Expression of a Synthetic Gene for the Major Cytotoxin (Cyt1Aa) of Bacillus thuringiensis subsp. israelensis in the Chloroplast of Wild-Type Chlamydomonas

Chlamydomonas reinhardtii (Chlamydomonas) strains that are toxic to mosquito larvae because they express chloroplast transgenes that are based on the mosquitocidal proteins of Bacillus thuringiensis subsp. israelensis (Bti) could be very useful in mosquito control. Chlamydomonas has several advantages for this approach, including genetic controls not generally available with industrial algae. The Bti toxin is produced by sporulating bacteria and has been used for mosquito control for >30 years without creating highly resistant mosquito populations. The suite of toxins is four main proteins: three Cry proteins and the cytotoxic Cyt1Aa (27 kDa). Cyt1Aa is not very toxic to mosquitoes by itself, but it prevents the development of resistance. The production of Cyt1Aa in other microbes, however, has been challenging due to its affinity for certain membrane phospholipids. Here we report on the production of recombinant Cyt1Aa (rCyt1A) in the chloroplast of photosynthetic Chlamydomonas at levels of at least 0.3% total protein. Live cell bioassays demonstrated toxicity of the rCyt1Aa Chlamydomonas to larvae of Aedes aegypti. We also expressed the chloroplast cyt1Aa gene in a wild-type Chlamydomonas strain (21 gr) that can grow on nitrate. These results have implications for developing a Chlamydomonas strain that will be toxic to mosquito larvae but will not induce strongly resistant populations.


Introduction
Mosquitoes are vectors for some important parasites and viruses that cause diseases in humans, which include the malarial parasite, plasmodium, and the emerging viruses, West Nile, chikungunya and Zika [1]. Since treatments for the aforementioned viruses are generally lacking, and notwithstanding the sexual transmission of Zika, vector control is essential for limiting the prevalence of these diseases. Mosquito control has proven to be an effective approach, that is both preventative and powerful because it blocks the transmission of multiple diseases simultaneously. The preferred strategy is integrated pest management (IPM), which targets all stages of the insect life cycle. The pesticides used in IPM are increasingly problematic, especially the chemicals that kill adult insects, which are toxic to non-target organisms, such as honey bees. Moreover, the development of in the chloroplast (using a wild-type background) at substantial levels and with only a minor effect on growth rate. In addition, we demonstrate chloroplast transformation and cyt1Aa expression in a wild-type strain of Chamydomonas that can grow on nitrate as sole nitrogen source (21 gr).

DNA Constructs
The cyt1Aa DNA sequence of B. thuringiensis israelensis (NCBI NC_010076.1, nt 17372-18121, gene ID: 5759908) was optimized with the program Optimizer [34] and a codon-usage table for chloroplast genes of Chlamydomonas reinhardtii. Also, an 8-aa FLAG peptide (DYKDDDDK) was added to the C-terminus, and A297 was changed to T297 to remove the internal Nde I site. The gene was synthesized using 26 DNA primers that were 50 nt in length and overlapped flanking primers by 25 nt [35]. The DNA polymerase was Phusion (NEB, Ipswich, MA, USA), and the thermocycle program was 55 cycles of: 30 s at 94 • C, 30 s at 52 • C, and 30 s at 72 • C. The product was purified and used for PCR with outside primers to produce only the full-length gene; the thermocycle program was 23 cycles of 30 s at 94 • C, 30 s at 50 • C, and 30 s at 72 • C. The product was double-digested and cloned into pBluescript using the Nde I (5 ) and Xba I (3 ) sites. The psbD m :cyt1Aa:psbA gene was assembled in the low-copy vector pET-16B: the optimized cyt1Aa sequence was excised from pBluescript by cleaving with Xba I, blunting with Klenow DNA polymerase, and digesting with Nde I. It was ligated to pET-16B that had been cut with Bam HI, blunted with Klenow DNA polymerase, and digested with Nde I. The new plasmid, pET-Cyt1Aa, was confirmed by Sanger sequencing (University of Texas at Austin DNA Facility). For chloroplast expression, the psbD m (5 ) and psbA (3 ) regions used for expressing Cry genes [19] were ligated to cyt1Aa in pET-Cyt1Aa, creating the psbD m :cyt1Aa:psbA gene.
For transformation of strain CC-373, we modified a plasmid used previously for that purpose, P-655 [33]. P-655 was cut with Xho I and Xba I to remove the rps7:aadA:rbcL gene, and the ends were blunted with Klenow (plus dNTP's), dephosphorylated, and ligated to the psbD m :cyt1Aa:psbA gene, which had been excised from pET-16B with Bam HI and blunted with Klenow. E. coli transformants were analyzed by restriction with Nde I to distinguish the orientation of the psbD m :cyt1Aa:psbA gene.
To transform the 21 gr wild-type strain (CC-1690), the psbD m :cyt1Aa:psbA gene was excised from pET-16B with Bam HI and cloned into the chloroplast expression vector, p322-483aadA. The p322-483aadA vector was created by inserting the recyclable selectable marker 483aadA [26] into plasmid P-322, which has a cpDNA fragment that extends from the Eco RI site in the 4th intron of psbA (Cr.psbA4) to the Xho I site in the Cr.LSU intron of the 23S rRNA gene (NCBI ref. NC_005353.1). The 483aadA marker has the atpA:aadA:rbcL gene that confers spectinomycin resistance flanked by a direct repeat (DR) of 483 bp. The insertion of the marker in p322 re-created the Bam HI site between the psbA and 5S genes, and this was used for inserting the Bam HI-excised psbD m :Cyt1Aa:psbA gene.

Chloroplast Transformation and DNA Analysis
Biolistic transformation using S550d gold carrier particles (Seashell Technology) was performed as described previously [19]. Selection was on TAP-minimal plates for strain CC-373 (because repair of the atpB deletion restores photosynthesis and growth on minimal medium), and on TAP + spectinomycin (300 µg mL −1 ) plates for transformation of wild-type strain CC-1690. For the CC-373 transformation, we obtained 20-100 transformants per plate, which equaled 40-200 transformants per shot, since each shooting was replated on two selective plates. Transformation of CC-1690 gave about half as many transformants as did CC-373.
For analysis of the CC-373 transformants, DNA was extracted from cells grown on agar plates as described [36] and analyzed by PCR with Taq DNA polymerase (NEB) and the manufacturer's buffer.
The initial PCR test was with Cyt1Aa-internal primers, 886 (5 -ccaggcatatggaaaatttaaatcattgtccattagaagatattaaag-3 ) and 899 (5 -tcatgtctagattatttgtcgtcgtcgtctttgtagtctaatg-3 ), which bound to the 5 and 3 ends, respectively, of the Cyt1Aa coding sequence. Subsequent reactions used external primers that annealed adjacent to the gene-insertion site; these were 1009 (5 -cctcctaacggagcattaaaatc-3 ), which bound to nt 159528-159550 of cpDNA, and 1008 (5 -ccaaattatatttgtcgtccacgag-3 ), which bound to nt 160160-160136 of cpDNA. The external primers were used to test for homoplasmicity, as they gave a PCR product of 633 bp if there was no psbD m :cyt1Aa:psbA gene inserted and a product of~1850 bp if it was present. As no antibiotic resistance marker was employed, 2 to 4 rounds of subcloning (selecting each time for the colonies with the most copies of the psbD m :cyt1Aa:psbA gene) was required for copy-correction between the 2 copies of the inverted repeat to produce homoplasmicity [33]. The subcloning was performed using regular TAP plates, rather than TAP-minimal, since homoplasmicity of the repaired atpB gene occurred quickly on the original TAP-minimal selection plate. Only about 10% or less of the transformants also became homoplasmic for the presence of the psbD m :cyt1Aa:psbA gene.
For analysis of the CC-1690 transformants, the DNA was also analyzed first by PCR with cyt1Aa-internal primers (886 and 899, described above), and then with external flanking primers to determine homoplasmicity. The flanking primers annealed upstream and downstream, respectively, of the co-inserted gene cassettes, psbD m :cyt1Aa:psbA and DR-atpA:aadA:rbcL-DR (where DR is a direct repeat of 483 bp) and were 863 (5 -taacccataaatagtttcaattggaa-3 , binding at position 147143-147168 of cpDNA) and 975 (5 -aatgcaaagtaccatcagatattgcta-3 , binding at position 147266-147241 of cpDNA), respectively. They gave a PCR product of 124 bp if there was no insertion (i.e., untransformed DNA) and~4300 bp with the inserted genes. Selection for spectinomycin resistance by growth on TAP plates containing 300 µg mL −1 spectinomycin (in the light) hastened the attainment of homoplasmicity, and most of the transformants that survived this selection became homoplasmic.

Protein Extraction and Western Blotting
For the extraction of total cellular protein, 50 mL of cell culture was centrifuged at 900× g for 10 min (at RT), and the cells were resuspended in 1 mL of 100 mM Tris-HCl pH 8.5, 100 mM DTT, 7 mM benzamidine, and 5 mM EDTA pH 8.0. Then, 0.6 mL was removed and added to 0.4 mL of SDS solution (5% sodium dodecylsulfate, 30% sucrose, 0.025% bromophenol blue), and the lysates were stored at −70 • C. Total chlorophyll was determined from an aliquot of the cells and used to estimate protein using a mass ratio that we previously determined to be close to 1:10 chlorophyll:protein.
For SDS-PAGE, the solubilized cells, plus purified Cry11Aa [19] in some samples, were heated to 100 • C for 5 min before loading on 10% polyacrylamide SDS-PAGE gels along with a pre-stained protein ladder (PageRuler, Fermentas, Waltham, MA, USA) as size markers. After electrophoresis, the gels were soaked in cold transfer buffer (25 mM Tris base, 192 mM glycine, 5% methanol) for 15 min and then electrotransferred to a PVDF membrane (Hybond-P, GE HealthCare, Pittsburgh, PA, USA) using a Genie blotter (Idea Scientific, Minneapolis, MN, USA) at 12 V for 2 h at 4 • C [37]. The membrane was stained with Ponceau S to check the protein transfer, and then blocked overnight at 4 • C with 5% nonfat dried milk in TBS-T (20 mM Tris-HCl pH 7.6,150 mM NaCl plus 0.05% Tween 20). After washing twice with 0.5% nonfat milk in TBS-T for 5 min each, the membrane was incubated for 2 h (RT) with an anti-FLAG monoclonal antibody (M2) coupled to alkaline phosphatase (Sigma-Aldrich A9469, St. Louis, MO, USA) that was diluted 1:5000 in TBS-T plus 0.5% nonfat milk. After antibody incubation, the membrane was rinsed twice (2 min ea) with 50 mL of 0.5% nonfat milk in TBS-T, and then four times (10 min ea) with 50 mL of TBS-T. Finally, the membrane was rinsed with purified water for 5 min, overlaid with 2 mL of CDP-Star substrate (Invitrogen, Carlsbad, CA, USA), and incubated for 3 min. The substrate was decanted, the edges of the membrane blotted with a Kimwipe (Kimberly-Clark, Irving, TX, USA), and then put in a sheet protector to prevent drying. After 40 min, the membrane was exposed to X-ray film for periods ranging from 10 s to 3 min.

Mosquito Bioassay
The algal strains used for the live cell bioassay with mosquito larvae were grown in liquid TAP medium to 2-5 × 10 6 cells/mL, except for the Cry11Aa strain which was grown in TAP-minus copper [19]. The cells were harvested by centrifugation, washed with 40 mL of once-distilled water, and resuspended in water to 1 × 10 7 cells/mL. For each assay trial, 17 mL of the resuspended cells was added to 17 mL of once-distilled water containing 14-16 mosquito larvae, giving a total vol of 34 mL (with 5 × 10 6 cells/mL of algae). The mosquitoes were 3rd instar larvae of Aedes aegypti (Galveston), which had been raised on mosquito diet (Carolina Biological) and then starved for 48 h prior to the assay. The plastic cups were covered with clear plastic wrap and secured with a rubber band to prevent the escape of adults. The assays were incubated at 27 • C under a 12 h:12 h light-dark cycle, and were monitored daily for 12 cycles, at which time all larvae had either died or developed into adults.

cyt1Aa Gene Engineering and Chloroplast Transformation
The codon-optimized cyt1Aa gene had a codon adaptive index (compared to Chlamydomonas chloroplast genes) of 1, but the same amino acid sequence as Bti Cyt1Aa. An 8-amino acid FLAG tag was added to the C-terminus to enable detection with a commercial antibody; and it was expected to have little or no effect on toxicity as both ends of Cyt1Aa are cleaved away by larval gut proteases [38]. The expression signals from the psbD (promoter and 5 UTR) and psbA (3 UTR) genes were used so that the cyt1Aa gene could be expressed using the inducible system-which requires that the target gene have a psbD 5 UTR and a special host strain [39]-or constitutively in a wild-type background. Since we obtained good expression in the constitutive system, we did not attempt the inducible system with this gene.
To achieve expression in a wild-type background, the psbD m :cyt1Aa:psbA gene was transformed into the chloroplast of the CC-373 strain, which has a deletion of the atpB gene [33]. This strain is non-photosynthetic due to the deletion, but the defect is complemented by transformation, thus producing a strain resembling wild-type, which can be selected on minimal medium. The transforming DNA simultaneously introduces the psbD m :cyt1Aa:psbA gene into the inverted repeat (IR) near the atpB gene ( Figure 1). Since the psbD m :cyt1Aa:psbA transgene went into the cpDNA plasmid in both orientations relative to the orientation of the atpB gene (i.e., Reverse and Forward), we used both types of plasmids for transformation (see Donor DNA in Figure 1). Homoplasmicity is when all copies of the genome are the same, and this was achieved quickly at the atpB locus. However, several rounds of replating were required to achieve homoplasmicity at the psbD m :cyt1Aa:psbA locus, which because of the copy-correction process that occurs with the IR, can be completely retained, or lost in different clones. As shown in Figure 1, we obtained homoplasmic transformants for all

Analysis of rCyt1Aa Protein Accumulation in the psbDm:cyt1Aa:psbA Transformants
Having obtained transformants that were essentially wild-type due to the complementation of the atpB gene, we examined the rCyt1Aa protein on western blots of total Chlamydomonas protein ( Figure 2). Strong signals, relative to the standard (lane 2), and of the correct size, ~28 kDa, were obtained for both orientations of the psbDm:cyt1Aa:psbA gene, although the forward orientation produced more protein (Figure 2). The transformant that lacked the psbDm:cyt1Aa:psbA gene due to copy-correction with the other half of the IR during the cpDNA sorting-out period, gave no corresponding protein signal, as expected. The faint diffuse band of ~53 kDa in all the lanes is probably due to some non-specific sticking of the antibody to the very abundant proteins that migrate at that location; it serves as a fortuitous indicator of similar protein loading and transfer for the lanes. For lane 2, 50 ng of purified FLAG-tagged Cry11Aa [19] was added to the Chlamydomonas protein as a standard to provide a lower-limit estimate of FLAG-tagged Cyt1Aa. The results suggest that the psbDm:cyt1Aa:psbA transformants have at least 60 ng of rCyt1Aa per 20 µg total protein, or at least 0.3% of total protein is rCyt1Aa.

Analysis of rCyt1Aa Protein Accumulation in the psbD m :cyt1Aa:psbA Transformants
Having obtained transformants that were essentially wild-type due to the complementation of the atpB gene, we examined the rCyt1Aa protein on western blots of total Chlamydomonas protein ( Figure 2). Strong signals, relative to the standard (lane 2), and of the correct size,~28 kDa, were obtained for both orientations of the psbD m :cyt1Aa:psbA gene, although the forward orientation produced more protein (Figure 2). The transformant that lacked the psbD m :cyt1Aa:psbA gene due to copy-correction with the other half of the IR during the cpDNA sorting-out period, gave no corresponding protein signal, as expected. The faint diffuse band of~53 kDa in all the lanes is probably due to some non-specific sticking of the antibody to the very abundant proteins that migrate at that location; it serves as a fortuitous indicator of similar protein loading and transfer for the lanes. For lane 2, 50 ng of purified FLAG-tagged Cry11Aa [19] was added to the Chlamydomonas protein as a standard to provide a lower-limit estimate of FLAG-tagged Cyt1Aa. The results suggest that the psbD m :cyt1Aa:psbA transformants have at least 60 ng of rCyt1Aa per 20 µg total protein, or at least 0.3% of total protein is rCyt1Aa.

Growth Rates for the psbDm:cyt1Aa:psbA Transformants
Although it was apparent that the psbDm:cyt1Aa:psbA transformants grew well, it was possible that the psbDm:cyt1Aa:psbA gene (or rCyt1Aa protein) could be having a negative effect on the cells, and that it might show up in growth rate analysis. Therefore, liquid cultures of the same transformants analyzed in Figures 1 and 2 were monitored for ~82 h during growth in the light in standard TAP medium ( Figure 3). As the data shows, the growth curve for the psbDm:cyt1Aa:psbA transformant in reverse orientation, Cyt1A (Rev), is very similar to the control strain lacking the cyt1Aa gene (no Cyt1A). However, the psbDm:cyt1Aa:psbA transformant in the forward orientation, Cyt1A (For), which had the higher level of Cyt1Aa protein (Figure 2), grew at a rate that was ~20% slower. The doubling time was ~10 h for the control (no Cyt1A) and Cyt1A (Rev) strains, and ~12 h for the Cyt1A (For) strain. This result suggests that there is a lack of rCyt1Aa toxicity at the level found in Cyt1A (Rev), but there is some host toxicity with the higher level of rCyt1Aa found in the Cyt1A (For) transformant.

Growth Rates for the psbD m :cyt1Aa:psbA Transformants
Although it was apparent that the psbD m :cyt1Aa:psbA transformants grew well, it was possible that the psbD m :cyt1Aa:psbA gene (or rCyt1Aa protein) could be having a negative effect on the cells, and that it might show up in growth rate analysis. Therefore, liquid cultures of the same transformants analyzed in Figures 1 and 2 were monitored for~82 h during growth in the light in standard TAP medium ( Figure 3). As the data shows, the growth curve for the psbD m :cyt1Aa:psbA transformant in reverse orientation, Cyt1A (Rev), is very similar to the control strain lacking the cyt1Aa gene (no Cyt1A). However, the psbD m :cyt1Aa:psbA transformant in the forward orientation, Cyt1A (For), which had the higher level of Cyt1Aa protein (Figure 2), grew at a rate that was~20% slower. The doubling time was~10 h for the control (no Cyt1A) and Cyt1A (Rev) strains, and~12 h for the Cyt1A (For) strain. This result suggests that there is a lack of rCyt1Aa toxicity at the level found in Cyt1A (Rev), but there is some host toxicity with the higher level of rCyt1Aa found in the Cyt1A (For) transformant.

Growth Rates for the psbDm:cyt1Aa:psbA Transformants
Although it was apparent that the psbDm:cyt1Aa:psbA transformants grew well, it was possible that the psbDm:cyt1Aa:psbA gene (or rCyt1Aa protein) could be having a negative effect on the cells, and that it might show up in growth rate analysis. Therefore, liquid cultures of the same transformants analyzed in Figures 1 and 2 were monitored for ~82 h during growth in the light in standard TAP medium ( Figure 3). As the data shows, the growth curve for the psbDm:cyt1Aa:psbA transformant in reverse orientation, Cyt1A (Rev), is very similar to the control strain lacking the cyt1Aa gene (no Cyt1A). However, the psbDm:cyt1Aa:psbA transformant in the forward orientation, Cyt1A (For), which had the higher level of Cyt1Aa protein (Figure 2), grew at a rate that was ~20% slower. The doubling time was ~10 h for the control (no Cyt1A) and Cyt1A (Rev) strains, and ~12 h for the Cyt1A (For) strain. This result suggests that there is a lack of rCyt1Aa toxicity at the level found in Cyt1A (Rev), but there is some host toxicity with the higher level of rCyt1Aa found in the Cyt1A (For) transformant.  At the indicated times, aliquots (1 mL) were withdrawn, and total chlorophyll was quantified and converted to cell density.

Bioassay with Aedes Aegypti Larvae
Although Cyt1Aa has relatively poor toxicity to mosquito larvae by itself [5,13,31], it has been reported to be detrimental and even lethal in several cases [5]. To assess our rCyt1Aa-producing algae, we performed larval bioassays where 14-16 3rd instar larvae of Aedes aegypti (Galveston) were incubated with live algal cells in dH 2 0. Under these conditions, the algae do not grow even though they are incubated under a 12 h:12 h light-dark cycle at 27 • C; moreover, they are readily consumed by hungry larvae. The initial assays ran for up to 72 h, and we saw a small increase in larval mortality with the rCyt1Aa strains (not shown). However, with a longer time frame,~1/3 of the larvae died (Table 1) with the rCyt1Aa strains, versus 1/8th for the No Cyt1A control. For a positive control, we used a highly toxic Cry11Aa strain created previously that produces Cry11Aa under inducing conditions tõ 0.3% of total protein (is rCry11Aa) [19]; that strain was again highly toxic, killing~88% of the larvae.  [19] was grown without copper to induce the psbD m :cry11A:psbA gene.

Transforming a Wild-Type (WT) Strain Capable of Growth on Nitrate
Nearly all published papers that involve transforming the chloroplast in Chlamydomonas use host strains that are derivatives of the 137c or Ebersold/Levine line of C. reinhardtii, which is incapable of growing on nitrate as sole nitrogen source because it lacks nitrate reductase due to mutations [24]. Since the ability to grow on nitrate could be important for a Chlamydomonas strain that was to be used for mosquito control, we have explored transforming the chloroplast of CC-1690, better known as Ruth Sager's wild-type strain, or 21 gr. This strain is wild-type for nitrogen assimilation and can grow on nitrate [24].
To transform the chloroplast of CC-1690 (21 gr), we used an antibiotic resistance marker (483aadA) that also has the potential to be removed later by growth under non-selective conditions [26]. We elected to introduce the psbD m :cyt1Aa:psbA gene, since it worked well in CC-373 above, and in both orientations relative to the aadA-based marker, which confers spectinomycin resistance. The genes were integrated at a different location than in CC-373 (Figure 1), between the psbA and 5S rRNA genes in the IR ( Figure 4A), but homoplasmic transformants were identified similarly, by amplifying the integration site using flanking primers ( Figure 4B). The small (~0.124 kb) product that would be expected with these primers if there was no integration is completely absent in the Cyt1A-Fwd and Cyt1A-Rev transformants, because they are homoplasmic for integration. The 4 kb product is the size expected for amplification of both genes in tandem, whereas the~1.9 kb product is an artifact of PCR, caused by the sizable (483 bp) direct repeat (DR) at the ends of the atpA:aadA:rbcL marker. Consistent with this notion is the fact that this product was also obtained with the plasmids that were introduced into the cells, and the result with pCyt1A-Fwd is shown ( Figure 4B, lane 5). In summary, homoplasmic transformants of CC-1690 with the psbD m :cyt1Aa:psbA gene were obtained for further analysis.

Expression and Growth Characteristics of the CC-1690 Transformant
Western blotting was used to gauge the expression of the psbDm:cyt1Aa:psbA gene in CC-1690 using the FLAG antibody and total Chlamydomonas protein. Since both orientations of the psbDm:cyt1Aa:psbA gene gave levels of rCyt1Aa protein on western blots that were indistinguishable (not shown), only one orientation was analyzed for Figure 5. A single major band of the expected size, ~28 kDa, was obtained, and comparing it to the FLAG-tagged standard (60 ng of Cry11A Std, lane 1) the level of rCyt1Aa appears to be about 60 ng per 20 µg total protein, or ~0.3% of total protein.
We also examined the growth rate of the rCyt1Aa strain compared to the parental WT strain (CC-1690) in TAP medium under continuous light. The initial analysis was performed over about 80 h and growth of the rCyt1Aa strain was slower than WT (not shown). However, to see if the Cyt1Aa strain would eventually make it to a final density similar to WT, we monitored another set of cultures for 150 h (6.25 days) ( Figure 5B). The rate of growth of the rCyt1Aa strain was again ~20% slower The arrows above the map indicate that the psbD m :cyt1Aa:psbA gene was cloned separately in both orientations, Reverse and Forward. The selectable marker gene, atpA:aadA:rbcL, which was only in the Reverse orientation, is flanked by 483-bp direct repeats (DR). E and X are restriction sites Eco RI and Xho I, respectively. (B) PCR analysis of homoplasmic transformants. PCR was performed with total DNA from the indicated strains (and plasmid) with primers (863 + 975) that flank the inserted DNA, to test for homoplasmicity. Lanes: M, DNA size markers; WT, CC-1690 host strain; Cyt1A-Fwd, transformant with psbD m :cyt1Aa:psbA in the Forward orientation; Cyt1A-Rev, transformant with psbD m :cyt1Aa:psbA in the Reverse orientation; pCyt1A-Fwd, plasmid that was transformed into CC-1690 to create the Cyt1A-Fwd transformant; -DNA, control reaction minus-DNA. Note: The~1.8 kb band in lanes 3-5 is most likely a PCR artifact mediated by the large direct repeat (DR) at the ends of the selectable marker, which causes a deletion of the sequence between the DRs plus one copy of the DR, or~2.5 kb.

Expression and Growth Characteristics of the CC-1690 Transformant
Western blotting was used to gauge the expression of the psbD m :cyt1Aa:psbA gene in CC-1690 using the FLAG antibody and total Chlamydomonas protein. Since both orientations of the psbD m :cyt1Aa:psbA gene gave levels of rCyt1Aa protein on western blots that were indistinguishable (not shown), only one orientation was analyzed for Figure 5. A single major band of the expected size,~28 kDa, was obtained, and comparing it to the FLAG-tagged standard (60 ng of Cry11A Std, lane 1) the level of rCyt1Aa appears to be about 60 ng per 20 µg total protein, or~0.3% of total protein.
We also examined the growth rate of the rCyt1Aa strain compared to the parental WT strain (CC-1690) in TAP medium under continuous light. The initial analysis was performed over about 80 h and growth of the rCyt1Aa strain was slower than WT (not shown). However, to see if the Cyt1Aa strain would eventually make it to a final density similar to WT, we monitored another set of cultures for 150 h (6.25 days) ( Figure 5B). The rate of growth of the rCyt1Aa strain was again~20% slower than WT, but it eventually reached a cell density (8 × 10 6 cells/mL) close to that of WT (8.2 × 10 6 cells/mL).
The reduced growth rate of the cyt1Aa transformant of CC-1690 ( Figure 5B) and of the cyt1Aa transformant of CC-373 (Cyt1A (For) in Figure 3) suggests that there is some weak toxicity associated with producing rCyt1Aa in the chloroplast at substantial levels. However, there was at least one other explanation, that the psbD m :cyt1Aa:psbA mRNA was competing with the psbD mRNA for the translational factors that bind to the psbD 5 UTR [40] and reducing psbD expression. As a simple test of this hypothesis, we examined growth curves of the cyt1Aa transformant (in CC-1690) and WT in the dark, where psbD expression levels should matter very little, because there is no photosynthesis. As Figure 5C shows, the growth rate of Cyt1Aa was still slower than WT, and possibly even more pronounced. This suggests that the reduction in growth rate of the cyt1Aa transformant compared to wild-type is probably due to some toxicity associated with producing Cyt1Aa in the chloroplast at substantial (>0.2%) levels.
The reduced growth rate of the cyt1Aa transformant of CC-1690 ( Figure 5B) and of the cyt1Aa transformant of CC-373 (Cyt1A (For) in Figure 3) suggests that there is some weak toxicity associated with producing rCyt1Aa in the chloroplast at substantial levels. However, there was at least one other explanation, that the psbDm:cyt1Aa:psbA mRNA was competing with the psbD mRNA for the translational factors that bind to the psbD 5′ UTR [40] and reducing psbD expression. As a simple test of this hypothesis, we examined growth curves of the cyt1Aa transformant (in CC-1690) and WT in the dark, where psbD expression levels should matter very little, because there is no photosynthesis. As Figure 5C shows, the growth rate of Cyt1Aa was still slower than WT, and possibly even more pronounced. This suggests that the reduction in growth rate of the cyt1Aa transformant compared to wild-type is probably due to some toxicity associated with producing Cyt1Aa in the chloroplast at substantial (>0.2%) levels.

Discussion
We have demonstrated that the cytotoxic Cyt1Aa protein from the mosquitocidal bacterium, Bti, can be produced in the chloroplast at reasonably high levels in wild-type photosynthetic cells.

Discussion
We have demonstrated that the cytotoxic Cyt1Aa protein from the mosquitocidal bacterium, Bti, can be produced in the chloroplast at reasonably high levels in wild-type photosynthetic cells. Although the CC-373 host strain was not photosynthetic before chloroplast transformation, it was after introducing the psbD m :cyt1Aa:psbA and atpB genes, the latter of which repaired the deletion in CC-373 and restored photosynthesis [33]. Although we probably could have achieved a higher level of rCyt1Aa with the inducible system [19,39]-because with that system the psbD m :cyt1Aa:psbA mRNA would not compete with psbD mRNA for translation factors that bind the psbD 5 UTR [40]-we would still have had to develop conditions for cyt1Aa expression in a wild-type background. Inducible expression of toxin would not be desirable in a mosquito-control strain, since induction conditions would probably not be present in diverse field sites. Thus, these results have direct implications for developing a larvacidal Chlamydomonas strain that will not engender strong resistance because it produces Cyt1Aa constitutively [5].
We also expressed the psbD m :cyt1Aa:psbA gene in the chloroplast of a wild-type strain (21 gr, CC-1690) that can grow on nitrate as sole nitrogen source. To our knowledge, this is one of a very few, if not the only, report of chloroplast transformation that does not involve a 137c (Ebersold/Levine) line. Chlamydomonas strains carried in captivity can change due to genetic drift, and the 137c line lost nitrate reductase, which is needed for nitrate assimilation, prior to it becoming the favored strain for chloroplast researchers [24]. Although Chlamydomonas can grow on ammonia or even certain amino acids as nitrogen sources, these compounds can be limiting for growth in many natural habitats. Thus, a strain with great versatility in nitrogen utilization, such as 21 gr, should be better at growing in diverse larval habitats.
We were fortunate that the psbD m :cyt1Aa:psbA gene expressed well in a wild-type strain, as high-level expression of transgenes in wild-type Chlamydomonas chloroplasts remains a challenge. Indeed, the Cry11Aa and Cry4A 700 genes that expressed well in the inducible system [19], expressed poorly when they were transformed into wild-type strains (S. Kang, O.W. Odom and D.L. Herrin, unpublished results). Why the psbD m :cyt1Aa:psbA gene expresses well in the presence of a normal psbD mRNA, while the psbD m :cry11A:psbA and psbD m :cry4A 700 :psbA genes do not is not clear, but it indicates that the coding region can be critical in transgene expression. Moreover, the answer is not as simple as a more-compatible coding region (with the psbD m :cyt1Aa:psbA gene), because expression of cyt1A using two other promoter-5 UTR combinations (atpA and 16S-atpA, respectively) did not yield high levels of rCyt1Aa in the CC-373 background (O.W. Odom and D.L. Herrin, unpublished results). Perhaps there are interactions between the 5 UTR and coding regions of certain chloroplast mRNAs that affect their levels and/or translational efficiency.
Toxicity to heterologous hosts, and in some cases instability, has hindered transgenic cyt1Aa expression; however, in bacteria this was overcome by co-expressing another protein from Bti called P20 [13,31], which may be a chaperonin-like protein [32]. Since the chloroplast generally lacks the membrane phospholipids that attract Cyt1Aa, we posited that host toxicity might be less of a problem with chloroplast expression, and that seems to be the case, as we readily obtained chloroplast transformants producing rCyt1Aa. Moreover, we have not seen evidence of genetic instability with these transformants.
However, in two of the three Cyt1Aa transformants that we examined closely, there was evidence of a reduced growth rate (10-20%) in liquid culture relative to the control strain. Although that would not be a desirable trait for a strain that was to be used for mosquito control, we are not overly concerned at this point, because we are looking to co-express cyt1Aa with either cry11Aa or cry4Aa 700 , which are known to interact with Cyt1Aa [5] and could mitigate damaging effects of rCyt1Aa on the chloroplast. If there is still evidence of rCyt1Aa toxicity in the co-expressing strains, we can localize rCyt1Aa within the organelle, to keep it away from membranes, and that may prevent such damage. To that end, we have explored binding rCyt1Aa to starch granules using a starch-binding domain [41]. In preliminary experiments, the starch-binding domain eliminated the growth reduction in the CC-1690 strain caused by cyt1Aa expression, but it also reduced the level of the Cyt1Aa fusion protein, making it difficult to determine what actually restored the growth rate (S. Kang, O.W. Odom and D.L. Herrin, unpublished results). Nonetheless, this remains a potential approach to deal with Cyt1Aa host toxicity.

Conclusions
We have shown that a synthetic cyt1Aa gene can be expressed constitutively in the chloroplast of an essentially wild-type Chlamydomonas, with the rCyt1A protein (28 kD) accumulating to levels of at least 0.3% total protein. The rCyt1Aa strain was also shown to be toxic to larvae of Aedes aegypti, although not nearly as toxic as a Cry11Aa-producing strain, which was expected. We also expressed the cyt1Aa gene in the chloroplast of a wild-type strain that can grow on nitrate as sole nitrogen source (CC-1690 or 21 gr), which is a trait that could be important for a strain that is to be used for larval control in aquatic habitats. To that end, we are currently working to produce strains that constitutively express genes for both Cry11Aa and Cyt1Aa (at high levels) in the same chloroplast, and in a wild-type strain such as CC-1690.