Genetics, Vol. 150, 1245-1256, November 1998, Copyright © 1998

Extreme Structural Heterogeneity Among the Members of a Maize Retrotransposon Family

Sylvestre Marillonnet1,a and Susan R. Wesslera
a Departments of Botany and Genetics, University of Georgia, Athens, Georgia 30602

Corresponding author: Susan R. Wessler, Department of Genetics, University of Georgia, Athens, GA 30602., sue{at}dogwood.botany.uga.edu (E-mail).

Communicating editor: V. SUNDARESAN


*  ABSTRACT
*TOP
 *ABSTRACT
*MATERIALS AND METHODS
 *RESULTS
 *DISCUSSION
 *LITERATURE CITED

A few families of retrotransposons characterized by the presence of long terminal repeats (LTRs) have amplified relatively recently in maize and account for >50% of the genome. Surprisingly, none of these elements have been shown to cause a single mutation. In contrast, most of the retrotransposon-induced mutations isolated in maize are caused by the insertion of elements that are present in the genome at 2–50 copies. To begin to understand what limits the amplification of this mutagenic class of LTR-retrotransposons, we are focusing on five elements previously identified among 17 mutations of the maize waxy gene. One of these elements, Stonor, has sustained a deletion of the entire gag region and part of the protease domain. Missing sequences were recovered from larger members of the Stonor family and indicate that the deletion probably occurred during retrotransposition. These large elements have an exceptionally long leader of 2 kb that includes a highly variable region of ~1 kb that has not been seen in previously characterized retrotransposons. This region serves to distinguish each member of the Stonor family and indicates that no single element has yet evolved that can attain the very high copy numbers characteristic of other element families in maize.

RETROTRANSPOSONS belong to a class of elements distinguished by their mode of transposition through an RNA intermediate. Two types of retrotransposons have been identified in eukaryotes: the retrotransposons characterized by the presence of long terminal repeats (LTRs) and the long interspersed nuclear elements, which do not contain LTRs. LTR-retrotransposons appear to be ubiquitously distributed in plants (FLAVELL et al. 1992 Down; VOYTAS et al. 1992 Down; HIROCHIKA and HIROCHIKA 1993 Down), where they represent the most abundant class of transposable elements. It has been estimated that 50% of the maize genome consists of retrotransposon sequences, with five large families contributing up to 25% of the genomic DNA (SAN-MIGUEL et al. 1996 Down). Not all retrotransposon families, however, are present at high copy numbers. Hundreds or perhaps thousands of families containing <100 elements may also be present in maize (SANMIGUEL et al. 1996 Down).

Interestingly, the elements found to cause mutations in maize are those that are present in the genome at relatively low copy numbers. These include Magellan (four to eight copies per haploid genome, PURUGGANAN and WESSLER 1994 Down), Hopscotch (two to six copies, WHITE et al. 1994 Down; S. E. WHITE and S. R. WESSLER, unpublished data), and Bs1 (one to five copies, JOHNS et al. 1985 Down). Members of the B5 family, with only two to four copies, are responsible for three independent mutations (wxB5, wxG, and bm-3; VARAGONA et al. 1992 Down; VIGNOLS et al. 1995 Down), while Magellan elements have been found in two mutant alleles (wxM and pl-987, PURUGGANAN and WESSLER 1994 Down; P. L. COOPER and K. C. CONE, personal communication). No member of any of the larger families of maize elements, such as Opie (30,000 copies), or Grande, Ji, or Huck (each with >10,000 copies), has caused any of the characterized maize mutations, despite the fact that some appear to be largely intact and capable of further retrotransposition (SANMIGUEL et al. 1996 Down). These data led SANMIGUEL et al. 1996 Down to suggest that there may be a cause-and-effect relationship between element family size and the propensity to insert into genes. That is, elements from very large families may thrive because they display a target site preference for other retrotransposons. This way, they avoid mutating maize genes. Whether this trend holds true for other plant species cannot be ascertained at this time because few spontaneous mutations outside of maize have been characterized. However, it was previously demonstrated that the retrotransposon Tos17, which is present at only one to two copies in the rice genome, prefers insertion into gene sequences after its induction into cell culture (HIROCHIKA et al. 1995 Down).

Just as the avoidance of gene targets may favor element amplification, an inability to avoid genes may prevent significant amplification of the smaller element families. To begin to understand this apparent correlation between the size of plant retrotransposon families and their propensity to insert into genes, we have characterized the Stonor family of maize. The Stonor element was first identified as an insertion into the intron5/exon6 junction of the maize wxStonor allele (Figure 1; VARAGONA et al. 1992 Down). In this study, we have sequenced the complete Stonor insertion and found that it is a defective element containing a deletion of part of its coding sequence. This observation prompted us to clone larger elements from the same family to identify sequences missing from Stonor. Analysis of the sequences at the deletion endpoints and in the remainder of the element indicates that Stonor is derived from a larger, presumably active element that sustained a deletion during retrotransposition. Characterization of several family members larger than Stonor showed that they differ from most other retrotransposons by the presence of an unusually long, untranslated leader region that contains a domain of highly variable sequence.



View larger version (17K):
In this window
In a new window
Download PPT slide
  		Figure 1. Structure of the wxStonor allele, including the Stonor element. Exons (open boxes) and introns (connecting lines) of the wx gene are shown with the start and stop codons of translation. Stonor is inserted into the splice acceptor site of intron 5. The LTR of Stonor is represented by black boxes containing white triangles. The internal domain extends from the PBS to the PPT, and it includes the protease (Prot), endonuclease (Endo), reverse transcriptase (RT), and RNaseH domains.


*  MATERIALS AND METHODS
*TOP
 *ABSTRACT
 *MATERIALS AND METHODS
*RESULTS
 *DISCUSSION
 *LITERATURE CITED

Strains:
Maize strains Black Mexican Sweet, Zapalote Chico 2840, and Zapalote Chico 217413 were obtained from the North Central Regional Plant Introduction Station (Ames, IA). M14, B37, W23, W22, and GA221 are standard inbred lines. Two hybrid lines, Q60 and B70, were obtained from Lane Arthur (University of Georgia). Teosinte strains Zea maize mexicana and Z. maize parviglumis were obtained from the National Clonal Germplasm Repository (Miami, FL).

Cloning procedures:
The 5' and 3' halves of the Stonor element were cloned previously (VARAGONA et al. 1992 Down), except for 170 bp located between the cloned fragments. A fragment including the 170 bp was isolated after PCR amplification using a primer located in Stonor sequences contained in the 5' end clone (primer STB: 5' GGGTTCACACAGAGAGAAGG 3') and a second primer located in wx sequences downstream of the insertion (wxexon72: 5' TTGAGGTAGCACGAGAGAGG 3'). PCR amplification was performed in 100-µl reactions with 200 ng genomic DNA, 70 ng primer, and 2.5 units DNA polymerase (Amplitaq; Perkin Elmer, Norwalk, CT) in PCR reaction buffer (10 mM Tris-HCl pH 8.3, 50 mM KCl, 1.5 mM MgCl2, and 0.01% gelatin). The reactions were cycled 40 times for 1 min at 95°, 2 min at 65°, 3 min at 72°, and then at 72° for 10 min. The amplified fragment was subcloned in pUC119 and sequenced.

Sto-12, Sto-14, and Sto-30 were cloned after PCR amplification using primer pairs ST6 (5' CGGATTGGTATTCTAGGGAC 3')/STC (5' GATGTACCACTGTCTGGAGG 3'), ST5 (5' GGAGATCGTCAAGAAGGAGG 3')/STC, and ST5/STC2 (5' CCACCTGTGGTCATAGTTGG 3'), respectively. PCR buffer and amplification conditions were the same as described above. Sto-3 and Sto-4 were cloned after PCR amplification using PCR primers ST5 and ST3 (5' GGTCGAGCGATTAGCGTTG 3'). To amplify the large sequences corresponding to entire elements (PCR product of ~6.5 kb), the Elongase enzyme mix from Bethesda Research Laboratories (BRL, Gaithersburg, MD) was used. PCR reactions (50 µl) contained 200 ng of genomic DNA, 0.2 mM of each primer, and 0.2 mM dNTPs in PCR buffer [60 mM Tris-SO4, pH 9.1, 18 mM (NH4)2SO4, and 1.5 mM MgSO4]. Reactions were heated to 94° for 30 sec, cycled 35 times for 30 sec at 94°, 30 sec at 62°, and 7 min at 68°, and then incubated at 68° for 10 min. PCR products were cloned in the pGEM-T vector (Promega, Madison, WI).

Sto-1, Stl-2, Stl-3, Stl-4, Stl-5, Stl-8, Stl-10, Stl-12, Stl-14, Stl-18, Stl-19, Stl-20, Stl-21, Stl-23, Stl-24, Stl-31, Stl-33, Stl-35, Stl-36, Stl-37, Stl-39, and Stl-40 were obtained after PCR amplification from maize genomic DNA using primers ST5 and STC2 (buffer and amplification conditions were identical as described for Sto-30), and they were cloned in the TA cloning vector (Invitrogen, San Diego, CA).

A genomic library was constructed by cloning maize genomic DNA digested with BamHI into the XhoI site of lambda ZAPII (Stratagene, La Jolla, CA). Insert and vector cohesive ends were rendered compatible by partial filling of the ends of the BamHI-digested DNA fragments with dGTP and dATP, and the ends of the XhoI-digested vector with dTTP and dCTP. Sto-17 was obtained by screening this library with a 556-bp SstI-PstI fragment derived from the internal domain of Sto-14.

The sequences determined from these clones have been deposited into the GenBank database under the accession numbers AF082127, AF082128, AF082129, AF082130, AF082131, AF082132, AF082133, AF082134, AF082135, AF082136, AF082137, AF082138, AF082139, AF082140, AF082141, AF082142, AF082143, AF082144, AF082145, AF082146, AF082147, AF082148, AF082149, AF082150, AF082151, AF082152, AF082153, AF082154, AF082155.

Plasmid construction and probe synthesis:
Plasmid pSto-wx was made by subcloning a 509-bp SalI-NsiI fragment from the wx gene and a 556-bp PstI-Sst-I fragment from Sto-14 into pUC119. A 1.05-kb fragment containing wx and Sto-14 sequences was amplified from pSto-wx using the m13-puc forward and reverse primers, and it was labeled using the Random Primers DNA Labelling System from BRL to produce probe PrSto-wx.

Sequence analysis and phylogenetic analysis:
Multiple alignments were made using the program PILEUP of the University of Wisconsin Genetics Computer Group (GCG) and displayed using the GCG program BOXSHADE. An alignment of inferred amino acid sequences corresponding to the reverse transcriptase and part of the RNaseH domains of Stonor [sequences located between nucleotide (nt) positions 2407 and 3630], Tnt1 (GenBank accession no. X13777), Tto1 (D83003), Ta1-3 (X13291), BARE-1 (Z17327), Osser (X69552), copia (X0-2599), PREM-2 (ZMU41000), Hopscotch (U12626), Tst1 (X52-387), and RIRE1 (D85597) was made using PILEUP. A phylogenetic tree was constructed using these sequences and the programs PROTPARS and SEQBOOT from the PHYLIP 3.5 package (using maximum parsimony analysis). A consensus tree was obtained after running 100 bootstraps.

DNA gel blot analysis:
Hybridizations and washes (with 0.1x SSC and 0.5% SDS) were performed at 65°.


*  RESULTS
*TOP
 *ABSTRACT
 *MATERIALS AND METHODS
 *RESULTS
*DISCUSSION
 *LITERATURE CITED

The wxStonor allele contains a copia-like retrotransposon:
The 4542-bp insertion in the wxStonor allele was cloned and sequenced (see MATERIALS AND METHODS) and found to have structural features characteristic of retrotransposons, including the following: (i) a perfect LTR of 560 bp flanking a 3422-bp internal domain, (ii) an 18-nt primer-binding site (PBS) adjacent to the 5' LTR and homologous to the 3' end of the wheat Met-tRNA (GHOSH et al. 1982 Down), and (iii) a polypurine tract (PPT) upstream of the 3' LTR (Figure 1). Internal sequences encode a putative 998-amino-acid open reading frame (ORF) that contains regions with similarity to several copia-like retrotransposons from Drosophila (copia, 1731) and plants (Tnt1, Tto1, Ta1-3, Tst1, Hopscotch, and BARE-1). The most extensive identity between Stonor and other retrotransposons includes the endonuclease, reverse transcriptase, and RNaseH domains (Figure 2A). However, Stonor appears to be a defective element that is missing the entire gag domain (normally found between the 5' LTR and the pol domain) and part of the protease domain (Figure 2B).



View larger version (54K):
In this window
In a new window
Download PPT slide
  		Figure 2. Amino acid sequences shared by Stonor (or Sto-4), Tnt1, and copia. (A) Alignment of the endonuclease, reverse transcriptase, RNaseH, and nucleic acid-binding domains. (B) Components of an autonomous retrotransposon. Plant retrotransposons contain a gag domain, which includes a nucleic acid-binding domain, located between the 5' LTR and pol. The gag and pol polyproteins are encoded in a single ORF (represented by the bent arrow). In Stonor, the sequences corresponding to the gag domain and part of the protease are deleted. Other features are as described in Figure 1.

Stonor is part of a multigene family in maize, teosinte, and Tripsacum:
Southern blot analysis was undertaken to identify additional Stonor elements in the maize genome (including complete elements) and to ascertain the species distribution of this retrotransposon. To this end, genomic DNAs were digested with EcoRI (which does not cut within available Stonor sequences), and the blot was probed with the Stonor LTR. Using high-stringency washes (see MATERIALS AND METHODS), many sequences related to Stonor were detected in maize, teosinte, and Tripsacum, but not in the genomes of rice, millet, soybean, and oat (Figure 3).



View larger version (123K):
In this window
In a new window
Download PPT slide
  		Figure 3. DNA gel blot analysis of Stonor sequences in monocotyledonous and dicotyledonous plants. EcoRI-digested genomic DNA (10 µg) of teosinte (lane 1), maize (lane 2), rice (lane 3), millet (lane 4), soybean (lane 5), Tripsacum dactyloides (lane 6), and oat (lane 7) were separated by gel electrophoresis, transferred to nylon membrane, and hybridized with a probe derived from Stonor LTR sequences. The sizes of molecular weight standards are given in kilobases on the left.

Identifying larger family members in the maize genome:
To determine whether elements larger than Stonor are present in the maize genome and, thereby, identify the sequences missing in Stonor, PCR was performed using primers located on either side of the putative deletion endpoints (primers ST5 and STC, Figure 4E). Genomic DNA from several maize strains and from teosinte was used in conjunction with a protocol designed to amplify products of <=10 kb (Elongase; BRL). PCR products were analyzed by Southern blot with a probe derived from the LTR. A fragment corresponding to the size expected for the Stonor element (2338 bp) was amplified from a strain containing the wxStonor allele (Figure 4A, lane 1), but not from any other maize strain. This observation suggests that the deletion in Stonor is not present in any other element from this family and may have been created during transposition of an active family member into the wx gene (see below).



View larger version (82K):
In this window
In a new window
Download PPT slide
  		Figure 4. DNA gel blot analysis of PCR products amplified from maize genomic DNA using primers located in LTR and internal element sequences. Blots were hybridized with a probe derived from the Stonor LTR. (A) Products obtained by PCR amplification using primers ST5 and STC (see E) with DNA from the following: a maize strain containing the wxStonor allele (lane 1), Q60 (lane 2), B70 (lane 3), Black Mexican Sweet (lane 4), M14 (lane 5), Zapalote Chico 217413 (lane 6), Zapalote Chico 2840 (lane 7), B37 (lane 8), w22 (lane 9), Z. maize mexicana (lane 10), Z. maize parviglumis (lane 11), and no DNA (lane 12). (B–D) Same as in A, except that primer pairs ST5/ST3, STB/ST3, and STA/ST3, respectively, were used for amplification. (E) Structure of some of the PCR products obtained by amplification using primer pairs ST5/STC, ST5/ST3, STB/ST3, and STA/ST3.

The only product amplified from all strains tested was also the largest, at ~5 kb (Figure 4A). Given the size of this fragment, the position of the primer within the LTR, and the size of the Stonor element, we estimate that a complete element should be ~7.2 kb. To confirm the size and presence of this element in the genomes of maize and teosinte, PCR was performed using a second pair of primers located in the 5' and 3' LTRs (primers ST5 and ST3, Figure 4E). The positions of these primers (with the ST5 sequence downstream of the ST3 sequence) were chosen to preclude amplification of a product from a single LTR. Consistent with the existence of an element of 7.2 kb was the amplification in most strains of a product of 6.5 kb (Figure 4B). A 6.5-kb fragment is not visible among the products from a strain containing the wxStonor allele. However, because of the efficient amplification of the PCR product corresponding to Stonor, a fivefold dilution of this reaction was loaded on the gel. More efficient amplification of the Stonor element may reflect its small size and/or the perfect match of primer sequences.

Unlike the 5' half of Stonor, the 3' half seems to be intact and representative of the other elements in the maize genome. Use of primers STB and ST3 (Figure 4E) resulted in the amplification of the expected 1.7-kb product (Figure 4C). Similarly, use of a more 5' primer, STA (obtained from the Sto-4 element, Figure 5A), with ST3 also resulted in a single product of 3.8 kb (Figure 4D and Figure E). Taken together, these data suggest that the structural variation of family members is largely restricted to the region just downstream of the 5' LTR.



View larger version (29K):
In this window
In a new window
Download PPT slide
  		Figure 5. Structure of the cloned family members. (A) Black areas correspond to parts of the elements that were cloned and sequenced, while gray areas represent regions cloned but not sequenced. The rectangles at the termini of each element represent the LTRs. Small arrows indicate the location of the primers used for PCR amplification. Large, bent arrows represent the extent of the ORF in the sequenced part of the elements. Asterisks indicate the location of frameshifts in the large ORFs. (B) Structure of the 5' region of the cloned elements. The LTRs (shown as black rectangles with a white triangle) are highly conserved (93–100% identity), as is the 140-bp region A (92–99% identity). Sequences downstream of region A are highly variable among the cloned elements and contain segments that can either be aligned with some of the other elements (B–F, H, and I) or regions that cannot be aligned with any other cloned sequence (G1–G6). Sequences located downstream of the variable region and upstream of the putative ORF (region J) are highly conserved (97–99.9% identity), as are downstream sequences in the ORF.

Structural characterization of Stonor-like sequences:
Both PCR products and fragments from genomic libraries were analyzed to define the Stonor deletion and to identify the putative 7.2-kb, full-length element(s) predicted to be in all maize strains analyzed. The structure of seven reconstructed elements and the Stonor element are shown in Figure 5A. The shaded regions of all elements, except Sto-17 at the top and Stonor at the bottom, were obtained as PCR fragments using the primer pairs shown. Sto-17, like Stonor, was isolated from a genomic library. Unshaded regions represent our reconstruction of what the rest of the element may look like in the genome, based on a comparison with the other elements and the PCR results summarized in Figure 4.

Based on a LTR length of 560 bp for Stonor (VARAGONA et al. 1992 Down), complete Sto-3 and Sto-4 elements are estimated to be 7.1 and 7.2 kb, respectively. Estimates for Sto-17, Sto-30, Sto-1, Sto-12, and Sto-14 were based on the sizes of the cloned fragments because we had previously determined that the 3' half of most family members was intact (Figure 4C and Figure D). Sto-4 encodes a single intact ORF of 1406 amino acids that is located 1668 bp downstream of the 5' LTR (Figure 5A). This ORF contains all the sequence features present in active elements, including the nucleic acid-binding domain (part of gag, Figure 2B) and a complete pol domain. In contrast, both Sto-17 and Sto-3 contain a frameshift in the sequenced part of their respective ORFs (Figure 5A), suggesting that they are defective elements. The PBS of all elements, except Stonor, is 11 nt. The Stonor PBS has been extended to 18 nt by the deletion (see below).

Surprisingly, restriction sites in one region of several independently derived clones were completely different (variable region, Figure 5A). Sequence analysis of Sto-17, Sto-3, Sto-4, Sto-30, and Sto-1 revealed a highly variable region that includes both unique sequences (represented by G1–G6, Figure 5B) and sequences held in common among different elements (represented by A–I, Figure 5B). More detailed characterization of the variable region is described below.

Structure of the deletion derivatives:
The sequences at deletion endpoints can be informative in deducing the mechanism of deletion formation. For the Stonor element, these endpoints are located in the PBS and the ORF, two regions that are virtually identical in all the larger family members (Figure 5B and Figure 6A). For this reason, it was a straightforward task to compare Stonor with the larger elements and identify where the sequences diverged. These regions are indicated by the juxtaposition of uppercase, bold letters and lowercase letters in the large element (Figure 6A, top). For Stonor, the sequences downstream of the PBS, beginning with the trinucleotide GTT, are homologous with the protease domain. In the large elements, these sequences are located 2.6 kb downstream of the 5' LTR. Stonor also contains 4 bp, CCAG, that cannot be aligned at either the 5' or 3' deletion breakpoints. DNA inserted between deletion endpoints is called filler DNA and is frequently encountered in maize deletions, where it is believed to result from double-slip mispairing during DNA replication (WESSLER et al. 1990 Down). According to this mechanism, filler DNA is copied from sequences flanking the deletion. However, it is unlikely that such a mechanism was involved in the generation of the deletion in Stonor because the sequences of the filler DNA and of the downstream trinucleotide GTT increase the length of sequences complementary to the Met-tRNA from 11 to 18 nt (Figure 6A). Instead, this structure suggests that the deletion in Stonor was generated during retrotransposition (see DISCUSSION).



View larger version (32K):
In this window
In a new window
Download PPT slide
  		Figure 6. Structure of the deletions in Stonor, Sto-14, and Sto-12. Sequences shown in lowercase letters are present in the large elements but absent in the deletion derivatives. (A) Structure of the deletion in Stonor. Boxed sequences are complementary to the Met-tRNA. Eighteen nucleotides of tRNA sequence complementary to the PBS of Stonor are shown. (B and C) The deletions in Sto-14 and Sto-12. Direct repeats at the deletion endpoints are boxed.

Like Stonor, both Sto-12 and Sto-14 have sustained deletions with one breakpoint downstream of the 5' LTR and the other in the ORF (Figure 5A and Figure B). However, the 5' breakpoints for Sto-12 and Sto-14 were not present in all the larger elements. For Sto-14, the deletion breakpoints were determined by comparison with Sto-1, which shares the AC motif, and were found to coincide with a 6-bp direct repeat (Figure 6B). For Sto-12, the 5' breakpoint was identified by comparison with Sto-1, which is the only other cloned family member containing the sequences present downstream of region F (Figure 5B, region G6). The 3' breakpoint of Sto-12 occurs in sequences that were not cloned in Sto-1, but that were cloned and sequenced in several of the other elements, including Sto-4, Stonor, and Sto-17 (Figure 6C). Based on such a comparison, the Sto-12 deletion breakpoints were also found to lie within a direct repeat, this one of 3 bp (Figure 6C).

Copy number of elements in the Stonor family:
The size of the Stonor family was estimated by analyzing maize genomic DNA on Southern blots using a probe that can hybridize both to element sequences and to an unrelated single copy sequence. Upon restriction with an enzyme that cuts twice in a conserved region of the internal domain, most elements should give rise to a fragment of the same size, while the unrelated single copy sequence should give rise to a fragment of a different size. Estimation of the copy number is determined by comparing the intensity of the hybridization signals.

Probe PrSto-wx contains 556 bp from the conserved region of the internal domain of one of the members of the Stonor family (Sto-14, see MATERIALS AND METHODS). This was then fused with a fragment of approximately the same size (509 bp) from the wx gene. When hybridized with a Southern blot of maize genomic DNA digested with EcoRI and SalI, a major band of 1.1 kb is seen in addition to several minor bands (Figure 7), one of which (2 kb) corresponds to the wx gene. Quantification of the hybridization signals with a PhosphorImager indicates a copy number of 20–25 elements for the 1.1-kb EcoRI fragment. At least 12 other sequences, which probably represent elements that lack an EcoRI site or that have undergone insertions or deletions, also hybridize to probe PrSto-wx. When the elements corresponding to these bands are included, at least 32 elements are estimated to be present in the maize genome. A similar experiment using a probe from the reverse transcriptase domain indicated a copy number of 30–40 elements (data not shown).



View larger version (61K):
In this window
In a new window
Download PPT slide
  		Figure 7. DNA gel blot determination of the copy number of elements in the Stonor family. (A) Ten micrograms of maize genomic DNA (from inbred GA221) digested with EcoRI and SalI were loaded in lane 1. Lanes 2–7 each contain half the amount of DNA loaded in the previous lane. The sizes of molecular weight standards are given in kilobases on the left. (B) The annealing of probe PrSto-wx to the Stonor family members. Annealing of the wx sequences present in this probe is not shown.

Are the large cloned elements representative of the Stonor family?
Analysis of PCR products (Figure 4) indicates that the Stonor family is comprised of large elements of ~7.2 kb with variable sequences downstream of the 5' LTR. In addition, smaller members have sustained deletions of sequences downstream of the 5' LTR. However, PCR products may not accurately reflect the composition of the Stonor family because PCR can preferentially amplify small elements or those that contain sequences more similar to the primer pairs. For these reasons, Southern blots were used to obtain an independent assessment of family composition. Ideally, this is accomplished by digesting genomic DNA with an enzyme that recognizes sites in the LTRs and by determining the size of the resulting fragments on Southern blots with a probe derived from the internal domain of the element. Unfortunately, no enzyme could be found with restriction sites only in the LTRs. Instead, we used AspI, which cuts once in each LTR and twice in conserved regions of the internal domain (Figure 8C). Southern blots were then hybridized with probes homologous to each resulting fragment (probes Pr1–Pr3) from regions conserved in the four cloned large elements. The strongest hybridization signals obtained with these probes identified fragments of ~2.1 kb, 430 bp, and 4 kb, consistent with the presence of abundant elements of ~7.2 kb (Figure 8A).



View larger version (44K):
In this window
In a new window
Download PPT slide
  		Figure 8. DNA gel blot analysis of Stonor family members in the maize genome. (A) Ten micrograms of maize genomic DNA (from inbred W22) digested with AspI was separated by gel electrophoresis, transferred to a nylon membrane, and hybridized with probes Pr1 (lane 1), Pr2 (lane 2), and Pr3 (lane 3). (B) Ten micrograms of maize genomic DNA was digested with either AspI (noted A) or EcoRI (noted E), and was probed with Ps-30, Ps-3, Ps-4, and Ps-17 (see Figure 5B for the position of these probes relative to regions G1–G4 and E). (C) Positions of the probes within Stonor family members. The AspI site noted with an asterisk may not be present in all family members.

Representation of the variable regions in genomic DNA:
Each of the large elements contains sequences that cannot be aligned with sequences from any of the other cloned elements. To determine whether these sequences are unique or repetitive in the maize genome, they were used as probes to analyze Southern blots of maize genomic DNA (Figure 8B). Probes Ps-3, Ps-4, and Ps-17 were derived from the variable regions of Sto-3, Sto-4, and Sto-17, respectively (Figure 5B and Figure 8C). Probe Ps-30 contains sequences that are shared by Sto-30 and Sto-4 (Figure 5B, region E), as well as other sequences that are unique to Sto-30 (Figure 5B, region G4).

Maize genomic DNA was digested with either AspI or EcoRI; the latter should digest once in genomic sequences upstream of the element and once in element sequences downstream of the four probes. Digestion with AspI should give rise to 2.1 to 2.3-kb fragments if all sequences hybridizing to the probe are part of elements with the same AspI sites as in their respective cloned elements (Figure 8C). Ps-30, Ps-3, Ps-4, and Ps-17 appear to hybridize to a subset of the sequences identified by the conserved Pr-1 probe (Figure 8B). Included among these bands are the predicted 2.1- to 2.3-kb fragments in addition to larger fragments that may represent elements that have undergone mutations of the AspI sites. Alternatively, these sequences may not reside in Stonor elements. All but one of the EcoRI fragments are >2.1–2.3 kb, as expected for an enzyme with sites in flanking upstream sequences. The one exception is a fragment of ~1.5 kb detected by Ps-30, a probe that also contains sequences held in common by another cloned element (Figure 5B, region E).

PCR amplification and sequence analysis of additional variable regions:
Maize genomic DNA from four strains was used in conjunction with the primer pair ST5 and STC2 to amplify variable regions that resided in members of the Stonor family (Figure 9A). Twenty-one independent clones containing amplified products were isolated from the following strains: B70, Stl-2, Stl-3, Stl-4, and Stl-5; BMS, Stl-12, Stl-14, Stl-18, Stl-19, and Stl-20; M14, Stl-21, Stl-23, and Stl-24; and W22, Stl-31, Stl-33, Stl-35, Stl-36, Stl-37, Stl-39, and Stl-40. For each clone, 1 kb of sequence was obtained from an overlapping region of variable and conserved sequences (Figure 9A, segment V includes 500 bp of variable region, and segment C includes region J and 100 bp of the ORF).



View larger version (50K):
In this window
In a new window
Download PPT slide
  		Figure 9. Sequence variability in the leader region of Stonor family members. (A) The regions amplified with primers ST5 and STC2. Cloned sequences are shown in dark gray (conserved) or light gray (variable). The sequenced regions of the cloned PCR products are underlined and labeled for variable (V) or conserved (C). (B) Phylogenetic tree of all cloned Stonor family members based on an alignment of sequences from segment V of the variable region. Multiple alignments were made with the program DNASTAR using the Clustal method with a gap penalty of 8 and a gap length penalty of 8. The percent divergence between elements is indicated by the horizontal axis. Clones exhibiting <0.9% divergence were grouped together. Each group number is indicated in parentheses. (C) Similarity matrix of 16 Stonor family members calculated for sequence segments V and C from the alignment described above. Numbers below the diagonal represent similarity between clones for the variable region V, while numbers above the diagonal represent similarity in the conserved region C.

Multiple sequence alignments were performed for these regions of the 21 new sequences and the five cloned elements containing variable domains. These 26 sequences fell into 16 distinct groups (Figure 9B). For 7 of these groups, several clones were found to be almost identical, with the exception of a few nucleotides. It is not known whether these sequence differences result from PCR errors that follow amplification of identical elements, or whether there are multiple copies of virtually identical elements in a single genome. For example, clones such as Sto-3 and Stl-12 or Stl-8 and Stl-37 were very similar (2- and 8-nt differences out of 1 kb, respectively) even though they were amplified from different maize strains. In contrast, most clones from different groups were found to be extremely divergent, with <60% identity (Figure 9C). For almost all sequence comparisons performed between elements of different groups, at least part of the variable region was so divergent that the sequences could not be aligned properly. In contrast to the variable region, all sequences in the conserved region were >90% conserved (Figure 9C).


*  DISCUSSION
*TOP
 *ABSTRACT
 *MATERIALS AND METHODS
 *RESULTS
 *DISCUSSION
*LITERATURE CITED

The wxStonor allele was shown previously to contain a retrotransposon insertion in the maize wx gene (VARAGONA et al. 1992 Down). The sequence of this insertion, now called the Stonor element, has been determined and found to contain a deletion that includes all of gag and a portion of the protease domain. This deletion could have arisen during or after element insertion into wx. As such, this element is probably no longer capable of further retrotransposition. Alternatively, retrotransposition of the defective Stonor element may have been mediated by the products of an autonomous element located elsewhere in the maize genome.

The Stonor deletion:
Several lines of evidence suggest that the Stonor deletion was sustained during retrotransposition and probably just before insertion into the wx gene; i.e., the Stonor element was recently active but is no longer capable of retrotransposition. First, what remains of the large ORF is still intact. This is expected if the Stonor element was recently active, but not if its movement was complemented in trans by the products of another family member. Second, we show that the Stonor element is restricted to strains carrying the wxStonor allele. This is consistent with the behavior of an element rendered defective upon insertion. An element that can be complemented in trans would have predated the wxStonor mutation and might be expected to be in other strains.

The strongest evidence for the generation of the Stonor deletion during retrotransposition is that it is reminiscent of similar structures formed during retrovirus retrotransposition. The relevant steps in both retrovirus and retrotransposon transposition are summarized in Figure 10A. According to PULSINELLI and TEMIN 1991 Down, in some instances, DNA synthesis may continue past the usual site in the tRNA, resulting in the addition of tRNA-derived nucleotides at the 3' end (Figure 10B, 6.2). After strand transfer, these extra nucleotides cannot hybridize with the existing PBS and may "seek out" complementary sequences downstream (Figure 10B, 7.1–7.2). DNA synthesis then continues, resulting in the structure observed in Stonor.



View larger version (28K):
In this window
In a new window
Download PPT slide
  		Figure 10. Model for the generation of the deletion in the Stonor element. Thin lines are RNA, and thick lines are DNA. (A) Steps involved in normal reverse transcription of LTR-retrotransposons or retroviruses. U3, R, and U5 comprise the LTR. (1) Annealing of a Met-tRNA molecule (the cloverleaf structure) to the element-encoded RNA. (2) The Met-tRNA serves as a primer for synthesis of a DNA fragment corresponding to the R and U5 regions (called minus-strand, strong-stop DNA). (3) Degradation of the R and U5 regions of the template with RNaseH and transfer of the minus-strand, strong-stop DNA to the other end of the RNA template via homology to the R repeat. (4) Minus-strand DNA synthesis continues. (5) RNaseH nicks the RNA template at the PPT (location of the nick shown as a white arrow). (6) After removal of RNA downstream of PPT with RNaseH, plus-strand DNA synthesis initiates at PPT and stops in the Met-tRNA after synthesis of the plus-strand, strong-stop DNA (containing U3, R, U5, and PBS). (7) Removal of RNA by RNaseH followed by plus-strand strong-stop DNA transfer. (8) Completion of plus-strand and minus-strand DNA syntheses by reverse transcriptase. (B) Model for the generation of the deletion in Stonor. Steps 1–5 are not shown because they occur as described in A. (6.1) Plus-strand DNA synthesis proceeds through the point where it normally stops and (6.2) copies an additional 7 nt from the tRNA template. (7.1) Plus-strand DNA transfer. (7.2) Annealing of the extra nucleotides to downstream sequences. (8) Completion of DNA synthesis.

Generation of the Stonor deletion by such a mechanism implies that the primer tRNA sequence was inherited during retrotransposition of Stonor into the wx gene. Such an event was thought to be associated with the propagation of full-length genomic copies of both retroviruses and retrotransposons until LAUERMANN and BOEKE 1994 Down demonstrated that the primer tRNA sequence was not inherited during retrotransposition of the yeast Ty1 element. Because Stonor is obviously a defective element, retention of the tRNA sequence may reflect the aberrant outcome of an unusual event. Alternatively, inheritance of tRNA sequences during reverse transcription may be a feature that distinguishes retrotransposition of the Stonor element, and perhaps other plant elements, from the yeast Ty1 element.

Other deleted members of the Stonor family:
One by-product of the search for larger members of the Stonor family was the isolation of two elements (Sto-14 and Sto-12, Figure 5A) that, like Stonor, have sustained deletions. However, unlike Stonor, both of these elements contain numerous frameshifts, stop codons, deletions, and/or insertions in the ORF, indicating that they have been inactive for a long time. Furthermore, the deletion endpoints of both Sto-12 and Sto-14 occur in a direct repeat sequence (Figure 6B and Figure C), a structure associated with deletion formation during both DNA replication (NALBANTOGLU et al. 1986 Down) and reverse transcription (TEMIN 1993 Down).

The larger family members:
The results of three independent experimental approaches have led us to conclude that the largest members of the Stonor family are ~7.1–7.4 kb. First, a PCR assay using primers located in the 5' and 3' LTRs generated a product of 6.5 kb from genomes of maize and teosinte. When the positions of the primers are taken into account, this corresponds to an element of ~7.2 kb. Second, when primers flanking the endpoints of the Stonor deletion are used, the largest PCR product obtained was 5 kb. This would also correspond to a reconstituted Stonor element of 7.2 kb. Finally, the largest Stonor family member isolated from a genomic library, Sto-17, would be 7.4 kb upon restoration of its missing 3' end (Figure 5A).

It is unknown at this time whether the active members of the Stonor family are of this size class. Because we have not been able to identify element-encoded transcripts (S. MARILLONNET and S. R. WESSLER, data not shown), the Stonor element provides our only link to the active progenitor (if we assume that it was active before sustaining the deletion, as described above). If this scenario is correct, then identification of the Stonor progenitor could lead us to an active family member. Unfortunately, both the 5' and 3' deletion breakpoints of Stonor reside in sequences held in common by all the cloned large elements (Figure 5B); i.e., the variable region has been deleted in Stonor, making it impossible to identify its direct progenitor.

Stonor and related elements have long 5' leaders:
The larger elements exhibit two unusual features downstream of the 5' LTR. First, if transcribed, they would have exceptionally long 5' leaders. The Sto-4 element has a single intact ORF of 1406 amino acids that encodes all the protein domains found in active retrotransposons. This ORF has a putative translation start ~1.6–1.8 kb from the 5' LTR, just downstream of region J (Figure 5B). A putative TATA box is located between nucleotide positions 132 and 138 in the 5' LTR (VARAGONA et al. 1992 Down). Transcripts that may initiate at this position would have a 2.0- to 2.2-kb leader, which is longer than all other 5' leaders reported to date. Only the BARE-1 element of barley and the RIRE1 element of wild rice have leaders of comparable length (SUONIEMI et al. 1996B Down). Interestingly, an analysis based on the alignment of the conserved regions of either the reverse transcriptase/RNaseH or protease domains of Stonor and several plant and Drosophila retrotransposons indicates that Stonor is more closely related to BARE-1 and RIRE1 than to other retrotransposons, including those from maize (Figure 11 and data not shown).



View larger version (17K):
In this window
In a new window
Download PPT slide
  		Figure 11. Maximum parsimony tree of copia-like retrotransposons from plants and Drosophila using copia as the outgroup. Numbers next to each node give the bootstrap value supporting this node.

The presence of very long leaders in three (distantly) related families may indicate that this feature has been conserved because it has functional significance. The BARE-1 leader and the four putative leaders of the large Stonor family members contain numerous ATG codons in all reading frames (49 in BARE-1, 32 in Sto-3, 34 in Sto-4, 41 in Sto-30, and 38 in Sto-17). Leader AUGs have been shown to inhibit translation of downstream ORFs in both plants and animals (DAMIANI and WESSLER 1993 Down; GEBALLE 1996 Down). However, while this may be a post-transcriptional mechanism to control retrotransposition of Stonor family members, it has not been very effective at regulating BARE-1, whose 30,000 copies comprise nearly 7% of the barley genome (SUONIEMI et al. 1996A Down). Several viruses, however, including picornaviruses, can bypass the inhibitory effect of upstream AUGs by using internal ribosome entry sites (IRES) (KAMINSKI et al. 1994 Down; EHRENFELD 1996 Down) that feature complex secondary structures. In picornaviruses, an IRES of 400–500 bp is required for internal initiation and is conserved within the viral species. In Stonor family members, ~400 bp of sequences located upstream of the large ORF are conserved and might potentially be involved in internal ribosome entry (Figure 5B, region J).

A variable domain in the larger elements:
The second unusual feature of the Stonor leader is a variable region that results in the structural uniqueness of each of the cloned elements. Variability of this type has not been reported previously for retrotransposons. Of interest is whether the variable sequences are an integral part of the element family, or whether they are random mutations that accumulate in inactive elements. The fact that these sequences are located at approximately the same position in all four elements and are flanked by sequences with >90% identity suggests the existence of a variable domain in the Stonor family.

What is the origin of this variable domain? It is unlikely to be caused by the insertion of transposable elements because the sequences do not have typical structural features of elements (e.g., terminal inverted repeats and direct repeats). Furthermore, the diversity of sequences would mean that many element families would have inserted into similar but not identical sites. In contrast, mechanisms have been proposed for the acquisition of nonelement sequences during retrotransposition. The variable regions may have been acquired by a template switch mechanism like that proposed for transduction of cellular genes by oncogenic retroviruses (SWAIN and COFFIN 1992 Down; ZHANG and TEMIN 1993 Down). A similar mechanism was proposed for the acquisition of part of a maize cellular gene by the Bs1 retrotransposon (BUREAU et al. 1994 Down; JIN and BENNETZEN 1994 Down). However, the presence of so many different sequences in approximately the same region of Stonor family members is hard to reconcile with any single mechanism.

Low- vs. high-copy-number families:
Stonor, like other elements responsible for mutations in maize, is a member of a relatively low-copy-number family of ~30–40 elements. Family members display an unusual amount of structural diversity, including numerous deletions of sequences downstream of the 5' LTR. The sequence of the Stonor deletion suggests that it was generated during retrotransposition. Similarly, other retrotransposon-induced mutations in maize contain defective elements, including wxM and pl-987 (defective Magellan elements, PURUGGANAN and WESSLER 1994 Down; P. L. COOPER and K. C. CONE, personal communication) and bm-3 (a defective B5 element, VIGNOLS et al. 1995 Down). Such examples may indicate that the generation of inactivating mutations may be one mechanism to minimize the copy number of these element families.

A post-transcriptional mechanism to control the Stonor copy number is also suggested by the identification of a long and variable leader with many upstream AUGs. However, the phylogenetically related BARE-1 element of barley has managed to attain very high copy numbers despite having a similar number of upstream AUGs. This could mean that the long leaders of Stonor, BARE-1, and RIRE1 are not conserved because they have a role in keeping copy number in check. Alternatively, the high copy number of the BARE-1 family may be caused by the amplification of a variant that was able to overcome the repressive effects of upstream AUGs. The finding that most of the BARE-1 family members appear to be full length and structurally homogeneous is consistent with this scenario (SUONIEMI et al. 1996A Down). However, for such an element to be successful, it also had to evolve a mechanism to avoid inserting into genes. Although nothing is known about a target site preference for BARE-1, almost 60% of the 17 characterized insertion sites were found to be in other retrotransposons (SUONIEMI et al. 1997 Down). The existence of the wxStonor mutation suggests that the Stonor family has not (as yet) evolved a mechanism to avoid inserting into genes. The inability to evolve such a mechanism may be the most potent selective force preventing the amplification of the Stonor family.


*  FOOTNOTES

1 Present address: The Sainsbury Laboratory, John Innes Centre, Norwich NR4 7UH, United Kingdom. Back


*  ACKNOWLEDGMENTS

We thank Drs. Kelly Dawe and Mike Scanlon for critical reading of the manuscript. This work was supported by a grant from the National Institutes of Health to S.R.W.

Manuscript received April 25, 1998; Accepted for publication July 29, 1998.


*  LITERATURE CITED
*TOP
 *ABSTRACT
 *MATERIALS AND METHODS
 *RESULTS
 *DISCUSSION
 *LITERATURE CITED

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

DAMIANI, R. D. and S. R. WESSLER, 1993  An upstream reading frame represses expression of Lc, a member of the R/B family of maize transcriptional activators. Proc. Natl. Acad. Sci. USA 90:8244-8248[Abstract/Free Full Text].

EHRENFELD, E., 1996 Initiation of translation by picornavirus RNAs, pp. 549–573 in Translational Control, edited by J. W. B. HERSHEY, M. B. MATHEWS and N. SONENBERG. Cold Spring Harbor Laboratory Press, Cold Spring Harbor, NY.

FLAVELL, A. J., E. DUNBAR, R. ANDERSON, S. R. PEARCE, and R. HARTLEY et al., 1992  Ty1-copia group retrotransposons are ubiquitous and heterogeneous in higher plants. Nucleic Acids Res. 20:3639-3644[Abstract/Free Full Text].

GEBALLE, A. P., 1996 Translational control mediated by upstream AUG codons, pp.173–197 in Translational Control, edited by J. W. B. HERSHEY, M. B. MATHEWS and N. SONENBERG. Cold Spring Harbor Laboratory Press, Cold Spring Harbor, NY.

GHOSH, M. P., K. GHOSH, M. SIMSEK, and U. L. RAJBHANDARY, 1982  Nucleotide sequence of wheat germ cytoplasmic initiator methionine transfer ribonucleic acid. Nucleic Acids Res. 10:3241-3247[Abstract/Free Full Text].

HIROCHIKA, H. and R. HIROCHIKA, 1993  Ty1-copia group retrotransposons as ubiquitous components of plant genomes. Jpn. J. Genet. 68:35-46[Medline].

HIROCHIKA, H., K. SUGIMOTO, Y. OTSUKI, and M. KANDA, 1995  Retrotransposons of rice involved in mutations induced by tissue culture. Proc. Natl. Acad. Sci. USA 93:7783-7788[Abstract/Free Full Text].

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

JOHNS, M. A., J. MOTTINGER, and M. FREELING, 1985  A low copy number, copia-like transposon in maize. EMBO J. 4:1093-1102[Medline].

KAMINSKI, A., S. L. HUNT, C. L. GIBBS and R. JACKSON, 1994 Internal initiation of mRNA in eukaryotes, pp.115–155 in Genetic Engineering, Vol. 16, edited by J. K. SETLOW. Plenum Press, New York.

LAUERMANN, V. and J. D. BOEKE, 1994  The primer tRNA sequence is not inherited during Ty1 retrotransposition. Proc. Natl. Acad. Sci. USA 91:9847-9851[Abstract/Free Full Text].

NALBANTOGLU, J., D. HARTLEY, G. PHAER, G. TEAR, and M. MEUTH, 1986  Spontaneous deletion formation at the aprt locus of hamster cells: the presence of short sequence homologies and dyad symmetries at deletion termini. EMBO J. 5:1199-1204[Medline].

PULSINELLI, G. A. and H. M. TEMIN, 1991  Characterization of large deletions occurring during a single round of retrovirus vector replication: novel deletion mechanism involving errors in strand transfer. J. Virol. 65:4786-4797[Abstract/Free Full Text].

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

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

SUONIEMI, A., K. ANAMTHAWAT-JONSSON, T. ARNA, and A. H. SCHULMAN, 1996a  Retrotransposon BARE-1 is a major, dispersed component of the barley (Hordeum vulgare L.) genome. Plant Mol. Biol. 30:1321-1329[Medline].

SUONIEMI, A., A. NARVANTO, and A. H. SCHULMAN, 1996b  The BARE-1 retrotransposon is transcribed in barley from an LTR promoter active in transient assays. Plant Mol. Biol. 31:295-306[Medline].

SUONIEMI, A., D. SCHMIDT, and A. H. SCHULMAN, 1997  BARE-1 insertion site preferences and evolutionary conservation of RNA and cDNA processing sites. Genetica 100:219-230[Medline].

SWAIN, A. and J. M. COFFIN, 1992  Mechanism of transduction by retroviruses. Science 255:841-845[Abstract/Free Full Text].

TEMIN, H. M., 1993  Retrovirus variation and reverse transcription—abnormal strand transfers result in retrovirus genetic variation. Proc. Natl. Acad. Sci. USA 90:6900-6903[Abstract/Free Full Text].

VARAGONA, M. J., M. PURUGGANAN, and S. R. WESSLER, 1992  Alternative splicing induced by insertion of retrotransposons into the maize waxy gene. Plant Cell 4:811-820[Abstract/Free Full Text].

VIGNOLS, F., J. RIGAU, M. A. TORRES, M. CAPELLADES, and P. PUIGDOMENECH, 1995  The brown midrib3 (bm3) mutation in maize occurs in the gene encoding caffeic acid o-methyltransferase. Plant Cell 7:407-416[Abstract].

VOYTAS, D. F., M. P. CUMMINGS, A. KONIECZNY, F. M. AUSUBEL, and S. R. RODERMEL, 1992  copia-like retrotransposons are ubiquitous among plants. Proc. Natl. Acad. Sci. USA 89:7124-7128[Abstract/Free Full Text].

WESSLER, S. R., A. TARPLEY, M. PURUGGANAN, M. SPELL, and R. OKAGAKI, 1990  Filler DNA is associated with spontaneous deletions in maize. Proc. Natl. Acad. Sci. USA 87:8731-8735[Abstract/Free Full Text].

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

ZHANG, J. and H. M. TEMIN, 1993  Rate and mechanism of non-homologous recombination during a single cycle of retroviral replication. Science 259:234-238[Abstract/Free Full Text].