النتائج (
العربية) 2:
[نسخ]نسخ!
The ITS and rbcL loci provide a baseline against which to compare other genes and intergenic spacers in our directed search for sequences to use in plant DNA barcoding. Besides ITS, those single-copy nuclear genes or their introns that are gaining prominence in species-level molecular systematics studies (e.g., leafy, waxy, pistillata, and RPB2), also were considered. However, because of the lack of universal primers (either published or with potential development by using current information) and poor success by using existing primers, these loci have been eliminated as potential barcode loci. The poor success by using existing primers is probably due to the difficulty of amplifying genes with low numbers of copies in degraded samples and the frequent need to clone PCR products before sequencing. We, therefore, turned our attention to the plastid genome in search of the most variable sequences that would also meet the criteria needed for maximum utility (i.e., variability, universal primers, and short length) and that could be used in place of or in addition to the ITS region. The significantly greater length of rbcL (usually 1,428 bp; Tables 1 and 2) causes problems because it is necessary to use four primers for double-stranded sequencing of the entire gene. Although this number of primers is equivalent to that needed if a two-loci system is used for barcoding purposes, the level of interspecific variation we observed in rbcL is less than the variation detected in either ITS or trnH-psbA alone (Table 2). Furthermore, this gene has been previously discounted for discrimination at the species level (e.g., refs. 31, 42, and 43).
We suggest that the trnH-psbA intergenic spacer is the best plastid option for a DNA barcode sequence that has good priming sites, length, and interspecific variation. In our trials across a diverse set of genera in seven plant families, three plastid regions (trnH-psbA, rp136-rpf8, and trnL-F) ranked highest with respect to amplification success and appropriate sequence length, but trnH-psbA demonstrated nearly 3 times the percentage sequence divergence of these other two regions (1.24% in trnH-psbA vs. 0.44% in both rp136-rpf8 and trnL-F; Table 2). The two spacers with the next highest mean sequence divergence after trnH-psbA (atpB-rbcL at 0.63% and trnC-ycf6 at 0.55%) could not be amplified in one or more of the test genera. In only one genus (Solidago; Asteraceae), exceptionally low sequence divergence in trnH-psbA prevented discrimination among the three species tested, although insertion/deletion differences still allowed us to distinguish among the species. This lack of sequence divergence between taxa was true for one or more species pairs in ITS and all other plastid spacers, except atpB-rbcL, in our test sample. In only 2% of our samples did homopolymer regions adversely affect sequence quality in trnH-psbA.
For a number of reasons, we refrained from a statistical test of differences among mean sequence divergences of the nine spacer regions. First, the sample size in our survey was too restricted to provide a meaningful statistical test (although the standard error of the mean of trnH-psbA does not directly overlap with the means of any of the other spacers). More importantly, as pointed out by Shaw et al. (33), genera within and between families of plants are phylogenetically nonequivalent, i.e., lineages recognized as genera may have quite different divergence rates depending on the various life history traits of the included species. Therefore, statistical comparisons between genera with respect to genetic distance are not valid or warranted at this time. Our intent in calculating these mean percent divergences across loci is to provide a qualitative evaluation of each spacer region for barcoding purposes. In this respect, we consider the high divergence value of trnH-psbA, which permits species discrimination in the largest number of taxa we tested (six of the eight genera and 11 of the 14 species pairs), as strong support for its use as a plant barcode.
The universality of trnH-psbA for differentiating among all flowering plant species clearly needs further investigation (see below), especially in taxa with extremely short spacers that may not contain enough sequence variation for species-level discrimination (e.g., Thalictrum and Solidago in our study and Minuartia in ref. 33). This spacer region also is present in other nonflowering land plants. In a search of GenBank, we found that the trnH-psbA spacer has been successfully amplified in angiosperms, gymnosperms, ferns, mosses, and liverworts, although we do not know at this time the degree of between-species divergence. Further study is needed to determine whether this plastid region is as variable in the nonflowering plants as we have shown for our test angiosperms, and therefore whether it is of broad utility as a barcode across the total spectrum of land plants.
Our findings on the properties of trnH-psbA agree with Shaw et al. (33) in their extensive survey of noncoding plastid DNA for phylogenetic purposes. By applying our barcode criteria (i.e., length considerations and universality) to the framework of their study, we conclude that trnH-psbA has greater potential for species-level discrimination than any other locus they analyzed. Similar to our results, they demonstrated that trnH-psbA amplified and sequenced easily with an average length of 465 bp across the 30 taxa they surveyed. Although this region was the second most variable of the 21 spacers they tested in terms of potentially informative characters, they ranked its utility for phylogenetic purposes as low (tier 3) because of its short length. Our analysis of the number of nucleotide substitutions within genera across all taxa in the 21 plastid regions presented by Shaw et al. (33) indicates that the trnH-psbA spacer has the highest percentage nucleotide difference (0.0135 difference per base pair), even though at least 8 of the 21 other regions showed a greater total number of nucleotide substitutions because of their longer length. The interspecific nucleotide differences in trnH-psbA ranged from 18% to 105% higher than that of the other eight most variable plastid regions. Because short sequence length is an important criterion for barcoding, the high frequency of nucleotide differences of trnH-psbA, in combination with its relatively short length, is a significant advantage. Other studies also have shown a high percentage of interspecific divergence for trnH-psbA, and in most cases, the highest in all plastid regions tested (e.g., refs. 44-48).
Despite this high level of interspecific variation, trnH-psbA has found only limited use in species-level phylogenetic reconstruction because of the short length as well as the difficulty of alignments resulting from a high number of indels (e.g., refs. 49-51). In contrast with the problems of indels for phylogenetic construction, we suspect that indels will ultimately enhance the information needed for species identifications, once the appropriate informatics tools for barcoding are developed. In the set of species we sampled, sequences were alignable within genera, but problematic above that rank. In the one case (Solidago) where sequence divergence was not sufficient to separate species, the presence of unique indels allowed easy discrimination among the taxa. Blaxter (34) advocates ease of alignment as a criterion when evaluating the utility of barcode loci. We do not consider difficulty of alignment to be a major obstacle to the applicability of either ITS or trnH-psbA for the primary purpose of DNA barcoding, i.e., identification. Although ease of alignment is desirable, it is not necessary for barcoding. Searches in GenBank by using our data from both loci with a BLAST search returned correct identities at both the gene and species level. BLAST searches are anchored and canalized by conserved regions in both loci, 5.8S in ITS and the small region of flanking exon for trnH-psbA. Intraspecific variation in both ITS and trnH-psbA is known to be relatively low, compared with interspecific variation (27, 52), although in the present study, our intraspecific sampling was insufficient to address this issue.
The extraction of DNA from specimens in herbarium collections was highly successful. This success may be due to the specimens having been air-dried and in a good state of preservation as evidenced by the generally green appearance of the leaves selected for extraction (Fig. 2). Plant voucher specimens vary in how and when they are dried after being pressed. If specimen-drying facilities are not immediately available, especially in humid tropical climates, botanists often treat pressed specimens with ethanol to temporarily preserve them against fungal attack and degradation. Alcohol has been shown to be detrimental to recovering high-quality DNA (53), although how it will affect the short sequences needed for barcoding is unknown. We are encouraged by the fact that museum specimens of insects dried from ethanol storage readily yield CO1 sequences. A more thorough investigation and optimization of methods to extract high-quality barcode DNA from herbarium collections in a high-throughput format will be critical to efficiently build a sequence-database library for plant DNA barcodes. Our positive results by using well preserved specimens indicate that the a priori selection of apparently undegraded plant samples will be an important determinant of success. Fortunately, herbaria often have more than one specimen per species among which to select for successful DNA barcoding.
We have shown here that there are gene sequences suitable for DNA barcoding of flowering plants. It may be necessary to employ more than one locus to attain species-level discrimination across all flowering plant species. Algorithms for combining barcoding sequences from two or more DNA regions to yield species-level unique iden
يجري ترجمتها، يرجى الانتظار ..