In 1991, Pierre Taberlet published what was probably the first article recommending 'universal' polymerase chain reaction (PCR) primers for use across plant genera and species, with a view to analysing intra-specific variation 1. The approach has been favourably adopted by the scientific community: a recent search identified 855 citing papers for the original publication (Scholar Google, 25 October 2006; up from 678 in January 2006). New sets of primers have subsequently been published that reflect Taberlet's original intention to study molecular variation among closely related species, or among separate sets of populations within species, by analysing introns and spacers [e.g., 2345]. This can be done since chloroplast genes evolve slowly, and primers can be designed with the purpose of working across species. In chloroplast genomes, gene order is highly conserved 235, whereas some spacers show even intra-species variation. Amplified fragments can be analysed by restriction analysis or DNA sequencing. The author's experience is that small insertions/deletions (indels) are relatively frequent, when compared to point mutations that result in restriction site changes 678. Exon sequences are generally highly conserved, but this depends on the gene in question. Molecular systematicists, starting with the highly conserved rbcL gene, and later expanding to e.g. matK, ndhF, rpl16, and atpB, have utilized PCR-amplified chloroplast gene sequences for establishing and verifying phylogenies. Sets of primers recommended for this purpose have also expanded in size [e.g., 9101112]. We have published a core set of 38 primer pairs useful in the amplification of the large single copy region in angiosperms, but also for fragments of this region in other plants 13.

As more and more partial and complete chloroplast DNA genome sequences become available, it is apparent that a balanced view on chloroplast sequence variation depends on the choice of many different sites along the genome 1011. It is interesting to notice that different groups of authors tend to work with alternative sets of primers. A central site for primer information should therefore help in making resources that are already there more widely known, and to encourage comparative studies across many laboratories.

Transferring PCR primers to new species has accelerated molecular research tremendously. However, it should be noted that from the early days of chloroplast genome research, when probes and blotting techniques were still in use, cross-species transfer of such gene probes was always possible [e.g., 23]. Nevertheless, venturing into unknown species with primers is still a 'trial and error' experience. It is hoped that the availability of a database that includes alternative, tested primers for a number of species will reduce the efforts in such cases. The database can now easily be searched or filtered: a text field is included which allows for free-text searching in any of the fields. For example, this text may be a part of a gene or primer name, an author name in the references (column 'Ref_Src'), or even a part of a primer nucleotide sequence. Additionally, the data can be sorted for the values in any of the fields. Numerical values (e.g., primer length or position in any of the 13 genomes) will be sorted arithmetically, and text values alphabetically. For instance, a search and filter operation for all primers associated with transfer RNA (trn) genes would require typing 'trn' into the filter; the results can be sorted e.g. by species (by entering the column number 11 for e.g. Acorus) or by gene names (column 3). Positioning the mouse over the column headings will show the full name of the columns.

Furthermore, it is possible to combine primers from different publications into newly assembled pairs. 'Generic' PCR conditions that favour successful amplification with such new combinations are standard PCR conditions with the use of 2 mM Mg²⁺ and a PCR program with 10 cycles at 70°C annealing, followed by 32 cycles at 55°C or 50°C annealing 7. With these primers and methods, the chloroplast genomes of hitherto unexplored species can be scanned in detail, and regions specific for different purposes picked [e.g. 78].

The author's recommendation for analysing uncharacterized chloroplast genomes is the following: (i) select a genome from the 13 listed which is phylogenetically close to the species of interest; (ii) sort the database for primer position in the selected genome (leaving the 'filter' field empty so that all primers will be displayed); (iii) select primer pairs in a suitable distance in the selected genome. In this last step, care must be taken to select primer pairs in the correct orientation. This can be done by comparing the gene order in the region of interest with the one in tobacco. In case of similar orientation of the genes, the 'F' and 'R' designations of the primers can be used as given in the database. In case of reverse orientation of the local gene order (relative to tobacco) in the species of interest, primer orientation is also reverse.

Recently, Dhingra and Folta 18 have suggested using overlapping PCR fragments for sequencing entire chloroplast genomes from total plant DNA preparations. While this approach holds some promise, it must be mentioned that a number of phenomena can interfere with it. Chloroplast DNA fragments are constantly transferred to the nucleus and to the mitochondrion 24, and extra-chloroplast DNA can lead to artefacts like apparent sequence polymorphisms. We have encountered such polymorphisms in shotgun sequencing of the black poplar (Populus trichocarpa) genomes [Heinze et al. in preparation; 27]. The quality of the DNA preparation (relative amounts of nuclear, mitochondrial, and chloroplast DNA) is key to the success of the PCR sequencing strategy. Prior purification of chloroplasts, which has always been a speed-limiting step, will determine the success to a large degree. The advantage of using primers from this database, as opposed to a fixed set of primer pairs, is that it is easy to switch from unsuccessful primers to alternatives, as very often, alternative primer positions close to each other, and degenerate sequences, have been proposed by different authors. It is also easier to generate overlapping sequence when primers are employed in varying combinations.

This version of the database offers the following improvements compared to earlier ones 26: the database can now be searched, filtered, and sorted online; there are now more than 700 primers (up from 500+); and primer positions are now given in 13 genomes (previously only five). In comparison to the single article introducing the highest number of primers 13, this database contains almost 10 times as many primers.

We have analysed collections of wild cherry (Prunus avium) and common ash (Fraxinus excelsior) DNAs for variation in chloroplast sequence, using either the PCR-RFLP 25 or a denaturing high-performance liquid chromatography approach [Heinze in preparation; 826]. In both cases, it was possible to quickly screen the major part of the large single copy region for variation between samples collected from different sites across the species ranges. In our laboratory, agarose PCR-RFLP is still a first screening method with acceptable throughput, when large sample numbers are analysed. After PCR, samples are scanned on agarose gels for successful amplification. An aliquot of the PCR is treated with restriction enzymes. Restriction polymorphisms and major indels can be detected in high-percentage agarose gels.

Some problems with universal PCR include the more rapid sequence evolution in some parts of the chloroplast; the identification of polymorphisms between conserved primer sites (introns), and occasional rearrangements, deletions, duplications in some genera, families, or higher taxonomic groups. However, rearrangements often only affect one spacer, leaving large blocks of genes with their order conserved. It is often at the breakpoints between conserved blocks of gene order where more sequence variation at lower taxonomic levels can be found.

It is tempting to speculate about poorly characterised chloroplast DNA regions that nevertheless yield conserved primer binding sites. This happens in some introns, but also in a number of spacers, open reading frames (ORFs) or hypothetical conserved reading frames (ycfs).

Conserved 'universal' primers and markers are possible for chloroplast DNA. Polymorphisms can be identified with tested primer sequences from the database. 'Generic' PCR conditions make possible the use of many primers in new combinations. Several opportunities exist for efficient detection and analysis of polymorphisms. It is hoped that this database will prove useful for many diverse problems and that our knowledge of mutation and evolution processes in chloroplasts will subsquently be enhanced, making it possible in the future to postulate informed predictions for poorly characterized species. Additions to the database (by e-mail to the author) are welcome.

The author(s) declare that they have no competing interests.