Study Of DNA Barcoding In Identifying Species Of Picea In Different Geographical Regions

Picea A. Dietrich is a genus belonged to Pinaceae and has 34 species, which are prominent component of the forests in the Northern Hemisphere with important economic and ecological value. They are mainly distributed in cool temperature zone, alpine and subalpine areas in temperate zone, of which 8 species distributed in Northern America, 2 in Europe, and 24 in Asia.Studies of the genus Picea had a history of more than 200 years, using lots of technique such as morphology, anatomy, biochemistry, karyotype, isozyme, RFLP, RAPD, and so on. However, owning to converging morphology and frequent hybridizations, relationship between species of Picea remains to be solved, especially between species distributed in Eurasia. In consequence, it is difficult to identify samples in the field and indoors. The emerging and developing of DNA barcoding provided a new method and a fresh perspective for the study of Picea. Searching for suitable barcodes in Picea could provide a basis for classification, introduction, cultivation and utilization of Picea. We compared nine candidate plant barcoding regions from the chloroplast for how well they discriminated the species of Picea. The markers comprised portions of seven loci recommended by the Consortium for the Barcode of Life (atpF-H, matK, trnH-psbA, psbK-I, rbcL, rpoB, rpoC1) and two alternatives (petA-psbJ, trnS-fM). The results may shed light on some questions when barcoding gymnosperms.The coding locus rpoC1 of some species had two peaks at a few bases in the sequence files with a low rate of variation,so rpoC1 was not evaluated in the following analysis. In the remaining six recommended loci, matK had the highest resolution rate of 35.7% with moderate length, followed by psbK-I (25%), which had the highest nucleotide diversity and the most haplotypes, but also had large intra-species variation. The noncoding locus trnH-psbA had a similar resolution with the psbK-I, and the next was the noncoding locus atpF-H (21.4%). rbcL and rpoB were most conservative and had lower resolution of 14.3% and 17.9%, respectively. While the additional two noncoding loci had good resolution, especially petA-psbJ, which had higher resolution than any other locus(57.1%). trnS-fM had the similar ability to discriminate species with atpF-H (21.4%).In all two-locus combinations of the recommended loci, trnH-psbA+psbK-I had the highest resolution(50%), followed by matK+trnH-psbA, matK+psbK-I and atpF-H+psbK-I(46.4%), while the matK+rbcL recommended by the Plant Working Group in 2009 just discriminated 42.9% of species. In all three-locus combinations of the recommended loci, atpF-H+psbK-I+trnH-psbA had the highest resolution(60.7%). The four-locus combination (atpF-H+matK+trnH-psbA+psbK-I) and the six-locus combination (atpF-H+matK+trnH-psbA+psbK-I+rbcL+rpoB) both identified 64.3% of species. While another two two-locus combinations petA-psbJ+trnS-fM and petA-psbJ+matK both had higher resolution (71.4%) than any other multilocus of the recommended loci, of which the ability to discriminate species was equal to that of the eight-locus combination in our study. Due to the polyT and polyG in the sequences of petA-psbJ and trnS-fM, respectively, it needed sequencing both ends of the loci, and it affected the feasibility of these two regions as DNA barcodes to some extent. But in view of the highest resolution of petA-psbJ, we suggested that matK and petA-psbJ should be the candidate DNA barcodes in Picea.The resolution rate of species distributed in North America was higher than that of Eurasia. Species distributed in North America were all resolved in the two-locus combination, but there were still 40% of species distributed in Eurasia remaining unresolved even when all eight loci combinated. It suggested that species distributed in North America may be good species and classified reasonably, or by reason that our samples of those species were limited. There may be some reasons to explain why eight species distributed in Eurasia were unresolved in all combinations.(i) ability of chloroplast genome to discriminate species may be limitedThe resolution had reached the highest rate when two loci combinated (petA-psbJ+matK and petA-psbJ+trnS-fM), and it was unchanged when any other loci added. It suggested that there may be an upper limit to use chloroplast genome to discriminate species of Picea, just as other plants.(ii) sampling strategySampling strategy has significant effect on resolution of DNA barcodes. Ideal sampling strategy for DNA barcoding is that five samples should be collected from different populations of each species, but in our study there were 12 species with just one individual. The percentage of species resolved should be changed when the sampling is perfected. (iii) frequent hybridizations in PiceaThere were frequent hybridizations between species of Picea, and closely related species can hybrid easily. For example, Li et al. (2010) speculated that P.purpurea may be hybrid of P.wilsonii and P.schrenkiana-P.likiangensis group, resulting in that P.purpurea shared cytoplasmic haplotypes with P.wilsonii, and these two species couldn't be differentiated by chloroplast DNA barcodes.(iv) recent adaptive radiationSpecies distributed in Himalaya–Hengduan Mountains (P.likiangensis,P.farreri,P.spinulosa,P.smithiana,P.schrenkiana) may be the products of recent adaptive radiation. Additionally, diverse geographical and ecological environment and special island habitats in Himalaya–Hengduan Mountains lead to that those species showed significant difference in phenotype and adaptability, but had no time to accumulate enough genetic variation.(v) taxonomic species are similarThe five species in P.asperata species group (P.asperata,P.crassifolia,P.obovata,P.meyeri,P.koraiensis) and P.abies distributed in Europe and P. koyamai in Japan were similar in phenotype, and they were clustered together in trees of chloroplast genes remaining unresolved. It is necessary to revise the taxonomy of those similar species when studying DNA barcoding in Picea further.