Molecular phylogeny of 15 Cuscuta species
Cuscuta has previously been split into three subgenera: subgenus Cuscuta, subgenus Grammica, and subgenus Monogyna (Yuncker, 1932
), and evidence for the monophyletic nature of each of these three subgenera has been provided more recently (Stefanovic et al., 2002). This analysis supports the presence of these three subgenera, although it appears that Grammica can be divided into two clades, one containing C. australis, C. epithymum, C. indecora, and C. epilinum and the other containing C. chinesis, C. pentagona, C. cephalanthi, C. gronovii, and C. compacta. The Australis clade contains taxa that were previously classified as being within the Grammica and Cuscuta subgenera (Yuncker, 1932). Thus Grammica and Cuscuta (according to Yuncker) are paraphyletic.
The molecular phylogeny proposed groups C. palaestina, C. europaea, and C. planiflora in one clade (equating to subgenus Cuscuta) and indicates that C. lupuliformis, C. reflexa, and C. monogyna form a weakly supported clade, agreeing with their proposed position within the subgenus Monogyna (Yunker, 1932; Stefanovic et al., 2002). However, contrary to the Yuncker phylogeny which positions both C. australis and C. indecora in the Grammica subgenus (with the group containing C. compacta in Fig. 1) these data indicate that C. australis is sister to C. epithymum and forms a separate clade with C. epilinum and C. indecora. This analysis provides a phylogenetic framework upon which to investigate loss of photosynthesis and changes in structure of the plastid genome within these 15 species of Cuscuta.
Alterations to amino acids within the large subunit of Rubisco
In the process of generating the molecular phylogeny, part of the rbcL coding region was sequenced. This information was used to infer the amino acid sequence of the large subunit of Rubisco (LSU) for each species of Cuscuta. Chase et al. (1993) analysed the amino acid composition of the Rubisco LSU from 499 species of land plant. Our analysis indicated that in the 15 Cuscuta species, there were no changes to amino acids in the Rubsico active site. However, in other parts of the protein alterations were detected that have not previously been reported, and these have been termed unique amino acid changes. The greatest number of unique amino acid changes was 11 in C. indecora, whereas only one was found in both C. lupuliformis and C. reflexa. The sequence of the rbcL transcript has been analysed in C. reflexa, and an additional region of 53–69 bp in the 3'-UTR reported, which was proposed to reduce stability of the rbcL transcript and therefore influence abundance of the protein (Haberhausen et al., 1992). No evidence was found for species with more amino acid substitutions possessing lower rates of photosynthesis (data not shown).
Size and gene content of the Cuscuta plastid genomes
DNA blot analysis using tobacco plastid DNA probes allowed the size of plastid genomes in Convolvulus arvensis and five Cuscuta species to be estimated. Interestingly, in C. arvensis, a non-parasitic species closely related to Cuscuta (Stefanovic et al., 2002), the plastome was estimated to be 186 kbp in size, 30 kbp larger than that of tobacco (Shinozaki et al., 1986). However, 186 kbp is not an unusually large chloroplast genome for photosynthetic land plants (Palmer, 1985).
The size of the plastid genome of each Cuscuta species was smaller than those of C. arvensis and tobacco. These plastomes were estimated to range from 52% to 78% the size of that in tobacco. However, because C. arvensis is more closely related to Cuscuta and appears to have a larger plastid genome than tobacco, the Cuscuta species have probably lost a greater proportion of their plastid DNA than estimates based on comparison directly with tobacco. Hybridization between probes from each major region of the plastid DNA molecule (LSC, SSC, and IR) was detected, and this agrees with previous analysis of plastid DNA in autotrophic and parasitic plants (Palmer, 1985; Wolfe et al., 1992; Haberhausen et al., 1992). Berg et al. (2003) showed by immobilizing the entire tobacco plastome as BamHI fragments onto nylon membrane, and then challenging these with radiolabelled DNA from Cuscuta species, that C. reflexa (which contains chlorophyll) possesses a larger plastome than C. odorata (which does not possess chlorophyll). However, because the DNA from the dodders was not size-fractionated by electrophoresis, it was not possible to obtain quantitative estimates of plastome sizes (Berg et al., 2003).
Fragments from the inverted repeat have been cloned and sequenced from C. reflexa, and the junction between IRA and the LSC of plastid DNA is present (Bommer et al., 1993). In addition, in E. virginana, the parasitic plant with the smallest plastome, the inverted repeat still exists (dePamphlis and Palmer, 1990). It was therefore assumed that the plastid genomes of all five Cuscuta species examined possess inverted repeats, and we tentatively propose the structure of each plastid genome. This analysis indicates that these plastomes have lost fragments of similar sizes, for example, the LSC region is reduced to either 58 kbp or 44 kbp, while the small single copy region is reduced to either 14 kbp or 5 kbp. Interestingly, the estimate for the size of the SSC region in C. reflexa, and C. epilinum of 5 kbp is the same as that determined for E. virginiana (Wolfe et al., 1992). This may indicate that the SSC has been reduced to the smallest stable size in both Epifagus and some species of Cuscuta.
Although DNA blot analysis indicated that relatively large regions of the Cuscuta plastome have been lost, dot blots showed that homologous DNA to most genes present in the plastome of tobacco are present in Cuscuta. Previous work has reported a small number of plastid genes in C. reflexa, for example, a 6 kbp region of plastid DNA has been cloned, and contains petG, trnV, trnM, atpE, and atpB (Haberhausen et al., 1992). It has also been shown that transcripts for 15 chloroplast genes are found in C. reflexa (Berg et al., 2003). DNA homologous to photosynthesis genes in C. reflexa was detected, but transfer RNA genes were not present on the dot blot. Previously it has been reported that ndhB is a pseudogene in C. reflexa, and the other members of this family are not detectable (Haberhausen and Zetsche, 1994). However, in C. reflexa, an homologous DNA sequence to all the tobacco ndh genes was detected; this may be due to differences between the stringency of the hybridization conditions used. In C. subinclosa, there is evidence that rbcL has been lost completely (Van der Kooij et al., 2000), and in C. odorata, the only photosynthesis gene to remain in the plastome is psbA (Berg et al., 2003). No evidence was found for the absence of these genes in the five species analysed.
It is possible that the presence of homologous sequences to all the tobacco plastid genes in C. reflexa is due to transfer and integration of genes from the plastid into the nucleus. There is good evidence that genetic material from the plastid genome is able to transfer to the nucleus at significant rates (Martin et al., 1998, 2002; Millen et al., 2001; Huang et al., 2003). While RNA produced from nuclear genes is polyadenylated, in chloroplasts this only occurs to a small proportion of messages prior to degradation (Hayes et al., 1999). Therefore cDNA was produced from total RNA of C. reflexa and C. australis using poly-dT primers, and it was assumed that cDNA produced in this way would represent nuclear transcripts. Using this approach, transcripts with polyA tails were detected for the following genes that are normally resident in the plastid: rps14, rps7, orf131, atpA, atpF, psaB, and rpl36, as well as some ribosomal RNA genes. While it is known that the rrn16 gene is highly expressed in C. reflexa, and it is therefore possible that a proportion of its transcripts are targeted for degradation via the addition of a polyA tail (Hayes et al., 1999), photosynthesis genes are not highly expressed (Berg et al., 2003). It is therefore proposed that the most likely reason for a significant proportion of these other transcripts being polyadenylated is that the genes have transferred to the nucleus and are transcribed there. This possibility needs to be tested rigorously. These genes were not detected using the same procedure with the closely related but photosynthetic and non-parasitic species C. arvensis. It is proposed that in Cuscuta faster rates of DNA movement from plastid genome to nucleus is possible because there is less need for plastid-based redox signalling between the photosynthetic apparatus and the plastid genome. This is thought to be the reason for the retention of photosynthesis genes in the plastid genome of higher plants (Pfannschmidt et al., 1999).
Randomized loss of photosynthetic ability amongst the 15 species of Cuscuta
The loss of chlorophyll and alterations to photosynthetic ability in the different species of Cuscuta was mapped onto their phylogenetic position to test the hypothesis that photosynthesis has been lost to a greater extent in the most rapidly evolving clade. However, these data provided no support for this hypothesis, and it appears that selection pressure which maintains functional photosynthetic apparatus is reduced randomly across the 15 species analysed. In fact, there appears to be a relatively unco-ordinated loss of photosynthetic ability amongst these 15 species of Cuscuta, as Fv/Fm, and the maximum rate at which CO2 was incorporated did not correlate with the amount of chlorophyll amongst the different species. It is entirely possible, because a relatively small proportion of the genus (
10%) has been sampled, that studying the entire genus may reveal clear relationships between plastome size, organization, and phylogenetic position.
Overall, this study's analysis indicates that loss of photosynthesis and alterations in the structure of the plastid genome, are occurring in an unco-ordinated manner within Cuscuta, and that there may be an enhanced flux of DNA from plastid to nucleus.
Answers to all your Biology Questions
