Abstract
The regulatory transposon of the Mutator system of transposable elements in maize is MuDR. MuDR elements produce two transcripts, from genes mudrA and mudrB, encoding proteins MURA and MURB, respectively. Like many other transposons, MuDR elements often undergo deletions, usually of internal sequences. Analysis of a deletion that is restricted to the region encoding MURB demonstrates that this gene is not required to cause excisions of a reporter element, although it may be required for transposition or suppression of suppressible alleles. Conversely, a derivative that lacks the region encoding MURA but that produces MURB is nonfunctional for all aspects of Mutator activity. Northern analysis of these derivatives reveals that each of the two transcripts can be independently transcribed, and analysis using an antibody specific for MURB reveals that mudrB transcript can also be successfully translated and its product appropriately localized in the absence of mudrA. A third deletion derivative provides evidence for a source of previously reported antisense transcript.
TRANSPOSABLE elements were first identified because of their propensity to introduce variation into the genomes of their hosts. It has become clear from a large number of studies (Berg and Howe 1989) that the activity of transposons can also lead to a great deal of variation in their own genomes as well. In the Ac/Ds system of maize, for instance, an autonomous Ac element at the Wx locus was observed to change into a nonautonomous Ds element (McClintock 1963). Subsequent molecular analysis of Ds elements demonstrated that this kind of change had resulted from the loss of sequences between the terminal inverted repeats of the originally active Ac element (Fedoroffet al. 1983; Doringet al. 1984; Pohlmanet al. 1984). Similarly, nonautonomous dSpm elements of the Suppressor-mutator system in maize are the result of deletions within autonomous Spm elements. Over the years several of these deletion derivatives have been characterized (reviewed by Fedoroff 1989). This phenomenon was also observed in Drosophila where, for instance, a large number of deletion derivatives of P elements have been identified (O'Hare and Rubin 1983; O'Hareet al. 1992; Glooret al. 1993; Rasmussonet al. 1993).
In some cases, the altered structure of transposable elements in turn can lead to new kinds of regulation. In the best-studied example of this kind of regulatory evolution, some of the derivatives of P elements exhibit a negative effect on the activity of full-length P elements (Blacket al. 1987; Robertson and Engels 1989; Glooret al. 1993; Rasmussonet al. 1993). Often those negative effects vary in their severity depending on the chromosomal position of the deletion derivative being examined (Glooret al. 1993). Among plant transposable elements, at least one deletion derivative of the maize Spm element encodes a product that can reduce the activity of a full-length version of the element (Cuyperset al. 1988).
With respect to the generation of diversity, the Mutator system of transposable elements in maize is of particular interest. The Mutator system is a diverse family of transposable elements in maize that was originally identified in a line of maize with an exceptionally high mutation rate (Robertson 1978). It is composed of at least five different classes of elements, all of which share homologous terminal inverted repeat sequences (TIRs), but each of which contains completely heterologous internal sequences (see Walbot 1991; Chandler and Hardeman 1992; and Bennetzen 1996 for reviews).
The regulatory transposon for the entire system is a member of the MuDR class of elements (Chometet al. 1991; Hershbergeret al. 1991; Qinet al. 1991; Jameset al. 1993). This 4.9-kbp transposon encodes two convergently transcribed transcripts of 2.8 and 1.0 kb, encoded by the genes mudrA and mudrB, respectively (Hershbergeret al. 1995). On the basis of limited sequence similarity to bacterial transposases, mudrA has been hypothesized to be the transposase (Eisenet al. 1994). The presence of MuDR is required for all aspects of the activity of the entire system of transposable elements, and in all Mu-active lines examined to date both transcripts are present, suggesting that both open reading frames (ORFs) encode proteins necessary for high levels of Mutator activity. Mutator activity is defined as a constellation of behaviors, including high levels of transposition and excision, demethylation of TIRs, and the presence of supercoiled forms of Mu (reviewed by Chandler and Hardeman 1992). Active MuDR elements may also alter the mutagenic effects of Mu transposon insertions near the promoters of genes such as hcf106 (Martienssenet al. 1990), a1-mum2 (Chometet al. 1991), knotted-1 (Greeneet al. 1994), and Lg3-Or 211 (Fowleret al. 1996). In each of these cases expression of a gene is suppressed in the presence of the putative transposase.
The Mutator system is the most mutagenic of the known plant transposable element systems; the high frequency of duplicative transposition of multiple elements in an active Mutator maize stock can result in mutation frequencies 50 times that of background (Robertson 1978). Duplication frequencies can approach or even exceed 100%, so that even on outcrossing to non-Mutator stocks the number of elements, including MuDR elements, can rapidly increase (Alleman and Freeling 1986; Lisch and Freeling 1994). Perhaps as a consequence of this high transposition rate, MuDR elements also undergo frequent internal deletions (Chometet al. 1991; Lisch and Freeling 1994; Hershbergeret al. 1995; Lischet al. 1995; Hsia and Schnable 1996). These deletions have been observed to occur at various times during development, resulting in either somatic or germinally transmitted events (Chometet al. 1991; Lischet al. 1995; Hsia and Schnable 1996).
Analysis of the structure of a number of deletion derivatives of MuDR elements has revealed that the junctions of these derivatives are consistent with a double-stranded gap repair mechanism (Hershbergeret al. 1995; Hsia and Schnable 1996). Lesions that largely delete mudrA eliminate both excision and suppression activity (Chometet al. 1991; Lisch and Freeling 1994; Lischet al. 1995; Hsia and Schnable 1996). In the absence of functional mudrA, residues in the termini of Mu1 elements were methylated, consistent with a lack of activity (Chometet al. 1991). However, these analyses did not reveal how deletions within MuDR elements affect expression of each of the two genes encoded by this element, and thus the degree to which these genes are independent of each other in expression or function. Further, most maize lines examined to date carry heterogeneous complements of nonfunctional MuDR elements. Although these elements are not expressed in inactive lines, it was not known whether or not they are transcriptionally active in active lines.
Here we describe the structure, expression characteristics, and behavior of three deletion derivatives of MuDR-1, the single MuDR element resident in our low copy minimal Mutator line. This line is exceptional in that it carries a single functional MuDR element (MuDR-1) or derivatives of that element, making it possible to examine the expression characteristics and functions of individual deletion derivatives in relative isolation.
MATERIALS AND METHODS
Maize lines: The a1-mum2 allele was originally isolated by D. S. Robertson and the line containing a1-mum2 in this study was obtained from S. Dellaporta. The a1-mum2 allele, which has a Mu1 insertion in the promoter of the a1 gene, is Musuppressible (Chometet al. 1991). Plants lacking MuDR-1 express red anthocyanin color, whereas plants carrying MuDR-1 show small red clonal sectors of tissue on a green background; kernels show small red clonal sectors on a pale yellow background. The suppression of the a1-mum2 allele is not readily apparent in the aleurone layer of the kernel, so that in the absence of MuDR-1 the kernel is pale yellow. The a1-dt sh2 tester line was obtained from B. McClintock. The a1-dt allele has a stable, colorless phenotype in the absence of Dt. All lines used in this report lacked the Dt element and for this reason we simply refer to the a1-dt allele as a1. Because the Sh2 gene is tightly linked (0.2 map units) to a1-dt, the sh2 allele serves as a marker for the presence of a1-dt. The tester, which lacks Mutator activity, will be referred to as the a1 sh2 tester. Progeny segregating for the MuDR-1 or dMuDR element(s) were obtained from the following crosses: +MuDR-1 (or dMuDR); a1-mum2 Sh2/a1 sh2 X a1 sh2/a1 sh2; –MuDR-1, -dMuDR or +MuDR-1 (or dMuDR); a1-mum2 Sh2/a1 sh2 X a1-mum2 Sh2/ a1 sh2; –MuDR-1, –dMuDR.
Nucleic acid samples: Maize DNA from leaves of 2-wk-old seedlings was purified according to Cocciolone and Cone (1993). RNA was extracted from immature ears according to Kloeckener-Gruissem et al. (1992), with the following modification: after the RNA was precipitated using lithium chloride, the supernatant from the subsequent centrifugation was saved. DNA from that supernatant was precipitated using 1/10 volume sodium acetate and 1 volume 2-propanol. This DNA was used to confirm the condition of the MuDR elements in the same samples as were analyzed by RNA gel blots.
DNA and RNA gel blot hybridizations: DNA hybridizations were performed as described in Chomet et al. (1991). The DNA blots were washed in 0.2× SSPE, 0.1% SDS at 65°. RNA gel blot hybridizations were performed under the same conditions as the DNA blot hybridizations. However, the RNA blots were washed in 2× SSPE, 0.1% SDS at 65° for 1.5 hr.
Reverse transcriptase-based PCR amplification: For the RT-PCR experiment, ∼10 μg of RNA from various samples was incubated with 1 unit of DNaseI and 5.6 units of RNase inhibitor for 30 min at 37°. Following DNaseI digestion the samples were extracted once with phenol/chloroform, precipitated in 2-propanol, and then resuspended in water to a concentration of ∼1 mg/ml. The resulting RNA solution (2 μg) was added to a reaction containing reverse transcriptase buffer, 10 mm DTT, primer 8f at a concentration of 0.5 mg/ml, dTTP, dATP, dCTP, and dGTP each at a concentration of 10 mm, and 2.8 units of RNase inhibitor to a total volume of 20 μl. This mix was incubated for 10 min at 70°, then put on ice for 5 min. Reverse transcriptase (2 units) was then added, and the reaction was incubated for 2 hr at 37°. The volume was then adjusted to 100 μl using water, and 5 μl of this mixture was used in a PCR reaction using primers 7f and 1r.
PCR primers and amplification conditions: The PCR primers for cDNA synthesis, PCR amplification, and sequencing included the following sequences: 1r = CTCTGCTCCTG TGCGGATGGATTGTCC; 4r = CTGTTTTCTGTGTTGTTG AG; 7f = CACCACAAAAAATTGGATCCC; 8f = CACCAAT AGCTAGAAGAGGTC.
PCR reactions were performed using ∼50 ng of genomic DNA or cDNA. The final concentrations of reagents in the 50-μl reactions were as follows: 200 mm dNTP, 1.5 mm MgCl2, 50 mm KCl, 10 mm Tris-HCl (pH 8.3), and 0.4 mm of each primer. Amplification of genomic and cDNA sequences used the following conditions: melting, 94° for 1 min; annealing, 50° for 1 min; extension, 72° for 2 min. Samples were subjected to 35 cycles of amplification.
DNA sequencing and analysis: Gel-isolated RT-PCR products were purified using the QIAquick Gel Extraction Kit (QIA-GEN Inc., Chatsworth, CA) after gel electrophoresis, and the resulting DNA was used for direct sequencing. DNA samples were sequenced at the DNA sequencing facility at the University of California at Berkeley using the double-stranded dye termination technique on an ABI sequencer (Applied Biosystems, Foster City, CA). Sequencing of the derivative d202 was performed using PCR primers 7f, 1r, and 4r as primers in the sequencing reactions.
Immunoblot and immunolocalizations: Immunoblot and immunocalizations were performed as described in Donlin et al. (1995). In each case in which a deletion derivative was subjected to these analyses, DNA from tissue samples from the same individual was subjected to DNA gel blot analysis to confirm the presence of the derivative. As a negative control, in each case either a sibling lacking the derivative or a closely related individual that lacked MuDR or dMuDR elements was also examined using each of the two methodologies.
RESULTS
MURB alone is not sufficient to cause excisions or suppression of a suppressible allele: Restriction mapping revealed that derivative MuDR-d112 (d112) carries only the sequences encoding mudrB (Figure 1). This derivative lacked an internal SacI fragment, consistent with the lack of one terminal inverted repeat. An EcoRI digest probed with a 5′ flanking probe (not shown) demonstrated that the EcoRI site is present in this derivative. The derivative lacked the 5′ XbaI site as well as both of the BamHI and HindIII sites. On the basis of the position of those sites in the flanking region outside of the element (not shown), we conclude that the deletion extends through one TIR to a point between the second BamHI site (which is missing in the derivative) and the 3′ XbaI site (which is present). As expected on the basis of the restriction map, the derivative fails to hybridize to probe H/H (data not shown), consistent with the lack of this internal fragment. On the basis of these results, we conclude that derivative d112 carries a deletion that includes nearly all of the region encoding mudrA as well as one of the two TIRs and ∼1.5 kbp of flanking sequence (Figure 1). On the basis of restriction mapping data as well as Northern blot analysis (below) we conclude that this deletion does not extend into the region of MuDR encoding mudrB. Derivative d112 is at the same chromosomal position as the full-length element first cloned on chromosome 2L (Chometet al. 1991).
Northern analysis reveals that only the mudrB transcript is expressed at detectable levels in individuals carrying this derivative (Figure 2). Immunolocalization also reveals that MURB is successfully translated and localized to the nucleus in the absence of MURA (Figure 3). Further, MURB is most prevalent in actively dividing cells, a pattern of localization that is very similar to that observed previously when this protein was produced from a full-length element (Donlinet al. 1995).
—Restriction mapping or sequencing data of three dMuDR elements. In the representation of d112, the dashed line represents the region in which the 3′ breakpoint of the deletion is found. In the representation of d201, the dashed line represents that region in which a deletion of 500 bp is found. The precise location of the deletion junction of derivative d202 is as indicated. Transcribed regions mudrA (left) and mudrB (right) are as indicated (from Hershbergeret al. 1995). ORFs encoding MURA and MURB are also as indicated. The shaded boxes on the ends represent the terminal inverted repeats. S, SacI; Sa, SalI; E, EcoRI; Xb, XbaI; EII, EcoRII; B, BamHI; H, HindIII. Solid bars above the restriction map indicate probes referred to in the text. PCR primers, the sequences of which are given in materials and methods, are indicated by small arrows below and above the map. Polyadenylation sites are designated A+; translational stop codons are designated with asterisks. a, mudrA; b, mudrB.
Individuals that carry derivative d112 show no signs of Mutator activity. Kernels carrying a1-mum2 and this derivative do not exhibit excisions, and plants grown from these kernels are red, consistent with the lack of suppression of a1-mum2. On the basis of these observations we conclude that although it can be transcribed, translated, and translocated in the absence of MURA, MURB by itself is not associated with these aspects of Mutator activity.
MURA alone is sufficient to cause excision of a reporter element: Derivative d201 lacks a large portion of the gene encoding MURB. This derivative contains a deletion of 500 bp of the 861 bp between the EcoRI site and the 3′ EcoRII site in MuDR (Figure 1). Since the element lacks the 3′ XbaI site, we conclude that the deletion must remove nucleotides encoding at least 54 amino acid residues from the carboxy-terminal portion of MURB as well as nearly all of the 3′ untranslated portion of the mudrB transcript and, at least potentially, a portion of the intergenic region. Derivative d201 is directly derived from the fully functional element first cloned on chromosome 2L. However, because d201 was derived from a transposed copy of that MuDR element, it is not located at that chromosomal position. Although some kernels display a Mu1 excision frequency comparable to that observed in kernels carrying full-length elements, in most cases, kernels carrying d201 exhibit few or no excisions of a reporter Mu1 element from a1-mum2 (Figure 4). This is also the case for the excision of a Mu7 element from bz-rcy (data not shown).
—RNA gel blot analysis of deletion derivatives. When two probes were used, as in the left-hand section of each part, RNA from a plant carrying the indicated element was probed with Mu*. In the right-hand section of these parts the same sample was probed with Mu-Sal. As a loading control, the bottom of each part is a photograph of the ethidium-stained samples before transfer. +, an individual carrying the indicated element; –, a sibling lacking the element.
Northern analysis of d201 demonstrates that although the mudrA transcript is present, the mudrB transcript is not detectable, indicating that the deleted region in d201 is required for the production of detectable levels of this transcript (Figure 2). Further, an antibody specific to MURB failed to detect the protein encoded by the small transcript (Figure 5). These data confirm that MURA by itself can cause excisions of a reporter element in the absence of detectable levels of mudrB transcript or MURB protein, demonstrating that MURB is not required for this function. They also demonstrate that production of MURA is not dependent on the presence of MURB.
A deletion derivative lacking mudrB is unable to prevent methylation and appears to be unable to transpose: HinfI sites within the termini of Mu elements are characteristically methylated in the absence of Mutator activity (Bennetzen 1987). To determine the effect of the presence of d201 on the status of those sites within the Mu1 element at a1-mum2, DNA from plants carrying d201 and siblings lacking the element were digested with HinfI, blotted, and probed with a fragment of the A1 gene (Figure 6). Successful digestion of the sites results in a fragment of 0.6 kb. A lack of digestion of both sites results in a 2.1-kbp fragment. Additional fragments present in only a subset of the samples (indicated by “a1-dt”) are those associated specifically with the a1-dt allele of A1, which was segregating in the families examined. In each case examined, the Mu1 at a1-mum2 was modified to the same degree in plants carrying d201 and siblings lacking the element, indicating that the presence of d201 does not alter the methylation status of this element. This was the case even though the plant containing d201 depicted in Figure 6 was grown from a kernel exhibiting a high frequency of Mu1 excisions. This suggests that MURA is unable by itself to prevent modification of nonautonomous elements.
Despite the sometimes-frequent excisions conditioned by d201, in no case has this element been observed to transpose in the absence of a full-length MuDR-1 element. This was the case even when kernels carrying d201 exhibited a high frequency of excisions of Mu1 from a1-mum2. A total of 45 plants grown from such high excision kernels were test crossed as females to a1-mum2/a1 sh2 testers. In the resulting ears the ratios of highly spotted to weakly spotted kernels were variable and quite similar to like ratios on the ears from which their parents were picked. Further, none of the resulting ears segregated for significantly more than 50% spotted kernels, suggesting that the element had not transposed (data not shown).
Molecular analysis of plants carrying d201 gave a similar result; no transposition of the element was observed. In order to assay for changes in copy number of d201, families segregating for the element were digested with SacI and EcoRI. Duplications of the element were detected as changes in the intensity of the internal 4.2-kbp SacI fragment or in the appearance of unique new EcoRI fragments. In the family portrayed in Figure 7, a family generated from a cross between a plant carrying d201 and an a1-mum2/a1 sh2 tester was subjected to Southern blot analysis. DNA from plants grown from spotted and pale kernels was extracted, digested with SacI, blotted, and probed with Mu*. As expected, all of the plants grown from spotted kernels carried d201. Further, a number of plants grown from nonspotted kernels also carried the element, indicating that d201 is not always able to cause somatic excision of Mu1. DNA from a total of 64 plants carrying d201 was digested with SacI and probed with Mu*. In all cases the 4.2-kbp internal fragment remained at a relative intensity, consistent with the presence of a single copy of the deletion derivative. Further, EcoRI digests of the DNA from 37 of these plants did not reveal the appearance of unique new fragments, consistent with a lack of duplicative transposition of d201. Finally, 84 progeny of individuals carrying d201 and at least one Mu1 element (the one at a1-mum2) that had been crossed to a1-mum2/a1 sh2 testers were examined by Southern blot for the presence of new Mu1 fragments. In no case was Mu1 observed to transpose. In contrast, the full-length MuDR-1 element at position 1 has been observed to duplicate itself and a Mu1 element at a frequency of between 10 and 20%, which was itself an exceptionally low transposition frequency for a MuDR element (Chometet al. 1991; Lischet al. 1995). These data suggest that d201 transposes at a frequency that is significantly lower than that reported for MuDR-1 at position 1 (P < 0.01). Indeed, it may be altogether incapable of causing transposition.
—Immunolocalization of MURB in a developing tassel. (a) Localization in the tassel primordium of a plant that is hemizygous for d112 in the minimal line. (b) Localization of d112 in a tassel primordium at lower magnification. (c) Localization in an a1-mum2/a1 sh2 tester that lacks MuDR-1 or its derivatives.
—Examples of the range of kernel excision frequencies of Mu1 from a1-mum2 associated with deletion derivatives (A) d201 and (B) d202.
Autonomous transposition of a deletion derivative: In contrast to all other derivatives examined to date, one deletion derivative, d202, is unique in that it has been observed to transpose autonomously. Like d201, derivative d202, which was initially identified because of its association with very low frequency excisions, can also condition a more typical Mu1 excision frequency (Figure 4). The same was true when this element was combined with bz-mum9, which carries a Mu1 element in the bronze gene (data not shown). Sequence analysis (see below) revealed that this derivative carries a deletion of 174 bp between and including residues 3094 and 3267 (Figure 1).
—Immunoblot analysis of derivative dMuDR-201 in the minimal line. Lane 1, a plant carrying MuDR-1. Lane 2, a plant carrying dMuDR-201. Lane 3, a sibling of the plant analyzed in lane 2 that carries no MuDR or dMuDR derivatives.
—(A) DNA gel blot of HinfI digests of DNA from plants carrying deletion derivatives d201 and d202. The blot was probed with a KpnI fragment from the A1 gene. Lane 1, d201 in a plant grown from a heavily spotted kernel. Lane 2, a sibling lacking d201. Lanes 3 and 4, two plants grown from heavily spotted kernels carrying d202. The fragments labeled a1-dt are those associated specifically with the a1-dt allele of A1, which was segregating in these families. (B) A restriction map of a1-mum2. The bars above the map indicate probes used in the analysis. K, KpnI; Hf, HinfI.
To determine if d202 could transpose, plants carrying the element were digested with HindIII and SacI. A transposition event would be expected to result in the presence of a new, unique HindIII fragment and an increased intensity of the diagnostic 4.52-kbp SacI fragment. Figure 8 shows a pair of DNA gel blots from individuals carrying d202. The DNA was digested with HindIII (Figure 8A) and with SacI (Figure 8B) and the resulting blot was hybridized to Mu*. The HindIII digests (Figure 8A) revealed the presence of a new MuDR-1-homologous fragment in DNA from one of these plants (indicated). Densitometric analysis of the SacI digests of DNA from that plant revealed that the diagnostic 4.52-kbp fragment exhibited a twofold increase in intensity, consistent with the duplication of d202 in this individual. When this plant was test crossed to an a1 sh2/a1-mum2 tester, ∼3/4 (99/120) of the progeny kernels carrying a1-mum2 were spotted, consistent with the independent segregation of two d202 elements.
—DNA gel blot of a family segregating for derivative d201. The first lane represents a related individual carrying two copies of MuDR-1.
Deletions can result in the production of antisense message: Northern analysis demonstrated that several transcripts were associated with d202 (Figure 2). Two aberrant transcripts were visible, one 1.2 kbp in length and the other 2.0 kbp. Both aberrant transcripts hybridized to probes from regions encoding each of the two ORFs (Figure 1, probes Mu-Sa and Mu*). Because the two transcripts are produced in opposite directions and from opposite strands, the presence of a 2.0-kbp transcript homologous to regions encoding both genes was consistent with the production of a hybrid transcript containing the sense version of one transcript and the antisense version of the other transcript. To test this hypothesis, reverse transcriptase-based PCR (RT-PCR) was used to specifically amplify the transcript expected to be produced if transcription were initiated from the mudrB promoter and continued through the intergenic region and into sequences normally expected to encode mudrA. This aberrant transcript would be expected to contain sequences identical to the normal mudrB transcript and sequences that are complementary to the 3′ end of the normal mudrA transcript.
—DNA gel blot of (A) HindIII and (B) SacI (B) digests of DNA from plants carrying d202. In both cases the blots were hybridized to Mu*. In A, the 6.5-kbp fragment is that previously identified as d202 at the position at which it was first identified. The arrow indicates the presence of a new MuDR-1-homologous fragment in lane 2. In B, the 4.52-kbp SacI fragment is the size expected for d202. The arrow indicates the fragment in lane 2 that is twice as intense as the same fragment in the other lanes.
To specifically amplify antisense product, both RNA and DNA from individuals carrying d202, and from siblings that lacked the derivative, were isolated. RNA and DNA from an individual carrying a full-length copy of MuDR-1 were also isolated. The RNA from each of these samples was digested with DNaseI, and DNA was synthesized from the RNA samples using primer 8f (Figure 1) to direct an initial round of first-strand cDNA synthesis using reverse transcriptase. Because this initial round of DNA synthesis used a primer that was identical to a sequence in the normal mudrA transcript, it would only be expected to result in a product if a transcript complementary to the normal transcript were present. The resulting product was subjected to PCR amplification using a nested primer (primer 7f, Figure 1) identical to a sequence in normal mudrA transcript and a second primer that is complementary to a sequence in the normal mudrB transcript (primer 1r, Figure 1). Controls included individual siblings that lacked the derivative as well as closely related individuals that carried only a full-length MuDR element. Each sample was also amplified either with or without DNaseI digestion.
Given that there are a number of MuDR-homologous sequences in this genetic background, we expected to see some amplification from all DNA samples not subjected to DNaseI digestion. As expected, in those samples a number of fragments resulted from our PCR amplification (Figure 9, lanes 3, 6, 9, and 11). Among these fragments, one was unique to the DNA sample from the individual carrying d202 (lane 3, arrow). This fragment was missing in a sibling lacking the derivative (lane 6) and in a related individual carrying a full-length element (lane 9); its size was that expected for a fragment amplified from the deletion derivative (∼1500 bp). The fragment unique to the individual carrying the full-length MuDR-1 element (lane 9, arrow) was of the size (∼1700 bp) expected for the amplification of this region from a full-length element.
To demonstrate that DNaseI digestion was complete and that our PCR reactions were not simply amplifying residual DNA, we amplified DNA samples that had been subjected to DNaseI digestion but had not undergone a round of cDNA synthesis. No product was amplified from these samples (lanes 2, 5, and 8). Similarly, when the same PCR reactions were performed on independently derived samples of MuDR-1-containing genomic DNA that had also been subjected to DNaseI treatment, no product was amplified (lane 10).
—RT-PCR analysis of transcript associated with d202. Lane 1, RNA from a plant carrying d202, digested with DNaseI, reverse transcribed, and then PCR amplified. Lane 2, the same RNA sample digested with DNaseI, not reverse transcribed, and then PCR amplified. Lane 3, DNA from the same plant amplified using the same PCR primers. Lane 4, RNA from a sibling that lacked the deletion derivative digested with DNaseI, reverse transcribed, and then PCR amplified. Lane 5, RNA from the same plant as is in lane 4 digested with DNaseI, not reverse transcribed, and then PCR amplified. Lane 6, DNA from the same plant PCR amplified. Lane 7, RNA from a plant carrying MuDR-1(p1) digested with DNaseI, reverse transcribed, and PCR amplified. Lane 8, the same sample digested with DNaseI, not reverse transcribed, and then PCR amplified. Lane 9, DNA from the same plant PCR amplified. Lane 10, DNA from a different plant also carrying MuDR-1(p1) digested with DNaseI and PCR amplified. Lane 11, DNA from the same plant as in lane 10 not digested with DNaseI and then PCR amplified. Lane 12, a water-only PCR amplification negative control for lanes 1–11.
Most importantly, of the RNA samples that had undergone DNaseI digestion, cDNA synthesis, and PCR amplification, only the one from the individual carrying d202 yielded a product (Figure 9, lane 1, arrow). That product was the size predicted if amplification was of the aberrant transcript associated with the deletion derivative.
To more precisely determine the nature of this aberrant transcript, the PCR product produced from the aberrant transcript associated with d202 was partially sequenced. The resulting sequence was identical to that reported for MuDR (Hershberger et al. 1991, 1995), with the exception of a deletion of 174 bp between and including residues 3094 and 3267. Because amplification was successful using primer 1r, we conclude that the transcript must extend at least as far as residue 4494 in the 3′ direction. As has been observed in a number of other deletion derivatives, the junction of the deletion includes one copy of a short, direct repeat (AAGAA in this case) found in the original sequence. Surprisingly, the deletion also removed the last 17 predicted amino acids of MURA as well as the normal translational stop codon. The derivative retained the normal mudrA polyadenylation sites, most of the 3′ untranslated region, and the entire intergenic region (Hershbergeret al. 1995).
This 2.0-kbp aberrant transcript included both the second and the third intron from mudrB, suggesting that neither of these introns is efficiently spliced from the aberrant transcript. In contrast, although the third intron of the normal mudrB transcript is only rarely spliced (<10%) from full-length elements, the second intron is normally spliced with high efficiency (Hershbergeret al. 1995).
On the basis of this analysis, it appears that the larger of the two aberrant transcripts associated with d202 contains nucleotide sequences that are identical to the 3′ end of mudrA and sequences that are complementary to at least part of mudrB, presumably due to transcription initiating from the promoter associated with mudrB and proceeding past its normal transcriptional stop signals into the region encoding the large transcript. As a consequence of this read-through, d202 produces an antisense mudrA mRNA. This derivative is active by both the excision (Figure 4) and the transposition assays (Figure 8). However, as is apparent in Figure 4, excisions driven by d202 were often less frequent and consistently more variable than those of d202's progenitor, MuDR-1. Finally, in contrast to the results obtained for d201, Mu1 at a1-mum2 is incompletely hypomethylated in the presence of d202, even in DNA from plants grown from heavily spotted kernels (Figure 6). This is most apparent when the plant carrying d202 was homozygous for a1-mum2.
DISCUSSION
Specific deletion derivatives are associated with specific transcripts: Although a large number of derivatives of MuDR have now been reported (Chometet al. 1991; Hershbergeret al. 1995; Lischet al. 1995; Hsia and Schnable 1996), the minimal line in which the above elements have been described provides some unique advantages. First and most importantly, the only transcripts observed in this line are those associated with a single previously cloned MuDR element or its derivatives (Figure 2). Thus, for the first time it is possible to associate specific transcripts with specific deletion derivatives and to ascribe specific roles to those transcripts. In addition, the use of the minimal line makes it possible to determine the source of aberrant transcripts, such as truncated or read-through messages. In more complex Mutator backgrounds, expression analysis of deletion derivatives has not been possible because of the presence of large numbers of additional, transcriptionally active elements. Finally, the minimal line used in these experiments is also well inbred (>10 generations) and so variations in various aspects of Mutator activity, such as excision frequency or suppression of reporter element, are most likely due to changes in the Mutator system itself rather than background effects.
MURB is not by itself associated with Mutator activity: Previous work (Lisch and Freeling 1994; Hsia and Schnable 1996) had demonstrated that deletions within the region encoding mudrA eliminated excision and suppression function, but it was not known how these deletions effected transcription from MuDR. The deletion in d112 extends through all of mudrA, but mudrB remains intact. Transcript analysis revealed the presence of a 1.0-kbp transcript that hybridizes with a mudrB-specific probe and that cosegregates with this derivative (Figure 2). Immunoblot analysis revealed that this transcript is translated and the protein is translocated to the nucleus primarily in mitotically active cells (Figure 3). These data demonstrate that expression of MURB is largely independent of MURA.
Because d112 is not associated with either excision of Mu1 from a1-mum2 or suppression of this suppressible allele, it does not appear that MURB is sufficient by itself to condition any aspect of Mutator activity. Of course, as we have not sequenced the entire derivative, we cannot say for certain that the small protein associated with d112 is not altered in some subtle way that has destroyed its normal function. However, we note that additional derivatives that carry large deletions in the regions encoding mudrA, but that can produce large quantities of mudrB transcript, are also incapable of causing excisions of reporter elements (D. Lisch, unpublished results; Lisch and Freeling 1994; Hsia and Schnable 1996), lending weight to the argument that MURB is not by itself capable of conditioning these events.
MURA is sufficient to cause excisions of a reporter element: Deletion derivative d201 is missing a large portion of mudrB. Although it contains a portion of the region encoding mudrB, including the promoter and all of the first exon, at a minimum, it lacks sequences including the third exon, the termination codon, and it may lack sites that are normally polyadenylated (Hershbergeret al. 1995) as well as a portion of the intergenic region. Derivative d201 appears to produce only mudrA transcript and is not associated with detectable amounts of mudrB transcript (Figure 2) or MURB (Figure 3). Our Northern blot analysis used total RNA, so there is the possibility that very small quantities of RNA are still produced. However, the lack of detectable truncated mudrB transcript suggests that in the absence of the deleted sequences mudrB transcript is produced in very small (or nonexistent) quantities compared to the normal level of expression. Alternatively, the expected truncated transcript may be unstable. Given the continued presence of the appropriate promoter elements in the Mu termini, the latter explanation is perhaps more likely. In either event, the net result is the lack of detectable MURB protein.
The observation that d201 is capable of causing excisions of Mu1, sometimes at a frequency approaching that associated with full-length elements (Figure 4), demonstrates that MURB is not required for excision of Mu1 elements. However, although it is capable of conditioning a high frequency of Mu1 excisions, d201 is not associated with a high or even moderate frequency of new insertions. An attractive explanation for this observation is that MURB is required for reinsertion but not excision of Mu elements. Because these data represent a negative result that may reflect only a reduced frequency of these events, it is possible to surmise only that d201 is significantly less efficient than a full-length MuDR-1 element at conditioning new insertions. Further, because d201 is at a new position, we cannot exclude the possibility that position effects are playing a role in the observed variations in the ability of this element to cause excisions of Mu1 (Figure 4).
In connection with our observation that MURA has independent activity, it is interesting to note that although genome database searches using mudrA sequences detect similar sequences in Arabidopsis and rice, there are no known sequences that share significant similarities to mudrB (D. Lisch, unpublished results). As the functional analysis of these sequences has not been performed, it is not known if these represent active elements. However, sequences previously identified as being similar to bacterial transposases are also conserved in the elements from Arabidopsis and rice, as is a conserved zinc finger domain, indicating functional similarities to mudrA (D. Lisch, unpublished results). It may be that mudrA and mudrB evolved independently and that mudrB function is specific in some way to the requirements of the Mutator system in maize. Alternatively, mudrB may have been lost due to internal deletions in these other species, or mudrB sequences simply may have diverged more due to less stringent selection for their preservation. It will be interesting to determine whether or not the elements in these heterologous species have full transpositional activity in the absence of mudrB.
The large variations in Mu1 excision frequency associated with d201 are as yet unexplained. The variation is quite sporadic; it does not segregate in a manner consistent with the presence of a few modifying loci, nor are high-frequency excisions heritable (D. Lisch, unpublished results). Alternatively, it may be that d201 is subject to some degree of epigenetic modification, and that variations in epigenetic modification of the element could explain the observed variations in excision frequency of Mu1. We also cannot exclude the possibility that there are additional factors in our minimal line that modify the activity of d201. We believe this is unlikely, however, given the extremely inbred nature of the line, as well as the absence of evidence for genetic segregation of variations in excision frequency. Regardless of the excision frequency associated with it, d201, like d112, is not associated with suppression of a1-mum2, suggesting that MURB may be necessary but not sufficient for suppression function. Nor is d201 associated with hypomethylation of Mu1 at a1-mum2 (Figure 6), suggesting some role for MURB in that process. Previous work has demonstrated that deletions within mudrA are associated with hypermethylation of Mu1 elements, suggesting that MURB is not sufficient to mediate hypomethylation of nonautonomous elements (Lischet al. 1995).
A deletion can result in the production of antisense message: Our analysis of the transcripts associated with d202 demonstrates one mechanism by which the previously reported MuDR antisense transcript can be produced (Hershbergeret al. 1995; Joaninet al. 1997). Four distinct transcripts are produced by this derivative: two transcripts characteristic of the full-length element and two aberrant transcripts (Figure 2). The larger aberrant transcript is composed of both sense mudrB and antisense mudrA sequences. Surprisingly, read-through from the mudrB promoter occurs despite the fact that the deletion is in the region containing stop signals not for mudrB, but for mudrA. Despite this, the mudrA transcript is present and is of the correct size (Figure 2), and it is transcription from the mudrB promoter that appears to extend past its normal stop signals. This result suggests that there are sequences in the transcribed 3′ region of mudrA that are necessary for the normal transcription of mudrB.
The region between the translational stop signals in the two normal transcripts is highly repetitive, and has a high potential for secondary structure (Hershbergeret al. 1995). It may be that in d202 disruption of this secondary structure has effected the normal termination of transcription of mudrB. Whatever the cause, the result is a read-through into mudrA and thus the production of antisense message.
Antisense message has been detected before in Mutator lines (Hershbergeret al. 1995). Here a particular antisense message is linked with a particular Mu element. All previously reported antisense messages were detected in lineages with multiple MuDR and dMuDR elements. In our minimal line this antisense message was not detected in individuals carrying only MuDR-1, suggesting that expression of antisense message (at least that complementary to primer 8f) is not necessarily associated with the normal function of full-length MuDR elements in the tissues analyzed.
Because the antisense message is at least potentially capable of interfering with the normal sense message, it has been suggested that the presence of antisense message may be indicative of a negative regulatory pathway (Hershbergeret al. 1995; Joaninet al. 1997). Because the excision frequency associated with d202 varies considerably (Figure 4B), and it appears that both normal and antisense transcripts are produced by this derivative (Figure 2), an intriguing possibility is that the observed variation may be due to changes in the relative quantities of normal vs. antisense transcript. If this were the case, then our expectation would be that the relative quantity of the antisense version of the transcript would be higher in those aleurones that show lower frequencies of Mu1 excisions. However, analysis would be complicated if the presence of antisense transcript resulted in post-transcriptional silencing of the element.
Alternatively, variations in excision frequency associated with d202 could be a result of direct changes in MURA. The deletion includes the last 17 predicted amino acids of MURA as well as the normal translational stop codon. Some of the residues that are lost are similar to nuclear localization signals in other plant genes (Hickset al. 1995). Thus, reductions in excision frequencies associated with d202 may be the result of reductions in the effective quantity of MURA within the nucleus. Changes in copy number of d202, as can be observed in Figure 8, could result in increased overall levels of protein that could compensate for a reduced efficiency of translocation. Indeed, the progeny of the individual carrying the transposed copy of d202 showed a marked increase in excision frequency (data not shown). This hypothesis would suggest that this derivative should show a marked decrease in the extent of nuclear localization of MURA compared with a fulllength element and that changes in excision frequency should be accompanied by quantitative changes in the level of MURA protein. The recent isolation of MURA protein in yeast (Benito and Walbot 1997) should make it possible to test these hypotheses directly.
Acknowledgments
We thank Vicki Chandler and David Selinger for critical reading of the manuscript. This work was supported by a National Science Foundation grant (MCB9219587) to M.F. D.L. was supported by a National Institutes of Health (N.I.H.) Training Grant (GM-07127-19). M.J.D. was supported by an N.I.H. postdoctoral fellowship (F32GM15870).
Footnotes
-
Communicating editor: J. A. Birchler
- Received November 15, 1996.
- Accepted October 5, 1998.
- Copyright © 1999 by the Genetics Society of America
