|Home | About | Journals | Submit | Contact Us | Français|
The major Rhesus (Rh) protein of the green alga Chlamydomonas reinhardtii, Rh1, is homologous to Rh proteins of humans. It is an integral membrane protein involved in transport of carbon dioxide. To localize a fusion of intact Rh1 to the green fluorescent protein (GFP), we used as host a white (lts1) mutant strain of C. reinhardtii, which is blocked at the first step of carotenoid biosynthesis. The lts1 mutant strain accumulated normal amounts of Rh1 heterotrophically in the dark and Rh1–GFP was at the periphery of the cell co-localized with the cytoplasmic membrane dye FM4-64. Although Rh1 carries a potential chloroplast targeting sequence at its N-terminus, Rh1–GFP was clearly not associated with the chloroplast envelope membrane. Moreover, the N-terminal half of the protein was not imported into chloroplasts in vitro and N-terminal regions of Rh1 did not direct import of the small subunit of ribulose bisphosphate carboxylase (SSU). Despite caveats to this interpretation, which we discuss, current evidence indicates that Rh1 is a cytoplasmic membrane protein and that Rh1–GFP is among the first cytoplasmic membrane protein fusions to be obtained in C. reinhardtii. Although lts1 (white) mutant strains cannot be used to localize proteins within sub-compartments of the chloroplast because they lack thylakoid membranes, they should nonetheless be valuable for localizing many GFP fusions in Chlamydomonas.
The best known Rh (Rhesus) proteins are those that compose the Rh blood group substance of humans. Despite their importance in human medicine and their abundance in the erythrocyte membrane, the primary function of Rh proteins remained unknown for 65years (Agre and Cartron, 1991). This was, in part, because Rh proteins are large, extremely hydrophobic membrane proteins that are difficult to handle. Though Chlamydomonas is usually considered a relative of vascular plants, it also has features of metazoan animals (e.g. flagella=cilia), which were lost in vascular plants (Merchant et al., 2007). Rh proteins are one of these features. They are found in Chlamydomonas and many other microbial eukaryotes, are present in all metazoan animals for which there are complete genome sequences, and were apparently lost in the moss Physcomitrella and vascular plants (JGI Eukaryotic Genomes, http://genome.jgi-psf.org/euk_home.html; Seack et al., 1997; Eichinger et al., 2005; Huang and Peng, 2005; Eisen et al., 2006). Chlamydomonas is one of the simplest organisms to have an Rh protein and we showed that its major Rh protein, Rh1, is involved in transport of carbon dioxide (Soupene et al., 2002, 2004; Fong et al., 2007). Recent evidence indicates that the human Rh blood group substance is also involved in CO2 transport (Endeward et al., 2006, 2008), and this is probably a function of its RhAG (Rh50) component (Huang and Peng, 2005). RhAG remains closest to ancestral Rh proteins like C. reinhardtii Rh1 and does not carry the immunologically problematic epitopes.
Knowing the sub-cellular localization of Rh1 is essential to interpreting future studies of its function. With one exception, all Rh proteins that have been localized are found in the cytoplasmic membrane (Huang, 1997; Liu et al., 2000; Weiner et al., 2003; Ji et al., 2006). The exception is the Rh50-like protein of the slime mold, Dictyostelium discoideum. This protein, which is highly expressed during vegetative growth, is found in the membrane of the contractile vacuole (Benghezal et al., 2001). Targeting requires clusters of acidic residues in its cytoplasmic C-terminal region (Mercanti et al., 2006) that are not present in Chlamydomonas Rh1. Initial in-silico analysis predicted that Chlamydomonas Rh1 had a cleavable chloroplast transit peptide at its N-terminus (Soupene et al., 2002). However, it is difficult to predict localization of membrane proteins in Chlamydomonas, particularly localization of a protein more typical of animals than of plants (Franzén et al., 1990), and hence we wanted to localize Rh1 experimentally. We here present evidence that it is in the cytoplasmic membrane, like most other Rh proteins, and discuss possible alternative roles for its N-terminal sequence.
We used 20 programs to predict the sub-cellular localization of the Chlamydomonas Rh1 protein. They split evenly between predicting a plasma membrane or non-organellar localization (10 programs) and an organellar localization, usually chloroplast (eight programs) (Table 1). To assess the efficacy of the programs for Chlamydomonas proteins, we used them to predict the localization of 16 known proteins (Supplemental Table 1). The 10 programs that did this best (10 or more correct predictions) were also split between predicting a plasma membrane localization (six programs) or chloroplast localization (four programs) for Rh1. Seven of the 10 programs that worked best for Chlamydomonas proteins analyzed the N-terminal sequence rather than the whole protein. The four of these that predicted a chloroplast localization for Rh1 focused on the 39 N-terminal amino acids, in particular the positive charge and the ‘SFFHS’ motif at amino acids 19–23 (Cline and Henry, 1996). However, two of them, which assigned likelihood values to their predictions, gave Rh1 scores of only 0.513 and 0.463 compared to threshold values of 0.500 and 0.420, respectively. One of the programs that predicted a plasma membrane localization for Rh1, PSORT, also predicted a thylakoid membrane localization as a close second. However, PSORT is a notoriously poor predictor of thylakoid membrane proteins (Gómez et al., 2003). Given that none of the programs predicted the localization of the Rh1 protein with high confidence, we explored its localization experimentally, focusing on the two compartments most commonly predicted—the plasma membrane and the chloroplast.
As described in Methods, we fused three portions of Rh1 (predicted transmembrane-spanning segments 1–4, 1–6, and 1–12) to a GFP protein adapted to C. reinhardtii nuclear codon usage (Fuhrmann et al., 1999) and later we also fused full-length Rh1 to this GFP (Figure 1). Fusions of TM1-6 and TM1-12 had been used previously to localize Rh proteins in other organisms (Liu et al., 2000; Ji et al., 2006). When we put the first three constructs into strain 4A+ and selected zeocin resistance, we saw little evidence of expression of fusion proteins in the transformants. Initially, we examined 25 transformants from each construct by microscopy (100–200 cells per transformant). Cultures were grown on TAP medium in the light, which yields low but detectable Rh1 expression (Soupene et al., 2002, 2004). We failed to detect GFP fluorescence above the background of chlorophyll fluorescence. We next used appropriate primers to amplify each RH1–GFP fusion from the 25 transformants to be sure it was present. We obtained PCR products from a subset of transformants in each case (Table 2) and, in all cases, their correctness was confirmed by DNA sequencing. However, only a further subset of the transformants that carried the fusions transcribed them. Though there may have been other problems in the 4A+ background, we subsequently found that the fusions of portions of Rh1 to GFP yielded very few fluorescent transformants in the lts1-204 (white) background, whereas we were able to obtain fluorescent transformants from a fusion of intact Rh1 to GFP (see below).
At a suggestion of K.K. Niyogi, we next explored the use of white mutant strains, lts1 (light sensitive 1) (McCarthy et al., 2004) as hosts for Rh1–GFP constructs. Strains with lesions in lts1 lack phytoene synthase, the first enzyme of carotenoid biosynthesis, and their lack of carotenoids results in degradation of chlorophyll (Herrin et al., 1992; McCarthy et al., 2004). This facilitates direct detection of GFP fluorescence microscopically and hence allows easy screening of many transformants. Because lts1 strains are restricted to growth on acetate in the dark, we first checked to see whether parental strain 4A+ expressed Rh1 in detectable amounts under these conditions and whether the lts1 strains expressed it in similar amounts. As shown in Figure 2, Western analysis indicated that strain 4A+ expressed Rh1 at low but detectable levels on acetate in the dark with respect to the high levels seen in cultures grown on TP and bubbled with air supplemented with 3% CO2 in the light (Soupene et al., 2002). (The Rh1 band in Figure 2 was identified by bubbling a TP-grown culture of 4A+ with CO2 for 4–5h (Lane 5 vs Lane 2).) The lts1-203 and lts1-204 strains appeared to have approximately as much Rh1 as 4A+ on acetate in the dark (Figure 2, Lane 3 vs Lane 1 and not shown). Unfortunately, bubbling dark-grown cultures of 4A+ and the two lts1 strains with CO2 in the dark did not increase the amount of Rh1 (Figure 2, Lanes 4 and 6 vs Lanes 1 and 3).
When we put constructs carrying portions of RH1 fused to GFP (see above) into strain lts1-204 (Inwood et al., 2008) and selected zeocin resistance, we again saw little evidence of expression of fusion proteins in the transformants (data not shown). As discussed above, we examined at least 25 transformants from each construct microscopically but did not perform other tests. Instead, we fused intact RH1 (~5kbp with the promoter region) to GFP. C. reinhardtii has a very high GC content of 67%, which makes it difficult to amplify large regions by PCR. Although we have been able to accomplish this (Kim et al., 2006), we were unable to obtain intact RH1 sequence without alterations (0/17). Hence, we selected three versions of the RH1–GFP construct with slight alterations (see Methods). When we put the intact RH1–GFP fusions into lts1-204, we found that 15 (~20%) of the 73 zeocin-resistant transformants we examined were fluorescent: six of these were from version 1 of the construct, eight from version 2, and one from version 3.
Only three transformants, all derived from version 2 of the intact RH1–GFP fusion, fluoresced strongly enough to be analyzed further by confocal fluorescence microscopy. We will refer to these as transformants 2a (CR505), 2b (CR507), and 2c (CR508). Version 2 was the cleanest of those we used because it contained no nucleotide changes in coding sequence for Rh1 (there were three changes in the 5’ untranslated region and three in introns). That there were no further changes when version 2 of the construct was integrated into the genome was confirmed by sequencing transformants 2a and 2b. Cells of transformant 2a were strongly fluorescent over 6months of examination, whereas cells of transformant 2c fluoresced strongly but intermittently. Cells of transformant 2b, although fluorescent initially, were only very weakly fluorescent for most of the study. Fluorescent cells of both transformants 2a and 2c were a small proportion of the total population (10–20%; Table 3). Occasionally, as many as 30–40% of the cells examined were fluorescent.
We used confocal microscopy to localize GFP fluorescence in ~200 transformant cells carrying intact Rh1–GFP (Table 4 and Figure 3). In almost all of the fluorescent cells, GFP appeared to localize peripherally. Peripheral GFP fluorescence was often more concentrated along the sides of the cell than at the posterior end and was always absent at the anterior end of the cell near the flagella. Peripheral GFP fluorescence was external to weak chlorophyll autofluorescence (red) (Figure 3A–3E) and, in cells of transformant 2a, was shown to co-localize with the red dye FM4-64, which stains the cytoplasmic membrane (Figure 3F). That green fluorescence was clearly exterior to the chloroplast membrane was shown in two-dimensional (Figure 4) and three-dimensional (Figure 5) reconstructions of the data. In addition to its peripheral location, about half the time, Rh1–GFP fluorescence was also found at internal cellular positions that were asymmetrically distributed (Figure 3D, 3E, and 3G–3K). Only rarely was internal GFP fluorescence seen in the absence of peripheral fluorescence (Table 4).
By contrast to the case for peripheral (cytoplasmic membrane) GFP fluorescence, there was no easily discernible pattern to internal GFP fluorescence, which was most intense in cells of transformant 2c (Figure 3D, 3E, and 3G–3K). Internal fluorescence appeared as round patches whose number varied and whose appearance was sharp or diffuse. These patches were sometimes in the cell interior and sometimes at more peripheral locations.
The patches of internal fluorescence did not co-localize with weak chlorophyll autofluorescence and hence were not in chloroplasts (Figure 3D and 3E). Chlorophyll autofluorescence had the same irregular pattern as in host strain lts1-204 (Inwood et al., 2008). Chlorophyll autofluorescence was usually seen around the perimeter of the cells and around starch granules, which could be very abundant and large, and was displaced to the anterior of the cell when starch granules were particularly large. Internal GFP fluorescence also did not co-localize with mitochondria (identified by staining with the dye MitoTracker orange; Figure 3G–3L).
To further test the predicted chloroplast localization of Rh1, we performed in-vitro chloroplast import assays. For this purpose, we used chloroplasts isolated from pea seedlings because they have been used generically for import of proteins from various organisms including Chlamydomonas (Mishkind et al., 1985). Our initial attempts to produce intact Rh1 (61.1kD) in a wheat germ extract were unsuccessful (data not shown). Hence, we tested whether the N-terminal portion of Rh1, which was predicted to carry a cleavable chloroplast targeting sequence, could target proteins to chloroplasts. The first tests were done with N-terminal fragments of Rh1 that carried the predicted transit peptide, namely the N-terminal 200 amino acids ending in loop 5 or the N-terminal 235 amino acids ending in loop 6 (Figure 6). There was no detectable import of [35S]-labeled truncations of Rh1 into chloroplasts (data not shown). The second tests were done with N-terminal fragments of Rh1 fused to the mature portion of the Rubisco small subunit (mSSU), specifically with Rh1(1–44)–mSSU and Rh1(1–105)–mSSU (Figure 6). A very small amount of Rh1(1–44)–mSSU was recovered in chloroplasts, although it showed no change in mobility on SDS–PAGE (Figure 7, compare Lanes 8 and 11, 12). No detectable Rh1(1–105)–mSSU was recovered in chloroplasts (Lanes 15–19).
Although Rh1(1–44)–mSSU was recovered in the pellet fraction after alkaline wash (Lane 11), which is normally indicative of an ‘integral association’ with chloroplast membranes, it was digested by thermolysin (compare Lanes 12 and 13), a protease that has access to proteins in the outer but not the inner membrane of the chloroplast envelope. Thus, the association of Rh1(1–44)–mSSU with membranes was apparently adventitious. By contrast to the fusions, the precursor of SSU (prSSU; Lane 1) was imported into the soluble fraction of chloroplasts as a smaller processed form (Lane 2). It was resistant to thermolysin digestion (compare Lanes 5 and 6) unless chloroplasts were disrupted with detergent (Lane 7). Import depended on the transit peptide because targeting of the mature form of SSU (mSSU, Lane 22) to chloroplasts was not detectable (Lanes 23–26). In fact, the behavior of mSSU was very similar to that of the Rh1–mSSU fusions (Figure 7 and data not shown). Collectively, these data indicate that the N-terminal region of Rh1 does not contain information for chloroplast targeting in vitro and does not function as a cleavable stromal targeting sequence. Although our failure to detect import of truncated or chimeric versions of Chlamydomonas Rh1 into pea chloroplasts in vitro might be attributed to the heterologous assay system, this is unlikely because the general chloroplast import system of Chlamydomonas appears to be much like that of vascular plants (Kalanon and McFadden, 2008).
To facilitate localization of fusions of Rh1 to GFP, we used as host an lts1 mutant strain of C. reinhardtii, which lacks carotenoids and is white (McCarthy et al., 2004). (See Results for difficulties in using a wild-type host.) A fusion of intact Rh1 to GFP localizes to the cytoplasmic membrane (Figures 3–5) and is depleted in the specialized region of membrane around the flagella (Bloodgood et al., 1986). Two- and three-dimensional reconstructions (Figures 4 and and5)5) show clearly that the fusion is external to the chloroplast. We have recently postulated that Rh proteins transport the hydrated form of CO2 (H2CO3=HCO3–+H+) (Fong et al., 2007) under conditions of high CO2 availability. That Chlamydomonas Rh1 would carry out this function across the cytoplasmic membrane seems a reasonable possibility. The lts1 strains have an intact chloroplast envelope membrane but lack stacked thylakoids (Inwood et al., 2008). Although we consider it unlikely, we cannot rule out the possibility that Rh1 is normally targeted to thylakoid membranes and defaults to the cytoplasmic membrane in their absence.
In agreement with the cytoplasmic membrane localization of the Rh1–GFP fusion in a C. reinhardtii lts1 strain, neither large N-terminal fragments of Rh1 nor fusions of short putative targeting sequences to SSU from pea were imported into chloroplasts in vitro. Although small amounts of one fusion appeared to be adventitiously associated with pea chloroplast membranes, the N-terminal sequence of this protein was not cleaved. Rather than targeting Rh1 to the chloroplast, its long N-terminal sequence may be a signal peptide or a signal anchor sequence (Shaw et al., 1988). The first transmembrane spanning segment of ammonium transport (Amt) proteins, which are the only known homologs of Rh proteins (Marini et al., 1997; Huang and Peng, 2005), is removed in proteobacteria and in archaea (Thomas et al., 2000; Khademi et al., 2004; Zheng et al., 2004; Andrade et al., 2005; Thornton et al., 2006). The first transmembrane spanning segment of the Rh protein from the bacterium Nitrosomonas europaea is also removed, at least in E. coli (Li et al., 2007; Lupo et al., 2007). Cleavage of the N-terminus, which leaves 11 transmembrane spanning segments, is apparently very unusual for bacterial membrane transport proteins and its mechanism remains to be defined (Pao et al., 1998; Saier, 2003; Thornton et al., 2006). In contrast to the case for the Nitrosomonas europaea Rh protein, all three components of the Rh blood group substance, RhAG, RhD, and RhCcEe, retain the N-terminal transmembrane spanning segment, which is followed by a large loop that may separate it from the remainder of the protein (Avent et al., 1992; Eyers et al., 1994) (Figure 8). It is not known whether the N-terminal segment is removed from C. reinhardtii Rh1.
In about half of the fluorescent cells we examined, Rh1–GFP fluorescence was localized internally as well as to the cytoplasmic membrane. Internal localization was to a variable number of patches in the cytoplasm that were readily distinguished from mitochondria (Figure 3G–3L). Whether this internal fluorescence is associated with the delivery machinery for cytoplasmic membrane proteins, the autophagous vacuoles abundant in lts1 strains (Inwood et al., 2008), the contractile vacuole, as is the case for a Dictyostelium discoideum Rh protein (Benghezal et al., 2001; Mercanti et al., 2006), or another organelle(s) remains to be determined. It likewise remains to be determined whether Rh1 is normally dually targeted. Dual localization has been well documented for other membrane transport proteins; for example, the general amino acid permease of the yeast Saccharomyces cerevisiae is found in both the cytoplasmic and vacuolar membranes (Risinger et al., 2006; Risinger and Kaiser, 2008).
Although lts1 strains lack thylakoid membranes (Inwood et al., 2008) and hence will not be useful for localization of proteins within sub-compartments of the chloroplast, their low levels of chlorophyll should make them valuable for localizing a variety of GFP fusions in Chlamydomonas. Rh1–GFP is one of the first cytoplasmic membrane protein fusions available in Chlamydomonas and, as such, may be useful for studying exclusion of cytoplasmic membrane proteins from specialized membrane regions around Chlamydomonas flagella (Bloodgood et al., 1986).
References, algorithms, and some basic parameters for the 20 programs we used to determine the localization of the Chlamydomonas Rh1 protein are listed in Table 1. Where possible, Chlamydomonas was treated as a plant. The 16 Chlamydomonas proteins of known localization used to assess the efficacy of the programs were: CSOA and CSOB, the opsin proteins localized in the plasma membrane associated with the eyespot; a PDR-like transport protein that was identified as a plasma membrane protein in Arabidopsis; PMA2, a plasma membrane ATPase; CAH1, a periplasmic carbonic anhydrase; CAH3, a thylakoid lumen carbonic anhydrase; CAH6, a chloroplast stromal carbonic anhydrase; mtCA, a mitochondrial carbonic anhydrase; SSU, the small subunit of ribulose bisphosphate carboxylase (Rubisco), which is in the chloroplast stroma; Lhcbm6, a light harvesting protein of photosystem II, which is in the thylakoid membrane of the chloroplast; ATPvA2 and ATPvA1, subunits of the vacuolar membrane ATPase; ATPC, ATPD, and ATPG, nuclear-encoded subunits of the chloroplast ATPase, which is in the thylakoid membrane; and ATP1A, a subunit of the mitochondrial ATPase, which is in the mitochondrial inner membrane.
Wild-type strain 4A+ and the lts1-203 and lts1-204 mutants were obtained from K.K. Niyogi (University of California, Berkeley). Wild-type strain CC-125, which was used to amplify RH1, was obtained from the Chlamydomonas Culture Collection (Duke University, Durham, North Carolina). Control strain Ble–GFP (Fuhrmann et al., 1999), which expresses a fusion between Ble and GFP in the nucleus, was obtained from S.P. Mayfield (The Scripps Research Institute, La Jolla, CA). Strains were maintained at 25°C on Tris Acetate Phosphate (TAP) agar medium (Harris, 1989, p. 26) containing NH4Cl (10mM) as nitrogen source, as were transformants carrying RH1 or portions of RH1 fused to GFP (see below). Liquid cultures were grown in the same medium unless otherwise specified. Strain 4A+ was also grown in medium without acetate (TP medium), as noted. Medium for Ble–GFP was supplemented with 250μM arginine. The lts1 strains, which have lesions in the gene for phytoene synthase, the first enzyme of carotenoid synthesis (McCarthy et al., 2004), were maintained in complete darkness, as were transformants carrying RH1–GFP fusions. Other strains were maintained under continuous illumination (125μmolphotonsm−2s−1). For microscopic examination, cells were grown in liquid culture with constant orbital shaking (~110rpm) at 25°C and harvested during the exponential phase.
Strains were grown in flat 1-l bottles containing 700ml of TAP–NH4Cl medium under constant illumination (~100μmolphotonsm−2s−1) or in darkness and bubbled with air (0.035% vol./vol. CO2). After growth to mid-exponential phase, some cultures were then subjected to an upshift by bubbling with air enriched with 3% vol./vol. CO2 for 4–5h. 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 incubated at 37°C for 30min. 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. Affinity-purified rabbit antibodies raised against a C-terminal peptide of Rh1 (Zymed) were used to detect Rh1, and affinity-purified rabbit antibodies (kindly provided by Martin H. Spalding, Iowa State University, Ames) were used to detect the periplasmic carbonic anhydrase Cah1. Membranes were incubated with an alkaline phosphatase-conjugated secondary antibody (Zymed).
The sequence of the RH1 gene was determined from the RH1/RH2 region of the Chlamydomonas genome (Joint Genome Institute v2.0) and was the same as that in strain CC-125. Chlamydomonas genomic DNA was prepared from strain CC-125 as described (Kim et al., 2005) and the 5′-UT region for RH1 (~1.4kb), portions of RH1 (through predicted transmembrane segments 4, 6, and 12; Figure 8), and intact RH1 were amplified by PCR using primers listed in Table 5. Amplified regions were cloned into the HindIII/XbaI sites of pCrGFP (Entelechon, Regensburg, Germany), in which the GFP gene is adapted for Chlamydomonas nuclear codon usage (Fuhrmann et al., 1999). Fragments carrying fusions were then excised from pCrGFP using HindIII/EcoRI and cloned into the EcoRV site of the pSP124s vector (gift of Heriberto Cerutti, University of Nebraska, Lincoln; described in Soupene et al., 2004).
Transmembrane segments of Rh1 were predicted by alignment with the human RhAG protein (Soupene et al., 2002) (Figure 8). Forward primer RH1–HIII/5’ was paired with each of the reverse primers RH–XbaI/3′, RH–XbaI/3′Ex7, and RH–XbaI/3′Ex10 for the various segments of RH1. Forward primer RH1–HIII/5′c was paired with reverse primer RH1–XbaI/3′Ex12 for intact RH1 (Figure 1). PCR of RH1 segments was performed using the Expand Long Template PCR System (Roche) in a 100-μl reaction mixture containing: genomic DNA from wild-type strain CC-125 (90ng); primers (300nM of each); dNTPs (500μM of each); DMSO (5%); MgCl2 (2.75mM); and proof-reading Expand Taq polymerase (3.75units). Amplification conditions were: 2min at 95°C; 30cycles of amplification (denaturation for 30s at 94°C/annealing for 1min at 55°C/elongation for 7min at 68°C; and 10min at 68°C). The only changes for PCR amplification of intact RH1 were: 150ng genomic DNA; 3.5mM MgCl2; 4.75units polymerase; and annealing at 65°C. Constructs in pSP124s were then transformed into strain 4A+ or strain lts1-204 that had been treated with autolysin as described previously (Soupene et al., 2004). They were transformed by a modified version of the glass bead method (Kindle, 1990). Polyethylene glycol was added during the transformation and was vortexed together with the plasmid and cells for 20s. The cells were then grown out in TAP/NH4Cl and subsequently concentrated and plated with selection for zeocin-resistance as described (Soupene et al., 2004). Plates were incubated in the light for 4A+ or the dark for lts1-204 for 4weeks before transformants were picked. Although we observed some green colonies for lts1-204 on transformation plates, these were obtained even when pSP124s, which did not carry GFP fusion inserts, was used. Green transformant colonies did not fluoresce and hence appeared to be accumulating more chlorophyll than lts1-204.
Although we obtained the correct sequences when portions of RH1 were fused to GFP, we were unable to do so when intact RH1 was fused to GFP. (We checked 17 sequences.) In the latter case, we chose three variants that had the fewest changes and those likely to be least damaging. Variant 1 carried one change in the 5′-UT, one change in an intron, and two silent changes in exons 2 and 12 (four changes total); variant 2 carried four changes in the 5′-UT and three changes in introns (seven changes total); variant 3 carried two changes in the 5′-UT, three changes in introns, and one change in exon 5, changing an isoleucine to a threonine (six changes total; Table 6).
A clone carrying the cDNA for Rh1 (Genbank accession number AY013257) in pBluescript KS (Cr.Rhp) was obtained from Cheng-Han Huang (Lindsley F. Kimball Research Institute, New York Blood Center, New York City) and was used for PCR amplification of cDNA. The cDNA coding for Rh1 was amplified with forward primer RH1–ATG and reverse primer RH1–TAG. The coding region plus a 1.4-kb region of the 3′-UT was excised from the original pBluescript plasmid and both were cloned into the EcoRV site of pGEM-T Easy (Promega). The cDNA coding for fragments of Rh1 ending after predicted transmembrane spanning segments 5 or 6 was amplified with forward primer RH1–ATG and reverse primers ‘LOOP 5’ or ‘LOOP 6’, respectively (Tables 5 and and7;7; Figure 8, L5 and L6). PCR products were cloned into pGEM-T Easy and constructs in which Rh1 cDNA was transcribed from the SP6 promoter were chosen. For fusions of putative chloroplast targeting sequences from Rh1 to mature SSU (mSSU) from pea, two segments of cDNA encoding N-terminal regions of Rh1 were amplified using primer A and primers B (fragment of 132 nucleotides) and C (fragment of 315 nucleotides) (Table 7 and Figure 8, a and b). PCR products were cloned into pGEM-T Easy and then excised with NcoI and SphI. They were cloned into the corresponding sites of pET–prSSU (Inoue et al., 2001) to fuse N-terminal codons of Rh1 to coding sequence for mature pea SSU from which N-terminal chloroplast targeting sequences had been deleted. cDNA for mSSU was amplified from pET–prSSU using primers mSSU–ATG and mSSU–TGA1 and cloned into pGEM-T Easy. The resultant protein was used as a negative control for chloroplast import. All constructs were transformed into DH5α (Invitrogen) and maintained on Luria broth supplemented with ampicillin and 0.2% glucose. When expressed from high copy number vectors such as pBluescript and pGEM-T Easy, Rh1 appears to be toxic to host cells, which grow worse without supplemental glucose. The correctness of all constructs was confirmed by sequencing.
Zeocin-resistant transformants potentially carrying GFP fusion constructs were first screened for detectable GFP fluorescence with a Zeiss Axiophot epifluorescence microscope with a Zeiss Fluar 40X objective. Transformants that showed Rh1–GFP fluorescence in preliminary tests were then examined further by confocal microscopy on a Zeiss 510 Meta confocal laser scanning microscope (Carl Zeiss, Oberkochen, Germany; CNR Biological Imaging Facility, UC Berkeley). To assure that we were studying live cells, we photographed only flagellated cells that had been observed to move as they were immobilized in warm agar (20μl cells+20μl low-melting-point agar at 40°C). They were observed through a Zeiss Plan-Geofluar 40X/1.35 oil immersion or a Zeiss Plan-APO chromat 100X/1.40 Oil DIC objective and were photographed through the latter. GFP fluorescence was excited using a 488-nm Argon laser with a 505–550-nm barrier filter before the photomultiplier tube (PMT) detector. Chlorophyll fluorescence was excited using a 543-nm Helium Neon laser with a 560-nm Long pass filter before the PMT. To resolve the closely related spectra of GFP fluorescence in transformants from chlorophyll autofluorescence, we first obtained reference spectra by taking lambda scans of strains 4A+, lts1-204, and Ble–GFP. These were used in linear unmixing of lambda scans from transformants. Differential interference (DIC) images were obtained using a 488-nm source and the transmitted light detector. Images were analyzed using Zeiss 510 meta software.
For cytoplasmic membrane staining, FM4-64 (Molecular Probes, Inc.) was diluted to 17μM final concentration from a stock solution of 82μM in Hanks balanced salt solution (Molecular Probes) and the cells were viewed immediately. For mitochondrial staining, MitoTracker Orange CMTMRos (Molecular Probes, Inc.) was diluted to 300nM and cells were incubated at room temperature for 15–40min. They were then pelleted by centrifugation and re-suspended in fresh medium. Stained cells were mixed with low-melting-point agar and viewed. Excitation was as for chlorophyll.
Precursor proteins were synthesized from the cDNA constructs described above using either T7 or SP6 RNA polymerase in a coupled transcription–translation system from wheat germ or rabbit reticulocytes (Promega, Madison, WI). They were labeled with [35S]methionine. Chloroplasts were isolated from 9–12-day-old pea seedlings as described (Bruce et al., 1994), and the import assay and subsequent fractionation and protease treatments were done as described by Inoue et al. (2006).
Supplementary Data are available at Molecular Plant Online.
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.). No conflict of interest declared.