In traditional plant genetic engineering, a foreign gene (referred to as a transgene) is inserted into the nuclear genome. A current controversy regarding plant genetic engineering is the possibility of transgene escape to wild relatives through cross-pollination. Chloroplast transformation is emerging as an alternative to nuclear transformation, and may address some of these concerns. In addition to gene containment, other advantages of chloroplast transformation may include the feasibility of obtaining high levels of protein production and the possibility of producing multiple proteins using polycistronic mRNAs.
In nuclear transformation, transgenes are integrated into the genome at random positions. The context of transgene insertion may influence its level of expression, a phenomenon known as “position effect”. In chloroplast transformation, on the other hand, transgenes are integrated by homologous recombination, allowing targeted insertion. Each plant cell has up to 10,000 identical copies of each plastid gene. Therefore, the expression of the transgene in transplastomics is many-fold and, once wild-type plastid genome copies are eliminated through repeated rounds of selection, a genetically stable population can be generated.
In the April issue of Current Opinion of Plant Biology (vol. 5: 164-71), Pal Maliga (Waksman Institute, Rutgers State University of New Jersey) discusses the current status of plastid genome transformation technology in higher plants. In the beginning of the review, the author points out that until recently plastid transformation was only carried out successfully in tobacco. The technology has now been tried in a few other crop species such as tomato and potato, with an appreciable degree of success.
The author describes several vector and selection systems used for plastid transformation. Vectors used for plastid transformation utilize left (LTR) and right (RTR) targeting regions to direct insertion of the transgene into plastid intergenic regions. Two recombination events targeted by the homologous regions direct insertion of the marker gene and the transgene into the LTR and RTR region of the plastid. The author describes several commonly used plasmid transformation vectors, such as the plasmid repeat vector (pPRV) and vectors pRB94 and pRB95, in some detail. In some of these systems, read-through transcription facilitates expression of ribosome binding site regions inserted at intergenic regions, allowing production of the protein(s) of interest from polycistronic mRNA transcripts. Selectable markers currently used in plastid transformation include spectinomycin-streptomycin resistance (conferred by the bacterial aadA gene) and kanamycin resistance (conferred by the neo gene). A positive selection marker is betaine aldehyde hydrogenase, which confers resistance to betaine aldehyde. The bacterial enzyme cytosine deaminase has been used as a negative selection marker system. Cells that express cytosine deaminase convert 5-fluorocytosine to the toxic compound 5-fluorouracil, and transformed seedlings can thus be identified using medium containing 5-fluorocytosine. The reporter genes beta-glucuronidase (GUS) and green fluorescent protein (GFP) have made it relatively easy to detect transient expression of the transgene as well as stable transformation events in chloroplasts. Enzymatic activity of GUS is visualized by histochemical staining, while GFP can be visualized by direct imaging under UV.
The author describes criteria required for production of high amounts of recombinant protein: strong expression (directed by the plasmid promoter) and stable mRNA transcript (determined by the 5’ untranslated region (UTR) and 3’ UTR of the transgene), efficient translation (determined by structure of the 5’ UTR), and protein stability. In light of these considerations, the author describes several plastid expression cassettes in current use. Regulatory regions in plastid expression cassettes include a 5′ regulatory region (PL cassette) and 3′ regulatory region (T cassette). Plastid genome promoters are recognized by plastid-encoded RNA polymerase (PEP) or nucleus-encoded plastid RNA polymerases (NEP). Most plastid transformation systems have used derivatives of the strong sigma-70-type PEP promoter of the rRNA operon promoter (Prrn). The mRNA 3′UTR is encoded by the T cassette, and typically includes an RNA stem-loop structure, which functions as an inefficient transcription terminator. Most T cassettes are derived from plastid pbsA, rbcLand rps16 genes, and differ in degree of stability that they confer to the transgene mRNA. In general, the expression cassettes are designed to optimize transcript stability and translation efficiency.
The author explains that mRNA secondary structure, particularly the 5’ UTR structure, plays an important role in translation efficiency and thus affects the yield markedly. The translation efficiency of chloroplast ribosomes is considerably influenced by the sequence around the AUG initiation codon. As evidence of this, silent mutations near the 5’ end of the coding region of NPTII (neomycin phosphotransferase) caused a 35% reduction in NPTII protein accumulation. Including the amino-terminus of the coding region of the source plastid gene enhances protein accumulation, possibly because sequences surrounding the initiation codon have evolved together to result in efficient translation in the plastid expression system. For example, NPTII accumulation comprised 10.8% of total soluble protein when expressed from a cassette including the rbcLcoding region amino-terminal segment, but only 4.7% in absence of this region.
The author emphasizes that codon usage differs between chloroplast and nuclear genes, necessitating careful optimization of transgene codons. However, despite the prokaryotic nature of the plastid expression system, chloroplast codon use preferences are also distinct from codon usage in E. coli. This was illustrated by experiments using a synthetic, codon-optimized version of CP4 5-enolpyruvylshikimate-3-phosphate synthase (ESPS) containing 77% plastid-preferred codons versus a non-optimized, bacterial CP4 ESPS version. The synthetic version of CP4 ESPS resulted in just 1.5 to 2-fold higher protein levels than the bacterial version. Other such examples indicate that care must be taken when incorporating heterologous coding regions in chloroplast transformation studies.
The author points out that the presence of marker genes in each plasmid genome may result in the marker protein comprising up to 10% of the total soluble protein in host cells. Due to concern about gene flow and possible health hazards of antibiotic resistance genes, the author states that it is desirable to eliminate marker genes after they have been used to select transformants. The author describes two methods developed to remove plastid marker genes. One, developed by Iamtham and Day, relies on loop-out via short, directly repeated sequences. The other, developed independently in two different laboratories (including that of the author), involves a Cre-lox site-specific recombination system.
The author then describes several applications of plasmid transformation. Recently, the rubiscoo large subunit (rbcL of tobacco was replaced with the Form II rubisco of a photosynthetic bacterium (Rhodospirillum rubrum), opening up the possibility of using chloroplast transformation for increasing photosynthetic efficiency of higher plants. The author also describes studies in which metabolic engineering was performed through transformation of the plastid genome to result in higher tryptophan production or increased lipid content in seeds. Applied research studies involving chloroplast transformation include the production of biologically active, soluble and disulfide-bonded somatotropin in chloroplasts, as well as the expression of cholera toxin gene in chloroplasts as a candidate oral vaccine antigen.
Plastid transformation has been successfully performed in several solonaceous species, such as potato and tomato, although plastid transformation in other systems remains problematic. Tomato plastid transformation has resulted in relatively high levels of protein production in tomato fruits, and holds promise for production of oral vaccines. Plastid transformation in Arabidopsis is presently feasible but inefficient, and although the plastid transformation in rice is feasible, regeneration of plants from cultured cells remains difficult.
In conclusion, the author lists some of the possible advantages of plastid transformation, including gene containment, expression of multiple genes, lack of position effects, high expression levels, and the possibility of expressing unmodified human and bacterial cDNAs. Noting the possible development of tools for chemically inducible promoter systems, the author expresses optimism that plastid transformation technology will continue to be a promising approach. The author also discusses the prospects for using plastid transformation to limit transgene flow to wild weedy relatives in greater detail.