Horizontal gene transfer in plants - The story of PgiC2

Abstract: Sheep's fescue, Festuca ovina, is a common grass found all over Eurasia. In populations in southern Sweden some individuals were found to have two copies of the gene for cytoplasmic phosphoglucose isomerase, PgiC. The second copy, PgiC2, was first believed to be a duplicate of the original gene, PgiC1, but comparisons showed that this is not probable. The PgiC2 locus comprises two PgiC2 genes that may be fully functional genes or inactive pseudogenes. Phylogenetic analyses based on PgiC sequences from several grasses show that PgiC1 clusters with sequences from other fine-leaved festucoids and that PgiC2 clusters with a group of polyploid Poa species. Thus, PgiC2 has its most likely origin among the Poas, and it is particularly similar to a sequence from P. palustris. The estimated time of transfer is less than 600 000 years ago. We have discovered that the PgiC2 locus is not a local phenomenon, but that it can be found in several populations around the Baltic region. Most notable is a population from Poland where 50% of the plants are transgenic. Studies of the up- and downstream flanking regions of the PgiC2 gene have been conducted. Sequences from PgiC1 and PgiC2 can be aligned for approximately 660 bp upstream of the gene; as PgiC2 is fully functional, these 660 bp probably contain all elements necessary for gene regulation and function. PgiC2 and PgiC from P. palustris can align for another 110 bp. Beyond the breakpoint of PgiC2 and PgiC from P. palustris there is a 145 bp long fragment we call the transgene characteristic fragment (TCF). Primers constructed to amplify this fragment generate a product when used on transgenic F. ovina plants, but not when F. ovina plants without PgiC2 are probed. Thus, it seems as if the TCF is specifically associated with the PgiC2 locus. Interestingly, when the same primers are used on the polyploid Poa species closest to PgiC2 in the phylogenetic analyses, a PCR product is generated even though the fragment must be positioned beyond the breakpoint between PgiC2 and PgiC in P. palustris. This could indicate that TCF is part of a transposable element present in the polyploid Poa species and that its localization close to PgiC2 in F. ovina is not coincidental. This element may have contributed to the transfer of PgiC2 into F. ovina.