Chloroplast Genetics
|
by: Sebastian Molnar
Abstract. The endosymbiotic gene transfer theory has become an important and interesting aspect in genetics and evolution, especially in plants. Chloroplasts are thought to be descendents from an ancestral cyanobacterium. Comparisons of chloroplast genomes and gene sequences (as well as plastid biochemistry and ultrastructure) have been the common methods of elucidating phylogenetic relationships. One of the big questions in plant evolution, is whether all plastids are monophyletic or polyphyletic. The concensus is that plastids in green plants and green algae are monophyletic. The rhodophytes (red algae) seem to have had an early divergence before the plant, fungus, and animal lineage, and therefore rhodoplasts may be polyphyletic to the green plastids. Of more fundamental concern is how those relationships have come about. What mechanisms are responsible? During endosymbiotic events and throughout evolution, there are gene transfers from symbiont (or "pre-organelle") to host, resulting in a tightly integrated intergenomic regulation between the organelles and the nucleus. Although chloroplast DNA (cpDNA) is, in general, highly conserved both in genomic structure and gene order, there are many species differences. Herein is a review of the current knowledge of chloroplast genetics in relation to endosymbiotic gene transfer and intercompartmental genetic interactions.
Introduction. The chloroplast genome has been a major focus in studying plant evolution and plant genetics (Golenberg et al., 1993; Clegg et al. 1994; Morton, 1995; Clegg et al. 1997; Morton, 1999; Stoebe and Kowllik, 1999). It is now commonly believed that chloroplasts are the consequence of an endosymbiotic event between a eukaryotic host cell and an ancestor of the cyanobacteria (Curtis and Clegg, 1984; Delwiche et al. 1995; Barbrook et al., 1998; Turmel et al. 1999). Plastids developed either from a primary endosymbiotic event (a photosynthetic prokaryote with a nonphotosynthetic eukaryote) or from a secondary event (a photosynthetic eukaryote with a nonphotosynthetic eukaryote). One of the main points of conjecture is the whether all plastids are monophyletic or polyphyletic (Gray, 1989). The evidence appears to overwhelmingly support a monophyletic origin (Delwiche et al., 1995), yet some cases are not so clear cut (Penny and O'Kelly, 1991; Lockhart et al. 1992). Plastids in the red algae appear to be of polyphyletic origin relative to the green plastid lineage. The rhodophytes probably acquired their plastids early on and diverged from the lineage of the green plants, fungi and animals (Douglas et al., 1991; Stiller and Hall, 1997). The green algae, from which green plants evolved, later acquired their plastids from a different cyanobacterial species, in which case they would be polyphyletic to the rhodophytes. Ultimately, all plastids are monophyletic -- assuming there was only one universal ancestor to all life. Depending on the species, there can be anywhere from 1 to 900 chloroplasts per plant cell (Frey, 1999). Several chloroplast genomes representing the major plant groups have been fully sequenced (Turmel et al., 1999), and there are several common features that should be noted. The chloroplast genome is circular and ranges in size from 30kbp to 200kbp (Martin and Hermann, 1998). In the angiosperms and most algae, the genomic structure (Fig-1) is arranged as two inverted repeats (IR) separating a large single copy (LSC) region from a small single copy (SSC) region (Curtis and Clegg, 1984; Turmel et al., 1999). This structure, as well as gene content, is highly conserved. Also, the substitution rates in cpDNA is much lower than in nuclear DNA (ncDNA) or in mitochondrial DNA (mtDNA) (Wolfe et al., 1987). And, the substitution rates in the inverted repeats is greatly reduced compared to the single copy regions (Wolfe et al., 1987). Throughout evolution, chloroplasts (and mitochondria) appear to have lost most of their ancestral genes. Table-1 shows a comparison of the genome size of a cyanobacterium (Synechocystis) with the chloroplast genomes of various land plants and algae. If chloroplasts are descendents from free living cyanobacteria, then there has been a major reduction in the genome sizes since their endosymbiotic origin. It is thought that many genes have been either transferred to the nucleus (i.e. by an 'endosymbiotic gene transfer') or lost completely (Martin and Hermann, 1998; Race et al., 1999). Transferred genes tend to adjust to the compartmental base composition and codon usage (Oliver et al., 1989). For example, cpDNA has a lower G-C content at the third codon position than in ncDNA, but when transferred to the plant nucleus, the G-C content of chloroplast genes increases. Codon usage of transferred chloroplast genes in the nucleus, also changes to match nuclear codon usage. How are genes transferred from organelles to the nucleus (or to other organelles)? For what reasons might these transfers occur? At the momment, an actual mechanism is not known, but a review of the literature might provide some clues. Gene Transfers and Intercompartmental Regulation. A few chloroplast genes have been extensively studied, such as the rbcL gene for the large subunit of Rubisco (reviewed by Clegg, 1993), and the psbA gene which encodes a thylakoid membrane protein involved in Photosystem II (PSII) (Frey, 1999; Clegg et al., 1994). In land plants and green algae, the rbcL locus is found on the LSC in the chloroplast genome, and the rbcS locus for the small subunit of Rubisco is found in the nuclear genome as a multicopy (Clegg et al., 1997). However, in cyanobacteria, the rbcL and rbcS genes are adjacent to one another and are cotranscribed (Nierzwicki-Bauer et al., 1984). Thus, the rbcS gene was probably transferred to the nucleus, and subsequently lost, from the chloroplast very early in the evolution of the plants. Another example of a gene transfer -- more recent than the rbcS gene transfer -- has been demonstrated in legumes (Gantt et al. 1991). The rpl22 gene, which encodes ribosomal protein L22, is present in the chloroplast genome of most angiosperms. Gantt et al. (1991) found that tobacco rpl22 probes hybridized in all angiosperm cpDNA tested, except in the legumes (Fig-2). Thus, the pea family (legumes) has lost rpl22 from the chloroplast genome. However, rpl22 was confirmed to be in the nucleus by a joint segregation analysis showing a 15% recombination distance with Acp-4, a gene on pea chromosome 6. In addition to chloroplast genes being transferred to the nucleus, some genes are thought to have been transferred to the mitochondria as well (Gray and Joyce, 1989; Menaud et al., 1998). For example, in Arabidopsis thaliana, a gene coding for methionyl-tRNA synthetase in the mitochondrial genome may have originated in the chloroplast (Menaud et al. 1998). And in wheat, three tRNA mitochondrial genes were found that show high sequence similarity to chloroplast genes (Gray and Joyce, 1989). Therefore the mitochondrial genome is a mosaic of genes with different origins (Gray and Joyce, 1989) -- as is the nuclear genome, and possibly the chloroplast genome. Gene products, rather than the genes themselves, with different origins may also be taken up by organelles. As a consequence, a particular organelle may be regulated by a different organelle or by the nucleus. In particular, several suppressor mechanisms have been found (Wu and Kuchka, 1995; Chen et al, 1997; Bennoun and Delosme, 1999). Defective chloroplast genes caused by mutation may be suppressed by the nucleus, allowing the defective chloroplast genes to become functional again. In one case, the mutation of a nuclear gene was found that reduced the synthesis of a component of PSII (D2, which is encoded in the chloroplast genome) in Chlamydomonas reinhardtii, and then another nuclear gene was found that suppressed the first mutation allowing for D2 synthesis (Wu and Kuchka, 1995). In another case, also in C. reinhardtii, a dominant mutation of a nuclear gene (SIM30) was discovered that suppressed a mutation for defective translational initiation in two chloroplast genes, petA and petD (Chen et al. 1997). petA codes for cytochrome b6/f and petD codes for cytochrome f. A mutation from AUG to AUU in both petA and petD was suppressed by SIM30, as was a mutation from AUG to AUC in petD. The authors believe SIM30 codes for a general chloroplast translation factor that interacts with the ribosome, since it is neither codon specific nor gene specific. The nucleus certainly exerts genetic control over the organelles, however chloroplasts have certain genes that may control their own division (Strepp et al. 1998; Wakasugi et al. 1997). These genes are are homologous to bacterial genes, providing further evidence for an endosymbiotic event between cyanobacteria and a eukaryotic host. Thus chloroplasts are "semi-autonomous" -- that is, they are partly responsible for their own metabolic processes and division. This is logical, since most genes are transferred to the nucleus from the organelles, rather than the other way around (Thorsness and Fox, 1990; Blanchard and Schmidt, 1995; Martin and Hermann, 1998). Can there also be organellar control over another organelle? The answer appears to be yes. A recent article by Bennoun and Delosme (1999) reveals evidence of chloroplast suppressors in Chlamydomonas that exerts control over mitochondrial and chloroplast mutations. Different plant species exhibit different inheritance patterns with respect to the segregation of chloroplast and mitochondrial organelles. The proposed mechanisms for the uniparental inheritance of organellar genes are diverse and no single mechanism is responsible for all types observed (Birky, 1995). They can be classed into one of four stages (Birky, 1995): i) prezygotic (e.g. unequal cell divisions, degradation of organelles or organellar DNA), ii) fertilization (e.g. exclusion of the organelles of one parent from the zygote), iii) zygotic deterministic/nonrandom (e.g. degradation of organelles or organellar DNA, exclusion of organelles from embryonic tissue, selective silencing), and iv) zygotic stochastic/random (e.g. exclusion of organelles from embryonic tissue). In Chlamydomonas, chloroplast and mitochondrial genes have uniparental-maternal (mt+) and uniparental-paternal (mt-) inheritance patterns respectively. This mating type pattern is caused by a nonrandom "selective silencing" mechanism in the zygote -- the enzymatic degradation of chloroplast DNA of the mt- gamete and the degradation of mitochondrial DNA of the mt+ gamete (Birky, 1995). The pattern, however, is not "strictly uniparental". Sometimes uniparental inheritance can break down -- there may be a number of different factors involved -- and biparental inheritance occurs, which may be segregated out after a few generations (Birky, 1995). Bennoun and Delosme (1999) used several strains of Chlamydomonas in the attempt to detect mutations in chloroplast genes that impaired mitchondrial functions. Through manganese mutagenic treatments (specific to mitochondria), the authors discovered a light-dependent (LD) mutant showing a slowed growth rate under heterotrophic conditions (i.e. in the dark and with acetate). LD was derived from the MUD2 strain, which has a mutation for myxothiazol resistance (used as a marker to control mitochondrial inheritance) in the mitochodrial cyt b gene. From the mutagenic treatment of the LD strain with fluodeoxyuridine (specific to chloroplasts), mutants exhibiting the wild type growth pattern (fast growth) were detected and named SUC1, for chloroplast suppressor 1. Bennoun and Delosme (1999) carried out several crosses with the LD and SUC1 mutants to determine their inheritance patterns and to determine whether any genetic interactions between the chloroplast and mitochondrial mutants. These crosses are summarized in Table-2. LD1 had a strictly uniparental-paternal inheritance pattern (Table-2, cross 1). SUC1 was found to have a non-strictly uniparental-maternal inheritance pattern (Table-2, cross 2), but SUC1 suppressed the mitochondrial LD1 (Table-2, crosses 3, 4, and 5). The authors also found a second chloroplast suppressor, SUC2, which suppressed both a mitochondrial mutation LD1 and a chloroplast point mutation in the rbcL gene (strain ac10-6C). Crosses carried out to determine the inheritance pattern and genetic interactions of SUC2 are summarized in Table-3. In several of the crosses from Tables 2 and 3, mixed patterns of inheritance were observed -- i.e. not strictly uniparental inheritance (Table 2, crosses 2, 4, and 5; Table-3 crosses 3, 4, 5, and 6). This is to be expected and there can be several factors involved. In Chlamydomonas, recombination of chloroplast DNA is frequent, and therefore may cause some biparental zygotes (Birky, 1995). Something to be noted from tables 2 and 3, is that some crosses were done several times while others only once. In any case, the samples sizes for all the crosses done by Bennoun and Delosme (1999) seems to be rather small (the most tetrads obtained in any cross was 45). Is this enough to be certain that the results aren't merely an artifact of small sampling? Perhaps it is enough, but just barely so. The frequencies of mixed patterns of uniparental and biparental inheritance can differ from one species to another, or even within the same species from one cross to another (Birky, 1995). Bennoun and Delosme indicate that their results are the first step towards a molecular characterization of these genetic interactions. The authors suspect that SUC2 is probably an informational suppressor, perhaps a suppressor tRNA. In C. reinhardtii, mitochondria genomes have only three tRNA genes, and therefore must import tRNAs from the cytosol. Many of these cytosolic tRNAs are encoded in the nucleus. Chloroplast genomes also have tRNA genes, which may be transported to the cytosol where they may be taken up by mitochondria -- this leads to the possibility of interorganellar genetic controls. The results of Bennoun and Delosme (1999) seem to point in that direction, but further research will be needed to confirm them. Mechanisms of Gene transfer. Why would genes be transferred to the nucleus in the first place? What selective advantage could there be? Muller's Ratchet has been suggested as a possible explanation (Martin et al., 1998). Asexual populations accumulate deleterious mutations and this may result in a decline in fitness. Also, the organelles with the fewest detrimental genes can be lost by drift, and this loss is irreversible in the absence of biparental inheritance and recombination (Birky, 1995). When a gene moves from the chloroplast to the nucleus, there is a change in context from an asexual to a sexual genome. Recombination can then take place to reduce genetic load (Race et al. 1999). However, in plant cells the mutation rate has been shown to be much lower in chloroplasts than in the nucleus (Wolfe et al. 1987), so Muller's Ratchet is not a sufficient explanation for gene transfer. Several suggestions have been made (Martin and Hermann 1999; Race et al, 1999), but there is no definitive answer yet as to why gene transfers occur. One example of divergence from the common cpDNA structure was recently found in the dinoflagellates (e.g. Heterocapsa triquetra) by Zhang et al. (1999). Dinoflagellate chloroplasts have 'mini-circles' each encoding a single gene ('one gene-one circle'), instead of the typical large multigenic circle. The mini-circles range from 2.0 kbp to 3.0 kbp and each has a conserved noncoding region, 9G-9A-9G, surrounded by 'less well conserved' regions (D1, D2, D3, D4) (Fig-3). So far, only nine genes were found in H. triquetra (seven protein coding genes and two rRNA genes): psaA and psaB for PSI; psbA, psbB, and psbC for PSII; atpA for ATPase; petB for cytochrome b6f; and genes for 23S rRNA and 16S rRNA. The rbcL gene appears to have been lost from dinoflagellate chloroplasts. Although only nine coding genes were detected, the authors acknowledge that other genes might be found later with a more extensive search. How the minicircles formed from a larger circle is not clear, but minicircles may provide insight into finding a mechanism for gene transfer. It could be that because they are compact, these single-gene minicircles can be easily packaged for transport (McFadden, 1999). This may explain why the dinoflagellate chloroplast genome is so highly reduced (McFadden, 1999): once the dinoflagellate chloroplast genome developed into single gene minicircles, these could be rapidly lost during evolution as compared to genes present within a larger circle. The question is, are there similar minicircles in plastids of other plant groups? If so, then this may be a key starting point into determining how endosymbiotic gene transfers occur. To date, minicircles have only been found in dinoflagellate chloroplasts and not in plastids of other plants, although single gene minicircles have been found in mitochondrial genomes of mesozoans (McFadden, 1999). A transferred gene faces several difficulties if it is to become functional in the nucleus (Gantt et al. 1991; Brennicke et al., 1993; Martin and Hermann, 1998). The theoretical steps involved are outlined in Fig-4. The gene must be released into the cytosol, taken up by the nucleus, and then integrate into the nuclear genome. Once integrated, the gene must also become transcriptionally activated (before mutations lead to permanent functional loss), and it must acquire a transit signal if the product is to return to the organelle from which it came. However, the gene need not be immediately lost from the organelle genome -- copies may coexist in both the nucleus and the organelle for some time period (Gantt et al., 1991). Also, acquiring a transit signal may be the least important (since some transferred genes need not necessarily be directed back to the organelle) and least difficult step (Martin and Hermann, 1998). Some genes may be lost altogether, and this would presumably be due to functional redundancy in the organelle and nuclear genomes (Martin and Herrmann, 1998). What mechanism is responsible for gene transfer? Although the possible steps involved have been elucidated (Fig-4), an actual mechanism has not. Gene transfers are ongoing throughout evolution (as indicated by ancient and recent transfers) and may occur by exporting either DNA or RNA from the organelle to the cytosol. Although RNA is a possible means of transfer (Blanchard and Schmidt, 1995), a gene may be directly transported as DNA without an RNA intermediate (Thorsness and Fox, 1990; Blanchard and Schmidt, 1995). It is likely that different genes were transferred by different mechanisms. For mitochondria, Thorsness and Fox (1990) determined that the rate of DNA transfer to the nucleus in Saccharomyces is about 2x10-5 per cell per generation. Whether a similar rate may also be found for chloroplast DNA transfer has yet to be determined. Temperature, certain chemicals (such as glycerol) and degradation of organelles may play a role in the escape of DNA from mitochondria, allowing release of mtDNA (Thorsness and Fox, 1990). Conclusion. Understanding chloroplast genomes and mechanisms of gene regulation will become increasingly important in the future. Several applications are already being put to use, such as making transgenic crops resistant to pesticides and herbicides (Stewart et al. 1996; Scott and Wilkinson, 1999). Creating transgenic crops may involve adding genes to the chloroplast genomes, rather than to the nucleus. Studying the unique chloroplast genome of Dinoflagellates (Zhang et al., 1999) may assist in developing drug therapy targeting the chloroplasts of the human parasites Plasmodium (which causes malaria) and Toxoplasma (McFadden, 1999). It has been demonstrated that chloroplasts are inherited maternally in certain crop species and that there is a very low probability of chloroplast gene flow into wild populations, such as in Brassica napus (Scott and Wilkinson, 1999). It is important for risk assessment to determine whether there is gene flow from transgenic crops into wild relatives (e.g. through pollen dispersal). The possibility of 'hybrid superweeds' resistant to pests and herbicides must be prevented, or at least minimized, from forming to avoid disruption of the local ecology. Thus, determining the inheritance patterns of organellar genes in crop species targeted for genetic modification is necessary in risk assessment. For example, if chloroplasts are maternally inherited in one particular species, a gene transfer to the nucleus (or to the mitchondria) could result in the paternal inheritance of a transgene. Since the rate may be high in some species (Thorsness and Fox, 1990), gene transfer will certainly need to be a part of risk assessment to prevent or avoid escape of transgenes to the environment. Another reason to look out for such events in transgenic crops is that a mechanism might be elucidated by catching a transfer event in the act, which could shed some light on the early evolution of endosymbionts with their hosts.
REFERENCES for Chloroplast Genetics
|
Get your own Free Home Page