensity and low-intensity of blue light and phot2 mediates phototropism only in high-intensity of blue light (Sakai et al., 2001; Jarillo et al., 2001). In order to make best use of sun light, Arabidopsis chloroplasts have an accumulation response to maximize light capturing under low light condition so that it can get enough energy for photosynthesis. Arabidopsis choloroplasts also show an avoidance response to minimize light illumination under very strong light condition so that the photo damage is reduced. Both phot1 and phot2 are in charge of the accumulation response of chloroplasts to low intensity blue light. Phot2 is not as sensitive as phot1 in low light condition. Phot2 level is up regulated by high light condition and is in charge of the avoidance response under high light condition while phot1 still mediates the accumulation response even under the high light condition (Sakai et al., 2001; Jarillo et al., 2001; Kagawa et al., 2001).

In Chlamydomonas, a phototropin was discovered in 2001 (Huang et al., 2002). It is a single copy gene and the amino acid sequence has high homology with phototropins from higher plants. The calculated molecular weight is 81.4kD, which is much smaller compared those phototropins from higher plants (~120kD).

Due to the failure of over expressing C. reinhardtii phot1-GFP fusion protein, Huang analyzed of the localization of phot-GFP fusion proteins in transiently transformed tobacco BY-2 protoplasts, and found that the fusion product associated with endogenous membranes (Huang et al., 2002). In parallel with the present work, Huang also found that phototropin existed in both cell body and flagella. In cell body, phototropin is located in the plasma membrane and the microsomal membrane. In flagella, phototropin attaches to the axoneme and is transported in flagella by IFT (intraflagella transport) ( Huang et al., 2004).

Figure 2.2.1 Amino acid sequence of Chlamydomonas phototropin gene (Huang et al., 2002)

The caged characters stand for three domains of phototropin. The big Cs in the LOV domains indicate the cysteine residues that have been shown to form an adduct with FMN in Phot 1 of Avena sativa. The sequence within the kinase domain marked in italics has not been ovserved in other phototropins. The deduced amino acid sequence of the C. reinhardtii phototropin was compared with those from Oryza sativa (PHOT1), acc. AB018443 Oryza sativa (PHOT2), acc. AB018444 Arabidopsis thaliana (PHOT1), acc. AF030864 Arabidopsis thaliana (PHOT2), acc AF053941 Avena sativa (PHOT1a), acc. AF033096 Avena sativa (PHOT1b), acc. AF033097 Zea mays (PHOT1), acc AF033263 Pisum sativum (PHOT1), acc. U83281 Adiantum capillusveneris (PHY3), acc AB012082 Adiantum capillus-veneris (PHOT1), acc. AB037188. An asterix indicates completely conserved residues, a colon indicates a conserved residue.

As shown in Figure 2.2.2, Chlamydomonas reinhardtii is heterothallic and isogamous. The mating type (mt+ or mt-) is permanently determined in a cell line and behaves as a single Mendelian locus in crosses. The gametes of mt+ and mt- are similar in size and appearance but different in ultra structural level.

Gametogenesis always appears when adverse growing conditions are provided. Two steps are necessary for gametogenesis. The first one required is nitrogen deprivation (Sager and Granick, 1954). Gametes are quite different from vegetative cells with respect to biochemistry, sub cellular morphology and behavior. Protein synthesis continues even in fully differentiated gametes (Jones et al., 1968; Jones and Chen, 1970) although net protein synthesis ceases shortly after nitrogen deprivation (Jones et al., 1968). Both cytoplasmic ribosomes and chloroplast ribosomes are degraded (Siersma and Chiang, 1971; Martin and Goodenough, 1975; Martin et al., 1976). New ribosomes, which have different sensitivity to antibiotics and ribosomal proteins with altered electrophoretic mobility, are synthesized. The degradation of old ribosomes provides the nitrogen source for de novo synthesis of proteins and nucleic acids which are essential for gametes (Jones, 1970; Siersma and Chiang, 1971). Two new organelles appear in gametes, one is a mating structure and the other is a special type of Golgi-derived vesicle (Friedman et al. 1968; Martin and Goodenough, 1975).

Figure 2.2.2 Life cycle of Chlamydomonas reinhardtii (Huang et al., 2003)

In the life cycle of C. reinhardtii, light is the essential factor in two step of the growth and development. The first step is the swich from pregamete to gamete and the second step is zygote germination.

After removal of the nitrogen source, vegetative cells become pregametes, which are still not mating competent. Different factors have been tested and light has been shown as required for gametogenesis (Sager and Granick, 1954; Kates and Jones, 1964). Light was defined as a necessary factor for gametogenesis by the work of Christoph F. Beck's group (Treier et al., 1989). DCMU, an inhibitor for photosystem II, was found to be able to prevent gametic differentiation in acetate-free medium but had no effect when acetate was present. Since acetate was used for growing vegetative cell in darkness, light was thought to be the energy source for gametogenesis (Treier et al., 1989). It was also found that the two signals, nitrogen starvation and light input, should be applied successively in gametogenesis. Treatment with the cytoplasmic protein synthesis inhibitor anisomycin or RNA synthesis inhibitor actinomycin D can stop the switch from pregametes to gametes while these two reagents do not effect the mating efficiency of mature gametes (Treier et al., 1989). Action spectrum for the light-dependent step in gametic differentiation was taken, and in the range between 321nm and 512nm, two peaks (450nm and 370nm) were found (Weissig and Beck, 1990). This result gives the hint of a potential blue light receptor attending in gametogenesis.

It was also reported that the mating competence was only maintained in the presence of both extrinsic signals: lack of nitrogen source and light. Mature gametes lost their mating competence when they were kept in darkness (Beck and Acker, 1992). Although the light induced switch from pregametes to gametes takes around 2 hours and require cytoplasmic protein synthesis, the conversion from dark-inactivated gametes to gametes is much faster and does not require protein synthesis (Pan et al., 1997). Among three wave lengths (451nm, 573nm, 655nm) tested, blue light was most effective in restoring the mating ability of dark-inactivated gametes. In pregametes, flagellar agglutinin activity was hardly detectable, the titer of flagellar agglutinin rose during pregamete to gamete conversion induced by light and decreased in the following dark treatment. Reillumination of dark-inactivated gametes helped them to regain this ability. While the cell body agglutinin activity remained constant in the process, the activity of the gamete lytic enzyme only raised in the process of pregamete-to-gamete switch and did not change in the dark. Reillumination gave very little increase in activity (Pan et al., 1997). The flagellar agglutinin activity could be the key factor in the switch between gametes and dark-inactivated gametes. Given the localization of phototropin in flagella (Huang et al., 2004), phototropin could have relation with the light-induced regaining of flagellar agglutinin activity.

Like in our experiments, an RNAi construct was made to reduce the phototropin level in Chlamydomonas reinhardtii. Compared to wild type cells, the selected Phot1- strain had a lower gamete formation percentage in the light-induced pregametes to gametes switch. Although the Phot1- strain also had a fluence dependent response for the pregametes to gametes conversion, the percentage of competent gametes was much lower compared to wild type strain (Huang et al., 2003). The GLE (encoding gametic lytic enzyme) mRNA levels increased when wild type pregametes were irradiated. In the Phot1- strain, such an increase was not observed (Huang et al., 2003). Reillumination of dark-inactivated gametes with low intensity of light gave a clear difference in the recovery of mating competence of wild type cells and Phot1- cells, while high intensity of light could accelerate the changing in Phot1- strain (Huang et al., 2003).

It has been shown that zygote germination was controlled by light (Gleockner and Beck, 1995). Under the same duration of illumination, zygotes produced from mating of wild type cells had a higher zygote germination than zygotes that have one Phot1- parental side. Zygotes which had both Phot1- parental strains had the lowest zygote germination (Huang et al., 2003). Since those zygotes were produced from Phot1- gametes, it has not been decided that the failure to germinate was caused by less amount of phototropin in zygote or the disability of germination caused by less amount of phototropin in gametes.

Among the different stages of C. reinhardtii life cycle, non-dividing vegetative cells, pregametes and mature gametes are motile. Vegetative cells are attracted to the preferred nitrogen source--ammonium (Sjoblad and Frederikse, 1981; Ermilova, 1993; Ermilova et al., 2000). However, mature gametes do not have chemotaxis to ammonium. The change of chemotaxis is closely related to the switch between pregametes and mature gametes. Nitrogen deprivation and light are the two sequential input signal required for the switch (Sager and Granick, 1954; Ermilova et al., 2003a; 2003b). Three mutants (lrg) which can form gametes in darkness (Gleockner and Beck, 1995) also showed independence of light in losing chemotaxis behavior (Ermilova et al., 2003a). The similarity suggested the possibility of sharing common components in the signal transduction of the two process (Ermilova et al., 2004). The minimal time for nitrogen starvation was around 4 hours before light illumination was applied to induce loss of chemotaxis. It was found that the reduction of fluence resulted in slower kinetics for the loss of chemotaxis activity. Among four wave lengths with identical fluence rates tested (blue (459 nm), green (540 nm), yellow (573 nm) and red (655 nm)), only blue light was effective in inducing the loss of chemotaxis (Ermilova et al., 2004). Phot1- strains needed longer light illumination after nitrogen starvation to lose chemotaxis. For Phot1- strains, re-illumination of dark-inactivated gamete to turn off chemotaxis also took longer time than in wild type cells (Ermilova et al., 2004).

Chlamydomonas reinhardtii, a unicellular organism, is attracting more and more attention these years and is regarded as a green yeast (Goodenough, 1992; Rochaix, 1995). Compared with other known expression systems, Chlamydomanas cells many beneficial characteristics. 1.) Nuclear transformation and chloroplast transformation of Chlamydomonas are reliable and easy to carry out. Unlike higher plants, the time between the initial transformation and final scale up for production is comparatively short. 2.) Chlamydomonas can form gametes, the process of gametogenesis is easy to handle. Genetic crosses between different strains of opposite mating type can be carried out and vegetative diploid cells can be produced. 3.) Chlamydomonas can be grown either phototrophically in synchronized or unsynchronized way, or it could be also grown in darkness with acetate as carbon source. 4.) The cost of growing Chlamydomonas is cheap and to grow big volume such as 500,000 L is affordable. 5.) Foreign proteins tend to be folded correctly in Chlamydomonas (Mayfeild and Franklin, 2005).

Chlamydomonas reinhardtii has a single, cup-shaped large chloroplast which counts for 40% of the cell volume. This chloroplast contains its own genome, which is a 196kb circular genome. One chloroplast contains 80 identical copies of this chromosome. The first successful chloroplast transformation was reported in 1988 (Boynton et al., 1988), in which Chlamydomonas was transformed by DNA coated tungsten particles with a particle gun. Integration of the transforming DNA takes place by homologous recombination. Screening of chloroplast transformant is carried out either by co-transformation with DNA conferring resistance to antibiotics (Goldschmidt-Clermont and Rahire, 1991; Fischer et al., 1996), or by transformation with DNA which helps to rescue phototrophy (Boynton et al., 1988). Chloroplast gene disruption could be either homoplasmic or heteroplasmic. Homoplasmic disruption requires that all 80 copies of chloroplast genome are changed by transformation while heteroplasic disruption means that only some copies of chloroplast genome are changed. Obviously stable transformation requires all 80 copies converted to recombinant form. The latter case happens for gene with essential functions (Rochaix, 1995).

However, the way leading to overexpression of foreign genes in Chlamydomonas chloroplast was not smooth at the beginning. Early work in Chlamydomonas chloroplast transformations mainly focused on complementation of mutants deficient in photosynthetic genes (Boynton et al., 1988), and on the expression of proteins with resistance to antibiotics (Goldschmidt-Clermont and Rahire, 1991; Fischer et al., 1996). Trials to overexpress common reporter gene such as β-glucuronidase (GSU) and Renilla luciferase result in low level of protein expressions (Savaldor et al., 1993a; Savaldor et al., 1993b; Ishikura et al., 1999; Minko et al., 1999). The reason was later found to be caused by the strong codon bias of Chlamydomonas chloroplast, the possibility of A or T appearing in the third position is around 80% (Mayfield and Franklin, 2005). The chloroplast codon optimized GFP gene transformant could increased expression level up to 80-folds compared with its non-optimized counterpart. And when driven by rbcL promoter and 5'UTR, it counts for 0.5% of total soluble protein in chloroplast (Franklin et al., 2002). However, independent homoplasmic transformants could have different level up to 5 folds, which could either be caused by mutagenic effect of random integration during transformation or by 5-fluoro-2-deoxyuridine (FdUdr) which is used to decrease chloroplast genome number prior to transformation (Mayfield et al., 2003). Recently, high level of expression and assembly of a human monoclonal antibody in Chlamydomonas reinhardtii was reported (Mayfield et al., 2003), which suggest the capability of Chlamydomonas chloroplast in expressing complex biomolecules in its active form.

Compared with overexpression of recombinant protein in Chlamydomonas chloroplast, over-expression based on nuclear transformation seems more difficult and meets lots of hindlerers although expression of foreign DNA sequences in Chlamydomonas was first reported in 1982 (Rochaix and van Dillewijn, 1982). They managed to express yeast ARG4 gene in cw-15 arg-7, and it was six years earlier than the first report of Chlamydomonas chloroplast transformation. For nuclear transformation, several well established methods are available. Cell wall deficient strains can be transformed with glass bead method readily (Kindle, 1990). Walled strains can either be treated with autolysin to shed the cell wall and then the glass bead method is applied, or silicon carbide can be used in place of glass bead for walled cell transformation (Dunahay 1993). Particle guns have also been used for transforming wild type Chlamydomonas cells (Debuchy et al., 1989; Kindle et al., 1989; Mayfield et al., 1990).

At the beginning, the selectable markers used for most studies are Chlamydomonas genes which are capable of complementing mutants to wild type phenotype. The genes employed are ARG7 (Debuchy et al., 1989), NIT1 (Kindle et al., 1989), OEE1 (Mayfield et al., 1990), atpC (Smart and Selman, 1993), NIC-7 (Ferris, 1995) and THI-10 (Ferris, 1995). Recently genes which confer Chlamydomonas resistance against antibiotics are widely used for cotransformation, such as CRY1 (Nelson et al., 1994), Ble (Stevens et al., 1996), APH-VIII (Sizova et al., 2001), APH-7 (Berthold et al., 2002).

However, the use of Chlamydomonas to express foreign genes has met great difficulties and there are only few successes with several reporter genes which are routinely used in other eukaryotic systems (Day et al., 1990; Blankenship and Kindle, 1992; Kindle and Sodeinde, 1994; Sizova et al. 1996; Stevens et al., 1996). The reason for such phenomenon could be certain characteristics of Chlamydomonas. During transformation, foreigh DNA randomly integrated into Chlamydomanas genome (Rochaix, 1995). Only those DNA fragments integrated into transcriptional active regions would get good expression. Similar to its chloroplast genome, Chlamydomonas nuclear genome also has a strong codon bias. It has high GC content (62%) and prefers to have G or C in the third position of the codon (Harris, 1989; Silflow, 1998). A codon optimized GFP gene got nice expression after transformation (Fuhrmannet al., 1999) in Chlamydomonas.

Another feature of Chlamydomonas gene is that most contain introns (Harris, 1989). It is possible to rescue metabolic deficient mutants with cDNA while the expression levels of cDNA are much lower than their genomic counterparts (Diener et al., 1993; Auchincloss et al., 1999; Perron et al., 1999; Boudreau et al., 2000). It was suggested that introns could take part in the regulation of transcription and it was also shown that insertion of the first intron of rbcS2 gene could increase the strength of rbcS2 promoter up to 30 folds. Intron 1 of rbcS2 gene functions as an enhancer and its enhancer function is independent of its position and orientation (Lumbreras, 1998).

Another problem of nuclear expression is that Chlamydomonas does not have very strong promoters. The rbcS2 promoter has been widely used and is regarded as the strongest promoter for heterologous gene expression (Goldschmidt-Clermont and Rahire, 1986; Berthold et al., 2002; Fuhrmann et al., 1999). Recently it was found that the strength of the rbcS2 promoter could be improved by placing the HSP70A promoter in front of it (Schroda et al., 2000). Thus the HSP70A promoter combined with rbcS2 promoter plus rbcS2 intron 1 is currently the most powerful promoter available for expression in Chlamydomonas.

The PsaD gene encodes a 20kD subunit of photosystem I (Zilber and Malkin, 1988; Chitnis et al., 1989). This gene contains no introns, which suggest that all the regulatory sequences leading to high expression are located in the flanking regions. The expression of the ble gene driven by the PsaD promoter is almost similar to the expression driven by HSP70A plus rbcS2 promoter with rbcS2 intron1. It is regarded as a very powerful tool in expressing recombinant cDNA in Chlamydomonas nucleus (Fischer and Rochaix, 2001).

RNAi (RNA interference) is a powerful tool for gene silencing in all model systems currently studied by biochemists. Fire et al., (1998) discovered it while investigating the use of antisense and sense RNA for gene inhibition in the nematode worm Caenorhabditis elegans. Later, RNAi was found to be a universal defense mechanism for eukaryotic cells against RNA viruses or proliferation of transposable elements that replicate via RNA intermediates. In a wide variety of organisms, such as fungi, plants, worms, mice and probably humans, RNA interference can be applied as useful technique to turn off the expression of individual cellular genes.

The most important factor in RNA interference is a so called short interfering RNA (siRNA). siRNAs are 21-23nt dsRNA duplexes with symmetric 2-3 nt 3'overhangs. The 5'end of siRNA normally carries a phosphate, while 3'end does not. siRNAs are created by Dicer, a member of RNase III family. Dicer uses the energy of hydrolysis of ATP and digests dsRNA into siRNA (Bernstein et al., 2000; Ketting et al., 2001; Billy et al., 2001). This process takes place in cytoplasm (Hutvagner and Zamore, 2002; Zeng and Cullen, 2002; Kawasaki and Taira, 2003).

Figure 2.4.1 The RNAi pathway. (Dykxhoorn et al 2003)

(A) Short interfering RNA (siRNA). An siRNA is composed of two 21-nt ssRNA with a 19-nt duplex region. The 5'-ends of both strand are phosphorylated. (B) The siRNA pathway. Long double-stranded RNA (dsRNA) is digested by Dicer in an ATP-dependant way. Then siRNAs are uptaken by RISC. ATPs are hydrolysed to help unwind of siRNA but the incorporation is ATP independant. The single-stranded antisense strand help RISC to find the target mRNA, and the mRNA is cleaved in the middle of the duplex region, 10 nt from the 5'end of the siRNA. (C) The micro (mi) RNA pathway. Dicer also cleave the ~70-nt imperfect hairpin RNA and produce ~22nt miRNA, those miRNA would bind to the 3'untraslated region (UTR) of their target mRNA and block translation.

siRNAs then bind to a protein complex called RISC (RNA-inducing silencing complex). 5'-phosphate of siRNA is strictly required for binding and those siRNA which lack 5'-phosphate will be quickly phosphorylated and incorporated into RISC (Nykanen et al., 2001; Schwarz et al., 2002). The two strands of siRNA separate and one strand continues binding to RISC. The single-stranded RNA which remains in the complex works as a guide to its homologous target mRNA for endonucleolytic cleavage. RISC will cut mRNA the in the center of the duplex region formed between the guide siRNA and target mRNA, thus the mRNA is degraded (Elbashir et al., 2001).

In mammals, endogenously expressed siRNA has not been discovered but a similar miRNA (micro-interfering RNA) pathway has been found in many organisms and cell types (Pasquinelli, 2002). miRNA is also generated by Dicer, but in contrast to miRNA, those short RNA species (~22nt) are produced from precursors about ~70nt with imperfect hairpin. They are single-stranded. miRNAs will recognize homologous regions in target mRNA and thus prevent translation.

Many viruses have RNA dependent RNA polymerase (RdRps) to replicate their RNA genomes. Many eukaryotes also encode RdRps and use them in a sequence specific, RNA-triggered gene silencing mechanism (Ahlquist, 2002). RdRps make RNAi a self-amplifying process. RdRps can use single-stranded siRNA as primer and target mRNA, degraded target mRNA or melted dsRNA as template to form new dsRNA. These dsRNAs would enter the siRNA pathway again.

There are many different ways to trigger the RNA interfering process. The way to generate RNAi can be divided into two major groups, one is to use RNAs that are prepared in vitro, the other is to use RNA which is generated in vivo. The first group includes using chemically synthesized siRNA, using long dsRNA, using siRNA-based hairpin RNA or using miRNA-based hairpin RNA. The RNAs would be injected into the host to silence a specific gene. The RNAs prepared in vitro only last for a limited period of time. The second group is based on transforming the host with genes that would form double-stranded RNA or hairpin RNA after transcription. Promoter either for RNA polymerase II or for RNA polymerase III are used, the sense and anti-sense chains are either under the control of same promoter or arranged in a tandem way and the two single-stranded RNAs will form pairs (Dykxhoorn et al., 2003).

In Chlamydomonas, the AR promoter (HSP70A plus rbcS2) is used preferentially to get strong transcription. A genomic piece of target gene is taken and the reverse cDNA counterpart is placed behind. A 3'UTR is placed behind the RNAi construct in some cases (Huang et al., 2002), but construct without 3'UTR also seems to work fine (Fuhrmann et al., 2001).

Figure 2.4.2 Self-amplifying process of RNAi and transitive RNA siliencing (Ahlquist, 2002).

(A) Degradative and synthetic pathway connecting dsRNA, siRNA and mRNA. The black arrows outline the classical degradative pathways in RNAi process and yellow arrows denote the synthetic pathways. RdRps may act on siRNA-primed dsRNA(1), siRNA-primed mRNA(2) or asRNA-primed mRNA. The product, dsRNA, will re-enter the RNAi process to generate more siRNA. (B) Transitive RNA silencing. When there is intermediary mRNA AB or another mRNA which contains high homologous region of gene B, the RNAi reaction which targeted gene B would also cause the silencing of gene A.

Figure 2.4.3 Methods to generate RNAs that silence gene expression (Dykxhoorn et al., 2003).

(A) Silencing by RNAs generated in vitro. Aa, Chemically synthesized siRNA which could function in absence of Dicer. Ab, Long dsRNA enters the RNAi process directly. Ac, Perfect duplex hairpin RNA induces siRNA pathway. Ad, Imperfect duplex hairpin RNA induces miRNA pathway. (B) Silencing by RNAs generated in vivo. Ba, RNA polymerase II promoter was used and perfectly symmetric sense and anti-sense chains are under the control of same promoter. Long hairpin RNA will form after transcription. Bb, Sense and anti-sense strands are under the control of tandem RNA polymerase III promoter. Bc, Single RNA polymerase III promoter is used to control the transcription of siRNA-based hairpin RNA. Bd, Single RNA polymerase II promoter is used to control the transcription of the miRNA-based hairpin RNA.

As described in Chapter 2.2, light, especially blue light, is of great importance in the sexual life of C. reinhardtii. But, it is impossible to study the involvement of phototropin in zygote germination separately from gametogenesis and maintenance of mating competence in the way described in Huang et al., (2003).

The common way to study zygote germination is very complicated and few labs are able to carry it out. First, gametes of both mating types should be generated according to the standard protocols. Then the gametes are mixed and mate. After two hours, the cells are diluted and plated on TAP plates with 4% agar and those plates are illuminated for 24 hours and then transferred to the dark for 5 days. The vegetative cells and unmated gametes are then removed with a razor blade, leaving the hard-walled zygospores adherent to the agar. The plates are then treated with chloroform vapor for 30 seconds (Harris, 1989). To study the relationship between light and zygote germination, those two steps should be carried in darkness. Then the plates are placed under light for 1-5 hours and transferred into darkness again. 11 days later, the numbers of germinated and non-germinated zygotes are counted under a dissecting microscope (Huang et al., 2003). There are several drawbacks of this method. 1.) The procedure is difficult to handle. 2.) Since there are many steps, manipulation on each plate can hardly be identical, which may cause errors in the final result. 3.)Vegetative cells and unmated gametes can hardly be removed totally. 4.) This method assumes that the only difference between zygotes produced from wild type cells and the zygotes produced from Phot1- strains was the level of phototropin. Given the involvement of phototropin in gametogenesis and maintenance of mating competence, this point was questionable.

Thus, a new mating assay was required to study phototropin and zygote germination. The test should be easily carried out, involve less steps and be able to distinguish zygotes from vegetative and unmated gamete more efficiently. In this thesis, a new mating assay was established which fulfilled all requirements.