MBE Advance Access originally published online on January 12, 2005
Molecular Biology and Evolution 2005 22(4):976-990; doi:10.1093/molbev/msi082
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
Research Article |
Mobilization of the Active MITE Transposons mPing and Pong in Rice by Introgression from Wild Rice (Zizania latifolia Griseb.)
* Laboratory of Molecular Epigenetics, Institute of Genetics and Cytology, Northeast Normal University, Changchun, China; Division of Biological Sciences, University of Missouri; and The National Centre of Plant Transgenic Research & Commercialization, Gongzhuling, China
E-mail: baoliu{at}nenu.edu.cn
 |
Abstract |
|---|
 TOP Abstract IntroductionMaterials and MethodsResultsDiscussionSupplementary MaterialAcknowledgementsReferences |
|---|
Hybridization between different species plays an important role in plant genome evolution, as well as is a widely used approach for crop improvement. McClintock has predicted that plant wide hybridization constitutes a "genomic shock" whereby cryptic transposable elements may be activated. However, direct experimental evidence showing a causal relationship between plant wide hybridization and transposon mobilization has not yet been reported. The miniature-Ping (mPing) is a recently isolated active miniature inverted-repeat transposable element transposon from rice, which is mobilized by tissue culture and
-ray irradiation. We show herein that mPing, together with its putative transposase-encoding partner, Pong, is mobilized in three homologous recombinant inbred lines (RILs), derived from hybridization between rice (cultivar Matsumae) and wild rice (Zizania latifolia Griseb.), harboring introgressed genomic DNA from wild rice. In contrast, both elements remain immobile in two lines sharing the same parentage to the RILs but possessing no introgressed DNA. Thus, we have presented direct evidence that is consistent with McClintock's insight by demonstrating a causal link between wide hybridization and transposon mobilization in rice. In addition, we report an atypical behavior of mPing/Pong mobilization in these lines, i.e., the exclusive absence of footprints after excision.
Key Words: wide hybridization • transposon mobilization • MITEs • genome evolution • rice
|
Introduction |
|---|
|
TOPAbstractIntroductionMaterials and MethodsResultsDiscussionSupplementary MaterialAcknowledgementsReferences |
|---|
The miniature inverted-repeat transposable elements (MITEs) have been uncovered in several eukaryotes, including plants, animals, and humans, and are numerically the most prominent type of all transposable elements (TEs) in plants (Feng et al. 2002; Feschotte, Jiang, and Wessler 2002). For example, sequence analysis of the rice chromosome 4 has showed that MITEs account for nearly 50% of all the numbers of repetitive DNA (Feng et al. 2002). MITEs represent a special group of DNA transposons (class II elements) and have been classified into two superfamilies, Tourist-like and Stowaway-like, based on the similarity of their terminal inverted repeats (TIRs) and target site duplications (TSDs) (Feschotte, Jiang, and Wessler 2002). Several characterized MITEs are found to preferentially reside in low-copy, genic regions of plant genomes, underscoring their possible roles in the evolution of structure and function of plant genes (Bureau and Wessler 1992, 1994a, 1994b; Bureau, Ronald, and Wessler 1996; Wessler, Bureau, and White 1995; Zhang, Arbuckle, and Wessler 2000). The recent characterization of an active MITE family, the miniature-Ping (mPing) from rice (Jiang et al. 2003; Kikuchi et al. 2003; Nakazaki et al. 2003), together with the identification of several autonomous, transposase-encoding partners for specific groups of MITEs (Feschotte, Jiang, and Wessler 2002; Zhang et al. 2004), has greatly furthered our understanding on mechanistic issues concerning origin and transposition of MITEs.
The rice mPing family is the only active MITEs so far characterized in any organism (Jiang et al. 2003). mPing is a 430–base pair (bp) DNA repeat with TIRs (15 bp) and TSDs (TAA or TTA) typical of a Tourist-like MITE (Jiang et al. 2003; Kikuchi et al. 2003). Albeit exceptionally low in copy number compared with other characterized MITE families in plants (Feschotte, Jiang, and Wessler 2002; Jiang et al. 2004), mPing can be effectively mobilized by tissue culture (Jiang et al. 2003; Kikuchi et al. 2003) and also by -ray irradiation (Nakazaki et al. 2003), with the mobilized copies preferentially inserting into single-copy genomic regions (Jiang et al. 2003). Because mPing has no coding capacity, the transposase required to catalyze its transposition has to be provided in trans (Feschotte, Jiang, and Wessler 2002; Jiang et al. 2003, 2004). Based on sequence homology and comobilization with mPing, either of the two mPing-related and transposase-encoding elements, called Ping and Pong, is implicated as a possible autonomous element responsible for mPing mobilization under the specific conditions (Jiang et al. 2003; Kikuchi et al. 2003).
For the mPing MITEs to be relevant to rice genome evolution, they likely have been capable of being activated under naturally occurring circumstances. Indeed, it was found that the copy number of mPing varies dramatically between the two subspecies of rice, japonica and indica, as well as between the two subgroups of the japonica rice, with temperate japonica cultivars possessing a markedly higher copy number of mPing (Jiang et al. 2003). This finding has been implicated to suggest stress-induced mobilization of mPing under extreme environmental conditions during domestication of temperate japonica cultivars (Jiang et al. 2003), a situation consistent with McClintock's "genomic shock" theory for cryptic transposon activation by stress (McClintock 1984). It has been shown for several long-terminal repeat (LTR) retrotransposons in plants that some biotic and abiotic stresses, such as pathogen attack and drought, may cause transcriptional element activation (Wessler 1996; Grandibastien 1998; Bennetzen 2000; Kalendar et al. 2000) and possibly transposition (Kalendar et al. 2000; Wendel and Wessler 2000). Recently, it was found that a novel transposon distantly related to Mutator, called Jattery, was mobilized in a maize inbred line by infection with barley mosaic strip virus (Xu et al. 2004). Similarly, environmental stresses, such as low temperature (Giraud and Capy 1996), often induce mobilization of transposons in Drosophila (Capy et al. 2000). It remains unknown, however, what specific factors in nature can lead to mobilization of a MITE transposon in any organism.
Hybridization between genetically differentiated natural plant populations is a frequent phenomenon, which contributes to genome evolution and can lead to speciation via allopolyploidy or at the homoploid level (Anderson and Stebbins 1954; Stebbins 1959; Grant 1981; Rieseberg 1995; Wendel 2000; Rieseberg et al. 2003; Arnold 2004). Introgression of uncharacterized DNA segments from a related but distinct species into a crop has also been a widely used approach for introducing useful traits. In plant breeding, usually attention is focused on the transfer of desired alleles (traits) from the donor species into the recipient, through hybridization and successive backcrossing, while potential impact of alien DNA segment integration, other than the transfer per se, on the recipient genome has not been studied. In this regard, it is notable that there have been several studies in animals demonstrating that integration of foreign DNA can cause the host genome to undergo extensive and genome-wide alterations in DNA methylation (mostly de novo methylation) of both cellular genes and transposon-associated DNA repeats (Heller et al. 1995; Remus et al. 1999; Muller, Heller, and Doerfler 2001). Similarly, we found recently that extensive and heritable changes in DNA methylation patterns also occurred in a set of homologous rice recombinant inbred lines (RILs) with introgressed DNA segments from wild rice (Zizania latifolia Griseb.) (Liu et al. 2004). Given the frequent correlation between a transposon's activity and its methylation state (Chandler and Walbot 1986; Schwartz and Dennis 1986; Banks, Masson, and Fedoroff 1988; Hirochika, Okamoto, and Kakutani 2000; Miura et al. 2001; Ros and Kunze 2001; Singer, Yordan, and Martienssen 2001; Cui and Fedoroff 2002; Kato et al. 2003; Lippman et al. 2003), it is possible that some quiescent TEs might have been activated in these plants, as was indeed the case for an LTR retrotransposon Tos17 (Liu and Wendel 2000). This observation is reminiscent of what McClintock envisioned that wide hybridization in plants might activate quiescent transposons and cause genome restructuring (McClintock 1984). There are several additional lines of evidence in both plants (Hanson et al. 1999; Comai 2000; Comai et al. 2000; Kashkush, Feldman, and Levy 2002, 2003) and animals (Capy et al. 1990; Labrador and Fontdevila 1994; Kidwell and Lisch 1998; R. J. W. O'Neill, M. J. O'Neill, and Graves 1998; Labrador et al. 1999; Kidwell and Lisch 2000), which are consistent with McClintock's insightful prediction. Nevertheless, all available studies in plants (cited above) have only demonstrated transcriptional activation of TEs, or in the case of tetraploid cotton (Hanson et al. 1999), contributory factors other than hybridization cannot be ruled out as a cause for TE transposition because the evolutionary history of natural cotton (Gossypium hirsutum) is more than a million years old (Senchina et al. 2003). Therefore, to our knowledge, unequivocal experimental evidence demonstrating a causal link between wide hybridization and transposon mobilization has not been reported in plants.
In this paper, we present direct evidence showing mobilization of the rice mPing MITE and one of its closely related, transposase-encoding elements, Pong, in three homologous rice RILs as a result of intergeneric hybridization and introgression from wild rice (Z. latifolia Griseb.). A salient feature of mPing and Pong excisions in these lines is the exclusive absence of footprints. We discuss possible causes for the wide hybridization–induced transposon mobilization and its implications for plant genome evolution and crop domestication.
|
Materials and Methods |
|---|
|
TOPAbstractIntroductionMaterials and MethodsResultsDiscussionSupplementary MaterialAcknowledgementsReferences |
|---|
Plant Materials
Three rice RILs (RZ1, RZ2, and RZ35) derived from hybridization between rice (cultivar Matsumae) and a local accession of wild rice (Z. latifolia Griseb.) by a novel, simple approach called "repeated pollination" (Liu et al. 1999; fig. 1) were used in the present study. The RILs are homogeneous in phenotype and exhibit heritable, novel morphological characteristics in multiple traits compared with their rice parental cultivar Matsumae (Liu et al. 1999). The RILs have been maintained along with their exact parental lines and two lines of the same parentage, i.e., sibling lines (RZ36 and RZ60), by strict selfing in our laboratory. These plant materials are available upon request for research purposes.
|
Southern Blot Analysis
Genomic DNA was isolated from expanded leaves of individual plants by a modified cetyltrimethyl ammonium bromide method (Kidwell and Osborn 1992) and purified by phenol extractions. Genomic DNA (
3 µg per lane) of the various plant lines was digested by EcoRI, HindIII, or XbaI (New England Biolabs Inc., Beverly, Mass.). Digested DNA was run through 1% agarose gels and transferred onto Hybond N+ nylon membranes (Amersham Pharmacia Biotech, Piscataway, N.J.) by the alkaline transfer recommended by the supplier. For molecular characterization of the RILs, putative DNA fragments isolated from the amplified fragment length polymorphism (AFLP) gels were reamplified with the same pair of primers as in the original AFLP amplifications, gel purified and used as probes. For estimating copy number of mPing, Pong, and Ping, the probe fragments were polymerase chain reaction (PCR) amplified by using the following element- or region-specific primers—(1) mPing (positions 6–430): forward, 5'-GTCACAATGGGGGTTTCACT, reverse, 5'-GGCCAGTCACAATGGCTAGT; (2) Pong-ORF2 (positions 3199–4255): forward, 5'-AACGAGGCTTCTGACCATCG, reverse, 5'-CAGGTTCCTGAACGGTTGAT; (3) Pong specific (positions 158–1372): forward, 5'-GGGGTGAAACAGCATTGAGA, reverse, 5'-TGTGGTTGCAAAGAAGACCA; and (4) Ping specific (positions 327–1513): forward, 5'-CTACGGAGTACACCGCAACC, reverse, 5'-AATGGATTGCCTACTGCTGACT. Identities of all probe fragments were confirmed by sequencing. The fragments were then gel purified and labeled with fluorescein-11-deoxyuridine triphosphate by the Gene Images random prime-labeling module (Amersham Pharmacia Biotech). Hybridization signal was detected by the Gene Images CDP-Star detection module (Amersham Pharmacia Biotech) after washing at a stringency of 0.2 x saline sodium citrate, 0.1% sodium dodecyl sulfate for 2 x 50 min. The filters were exposed to X-ray film for 1–3 h depending on signal intensity.
PCR-Based Locus Assay on mPing and Pong Excision
To detect possible excisions of mPing and Pong, a set of 53 pairs of locus-specific primers each bracketing an intact mPing and four pairs of locus-specific primers each bracketing an intact Pong in the standard laboratory cultivar for the japonica rice ssp., Nipponbare (http://rgp.dna.affrc.go.jp), was designed by the Primer 3 software (http://biocore.unl.edu/cgi-bin/primer3/primer3_www.cgi). Loci containing a member of mPing or Pong in rice cultivar Matsumae (parent for the RILs) were identified, and the corresponding primers were given in supplementary table 1. An additional six loci in cultivar Matsumae flanking the 5' end of mPing (three loci) or Pong (three loci) were isolated by thermal asymmetric interlaced PCR (TAIL-PCR) (Liu et al. 1995) using the mPing or Pong subterminal-specific primers as reported (Jiang et al. 2003). The contiguous 3'-flanking sequences of these loci were identified based on the Nipponbare genome sequence information (http://rgp.dna.affrc.go.jp) by BlastN search. Locus-specific primers for these mPing- or Pong-bracketing loci were designed, as above, and listed in supplementary table 1. PCR amplifications were performed at annealing temperatures ranging from 58°C to 62°C depending on different pairs of primers. To ensure that the presence or absence of an expected PCR product in Matsumae and the RILs by using element-bracketing flanking primers was not due to PCR artifact or bias (particularly for the longer element Pong), PCR amplifications were also performed at all element-containing loci using mPing- or Pong-internal primers in combination with each of the locus-specific, 3'-flanking primers. The amplicons were visualized by ethidium bromide staining after electrophoresis through 2% agarose gels. All identified empty donor sites for mPing and Pong excisions were isolated and sequenced, together with their corresponding element-containing loci.
Isolation of mPing and Pong Insertion Sites in the RILs by Transposon Display
Transposon Display (Van den Broeck et al. 1998; Casa et al. 2000) was performed by combining the mPing or Pong subterminal-specific primers (Jiang et al. 2003) with a set of intersimple sequence repeat (ISSR) primers (available upon request) and visualizing the amplicons on 4% agarose gels by ethidium bromide staining, a method similar to the retrotranspon Display in barley (Schulman, Flavell, and Ellis 2004). Novel bands appeared in an RIL in the ISSR-mPing/Pong amplifications but absent in the corresponding ISSR-alone amplifications were considered as putative mPing or Pong de novo insertions and isolated for sequencing. The element insertions were then confirmed by PCR amplifications using both flanking primers and an element-specific internal primer together with the 3'-flanking primers, as described above for the excision analysis.
|
Results |
|---|
|
TOPAbstractIntroductionMaterials and MethodsResultsDiscussionSupplementary MaterialAcknowledgementsReferences |
|---|
Molecular Characterization of the Rice RILs
Generation of a series of rice lines with possible introgressed genomic DNA from wild rice (Z. latifolia Griseb.) by a novel, simple sexual hybridization approach called repeated pollination was described earlier (Liu et al. 1999) and is diagramed in figure 1. The mechanism by which introgression occurred remains elusive, but it may bear similarity to the reported interspecific transfer of nuclear DNA from irradiation-killed pollens in tobacco (Pandey 1975). Three typical rice lines (designated as RZ1, RZ2, and RZ35), currently at the 9th–11th selfed generations, are homogeneous in phenotypes, yet exhibit heritable, novel morphological characteristics in multiple traits compared with their rice parent cultivar Matsumae (e.g., fig. 2A and B). The three lines were thus virtually RILs. The lines were characterized by genome-wide AFLP fingerprinting at >6,000 loci as possessing <0.1% genomic DNA putatively derived from Z. latifolia (fig. 2C, D, and E; Y. Wang, Z. Dong, Z. Zhang, Y. Shen, and B. Liu, in preparation). Apart from potential introgression, it is noted that in the AFLP patterns, genomic rearrangements, i.e., loss of rice parental bands and gain of novel bands in the RILs, are also widely occurring (e.g., fig. 2C, D, and E, marked by triangles; Y. Wang, Z. Dong, Z. Zhang, Y. Shen, and B. Liu, in preparation). Importantly, when several putative Zizania-specific bands (bearing no homology to the complete rice genome sequence at http://rgp.dna.affrc.go.jp) isolated from the AFLP lanes of the RILs (e.g., fig. 2C, D, and E, arrowed) were used as probes for Southern blot analysis, all three RILs were found to contain genomically integrated Zizania-specific DNA repeats (e.g., fig. 2F and G), thus verified their identity as bona fide introgression lines. Interestingly, in some cases apparent amplification of Zizania-specific sequences (e.g., fig. 2F for RZ35) and loss of rice parental bands (e.g., fig. 2G for RZ2) occurred in the RILs.
|
Concordant Alterations in the Banding Patterns of mPing and Pong in the Rice RILs
The mPing MITE transposon (GenBank accession number AB087615) does not contain HindIII and XbaI restriction sites (Kikuchi et al. 2003); thus, Southern blot hybridization with these enzyme digestions should enable a conservative estimation on the element copy number and its possible mobilization or rearrangement. Evidently, the three RILs contained similar copy numbers of mPing as their rice parent Matsumae but showed marked difference in banding patterns from each other and from that of their rice parent in both enzyme digests (fig. 3A). The banding pattern for a given RIL versus that of the parent is most readily deciphered as loss of parental bands and gain of novel bands, a feature conform with mobilization of type II DNA transposons (Feschotte, Jiang, and Wessler 2002; Casacuberta and Santiago 2003; Kikuchi et al. 2003) in the RILs. Furthermore, the wild rice (Z. latifolia) does not hybridize with mPing under the stringency used (fig. 3A, lanes Zizania), and hence, none of the changing patterns in the RILs should be attributable to elements directly transferred from this donor species. To further verify this point, PCR amplification with mPing-specific primers was conducted on three independent isolates of the local accession (used in the initial crossing) of Z. latifolia, together with the rice lines. Results indicated that a band of the expected size (425 bp) was amplified from all rice lines, but no amplicon was obtained from the Zizania plants (fig. 3B), thus confirming absence of any mPing homolog in this wild species. This therefore validated that the changed Southern blot profiles of mPing in the RILs relative to Matsumae (fig. 3A) were not due to transfer from Z. latifolia. To investigate homo- or heterogeneity of the banding patterns of mPing, four to five randomly chosen individuals from each line (parent and RILs) were subjected to the same Southern blot analysis. Identical patterns were observed among all individuals for a given line in both enzyme digests (fig. 3C and data not shown), indicating lack of heterozygosity and, hence, current stability of mPing in these lines.
|
When the same blots were probed with an isolated fragment in the second open reading frame (ORF2) of the mPing-related and transposase-encoding element Pong (Jiang et al. 2003; GenBank accession number BK000586, positions 3199–4255, which contains no restriction site for XbaI and two sites for HindIII, with both being upstream of position 3199), conservation in element copy number but apparent alterations in banding pattern were also apparent in the RILs (fig. 3D). Several features characterizing the pattern changes of Pong in the RILs were concordant with mPing, including lack of hybridization signal in Z. latifoia (fig. 3D, lane Zizania), a loss or gain type of alteration (fig. 3D), and homogeneity among randomly chosen individuals for a given line (fig. 3E). To further verify that the changing Southern blot patterns of the Pong-ORF2 fragment in the RILs were not caused by direct transfer from Z. latifolia, PCR amplifications were performed with both the Pong-ORF2 and Pong-specific (see the following section) primers on the same Z. latifolia isolates as for mPing described above. Consequently, only weaker bands with smaller sizes than the corresponding rice Pong-ORF2 or Pong-specific fragment were amplified from Z. latifolia (fig. 3F and G). Sequence analysis showed that none of the Z. latifolia–derived fragments shared meaningful homology to any region of Pong (data not shown). Therefore, as in the case of mPing, the changing patterns of Pong in the RILs could not have resulted from elements transferred from Z. latifolia.
Because the ORF2 of Pong shares significant homology (85% identity) with the corresponding region of the direct donor of mPing, called Ping (Jiang et al. 2003; Kikuchi et al. 2003), it is important to distinguish the two elements. We thus designed element-specific primers (Pong specific and Ping specific) in the region upstream of ORF1, wherein no sequence homology exists between the two elements or between either of them and mPing, and attempted PCR amplification for the corresponding fragments. A product of the expected size was generated for Pong in all four lines (parental and RILs, fig. 3G), but no amplification was observed for Ping in any line even after repeated efforts. This, together with the fact that a Ping fragment of the expected size was readily obtained with the same pair of Ping-specific primers from cultivar Nipponbare that is known to harbor at least one copy of the element (Jiang et al. 2003), suggests the absence of the Ping element in the rice lines we used. When the same blots (XbaI digest, whose restriction site is absent from both Pong and Ping) used above were hybridized, respectively, with the Pong-specific (amplified from Matsumae) and Ping-specific (amplified from Nipponbare) fragments, a pattern similar to that of the Pong-ORF2 probe was observed for the Pong-specific fragment (fig. 3H), whereas no hybridization signal was detected in any line for the Ping-specific fragment (not shown), thus confirming the absence of Ping in these rice lines.
Detection of mPing and Pong Excision and Insertion in the Rice RILs
The foregoing shows marked and probably nonrandom banding-pattern alterations in both mPing and Pong in the rice RILs. Two possible explanations can be conceived for this observation: (1) transit concomitant mobilization of the two elements via a "cut and paste model" followed by efficient element repression and stable inheritance and (2) genomic rearrangement in chromosomal regions involving mPing and Pong, followed by homogenization through continued selfing. Given the remarkable concordance of the changing patterns in the RILs not only between mPing and Pong (fig. 3C and E) but also between different probe regions within Pong (fig. 3D and H), we deemed that the possibility for genomic rearrangement as the sole cause was unlikely. To test for possible element excision, we designed a set of locus-specific primers bracketing mPing (53 pairs) and Pong (4 pairs) of the japonica rice standard laboratory cultivar Nipponbare based on its available whole genome sequence (http://rgp.dna.affrc.go.jp) and performed PCR amplifications with genomic DNA of the parental line Matsumae. By testing all these primer pairs, we identified eight candidate mPing-containing loci (supplementary table 1) in this cultivar, as PCR products identical in size to those amplified from Nipponbare were generated, which were of the expected sizes encompassing a copy of mPing, but no Pong-containing locus was obtained, as only small-sized bands denoting the absence of the element were amplified. Sequencing confirmed that all eight candidate mPing-containing loci indeed encompassed an intact mPing element with conserved 15-bp TIRs (GGCCAGTCACAATGG) and trinucleotide TSDs (TAA or TTA) (Jiang et al. 2003; Kikuchi et al. 2003), implicating that they were potentially mobile copies. By TAIL-PCR amplification (Liu et al. 1995), coupled with querying the Nipponbare genome sequence information, we designed more locus-specific primers and identified another three distinct loci bracketing mPing and three loci bracketing Pong in Matsumae (supplementary table 1). PCR assay at all 11 mPing-bracketing and 3 Pong-bracketing loci showed that 10 of the 11 mPing-bracketing loci and 2 of the 3 Pong-bracketing loci showed evidence for element excision in the RILs, as smaller PCR products consistent with loss of an mPing or a Pong copy were amplified (fig. 4A and B, upper panels, and supplementary table 1). To further verify that the amplification of a larger versus a smaller PCR product in Matsumae and the RILs by using the element-bracketing flanking primers was not due to PCR bias (particularly for the longer element Pong), PCR amplifications were also performed at all 14 loci using mPing- or Pong-internal primers in conjunction with each of the locus-specific 3'-flanking primers (see Materials and Methods). It was found that in all primer combinations, amplifications with the mPing/Pong-internal primers completely corroborated those amplified with the element-bracketing primers (fig. 4A and B, lower panels; data not shown). With regard to the amplification by these locus-specific primers on the donor species, Zizania, it was found that all primer pairs for these 14 mPing- or Pong-bracketing loci did not amplify a distinct band of the expected sizes (large or small) from Z. latifolia (fig. 4A and B, upper panels; data not shown), indicating that the appearance of the smaller bands in the RILs (relative to larger bands in Matsumae) were not due to transfer of element-empty sites from Zizania. This observation was also completely verified by the element-internal primers (fig. 4A and B, lower panels). The sole exception was amplification by a Pong-internal primer coupled with a 3'-flanking region primer (TAIL-Pong2) that generated a band from Zizania around the size of 209 bp (the expected size for the presence of a Pong copy, as in Matsumae) (fig. 4B, the lower rightmost panel). Sequencing, however, showed that this amplified fragment from Zizania had no homology to mPing or the flanking region (data not shown). Taken together, it can be concluded that all element-empty sites in the RILs were the result of de novo excisions of mPing or Pong. Specifically, mPing at two loci excised in all three RILs and at eight loci in two RILs (seven in RZ1/RZ2 and one in RZ2/RZ35) (e.g., fig. 4A and supplementary table 1). One locus (mPL9) showed genomic changes that might not be due to mPing excision: this locus amplified a mPing-containing fragment in Matsumae and RZ35 but did not amplify any product (upper or lower band) in RZ1 and RZ2 (data not shown), indicating regional elimination, insertion of Zizania DNA, or sequence change at the primer regions in these two lines. For the three Pong-bracketing loci, two showed evidence for element excision in two or all three RILs as smaller fragment consistent with the loss of Pong was amplified from these lines (fig. 4B and supplementary table 1).
|
It has been shown that mPing and Pong often excise imprecisely, i.e., leaving footprints, both under tissue culture conditions (Kikuchi et al. 2003) and by
-ray irradiation (Nakazaki et al. 2003). Thus, to analyze the mPing and Pong excision patterns in the RILs, we sequenced all 22 identified empty donor loci for mPing and five loci for Pong, isolated from the RILs, together with all corresponding element-containing loci isolated from the parental line Matsumae. Sequence comparison showed that, contrary to our expectation and in sharp contrast to previously reported results (Kikuchi et al. 2003; Nakazaki et al. 2003), for all the mPing and Pong empty donor loci analyzed in the RILs, element excision did not leave behind any footprints (supplementary table 1). To test if the excision of mPing and Pong was accompanied by element de novo insertion in the RILs, ISSR-mPing and ISSR-Pong PCR amplifications, a modified method of the transposon Display technique (Van den Broeck et al. 1998; Casa et al. 2000) was used to look for novel bands present only in the RILs. By using about 40 ISSR primers (sequence available upon request) in combination with mPing or Pong, 16 and 3 candidate bands, respectively, for mPing and Pong, were isolated from the RILs (e.g., fig. 5A). Sequence analysis identified, respectively, seven and two distinct clones that contain, at their 5' termini, the expected partial mPing and Pong sequences with typical TIRs and TSDs (supplementary table 2), implying that they were likely mPing and Pong de novo insertions in the RILs. All the nine mPing- or Pong-containing clones were mapped to unique-copy regions of different chromosomes in the Nipponbare genome (supplementary table 2). Taking advantage of the Nipponbare genome sequence information, we again designed locus-specific primers for all these nine clones and performed PCR amplifications. We found that all nine loci generated amplification products in the range consistent with lack of mPing or Pong in the parental line Matsumae. On the other hand, larger fragments coinciding with adding an intact copy of mPing or Pong were amplified from the RILs (e.g., fig. 5B and C, leftmost panel) from which the band was initially isolated from the ISSR-mPing or -Pong gels, as was verified by hybridizing blots of the gels with the mPing or Pong (ORF2) probes (e.g., fig. 5B, right panel, and 5C, middle panel). To ensure that the failure to amplify the Pong-containing band (>5 kb) in Matsumae was not caused by PCR bias, a Pong-internal primer was used together with each of the 3'-flanking locus-specific primers of all the analyzed loci. As in the case for element excisions, for both analyzed loci, de novo insertion of a Pong copy was verified (e.g., fig. 5C, rightmost panel). With regard to the amplification by these locus-specific primers on Z. latifolia, though in some cases faint fragments were amplifiable from this species, in no case the band was of the same size as those in the RILs containing a member of mPing or Pong element. This suggested that the amplifications from Z. latifolia were nonspecific, as was indeed confirmed by gel blot analysis (fig. 5B and C and data not shown). Further sequencing of either the full-length sequence (in the case of mPing-containing loci) or both the 5'- and 3'-flanking sequences (in the case of Pong-containing loci) of the PCR products amplified from the RILs confirmed that all harbored the mPing- and Pong-specific 15-bp TIRs (GGCCAGTCACAATGG) boarded by the TAA or TTA TSDs (supplementary table 2). Therefore, all larger amplification products for a given locus in one or more RILs were the consequence of mPing or Pong de novo insertions. In agreement with the genomic Southern blot analysis result (fig. 3), all mPing and Pong insertions revealed by PCR amplifications were conserved among randomly chosen individuals for a given line (fig. 5B and data not shown), indicating that the insertions likely occurred at initial stages of the F1 plant (Liu et al. 1999), and/or of a non-random nature, followed by rapid immobilization and stable inheritance of the insertion sites.
|
Taken together, the high frequency of mPing/Pong excision-insertion events that occurred in the RILs probably could account for most, if not all, of the altered banding patterns detected by Southern blot analysis with these elements (fig. 3). The mobilization of mPing/Pong was most probably triggered by introgression from Zizania because both Southern blotting with mPing/Pong probes and locus assay at all 11 loci containing highly mobile mPing copies (supplementary table 1) failed to detect banding-pattern changes or excision events (exemplified in fig. 6A and B) in two rice lines (RZ36 and RZ60) sibling to the three RILs. These two lines, RZ36 and RZ60, were derived from the same hybridization cross by repeated pollination (fig. 1) as the three RILs but contained no integrated DNA from Zizania based on AFLP analysis and Southern blot hybridization with Zizania-specific DNA repeats (data not shown); instead, they were genomically and phenotypically indistinguishable from the parental line Matsumae.
|
Tissue Culture–Mobilized mPing and Pong in Rice Cultivar Matsumae Also Are Devoid of Excision Footprints
Because it was shown earlier that mPing excisions induced by either tissue culture (Kikuchi et al. 2003) or
-ray irradiation (Nakazaki et al. 2003) often left footprints, results of the present study that none of the 27 mPing/Pong excisions in the RILs left footprint are unexpected and atypical. To test if this unusual phenomenon—the exclusive absence of mPing and Pong excision footprints in the RILs—was due to difference in the elicitors (introgression vs. tissue culture or -ray irradiation) or a genotype-dependent trait, we subjected the three RILs and their parent, Matsumae, to tissue culture. Regenerated plants (regenerants) from 3- to 12-month-old calli of the RILs and the parent were subjected to Southern blot analysis using mPing and Pong-ORF2 as probes. Interestingly, tissue culture seemed capable of mobilizing mPing and Pong only in the parental cultivar Matsumae, as was evidenced by loss of several original bands and appearance of novel bands in the regenerants of this cultivar (fig. 7A); whereas no change in the hybridization patterns denoting transpositional activity for mPing or Pong was detected in regenerants of the RILs cultured for the same periods as for Matsumae (fig. 7B and data not shown). This differential response to tissue culture between Matsumae and the RILs with regard to mPing/Pong activation was also confirmed by the PCR-based locus assay at all available loci (supplementary table 1) that harbored the elements for Matsumae and each of the RILs (fig. 7C and D). A notable feature for mPing/Pong mobilization in the regenerants of Matsumae was the near-identical banding patterns among four randomly chosen regenerants from calli subcultured for different periods (fig. 7A), suggesting that the mobilization events occurred very early, at the initial callus cells, and/or both the excisions and insertions were highly nonrandom. Further investigations by analyzing more element-containing loci on a large number of independent regenerants may distinguish the two possibilities.
|
To investigate whether the tissue culture–induced mPing and Pong excisions in the regenerants of Matsumae would leave footprints or not, we performed PCR amplifications on 9–10 independent regenerants of this cultivar with locus-specific primers at 11 loci that each bracketed an mPing and three loci that each bracketed a Pong in uncultured seed plant of Matsumae (supplementary table 1). We found that four mPing-bracketing loci and one Pong-bracketing locus showed element excision in all the regenerants studied (fig. 7C and data not shown), while the rest mPing/Pong loci did not show excision in any regenerant. In contrast, all three RILs at the analyzed mPing/Pong-containing loci showed complete stability (e.g., fig. 7D and data not shown), an observation mirroring the gel blot data (fig. 7B). The "all-or-none" fashion of element excision in independent regenerants of Matsumae at the analyzed loci is interesting and merits further investigation. Totally, we isolated 31 and 10 empty donor loci, respectively, for mPing and Pong in these locus-regenerant combinations (supplementary table 3). Sequence analysis showed that none of these independent excisions (31 for mPing and 10 for Pong) had left behind any footprint (supplementary table 3), a feature identical with the situation in the RILs derived from Matsumae, as described above. The exclusive absence of footprints upon mPing/Pong excision, regardless of being induced by introgression or tissue culture, in cultivar Matsumae provided a sharp contrast to previous results on other cultivars (Kikuchi et al. 2003; Nakazaki et al. 2003). This discrepancy, however, can be readily reconciled if the presence or absence of footprint upon mPing/Pong excision is largely a genotype-dependent trait.
|
Discussion |
|---|
|
TOPAbstractIntroductionMaterials and MethodsResultsDiscussionSupplementary MaterialAcknowledgementsReferences |
|---|
mPing/Pong Mobilization in the Rice RILs Is Induced by Introgression from Wild Rice
It has been documented that mPing and Pong transposons can be activated by tissue culture (Jiang et al. 2003; Kikuchi et al. 2003) and
-ray irradiation (Nakazaki et al. 2003). If these transposons have played a role in rice genome evolution, they should be capable of being activated by naturally occurring factors. We have shown in this paper that introgression from wild rice (Z. latifolia Griseb.) could apparently mobilize mPing and Pong. This conclusion is confidently drawn because the presence of Zizania introgression and because no alternative is conceivable to explain the observed results. (1) The rice parental line cultivar Matsumae is a genetically pure line, as rice is a predominantly self-pollinating plant, and moreover, because the initial purpose for making the cross was to look for possible introduced phenotypic traits and phenotypic variations (Liu et al. 1999), the specific plants used for the work had been further selfed (bagging) for five successive generations to ensure their inbred nature before making the cross. In fact, the inbred nature of Matsumae has also been corroborated in this study with respect to all the analyzed mPing and Pong loci, as in no case heterozygosiy was observed in either the Southern blot analysis or the PCR-based locus assay on random individuals of Matsumae (figs. 3 and 5). Therefore, parental heterozygosity can be confidently ruled out as a causal factor for the changing patterns of mPing and Pong in the RILs compared with their rice parent. (2) Contamination by pollens of other rice cultivars was deemed extremely unlikely because extensive precautions were taken both in the hybridization manipulations (emasculation and pollination) and in later propagations of the RILs. Moreover, the following two lines of evidence testify that contamination could not have occurred at any stage. First, all the hybridization manipulations to produce the F1 hybrid (fig. 1) were performed in an isolated greenhouse where only Matsumae and Z. latifolia plants were grown. Thus, contamination could not occur at the F1 stage. Second, the three studied RILs, each being a progeny of an independent F2 individual derived from the same phenotypically novel F1 plant (fig. 1), showed mPing and Pong mobilizations (see Results). Therefore, if contamination were a cause for the changed mPing and Pong patterns, then by logic, the three RILs had to be all contaminated independently in the F2 or later generations. Given the strict precautions taken, it apparently was not possible for this to occur. Thus, the two lines of evidence have confidently ruled out contamination as a cause for the changed patterns of mPing and Pong in the RILs. (3) Since production, all relevant plant materials (the RILs, their sibling lines without introgression, along with the rice parental line) have been grown under the same normal conditions (Liu et al. 1999), that is, they have never been subjected to any environmental extremes. (4) Southern blot analysis with the mPing/Pong probes and locus assay at 11 mPing-containing loci on two lines (RZ36 and RZ60) sibling to the RILs, which were derived from the same cross manipulations (Liu et al. 1999; fig. 1) but with no evidence for Zizania introgression or genomic and phenotypic alterations relative to the parental line Matsumae (unpublished data), showed no evidence for mPing activity (fig. 6), indicating that the repeated pollination procedure (Liu et al. 1999) per se is unlikely a condition to cause mPing and Pong activation. Taken together, all evidence supports that Zizania introgression is the only possible cause for the mobilization of mPing and its putative transposase-encoding partner, Pong, in the rice RILs. Nonetheless, it presently is not clear what specific Zizania DNA segments and/or specific integration sites are responsible for the mobilization of mPing and Pong. Of the several putative Zizania-specific AFLP bands identified among the >6,000 assayed loci in the three RILs, none showed significant homology to known sequences or possess special nucleotide composition or structures that would imply a possible mechanism (unpublished data). Nevertheless, this number of analyzed loci is far from exhaustive, and it is possible that the true instigating sequence and/or integration sites remain to be identified. In any case, it is remarkable that an amount of introgression of less than 0.1% has such a dramatic impact with regard to induction of mobilization of several kinds of transposons, as based on copy number estimation, an endogenous LTR retrotransposon (Tos17) was also found transitorily activated in some RILs (Liu and Wendel 2000). That tiny amount of introgression may impose a potent effect on mobility of a host endogenous transposon has been documented earlier in Drosophila, where it was found that the amount of D. koepferae DNA introgressed into the D. buzzatii genome was not correlated to the degree of enhanced mobilization of the LTR retrotransposon Osvaldo in the hybrid lines (Labrador et al. 1999). These authors concluded that it was qualitative factors rather than quantitative loci that controlled the element transposition.
The Lack of Excision Footprints for mPing and Pong in the Rice RILs Is a Genotype-Dependent Attribute
Although it is a general characteristic of class II transposons to leave footprints upon excision, in certain cases the chromatid breaks generated by transposon excision can be efficiently repaired via gene conversion (gap repair) and, hence, leaves no footprints (Engels et al. 1990; Xu et al. 2004). In fact, it has been suggested that one possible reason for MITEs to attain their unusual high copy numbers (relative to other class II elements) is their possible excisions by a gap repair mechanism (Zhang, Arbuckle, and Wessler 2000). Being the only characterized active MITEs to date, mPing and Pong were found to predominantly leave footprints upon excision (Kikuchi et al. 2003; Nakazaki et al. 2003). Thus, it was unexpected that we found none of the 27 mPing/Pong excisions in the RILs left footprint. Two possible causes can be conceived to explain the discrepancy, i.e., difference in the elicitors (introgression vs. tissue culture or -ray irradiation) and difference in host genotypes. To distinguish the two possibilities and to further confirm the unusual phenomenon, we cloned and sequenced 41 independent empty donor sites of mPing/Pong in regenerated plants of Matsumae, and we found, again, none to leave footprints. This strongly suggests that leaving footprints or not by mPing/Pong excisions in rice is not dependent on the different elicitors; rather, it appears largely determined by the host genotype. It has been reported that both direct ligation and gap repair mechanisms may be involved in chromatid repair after transposon excision (Arca et al. 1997). Thus, the genotypic difference in mPing/Pong transpositional behavior (leaving footprints or not) may reflect the difference in the relative prevalence or titration of these two types of repair systems among rice cultivars. That, depending on the cultivars, the same MITE transposon may or may not leave footprints upon excision has apparent bearing on judging current activity-stability of the transposons (Zhang, Arbuckle, and Wessler 2000). To our knowledge, the present finding represents the first unequivocal evidence that a MITE transposon may indeed actively transpose without leaving any footprints.
Possible Causes for mPing/Pong Mobilization in Wide Hybridization-Derived RILs and Its Implications in the Context of Plant Genome Evolution and Breeding
The mechanism for mPing MITE mobilization in the rice genome as a result of tissue culture, -ray irradiation and introgression remains mysterious. An apparent effect of plant tissue culture (complete dedifferentiation during callus induction and maintenance) is the breakdown of normal controls on intrinsic chromatin state (Phillips, Kaeppler, and Olhoft 1994; S. M. Kaeppler, H. F. Kaeppler, and Rhee 2000) and, hence, frequently resulting in an array of genetic and epigenetic alterations including activation of dormant transposons (Grandibastien 1998; Okamoto and Hirochika 2001). It is conceivable that, similar to the situation of tissue culture, wide hybridization and/or subsequent introgression may also cause disturbance of the repressive chromatin state in the hybrid or introgression line, causing malfunction of cellular regulatory mechanisms on transposon activity (reviewed in Lozovskaya, Hartl, and Petrov 1995; Capy et al. 2000; Comai et al. 2003). This has been well documented in several systems of hybrid dysgenesis in Drosophila (Petrov et al. 1995; Evgen'ev et al. 1997) and in a mammalian F1 hybrid between two wallaby species (R. J. W. O'Neill, M. J. O'Neill, and Graves 1998). In plants, transcriptional activation of silent transposons has been documented in several wheat interspecific F1 hybrids or their derived allopolyploids (Kashkush, Feldman, and Levy 2002, 2003), and transcriptional instability of a transposon-related element was also detected in newly synthesized Arabidopsis hybrid plants (Comai et al. 2000). Thus, it is likely that the mobilization of mPing and Pong in the rice RILs also resulted from introgression-induced malfunction of normal cellular control systems in the rice genome.
The observation that all studied independent regenerants of Matsumae from calli subcultured for different periods (from 3 to 12 months) showed near-perfect conservation in both the Southern blot patterns and the analyzed excisions is surprising. This implies that the mobilizations of mPing and Pong were transitory and followed by rapid and complete repression while still at the callus stage. In this regard, the remarkable homogeneity of mPing and Pong hybridization patterns among random individual plants within a given line of all three RILs studied suggests that element activity in these lines was also ephemeral and followed by rapid and complete repression. It is notable that given the lack of marked element copy number elevation, the silencing mechanism for mPing and Pong activity in these lines may be different from that proposed for the LTR retrotransposons, which is regulated at the transcriptional level and mainly triggered by significant increase in element copy number (Hirochika, Okamoto, and Kakutani 2000; Nakayashiki et al. 2001). Instead, it is likely that mPing and Pong immobilization in these rice lines is accomplished by decreasing or abolishing accumulation of the required transposase encoded by a partner element, like Pong, at posttranscriptional and/or translational levels, as some other class II elements (Okamoto and Hirochika 2001).
Another result of note was the differential response between the RILs and their rice parental cultivar Matsumae with regard to mobility of mPing and Pong by tissue culture: whereas there was a marked level of mobilization of both elements in regenerants of Matsumae as judged by both the Southern blotting patterns and PCR-based locus assay for excision and insertions, there was no activity of either element in regenerants of the RILs (see Results). Previous studies have shown that a similar sharp difference in mPing and Pong activity (mobilization vs. complete stability) under tissue culture conditions exists in two rice cultivars, Nipponbare and C5924, respectively, representing the two rice subspecies, japonica and indica (Jiang et al. 2003). Although the underlying genetic and molecular basis of the differential response is currently unknown, this result may suggest a high degree of genetic divergence between the RILs and their rice parental cultivar Matsumae, as was indeed revealed by the genome-wide AFLP analysis (Y. Wang, Z. Dong, Z. Zhang, Y. Shen, and B. Liu, in preparation).
Given the prevalence of hybridization and introgression in natural plant populations (Anderson and Stebbins 1954; Stebbins 1959; Rieseberg 1995; Wendel 2000; Rieseberg et al. 2003; Arnold 2004), our findings on transposon mobilization induced by introgression bear significant implications for genomic and organismal evolution in plants. It is increasingly clear that TEs are particularly abundant in plant genomes and have played a significant role in the host genome evolution. The findings of this paper have provided circumstantial evidence that the role of transposons in plant genome evolution can be facilitated by hybridization and introgression. In this respect, owning to their often-intimate association with low-copy genic regions in a plant genome (Bureau and Wessler 1992, 1994a, 1994b; Wessler, Bureau, and White 1995; Zhang, Arbuckle, and Wessler 2000; Jiang et al. 2003), the mobilization of MITEs by introgression may be particularly noteworthy because both the excisions and insertions may potentially affect gene expression and, hence, might have phenotypic consequences, as was clearly shown for a -ray mobilized mPing copy that inserted into the rice homolog of the CONSTANS gene and caused quantitative changes in flowering time (Nakazaki et al. 2003).
Another potential implication of our results is in the context of plant breeding involving wide hybridization and introgression. Although it is likely that not all wide hybridization and introgression would cause activation of transposons, based on findings of the present paper, an important consideration in evaluating plant lines derived from hybridization is that, apart from transfer of the desired genes or traits from the donor species, these lines may contain additional genomic variations including those induced by mobilized transposons. The findings of this paper on transposon mobility, together with our earlier demonstration on extensive alterations in DNA methylation pattern in these rice lines (Liu et al. 2004), clearly indicate that hybridization and introgression have a broader effect than hitherto recognized. In this respect, we note that the rice RILs have expressed multiple phenotypic novelties, including changes in the overall morphology (fig. 2A), flowering time, yield component traits and disease resistance (unpublished data), that probably far exceed the scope accountable by the trace amount of introgression from Z. latifolia. We are actively investigating whether any of the phenotypic variations in these lines are caused by mobilization of transposons like mPing and Pong.
|
Supplementary Material |
|---|
|
TOPAbstractIntroductionMaterials and MethodsResultsDiscussionSupplementary MaterialAcknowledgementsReferences |
|---|
Supplementary tables are available at Molecular Biology and Evolution online www.molbiolevol.org).
|
Acknowledgements |
|---|
|
TOPAbstractIntroductionMaterials and MethodsResultsDiscussionSupplementary MaterialAcknowledgementsReferences |
|---|
This study was supported by the National Science Award for Distinguished Young Investigators in China (30225003) and the National Natural Science Foundation of China (30430060). We are grateful to Moshe Feldman and Keith Adams, as well as two anonymous reviewers, for their critical comments and valuable suggestions to improve the manuscript. This paper is dedicated to the memory of the late Prof. Hengmao Piao, who pioneered the hybridization between rice and Zizania latifolia Griseb.
|
Footnotes |
|---|
1 Equal contribution to this work.
|
References |
|---|
|
TOPAbstractIntroductionMaterials and MethodsResultsDiscussionSupplementary MaterialAcknowledgementsReferences |
|---|
Anderson, E., and G. L. Stebbins Jr. 1954. Hybridization as an evolutionary stimulus. Evolution 8:378–388.
Arca, B., S. Zabalou, T. G. Loukeris, and C. Savakis. 1997. Mobilization of a Minos transposon in Drosophila melanogaster chromosomes and chromatid repair by heteroduplex formation. Genetics 145:267–279.[Abstract]
Arnold, M. L. 2004. Transfer and origin of adaptations through natural hybridization: were Anderson and Stebbins right? Plant Cell 16:562–570.
Banks, J. A., P. Masson, and N. Fedoroff. 1988. Molecular mechanisms in the developmental regulation of the maize Suppressor-mutator transposable element. Genes Dev. 2:1364–1380.
Bennetzen, J. L. 2000. Transposable element contributions to plant gene and genome evolution. Plant Mol Biol. 42:251–269.[CrossRef][ISI][Medline]
Bureau, T. E., P. C. Ronald, and S. R. Wessler. 1996. A computer-based systematic survey reveals the predominance of small inverted-repeat elements in wild-type rice genes. Proc. Natl. Acad. Sci. USA 93:8524–8529.
Bureau, T. E., and S. R. Wessler. 1992. Tourist: a large family of small inverted repeat elements frequently associated with maize genes. Plant Cell 4:1283