Genome-Wide Analysis of mariner-Like Transposable Elements in Rice Reveals Complex Relationships With Stowaway Miniature Inverted Repeat Transposable
a Departments of Plant Biology and Genetics, The University of Georgia, Athens, Georgia 30602., http://www.100md.com
ABSTRACT., http://www.100md.com
Stowaway is a superfamily of miniature inverted repeat transposable elements (MITEs) that is widespread and abundant in plant genomes. Like other MITEs, however, its origin and mode of amplification are poorly understood. Several lines of evidence point to plant mariner-like elements (MLEs) as the autonomous partners of the nonautonomous Stowaway MITEs. To better understand this relationship, we have taken advantage of the nearly complete genome sequences of two rice subspecies to generate the first inventory of virtually all MLEs and Stowaway families coexisting in a single plant species. Thirty-four different MLEs were found to group into three major clades and 25 families. More than 22,000 Stowaway MITEs were identified and classified into 36 families. On the basis of detailed sequence comparisons, MLEs were confirmed to be the best candidate autonomous elements for Stowaway MITEs. Surprisingly, however, sequence similarity between MLE and Stowaway families was restricted to the terminal inverted repeats (TIRs) and, in a few cases, to adjacent subterminal sequences. These data suggest a model whereby most of the Stowaway MITEs in rice were cross-mobilized by MLE transposases encoded by distantly related elements.
Tc1/mariner is a diverse and widespread superfamily of eukaryotic class 2 transposable elements (reviewed in CAPY et al. 1998 ; PLASTERK et al. 1999 ; PLASTERK and VAN LUENEN 2002). One hallmark of the superfamily is insertion into the dinucleotide TA that is duplicated upon insertion and flanks the element as a target site duplication (TSD). Tc1/mariner elements are relatively short (1.2–3.5 kb) and are simple in structure with terminal inverted repeats (TIRs) and a single gene encoding the transposase. A common model for the transposition mechanism of Tc1/mariner elements has emerged from the functional study of a limited number of animal transposases (PLASTERK and VAN LUENEN 2002 ). The N-terminal region of Tc1/mariner transposases contains DNA-binding domain(s) that bind specifically to the TIRs (PLASTERK et al. 1999 ; LAMPE et al. 2001 ; ZHANG et al. 2001 ). A C-terminal domain is characterized by an amino acid signature called the DDE/D motif consisting of two aspartic acid residues and a glutamic acid residue (or a third D). This motif is required for catalysis of both the DNA cleavage and the strand transfer steps of the "cut and paste" transposition reaction (reviewed in HARTL et al. 1997 ; PLASTERK and VAN LUENEN 2002 ).
Tc1/mariner elements were recently found to be widespread in plants (reviewed in FESCHOTTE et al. 2002A ). The first reported plant members were Soymar1, a mariner-like element (MLE) from soybean (JARVIK and LARK 1998 ) and Lemi1, a pogo-like element from Arabidopsis thaliana (FESCHOTTE and MOUCHES 2000 ). Three additional rice MLEs were subsequently identified by database searches, but none were characterized further (TARCHINI et al. 2000 ; SHAO and TU 2001 ; TURCOTTE et al. 2001 ; FESCHOTTE and WESSLER 2002 ). These five elements were used to derive plant-specific primers that successfully amplified MLE transposase genes in PCR assays with DNA from a wide spectrum of flowering plant genomes (FESCHOTTE and WESSLER 2002 ). For the majority of genomes assayed, multiple divergent lineages of transposases were amplified from single species.%, 百拇医药
Demonstration that MLEs are widespread and diverse in plants provided support for the hypothesis that MLEs are the autonomous elements responsible for the origin and spread of Stowaway, a large group of miniature inverted repeat transposable elements (MITEs; BUREAU and WESSLER 1994 ). MITEs are structurally reminiscent of class 2 nonautonomous elements with their small size (<600 bp), lack of coding capacity, and TIRs (reviewed in FESCHOTTE et al. 2002B ). However, their high copy number and structural homogeneity have served to distinguish them from most of the previously described class 2 elements (WESSLER et al. 1995 ). MITEs were first discovered in plants, where they are now recognized as the predominant type of transposable element associated with the noncoding regions of plant genes. This is particularly evident in the cereals, including rice, maize, barley, and wheat (BENNETZEN 2000 ; FESCHOTTE et al. 2002A ; GOFF et al. 2002 ; YU et al. 2002 ). Vast amounts of MITEs have also been discovered in many invertebrate and vertebrate genomes (reviewed in FESCHOTTE et al. 2002B ).
Most of the tens of thousands of MITEs in plant genomes have been divided into two groups on the basis of the similarity of their TIRs and TSDs: Tourist-like MITEs and Stowaway-like MITEs (WESSLER et al. 1995 ; FESCHOTTE et al. 2002B ). That Stowaway-like MITEs and plant MLEs share similar terminal sequences (5'-CTCCCTCCRT-3', where R stands for A or G) and target site preference (TA) strongly suggested that Stowaway MITEs were mobilized in trans by transposases encoded by MLEs (TURCOTTE et al. 2001 ; FESCHOTTE et al. 2002B ). A model was formulated that hypothesized that Stowaway elements originated by internal deletion(s) from a larger autonomous element (like previously described nonautonomous DNA elements) and were amplified to very high copy number by the transposase encoded by the autonomous element (FESCHOTTE et al. 2002B ). The diversity of Stowaway families observed in a single genome was explained by proposing that the families originated as deletion derivatives of distinct lineages of MLEs (FESCHOTTE et al. 2002A , FESCHOTTE et al. 2002B ). If this model is correct, one should encounter Stowaway families that have extensive sequence similarity (i.e., not just in their termini) with MLEs present in the same genome. In addition, the diversity of Stowaway families should correspond with a similar diversity of MLEs in that same genome. Failure to match Stowaway families with MLEs would indicate that the model was incorrect or overly simplistic.
Comparison of all of the MLEs and Stowaway elements in a genome is possible only for Arabidopsis and rice for which entire genome sequences are available. Although remnants of MLE transposases are still recognizable in the sequence of A. thaliana, no full-length MLEs are identifiable (SHAO and TU 2001 ; FESCHOTTE and WESSLER 2002 ; C. FESCHOTTE, unpublished data). Furthermore, Stowaway MITEs are relatively scarce in this species (at least in the sequenced ecotype), with <250 copies organized into fewer than five families (LE et al. 2000 ; C. FESCHOTTE, unpublished data). In contrast, previous searches of a limited amount of rice genomic sequence identified numerous families of Stowaway MITEs and full-length MLEs (BUREAU et al. 1996 ; JIANG and WESSLER 2001 ; SHAO and TU 2001 ; TURCOTTE et al. 2001 ; FESCHOTTE and WESSLER 2002 ). For these reasons, the goal of this study was to characterize all MLE and Stowaway families in rice and determine the extent of sequence relatedness between these two groups.
A semiautomated computational approach was used to identify and compare MLEs and Stowaway MITEs in the two draft genome sequences of rice (GOFF et al. 2002 ; YU et al. 2002 ). In this way 34 MLEs were identified, with 22 considered full-length, as they contain a complete transposase coding region, TIRs, and TSD. Phylogenetic analysis and other criteria, such as the presence or absence of introns, led to their grouping into 25 distinct families falling into three major clades. In addition, up to 33,000 Stowaway MITEs were identified, with the high-copy-number elements grouping into at least 36 families. Surprisingly, none of the 25 MLE families could be associated by simple internal deletion with any of the Stowaway families. Instead, sequence similarity between Stowaway and MLE families was restricted to the TIRs and, in a few cases, to some adjacent subterminal sequence. These data have led us to conclude that most of the Stowaway MITEs in rice were probably cross-mobilized by MLE transposases encoded by distantly related elements.
MATERIALS AND METHODS(w8jki, 百拇医药
Semiautomated mining of full-length rice MLEs:(w8jki, 百拇医药
A series of Perl scripts was written to automate the process of identifying and fetching full-length elements related to a particular transposase. In a first step, the transposase amino acid sequence is used as a query in a local WU-TBLASTN search () against a genomic database. The output file is parsed and the significant hits (in this study, E values <10-5) are extracted from the database along with up to 10 kb of flanking DNA sequence. In a second step, the flanking sequences are searched for the possible ends of the elements using a subroutine called MATCH-TIR. This program scans the 5' and 3' flanking regions of each hit with a 16-mer sliding window for the presence of a consensus motif corresponding to the 5' and 3' ends of the element plus the expected target site duplications (user input). MATCH-TIR extracts 5' and 3' hits (sequences with >80% similarity to the motif) along with 50 nucleotides internal to the hits and produces pairwise alignments between 5' extended hits and the reverse complement of 3' extended hits. The alignments are inspected visually and the best matching pairs (usually fewer than four mismatches in the first 22 nucleotides) are considered as the TIRs of the element. In this study, the Osmar1 transposase sequence was used as the query in a WU-TBLASTN search against two databases. The first database contained(Cédric Feschotte Lakshmi Swamy and Susan R. Wessler)
ABSTRACT., http://www.100md.com
Stowaway is a superfamily of miniature inverted repeat transposable elements (MITEs) that is widespread and abundant in plant genomes. Like other MITEs, however, its origin and mode of amplification are poorly understood. Several lines of evidence point to plant mariner-like elements (MLEs) as the autonomous partners of the nonautonomous Stowaway MITEs. To better understand this relationship, we have taken advantage of the nearly complete genome sequences of two rice subspecies to generate the first inventory of virtually all MLEs and Stowaway families coexisting in a single plant species. Thirty-four different MLEs were found to group into three major clades and 25 families. More than 22,000 Stowaway MITEs were identified and classified into 36 families. On the basis of detailed sequence comparisons, MLEs were confirmed to be the best candidate autonomous elements for Stowaway MITEs. Surprisingly, however, sequence similarity between MLE and Stowaway families was restricted to the terminal inverted repeats (TIRs) and, in a few cases, to adjacent subterminal sequences. These data suggest a model whereby most of the Stowaway MITEs in rice were cross-mobilized by MLE transposases encoded by distantly related elements.
Tc1/mariner is a diverse and widespread superfamily of eukaryotic class 2 transposable elements (reviewed in CAPY et al. 1998 ; PLASTERK et al. 1999 ; PLASTERK and VAN LUENEN 2002). One hallmark of the superfamily is insertion into the dinucleotide TA that is duplicated upon insertion and flanks the element as a target site duplication (TSD). Tc1/mariner elements are relatively short (1.2–3.5 kb) and are simple in structure with terminal inverted repeats (TIRs) and a single gene encoding the transposase. A common model for the transposition mechanism of Tc1/mariner elements has emerged from the functional study of a limited number of animal transposases (PLASTERK and VAN LUENEN 2002 ). The N-terminal region of Tc1/mariner transposases contains DNA-binding domain(s) that bind specifically to the TIRs (PLASTERK et al. 1999 ; LAMPE et al. 2001 ; ZHANG et al. 2001 ). A C-terminal domain is characterized by an amino acid signature called the DDE/D motif consisting of two aspartic acid residues and a glutamic acid residue (or a third D). This motif is required for catalysis of both the DNA cleavage and the strand transfer steps of the "cut and paste" transposition reaction (reviewed in HARTL et al. 1997 ; PLASTERK and VAN LUENEN 2002 ).
Tc1/mariner elements were recently found to be widespread in plants (reviewed in FESCHOTTE et al. 2002A ). The first reported plant members were Soymar1, a mariner-like element (MLE) from soybean (JARVIK and LARK 1998 ) and Lemi1, a pogo-like element from Arabidopsis thaliana (FESCHOTTE and MOUCHES 2000 ). Three additional rice MLEs were subsequently identified by database searches, but none were characterized further (TARCHINI et al. 2000 ; SHAO and TU 2001 ; TURCOTTE et al. 2001 ; FESCHOTTE and WESSLER 2002 ). These five elements were used to derive plant-specific primers that successfully amplified MLE transposase genes in PCR assays with DNA from a wide spectrum of flowering plant genomes (FESCHOTTE and WESSLER 2002 ). For the majority of genomes assayed, multiple divergent lineages of transposases were amplified from single species.%, 百拇医药
Demonstration that MLEs are widespread and diverse in plants provided support for the hypothesis that MLEs are the autonomous elements responsible for the origin and spread of Stowaway, a large group of miniature inverted repeat transposable elements (MITEs; BUREAU and WESSLER 1994 ). MITEs are structurally reminiscent of class 2 nonautonomous elements with their small size (<600 bp), lack of coding capacity, and TIRs (reviewed in FESCHOTTE et al. 2002B ). However, their high copy number and structural homogeneity have served to distinguish them from most of the previously described class 2 elements (WESSLER et al. 1995 ). MITEs were first discovered in plants, where they are now recognized as the predominant type of transposable element associated with the noncoding regions of plant genes. This is particularly evident in the cereals, including rice, maize, barley, and wheat (BENNETZEN 2000 ; FESCHOTTE et al. 2002A ; GOFF et al. 2002 ; YU et al. 2002 ). Vast amounts of MITEs have also been discovered in many invertebrate and vertebrate genomes (reviewed in FESCHOTTE et al. 2002B ).
Most of the tens of thousands of MITEs in plant genomes have been divided into two groups on the basis of the similarity of their TIRs and TSDs: Tourist-like MITEs and Stowaway-like MITEs (WESSLER et al. 1995 ; FESCHOTTE et al. 2002B ). That Stowaway-like MITEs and plant MLEs share similar terminal sequences (5'-CTCCCTCCRT-3', where R stands for A or G) and target site preference (TA) strongly suggested that Stowaway MITEs were mobilized in trans by transposases encoded by MLEs (TURCOTTE et al. 2001 ; FESCHOTTE et al. 2002B ). A model was formulated that hypothesized that Stowaway elements originated by internal deletion(s) from a larger autonomous element (like previously described nonautonomous DNA elements) and were amplified to very high copy number by the transposase encoded by the autonomous element (FESCHOTTE et al. 2002B ). The diversity of Stowaway families observed in a single genome was explained by proposing that the families originated as deletion derivatives of distinct lineages of MLEs (FESCHOTTE et al. 2002A , FESCHOTTE et al. 2002B ). If this model is correct, one should encounter Stowaway families that have extensive sequence similarity (i.e., not just in their termini) with MLEs present in the same genome. In addition, the diversity of Stowaway families should correspond with a similar diversity of MLEs in that same genome. Failure to match Stowaway families with MLEs would indicate that the model was incorrect or overly simplistic.
Comparison of all of the MLEs and Stowaway elements in a genome is possible only for Arabidopsis and rice for which entire genome sequences are available. Although remnants of MLE transposases are still recognizable in the sequence of A. thaliana, no full-length MLEs are identifiable (SHAO and TU 2001 ; FESCHOTTE and WESSLER 2002 ; C. FESCHOTTE, unpublished data). Furthermore, Stowaway MITEs are relatively scarce in this species (at least in the sequenced ecotype), with <250 copies organized into fewer than five families (LE et al. 2000 ; C. FESCHOTTE, unpublished data). In contrast, previous searches of a limited amount of rice genomic sequence identified numerous families of Stowaway MITEs and full-length MLEs (BUREAU et al. 1996 ; JIANG and WESSLER 2001 ; SHAO and TU 2001 ; TURCOTTE et al. 2001 ; FESCHOTTE and WESSLER 2002 ). For these reasons, the goal of this study was to characterize all MLE and Stowaway families in rice and determine the extent of sequence relatedness between these two groups.
A semiautomated computational approach was used to identify and compare MLEs and Stowaway MITEs in the two draft genome sequences of rice (GOFF et al. 2002 ; YU et al. 2002 ). In this way 34 MLEs were identified, with 22 considered full-length, as they contain a complete transposase coding region, TIRs, and TSD. Phylogenetic analysis and other criteria, such as the presence or absence of introns, led to their grouping into 25 distinct families falling into three major clades. In addition, up to 33,000 Stowaway MITEs were identified, with the high-copy-number elements grouping into at least 36 families. Surprisingly, none of the 25 MLE families could be associated by simple internal deletion with any of the Stowaway families. Instead, sequence similarity between Stowaway and MLE families was restricted to the TIRs and, in a few cases, to some adjacent subterminal sequence. These data have led us to conclude that most of the Stowaway MITEs in rice were probably cross-mobilized by MLE transposases encoded by distantly related elements.
MATERIALS AND METHODS(w8jki, 百拇医药
Semiautomated mining of full-length rice MLEs:(w8jki, 百拇医药
A series of Perl scripts was written to automate the process of identifying and fetching full-length elements related to a particular transposase. In a first step, the transposase amino acid sequence is used as a query in a local WU-TBLASTN search () against a genomic database. The output file is parsed and the significant hits (in this study, E values <10-5) are extracted from the database along with up to 10 kb of flanking DNA sequence. In a second step, the flanking sequences are searched for the possible ends of the elements using a subroutine called MATCH-TIR. This program scans the 5' and 3' flanking regions of each hit with a 16-mer sliding window for the presence of a consensus motif corresponding to the 5' and 3' ends of the element plus the expected target site duplications (user input). MATCH-TIR extracts 5' and 3' hits (sequences with >80% similarity to the motif) along with 50 nucleotides internal to the hits and produces pairwise alignments between 5' extended hits and the reverse complement of 3' extended hits. The alignments are inspected visually and the best matching pairs (usually fewer than four mismatches in the first 22 nucleotides) are considered as the TIRs of the element. In this study, the Osmar1 transposase sequence was used as the query in a WU-TBLASTN search against two databases. The first database contained(Cédric Feschotte Lakshmi Swamy and Susan R. Wessler)