|Home | About | Journals | Submit | Contact Us | Français|
Chlamydomonas reinhardtii strains lacking phytoene synthase, the first enzyme of carotenoid biosynthesis, are white. They lack carotenoid pigments, have very low levels of chlorophyll, and can grow only heterotrophically in the dark. Our electron and fluorescence microscopic studies showed that such a mutant strain (lts1-204) had a proliferated plastid envelope membrane but no stacks of thylakoid membranes within the plastid. It accumulated cytoplasmic compartments that appeared to be autophagous vacuoles filled with membranous material. The lts1 mutants apparently lacked pyrenoid bodies, which normally house ribulose bisphosphate carboxylase–oxygenase (Rubisco), and accumulated many starch granules. Although these mutant strains cannot synthesize the carotenoid and carotenoid-derived pigments present in the phototactic organelle (eyespot), the mutant we examined made a vestigial eyespot that was disorganized and often mislocalized to the posterior end of the cell. The absence of a pyrenoid body, the accumulation of starch, and the disorganization of the eyespot may all result from the absence of thylakoids. The ultrastructure of lts1 mutant strains is similar to but distinct from that of previously described white and yellow mutant strains of C. reinhardtii and is similar to that of naturally colorless algae of the Polytoma group.
To facilitate localization of a fusion between an integral membrane protein of Chlamydonomas reinhardtii and the green fluorescent protein, we used an lts1 (light sensitive 1) mutant strain, which has a lesion in the phytoene synthase (PSY1) gene and is white (McCarthy et al., 2004). This led us to examine the structure of two lts1 mutant strains by fluorescence microscopy and the ultrastructure of one of them by electron microscopy. Phytoene synthase is the first enzyme of carotenoid biosynthesis, and lts1 mutant strains lack carotenoids and have very little chlorophyll. Carotenoids are known to be required for the accumulation of photosynthetic reaction centers, which are central components of thylakoid membranes (Herrin et al., 1992). We here show that the plastids of an lts1 null mutant lack stacks of thylakoid membranes and are aberrant in a number of other ways that appear to be related to the lack of thylakoids. Their defects are more extreme than those that have been reported for other pigment mutants.
Classic studies of mutant strains of C. reinhardtii with defects in pigment synthesis or accumulation go back to Sager and Palade (1954) and Sager and Zalokar (1958). They revealed the intimate connections between photosynthetic pigments and stable assembly of thylakoid membranes (Ohad et al., 1967a, 1967b; Wang et al., 1974, 1975; Wilson et al., 1980). Although Sager and Zalokar (1958) found that their white mutant, No. 95, lacked carotenoids, this strain is not available from the Chlamydomonas Genetic Stock Center and was not studied further. The white mutant of Wilson et al. (1980), U3A, was clearly different from that of Sager and Zalokar (1958) and those of other early workers because it and the yellow mutant U3N both derived from the unstable U3 strain. U3A was lost. Three of the early described white mutants of Chlamydomonas, lts1-30 (Chemerilova, 1978; Iroshnikova et al., 1982; Ladygin et al., 1982), fn68 (Foster et al., 1984), and w7 (Spreitzer and Mets, 1981), are still available for study. As shown by McCarthy et al. (2004), all have lesions in the PSY1 gene and augment a collection of recently isolated strains with lts1 lesions. Given their block at the beginning of carotenoid synthesis and the availability of sequenced null alleles in a defined background (McCarthy et al., 2004), lts1 strains are a unique resource for studying effects of the absence of carotenoids on ultrastructure. Early reports showed that these strains had a defect in thylakoid accumulation (Sager and Zalokar, 1958; Ladygin et al., 1982), and a phytoene synthase mutant of the chlorophyte Scenedesmus obliquus, C6E, lacked stacked thylakoids (Wellburn et al., 1980; Sandmann et al., 1997).
It has been noted that pigment mutants of C. reinhardtii have a strong resemblance to colorless algae of the Polytoma group (Lang, 1963; Siu et al., 1976). These are organisms with a single plastid (leucoplast) that lacks stacked thylakoids and it was hypothesized that they arose from a Chlamydomonas-like ancestor. Recently, this has been corroborated phylogenetically and it has been shown that Polytoma strains originated from Chlamydomonas-like ancestors at least twice and hence are polyphyletic (Rumpf et al., 1996).
The lts1-204 strain, which we used for most of our studies, is in the background of parental strain 4A+, a wild-type strain that grows well in the dark (McCarthy et al., 2004). Although it was obtained from a culture mutagenized with ethylmethane sulfonate, lts1-204 carries an insertion of the TOC1 transposon (Day and Rochaix, 1991) in exon 1 of the PSY1 gene (Figure 1) at a position with features similar to those described previously (Kim et al., 2006). The transposon is inserted between the codons for amino acid residues 91 and 92 of phytoene synthase, which should result in a null phenotype. Indeed, the lts1-204 strain had the lowest levels of α-tocopherol and chlorophyll a of the 12 lts1 strains characterized by McCarthy et al. (2004).
The lts1-204 strain grew almost as well as parental strain 4A+ on acetate in the dark (doubling time of 16 versus 13h) (Figure 2), but it failed to grow in even dim light (McCarthy et al., 2004). Its cell dimensions and shape appeared normal. Likewise, it appeared to swim normally and have a normal overall level of activity when observed at low magnification (100× and less). However, in agreement with the studies of Foster et al. (1984), Hegemann et al. (1991), and Lawson and Satir (1994), it showed neither positive nor negative phototaxis.
The cell wall of lts1-204 had the same appearance as that of parental strain 4A+, a multi-layered structure of 100–200-nm thickness (Figures 3–5). Likewise, the cytoplasmic membrane of the mutant had a typical thickness (~100Å) and appearance: two dark layers with a lighter layer between. Flagella appeared normal (not shown) but were not examined in detail, since the mutant cells appeared to swim normally. The nucleus was of typical shape and size (2–3μm in diameter; Figures 3 and and4).4). It was surrounded by twin membranes perforated by nuclear pores and a darker staining nucleolus was often observed within it. Although the morphology of mitochondria was grossly normal, the mutant had more mitochondria than parental strain 4A+, they were larger (see the particularly large mitochondrion of 3μm in length at the right of Figure 5A), and they had more elaborate internal structure. The cytoplasm of lts1-204 contained very large numbers of ribosomes, as did that of 4A+. In both cases, some were associated with the rough endoplasmic reticulum (Figure 6A and not shown) and, in both cases, there were membranes and vesicles of the Golgi apparatus between the nucleus and the plastid (Figures 3 and and44).
Vacuoles of the 4A+ cell were variable in size and number, and their contents were lightly stained (Figures 3A and and4A).4A). The lts1-204 strain also had a number of vacuoles in the cytoplasm, some of which were nearly as large as the nucleus (Figures 3B, ,4B,4B, and and5C).5C). They often contained membranous or dark granular material (Figures 3B, ,4B,4B, and and5C)5C) that resembled the material of the wild-type pyrenoid body (Figure 4A) and appeared to be autophagous vacuoles (Wilson et al., 1980). The electron dense object in the giant vacuole in Figure 4B is about 1μm in diameter.
The plastid envelope of lts1-204 mutant cells appeared typical. It was composed of an outer membrane that resembled a slightly thinner version of the cytoplasmic membrane and an inner membrane that appeared as a single dark line about 50Å thick (but see below regarding its collapse and proliferation). However, there were no stacks of thylakoid membranes in the plastid in any sections of the approximately 1000 cells of lts1-204 that we examined, whereas such stacks were conspicuous in plastids of parental strain 4A+ (Figures 3–6). Occasionally, remnants of lamellae of the thylakoid membrane were seen in strain lts1-204 (Figures 4C, ,5B,5B, and and5E;5E; best shown in Figure 5B). For both parental and mutant strain, ribosomes within the matrix of the plastid were slightly smaller than cytoplasmic ribosomes, which allowed us to distinguish between different compartments in the complex structure of lts1-204 mutant cells.
The lts1-204 mutant strain appeared to lack pyrenoid bodies (Figures 4 and and5).5). These are normally spherical with a diameter of 1.5–2.0μm and appear to be composed of granular, tightly packed material surrounded by starch plates (Sager and Palade, 1957). Rawat et al. (1996) described the pyrenoid body as an ‘anastomosing tubular membranous network all embedded in a matrix of granular material’ whose electron density varied with growth conditions but whose membrane network was a constant indication of pyrenoid presence. As reported previously (Rawat et al., 1996) and predicted from the relative dimensions of pyrenoid and cell, we observed a pyrenoid body in about 20% of the median sections of parental strain 4A+ (e.g. Figure 4A). We found no pyrenoids with this appearance in the approximately 1000 cells of lts1-204 that we examined. Because lts1-204 lacked thylakoid membranes, we relied on electron density and location to search for pyrenoid bodies. The only two candidates we found, based on location and association with starch granules, are shown in Figure 5E and 5G. The pyrenoid-like feature in Figure 5E has membrane remnants that resemble pyrenoid tubules.
The portion of the wild-type eyespot visible by electron microscopy, which is required for efficient phototaxis (Dieckmann, 2003), is composed of ordered rows of granules of similar size and texture (Figure 3A). The granules lie immediately under the chloroplast envelope at a point where the envelope is directly apposed to the cytoplasmic membrane and are found equatorially at the base of the flagellar root (Melkonian and Robenek, 1980). Each layer of granules is subtended by a thylakoid membrane (Dieckmann, 2003). The eyespot of lts1-204 was disordered (Figures 3–5). The granules appeared to differ in size, often had lightly stained centers, and usually appeared in only a single, jagged row. Although the eyespot lay immediately under the chloroplast envelope membrane, as it does in wild-type cells, there was no associated underlying membrane structure. Nevertheless, the space beneath the eyespot was only lightly stained, as it was in wild-type cells. The eyespot was sometimes found equatorially in what appeared to be its normal location (Figures 3B and and5B),5B), but it was often found at the posterior end of the cell (Figures 4B, 5A, 5C, 5E, and 5G). Of 15 cases in which eyespots of lts1-204 cells were visible, three were correctly placed, three were indeterminate in location, and nine were mislocalized to the posterior end of the cell.
Starch granules in the parental 4A+ strain were usually small, regular in shape, and sandwiched between thylakoid membranes (Figures 3A and and4A).4A). By contrast, starch granules in the lts1-204 mutant were sometimes very large, irregular in shape, and often numerous (Figures 4B, 4C, and 5B–5D). They appeared to occupy a greater proportion of plastid and cell volume than in parental cells (e.g. the large granule in Figure 5C appears to occupy ~8% of total cell volume) and, in one instance, we observed one granule apparently nested within another (Figure 5D).
Although the plastid envelope of lts1-204 was normal in appearance, two envelopes were often apposed to one another as if the plastid between them had collapsed. These pairs of membranes folded back on themselves in sheets or whorls (Figures 3B, 3D, ,5A,5A, and and6),6), which were presumably produced by pleating or rolling, respectively. In some cases, the apposed membranes folded around starch granules and vacuoles. The alternation of cytoplasmic and plastid ribosomes between layers of envelope allowed us to distinguish between the cytoplasm and the chloroplast stroma and thereby to distinguish between folded chloroplast envelope membrane (Figure 3D) and stacked thylakoids (Figure 3C). To determine the precise shape of the plastid envelope of lts1-204 will require three-dimensional reconstruction.
Due to the symmetries of the wild-type chloroplast (Figure 4A), the lengths of chloroplast envelope observed in any median section were relatively constant. This length was approximately 75% greater than the length of the cytoplasmic membrane. However, the length of plastid envelope observed in lts1-204 cells was highly variable in median sections. In 11 of 32 median sections of whole cells, the chloroplast envelope was folded back on itself, as described above. The length of plastid envelope in these sections was 50–100% greater than the length of envelope in a typical wild-type cell. In the remaining 21 median sections, the overall length of envelope was normal, but the length of envelope at the anterior end of cells was often reduced.
Unlike the case for strain 4A+, in which ribulose bisphosphate carboxylase–oxygenase (Rubisco) was concentrated in a 1-μm spot at the posterior end of the cell (green staining in Figure 7A and 7C), Rubisco in mutant strains lts1-204 and lts1-203 was distributed throughout the chloroplast stroma around the starch granules (green staining in Figure 7D and 7F). Rubisco was localized in <0.15% of lts1 mutant cells examined, whereas it was localized in 80% of 4A+ cells (Table 1). The diffuse localization of Rubisco in lts1 mutant strains was coincident with their weak chlorophyll autofluorescence (not shown). The lts1 mutant cells contained many large starch granules (Figure 7D–7F).
Herrin and colleagues had established that an lts1 strain (lts1-202=fn68=fl) did not accumulate photosynthetic reaction centers, which contain carotenoids. We confirmed that lts1-204 lacked the D1 protein, an integral component of the PSII core complex in the thylakoid membrane (Figure 8 and Table 2). The lts1-204 strain apparently had less cytochrome f and the Rieske iron–sulfur protein, soluble components of the cytochrome b6f complex, than 4A+. The cytochrome b6f complex also contains a molecule of carotenoid (Zhang et al., 1999; Choquet and Vallon, 2000). As Herrin et al. (1992) had observed for lts1-202, lts1-204 retained the soluble α and β components of the thylakoid membrane ATP synthase. It also had apparently normal or perhaps elevated amounts of the integral membrane proteolipid component of the ATP synthase (Figure 8 and not shown). Although antibodies to the α and β components can cross-react with the mitochondrial ATPase, the proteolipid component is unique to the chloroplast enzyme. Where thylakoid membrane proteins are located in lts1-204 is not clear.
Electron microscopic studies showed that the plastid envelope of the lts1-204 mutant strain was highly contorted and that the plastid lacked stacks of internal thylakoid membranes. Both lts1-204 and lts1-203 lacked pyrenoid bodies and accumulated very large amounts of starch. Despite the fact that lts1-204 lacks carotenoids, which fill the granules of the eyespot, and retinal, the carotenoid-derived pigment for the photoreceptors in the overlying plasma membrane, the lts1-204 strain retained a vestigial eyespot. It was disorganized and, though sometimes found equatorially, was usually found at the posterior end of the cell. To our knowledge, a posterior location is rare (see below). Thylakoid membranes are an essential component of the pyrenoid body, normally subtend the starch plates that surround the pyrenoid and other starch granules, and serve to localize the granules of the eyespot in regular rows just underneath the chloroplast envelope membrane. Hence, we postulate that the lack of thylakoids accounts for several of the other ultrastructural abnormalities of lts1 mutant strains. Each abnormality is elaborated briefly below. The large numbers of membrane-containing vacuoles in lts1-204, which are probably autophagous vacuoles, may arise as a consequence of aborted synthesis of thylakoids and the attempt to recycle their components. Given the degree to which the ultrastructure of lts1 mutant strains is aberrant, lts1-204 grew remarkably well in the dark.
Previous biochemical studies of lts1 mutant strains provide a possible rationale for why they lack thylakoid membranes. The major protein components of thylakoids are the photosynthetic reaction centers, and carotenoids are essential components of both the core and light-harvesting complexes that constitute these centers. Using primarily lts1-202 (fn68=f1), Herrin et al. (1992) showed that photosynthetic reaction centers fail to accumulate in the absence of carotenoids and that chlorophyll is degraded as a consequence of this. The same was apparently true of lts1-204. In the absence of stable photosynthetic reaction centers, thylakoid synthesis may be aborted. Whether proliferation of the chloroplast envelope membrane of lts1 strains (Ladygin et al., 1982; this work), a phenomenon that has seldom been described (Wilson et al., 1980), is a consequence of aborted synthesis of thylakoids remains to be determined. Likewise, it remains to be determined whether components of the thylakoid membrane ATPase and cytochrome b6f, which continue to accumulate in lts1 strains, reside in this proliferated envelope.
The pyrenoid body, which normally houses Rubisco in a network of thylakoid-derived membranes, is found in a wide variety of algae and in terrestrial plants of the hornwort group (Gibbs, 1962a, 1962b; Ohad et al., 1967a; Vaughn et al., 1992; Rawat et al., 1996; Borkhsenious et al., 1998). Like the chloroplast, the pyrenoid reproduces by fission during cell division (Goodenough, 1970). In the absence of thylakoids, lts1-204 is apparently unable to form a pyrenoid body, and Rubisco assumes a diffuse localization throughout the plastid in both lts1-204 and lts1-203. Rawat et al. (1996) have shown that the physical presence of Rubisco is also necessary for the formation of the pyrenoid of C. reinhardtii but that neither CO2 fixation nor an active Rubisco is required. Other strains that lack both Rubisco and other chloroplast proteins also lack pyrenoids (Togasaki and Levine, 1970; Goodenough and Levine, 1970) and strains that are deficient in chlorophyll and have a severe reduction in the internal membrane structure of the chloroplast have only poorly defined pyrenoids (Wang et al., 1974, 1975). Finally, strains with a lesion in the ccm1 gene (also called cia5), which codes for a primary regulatory protein controlling the carbon concentrating mechanism, fail to form pyrenoid bodies (Moroney et al., 1989; Fukuzawa et al., 2001; Xiang et al., 2001).
The eyespot, a remarkable phototactic organelle, has components in both the cytoplasmic membrane and the chloroplast (Dieckmann, 2003). As in the case of magnetosomes, which can be assembled in the absence of magnetite (Komeili et al., 2004), vestigial eyespot granules that appear to be of irregular size and shape can be formed in the absence of carotenoids (Lawson and Satir, 1994; this work). Moreover, when all-trans retinal was provided exogenously to strain fn68, which cannot synthesize carotenoids, both positive and negative phototaxis and the photophobic response could be reconstituted (Foster et al., 1984; Hegemann et al., 1991; Takahashi et al., 1992; Zacks et al., 1993; Lawson and Satir, 1994). The disorganization of the granules in the eyespot of lts1-204 (also observed in fn68 but not remarked on (Lawson and Satir, 1994)) is likely to be a consequence of the lack of thylakoids, which normally subtend them. In wild-type, the granules are of uniform size and hexagonally packed. The granules may also be of uniform size in lts1-204 and their disorganization may be accounted for by their not being ordered in a row. Alternatively, they may actually be of different sizes. Unlike their disorganization, the frequent mislocalization of the eyespot granules in lts1-204 to the posterior end of the cell is more difficult to understand. To our knowledge, the latter location has been reported previously only when microtubules were disassembled with colchicine (Walne, 1967) and in the U3A mutant (Figure 3 of Wilson et al., 1980; not remarked on). It will be of interest to determine whether the rhodopsin photoreceptors themselves, which normally lie in the cytoplasmic membrane above the eyespot granules, are also mislocalized to the posterior end of the cell in lts1 mutant strains. Disorganized eyespots have been seen in strains with defects other than those in carotenoid biosynthesis, as have eyespots mislocalized close to the region of the flagellar basal body (Lamb et al., 1999; Dieckmann, 2003).
Starch appeared in place of glycogen in photosynthetic eukaryotes (Ball, 1998). Over-accumulation of starch has been seen in paralyzed mutants of C. reinhardtii (Hamilton et al., 1992) and profound over-accumulation has been seen previously in pigment mutants (Ohad et al., 1967a for y-1; Wang et al., 1974 for brs-1; Wilson et al., 1980 for U3A). Colorless algae of the Polytoma group also accumulate very large amounts of starch (Lang, 1963; Siu et al., 1976). Since our parental strain 4A+ has fewer, smaller starch granules than lts1-203 or lts1-204 when grown in the dark, over-accumulation in the lts1 mutant strains does not appear to have a metabolic cause. Enlargement of starch granules may be due to the absence of physical constraints on their size in the absence of stacked thylakoid membranes or may represent a storage form for materials normally found in stacked thylakoids. Why there would be increased nucleation of starch granules in the absence of thylakoids is less clear. Starch is not required for viability of C. reinhardtii (Zabawinski et al., 2001) (or for the synthesis of flagella and fertility (Steven G. Ball, personal communication)). It will be of interest to see whether starch accumulation is essential in the lts1 background.
Like lts1 strains, other white and yellow mutant strains of C. reinhardtii lack or have decreased amounts of stacked thylakoids, have disorganized eyespots, and, as noted above, often accumulate large amounts of starch (Sager and Palade, 1954; Ohad et al., 1967a; Wang et al., 1974; Wilson et al., 1980). Presumably, the basis for these defects is the same as that which we have presented for the lts1 strains: when synthesis of carotenoids or chlorophyll is curtailed, reaction centers cannot accumulate and synthesis of stacked thylakoids cannot be completed. However, unlike lts1 null strains, the other well studied pigment mutants usually have a pyrenoid body or some vestige of it. Thus, assembly of the thylakoid membranes of the pyrenoid may somehow be privileged (Uniacke and Zerges, 2007). In our hands, the yellow-in-the-dark strain y-1 lacked a pyrenoid body when grown in the dark (Rubisco detection assessed by fluorescence microscopy; C. Yoshihara, unpublished) and Sager and Palade (1954) reported the same. However, when y-1 was grown in continuous culture in the dark (Ohad et al., 1967a), it had a pyrenoid body. None of the white and yellow mutant strains described previously had the apposed plastid membranes seen in lts1-204, which appeared to fold back on themselves by rolling or pleating.
Apart from their disorganization, such as the presence of autophagous vacuoles and folded chloroplast envelope membrane, cells of lts1-204 show a remarkable resemblance to colorless algae of the Polytoma group (Lang, 1963; Siu et al., 1976). Like the lts1 strain, Polytoma strains lack a pyrenoid body and accumulate a great deal of starch. Like lts1-204, at least one member of the Polytoma group retains an eyespot (stigma; Lang, 1963). It is tempting to speculate that phytoene synthase mutations may have been among those that initiated the origin of Polytoma species from their Chlamydomonas-like ancestors (Rumpf et al., 1996).
Wild-type strain 4A+ and the lts1-203 and lts1-204 mutants were obtained from K.K. Niyogi (University of California, Berkeley). They were maintained at 25°C on Tris-Acetate-Phosphate (TAP) agar medium (Harris, 1989, pp. 25–29) containing NH4Cl (10mM) as nitrogen source. The lts1 strains were maintained in complete darkness. The lts1-203 lesion is a missense mutation (McCarthy et al., 2004) and is in the background of parental strain 4A+, which grows well in the dark (Soupene et al., 2004). The lts1-202 lesion (fn68; Foster et al., 1984) is a frameshift mutation in a different background (McCarthy et al., 2004) and the lts1-30 lesion (Chemerilova, 1978) is a nonsense mutation at position 125 (McCarthy et al., 2004), also in a different background.
Cultures were grown in the dark at 25°C on TAP medium with 10mM NH4Cl as nitrogen source and were bubbled with air for growth curves or were shaken in flasks for electron microscopy.
Cultures were grown, DNA was isolated, and PCR was performed as described in Kim et al. (2005). The primers used are listed in Table 3. The DNA sequence of the PSY1 gene was obtained from JGI version 3.0 of the Chlamydomonas genome (Scaffold_62:413352–418669 Model estExt_GenewiseH_1.c_620008; AY604701).
Cells in early exponential phase that were swimming vigorously were placed on a concave slide and observed with a Leica MZ 12 5 compound dissecting microscope. As reported by Foster and Smyth (1980), cells of 4A+ swam down or up, depending on the intensity of illumination from beneath, and right or left, depending on the intensity of illumination from the side. Cells of lts1-204 were unresponsive to any light intensity, whether lighted from beneath or the side.
Cultures of 4A+ and lts1-204 were grown to mid-exponential phase (5×106cellsml−1 by cell count), concentrated by centrifugation, and fixed overnight with 2% glutaraldehyde in 0.1M sodium cacodylate buffer, pH7.2. Specimens were washed and post-fixed with a sodium cacodylate buffer containing 1% osmium tetroxide and 0.8% potassium ferricyanide. After rinses, specimens were treated with 1.0% uranyl acetate and dehydrated with increasing concentrations of acetone. They were then infiltrated and embedded in resin. Sections were cut to approximately 500μm in diameter and 60nm in thickness. Therefore, 100 sections were required to traverse a cell along its short axis (~6μm).
Sections were collected on Maxtaform copper slot grids (2×1-mm oval hole) that had been coated with 0.5% formvar. Sections were post-stained with uranyl acetate and lead citrate. Some sections were also stained with tannic acid to visualize starch granules. They were examined using an FEI Tecnai 12 Transmission Electron Microscope.
We examined >500 cells of strain 4A+ and >1000 cells of lts1-204 in multiple sections. We photographed and studied in detail 13 cells of 4A+ and 45 cells of lts1-204.
Anti-Rubisco antibody was obtained from P. Schurmann (Université de Neuchâtel) through B. Buchanan (University of California, Berkeley). Cultures were grown to mid-exponential phase as described above, collected by centrifugation at low speed, and suspended in PME buffer (50mM Pipes, 5mM EGTA, 2mM MgSO4, pH6.2). They were placed on a multi-well slide (ICN Biomedicals, Inc.) that had been treated with poly-L-lysine (Sigma), incubated for 20min, and then fixed with 4% formaldehyde (Polyscience Inc.) in PME for 45min. After three washes with PME, cells were permeabilized with 0.5% Triton-X100 for 10min and immersed in methanol at –20°C to extract chlorophyll. The slide was dried and rehydrated with Phosphate-buffered saline (PBS) for 5min. Cells were blocked with 2% BSA in PBS for 15min and incubated overnight at 4°C in a 1:100 dilution of anti-Rubisco antibody in BSA/PBS. The slide was then washed 10 times with PBS and incubated with goat anti-rabbit secondary antibody conjugated to FITC (Molecular Probes) in 2% BSA/PBS for 3h at room temperature in the dark. After eight washes with PBS, a drop of Slow Fade@antifade gold (Molecular Probes) was added before slides were covered.
All images were acquired with a Zeiss 510 Meta confocal laser scanning microscope. FITC was excited using a 488-nm Argon laser with a 505–550-nm barrier filter before the photo multiplier tube (PMT) detector. Differential interference contrast (DIC) images were obtained using a 488-nm source and the transmitted light detector. Images were analyzed using Zeiss 510 meta software.
Strains were grown in flat 1-l bottles containing 700ml of TAP-NH4Cl medium in darkness and bubbled with air (0.035% vol./vol. CO2). Cell densities were measured with a hemocytometer and were matched for different samples. A 10-ml portion of cells was harvested by centrifugation at 8000g for 5min at 4°C. Cells were suspended in NuPAGE LDS sample buffer (Invitrogen), and lysates were boiled for 1min. Cell extracts (~4μg protein, measured before samples were put in lysis buffer) were subjected to SDS–PAGE (NuPAGE 4–12% Bis-Tris gel; Invitrogen) and proteins were transferred to a nitrocellulose membrane. Antibodies to D1, the Rieske Fe–S protein (antibody 1), the α subunit of the chloroplast membrane ATPase (CF1α), and cytochrome f (antibody 1) were a generous gift of Dr Sungsoon Park (laboratory of Dr Anastasios Melis, Department of Plant and Microbial Biology, University of California, Berkeley). All were affinity-purified rabbit antibodies. The first three were made to portions of the respective proteins from Arabidopsis that had been expressed in E. coli. The cytochrome f antibody, originally obtained from Dr Paraq Chitnis, was raised against a cyanobacterial protein. Antibodies to the Rieske Fe–S protein (antibody 2), cytochrome f (antibody 2), the β subunit of the ATPase (CF1β), and an integral membrane component of the ATPase (CF0-suIII) were kindly provided by Dr Olivier Vallon (CNRS/Université Paris 6, Institut de Biologie Physico-Chimique, Paris, France). Antisera were made in rabbit except for that to CF0-suIII, which was mouse ascites. The Rieske Fe–S protein was from spinach and the remaining proteins were from C. reinhardtii. Alkaline phosphatase-labeled goat anti-rabbit and rabbit anti-mouse were used as secondary antibodies. Markers (MW) were BenchMark Pre-stained Protein Ladder (Invitrogen).
This work was funded by a National Institutes of Health grant (GM38361 to S.K.) and a grant from the Torrey Mesa Research Institute, Syngenta Research and Technology, La Jolla, California (S.K.).
We thank Krishna K. Niyogi for suggesting we use lts1 mutants to localize Rh1-GFP, Kent McDonald for assistance with fixation and staining protocols, Andrei Osterman for translating articles in Russian, Sungsoon Park and Olivier Vallon for antibodies and help with Western blotting, and Denise Schichnes of the CNR Biological Imaging Facility at UC Berkeley for help with fluorescence microscopy. No conflict of interest declared.