Potential Retroviruses in Plants: Tat1 Is Related to a Group of Arabidopsis thaliana Ty3/gypsy Retrotransposons That Encode Envelope-Like Proteins

Corresponding author: Daniel F. Voytas, 2208 Molecular Biology Building, Iowa State University, Ames, IA 50011, voytas{at}iastate.edu (E-mail).

Communicating editor: V. SUNDARESAN

	ABSTRACT

TOP ABSTRACT MATERIALS AND METHODS RESULTS DISCUSSION LITERATURE CITED

Tat1 was originally identified as an insertion near the Arabidopsis thaliana SAM1 gene. We provide evidence that Tat1 is a retrotransposon and that previously described insertions are solo long terminal repeats (LTRs) left behind after the deletion of coding regions of full-length elements. Three Tat1 insertions were characterized that have retrotransposon features, including a primer binding site complementary to an A. thaliana asparagine tRNA and an open reading frame (ORF) with ~44% amino acid sequence similarity to the gag protein of the Zea mays retrotransposon Zeon-1. Tat1 elements have large, polymorphic 3' noncoding regions that may contain transduced DNA sequences; a 477-base insertion in the 3' noncoding region of the Tat1-3 element contains part of a related retrotransposon and sequences similar to the nontranslated leader sequence of AT-P5C1, a gene for pyrroline-5-carboxylate reductase. Analysis of DNA sequences generated by the A. thaliana genome project identified 10 families of Ty3/gypsy retrotransposons, which share up to 51 and 62% amino-acid similarity to the ORFs of Tat1 and the A. thaliana Athila element, respectively. Phylogenetic analyses resolved the plant Ty3/gypsy elements into two lineages, one of which includes homologs of Tat1 and Athila. Four families of A. thaliana elements within the Tat/Athila lineage encode a conserved ORF after integrase at a position occupied by the envelope gene in retroviruses and in some insect Ty3/gypsy retrotransposons. Like retroviral envelope genes, this ORF encodes a transmembrane domain and, in some insertions, a putative secretory signal sequence. This suggests that Tat/Athila retrotransposons may produce enveloped virions and may be infectious.

THE eukaryotic retrotransposons are divided into two distinct classes of elements on the basis of their structure: the long terminal repeat (LTR) retrotransposons and the LINE-like or non-LTR elements (DOOLITTLE et al. 1989 ; XIONG and EICKBUSH 1990 ). These element classes are related by the fact that each must undergo reverse transcription of an RNA intermediate to replicate, and each generally encodes its own reverse transcriptase. The LTR retrotransposons replicate by a mechanism that resembles that of the retroviruses (BOEKE and SANDMEYER 1991 ). They typically use a specific tRNA to prime reverse transcription, and a linear cDNA is synthesized through a series of template transfers that require redundant LTR sequences at each end of the element mRNA. This all occurs within a virus-like particle formed from proteins encoded by the retrotransposon mRNA. After reverse transcription, an integration complex directs the resulting cDNA to a new site in the genome of the host cell.

Phylogenetic analyses based on reverse transcriptase amino acid sequences resolve the LTR retrotransposons into two families: the Ty3/gypsy retrotransposons (Metaviridae) and the Ty1/copia elements (Pseudoviridae) (BOEKE et al. 1998A ; BOEKE et al. 1998B ; XIONG and EICKBUSH 1990 ). Although distinct, Ty3/gypsy elements are more closely related to the retroviruses than to the Ty1/copia elements. They also share a similar genetic organization with the retroviruses, principally in the order of integrase and reverse transcriptase in their pol genes. For the Ty3/gypsy elements, reverse transcriptase precedes integrase, and this order is reversed for the Ty1/copia elements. In addition, some Ty3/gypsy elements have an extra open reading frame (ORF) similar to retroviral envelope (env) proteins, which is required for viral infectivity. The Drosophila melanogaster gypsy retrotransposons encode an env-like ORF and can be transmitted between cells (KIM et al. 1994 ; SONG et al. 1994 ). Thus there are two distinct lineages of infectious LTR retroelements, the retroviruses, and those Ty3/gypsy retrotransposons that encode envelope-like proteins. The Ty3/gypsy elements have been divided into two genera, the metaviruses and the errantiviruses, the latter of which include all elements with env-like genes (BOEKE et al. 1998A ).

In plants, retrotransposons have been extremely successful (BENNETZEN 1996 ; VOYTAS 1996 ). The enormous size of many plant genomes demonstrates a great tolerance for repetitive DNA, a substantial proportion of which appears to be composed of retrotransposons. Because of their abundance, retrotransposons have undoubtedly influenced plant gene evolution. They can cause mutations in coding sequences (GRANDBASTIEN et al. 1989 ; HIROCHIKA et al. 1996 ; PURUGGANAN and WESSLER 1994 ), and the promoter regions of some plant genes contain relics of retrotransposon insertions that contribute transcriptional regulatory sequences (WHITE et al. 1994 ). Retrotransposons also generate gene duplications: Repetitive retrotransposon sequences provide substrates for unequal crossing over, and such an event is thought to have caused a zein gene duplication in maize (WHITE et al. 1994 ). Occasionally, cellular mRNAs are reverse transcribed and the resultant cDNA recombines into the genome giving rise to new genes, or more frequently, cDNA pseudogenes (MAESTRE et al. 1995 ). The transduction of gene sequences during reverse transcription, which produced the oncogenic retroviruses, has also been documented to occur for a plant retrotransposon (BUREAU et al. 1994 ; JIN and BENNETZEN 1994 ); a maize Bs1 insertion in Adh1 carries part of an ATPase gene and is the only known example of a retrotransposon-mediated gene transduction event.

Arabidopsis thaliana is unusual among plants in that its genome is small and its retrotransposon families are of low copy number (KONIECZNY et al. 1991 ; VOYTAS and AUSUBEL 1988 ; VOYTAS et al. 1990 ; WRIGHT et al. 1996 ). Of the 28 Ty1/copia and non-LTR retrotransposon families identified in our laboratory, none appear transpositionally active. Each family typically has three or fewer insertions in a given ecotype, and the structure of many insertions has been compromised by mutation or deletion. Furthermore, the chromosomal locations of elements and their copy numbers often do not differ between ecotypes, suggesting that they have not transposed recently. It seems that retrotransposon activity has been suppressed in A. thaliana or that most repetitive DNA has been lost (VOYTAS 1996 ).

The transposable elements Tat1 and Athila are the only known A. thaliana elements of moderate copy number. These families are represented in some ecotypes by about 10 and 30 copies, respectively (PELEMAN et al. 1991 ; PELISSIER et al. 1995 ). Tat1 and Athila were chance discoveries; each was found in a section of sequenced DNA. Athila is flanked by LTRs typical of retrotransposons; however, none of the insertions characterized encode gag or pol homologs (PELISSIER et al. 1995 ). Tat1 was initially discovered as a 431-base insertion in one of 11 genomic clones of the S-adenosylmethionine synthetase gene (SAM1) isolated from a -phage library (PELEMAN et al. 1991 ). Its presence in only one of the characterized clones suggested that it transposed into this site within the population of plants from which DNA was extracted for library construction. Because of its small size and lack of coding sequences, Tat1 was thought to be a degenerate DNA transposon.

We favored an alternative hypothesis to describe Tat1, namely that it is a retrotransposon solo LTR. Solo LTRs arise when the two LTRs of an integrated retrotransposon recombine, deleting the internal region and leaving behind a single LTR flanked by a target site duplication. Tat1 shares features with retrotransposon solo LTRs: It has LTR dinucleotide end-sequences (5'TG-CA3'), which are part of a 12-base inverted terminal repeat, and it created a 5-base target site duplication upon integration, typical of plant retrotransposons. In this study, we demonstrate that Tat1 is a retrotransposon and a member of a group of related retrovirus-like Ty3/gypsy elements present in the genomes of monocots and dicots. Some of these elements encode a conserved env-like gene, suggesting that infectious LTR retroelements exist in plants.

	MATERIALS AND METHODS

TOP ABSTRACT MATERIALS AND METHODS RESULTS DISCUSSION LITERATURE CITED

Plant material and Southern hybridizations:
The Arabidopsis Information Service supplied the following seed stocks (KRANZ and KIRCHHEIM 1987 ): Col-0, La-0, Kas-1, Co-4, Sei-0, Mv-0, Ll-0, Cvi-0, Fi-3, Ba-1, Hau-0, Aa-0, Ms-0, Ag-0, Ge-0, No-0 and Mh-0. Genomic DNA was extracted using genomic tips and protocols supplied by Qiagen (Valencia, CA). For Southern hybridizations, the resulting DNA was digested with EcoRI, electrophoresed on 0.8% agarose, and transferred to Gene Screen Plus membranes using the manufacturer's alkaline transfer protocol (New England Nuclear, Boston, MA). All hybridizations were performed as described (CHURCH and GILBERT 1984 ).

Library screening, probe preparation and PCR:
Tat1 clones were obtained by screening a Landsberg erecta (La-0) -phage library (VOYTAS et al. 1990 ), using a probe derived by PCR amplification of La-0 DNA. The primers for probe amplification were based on published Tat1 sequences: (DVO158, 5'-GGGATCCGCAATTAGAATCT-3'; DVO159, 5'-CGAATTCGGTCCACTTCGGA-3') (PELEMAN et al. 1991 ). Subsequent probes were restriction fragments of cloned Tat1 elements (Figure 1), and all probes were radiolabeled by random priming (Promega, Madison, WI). Long PCR was performed using the Expand Long Template PCR System (Boehringer Mannheim, Indianapolis) with LTR-specific primers (DVO354, 5'-CCACAAGATTCTAATTGCGGATTC-3'; DVO355, 5'-CCGAAATGGACCGAACCCGACATC-3'). The protocol used was for PCR amplification of DNA up to 15 kb. The following PCR primers were used to confirm the structure of Tat1-3: DVO405 (5'-TTTCCAGGCTCTTGACGAGATTTG-3') for the 3' non-coding region, DVO385 (5'-CGACTCGAGCTCCATAGCGATG-3') for the second ORF of Tat1-3 (note that the seventh base was changed from an A to a G to make an XhoI and a SalI restriction site) and DVO371 (5'-CGGATTGGGCCGAAATGGACCGAA-3') for the 3' LTR.

View larger version (27K): In this window In a new window Download PPT slide   Figure 1. —Genomic organization of Tat1 elements. Boxes with black triangles represent long terminal repeats (LTRs), and shaded portions of LTRs denote DNA sequences missing from particular clones. Open boxes indicate ORFs; the short lines are methionine codons and long lines are stop codons. Offset boxes represent a change in reading frame. Percentage nucleotide identified is indicated in the shaded regions between ORFs. The 3' noncoding regions are depicted as thin lines; they share 98% identity between Tat1-1 and Tat1-2, 96% identity between Tat1-2 and Tat1-3, and 89% identity between Tat1-3 and Tat1-4. The narrow boxes represent insertions unique to Tat1-3 and Tat1-4. Black bars below Tat1-1 indicate regions used for hybridization probes. Short, thin lines denote locations of primers used for PCR amplification of Tat1-3 and Tat1-4. Abbreviations are as follows: PBS, primer binding site; PPT, polypurine tract. The scale is in base pairs.

DNA sequencing:
Clones were sequenced either by the DNA sequencing facility at Iowa State University or with the fmol sequencing kit (Promega). DNA from the -phage clones was initially subcloned into the vector pBluescript II KS- and transformed into the E. coli host strain XL1 Blue (Stratagene, La Jolla, CA) (AUSUBEL et al. 1987 ). Subclones in the vector pMOB were used for transposon mutagenesis with the TN 1000 sequencing kit (Gold Biotechnologies, St. Louis, MO). Transposon-specific primers were used for DNA sequencing reactions.

Sequence analysis:
Sequence analysis was performed using the GCG software package (DEVEREUX et al. 1984 ), DNA Strider 1.2 (MARCK 1988 ), the BLAST search tool (ALTSCHUL et al. 1990 ), and the tRNAscan-SE 1.1 program (LOWE and EDDY 1997 ). Phylogenetic relationships were determined by the neighbor-joining distance algorithm using Phylip (FELSENSTEIN 1993 ; SAITOU and NEI 1987 ) and were based on reverse transcriptase amino-acid sequences that had been aligned with ClustalW1.7 (THOMPSON et al. 1994 ). Transmembrane helices were identified using the PHDhtm program (ROST et al. 1995 ). All DNA sequences have been submitted to the DDBS/EMBL/GenBank databases under the accession numbers AF056631, AF056632, AF056633, and AF056634.

	RESULTS

TOP ABSTRACT MATERIALS AND METHODS RESULTS DISCUSSION LITERATURE CITED

Tat1 is a retrotransposon:
Tat1 insertions share features with retrotransposon solo LTRs. We reasoned that if Tat1 is a retrotransposon, then there should be full-length elements in the genome consisting of two Tat1 sequences flanking an internal retrotransposon coding region. To test this hypothesis, additional Tat1 elements were isolated by screening a Landsberg (La-0) genomic DNA library with a Tat1 probe. Twenty-one -phage clones were isolated and Southern analysis revealed two clones (pDW42 and pDW99) each with two copies of Tat1 (data not shown). The two Tat1 elements in each clone were sequenced, along with the intervening DNA (Figure 1). All Tat1 sequences shared >89% nucleotide identity with the previously characterized Tat1a - Tat1c elements (PELEMAN et al. 1991 ). In clone pDW99, the 5' and 3' Tat1 sequences were 433 bases in length and only differed at two base positions. These Tat1 sequences also had conserved features of LTRs, including the dinucleotide end-sequences (5'-TG-CA-3') that were part of 12-base inverted terminal repeats. If the two Tat1 elements in clone pDW99 were retrotransposon LTRs, then both, along with the intervening DNA, should be flanked by a target site duplication. A putative 5-base target site duplication (TATGT) was present immediately adjacent to the 5' and 3' Tat1 elements, supporting the hypothesis that they and the intervening DNA inserted as a single unit. In clone pDW42, the 5' Tat1 was 432 bases in length and shared 98% nucleotide sequence identity to the 3' Tat1. The last ~74 bases of the 3' Tat1 was truncated during library construction and lies adjacent to one phage arm. A target site duplication, therefore, could not be identified in this clone.

DNA sequences were analyzed for potential coding information between the 5' and 3' Tat1 elements. Nearly identical ORFs of 424 and 405 amino acids were found encoded between the Tat1 sequences in pDW42 and pDW99, respectively (Figure 1). The derived amino-acid sequences of these ORFs were used to search the DNA sequence database with the BLAST search tool, and significant similarity was found to the Zea mays retrotransposable element Zeon-1 (p = 4.4e-08) (HU et al. 1995 ). The ORFs have ~44% similarity across their entirety to the 628-amino-acid ORF encoded by Zeon-1 (see below). The Zeon-1 ORF includes a zinc finger motif characteristic of retrotransposon gag protein RNA binding domains (HU et al. 1995 ). Although the Tat1 ORFs do not include the zinc finger motif, the degree of similarity suggests that they are part of a related gag protein.

If the Tat1 sequences in pDW42 and pDW99 defined retrotransposon insertions, a primer binding site (PBS) would be predicted to lie adjacent to the 5' Tat1 elements in both clones. The putative Tat1 PBS shares similarity with PBSs of Zeon-1 and another maize retrotransposon called Cinful (see below), but it is not complementary to an initiator methionine tRNA as is the case for most plant retrotransposons. Additionally, a possible polypurine tract (PPT), the primer for second-strand cDNA synthesis, was observed 1 base upstream of the 3' Tat1 sequence in both phage clones (5'-GAGGACTTGGGGGGCAAA-3'). We concluded from the available evidence that Tat1 is a retrotransposon, and we have designated the 3960-base insertion in pDW42 as Tat1-1 and the 3879-base insertion in pDW99 as Tat1-2 (Figure 1). It is apparent that both Tat1-1 and Tat1-2 are nonfunctional. Their ORFs are truncated with respect to the coding information found in transposition-competent retrotransposons, and they lack obvious pol motifs.

In light of our findings, the previously reported Tat1 sequences can be reinterpreted. Tat1a and Tat1b, which are flanked by putative target site duplications, are solo LTRs. Tat1c, the only element without a target site duplication, is actually the 5' LTR and part of the coding sequence for a larger Tat1 element (Figure 1).

Copy number of Tat1 among A. thaliana ecotypes:
To estimate Tat1 copy number, the 5' LTR, gag, and the 3' noncoding region were used as separate probes in Southern hybridizations (Figure 2). The Southern filters contained genomic DNA from 17 ecotypes representing wild populations of A. thaliana from around the world. This collection of ecotypes had previously been used to evaluate retrotransposon population dynamics (KONIECZNY et al. 1991 ; VOYTAS et al. 1990 ; WRIGHT et al. 1996 ). Based on the hybridization with the gag probe, element copy number ranges from two to approximately ten copies per ecotype (Figure 2). The copy number of the LTRs is higher, likely due to the presence of two LTRs flanking full-length elements or solo LTRs scattered throughout the genome. The Tat1 copy number contrasts with the copy numbers (typically less than three per ecotype) observed for 28 other A. thaliana retrotransposon families (KONIECZNY et al. 1991 ; VOYTAS et al. 1990 ; WRIGHT et al. 1996 ). In addition, the Tat1-hybridizing restriction fragments are highly polymorphic among strains. This degree of polymorphism, coupled with the high copy number, suggested that Tat1 has been active in transposition since the separation of the ecotypes.

View larger version (50K): In this window In a new window Download PPT slide   Figure 2. —Copy number of Tat1 in 17 ecotypes of A. thaliana. Ecotype DNAs were digested with EcoRI and the Southern filter was hybridized with radiolabeled probes for the Tat1-1 LTR (A), gag (B), and 3' noncoding region (C). Lanes in A are labeled with the corresponding ecotypes; the same filter was stripped and used for all three hybridizations.

The Tat1 3' noncoding region contains DNA sequences from elsewhere in the genome:
In an attempt to identify a complete and functional Tat1 element, LTR-specific primers were used in PCR reactions optimized for amplification of large DNA fragments. Most full-length retrotransposable elements are between 5 and 6 kb in length. DNAs from all 17 ecotypes were used as templates, and each gave amplification products of ~3.2 kb, the size predicted for Tat1-1 and Tat1-2 (data not shown). In La-0, however, a 3.8-kb PCR product was also recovered. This PCR product was cloned, sequenced and called Tat1-3. This insertion is expected to be about 4.6 kb in total length if the LTR sequences are included (Figure 1).

Tat1-3 differed from Tat1-1 and Tat1-2 in that it had two ORFs separated by stop codons and a 477-base insertion in the 3' noncoding region. The first ORF (365 amino acids) was similar to but shorter than the ORFs of the other Tat1 elements (Figure 1). The sequences constituting the second ORF (188 amino acids) were not present in the other Tat1 insertions and were not related to other sequences in the DNA databases. Database searches with the 477-base insertion in the 3' non-coding region, however, revealed three regions of similarity to other genomic sequences (Figure 3). A region of 113 bases matched a region of 26-bp repeats in the 5' untranslated sequence of the AT-P5C1 mRNA, which encodes pyrroline-5-carboxylate reductase (p = 2.1e-19) (Figure 3B) (VERBRUGGEN et al. 1993 ). In addition, 50 bases appear to be a remnant of another retrotransposon related to Tat1. These 50 bases are 71% identical to the 3' end of the Tat1-3 LTR and the putative primer binding site (Figure 3C). The putative primer binding site, however, is more closely related to those of other plant retrotransposons such as Huck-2 (SANMIGUEL et al. 1996 ) (Figure 3C). Finally, sequences in the remainder of the insertion showed significant similarity to a region on chromosome 5 (data not shown). To confirm that Tat1-3 was not a PCR artifact, two additional primer pairs were used in separate amplifications. Both amplifications gave PCR products of the predicted sizes, which were cloned and confirmed to be Tat1-3 by DNA sequencing.

View larger version (33K): In this window In a new window Download PPT slide   Figure 3. —Putative transduced sequences in the 3' noncoding region of Tat1-3. (A) The genomic organization of Tat1-3 as described in Figure 1. B and C designate sequences in the 477-base insertion that match sequences elsewhere in the A. thaliana genome and are shown in detail. The number 5 denotes sequences found on chromosome 5. (B) The lines over the alignment denote a 26-base repeat found in the 5' nontranslated region of AT-P5C1, a gene encoding pyrroline-5-carboxylate reductase (VERBRUGGEN et al. 1993 ). Numbers indicate base positions within Ta1-3 and AT-P5C1 (GenBank accession number M76538). (C) A sequence that matches the primer binding site of the Z. mays retrotransposon Huck1-2, and to a lesser degree, the PBS and 5' LTR end-sequences of Tat1-3. Underlined sequences represent the Tat1-3 LTR or LTR-similar sequences in the 3'-end of Tat1-3.

PCR amplifications with the additional primer pairs also yielded a product 0.8 kb longer than that expected for Tat1-3. This product was cloned, sequenced, and found to be another Tat1 element, designated Tat1-4 (Figure 1). This element has sequences similar to a Tat1 LTR, polypurine tract, and the second ORF of Tat1-3. In Tat1-4, 1182 bases of DNA are found in the 3' non-coding region at the position corresponding to the 477-base insertion in Tat1-3. This region does not match any sequences in the DNA databases.

Other Tat1-like elements in A. thaliana:
A BLAST search of DNA sequences generated by the A. thaliana genome project identified two more solo LTRs similar to Tat1 (Figure 4). All share similarities throughout, but most strikingly, they are very well conserved at the 5' and 3' ends where it is expected integrase would bind (BRAITERMAN and BOEKE 1994 ). These conserved end-sequences suggest that the integrases encoded by full-length elements are also related, and that the LTRs have evolved under functional constraints; that is, they are not simply degenerate Tat1 LTRs. The two new LTRs are designated as Tat2-1 and Tat3-1. Tat2-1 is 418 bases long, is flanked by a 5-base target site duplication (CTATT) and is ~63% identical to the Tat1-2 5' LTR. Tat3-1 is 463 bases long and is also flanked by a target site duplication (ATATT). Tat3-1 is ~53% identical to the Tat1-2 5' LTR.

View larger version (55K): In this window In a new window Download PPT slide   Figure 4. —Additional A. thaliana Tat retrotransposons. Alignment of the 5' LTR from Tat1-2 with two related solo LTRs found in the available sequence from the A. thaliana genome project. Tat2-1 is from the ESSA I contig fragment 4 (Accession Z97339, bases 115,028-115,445) and Tat3-1 is from BAC F11P17 (Accession AC002294, bases 81,040-81,502). The solo LTRs are flanked by the target site duplications CTATT and ATATT, respectively (not shown).

Tat1 and Athila are related to Ty3/gypsy retrotransposons:
Further analysis of data from the A. thaliana genome project revealed two slightly degenerate retrotransposons with similarity to the Tat1 ORF. These elements were identified within the sequence of the P1 phage clones MXA21 (Accession AB005247; bases 54,977–66,874) and MX110 (Accession AB005248; bases 24,125–35,848). Each has two LTRs, a putative PBS, and long ORFs between their LTRs. The genetic organization of these elements is depicted in Figure 5A and Figure 6A. Amino-acid sequence analysis indicated the presence of an RNA binding domain that defines gag in both elements (data not shown). This region is followed by conserved reverse transcriptase, RNase H, and integrase amino-acid sequence domains characteristic of pol (data not shown). Classification of eukaryotic retrotransposons into the Ty1/copia elements (Pseudoviridae) and Ty3/gypsy elements (Metaviridae) is based on pol gene structure (BOEKE et al. 1998A ; BOEKE et al. 1998B ). The domain order of the pol genes (reverse transcriptase precedes integrase) and similarities among their encoded reverse transcriptases (see below) identifies these elements as the first full-length A. thaliana Ty3/gypsy elements.

View larger version (65K): In this window In a new window Download PPT slide   Figure 5. —Genomic organization of plant Ty3/gypsy retrotransposons related to Tat1. (A) Element features are as described in Figure 1. Tat4-1 is from A. thaliana and Zeon-1 is from Z. mays. The numbers in the gray areas between elements reflect percentage amino-acid similarity to Tat1-2. The scale is in base pairs. Abbreviations not in Figure 1 are as follows: RB, RNA binding domain; PR, protease; RT, reverse transcriptase; IN, integrase. (B) Amino-acid sequence alignment of the gag genes of Tat4-1, Tat1-2 and Zeon-1. Black boxes identify identical amino-acid residues; gray boxes are similar residues (I=L=V, K=R, D=E).

View larger version (54K): In this window In a new window Download PPT slide   Figure 6. —Genomic organization of plant Ty3/gypsy retrotransposons related to Athila. (A) Element features are as described in Figure 1. The numbers in the gray areas between elements reflect percentage amino-acid similarity to Athila. Asterisks denote potential transmembrane domains. Athila1-2 is from BAC IG009D12, Athila2-1 is from BAC IG007N22 (bases 59,949-71,254) and Athila3-1 is from BAC TM025C13. The scale is in base pairs. (B) Amino-acid sequence alignments of the regions encoding putative transmembrane domains. The topmost alignment is for the 5'-most transmembrane domain in Athila and Athila2-1, which may serve as a secretory signal sequence. The lower alignment is for the transmembrane domain found in the second ORF of the four different retrotransposon families. Boxed amino-acid residues were predicted to be part of the transmembrane domain by the PHDhtm program (70–90% confidence) (ROST et al. 1995 ).

Because the characterized Tat1 insertions do not encode pol genes, this element family could not be classified. However, the amino-acid sequence of the Tat1-2 ORF is 51% similar to the gag region of the MXA21 retrotransposon (Figure 5B). Since plant retrotransposons within the Ty1/copia or Ty3/gypsy families, even those with highly similar pol genes, share little amino-acid sequence similarity in their gag regions, Tat1 is likely a Ty3/gypsy element. This conclusion is further supported by the report that the Tat-like Zeon-1 retrotransposon is very similar to a Z. mays Ty3/gypsy element called cinful (BENNETZEN 1996 ); however, only the 5' LTR and putative PBS sequences are available in the sequence database for analysis (Accession U68402). Because of the extent of similarity to Tat1, we have named the MXA21 insertion Tat4-1.

The gag region of the MX110 element is 62% similar (p = 1.1e-193) to the first ORF of Athila, which has previously been unclassified (PELISSIER et al. 1995 ) (data not shown). This implies that Athila is also a Ty3/gypsy element, and we have designated the MX110 insertion as Athila1-1 (Figure 6A). Our classification of Athila as a Ty3/gypsy element is further supported by the observation that the Athila gag amino-acid sequences share significant similarity to the gag protein encoded by the cyclops-2 Ty3/gypsy retrotransposon of pea (Accession AJ000640; p = 1.1e-46; data not shown). Further analysis of the available A. thaliana genome sequences identified three additional Athila homologs. They include an additional Athila1 element, designated Athila1-2, and two more distantly related Athila-like elements, designated Athila2-1 and Athila3-1 (Figure 6A).

In addition to similarities among their gag amino-acid sequences, the Tat elements have short LTRs (<550 bp) and long 3' noncoding regions (>2 kb) (Table 1, Figure 5A). In contrast, the Athila-like elements have long LTRs (>1.2 kb) and are very large retrotransposons (>11 kb) (Table 1, Figure 6A). One additional feature to note about both the Athila-like and Tat-like elements is the high degree of sequence degeneracy of their internal coding regions. This contrasts with the near sequence identity of their 5' and 3' LTRs, which is typically greater than 95% (Table 1). Because a single template is used in the synthesis of both LTRs, LTR sequences are usually identical at the time of integration. The degree of sequence similarity between the LTRs suggests that most elements integrated relatively recently. The polymorphisms observed in the internal domains of these insertions, therefore, may have been present in their progenitors, and these elements may have been replicated in trans.

A novel, conserved coding region in Athila elements:
A surprising feature of Athila1-1 is the presence of an additional ORF after integrase (Figure 6A). Like gag, this ORF shares significant similarity across its entirety (p = 3.8e-08) to the second ORF of Athila. This ORF is also encoded by the Athila2-1 and Athila3-1 elements, although it is somewhat more degenerate. The presence of this coding sequence among these divergent retrotransposons suggests that it plays a functional role in the element replication cycle. However, the ORF shows no similarity to retrotransposon gag or pol genes. The retroviruses and some Ty3/gypsy retrotransposons encode an env gene after integrase. Although not well conserved in primary sequence, both viral and retrotransposon envelope proteins share some structural similarities. They are typically translated from spliced mRNAs and the primary translation product encodes a signal peptide and a transmembrane domain near the C terminus. All four families of Athila elements encode a domain near the center of the ORF that is strongly predicted to be a transmembrane region (70–90% confidence, depending on the element analyzed) (ROST et al. 1995 ) (Figure 6B). Two retrotransposons, Athila and Athila2-1, also have a hydrophobic transmembrane domain near the 5' end of their env-like ORFs, which may serve as a secretory signal sequence (VON HEIJNE 1986 ).

Two lineages of plant Ty3/gypsy retrotransposons:
Relationships among Ty3/gypsy retrotransposons from A. thaliana and other organisms were assessed by constructing a neighbor-joining tree of their reverse transcriptase amino-acid sequences (Figure 7). Included in the analysis were reverse transcriptases from two additional families of A. thaliana Ty3/gypsy elements that we identified from the unannotated genome sequence data (designated Tma elements; Tma1-1 and Tma3-1); two other Tma element families were identified in the genome sequence that did not encode complete reverse transcriptases (Tma2-1 and Tma4-1; Table 1). Also included in the phylogenetic analyses were reverse transcriptases from a faba bean retrotransposon and the cyclops-2 element from pea. The plant Ty3/gypsy group retrotransposons resolved into two lineages: One was made up of del1 from lily, the IFG7 retrotransposon from pine, reina from Z. mays, and Tma1-1 and Tma3-1. This group of elements formed a single branch closely related to numerous fungal retrotransposons (branch 1). The second branch (branch 2) was well separated from all other known Ty3/gypsy group elements, and was further resolved into two lineages: Athila1-1, cyclops-2 and the faba bean reverse transcriptase formed one lineage (the Athila branch), and Tat4-1 and Grande1-4 from Zea diploperennis formed a separate, distinct branch (the Tat branch).

View larger version (25K): In this window In a new window Download PPT slide   Figure 7. —Phylogenetic relationships of Ty3/gypsy retrotransposons. Relationships of reverse transcriptase amino-acid sequences were determined by the neighbor-joining distance algorithm using Phylip (FELSENSTEIN 1993 ; SAITOU and NEI 1987 ). The tree was rooted using retrovirus sequences. Bootstrap values of greater than 50% of 100 replicates are shown as numbers above branch nodes. Nonplant elements containing env-like ORFs are labeled as errantiviruses. The remaining nonplant elements do not encode env-like ORFs and are metaviruses. All plant elements with env-like ORFs are in branch 2. Elements are as follows: del1, Lilium henryi (X13886); Tma3-1, A. thaliana (BAC IG009D12); Tma1-1, A. thaliana (BAC T32N15)(AC002534); IFG7, Pinus radiata (XIONG and EICKBUSH 1990 ); Reina, Zea mays (U69258); Tf1, Schizosaccharomyces pombe (M38526); CfT-1, Cladosporium fulvum (Z11866); Skippy, Fusarium oxysporum (L34658); Maggy, Magnaporthe grisea (L35053); Grasshopper, Magnaporthe grisea (M77661); Faba bean element, Vicia faba (AB007466); Cyclops-2, Pisum sativum (AJ000640); Athila1-1, A. thaliana (P1 clone MX110) (AB005248); Grande1-4, Zea diploperennis (X97604); Tat4-1, A. thaliana (P1 clone MXA21)(AB005247); Ty3, Saccharomyces cerevisiae (M23367); SURL, Tripneustis gratilla (M75723); Mag, Bombyx mori (X17219); Woot, Tribolium castaneum (U09586); Ulysses, Drosophila virilis (X56645); 412, D. melanogaster (X04132); mdg1, D. melanogaster (X59545); TED, Trichoplusia ni (M32662); Tom, Drosophila ananassae (Z24451); 297, D. melanogaster (X03431); 17.6, D. melanogaster (X01472); Yoyo, Ceratitis capitata (U60529); gypsy, D. melanogaster (M12927); Cer1, Caenorhabditis elegans (U15406).

Primer binding sites:
Most plant Ty1/copia retrotransposons as well as the branch 1 Ty3/gypsy elements have PBSs complementary to the 3' end of an initiator methionine tRNA. This is not the case for any of the branch 2 Ty3/gypsy elements. We compared the putative PBSs of Tat-branch and Athila-branch elements to known plant tRNA genes as well as to the 11 tRNA genes that had been identified to date in sequences generated by the A. thaliana genome project. In addition, we searched the unannotated A. thaliana genome sequences and identified 30 more A. thaliana tRNA genes using the program tRNAscan-SE (LOWE and EDDY 1997 ) (data not shown). The PBS of Tat1 is complementary to 10 bases at the 3' end of the asparagine tRNA for the AAC codon; these 10 bases are followed by a 2-base mismatch and 6 additional bases of perfect complementarity (Figure 8A). The Tat4-1 PBS is complementary to 20 bases at the 3' end of the arginine tRNA for the AGG codon with one mismatch 10 bases from the 3' end; Huck-2, Grande-zm1, Grande1-4, and the retrotransposon-like insertion in the 3' noncoding region of Tat1-3 all have 20-base perfect complementarity to this tRNA (Figure 8B). The PBS of Athila1-1 is perfectly complementary to 15 bases at the 3' end of the aspartic acid tRNA for the GAC codon, and Athila and Athila2-1 have 13 bases of complementarity to this tRNA (Figure 8C). At this time there is no known plant tRNA complementary to the PBS of Zeon-1, which has the same PBS as the maize retrotransposon cinful. As more tRNA sequences become available, a candidate primer may be identified for these elements.

View larger version (50K): In this window In a new window Download PPT slide   Figure 8. —Putative primer binding sites of plant Ty3/gypsy retrotransposons. In all panels, the shaded residues in the top figure are identical nucleotides shared among the various retrotransposon PBSs. (A) Complementarity between the 3' end of an A. thaliana Asn tRNA and the Tat1 PBS. (B) Similarities among the PBSs of Tat4-1, Huck-2, Grande-zm1, Grande 1-4 and the insertion in the noncoding region of Tat1-3. Complementarity is shown between an A. thaliana Arg tRNA and the PBS of Tat4-1. (C) Similarities between the PBS of Athila-like elements, and base-pairing between the 3' end of an A. thaliana Asp tRNA with the Athila1-1 PBS.

	DISCUSSION

TOP ABSTRACT MATERIALS AND METHODS RESULTS DISCUSSION LITERATURE CITED

Tat1 is related to plant Ty3/gypsy retrotransposons:
Tat1 was originally identified as an insertional polymorphism downstream of the A. thaliana SAM1 gene (PELEMAN et al. 1991 ). Of 11 genomic clones characterized from a -phage library, one contained a 431-base insertion that was designated Tat1a. The small size of this insertion suggested that it was a DNA transposon. Because Tat1a was present in only one of the SAM1 clones characterized, it was thought to have transposed to this site in one of the plants from which DNA was extracted for library construction. Tat1, therefore, was considered a likely candidate for an active A. thaliana transposon.

We considered a different interpretation of the Tat1 data, namely that the 431-base Tat1a insertion was a retrotransposon solo LTR. Solo LTRs are left behind as a consequence of recombination between LTRs of full-length elements. The characterization of additional Tat1 insertions supported our hypothesis. For example, we identified an insertion, designated Tat1-2, which has two 433-base LTRs (each >91% identical to Tat1a) and has a flanking 5-base target site duplication. Three bases after the Tat1-2 5' LTR is a putative PBS with 10 bases of complementary to the 3' end of an A. thaliana asparagine tRNA. One base upstream of the 3' LTR is a polypurine tract. Tat1 elements encode a short ORF that is highly similar (~44%) to the gag protein of the maize element Zeon-1. This ORF is even more similar (~51%) to the gag protein of the A. thaliana Ty3/gypsy element Tat4-1, which we identified from the DNA sequence of the A. thaliana P1 phage clone MXA21. Since the gag proteins of retrotransposons are generally not well conserved, this suggests that Tat1 is a Ty3/gypsy retrotransposon. By this reasoning, the previously characterized Athila element also appears to be a Ty3/gypsy retrotransposon. It shares ~62% amino-acid similarity between its first ORF and the gag protein of Athila1-1, an A. thaliana Ty3/gypsy element that we identified from the sequence of the P1 phage clone MX110. Although it is possible that the gag sequence similarity between these elements is the consequence of xenologous recombination, we do not believe this is the case, because Tat1 and Athila share a number of other features with related plant Ty3/gypsy elements (see below). With the exception of degenerate Ty3/gypsy reverse transcriptase sequences in the A. thaliana mitochondrial genome (KNOOP et al. 1996 ), the elements described in this report are the first A. thaliana Ty3/gypsy retrotransposons and are among a handful described to date in plants.

None of the three characterized Tat1 insertions encode reverse transcriptase or integrase motifs typically associated with functional retrotransposons. This may be the result of internal deletions, as suggested by size polymorphisms among the elements and their encoded ORFs. The related maize element Zeon-1 also does not encode proteins necessary for transposition (HU et al. 1995 ). Both Tat1 and Zeon-1, therefore, may be replicated by one or more master elements that provide functions in trans. This mechanism for transposition is further supported by the observation that all of the characterized A. thaliana Ty3/gypsy insertions have LTRs that share >95% nucleotide identity. This contrasts with often highly degenerate internal coding sequences. For the A. thaliana Ta1 elements, a family of Ty1/copia retrotransposons, LTR sequences of given insertions also share ~95% nucleotide identity, yet their coding regions are largely intact and carry only a few premature stop codons or frameshifts (VOYTAS et al. 1990 ). The extent of internal coding sequence degeneracy among the Ty3/gypsy elements relative to the near identity of their LTRs implies that transcripts from defective elements were acted upon in trans to generate these insertions. We were unable to identify candidate Tat1 master elements. Nonetheless, because Tat1 integrated recently near SAM1, a master element or a related retrotransposon that can act on Tat1 mRNA is likely present in the A. thaliana genome.

Southern hybridization analyses also suggest that Tat1 is transpositionally active. Up to 10 copies of Tat1 insertions are found in the genomes of the 17 diverse A. thaliana ecotypes analyzed. The extensive levels of observed restriction fragment length polymorphism among ecotypes is also consistent with transposition, although polymorphisms generated by recombination cannot be excluded. Like Tat1, the Athila elements are highly polymorphic and of moderate copy number (up to 30) (PELISSIER et al. 1995 ). The copy number and polymorphic nature of the Tat1 and Athila elements contrasts sharply with the 28 LINE-like and Ty1/copia elements previously characterized (KONIECZNY et al. 1991 ; VOYTAS et al. 1990 ; WRIGHT et al. 1996 ). These elements typically number no more than three insertions per ecotype), and they generally exhibit very uniform hybridization patterns; most differences can be explained by restriction site gain or loss. Ty3/gypsy elements, therefore, may be more transpositionally active than other classes of A. thaliana retrotransposons.

Tat1 elements may transduce genomic sequences:
Tat1, Zeon-1, and Tat4-1 have large 3' noncoding regions (from 2 to 4.5 kb). For the Tat1 insertions, this region is highly polymorphic and is characterized by numerous insertions/deletions. Imbedded within the 3' noncoding region of Tat1-3 is a 477-base insertion that contains four and a half iterations of a 26-base motif found in the leader sequence of AT-P5C1, an A. thaliana pyrroline-5-carboxylate reductase gene (VERBRUGGEN et al. 1993 ). In addition, this 477-base insertion contains sequences that resemble part of the LTR and PBS of a putative Ty3/gypsy element, as well as sequences highly similar to a region of chromosome 5. Tat1-4 contains 1.8 kb of DNA sequence of unknown origin in place of the 477 base Tat1-3 insertion. The large 3' noncoding regions of Zeon-1 and Tat4-1 suggest that they also may carry genomic sequences. Transduction may be one mechanism by which genomic DNA sequences are incorporated into the 3' noncoding regions of these elements. Transduction is well documented for the retroviruses, some of which are oncogenic as a consequence of having incorporated genes for cellular growth factors. Transduction events among retrotransposons, however, are rare. The only documented retrotransposon-mediated transduction event is for the maize retroelement Bs1; a Bs1 insertion within the Adh1 gene encodes part of a cellular ATPase gene (BUREAU et al. 1994 ; JIN and BENNETZEN 1994 ). As the A. thaliana genome is sequenced, it will be possible to explore more definitively the origin of the sequences in the 3' regions of these elements and the likelihood that they arose by transduction.

Plant Ty3/gypsy retrotransposons:
The ongoing A. thaliana genome project has increased our understanding of plant transposable element diversity. From the available genomic DNA sequences, we have identified five A. thaliana Ty3/gypsy elements (by their characteristic reverse transcriptase sequences and pol gene organization) and 10 partial A. thaliana Ty3/gypsy insertions closely related to these elements. Phylogenetic analyses based on Ty3/gypsy reverse transcriptase amino-acid sequences resolved the plant retrotransposons into two major lineages. One is composed of del1 from lily, reina from Z. mays, IFG7 from pine, and two A. thaliana Ty3/gypsy elements. These retrotransposons are all closely related to a group of fungal Ty3/gypsy retrotransposons. The second lineage includes Tat4-1, Athila1-1, and their homologs. An unusual feature of some elements in the Tat/Athila lineage is the presence of an additional, well-conserved ORF after the pol gene.

A nomenclature system proposed for the retrotransposons divides the Ty3/gypsy elements into two genera, the Metavirus and the Errantivirus (BOEKE et al. 1998A ). This classification is based principally on the presence of an additional ORF or env gene in the errantiviruses. Reverse transcriptase sequences also differ among the genera. Although little confidence can be placed in the relationships of the more basal branches of the Ty3/gypsy reverse transcriptase tree, the metaviruses and errantiviruses never cluster with each other, even for elements from the same species, such as those from D. melanogaster. Because plant Ty3/gypsy retrotransposons resolve into two distinct lineages, one of which contains elements with env-like ORFs, we propose that both genera of Metaviridae are present in plants. Specifically, we propose that del1 and related elements are metaviruses (branch 1, Figure 7) and that Tat4-1, Athila1-1 and their homologues are errantiviruses (branch 2, Figure 7). The plant errantiviruses further resolve into two major lineages, one containing Tat4-1 (which we refer to as the Tat branch) and the other containing Athila1-1 (which we refer to as the Athila branch). Not all of the insertions that we have classified as plant errantiviruses encode clear env-like ORFs, likely due to deletion events and sequence degeneracy. Hopefully, the ongoing genome sequencing efforts will reveal additional, more intact members of these element families to determine how well this classification system is supported.

Elements within the Tat and Athila branches share several other distinguishing features: Tat-branch elements have short LTRs (<550 bp) and long 3' non-coding regions (>2 kb). Elements in the Athila branch have long LTRs (>1.2 kb) and are generally very large retrotransposons (>11 kb). An additional, highly polymorphic feature of the plant errantiviruses are the sequences of their putative primer binding sites. Plant Ty1/copia elements and all characterized plant metaviruses have PBSs complementary to an initiator methionine tRNA. This is not the case for elements in the Tat or Athila lineages. We identified at least three possible primer tRNAs for these retrotransposons among tRNA genes that we identified in the emerging A. thaliana genome sequence. Potential primers include an aspartic acid tRNA (for the Athila branch elements) an arginine tRNA (for Tat4-1 and the Zea elements Huck-2, Grande-zm1 and Grande1-4) and an asparagine tRNA (for Tat1).

Plant retroviruses?
What is the function of the additional ORF encoded by the plant errantiviruses? Two lines of evidence suggest that it plays a role in the replication cycle of these elements: The ORF is found in multiple distinct element families, and within these elements it has evolved under functional constraints. For example, between Athila and Athila1-1, the env-like ORF shares ~34% similarity over >400 amino acids. Second, the ORF has a transmembrane domain, which is the most universal feature of retrovirus and animal errantivirus envelope proteins and suggests that it encodes components of a viral envelope. Athila and the closely related retrotransposon Athila2-1 also encode a transmembrane domain near the N terminus of the ORF at a position typically occupied by secretory signal sequences in envelope proteins. Envelope proteins of mammalian retroviruses and animal errantiviruses share other features in common; env is typically encoded by a subgenomic spliced transcript, and the protein is cleaved by a cellular endopeptidase to give rise to the glycosylated surface protein and transmembrane protein of the infectious virus. Putative glycosylation sites and endopeptidase cleavage domains can also be identified in the env-like genes of the plant errantiviruses (data not shown). However, until a replication-competent plant errantivirus is identified, their significance remains speculative. The possibility of retroviruses in plants has been previously suggested (BENNETZEN 1996 ). If they do exist, they have likely evolved unique mechanisms for transmission, including the ability to overcome the obstacle for infection presented by the plant cell well. At this point we cannot exclude the possibility that these elements have originated from nonplant hosts via horizontal transfer. Nonetheless, they appear to be widespread among plant genomes, as they are prevalent in the genomes of both monocots (maize) and dicots (A. thaliana, pea, faba bean).

Most LTR retrotransposons replicate strictly within the confines of their host cells. The finding that the gypsy retrotransposon of D. melanogaster has an infectious extracellular stage, however, has made it evident that infectious LTR retroelements are not limited to the vertebrate retroviruses (KIM et al. 1994 ; SONG et al. 1994 ). Our discovery of retrotransposons with a third ORF in the plant kingdom suggests that infectious LTR retroelements are pervasive. Their presence, coupled with evidence that some Tat1 elements and the maize Bs1 elements transduce genomic sequences, argues that barriers for interspecies gene flow may not be very rigid in plants. Animal retroviruses are notorious causal agents of disease, and if plant retroviruses exist, they may play a role in plant disease or disease processes. Infectious retrotransposons would offer many potential applications in plant biotechnology. For example, the unusually large sizes of these elements, and in particular, their large 3' noncoding regions, suggest that they can be modified to carry additional genes for use as vectors for plant gene transfer. It is our hope that the plant genome efforts will uncover replication-competent members of this unusual group of retrotransposons that can be used to test directly the biological significance of their envelope-like ORFs and the likelihood that these elements are transmitted extracellularly.

	ACKNOWLEDGMENTS

We thank members of the Voytas lab for helpful comments on the manuscript. This is Journal Paper No. J-17759 of the Iowa Agriculture and Home Economics Experiment Station, Ames, Iowa, Project No. 3120, and was supported by Hatch Act and State of Iowa funds.

Manuscript received December 29, 1997; Accepted for publication March 9, 1998.

	LITERATURE CITED

TOP ABSTRACT MATERIALS AND METHODS RESULTS DISCUSSION LITERATURE CITED

ALTSCHUL, S., W. GISH, W. MILLER, E. MYERS, and D. LIPMAN, 1990  Basic local alignment search tool. J. Mol. Biol. 215:403-410[Medline].

AUSUBEL, F. M., R. BRENT, R. E. KINGSTON, D. D. MOORE, J. G. SEIDMAN et al., 1987 Current Protocols in Molecular Biology. Greene/Wiley Interscience, New York.

BENNETZEN, J. L., 1996  The contributions of retroelements to plant genome organization, function and evolution. Trends Microbiol. 4:347-353[Medline].

BOEKE, J. D., T. EICKBUSH, S. B. SANDMEYER and D. F. VOYTAS, 1998a Metaviridae, in Virus Taxonomy: ICTV VIIth Report, edited by F. A. MURPHY. Springer-Verlag, New York.

BOEKE, J. D., T. EICKBUSH, S. B. SANDMEYER and D. F. VOYTAS, 1998b Pseudoviridae, in Virus Taxonomy: ICTV VIIth Report, edited by F. A. MURPHY. Springer-Verlag, New York.

BOEKE, J. D., and S. B. SANDMEYER, 1991 Yeast transposable elements, pp. 193–261 in The Molecular and Cellular Biology of the Yeast Saccharomyces, edited by J. BROACH, E. JONES and J. PRINGLE. Cold Spring Harbor Laboratory, Cold Spring Harbor, N.Y.

BRAITERMAN, L. T. and J. D. BOEKE, 1994  Ty1 in vitro integration: affects of mutations in cis and in trans.. Mol. Cell. Biol. 14:5731-5740[Abstract/Free Full Text].

BUREAU, T., S. WHITE, and S. WESSLER, 1994  Transduction of a cellular gene by a plant retroelement. Cell 77:479-480[Medline].

CHURCH, G. M. and W. GILBERT, 1984  Genomic sequencing. Proc. Natl. Acad. Sci. USA 81:1991-1995[Abstract/Free Full Text].

DEVEREUX, J., P. HAEBERLI, and O. SMITHIES, 1984  A comprehensive set of sequence analysis programs for the VAX. Nucleic Acids Res. 12:387-395.

DOOLITTLE, R. F., D.-F. FENG, M. S. JOHNSON, and M. A. MCCLURE, 1989  Origins and evolutionary relationships of retroviruses. Q. Rev. Biol. 64:1-30[Medline].

FELSENSTEIN, J., 1993 PHYLIP (Phylogeny Inference Package). Department of Genetics, University of Washington, Seattle.

GRANDBASTIEN, M. A., A. SPIELMANN, and M. CABOCHE, 1989  Tnt1, a mobile retroviral-like transposable element of tobacco isolated by plant cell genetics. Nature 337:376-380[Medline].

HIROCHIKA, H., K. SUGIMOTO, Y. OTSUKI, H. TSUGAWA, and M. KANDA, 1996  Retrotransposons of rice involved in mutations induced by tissue culture. Proc. Natl. Acad. Sci. USA 93:7783-7788[Abstract/Free Full Text].

HU, W., O. P. DAS, and J. MESSING, 1995  Zeon-1, a member of a new maize retrotransposon family. Mol. Gen. Genet. 248:471-480[Medline].

JIN, Y.-K. and J. L. BENNETZEN, 1994  Integration and nonrandom mutation of a plasma membrane proton ATPase gene fragment within the Bs1 retroelement of maize. Plant Cell 6:1177-1186[Abstract].

KIM, A., C. TERZIAN, P. SANTAMARIA, A. PELISSON, and N. PURD'HOMME et al., 1994  Retroviruses in invertebrates: the gypsy retrotransposon is apparently an infectious retrovirus of Drosophila melanogaster.. Proc. Natl. Acad. Sci. USA 91:1285-1289[Abstract/Free Full Text].

KNOOP, V., M. UNSELD, J. MARIENFELD, P. BRANDT, and S. SUNKEL et al., 1996  Copia-, gypsy- and LINE-like retrotransposon fragments in the mitochondrial genome of Arabidopsis thaliana.. Genetics 142:579-585[Abstract].

KONIECZNY, A., D. VOYTAS, M. CUMMINGS, and F. AUSUBEL, 1991  A superfamily of retrotransposable elements in Arabidopsis thaliana.. Genetics 127:801-809[Abstract].

KRANZ, A., and B. KIRCHHEIM, 1987 Genetic resources in Arabidopsis. Arabidopsis Inf. Serv. 24.

LOWE, T. M. and S. R. EDDY, 1997  RNAscan-SE: a program for improved detection of transfer RNA genes in genomic sequence. Nucleic Acids Res. 25:955-964[Abstract/Free Full Text].

MAESTRE, J., T. TCHENIO, O. DHELLIN, and T. HEIDMANN, 1995  RNA retroposition in human cells: processed pseudogene formation. EMBO J. 14:6333-6338[Medline].

MARCK, C., 1988  DNA Strider: a C program for the fast analysis of DNA and protein sequences on the Apple Macintosh family of computers. Nucleic Acids Res. 16:1829-1836[Abstract/Free Full Text].

PELEMAN, J., B. COTTYN, W. VAN CAMP, M. VAN MONTAGU, and D. INZE, 1991  Transient occurrence of extrachromsomal DNA of an Arabidopsis thaliana transposon-like element, Tat1. Proc. Natl. Acad. Sci. USA 88:3618-3622[Abstract/Free Full Text].

PELISSIER, T., S. TUTOIS, J. DERAGON, S. TOURMENTE, and S. GENESTIER et al., 1995  Athila, a new retroelement from Arabidopsis thaliana.. Plant Mol. Biol. 29:441-452[Medline].

PURUGGANAN, M. D. and S. R. WESSLER, 1994  Molecular evolution of magellan, a maize Ty3/gypsy-like retrotransposon. Proc. Natl. Acad. Sci. USA 91:11674-11678[Abstract/Free Full Text].

ROST, B., R. CASADIO, P. FARISELLI, and C. SANDER, 1995  Prediction of helical transmembrane segments at 95% accuracy. Protein Sci. 4:521-533[Abstract].

SAITOU, N. and M. NEI, 1987  The neighbor-joining method: a new method for reconstructing phylogenetic trees. Mol. Biol. Evol. 4:406-425[Abstract].

SANMIGUEL, P., A. TIKHONOV, Y. JIN, N. MOTCHOULSKAIA, and D. ZAKHARROV et al., 1996  Nested retrotransposons in the intergenic regions of the maize genome. Science 274:765-768[Abstract/Free Full Text].

SONG, S., T. GERASIMOVA, M. KURKULOS, J. BOEKE, and V. CORCES, 1994  An env-like protein encoded by a Drosophila retroelement: evidence that gypsy is an infectious retrovirus. Genes Dev. 8:2046-2057[Abstract/Free Full Text].

THOMPSON, J. D., D. G. HIGGINS, and T. J. GIBSON, 1994  CLUSTAL W: improving the sensitivity of progressive multiple sequence alignment through sequence weighting, positions-specific gap penalties and weight matrix choice. Nucleic Acids Res. 22:4673-4680[Abstract/Free Full Text].

VERBRUGGEN, N., R. VILLARROEL, and M. VAN MONTAGU, 1993  Osmoregulation of a pyrroline-5-carboxylate reductase gene in Arabidopsis thaliana.. Plant Physiol. 103:771-781[Abstract].

VON HEIJNE, G., 1986  A new method for predicting signal sequence cleavage sites. Nucleic Acids Res. 14:4683-4690[Abstract/Free Full Text].

VOYTAS, D. F., 1996  Retroelements in genome organization. Science 274:737-738[Free Full Text].

VOYTAS, D. F. and F. M. AUSUBEL, 1988  A copia-like transposable element family in Arabidopsis thaliana.. Nature 336:242-244[Medline].

VOYTAS, D. F., A. KONIECZNY, M. P. CUMMINGS, and F. M. AUSUBEL, 1990  The structure, distribution and evolution of the Ta1 retrotransposable element family of Arabidopsis thaliana.. Genetics 126:713-721[Abstract].

WHITE, S. E., L. F. HABERA, and S. R. WESSLER, 1994  Retrotransposons in the flanking regions of normal plant genes: a role for copia-like elements in the evolution of gene structure and expression. Proc. Natl. Acad. Sci. USA 91:11792-11796[Abstract/Free Full Text].

WRIGHT, D. A., N. KE, J. SMALLE, B. M. HAUGE, H. M. GOODMAN, and D. F. VOYTAS, 1996  Multiple non-LTR retrotransposons in the genome of Arabidopsis thaliana.. Genetics 142:569-578[Abstract].

XIONG, Y. and T. H. EICKBUSH, 1990  Origin and evolution of retroelements based upon their reverse transcriptase sequences. EMBO J. 9:3353-3362[Medline].

This article has been cited by other articles:

				C. Llorens, R. Futami, D. Bezemer, and A. Moya The Gypsy Database (GyDB) of mobile genetic elements Nucleic Acids Res., January 11, 2008; 36(suppl_1): D38 - D46. [Abstract] [Full Text] [PDF]

				D. Holligan, X. Zhang, N. Jiang, E. J. Pritham, and S. R. Wessler The Transposable Element Landscape of the Model Legume Lotus japonicus Genetics, December 1, 2006; 174(4): 2215 - 2228. [Abstract] [Full Text] [PDF]

I. S. Oh, A. R. Park, M. S. Bae, S. J. Kwon, Y. S. Kim, J. E. Lee, N. Y. Kang, S. Lee, H. Cheong, and O. K. Park Secretome Analysis Reveals an Arabidopsis Lipase Involved in Defense against Alternaria brassicicola PLANT CELL, October 1, 2005; 17(10): 2832