The complete Daucus carota plastid genome is 155,911 base pairs (bp) in length (Fig. 1). The inverted repeat is 27,051 bp and the two copies are separated by two single copy regions; the large single copy region is 84,242 bp long and the small single copy region is 17,567 bp. There are a total of 136 predicted coding regions, 115 of which are unique and 21 are duplicated in the IR. On LSC/IRb boundary, the IR extends into rps19, resulting in the duplication of a portion of this gene. There are 81 unique protein-coding genes, 10 of which are duplicated in the IR. Also in the IR region is the ribosomal operon, which includes all 4 rRNA genes as well as tRNA-Ile and tRNA-Ala. There are five additional tRNAs within the IR resulting in a total of 37 tRNA genes, 30 of which are unique. There are 18 genes containing introns, with 15 of these with only a single intron (Table 1). Non-coding sequences, including IGS regions and introns, comprise 43.61 % of the carrot plastome. The overall nucleotide composition is 62.34 % AT and 37.66 %.

Repeat analysis identified 12 direct repeats and 2 palindromes of ≥ 30 bp with a sequence identity of ≥ 90 % (Hamming distance of 3). Repeated sequences were found in IGS regions, introns and within coding sequence (Table 2). There are 4 direct repeats in ycf2, with repeated sequences ranging up to 70 bp in length.

Our phylogenetic data set included 61 protein-coding genes for 31 taxa (Table 3), including 29 angiosperms and two gymnosperm outgroups (Pinus and Ginkgo). The data set comprised 45,582 nucleotide positions but when the gaps were excluded to avoid regions with ambiguous alignment due to length variation there were 39,490 characters.

Maximum Parsimony (MP) analyses resulted in a single, fully resolved tree with a length of 54,140, a consistency index of 0.44 (excluding uninformative characters) and a retention index of 0.60 (Fig. 2). Bootstrap analyses indicated that 26 of the 28 nodes were supported by values ≥ 95% and 19 of these had a bootstrap value of 100%. Maximum likelihood (ML) analysis resulted in a single tree with – lnL = 312205.340. ML bootstrap values also were also high, with values of = 95% for 24 of the 28 nodes and 22 nodes with 100% bootstrap support. The ML and MP trees had very similar topologies but differed in three places. (1) The MP tree placed Amborella as the sister group to all other angiosperms, whereas the ML tree placed Amborella sister to the Nymphaeales, and together this group formed the sister group of all other angiosperms. Support for the relationships of basal angiosperms in the MP tree is strong (100%) but only moderate in the ML tree (65%). (2) The MP tree placed Calycanthus sister to the eudicots, whereas the ML tree positioned Calycanthus as sister to a large clade that included both monocots and eudicots. Support for the different placements of Calycanthus was weak in both MP and ML analyses. (3) Relationships among the rosids differed, especially the position of Cucumis and the monophyly of eurosids I. The MP tree (Fig. 2) provides strong support for the monophyly of the eurosid I clade because Cucumis is sister to the three legume taxa. In contrast, the ML tree (Fig. 3) places Cucumis sister to the two examined taxa of Myrtales, and support for this relationship is not as strong (88% bootstrap value). These three differences were detected in recent phylogenies based on complete plastid genome sequences of basal angiosperms 49 and rosids 35. The remaining angiosperms formed two major clades, one including monocots and a second including the eudicots (Figs. 2, 3). Monophyly of the monocots was strongly supported (100% bootstrap value for both MP and ML). Ranunculales were strongly supported as sister to the remaining eudicots. There were two major clades of core eudicots, one including the rosids and the second including the Caryophyllales + asterids. Both MP and ML trees provide very strong support (100% bootstrap) for the sister relationship of Daucus with Panax in the euasterid II clade.

An important agricultural crop worldwide, carrot has long attracted attention from the research community. It was the first crop species in which somatic embryogenesis was demonstrated 59. This ability, to regenerate entire plants from cell or tissue cultures, has helped to maintain interest in carrot as our technology has advanced to include the improvement of agronomic species via genetic manipulation. The carrot nucleus has been the recipient of foreign genes to confer input characteristics such as pathogen resistance 60 and herbicide resistance 61. Recently an extensive analysis of four Agrobacterium rhizogenes strains and twelve Daucus carota genotypes examined the utility of green fluorescent protein (GFP) as a selectable marker for nuclear transformants 62 demonstrating continued interest in this system for expression of foreign proteins.

Most interesting has been the exploration of carrot as an ideal platform for the production of proteins of significance to pharmacology. The small isoform of human glutamic acid decarboxylase (GAD65) has been identified as a major autoantigen contributing to the onset of insulin-dependent diabetes mellitus (IDDM) 63. Expression of the GAD65 cDNA in transgenic carrot and tobacco resulted in an immunoreactive product which retained appropriate enzymatic function. Unfortunately levels of expression were quite low for both tobacco leaves and carrot taproots, on the order of 0.040 % and 0.012 % of total soluble protein (TSP), respectively 64. A heterologous version of human GAD65 having the N-terminus substituted with GAD67 from rat was expressed in tobacco and was able to achieve stable accumulation of functional immunoreactive product up to 0.19% of the TSP. Although oral administration of disease associated autoantigens such as GAD65 can lead to the induction of tolerance in the murine model, dosage on the order of milligrams per week per mouse are required 65. Expression levels and stable accumulation will have to be improved by orders of magnitude to make oral dosage truly feasible.

Vaccine antigens have been expressed in a number of plant species 66676869, including carrot. Transformation experiments have introduced a hemagglutinin glycoprotein 70 and a novel chimeric polyepitope antigen 71, both for neutralizing immunization against measles virus into the carrot nuclear genome. Extracted proteins demonstrated immunogenicity, raising antibody titers in sera of injected mice. However, quantitative data on protein accumulation in transgenic plants is lacking, especially for taproots. An estimate based on ELISA of 2% of membrane fraction in crude membrane preparations from carrot leaf is offered, but no mention is made of protein content in the root 70. It is noteworthy that these extracts were homogenized with Freund's adjuvant (1:1) prior to initial injection and at each boost. Freund's adjuvant is employed to enhance antibody formation suggesting that plant extracts from nuclear transformants were insufficient to induce the desired immune response.

The need for an alternative expression system to obtain high levels of protein accumulation in carrot roots becomes apparent if efficacious oral delivery is to be accomplished. Furthermore a system for oral delivery of antigens should ideally include the adjuvant, limiting further the need for post harvest processing. Transformation of the plant plastid has demonstrated the capacity to produce substantial quantities of functional and immunoreactive proteins [reviewed in 29].

Stable transformation of the Daucus carota plastid genome via somatic embryogenesis has been demonstrated recently 32. Integration of foreign genes in these experiments was accomplished through the use of flanking sequences that were PCR-amplified from the native carrot plastid ribosomal operon, whereas the regulatory sequences used to facilitate expression of the transgenes were derived from tobacco and bacteriophage T7. When assayed for BADH enzyme function, roots of carrot plastid transformants showed activity up to 74.8 % of leaf tissue. In root tissues, plastome copy number is generally about 5 % of the level in mature leaves. The notably high activity is probably due to the elevated concentration of root chromoplasts in carrot, the plastid type responsible for the orange coloration. With the availability of the entire carrot plastid genome sequence, it will now be possible to incorporate native translation regulatory sequences into transformation constructs to further enhance foreign protein accumulation in carrot plastids. Additionally, detailed knowledge of this genome will allow the identification of optimal intergenic spacer regions for the integration of transgenes.

Receptor-mediated translocation of antigens and other pharmaceutical proteins across the intestinal mucosa offers the potential to make plant-produced, orally delivered vaccines and therapeutics a reality. The toxin of Vibrio cholerae (CT) is recognized as one of the most potent mucosal adjuvants. The holotoxin is composed of the A subunit, responsible for toxicity, and the non-toxic homopentameric B subunit (CTB), which facilitates entry into epithelial cells of the intestine by binding the GM1 receptor followed by endocytosis. Recombinant forms excluding the A subunit are rendered non-toxic, and when fused to another antigen, the B subunit cannot only carry this antigen across the intestine, but also strongly potentiate the antigen's immunogenicity 7273. Recently a fusion construct of CTB and GFP expressed in transplastomic tobacco demonstrated the efficacy of CTB to deliver foreign proteins to the circulatory system of mice, which were fed pulverized leaf tissue from plastid transformants. Between the two protein sequences investigators included the cleavage site for the ubiquitous protease furin to facilitate the intracellular cleavage of GFP. Quantitative ELISA revealed accumulation of CTB-GFP in transgenic plants ranging from 19.09 to 21.3% of TSP. Following oral administration of CTB-GFP expressing leaf material to mice, fluorescence microscopy and immunohistochemical analyses confirmed the presence of GFP in the mouse intestinal mucosa, liver and spleen while CTB remained in the intestinal cells 74. Remarkable levels of protein accumulation coupled to a receptor-mediated oral delivery mechanism offers realistic hope for the possibility of plant-derived, orally delivered therapeutic proteins.

The Daucus genome with two copies of an IR separating the SSC and LSC regions is identical in architecture to most sequenced angiosperm plastid genomes 2. The size of the genome at 155,911 bp is also within the known range for angiosperms, which generally vary from 150,519 75 to 162,686 bp 47 for taxa that have both copies of the IR. The size of the Daucus IR at 27,051 bp is at the upper end of the size range of other sequenced genomes, which vary from 23,302 (Calycanthus) 76 to 27,807 (Oenothera) bp 77. Gene content and order of the Daucus plastid genome are identical to Panax 58, the only other published euasterid II genome.

A number of recent comparisons of plastid genomes of angiosperms have identified dispersed direct and inverted repeats 35363778. The carrot genome contains similar numbers and sizes of repeats to these other angiosperms (Table 2, Fig. 1). In most cases, these repeats are located in intergenic spacer regions and in introns but several also occur in tRNAs and protein-coding genes. Examination of repeats in highly rearranged algal and angiosperm genomes have demonstrated a correlation between both the number and location of the repeats and the propensity for rearrangements 7980. The role of dispersed repeats in unrearranged plastid genomes remains unknown.

The phylogenies based on 61 protein-coding plastid genes for 29 angiosperms (Figs. 2, 3) are largely congruent with relationships suggested by previous studies based single and multiple genes 56 and a number of recent phylogenies based on complete plastid genome sequences 35364849507681. There is strong support for the monophyly the major clades of angiosperms, including monocots, eudicots, rosids, asterids, eurosids II, asterids I and asterid II. The three areas of incongruence between the MP and ML trees regarding relationships of basal angiosperms, Calycanthus, and eurosids I were identified previously 3549, and are likely due to limited taxon sampling and misspecification of model parameters in large concatenated, multigene data sets.

The position of Caryophyllales within angiosperms has been controversial in the past. Previous molecular phylogenies clearly indicated that this order is a member of the eudicot clade 56, but relationships of Caryophyllales to other major eudicot lineages remains uncertain. The order has been suggested to be allied to rosids, asterids, or simply as an unresolved major eudicot clade sister to the Dilleniaceae 82. Two recent phylogenies based on 61 shared plastid gene sequences provided support for a sister relationship between the Caryophyllales and asterids 3536, however, only a single representative of the euasterid II clade was included. The addition of the Daucus plastid genome to this data set increases the level of support for a sister relationship of asterids and Caryophyllales (Figs. 2, 3).

Finally, the multiple plastid gene phylogenies also provide strong support for the monophyly of the euasterid II clade (Figs. 2, 3). This result is not surprising given that the Araliaceae (Panax) and Apiaceae (Daucus) have been considered sister families for a long time based on both morphological and molecular data 56. Expanded taxon sampling of the other three major clades of euasterids II is needed to further test the monophyly and relationships of this large, diverse angiosperm clade.

Daucus carota L. cv half long plants were purchased fresh from a local market with leaves intact. Leaf tissue (10 g) was collected for plastid isolation based on the sucrose step gradient centrifugation method of Palmer 83. Isolation was followed by whole plastid genome Rolling Circle Amplification (RCA) using the Repli-g RCA kit (Qiagen, Inc.) following the methods outlined in 84. After incubation at 30°C for 16 hr, the reaction was terminated with 10-minute incubation at 65°C. Digestion of the RCA product with BstXI, EcoRI and HindIII allowed verification of successful RCA amplification of the plastome, as well as assessment of its quality prior to genome sequencing.

DNA was sheared by nebulization, size fractionated to 4–6 kb, linker ligated and cloned into pHOS2, a TIGR medium copy vector. A total of 1231 high quality reads with an average length of 808 bases was generated during the random (1126 reads) and closure (105 reads) phases of sequencing. Sequences were assembled using TIGR assembler 85 and scaffolded using Bambus 86. Sequence finishing included directed PCR to span gaps and directed primer walking of clones to cover the entire genome and to complete regions of low depth of coverage.

The Daucus carota plastid genome was annotated using DOGMA 87, which performs BLASTX searches against a custom database of previously published plastid genomes to identify putative coding sequences. The user submits a FASTA-formatted input file of the complete plastid genome sequence for analysis. DOGMA identified putative start and stop codons, which must then be confirmed by the user for each putative protein-coding gene. Identification of intron and exon boundaries and tRNAs and rRNAs must also be confirmed. The fully annotated plastid genome of Daucus carota was submitted to NCBI GenBank with the following accession number [GenBank:DQ898156 ].

Analysis of repeat structure was carried out using Comparative Repeat Analysis (CRA) 88. The settings for identifying direct and inverted (palindromic) repeats included a size range between 30–5000 bp and a Hamming distance of 3 (limiting hits to sequence identity of ≥ 90%).

The 61 genes included in the analyses of Goremykin et al. 47, Leebens-Mack et al. 49, Lee et al. 36, and Jansen et al. 35 were extracted from the plastid genome sequence of Daucus using the DOGMA 87. The same set of 61 genes was extracted from plastid genome sequences of 30 other sequenced plastid genomes (see Table 3 for complete list of genomes examined). All 61 protein-coding genes of the 31 taxa were translated into amino acid sequences, aligned using MUSCLE 89 followed by manual adjustments, and then nucleotide sequences of these genes were aligned by constraining them to the aligned amino acid sequences. A Nexus file with character sets for phylogenetic analyses was generated after nucleotide sequence alignment was completed. The complete nucleotide alignment is available online at Chloroplast Genome Database 90.

Phylogenetic analyses using maximum parsimony (MP) and maximum likelihood (ML) were performed with PAUP* version 4.10b10 91. Phylogenetic analyses excluded gap regions to avoid alignment ambiguities in regions with variation in sequence lengths. All MP searches included 100 random addition replicates and TBR branch swapping with the Multrees option. Modeltest 3.7 92 was used to determine the most appropriate model of DNA sequence evolution for the combined 61-gene dataset. Hierarchical likelihood ratio tests and the Akaike information criterion were used to assess which of the 56 models best fit the data, which was determined to be GTR + I + Γ by both criteria. For ML analyses we performed an initial parsimony search with 100 random addition sequence replicates and TBR branch swapping, which resulted in a single tree. Model parameters were optimized onto the parsimony tree. We fixed these parameters and performed a ML analysis with three random addition sequence replicates and TBR branch swapping. The resulting ML tree was used to re-optimize model parameters, which then were fixed for another ML search with three random addition sequence replicates and TBR branch swapping. This successive approximation procedure was repeated until the same tree topology and model parameters were recovered in multiple, consecutive iterations. This tree was accepted as the final ML tree (Fig. 3). Successive approximation has been shown to perform as well as full-optimization analyses for a number of empirical and simulated datasets 93. Non-parametric bootstrap analyses 94 were performed for MP analyses with 1000 replicates with TBR branch swapping, 1 random addition replicate, and the Multrees option and for ML analyses with 100 replicates with NNI branch swapping, 1 random addition replicate, and the Multrees option.