Discovery of novel plastid phenylalanine (trn F) pseudogenes defines a distinctive clade in Solanaceae

The plastid trn T-trn F region has been widely applied to resolve phylogeny of embryophytes (Quandt and Stech 2004; Zhao et al. 2011) and to address various questions of population genetics since the development of universal primers by Taberlet et al. (1991). This marker is located in the large single copy region of the chloroplast genome and contains a co-transcribed region consisting of three highly conserved exons that code the transfer RNA (tRNA) genes for threonine (UGU), leucine (UAA) and phenylalanine (GAA). The region is interspersed by two intergenic spacers and by a group I intron intercalated within the first and second exon of the trn L_(UAA) gene. Phylogenetic results obtained with the trn T-trn F region (or part of it) should be treated with caution. This is due to the fact that some recent studies (e.g. Koch et al. 2005; Pirie et al. 2007; Schmikl et al. 2009; Vivjerberg and Bachmann 1999) have shown that there are clearly several copies of certain parts of this region. If this is ignored, it will easily lead to situations where basic requirement of homology of the characters used for phylogenetic analyses is compromised. This might lead to false hypotheses of phylogeny, especially when they are based on the analyses of only this region.

Larger structural changes (>50 bp) rarely occur in the plastome. However, duplications of the rpl 2 or rpl 23 genes (Bowman et al. 1988) or even the duplication of tRNAs (pseudogenes) are occasionally reported. The later are extremely rare in angiosperms and so far they have only been described from Asteraceae (Vijverberg and Bachmann 1999; Witzell 1999), Annonaceae (Pirie et al. 2007), Brassicaceae (Ansell et al. 2007; Koch et al. 2007; Tedder et al. 2010) and Juncaceae (Drábkova et al. 2004). In our recent study we reported a tandem repeat comprising of two to four pseudogene copies upstream of the original trn F gene in four Solanum (Solanaceae) species (Poczai and Hyvönen 2011a). We have characterized these structural duplications and shown that they consist of several highly structured motifs, which are partial residues, or entire parts of the anticodon, T- and D-domains of the original gene, but all lack the acceptor stems at the 5′ or 3′. We were further interested to evaluate the possible occurrence of complete or partial trn F pseudogenes in Solanaceae. This family contains many economically important plant species, e.g., potato (Solanum tuberosum L.), tomato (Solanum lycopersicum L.) and paprika (Capsicum annuum L.) and is under intensive phylogenetic investigation and the trn T-F plastid marker is commonly used in these studies. These sequences together with the results of molecular breeding programs provide large amount of data that is available in GenBank. During data mining we concentrated on a structured dataset generated in previous phylogenetic studies (Fukuda et al. 2001; Garcia and Olmstead 2003; Santiago-Valentin and Olmstead 2003; Bohns 2004; Clarkson et al. 2004; Levin and Miller 2005; Levin et al. 2005; Weese and Bohns 2007; Olmstead et al. 2008) that contained 195 taxa and 390 sequences. This dataset provided the basis for the latest robust phylogenetic hypothesis of the Solanaceae including 89 from the 98 (Olmstead and Bohns 2007) recognized genera. Manual search using the anticodon domain of the original trn F gene and automated tRNA recognition by CENSOR (Kohany et al. 2006) indicated the presence of pseudogene repeats in numerous genera of Solanaceae.

We used the core trn L-F dataset to map the occurrence of pseudogenic repeats on the phylogenetic tree of Solanaceae. As presented in Figure 1 the distribution of pseudogenic duplications is in congruence with the previously published phylogeny of the Solanaceae (Olmstead et al. 2008), and it is obvious that the first pseudogenic copy evolved only once at the base of a highly supported clade within the subfamily Solanoideae. Among the members of this lineage, referred here as the Pseudosolanoid clade, the anticodon domain of the trn F gene exhibits extensive gene duplications with one to seven tandemly repeated copies in close 5′-proximity of the original functional gene (Table 1). The size of each pseudogenic copy ranged between 32 and 73 bp and the anticodon domain was identified as the most conserved element. A common ATT(G)_n motif is of particular interest and its modifications were found to border the 5′ of the duplicated regions in the same way as found in Brassicaceae (Ansell et al. 2007; Koch et al. 2005 and 2007; Schmikl et al. 2009; Tedder et al. 2010). Other motifs were partial residues or entire parts of the T- and D-domains. The residues of the 3′ and 5′ acceptor stems were rarely found among the copies (see Table 1). The D-domain was more conserved than the T-domain among the copies and other internal repeats (AT, AAT, ATT, AATCC) were intercalated within this region for example in genus Lycianthes (Dunal.) Hassl. In addition to these newly discovered pseudogenes we were also able to characterize putative promoter motifs showing high similarity to a sigma⁷⁰-type bacterial promoter. These two elements (−35 TTGACA/-10 GAGGAT) are consistently found in the trn L-F spacer region of embryophytes, and they are believed to represent the ancient and original trn F gene promoter (Quandt et al. 2004). Interestingly, pseudogenic repeats were found to be exclusively inserted after such motifs in Solanaceae, contrary to Brassicaceae, where similar pseudogenic repeats were found only between promoter motifs in the trn L-F intergenic spacer region (Koch et al. 2005). The later finding lead Koch et al. (2005) to support the conclusion by Kanno and Hirtai (1993) that these elements should be non-functional due to the intercalated position of pseudogenes between promoters. However, this may be challenged by the position of Solanaceae pseudogenes following the −10 and −35 promoters, which are also variable in number and composition.

The occurrence of pseudogenes provides strong evidence of relationships among some groups that had low support values in the previous analyses (e.g. Olsmtead et al. 2008). This event robustly separates the (1) Atropina (Hyoscyameae, Lycieae, Jabrosa, Latua, Nolana and Scleraphylax) and (2) Juanulloeae clades from the Pseudosolanoid clade composed by (3) Solaneae, Capsiceae, Physaleae and Datureae and (4) Salpichroina (Salpichroa Miers and Nectouxia Kunth). In clades (1) and (2) pseudogenes are absent while they appear at the basal node of clade (3) and (4). This lineage where pseudogene copies have been found includes 29 genera; here belongs also the clade of Solanum L. and Capsicum L. with many economically important plant species. However, sequence information was lacking for the genera Mellissia Hook. f. and Athenaea Adans. to confirm the presence of trn F pseudogenes. This is not surprising as available plant material of these taxa is very restricted. For example Mellissia is a genus with a single species, Mellissia begoniifolia (Roxb.) Hook. f. which is critically endangered and endemic to the island of Saint Helena. The larger clade of Solanoideae also includes several branches with low support values composed of small genera (Exodeconus Raf., Mandragora L., Nicandra (L.) Gaerten., Schultesianthus Hunz., Solandra Sw.) in the phylogeny proposed by Olmstead et al. (2008). These lineages are from the early diversification of the Solanoideae with no close relatives and all lack pseudogene repeats that could be informative to trace their ancestry.

The latest large scale phylogenetic analysis of the Solanaceae (Olmstead et al. 2008) established major clades of the family but sampling in some of the lineages can still be improved. Goldberg et al. (2010) analyzed a larger data set but they did not focus on taxonomic relationships but rather on the evolution of self-compatibility. Some studies have attempted to calibrate a molecular clock for various groups within Solanaceae, but all of these used the same (Paape et al. 2008; Poczai and Hyvönen 2011b), or only few fossil records (Dillon et al. 2009; Tu et al. 2010). Fossil record of the Solanaceae has not been reviewed recently. This urges for the re-assessment of the specimens and could potentially provide more robust calibration points for the family (Särkinen, personal communication). Latest current estimates show the age of the Pseudosolanoids to be approximately 20 My (Särkinen, personal communication), and thus the origin of the pseudogene duplications of Solanaceae to be approximately of the same Miocene age as in Brassicaceae (16–21 My; Koch et al. 2005).

Despite of the extensive studies based on sequence level characters the taxonomy of the Solanaceae is not yet completely understood. However, there is ongoing work on different levels by multiple groups to resolve phylogenetic relationships (Fukuda et al. 2001; Garcia and Olmstead 2003; Santiago-Valentin and Olmstead 2003; Bohs 2004; Clarkson et al. 2004; Levin and Miller 2005; Levin et al. 2005; Weese and Bohs 2007; Olmstead et al. 2008). There are a number of questions that should be answered regarding the discovery of trn F pseudogenes, for example: How did the duplications originate? Are the pseudogene copy numbers a useful character for phylogenetic inference? To what extent does the number of pseudogene copies vary within a single species? The evolution and structure of pseudogenic copies should be compared with others reported from different plant families especially from Brassicaceae. The potential of trn F pseudogenes as phylogenetic markers need to be investigated further in the future for better understanding of the evolution of Solanaceae. These investigations could answer what are the wider implications of the pseudogene repeats for Solanaceae studies that utilize the trn L-F spacer region.