At present, plant molecular systematics and DNA barcoding techniques rely heavily on the use of chloroplast gene sequences. Because of the relatively low evolutionary rates of chloroplast genes, there are very few choices suitable for molecular studies on angiosperms at low taxonomic levels, and for DNA barcoding of species. We scanned the entire chloroplast genomes of 12 genera to search for highly variable regions. The sequence data of 9 genera were from GenBank and 3 genera were of our own. We identified nearly 5% of the most variable loci from all variable loci in the chloroplast genomes of each genus, and then selected 23 loci that were present in at least three genera. The 23 loci included 4 coding regions, 2 introns, and 17 intergenic spacers. Of the 23 loci, the most variable (in order from highest variability to lowest) were intergenic regions ycf1-a, trnK, rpl32-trnL, and trnH-psbA, followed by trnS^UGA-trnG^UCC, petA-psbJ, rps16-trnQ, ndhC-trnV, ycf1-b, ndhF, rpoB-trnC, psbE-petL, and rbcL-accD. Three loci, trnS^UGA-trnG^UCC, trnT-psbD, and trnW-psaJ, showed very high nucleotide diversity per site (π values) across three genera. Other loci may have strong potential for resolving phylogenetic and species identification problems at the species level. The loci accD-psaI, rbcL-accD, rpl32-trnL, rps16-trnQ, and ycf1 are absent from some genera. To amplify and sequence the highly variable loci identified in this study, we designed primers from their conserved flanking regions. We tested the applicability of the primers to amplify target sequences in eight species representing basal angiosperms, monocots, eudicots, rosids, and asterids, and confirmed that the primers amplified the desired sequences of these species. Chloroplast genome sequences contain regions that are highly variable. Such regions are the first consideration when screening the suitable loci to resolve closely related species or genera in phylogenetic analyses, and for DNA barcoding.

A DNA barcode is a DNA fragment used to identify species. For land plants, DNA fragments of plastid genome could be the primary consideration. Unfortunately, most of the plastid candidate barcodes lack species-level resolution. The identification of DNA barcodes of high resolution at species level is critical to the success of DNA barcoding in plants. We searched the available plastid genomes for the most variable regions and tested the best candidates using both a large number of tree species and seven well-sampled plant groups. Two regions of the plastid gene ycf1, ycf1a and ycf1b, were the most variable loci that were better than existing plastid candidate barcodes and can serve as a barcode of land plants. Primers were designed for the amplification of these regions and the PCR success of these primers ranged from 82.80% to 98.17%. Of 420 tree species, 357 species could be distinguished using ycf1b, which was slightly better than the combination of matK and rbcL. For the well-sampled representative plant groups, ycf1b generally performed better than any of the matK, rbcL and trnH-psbA. We concluded that ycf1a or ycf1b is the most variable plastid genome region and can serve as a core barcode of land plants.

Plastid genomes are an invaluable resource for plant biological studies. However, the number of completely sequenced plant plastid genomes is still small compared with the vast number of species. To provide an alternative generalized approach, we designed a set of 138 pairs of universal primers for amplifying (termed “short-range PCR”) and sequencing the entire genomes of the angiosperm plastid genomes. The universality of the primers was tested by using species from the basal to asterid angiosperms. The polymerase chain reaction (PCR) success rate was higher than 96%. We sequenced the complete chloroplast genome of Liquidambar formosana as an example using this method and compared it to the genomes independently determined by long-range PCR (from 6.3 kb to 13.3 kb) and next-generation sequencing methods. The three genomes showed that they were completely identical. To test the phylogenetic efficiency of this method, we amplified and sequenced 18 chloroplast regions of 19 Saxifragales and Saxifragales-related taxa, as a case study, to reconstruct the phylogeny of all families of the order. Phylograms based on a combination of our data, together with those from GenBank, clearly indicate three family groups and three single families within the order. This set of universal primers is expected to accelerate the accumulation of angiosperm plastid genomes and to make faster mass data collection of plastid genomes for molecular systematics.