Class 2 DNA transposons were discovered in maize over 60 years ago with the genetic characterization of the Ac/Ds family of autonomous and nonautonomous elements by McClintock . Since then, DNA transposons have been found in all kingdoms of life and have been characterized into at least 10 superfamilies, based on the sequence of the element-encoded transposase protein . The newest superfamily is PIF/Harbinger, whose existence only came to light in the last decade. PIF/Harbinger derives its name from the two founding elements: Harbinger from Arabidopsis thaliana and PIF from Zea mays, discovered by computational and genetic analyses, respectively [3, 4].
Several features of transposition distinguish PIF/Harbinger from the other superfamilies. First, virtually all coding elements characterized to date contain two genes, ORF1 and TPase [5, 6]. Unlike CACTA elements where alternative splicing produces multiple proteins [7, 8], the two genes of PIF/Harbinger elements appear to be independent [5, 6]. Both the ORF1 and TPase proteins are required for transposition [9, 10]. Second, where analysed, excision is usually perfect as both the element and one copy of the 3 bp target site duplication (TSD) generated upon insertion is excised from the donor site [9, 10]. This differs from all previously characterized plant transposable elements where the majority of excision events leave a footprint or deletion at the excision site . Third, PIF/Harbinger elements display an extended target sequence preference: 9 bp in plants [4, 9, 12] and 15 bp in vertebrates [6, 10].
Another distinguishing feature of this superfamily is that PIF/Harbinger elements are responsible for the generation and amplification of Tourist-like miniature inverted repeat transposable elements (MITEs), one of the two predominant MITE families (the other being Stowaway). MITEs are small (100-500 bp) non-coding elements with the ability to amplify rapidly from one or a few near-identical elements to hundreds or thousands of copies . MITEs comprise ~5% of the rice genome  and are abundant in the genomes of some animals including mosquitoes , zebrafish  and humans [17, 18]. Where MITEs have been analysed on a genome-wide basis, they appear to play a significant role in gene evolution as they are abundant and insert preferentially into or near genes . In order to understand the success of Tourist-MITEs we need to first understand the transposition mechanism of PIF/Harbinger elements. With this goal in mind, we focus in this study on the rice mPing element, which is the only known active MITE.
Computational analysis of the sequenced genome of the rice (Oryza sativa, japonica) cultivar Nipponbare led to the identification of mPing . mPing was independently discovered to be actively transposing in the rice strain Gimbozu/EG4  and in rice anther culture . Further analysis revealed that this 431 bp Tourist-like MITE is a perfect deletion derivative of the 4.5 kb Ping element, which is present as a single copy in the Nipponbare genome and is a member of the PIF/Harbinger superfamily . Thus, it came as a surprise to find that mPing was actively transposing in an indica rice cell culture line that lacked the Ping element . The most likely source of TPase was determined to be the closely related Pong element, which is present in multiple copies in all tested strains of Oryza sativa. Subsequent studies with transgenic Arabidopsis confirmed that either Ping or Pong proteins were able to mobilize mPing and that transposition required functional copies of ORF1 and TPase .
Heterologous assay systems in plants and human cell culture have provided clues to the function of ORF1 protein and the reason why two proteins are required for transposition [9, 10]. For most class 2 elements, the TPase contains a conserved catalytic domain (DDE) and a DNA binding domain that recognizes and binds to the terminal inverted repeat (TIR) and/or subterminal regions . In contrast, the TPase of characterized members of the PIF/Harbinger superfamily lacks an obvious binding domain. Instead, a conserved Myb-like domain in ORF1 protein was hypothesized to be involved in DNA binding [5, 6]. This model was supported by studies with Harbinger3N_DR, an artificial element whose reconstruction was guided by building consensus sequences from the zebrafish genome . Harbinger3N_DR was mobilized in human cells only when both the reconstructed Harbinger3_DR TPase and ORF1 proteins were co-expressed . Furthermore, ORF1 protein was shown to bind the Harbinger3N_DR TIRs and interact with TPase. Finally, this study found that interaction with ORF1 protein was required for nuclear localization of TPase . These results suggest that ORF1 protein plays a critical role by positioning the TPase both in the nucleus and at the TIR where excision occurs.
We were motivated to develop a more facile assay system as a first step in the understanding of the amplification of MITEs and to dissect the complex transposition mechanism catalyzed by the PIF/Harbinger proteins. Here we report a yeast assay that recapitulates all of the features of excision and reinsertion of mPing as seen in plant systems. Furthermore, we demonstrate the validity of the yeast model by showing that a mutation of a putative nuclear export signal (NES) in the TPase increased transposition both in yeast and plants. These results provide a platform for further analysis of the regulation of other PIF/Harbinger elements and their relationships to Tourist-like MITEs.