Coe1 in Beta vulgaris L. Has a Tnp2-Domain DNA Transposase Gene within Putative LTRs and Other Retroelement-Like Features

We describe discovery in Beta vulgaris L. of Coe1, a DNA transposase gene within putative long terminal repeats (LTRs), and other retrotransposon-like features including both a retroviral-like hypothetical gene and an Rvt2-domain reverse transcriptase pseudogene. The central DNA transposase gene encodes, in eight exons, a predicted 160-KDa protein producing BLAST alignments with En/Spm-type transposons. Except for a stop signal, another ORF encodes a Ty1-copia-like reverse transcriptase with amino acid sequence domain YVDDIIL. Outside apparent LTRs, an 8-mer nucleotide sequence motif CACTATAA, near or within inverted repeat sequences, is hypothetical extreme termini. A genome scan of Arabidopsis thaliana found another example of a Tnp2-domain transposase gene within an apparent LTR-retrotransposon on chromosome 4.


INTRODUCTION
Since the discovery of transposable elements (TEs) in corn [1], DNA sequencing has revealed that genomes of eukaryotic organisms are largely comprised of evolutionarily significant TEs responsible for creation of considerable genetic diversity [2]. The movement of transposable elements is either autonomous or dependent on other elements. Classified according to mode of transposition, Class I TEs, or retrotransposons, are retroviral-type elements which may or may not have long terminal (direct) repeats (LTRs) [3]. Movement of Class I elements necessarily involves an RNA inter-mediate in what can appropriately be termed "replicative" transposition. Retrotransposons are transcribed into RNA, and then the reverse transcriptase and integrase make and insert a DNA copy at a secondary genomic location. Class II TEs are often called "DNA transposons," but it is important to note that Class I retrotransposons are also comprised of DNA except during transposition. Transposases permit Class II transposable elements to move by a "cut and paste" process, first excising from one site and then reintegrating at another. DNA replication is not required. "Footprints," telltale evidence for a previous DNA transposon insertion, result from imprecise excision. Class II transposons characteristically have relatively short inverted repeat sequences near their termini and an excision site at each end recognized by the transposase.
One of the first plant transposons McClintock described [1], En1, or the maize suppressor-mutator (Spm), is the original example of a "CACTA" class, or superfamily, of transposons. CACTA transposons were thought until recently to be found only in plants, but a similar element was discovered [4] in the genome of Schistosoma mansoni, the causative agent of schistosomiasis.
Evidence that retrotransposons account for much of the sugar beet (Beta vulgaris L.) genome was first obtained by Schmidt et al. [5], who described (1) repetitive DNA sequences in Beta vulgaris similar to long interspersed nuclear elements (LINEs), a type of retrotransposon without LTRs, and (2) other repetitive DNA sequences that resembled LTR retrotransposons of the Ty1-copia class. Vulmar1, a mariner-class transposon in Beta vulgaris, [6], 3 909 bp in length, has 32 bp terminal inverted repeats and carries a 2 International Journal of Plant Genomics single ORF that encodes a transposase with a characteristic DDE signature motif in a single exon.
Our interest in repetitive DNA developed from our recent discovery of a number of LTRs and retrotransposon genes as well as a transposase gene in the region between two clusters of core plant genes on a 130 Kb sugar beet BAC [7,8]. One gene cluster has an NPR1-class disease resistance-potentiating gene adjacent to another core plant gene whose predicted product has high similarity to a heat shock factor protein. The other cluster consists of a signal peptide calmodulin-binding protein kinase gene located near a CK1-class protein kinase gene. In this communication, we report the discovery in Beta vulgaris of Coe1, a Class II DNA transposase gene within putative LTRs and other features that are characteristic of Class I LTR-retrotransposons. Also, a genome scan of Arabidopsis thaliana found a similar arrangement of transposon and retro-element genes on chromosome 4.
A new multicopy direct tandem repeat (MDTR) within an intronic region within the Coe1 transposase gene was identified by BLAST of the intron against the entire 130 kb BAC and plotting a diagram of repeat versus DNA base positions with respect to the BAC. In order to find the starting points of the repeats, a window which was of a constant length less than one repeat was used in a sliding window technique. The first base used in this window was the putative starting point of the first repeat in the MDTR, as determined by the BLAST output. This window was BLASTed against the whole intron. If BLAST found any repeated DNA using this template, the window was actually still within the repeated segment; therefore, the starting base of the first repeat was deduced to be further towards the 5th end. The window was then moved a few bases in the 5th direction and the amended sequence was subjected to another BLAST. This process continued until BLAST no longer identified repeats.
GenBank accession EF101866 provides annotation of the 130 Kb NPR1-carrying BAC derived from thesugar beet genome. Conserved microsynteny of four core plant genes was observed with other eudicots (Kuykendall et al., submitted).
To scan the Arabidopsis thaliana genome for a Coe1like element(s), each chromosome was individually subjected to LTR STRUC analysis, and then each putative LTRretrotransposon element was examined for both a DNA transposase gene and a retrotransposon-like integrase or reverse transcriptase gene within its LTRs.

RESULTS AND DISCUSSION
BLAST and LTR STRUC analyses performed on an annotated 130 Kb NPR1 gene-carrying sugar beet BAC (GenBank accession EF 101866) revealed the presence of Coe1, which appears to be a new and unique composite of Class I and Class II transposable elements. Coe1 was chosen as its name to honor Dr. Gerald Coe who originally bred and developed a new U.S. hybrid sugar beet genotype, US H20 [10]. US H20 was the source of genomic DNA for a sugar beet Bacterial Artificial Chromosome (BAC) library 8 from which a BAC clone carrying the NPR1 disease resistance control gene was recently identified [7]. Initially detected by LTR STRUC analysis as a LTR-retrotransposon, Coe1, defined as 14.5 Kb by two putative 169 bp LTRs, has both an Rvt2-domain reverse transcriptase pseudogene and another retroviral-like hypothetical gene. However, a DNA transposase gene was found within its central region ( Figure 1). In addition to Coe1, LTR STRUC analysis performed on the 130-Kb NPR1 BAC identified at least two other LTR-retrotransposons, briefly: (1) a copia or Ty1-like retroelement, BvRTR1, which has a reverse transcriptase with active site YVDDIIF; and (2) BvRTR2, a gypsy or Ty3-like retroelement with active site FIDDILI in its conserved Rvt1 domain (unpublished). Precedence in the literature exists for similar yet considerably smaller repetitive DNA sequences from sugar beet largely uncharacterized except for genomic distribution.
The question then arises of whether the transposase gene of Coe1 represents a Class II transposon inserted into a Class I LTR-retrotransposon. This is probably how Coe1 originated. In any case, Coe1 has salient features of a Class II DNA transposon within a Class I retrotransposon ( Figure 1) as described below.
The Tnp2-domain transposase gene central to Coe1 consists of eight exons. The transposasegene of Coe1 has a predicted protein product that is evidently a CACTA superfamily DNA transposase as deduced from the results of BLAST amino acid sequence alignments of the predicted protein product with En/Spm-type DNA transposons ( Figure 2 (expressed sequence tags)(ESTs) [11] whose nucleotide sequence aligned by BLAST with Coe1's DNA transposase gene: CF542726 (e = 0.0) and BQ595658 (2e − 96). The Coe1 DNA transposase gene is flanked by inverted repeats and a CACTA sequence motif ( Figure 1). In addition to Coe1, the prototypical CACTA superfamily transposon En/Spm of corn, Tam1 of snapdragon [12] and seven other Tnp2-domain transposons, from various plant species, were compared using MegAlign. Figure 2 shows the amino acid sequence alignments obtained with one of two conserved domains. Cluster analysis of these data (Figure 3), a neighbor joining analysis tree, indicates that the Coe1 transposase falls into a group we designate as I subgroup A, with other plant Tnp2-domain transposases from Arabidopsis thaliana (BAB09502), Cleome spinosa (ABD969441), and Brassica rapa (BAA85462.1). Another subgroup of group I, B has En/Spm of Zea mays (AAA66266) and Oryza sativa japonica NP 001062816. In the amino acid sequence alignments of this particular conserved region (Figure 2), two other dissimilar groups (II and III) had the remaining four DNA transposases (Figure 3). The En/Spm-like superfamily of plant transposons, exemplified by Barbara McClintock's suppressor/mutator transposons of corn, has been named CACTA for the sequence motif recognized for excision. Conservation is well established over the taxonomic divide between eudicots, and monocots.
Coe1 has a centrally located transposase gene, flanked downstream by a pair of imperfect 51 bp I14 inverted repeat sequences (94% match) separated by only 10 bp, and upstream by another repeated sequence, I24, that aligns with I14 with about 75% identity over 41 bp. These distal I24/I14 inverted repeats are each flanked by a CACTA sequence motif.
Coe1 has a total of three ORFs: a retroviral-like hypothetical gene ORF1, the Tnp2-domain transposase gene and an apparent Rvt2-domain reverse transcriptase pseudogene, ORF2. Coe1 has putative long terminal repeats characteristic of LTR retrotransposons (Figure 1). These relatively short 169-bp direct repeats share only 96.4% identity. The 3.6% sequence divergence in the LTRs is consistent with possibility that the retroelement-like features of Coe1 are no longer active. The 5th end of the Coe1 positive strand has a pair of I13 inverted repeats 173 bp apart with 76% match over 190 bp and an internal CACTATAA sequence motif. I15 inverted repeats are found downstream of Coe1 and these inverted repeats, near another CACTATAA sequence motif, are 52 bp apart with about 74% match over 51 bp.
ORF1, the first retroelement-like gene of Coe1, encodes, in a single exon, a hypothetical protein for which no significant BLAST alignment is currently found. The predicted protein product of ORF1 had initially produced a significant BLAST with a "polynucleotidyl transferase" but that accession has been withdrawn.
ORF2 produced only a relatively weak nucleotide BLAST alignment (2e − 25) to EST BI643401. ORF2 is apparently a pseudogene since a stop signal occurs in the sequence prior to that part of the sequence that would otherwise encode the active site of an Rvt2 domain reverse transcriptase. Although a sequencing error is possible, it is unlikely; therefore, it is reasonable to deduce from the sequence data that ORF2 of Coe1 is a reverse transcriptase pseudogene. Disregarding the stop signal, the predicted protein product of the Rvt2-like gene of Coe1 aligned well with other Rvt2domain gene products (Figure 4). The Coe1 Rvt2 domain reverse transcriptase has a Ty1-copia-like YVDDIIL active site which is most highly conserved in comparison with that of the Medicago truncatula accession (Figure 4). Among the protein alignments performed, the hypothetical sugar beet Rvt2-domain containing gene product also showed higher similarity with retrotransposon-type reverse transcriptase proteins encoded by genes in two subspecies (indica and japonica) of Oryza sativa than with most others from a wide taxonomic range (Figure 4).
We recently performed a genome-wide scan or survey of the Arabidopsis thaliana genome looking specifically for a composite DNA transposon within LTR retrotransposon features similar to Coe1, and a similar single Tnp2domain transposase gene flanked by putative LTRs and other retrotransposon-like features was identified, as described below.      An Arabidopsis thaliana DNA transposase gene within putative LTRs and between LTR-retrotransposon genes is depicted in Figure 5. The apparent LTR-retrotransposon is 9 078 bp and has 506 bp and 471 bp LTR regions with about 90% identity. This element was found on BAC T26N6 from chromosome IV at 19.3 cM (accession AF07243). The first ORF (At4g04426) appears to be a highly degraded pseudogene of a reverse transcriptase, the central ORF (At4g04430) is a CACTA-class transposase gene, and the third ORF (At4g04440) has a predicted protein product with an Rvt2 reverse transcriptase domain similar to Ty1-copialike retrotransposons. Alignment of a conserved region of the predicted product of the reverse transcriptase and a neighbor joining tree from cluster analysis (not shown) revealed a number of similar BLAST hits but none of these was very nearly identical to the predicted product of the Arabidopsis element. The finding of a Coe1-like element in Arabidopsis supports the concept that Coe1 in Beta vulgaris is not a unique instance of a Class II transposon within a Class I retrotransposon.
Regarding the subject of Coe1 in Beta vulgaris, it is noteworthy that within the transposase gene most of the DNA sequence of an intron, when BLAST against itself yielded multipl, relatively small regions of imperfect DNA sequence identity. These small regions aligned, albeit imprecisely, with each other. These results, when graphed, showed a distinct pattern ( Figure 6) that suggested a whole series of multicopy direct tandem repeats (MDTRs).
Repeats of the apparent MDTR found by BLAST were of different lengths, but the first repeat in each set began at the same position, and the second repeat in each set ended at the same position. Each successive set of repeats was shorter by about the same number of bases. The intron was thus divisible into visually evident repeats ( Figure 6).
An intron within Coe1's transposase gene was thus determined to have 15 complete 173 bp MDTR and a partial repeat at the 3 end. The sequence of this MDTRcarrying intron did not yield any MegaBLAST hits against the NCBI database. One or more individual repeat(s) could have transposed into the intron followed by duplication(s), or possibly the entire series of multicopy tandem direct repeats transposed into the intron altogether. The latter possibility may be supported by the observation that the last repeat is partial, presumably a deletion.
How did Coe1, or an arrangement of transposable genes that resembles Coe1, originate? A DNA transposon could have moved into the middle of an LTR-retrotransposon. In other words, Coe1's central transposase gene, flanked by inverted repeats and CACTA sequence motifs, could have transposed into an LTR-retrotransposon.
The CACTATAA sequence motif, located 16.3 Kb apart, outside the putative LTRs near large inverted repeats, could perhaps be extreme boundaries of Coe1 instead of the two putative LTRs (Figure 1) separated by 14.5 Kb.
It is possible that the intact larger and more complex composite transposon could move about using DNA transposase. The LTRs are flanked by pairs of inverted repeat sequences that may be nonautonomous, miniature inverted [repeat] sequence transposable elements (MITEs) (Figure 1). MITEs, sometimes called "Class III transposons," are dependent on DNA transposase. The 8-mer sequence motif CACTATAA flanks the Coe1 LTRs near or within relatively large inverted repeats that are perhaps MITEs.
We hypothesize that originally a Class I LTR-retrotransposon inserted between pairs of MITEs. Then, a Class II DNA transposon moved into the middle of the LTRretrotransposon, and voila, a composite of Class II and Class I elements resembling Coe1, at least conceptually. To summarize, Coe1 has a Tnp2-domain transposase gene flanked by putative LTRs and between two retrotransposonlike genes, all within CACTATAAs near or within pairs of inverted repeats (Figure 1).
One may ask, "what possible selective advantage would a CACTA DNA transposon within an LTR retrotransposon have over a simple Class I or Class II element alone?" A combination of Class I and Class II features may offer little if any selective advantage, and thus such a composite might be unique. The finding of a similar gene arrangement in Arabidopsis provides a second example of a composite Class II transposon within a Class I retrotransposon.
A change in the expression of a gene when placed under the control of another can confer a selective advantage on  the host plant. Increased fitness might also be characteristic of the host of a versatile element which can hypothetically transpose in either of two known mechanisms. Such versatility could facilitate more rapid genetic change due both to transposition and to subsequent blockage of gene conversion.

CONCLUSION
In conclusion, based on results of in silico analyses, Coe1, found in the sugar beet genome, can be viewed as an incipient or emerging CACTA super-family DNA transposon amalgamated within an LTR-retrotransposon. A similar arrangement of a central Tnp2-domain transposase gene within LTRs and between retrotransposon genes was found in chromosome 4 of Arabidopsis thaliana by a genome scan. More DNA sequencing of the sugar beet genome, either of larger stretches or of the complete genome, is likely to be needed in order to distinguish whether Coe1 represents an evolutionarily significant gene arrangement or a mere coincidental merging of transposable genes.
In either case, as far we know, the two examples shown here of a Class II DNA transposon within a Class I retrotransposn are novel.