By continuing to browse this site you agree to us using cookies as described in About Cookies
Notice: Wiley Online Library will be unavailable on Saturday 7th Oct from 03.00 EDT / 08:00 BST / 12:30 IST / 15.00 SGT to 08.00 EDT / 13.00 BST / 17:30 IST / 20.00 SGT and Sunday 8th Oct from 03.00 EDT / 08:00 BST / 12:30 IST / 15.00 SGT to 06.00 EDT / 11.00 BST / 15:30 IST / 18.00 SGT for essential maintenance. Apologies for the inconvenience.
Reporter genes have been successfully used in chloroplasts of higher plants, and high levels of recombinant protein expression have been reported. Reporter genes have also been used in the chloroplast of Chlamydomonas reinhardtii, but in most cases the amounts of protein produced appeared to be very low. We hypothesized that the inability to achieve high levels of recombinant protein expression in the C. reinhardtii chloroplast was due to the codon bias seen in the C. reinhardtii chloroplast genome. To test this hypothesis, we synthesized a gene encoding green fluorescent protein (GFP) de novo, optimizing its codon usage to reflect that of major C. reinhardtii chloroplast-encoded proteins. We monitored the accumulation of GFP in C. reinhardtii chloroplasts transformed with the codon-optimized GFP cassette (GFPct), under the control of the C. reinhardtii rbcL 5′- and 3′-UTRs. We compared this expression with the accumulation of GFP in C. reinhardtii transformed with a non-optimized GFP cassette (GFPncb), also under the control of the rbcL 5′- and 3′-UTRs. We demonstrate that C. reinhardtii chloroplasts transformed with the GFPct cassette accumulate ≈80-fold more GFP than GFPncb-transformed strains. We further demonstrate that expression from the GFPct cassette, under control of the rbcL 5′- and 3′-UTRs, is sufficiently robust to report differences in protein synthesis based on subtle changes in environmental conditions, showing the utility of the GFPct gene as a reporter of C. reinhardtii chloroplast gene expression.
Reporter genes have greatly enhanced our ability to monitor gene expression in a number of biological organisms. In chloroplasts of higher plants, β-glucuronidase (uidA, Staub and Maliga, 1993); neomycin phosphotransferase (nptII, Carrer et al., 1993); adenosyl-3-adenyltransferase (aadA, Svab and Maliga, 1993); and the green fluorescent protein of Aequorea aequorea (gfp, Reed et al., 2001; Sidorov et al., 1999) have been used as reporter genes (Heifetz, 2000). Each of these genes has attributes that makes it a useful reporter of chloroplast gene expression, such as ease of analysis, sensitivity, or the ability to examine expression in situ. Based on these studies, other heterologous proteins have been expressed in the chloroplasts of higher plants, such as Bacillus thuringiensis Cry toxins, conferring resistance to insect herbivory (Kota et al., 1999), or human somatotropin (Staub et al., 2000), a potential biopharmaceutical.
Several reporter genes have been expressed in the chloroplast of the eukaryotic green alga Chlamydomonas reinhardtii, although with varying degrees of success. These include aadA (Goldschmidt-Clermont, 1991; Zerges and Rochaix, 1994); uidA (Ishikura et al., 1999; Sakamoto et al., 1993); Renilla luciferase (Ren Luc, Minko et al., 1999); and the amino glycoside phosphotransferase from Acinetobacter baumanii, aphA6 (Bateman and Purton, 2000). The amount of recombinant protein produced was reported for the uidA gene only (Ishikura et al., 1999), and from these reports, and based on Western blot analysis or activity measurements, it seems clear that the amounts produced were very low. We hypothesized that the inability to achieve high levels of recombinant protein expression in the C. reinhardtii chloroplast was due to the codon bias seen in the C. reinhardtii chloroplast genome (Nakamura et al., 1999). Previously, Fuhrman et al. (1999) had demonstrated that a synthetic GFP gene reflecting C. reinhardtii nuclear codon bias was expressed in C. reinhardtii and allowed for visualization of the recombinant protein as a fusion with the zeocin binding protein ble (Drocourt et al., 1990).
The C. reinhardtii chloroplast genome displays a strong codon bias, with A or U preferred at the third position (Nakamura et al., 1999). To test the hypothesis that codon usage might affect the level of expression of recombinant proteins in the C. reinhardtii chloroplast, we synthesized a gene encoding GFP de novo, optimizing its codon usage to reflect that of major C. reinhardtii chloroplast-encoded proteins. We monitored the accumulation of GFP in a C. reinhardtii strain whose chloroplast was transformed with the codon-optimized GFP cassette (GFPct), under the control of the C. reinhardtii rbcL 5′- and 3′-UTRs. We compared this expression with the accumulation of GFP in a strain of C. reinhardtii transformed with a non-optimized GFP cassette (GFPncb), also under the control of the rbcL 5′- and 3′-UTRs. In this report, we demonstrate that C. reinhardtii chloroplasts transformed with the GFPct cassette accumulate approximately 80-fold more GFP than a GFPncb-transformed strain, indicating that codon usage has a profound effect on the expression of heterologous proteins in the C. reinhardtii chloroplast. We further demonstrate that expression from the GFPct cassette, under control of the rbcL 5′- and 3′-UTRs, is sufficiently robust for the detection of fluctuations in gene expression due to environmental effects, making the GFPct cassette a generally useful reporter gene for C. reinhardtii chloroplasts.
De novo synthesis of a GFP gene in C. reinhardtii chloroplast codon bias
To develop a robust reporter gene for expression in the C. reinhardtii chloroplast, we synthesized a GFP gene whose codon usage was optimized to reflect that of the C. reinhardtii chloroplast genome. We designed two amino acid changes to the native GFP (GFPncb)-coding region to enhance the fluorescent and expression properties of the protein. The first of these amino acid changes, a serine-to-alanine change at amino acid position 2, was a consequence of introducing an NdeI restriction site at the 5′ end of the gene. Earlier work had demonstrated that such a change was unlikely to alter the stability or spectral qualities of the protein (Dopf and Horiagan, 1996). The second change, a serine-to-threonine change at amino acid 65, was made to enhance the amplitude of excitation at 485 nm relative to native GFP (approximately sixfold), while at the same time reducing excitation at 395 nm (Heim et al., 1995). This change was introduced into the GFPct coding sequence to improve our chances for fluorescent detection using visible light. As shown in Figure 1, there is also an amino acid change, Q80R in the GFPncb gene that is not in the wild-type GFP gene. This alteration was introduced during PCR amplification of the native GFP gene, prior to our obtaining the clone. This Q80R mutation is a common alteration found on amplification of native GFP coding sequences using PCR (Tsien, 1998), and has been shown to have no effect on protein function, therefore this change was included in the GFPct gene for consistency.
Characterization of E. coli-expressed GFPct and GFPncb
To determine if the GFPct and GFPncb genes were capable of producing functional GFP protein, we examined Escherichia coli cell lysates prepared from cells transformed with either pETGFPct or pETGFPncb. As shown in Figure 2 (left panel), Ni affinity chromatography of E. coli lysates produced proteins of the correct molecular mass for GFP. Direct fluorescence assays of SDS–PAGE-separated E. coli-produced proteins revealed that both proteins fluoresce under blue light illumination (Figure 2, central panel), and show slightly different fluorescent properties consistent with the introduced amino acid changes. Scanning excitation/emission spectroscopy revealed that the S65T alteration to the GFPct protein resulted in a greatly enhanced level of fluorescence at 485 nm, while its fluorescence at 395 nm was greatly reduced (Figure 2, central panel). Only one-fifth of the amount of E. coli-expressed GFPct protein was used in this assay relative to GFPncb protein. Western blot analysis, using a mouse polyclonal antibody raised against native GFP, showed a similar signal for both GFPct and GFPncb (Figure 2, right panel). This result is particularly important given the fact that we had intentionally enhanced the spectral qualities of the GFPct protein relative to the GFPncb protein. Thus, while fluorescence detection based on excitation in the visible (485 nm) would favor GFPct detection, immunolabeling is non-discriminatory, allowing for the direct comparison of GFPct and GFPncb protein accumulation in C. reinhardtii chloroplasts.
Southern and Northern blot analysis of GFPct and GFPncb transformants
Once we had demonstrated that both GFPct- and GFPncb-coding sequences were capable of producing functional GFP proteins, we proceeded to transform C. reinhardtii chloroplasts with pExGFPct and pExGFPncb. Cells were transformed with both the GFP plasmids and the selectable marker plasmid p228, conferring resistance to spectinomycin. Primary transformants were screened by PCR, followed by Southern blot analysis, and positive transformants were taken through additional rounds of selection to isolate homoplasmic lines, in which all copies of the chloroplast genome contained the introduced GFP gene. Two homoplasmic GFPct transformants, 18.3 and 21.2, and two homoplasmic GFPncb transformants, 5.8 and 12.1, were selected for further analysis. Figure 3(a) shows the GFPct and GFPncb constructs with relevant restriction sites indicated. Correct integration of the 7.1 kb Eco/Xho region of plasmids pExGFPct and pExGFPncb into the chloroplast genome was ascertained using the probes indicated on the map of the genes (Figure 3b). Genomic DNA from the wild type and the GFPct and GFPncb transformants was digested with EcoRI and XhoI, fractionated on agarose gels, and subjected to Southern blot analysis. As shown in Figure 3(a), the rbcL 5′-UTR contains an EcoRI restriction site, thus digestion of transformant DNA with EcoRI/XhoI should result in a smaller fragment hybridizing to either the 5′ or 3′ p322 probes relative to wild-type DNA. As shown in Figure 4, the 5′ p322 32P-labeled probe (left panel) and the 3′ p322 32P-labeled probe (right panel) hybridized to EcoRI fragments of 3.7 and 3.3 kb, respectively, in the GFPct and GFPncb transformants. These same probes, however, hybridize to a 5.7 kb EcoRI/XhoI fragment in the non-transformed wild-type C. reinhardtii strain, as expected. The DNA blots were stripped and re-probed with GFPct- and GFPncb-specific probes. An EcoRI/XhoI fragment of 3.3 kb was detected in transformants 5.8 and 12.1 using the GFPncb probe (Figure 4, central panel), and a similarly sized fragment was identified in transformants 18.3 and 21.2 using the GFPct probe (Figure 4, central panel). No signal was seen in wild-type C. reinhardtii DNA using either GFP probe.
Accumulation of GFP mRNA in transgenic strains
To determine if the GFPct and GFPncb genes were transcribed in transgenic C. reinhardtii chloroplasts, Northern blot analysis of total RNA was used. Ten µg of total RNA, isolated from wild-type and transgenic lines 5.8, 12.1, 18.3 and 21.2, was separated on denaturing agarose gels and blotted to nylon membrane. Duplicate filters were hybridized with either a 32P-labeled psbA or rbcL cDNA probe. As shown in Figure 5, each strain accumulates psbA and rbcL mRNAs to similar levels, demonstrating that equal amounts of RNA were loaded for each lane, and that chloroplast transcription and mRNA accumulation are normal in the transgenic strains. The filters were stripped and re-probed with the GFPct- and GFPncb-specific probes. As shown in the bottom panels of Figure 5, strains 5.8 and 12.1 accumulate GFPncb mRNA, while strains 18.3 and 21.2 accumulate GFPct mRNA. No GFP signal is observed in wild-type cells, as expected. All four cDNA probes were labeled to approximately the same specific activity, and while the GFPct and GFPncb signals were similar, both GFP-probed filters required longer exposures (approximately four times) to obtain a signal similar to the rbcL probe. These data indicate that the GFP mRNAs accumulate to roughly a quarter of the level of endogenous rbcL mRNA.
Analysis of GFP accumulation in transgenic C. reinhardtii chloroplasts
To determine the levels of GFPct and GFPncb protein accumulation in the transgenic lines, GFP was measured by both fluorescence and Western blot analysis. Six µg of total soluble protein (tsp) was separated by SDS–PAGE and the resulting gels subjected to either Coomassie staining, fluorescence imaging or Western blot analysis. The Coomassie-stained gel indicates that equal amounts of protein were loaded in each lane (Figure 6, left panel). The fluorescence gel imaged at 485 nm excitation, 535 nm emission, shows a signal only for the GFPct transformants 18.3 and 21.2 (Figure 6, central panel). No fluorescent signal was observed for any GFP transformant when excited at 366 nm (data not shown). Western blot analysis of the same samples showed similar results to the fluorescence analysis, with no GFP detected in the GFPncb transformants, and a good signal in the GFPct strains (right panel). To ascertain more precisely the difference in GFP accumulation between GFPct and GFPncb transformants, we carried out the titration shown at the bottom of Figure 6. Twenty µg tsp from GFPncb transformants 5.8 and 12.1 were separated along with tsp from GFPct transformant 21.2. For the GFPct strain, protein concentrations ranged from 20 µg (lane 1/1) to 250 ng (lane 1/80). A comparison between lanes 3 (1/80 the tsp loaded for GFPncb strains) and lanes 1 and 2 indicate that the level of GFPct accumulation in the 21.2 transformant is ≈80-fold higher than that seen in either of the GFPncb transformants.
We also developed a more rapid assay for quantifying GFP accumulation in soluble protein extracts of C. reinhardtii. As shown in Figure 7, we were able to reproducibly quantify differences in fluorescence between transformed strains 18.3 and 21.2 as well as the wild-type strain, 137c(+) utilizing a microtiter plate format. Figure 7(a) shows a fluorescence image obtained comparing soluble proteins extracted from wild-type and transgenic strains 18.3 and 21.2. Figure 7(b) shows quantitative data from the same experiment performed in triplicate. These results were in agreement with Western blot data which showed an approximately four- to fivefold difference in the level of GFP accumulation in transgenic strain 18.3 versus 21.2 (Figure 6).
Use of chloroplast optimized GFP as a reporter of chloroplast gene expression
To ascertain the utility of the GFPct gene as a reporter of chloroplast gene expression in C. reinhardtii, we examined the effects of different growth conditions on GFPct accumulation in transgenic lines. Chlamydomonas reinhardtii GFPct transgenic strain 21.2 was maintained under constant illumination at a cell density of 1 × 106 ml−1 at either 5000 lux (high light) or 450 lux (low light), prior to harvesting. A Western blot analysis was carried out on 1 µg tsp from each treatment. As shown in Figure 8, cells maintained at 1 × 106 ml−1 under constant illumination of 5000 lux accumulate roughly 10% as much GFPct as cells maintained at 1 × 106 ml−1 under low light flux. When a third flask was maintained at a cell density of 1 × 107 ml−1 under 5000 lux, constant illumination, GFP again accumulated to high levels, as the high cell density acted to reduce light intensity within the growing culture, in essence creating a low light environment. Western blots also revealed that steady-state levels of D1 protein and large subunit of RUBP carboxylase remained constant regardless of the light regime (data not shown). These data show that the GFPct gene can be used to report differences in protein synthesis based on subtle changes in environmental conditions, and demonstrate the utility of the GFPct gene as a reporter of chloroplast gene expression.
Finally, we were curious as to how GFP accumulation in our chloroplast transformants compared with that in two nuclear transformants expressing GFP fusion proteins. The first of these transformants, ble GFP, expresses a fusion between ble and nuclear codon-optimized GFP which has been shown to localize to the nucleus of C. reinhardtii (Fuhrman et al., 1999). This fusion protein is expected to have a molecular weight of ≈40 kDa. The second nuclear transformant we analyzed, SPcP76, expresses the same nuclear-optimized GFP, but as a c-terminal fusion with the gene encoding the chloroplast localized poly(A)-binding protein, RB47 (Yohn et al., 1998). The RB47 GFP fusion protein is expected to have a molecular weight of ≈96 kDa. While we were able to detect fluorescence in the nucleus of cells expressing ble GFP via fluorescence microscopy, we were not able to detect fluorescence in either of the other two GFP-expressing strains.
To determine the relative amounts of GFP in each of the transgenic lines, we carried out Western blot analysis on soluble proteins extracted from the three lines, as shown in Figure 9. Surprisingly, GFPct strain 21.2 accumulates approximately fivefold more GFP than either of the other two strains, ble GFP and SPcP76, which appear to accumulate roughly equal amounts of GFP. We also noted that the same relative level of GFP accumulation in the three transgenic lines was reflected by analysis of fluorescence gels (data not shown).
Several heterologous genes have been employed as reporters of chloroplast gene expression in C. reinhardtii, but their utility has been limited due to low levels of protein expression. There are several possible explanations for the low levels of heterologous protein expression in C. reinhardtii chloroplasts. First, it is possible that, in some instances, the promoters used to drive transcription of these genes might have resulted in low levels of transcription. Alternatively, some of these reporter mRNAs might be inherently unstable, resulting in low levels of mRNA accumulation. Another possibility is that RNA elements required for translation might be lacking from these chimeric mRNAs. Finally, the possibility exists that the strong codon bias observed in C. reinhardtii chloroplast genes might preclude the translation of heterologous mRNAs.
Although promoter activity and mRNA stability have been shown to greatly impact gene expression in chloroplasts, analysis of transgenic C. reinhardtii chloroplasts has shown sufficient heterologous mRNA accumulation to support high levels of protein synthesis (Blowers et al., 1990; Salvador et al., 1993). Additionally, in most cases C. reinhardtii 5′- and 3′-UTRs were used in construction of the chimeric genes, therefore it seems unlikely that critical RNA elements were lacking from these reporter mRNAs. We therefore decided to alter codon usage as a means of enhancing heterologous protein accumulation in the C. reinhardtii chloroplast, and chose the well characterized reporter gene encoding GFP from A. aequorea for these studies.
We engineered the coding region of GFP to match the codon usage of protein-coding sequences from the C. reinhardtii chloroplast genome. We placed the expression of this GFPct gene, as well as a native GFP gene (GFPncb), under the control of the C. reinhardtii chloroplast rbcL 5′- and 3′-UTRs. We show that both the GFPncb gene and the GFPct gene are transcribed and accumulate mRNA to similar levels in transgenic C. reinhardtii chloroplasts. Transgenic strains expressing GFPct accumulate ≈80-fold more GFP than those expressing GFPncb. The GFPct-producing strain 21.2 accumulates GFP to ≈0.5% of the total soluble protein under optimal growth conditions. This level of protein expression is adequate for analysis of GFP expression by fluorescence assays of total cellular proteins. Previous reports of uidA (GUS) expression in C. reinhardtii chloroplast under the control of the rbcL 5′- and 3′-UTRs showed low levels of protein expression, ≈0.01% of soluble protein (Ishikura et al., 1999). This level of GUS accumulation is similar to the level of GFP accumulation obtained with the GFPncb gene we obtained using the same rbcL control elements. These authors also reported relatively low levels of rbcL-GUS mRNA accumulation (Ishikura et al., 1999), again similar to the low levels we observed for rbcL-GFP mRNA.
We found that GFP fluorescence was not easily visualized in the chloroplast of transformant 21.2 using fluorescence microscopy, while the nucleus was easily discernible by fluorescence microscopy in the nuclear transformant expressing ble GFP. This was despite the fact that both Western blot analysis (Figure 9) and fluorescence gel assays confirmed that the chloroplast transformant 21.2 accumulates approximately fivefold more GFP than the ble GFP transformant. We attribute the inability to detect GFP fluorescence in situ in chloroplast transformants primarily to GFP's localization to the chloroplast. First, chlorophyll and other pigments found within the chloroplast are expected to absorb much of the incident light targeted to GFP. In addition, some of the light emitted by chloroplast-localized GFP might in fact be re-absorbed by chlorophyll and other pigments, further diminishing the fluorescence signal. While in situ visualization of GFP in higher plant chloroplasts has been achieved (Reed et al., 2001; Sidorov et al., 1999), the plastids were visualized only when levels of GFP accounted for over 5% of the total soluble protein. This level of accumulation is approximately 10-fold higher than those we have presently achieved in C. reinhardtii chloroplasts. Hence, increasing levels of GFP expression five- to 10-fold above that which we currently have should allow for the in situ visualization of GFP in C. reinhardtii chloroplasts as well.
Alternatively, we were able to develop a fairly rapid microtiter-based method for analysis of GFP accumulation in transgenic C. reinhardtii strains. The entire procedure, from isolation of soluble proteins through their visualization with a ccd camera, takes only 1 h and could be performed in microtiter plates allowing for the quantification of up to 96 samples simultaneously. The GFP-accumulation values obtained by fluorescence assay in microtiter plates was in good agreement with those generated by the much more time-consuming Western blot method.
Comparison of the GFPct gene with the GFPncb gene reveals a total of 123 codon changes in the optimized versus the native gene. Of these, 122 are synonymous codon changes, with the other two codons resulting in the amino acid substitutions discussed previously. Of the 122 synonymous codon changes, 68 changes represent only a modest shift toward a more optimized codon usage. The remaining 54 codons were changes that result in an infrequently used codon being replaced with a frequently used one. Codon optimization is fairly evenly distributed throughout the GFP gene, with 16 alterations in the first third of the coding region, 19 in the second third, and 19 in the final third. An analysis of genes previously expressed in C. reinhardtii chloroplasts, including Renilla luciferase (Minko et al., 1999); uidA (Sakamoto et al., 1993); aadA (Goldschmidt-Clermont, 1991); and aph A6 coding sequences (Bateman and Purton, 2000) revealed 61, 252, 121 and 65 non-preferred codons in each of these respective genes. If we express the number of non-preferred codons in these reporter genes as a percentage of their total codons we obtain values of 20, 42, 46 and 25%, respectively. This compares with the GFPncb gene, where non-preferred codons account for 23% of the total codons. These data suggest that expression of other reporter genes in C. reinhardtii chloroplasts could be greatly enhanced by altering codon usage.
As we had significantly changed the base composition of the GFP sequence, we examined the effect of these changes on the structure of the mRNA in the GFPct and GFPncb mRNAs. This analysis was primarily run to ensure that the enhancement of translation in the GFPct mRNA was due to the differences in codon usage, rather than to some effects of mRNA secondary structure that might preclude loading of the GFPncb onto ribosomes. We subjected the first 250 nucleotides of the GFPct and GFPncb mRNAs to the RNA-folding program mfold (Mathews et al., 1999; Zuker et al., 1999). No significant secondary structure differences were predicted between the two genes, with the free energy of the most favorable structures being −42 kcal for GFPct and a similar −38 kcal for the GFPncb sequence.
Taken together, our data clearly demonstrate the utility of the optimized GFPct gene as a reporter of chloroplast gene expression in C. reinhardtii. These data also show the importance of codon optimization in achieving high levels of recombinant protein expression in C. reinhardtii, and suggest that the utility of other C. reinhardtii reporter genes could be enhanced by codon optimization. The relatively low levels of GFP mRNA accumulation, compared to the endogenous rbcL mRNA, suggests that optimizing promoter activity and mRNA stability of GFPct could enhance the signal of GFPct to levels higher than those reported here. We are presently using the GFPct gene to optimize transcription, mRNA stability and translation of GFP in C. reinhardtii chloroplasts.
Chlamydomonas reinhardtii strains, transformation and growth conditions
All transformations were carried out on Chlamydomonas reinhardtii strain 137c (mt+). Cells were grown to late log phase (≈7 days) in the presence of 40 mm 5-fluorodeoxyuridine in TAP medium (Gorman and Levine, 1965) at 23°C under constant illumination of 450 lux on a rotary shaker set at 100 r.p.m. Fifty ml of cells were harvested by centrifugation at 4000 g at 4°C for 5 min. The supernatant was decanted and cells resuspended in 4 ml TAP medium for subsequent chloroplast transformation by particle bombardment, as described previously by Cohen et al. (1998). All transformations were carried out under spectinomycin selection (150 µg ml−1) in which resistance was conferred by co-transformation with the spectinomycin resistance ribosomal gene of plasmid p228 (Chlamydomonas Stock Center, Duke University, USA; http://www.biology.duke.edu/chlamy).
Cultivation of C. reinhardtii transformants for expression of GFP was carried out in TAP medium (Gorman and Levine, 1965) at 23°C under constant illumination of 5000 lux on a rotary shaker set at 100 r.p.m. unless stated otherwise. Cultures were maintained at a cell density of 1 × 107 ml−1 for at least 48 h prior to harvest.
All DNA and RNA manipulations were carried out essentially as described by Sambrook et al. (1989) and Cohen et al. (1998). The coding region of the GFP gene was amplified via PCR from a plasmid containing the native GFP (GFPncb) sequence (Tsien, 1998). PCR primers were designed to generate a 5′-NdeI site and a 3′-XbaI site immediately outside the coding region, to facilitate subsequent cloning. The 5′-GFPncb primer has the sequence 5′-CATATGAGTAAAGGAGAAGAAC-3′, while the 3′GFPct primer has the sequence 5′-TCTAGATTATTTGTATAGTTCATCC-3′. The coding region of the GFPct gene was synthesized de novo according to the method of Stemmer et al. (1995) from a pool of primers, each 40 nucleotides in length. The 5′ and 3′ terminal primers used in this assembly contained restriction sites for NdeI and XbaI, respectively. The resulting 717 bp PCR products containing the GFPct and GFPncb genes were then cloned into plasmid pCR2.1 TOPO (Invitrogen Corp., Carlsbad, CA, USA) according to the manufacturer's protocol to generate plasmids pCrGFPct and pCrGFPncb, respectively. The rbcL 3′-UTR was generated via PCR using a 1.6 kb HindIII fragment of C. reinhardtii chloroplast genomic DNA, cloned into plasmid pUC19, as the template. The PCR primer, corresponding to the 5′ end of the rbcL 3′-UTR and a portion of the pUC19 polylinker, including the XbaI site, has the sequence 5′-TCTAGAGTCGACCTGCAG-3′. The PCR primer, corresponding to the 3′ end of the rbcL 3′-UTR, has the sequence 5′-GGATCCGTCGACGTATG-3′ and includes a BamHI restriction site for subsequent cloning. The resulting 433 bp product was cloned into plasmid pCR2.1 TOPO to generate plasmid p3rbcL. The rbcL 5′-UTR was generated by PCR using C. reinhardtii genomic DNA as template. The PCR primer, complementary to the 5′ end of the rbcL gene beginning at position −189 relative to the translational start site, has the sequence 5′-GAATTCATATACCTAAAGGCCCTTTCTATGC-3′ and contains an EcoRI restriction site. The PCR primer complementary to the 3′ end of the rbcL 5′-UTR begins at the translation initiation site and has the sequence 5′-CATATGTATAAATAAATGTAACTTC-3′ and contains an NdeI restriction site. The resulting 241 bp PCR product was cloned into pCR2.1 TOPO to generate plasmid p5rbcL. The plasmid p5rbcL was digested with BamHI and NdeI, and the resulting fragment was ligated into either pCrGFPct or pCrGFPncb digested with BamHI and NdeI to generate plasmids p5CrGFPct and p5CrGFPncb, respectively. Finally, p5CrGFPct and p5CrGFPncb were digested with BamHI and XbaI, and the resulting 958 bp fragments were ligated into p3rbcL, also digested with BamHI and XbaI, to generate plasmids p53rGFPct and p53rGFPncb. Both p53rGFPct and p53rGFPncb were digested with NdeI and BamHI and the 1.2 kb fragments were ligated into pET19b (Novagen Inc., Madison, WI, USA) to generate plasmids pETGFPct and pETGFPncb, respectively, for expression in E. coli. p53rGFPct and p53rGFPncb were next digested with BamHI and the 1.43 kb fragments were ligated into the C. reinhardtii chloroplast transformation vector p322, which contains the 5.7 kb EcoRI/XhoI restriction fragment from the C. reinhardtii inverted repeat region (Chlamydomonas Stock Center) to form plasmids pExGFPct and pExGFPncb.
Southern and Northern blots
Southern blots and 32P-labeling of DNA for use as probes were carried out as described by Sambrook et al. (1989). Radioactive probes used on Southern blots included the 2.2 kb BamHI/PstI fragment of p322 (probe 5′ p322); the 2.0 kb BamHI/XhoI fragment of p322 (probe 3′ p322); and the 717 bp NdeI/XbaI fragments from p53rGFPct (probe GFPct) or p53rGFPncb (probe GFPncb). These latter two probes were also used to detect GFPct and GFPncb mRNAs on Northern blots. Additional radioactive probes used in Northern blot analysis included the psbA and rbcL cDNAs. Northern and Southern blots were visualized utilizing a Pakard Cyclone Storage Phosphor System equipped with optiquant software.
Protein expression, Western blotting and fluorescence gels
Plasmids pETGFPct and pETGFPncb were transformed into E. coli strain BL21 and 6-His-tagged GFPct or GFPncb protein expression induced by Isopropyl B-D-thiogalactopyranoside (IPTG) according to the manufacturer's protocol (Novagen). Purification of His-tagged proteins was carried out using Ni-agarose affinity chromatography (Qiagen, Valencia, CA, USA). Proteins were isolated from C. reinhardtii utilizing a buffer containing 750 mm Tris–Cl pH 8.0, 15% sucrose (wt/vol), 100 mmβ-mercaptoethanol and 1 mm phenylmethylsulfonylfluoride (PMSF). Samples were then centrifuged for 20 min at 13 000 g at 4°C, with the resulting supernatant used in Western blot or fluorescence gel analysis. Chlamydomonas reinhardtii proteins for use in microtiter plate assays of GFP accumulation were prepared in the same fashion, except that the crude lysate was centrifuged for 30 min at 40 000 g at 4°C to remove contaminating chlorophyll. Microtiter assays were carried out on volumes of 100 µl with samples diluted in protein extraction buffer. Protein concentrations were determined using Bio-Rad's Protein assay reagent (Bio-Rad Laboratories Inc., Hercules, CA, USA). Western blots were carried out as described by Cohen et al. (1998) using a rabbit anti-GFP primary antibody (Clonetech, Palo Alto, CA, USA) and an alkaline phosphatase-labeled goat anti-rabbit secondary antibody (Sigma, St Louis, MO, USA). Fluorescence gels were run in the same fashion as gels intended for Coomassie staining or Western transfer, except that proteins were not boiled prior to loading. Green fluorescent protein was visualized in gels or microtiter plates by viewing with a Berthold Night Owl CCD camera, model LB 981, equipped with 485 nm excitation and 535 nm emission filters (Chroma Technology Corp., Brattleboro, UT, USA). Exposure times of 250 msec were sufficient to visualize GFP fluorescence in most cases. Images were generated using winlight software.
Generation of excitation spectra for GFPct and GFPncb
Excitation spectra were generated with affinity-purified GFPct or GFPncb proteins on a Perkin Elmer Luminescence Spectrometer Model LS50 (Perkin Elmer, Shelton, CT, USA). Recombinant proteins were diluted in 50 mm NaH2PO4, 300 mm NaCl, 250 mm imidazole pH 8.0, prior to reading on the spectrometer. Excitation spectra were generated by scanning illumination from 350 to 550 nm, while monitoring emission at 510 nm.
We thank Emma Brown and Aravind Somanchi for a critique of the manuscript, and Jeff Harper for the original GFPncb clone. We also thank Peter Hegemann for supplying the ble-GFP-expressing strain, and Chris Fong and Susana Prieto for generation of the SPcP76 strain. This work was supported by funds from Sea Grant to S.F. and S.P.M., and from the Department of Energy to S.P.M.