Abstract
Alternative transposition can induce genome rearrangements, including deletions, inverted duplications, inversions, and translocations. To investigate the types and frequency of the rearrangements elicited by a pair of reversed Ac/Ds termini, we isolated and analyzed 100 new mutant alleles derived from two parental alleles that both contain an intact Ac and a fractured Ac (fAc) structure at the maize p1 locus. Mutants were characterized by PCR and sequencing; the results show that nearly 90% (89/100) of the mutant alleles represent structural rearrangements including deletions, inversions, translocations, or rearrangement of the intertransposon sequence (ITS). Among 37 deletions obtained, 20 extend into the external flanking sequences, while 17 delete portions of the intertransposon sequence. Interestingly, one deletion allele that contains only a single nucleotide between the retained Ac and fAc termini is not competent for further alternative transposition events. We propose a new model for the formation of intertransposon deletions through insertion of reversed transposon termini into sister-chromatid sequences. These results document the types and frequencies of genome rearrangements induced by alternative transposition of reversed Ac/Ds termini in maize.
TRANSPOSABLE elements are the major component of many eukaryotic genomes, where they play important roles in plant gene evolution and genome reorganization (Bennetzenet al. 2005; Dooner and Weil 2007). In maize, nearly 85% of the genome is composed of mobile DNA and other repeated sequences (Schnableet al. 2009). Transposable elements are broadly classified as either retroelements (class I) or DNA transposons (class II) (Wickeret al. 2007). Retroelement transposition occurs via a “copy and paste” mechanism, which can lead to massive amplification and is thought to be a major contributor to genome size differences. In contrast, most DNA elements move by a “cut and paste” mechanism. Although DNA elements are less abundant compared with retroelements, they can lead to a variety of genome structural changes (Rosset al. 1979; Gray 2000; Watanabeet al. 2007; Huang and Dooner 2008) including large chromosomal rearrangements, which may contribute to genome evolution (McClintock 1951; McClintock 1978; Listeret al. 1993; Zhanget al. 2009).
The well-known Ac/Ds elements in maize were initially identified by Barbara McClintock through their ability to induce chromosome breakage. In addition, McClintock identified a number of major chromosome rearrangements that were apparently induced by the Ac/Ds system (McClintock 1951; McClintock 1978). Subsequent genomic cloning and sequence analysis suggested that the original chromosome-breaking Ds element has a “double Ds” structure, i.e., one Ds element inserted into a second identical Ds, in the opposite orientation (Doringet al. 1984; Kleinet al. 1988; Martinez-Ferez and Dooner 1997). Chromosome breakage is thought to occur when the Ac transposase attempts to transpose a pair of directly oriented Ds 5′ and 3′ termini located on different sister chromatids (Englishet al. 1993). This reaction, termed sister-chromatid transposition, not only breaks chromosomes but can also generate deficiencies and inverted duplications (Zhang and Peterson 1999). A second type of alternative transposition reaction involves reverse-oriented 5′ and 3′ Ac/Ds termini; for example, two directly oriented Ac or Ds elements will have their apposed 5′ and 3′ termini in reversed orientation. Previously our lab isolated and characterized the P1-rr11 allele, which contains a full-length Ac element and a terminally deleted Ac (fAc) inserted in the maize p1 locus. The 5′ end of Ac and the 3′ end of fAc are in reversed orientation, separated by a ∼13-kb intertransposon segment (ITS). In this configuration, reversed-ends transposition can induce a variety of chromosome rearrangements, including deletions, inversions, translocations, and ITS rearrangements (Zhang and Peterson 2004; Zhanget al. 2009). In addition, this reversed-ends configuration can induce chromosome breakage (Huang and Dooner 2008; Yuet al. 2010).
In this article, we investigated the types and frequency of genome rearrangements produced by reversed-ends transposition at the p1 locus. The results show a high relative frequency of deletions and inversions, in agreement with previous reports (Zhang and Peterson 2004; Huang and Dooner 2008). In addition, we identified a substantial class (17%) of previously uncharacterized new mutant alleles, which contain deletions of various segments of the DNA between the reversed Ac/Ds termini. On the basis of structures of these deletions and the known propensity of Ac/Ds elements to transpose during DNA replication (Greenblatt and Brink 1962; Greenblatt 1968, 1974, 1984), we propose a model for their formation. Furthermore, we investigated the frequency of chromosome breakage induced by the resulting alleles that contain various lengths of DNA separating the reversed Ac/Ds termini. Interestingly, the chromosome-breakage frequency (and hence the alternative transposition frequency) declines precipitously for alleles in which the reversed Ac/Ds termini are in very close proximity (91 and 1 bp).
MATERIALS AND METHODS
Maize stocks and crosses:
The maize alleles P1-rr11, P1-rr910, and P1-ovov454 were isolated and described previously (Zhang and Peterson 2004; Zhanget al. 2009; Yuet al. 2010). Plants in which the P1-rr11, P1-rr910, or P1-ovov454 alleles were heterozygous with an allele for colorless kernel pericarp (p1-ww or p1-wr) were crossed by pollen from plants of genotype p1, r1-m3::Ds. The r1-m3::Ds allele contains a Ds element inserted in the r1 gene required for kernel aleurone pigmentation; Ac-induced excision of Ds from r1-m3::Ds results in purple aleurone sectors (Kermicle 1980). The mature ears were screened for kernels with colorless pericarp and spotted aleurone, as these were predicted to have undergone a loss of p1 function, but to have retained Ac activity. Selected kernels were sown, and pollen from mature plants was checked using a hand microscope to estimate pollen abortion frequency. Plants were self-pollinated to make the new mutant alleles homozygous for molecular analyses.
DNA extraction and PCR analyses:
Young leaves of individual plants were ground in liquid nitrogen, and genomic DNA was extracted with CTAB (cetyltrimethylammonium bromide) reagent (Saghai-Maroofet al. 1984). HotMaster Taq polymerase from Eppendorf (Hamburg, Germany) was used in the PCR reaction. The PCR mix was heated at 94° for 3 min to denature the DNA template, followed by 35 cycles of 20 sec at 94°, 30 sec at 60°, and 1 kb/1 min at 65°, and one cycle of 8 min at 65°. Primers used for PCR anlayses are listed in Table 1. Candidate ITS rearrangement alleles were analyzed using the following primer pairs: P1-2915r and P1-15588f (for alleles derived from P1-rr11) or P1-7061r and P1-15588f (for alleles derived from P1-rr910). Candidate internal deletions flanking the Ac 5′ terminus were analyzed using the following primers: Ac120r plus one of Ac4436f, P1-12046r, or P1-7061r (for alleles derived from P1-rr11); and Ac120r plus one of Ac4436f or P1-12046r (for alleles derived from P1-rr910). Candidate internal deletions flanking the fAc 3′ terminus were analyzed using the following primers: Ac4436f plus one of Ac120r, 9D9A5537f, P1-7869f, or 9D9A33680f (for alleles derived from P1-rr11); and Ac4436f plus one of Ac120r, P1-7869f, or 9D9A33680f (for alleles derived from P1-rr910). PCR products were excised from agarose gels, purified using the Pefectprep gel cleanup system (Eppendorf HotMaster Taq, Westbury, NY), and sequenced by the DNA Synthesis and Sequencing Facility, Iowa State University. Sequence of P1-ovov454 is available in GenBank (accession number GU595146).
PCR primers and their sequences used in this study
Evaluation of chromosome-breakage frequency:
The maize Dek1 gene is required for differentiation of kernel aleurone cells; endosperm cells lacking Dek1 function fail to differentiate into aleurone cells and thus cannot synthesize anthocyanin pigments (Lidet al. 2002). Wild-type Dek1 C1 R1 were crossed as ear parents by pollen from plants homozygous for the weak dek1-Dooner allele (Becraftet al. 2002) to generate a stock of heterozygous dek1-Dooner/Dek1 tester plants. The dek1/Dek1 plants were then crossed as ear parents by pollen from plants homozygous for the candidate p1 alleles to be tested. Kernels of mature ears were examined and the numbers of colorless sectors were used as an index to the chromosome-breakage frequency. The chromosome-breakage frequency grading is evaluated according to Yuet al. (2010).
RESULTS
Inversion and deletion are the major products of reversed-ends transposition:
The P1-rr11 and P1-rr910 alleles condition predominantly red kernel pericarp pigmentation with white (colorless) sectors of variable size (Zhang and Peterson 2004; Zhanget al. 2009). Both alleles contain an identical fAc element in the second intron of the p1 gene and a full-length Ac element inserted upsteam of the p1 transcription start site. The distances between the Ac 5′ terminus and fAc 3′ terminus in P1-rr11 and P1-rr910 are approximately 13 and 9 kb, respectively. In both alleles, p1 exons 1 and 2 are located between the Ac/fAc termini and are predicted to be deleted by reversed-ends transposition (Zhang and Peterson 2004; Zhanget al. 2009). Therefore, we reasoned that new alleles arising from reversed-ends transposition could be selected on the basis of loss of p1 function (colorless kernel pericarp) and retention of Ac activity (purple aleurone sectors) (materials and methods). In this way we selected and characterized a total of 100 putative mutant alleles (42 and 58 derived from P1-rr11 and P1-rr910, respectively).
To determine what types of structural changes had occurred in the mutant alleles, we performed PCR using four pairs of primers specific for the four junctions of Ac/fAc and the flanking genomic DNA: (1) the Ac 3′ terminus; (2) the Ac 5′ terminus; (3) the fAc 3′ terminus; and (4) the fAc 5′-end junction (Figure 1A and Table 2). Representative gel analysis results are shown in Figure 1B, and a summary of the data are shown in Table 3.
PCR analyses of rearrangement alleles. (A) Schematic structure of the parental alleles, showing the locations of the four PCR primer pairs. (B) PCR results from 19 candidate rearrangement alleles (samples 1–19) and parental allele P1-rr11 (sample 20). See Tables 1 and 2 for primer names and sequences.
Combinations of PCR primers used to determine allele structures
PCR patterns and inferred structures
The largest class (52%) was composed of alleles in which PCR primer pairs 1 and 4 were both positive while pairs 2 and 3 were negative (pattern + − − +). This pattern could be produced by one of several different types of rearrangements including translocations, inversions, and ITS rearrangements (classes a, b, and c).
a. Translocations and large inversions (five cases). Chromosome translocations and very large inversions would be expected to exhibit a higher than normal percentage of pollen abortion (50% for translocations, and up to 50% for inversions depending on size). By screening for pollen abortion followed by further molecular characterization we identified four translocations (p1-wwC15, p1-wwB47, p1-wwB33, p1-wwB1023) and one pericentric inversion (p1-wwC30). These cases have been described elsewhere (Zhanget al. 2009).
b. Small inversions (46 cases). Alleles with a + − − + pattern and normal pollen abortion frequency were classified as putative inversions of less than 10 cM in size. Larger inversions would be expected to exhibit significant levels of pollen sterility resulting from crossovers within the inverted segment. Among these 46 alleles, the breakpoint sequences of 6 representative cases were cloned by the “Ac casting” method (Singhet al. 2003; Zhanget al. 2009). The p1-wwB53 and p1-wwB58 alleles contain inversions of less than 10 kb, with breakpoints located in repeat sequences flanking the p1 3′ region. The p1-wwB9, p1-wwC14, and p1-wwC4 alleles contain inversions with breakpoints located 138, 142, and 370 kb distal to p1, respectively. Finally, the p1-wwC10 allele contains an ∼1-Mb inversion proximal to p1.
c. ITS rearrangements (1 case). ITS rearrangements are defined as those cases in which the ITS have been circularly permuted; i.e., the sequences between the transposon termini have remained, but they have changed their relative position and, in some cases, their orientation. These permutations result from reversed-ends transposition followed by insertion into the ITS of the same chromatid (Zhang and Peterson 2004; Huang and Dooner 2008). Candidate ITS rearrangement alleles were subject to one additional PCR test designed to detect the junction formed by excision of the Ac/fAc termini (materials and methods). Only one case (derived from P1-rr11) was confirmed as an ITS rearrangement.
d. External deletion (20 cases). Alleles in which only PCR primer pair 1 or pair 4 gave a positive result (pattern − − − + or + − − −) were classified as probable p1 left- or right-side flanking deletions. For example, one allele (p1-wwB4) has a deletion which removed the ITS region, the 2.0-kb fAc element plus 105 bp of the flanking p1 gene sequence.
e. Internal deletions (17 cases). The PCR pattern of this class is + − + + or + + − +. These alleles are described in detail below.
f. No Ac/fAc structure (7 cases). This class produced no positive reaction with any of the four primer pairs (− − − − pattern). These may represent segregation of the p1-ww or p1-wr allele together with a transposed Ac. They were not investigated further.
g. Ac excision (two cases). Two alleles (one each from P1-rr11 and P1-rr910) gave − − + + PCR patterns, which could be obtained by excision of Ac from the parental alleles. However, a simple transposition would not elicit a colorless pericarp phenotype; possibly, excision was accompanied by a deletion of essential p1 sequences. These cases were not characterized further.
h. Unknown (two cases). Two cases show PCR patterns (+ + − − and + + + −) that are not expected from any known single standard or alternative transposition event. The origin of these cases is unknown and is not considered further here.
Previous research has identified examples of “fused ends”; i.e., cases in which the ITS is deleted and the 5′ and 3′ Ac/Ds termini are ligated together (Gorbunova and Levy 1997; Krishnaswamyet al. 2008). The presence of fused ends among the mutant alleles tested here could be detected by PCR with Ac 5′ end primer (Ac120r) and Ac 3′ end primer (Ac4436f) (Table 2). Three alleles (p1-wwB42, p1-wwB54, p1-wwB59) from P1-rr11 showed a single strong band, but subsequent sequence analysis showed that each of these cases was in reality an internal deletion (+ + − + pattern). In summary none of the alleles we identified here contains fused Ac/fAc ends.
Sequence analysis of the flanking Ac 5′-end or fAc 3′-end deletion lines:
The breakpoints of 12 of the other 14 internal deletion alleles were isolated by PCR and sequencing (materials and methods). DNA sequence results show that in each case, the Ac/fAc structure is intact, and the deletions are confined to the intertransposon sequences flanking either Ac or fAc (Figure 2). Each deletion extends precisely from either the Ac 5′ end or the fAc 3′ end to a site in the intervening DNA. None of the alleles has a deletion of sequences flanking both the Ac and fAc termini. Because the deletions were selected on the basis of colorless kernel pericarp (i.e., lack of p1 function), it is expected that each deletion should remove some sequences essential for p1 expression. Indeed, all deletions but one (p1-wwB3) have removed part or all of p1 gene exons 1 and 2, which encode the Myb-homologous DNA binding domain of the p1 transcriptional activator protein. The p1-wwB3 allele removes a ∼2.5-kb region of the distal p1 promoter region. Approximately half of the deletions have endpoints within a 1-kb region including p1 exons 1 and 2; this region is a hot spot for Ac insertion as previously reported (Athmaet al. 1992). One allele (p1-wwB42) has only a single nucleotide (C) remaining between the Ac and fAc termini; it is not possible to determine from which side (Ac or fAc) the deletion originated, because the same nucleotide is present flanking both the Ac and fAc termini in the progenitor P1-rr11 allele.
Structures of progenitor and ITS deletion alleles. Schematic structures of P1-rr11 (A), P1-rr910 (B), and P1-ovov454 (C) are shown. The solid black boxes indicate p1 gene exons 1, 2, and 3 (left to right). The open and solid red triangles represent the 3′ and 5′ termini of Ac/fAc, respectively. Short horizontal arrows indicate PCR primers. Extent of ITS deletions (gapped lines) in alleles derived from P1-rr11 and P1-rr910 are shown in A and B, respectively.
Deletions flanking the external Ac 3′ end occur at much lower frequency than intertransposon events:
The results described above show that deletions flanking the Ac 5′ end and fAc 3′ end occur at a significant frequency. We wondered if deletions flanking the Ac 3′ end can also occur at similar frequency (Figure 2). However, we could not readily detect deletions from this terminus in the P1-rr11 and P1-rr910 alleles because such deletions would not remove sequences essential for p1 function. Therefore we screened for deletions using a third allele termed P1-ovov454, which conditions orange variegated pericarp and cob, i.e., orange tissue with red and colorless sectors. In the P1-ovov454 allele, both Ac and fAc are located in the p1 gene intron 2 separated by a distance of approximately 0.8 kb (Figure 2) (Zhanget al. 2009). Importantly, deletions extending from the Ac 3′ end into the upstream flanking sequence could remove p1 exons 1 and or 2, both of which are essential for p1 function. To screen for such deletions, we selected 133 kernels with colorless pericarp and spotted aleurone from P1-ovov454 heterozygotes. These kernels should carry a nonfunctional p1 gene and retain Ac activity. Plants grown from these kernels were screened by PCR using primers Ac4436f and Ac120r to identify those plants that retained the Ac/fAc structure of the progenitor allele. Only 3 of the 133 plants tested positive for this band. The other 130 plants most likely contain deletions of the ITS as a consequence of reversed-ends transposition involving the Ac 5′ end and fAc 3′ end. The three plants that retained the Ac/fAc structure were then tested using an Ac 3′ end primer (Ac4436f) together with a flanking sequence primer (PA-A11) to detect possible one-sided deletions. Two of the plants were positive for this product; i.e., they retained an intact Ac 3′ junction and hence did not contain flanking deletions. One of the three tested plants was negative for the Ac 3′ flanking sequence product, and therefore likely represents an Ac 3′ flanking deletion. These results indicate that one-sided deletions flanking the Ac 3′ end occur less frequently (1/133) than intertransposon deletions flanking the fAc 3′ end (8/100).
Reversed Ds ends in very close proximity exhibit reduced alternative transposition in maize:
Previously we reported that alleles containing pairs of reverse-oriented Ac/Ds ends can cause chromosome breakage, and chromosome-breakage frequency is inversely proportional to the distance between the element termini (Yuet al. 2010). The previous study examined pairs of Ac/Ds elements separated by distances ranging from 0.8 to 13 kb. To examine the effects of shorter intertransposon distances on chromosome breakage, we tested the intertransposon deletion alleles described here for the frequency of loss of a distal visible marker gene (Dek1; materials and methods). Most of the intertransposon deletion alleles derived from P1-rr11 or P1-rr910 show increased chromosome breakage as the distance between the reversed Ac/Ds ends decreases (Table 4 and Figure 3). For example, allele p1-wwB54 has an Ac/fAc separation distance of 331 bp and exhibits a relatively high chromosome-breakage frequency (grade 2.9). However, two alleles with shorter element separation distances have drastically reduced chromosome breakage: p1-wwB59 (91 bp between Ac/fAc) exhibits a relatively low chromosome-breakage frequency (grade 1.4). This frequency is similar to that of the progenitor allele P1-rr11, in which the Ac/fAc termini are separated by 13 kb. Even more striking, p1-wwB42, which has only one nucleotide separating the Ac/fAc termini, exhibits little or no detectable chromosome breakage (grade 0). To determine whether these alleles have mutation(s) in the Ac/fAc termini that would render them immobile, ∼450 bp of the 5′ Ac and 3′ fAc termini and subterminal regions were sequenced, but no changes were found. Both p1-wwB59 and p1-wwB42 alleles show normal Ac-transposase gene activity by crossing with the r1-m3::Ds tester line. We then tested whether the Ac elements in the p1-wwB59 and p1-wwB42 alleles were still capable of standard transposition. PCR analysis using specific primers flanking the Ac insertions produced bands as expected for normal Ac excision (data not shown). We conclude that reversed Ac/Ds ends in very close proximity (<100 bp) have reduced frequencies of alternative transposition.
Chromosome-breakage frequency exhibited by different alleles
Representative ears for evaluating the frequency of chromosome breakage. Ears of dek1/+ tester plants were crossed with pollen from plants homozygous for the indicated p1 alleles. The functional Dek1 gene conditions solid purple aleurone color; chromosome breakage at the proximal p1 locus results in loss of Dek1 as evidenced by colorless aleurone sectors. The ears were borne on heterozygous dek1/+ tester plants, hence only half of the kernels can show Dek1 loss. The number and size of colorless sectors indicate the chromosome-breakage frequency (Yuet al. 2010). The photographs show grade 0 to grade 3 representative ears produced by crossing with pollen from plants homozygous for the following alleles: p1-wwB42 (grade 0); p1-wwB59 (grade 1); p1-wwB8 (grade 2); p1-wwB34 (grade 3).
DISCUSSION
The aim of this study is to investigate the types and frequency of genome rearrangements induced by reversed-ends transposition at the maize p1 locus. The results indicate that about half of rearrangements are chromosome inversions, translocations, or ITS rearrangements; 20% are chromosome external deletions; and 17% are internal deletions (ITS deletion). Our previous data indicate that the inversions, translocations, external deletions, and ITS rearrangements were generated via reversed Ac end transposition (Zhang and Peterson 2004; Zhanget al. 2009), but it is not clear how the internal deletions were generated. Here, we propose a mechanism for the generation of small internal deletions. We also discuss the implications of our results for the development of transposon-based genome rearrangement tools.
Model for generation of transposon-induced flanking deletions:
We have previously presented evidence that pairs of reversed Ac/fAc termini on the same chromatid are competent for transposition (Zhang and Peterson 2004). To explain the high frequency of intertransposon segment deletions, we propose that, during reversed-ends transposition, the active Ac/fAc termini can insert into the region between Ac and fAc on the sister chromatid; this scenario would produce two daughter chromatids containing reciprocal deletions of the intertransposon segment. Each resulting chromatid would have a deletion flanking either the 5′ Ac or 3′ fAc ends (Figure 4). Insertion of the Ac/fAc termini in the opposite orientation would lead to formation of a chromatid bridge and chromosome breakage, as previously proposed (Huang and Dooner 2008; Yuet al. 2010). This model fits previous observations, including those that: (1) Ac/Ds elements exhibit a pronounced preference for local transposition; (2) the Ac element from the standard p1-vv allele often inserts into the sister chromatid (Greenblatt and Brink 1962); (3) the maize p1 gene promoter and exon 1 sequences are preferred sites for Ac insertion (Athmaet al. 1992), and this region is located within the Ac/fAc intertransposon segment in the P1-rr11 and P1-rr910 alleles.
Alternative transposition model for formation of ITS deletions. Symbols are the same as in previous figures. (1) The chromosome containing the Ac/fAc locus is partially replicated. Ac transposase (open, black circles) recognizes the reversely oriented Ac/fAc termini on the same chromatid; cleavage by transposase occurs at Ac and fAc termini. (2) Excised 5′ and 3′ termini of Ac/fAc insert into a site between Ac/fAc in the sister chromatid. (3–4) Reciprocal deletion alleles are generated following completion of DNA replication.
There are at least three other existing models for the formation of transposon-flanking deletions. For the bacterial Tn5 transposon, flanking deletions were proposed to occur by abortive transposition of a single transposon end (Jilket al. 1993). If this model were correct for the maize Ac/Ds elements, then the frequency of “one-sided” deletions flanking the Ac 3′ end should be equal to or greater than that of deletions flanking the fAc 3′ end. This is based on the premise that transposition of the Ac 3′ end should occur at least as frequently as that of the fAc 3′ end, because (1) the Ac 3′ end is present in the context of a complete Ac element, whereas the fAc 3′ end is present on a smaller (2059 bp) Ac fragment, and (2) the fAc 3′ end is located farther away from the Ac 5′ end (9 and 13 kb in P1-rr910 and P1-rr11, respectively) than is the 3′ end of an intact Ac element (4565 bp). However, we detected significantly more “one-sided” deletions flanking the fAc 3′ end (total 8/100 from P1-rr910 and P1-rr11) than the Ac 3′ end (1/133 from P1-ovov454) (corrected P = 0.0056; odds ratio = 11.38 with a 95% confidence interval [1.48, 511.80] by the Fisher exact test). These results do not fit the predictions of the single-end abortive transposition model.
An alternative model (Pageet al. 2004) proposed that deletions flanking Ds elements in Arabidopsis result from a hybrid transposition mechanism involving two sequential steps: First, a newly replicated Ds element inserts into a nearby unreplicated region and forms a pair of Ds elements in reversed orientation. Second, the 5′ end of one Ds and the 3′ end of the other Ds excise, leading to deletion of the internal sequence. The chromosome arm is repaired by ligation of the two broken ends. This model is not consistent with the deletions we isolated for at least three reasons: (1) The fAc in the P1-rr11 and P1-rr910 alleles is immobile and hence cannot participate in the first transposition reaction of the Page et al. model (2) the deletion junctions we isolated are precise and exhibit no evidence of the involvement of NHEJ-type repair, and (3) this model does not explain the difference in frequency of deletions flanking the external Ac 3′ side vs. the intertransposon fAc 3′ end.
In Drosophila, the P element can undergo a type of alternative transposition involving the termini of P elements located on homologous chromosomes (hybrid element insertion model) (Prestonet al. 1996; Parkset al. 2004). If such a hybrid element inserted into a nearby site, a “one-sided” flanking deletion could be generated (Figure 5). However, no evidence for a similar hybrid element mechanism was detected in two different experiments in maize (Yuet al. 2010). Previous work has shown that pairs of Ac/Ds elements on the same chromosome can induce chromosome breakage by an alternative transposition mechanism (Weil and Wessler 1993; Huang and Dooner 2008; Yuet al. 2010). When the distance between the two interacting Ac/Ds/fAc termini is increased to >200 kb, the chromosome-breakage frequency is highly reduced (Huang and Dooner 2008). Therefore, even if alternative transposition reactions involving hybrid elements were to occur, the frequency is expected to be too low to explain the relatively high frequency of intertransposon segment deletions observed for the P1-rr11 and P1-rr910 alleles.
P-element-induced flanking deletion model (modified from Parkset al. 2004). (1) Transposase (open, black circles) recognizes and binds the termini of two different P elements located on homologous chromosomes. (2) P element termini are cleaved and inserted into a flanking target site. (3) Structure of the deletion chromosome formed.
Generation of inversions vs. deletions by reversed-ends transposition:
The Ac/Ds transposon system is well known for having a preference for transposition to closely linked sites (Athmaet al. 1992; Vollbrechtet al. 2010). In contrast, the orientation of insertion of the Ds termini appears to be random in several large collections produced for gene tagging (Kolesniket al. 2004; Nishalet al. 2005; Panet al. 2005). If the insertion orientation of reversed-ends transposition is also random, then one would expect that inversions and deletions would be generated at a similar frequency. However, in the experiments described here, inversions were recovered at a twofold greater frequency than deletions. This difference may be due in part to the fact that deletions that extend from the 3′ fAc upstream beyond the Ac element would not have been included in our analysis due to the initial selection for spotted kernels, i.e., kernels that retain Ac activity. Moreover, very large deletions are known to have reduced transmission frequency (Birchler and Levin 1991; Linet al. 1997), presumably due to the loss of essential genes, whereas inversions have a normal gene balance and should transmit normally. Consistent with this point, a higher frequency of inversions than deletions was reported at the maize Bz locus (Huang and Dooner 2008).
Reversed Ds ends distance and chromosome rearrangement tool development:
This and earlier reports indicate that reversed Ac/Ds termini can produce chromosome rearrangements such as inversions, deletions, and translocations (Zhang and Peterson 2004; Huang and Dooner 2008; Zhanget al. 2009). The ability to generate chromosome rearrangements may be valuable for plant chromosome engineering and/or functional genomics analysis, and it will be important to determine the parameters affecting the frequency of alternative transposition events. Previous reports have shown that chromosome-breakage frequency is inversely related to the distance between the interacting Ac/Ds termini (Dooner and Belachew 1991; Yuet al. 2010). Here, we have shown that alleles in which the Ac/fAc termini are separated by 0.3 kb to 3 kb segments have the highest chromosome-breakage frequency (Table 4). Strikingly, alleles in which the Ac/fAc termini are separated by less than 100 bp exhibit much reduced chromosome-breakage frequency (Table 4, alleles p1-wwB59 and p1-wwB42). Chromosome breakage in these alleles can be fully explained as a consequence of reversed Ac/Ds ends transposition, and hence should be a good indicator of alternative transposition frequency. Interestingly, somatic extrachromosomal circular Ds molecules were previously reported in maize and transgenic tobacco tissues; these circular Ds elements appeared to be incapable of transposition and were considered likely to be abortive transposition products (Gorbunova and Levy 1997; Gorbunova and Levy 2000). Consistent with this idea is the fact that the p1-wwB42 allele, which has only a single nucleotide separating the Ac and fAc termini, exhibits little or no detectable chromosome breakage. Together these observations suggest that structures in which the Ac/Ds/fAc termini are separated by less than 0.1 kb are not competent for transposition. While other factors such as epigenetic modifications and local sequence context are known to affect transposition frequency, our results indicate that the spacing of Ac/Ds transposon termini is also important.
Acknowledgments
We thank Amber Newman, Dafang Wang, Pooja Gupta, Cierra Pairett, Tim Bruihler, Lisa Coffey and Peter Howe for laboratory and field assistance. This research was supported by the National Science Foundation (MCB 0450243) to T.P. and J.Z.
Footnotes
Communicating editor: G. P. Copenhaver
- Received January 13, 2011.
- Accepted February 2, 2011.
- Copyright © 2011 by the Genetics Society of America
