A Test for Transvection in Plants: DNA Pairing May Lead to trans-Activation or Silencing of Complex Heteroalleles in Tobacco

Corresponding author: Marjori Matzke, Institute of Molecular Biology, Austrian Academy of Sciences, Billrothstrasse 11, A-5020 Salzburg, Austria., mmatzke{at}imb.oeaw.ac.at (E-mail)

Communicating editor: J. A. BIRCHLER

	ABSTRACT

TOP ABSTRACT MATERIALS AND METHODS RESULTS DISCUSSION LITERATURE CITED

To study whether DNA pairing that influences gene expression can take place in somatic plant cells, a system designed to mimic transvection was established in transgenic tobacco. Pairing was evaluated by testing whether an enhancerless GUS gene on one allele could be activated in trans by an enhancer on the second allele. The required heteroalleles were obtained at four genomic locations using Cre-lox-mediated recombination. In one transgenic line, elevated GUS activity was observed with the heteroallelic combination, suggesting that trans-activation occurred. Conversely, when the unaltered allele was homozygous, GUS activity dropped to hemizygous levels in a silencing phenomenon resembling dosage compensation. Double-stranded GUS RNAs or small GUS RNAs indicative of RNA-based silencing mechanisms were not detected in plants displaying reduced GUS activity. These results suggested that a transgene locus capable of pairing, as revealed by trans-activation, could also become silenced in an RNA-independent manner, thus linking DNA pairing and gene silencing. The transgene locus was complex and comprised an inverted repeat, which possibly potentiated allelic interactions. The locus was unable to trans-activate transgenes at ectopic sites, further implicating allelic pairing in the transvection effects.

HOMOLOGY-DEPENDENT gene silencing (HDGS) refers to a type of epigenetic inactivation that is based on recognition of nucleic acid sequence identity at either the DNA or RNA level. HDGS phenomena have been described in diverse organisms and are probably common to most eukaryotes (GARRICK et al. 1998 ; BIRCHLER et al. 2000 ; FAGARD and VAUCHERET 2000 ; FAUGERON 2000 ; WIANNY and ZERNICKA-GOETZ 2000 ). Cosuppression, quelling, and RNAi are all terms that have been used to refer to a post-transcriptional gene silencing (PTGS) process in which double-stranded (ds) RNA induces the degradation of homologous RNAs in the cytoplasm (COGONI and MACINO 2000 ; MEINS 2000 ; PLASTERK and KETTING 2000 ). Other HDGS effects act at the genome level to trigger transcriptional gene silencing and DNA modifications. There is evidence from various organisms suggesting that sequence-specific DNA methylation can be triggered by DNA-DNA or RNA-DNA interactions. RNA-directed DNA methylation has been documented so far only in plants (WASSENEGGER 2000 ; MATZKE et al. 2001 ) but could conceivably be involved in mammalian X-chromosome inactivation and some cases of parental imprinting that involve overlapping sense/antisense RNAs and DNA methylation (WOLFFE and MATZKE 1999 ).

Evidence that DNA-DNA interactions can trigger the modification of linked and unlinked sequence duplications is provided most convincingly by the repeat- induced point mutation (RIP) and methylation induced premeiotically (MIP) phenomena observed, respectively, in the filamentous fungi Neurospora crassa (SELKER 1997 ) and Ascobolus immersus (FAUGERON 2000 ). RIP and MIP take place at a specific stage in the sexual cycle of these fungi and probably do not have strict counterparts in other organisms. Nevertheless, these specialized processes have provided precedents for the involvement of DNA sequence interactions in triggering epigenetic silencing and genome modifications in higher eukaryotes.

Pairing-dependent genetic phenomena on the chromosomal level have long been known to occur in Drosophila. The general terms transvection or trans-sensing effects are used to refer to cases in which homolog pairing influences gene expression (HENIKOFF and COMAI 1998 ; PIRROTTA 1999 ; WU and MORRIS 1999 ). Different types of transvection effects involving either gene activation or silencing in trans have been reported in Drosophila (HENIKOFF and DREESEN 1989 ; GOLDSBOROUGH and KORNBERG 1996 ; MORRIS et al. 1999 ). Such pairing-dependent phenomena are not surprising in this organism, where stable associations of homologous chromosomes are observed in somatic cells. Cytologically visible pairing does not appear to be the rule in most mammalian and plant cells. However, persistent pairing might not be necessary for allelic cross talk (TARTOF and HENIKOFF 1991 ) or for transmission of epigenetic states between homologous sequences (COLOT et al. 1997 ). Pairing thus remains an attractive hypothesis to explain certain epigenetic silencing effects in mammals and plants. Several cases of parental imprinting in mammals have been associated with pairing of chromosomal domains (LASALLE and LALANDE 1996 ; DUVILLIE et al. 1998 ) and chromosome pairing has been proposed to play a role in mammalian X-chromosome inactivation (MARAHRENS 1999 ).

In plants, DNA-DNA pairing has long been postulated to occur in various cases of HDGS that involve inactivation and/or methylation of repeated sequences on different chromosomes (JORGENSEN 1992 ; MATZKE and MATZKE 1995 ; BAULCOMBE and ENGLISH 1996 ; HOLLICK et al. 1997 ; BENDER 1998 ). Pairing is suggested by reports of coextensive methylation of unlinked homologous sequences, which indicates that sequence interactions trigger DNA modification (STAM et al. 1998 ; LUFF et al. 1999 ). However, because the involvement of RNA-directed methylation has not been completely ruled out in these cases (SELKER 1999 ), it is still uncertain whether pairing of homologous (trans)gene sequences actually occurs and leads to alterations of gene expression in somatic plant cells.

DNA-DNA pairing, particularly if it occurs transiently, might be difficult to detect in plant somatic nuclei by standard techniques such as fluorescence in situ hybridization (FISH). Moreover, FISH alone cannot establish whether any pairing interactions observed are functionally significant. To obtain a positive measure of pairing that leads to altered gene activity, we have set up a system to mimic transvection in transgenic tobacco. Heteroalleles of a 35SGUS reporter gene that either contain or lack a 35S enhancer were generated using Cre-lox-mediated site-specific recombination. The ability of the 35S enhancer on one allele to activate GUS expression in trans was then tested. In one tobacco line containing a large complex transgene locus, augmentation of GUS expression was observed with the heteroallelic combination, suggesting that pairing of transgene alleles occurred at this locus. As a possible consequence of pairing, plants homozygous for the 35SGUS unaltered allele exhibited gene silencing that did not appear to be mediated by dsRNA. These results suggest that allelic interactions capable of influencing gene expression in both positive and negative ways can take place in somatic cells.

	MATERIALS AND METHODS

TOP ABSTRACT MATERIALS AND METHODS RESULTS DISCUSSION LITERATURE CITED

Vector construction:
To create a 35SGUS chimeric gene that was manipulatable by site-specific recombination, the plasmid pEMBL-GUS-46 (BENFEY et al. 1989 ) was cut with BglII and a 35S enhancer (PIETRZAK et al. 1986 ) fragment flanked by synthetic lox sites (DALE and OW 1991 ) in direct orientation, and an FRT site (ODELL and RUSSELL 1994 ) was inserted. The orientation of the first FRT site was the same as for a second FRT site that was inserted at the end of the GUS-coding region-3C transcription terminator in pEMBL-GUS-46. This modified plasmid was linearized with BamHI and inserted into the BglII site of the BV4 binary vector (SCHERNTHANER et al. 1988 ). Further steps were carried out as described by MATZKE and MATZKE 1986 .

Plant transformation and reporter gene assays:
Leaf disk transformation and seed germination assays on kanamycin-containing medium to determine the number of segregating transgene loci were performed as described previously (MATZKE et al. 1989 ). Nopaline was detected using high-voltage paper electrophoresis and phenanthrenequinone staining (MATZKE et al. 1989 ). Fluorometric and histochemical GUS assays were performed on whole individual seedlings 6 weeks after germination as described previously (JEFFERSON 1987 ). The protein concentration in seedling extracts was determined using the Bradford assay (BRADFORD 1976 ). Assays were carried out for 1 hr at 37°. Statistical analyses were performed with the t-test using Microsoft Excel.

DNA blot analysis:
Plant DNA isolation and DNA blot analysis were carried out as described previously (MATZKE et al. 1989 ).

RNA analyses:
Total RNA was extracted from young expanding tobacco leaves using the Hybaid-AGS RNAClean system (Chemomedica, Vienna) including the RNAClean extension protocol according to the manufacturer's instructions. Northern blots were performed following standard protocols (METTE et al. 1999 ). Protocols for detection of double-stranded RNAs and small RNAs have been detailed by METTE et al. 2000 .

-cloning, rescue cloning, and nucleotide sequence analysis:
The plasmid pEMBL was included in the transgene construct to allow the transgene inserts to be recovered from the genomes of the tranformed tobacco lines by rescue cloning (PAPP et al. 1996 ). This procedure was successful with lines 17 and 19, which contained a complete T-DNA region. For lines 13 and 14, rescue cloning was unsuccessful and the respective transgene inserts were isolated by -cloning and sequenced as described previously (METTE et al. 1999 ).

	RESULTS

TOP ABSTRACT MATERIALS AND METHODS RESULTS DISCUSSION LITERATURE CITED

The construct used in these experiments contained a 35SGUS reporter gene positioned between selection and screening marker genes encoding kanamycin resistance (KAN) and nopaline synthase (NOS) activity, respectively (Fig 1A). The 35S enhancer was flanked by lox sites to allow its removal by the Cre recombinase (ODELL and RUSSELL 1994 ). The GUS coding region together with a minimal promoter (P-GUS) was flanked by FRT sites to permit excision by the FLP recombinase (KILBY et al. 1995 ). Heteroalleles that lacked either the 35S enhancer (35Senh) or the P-GUS gene (GUS) could thus be created in planta by site-specific recombination (Fig 1B). Theoretically, when the two heteroalleles are combined in the same genome, GUS activity should be obtained only when the 35S enhancer on one allele trans-activates the enhancerless P-GUS gene on the other allele and this would require physical pairing of the alleles [allelic transvection (A-TV); Fig 1B]. Pairing interactions between unlinked loci are also conceivable [ectopic transvection (E-TV)]. The flanking KAN and NOS genes provide extra sequence homology to promote DNA pairing interactions that could potentially affect the 35SGUS gene.

View larger version (19K): In this window In a new window Download PPT slide   Figure 1. (A) Transgene construct. A minimal promoter-ß-glucuronidase (GUS) gene, flanked by FRT sites, was positioned downstream of a 35S enhancer (35S) flanked by lox sites. To the right and left were genes encoding nopaline synthase (NOS) and resistance to kanamycin (KAN), respectively, both under the control of the NOS promoter (N). A bacterial origin of replication, pEMBL, was present in the construct and could be used to rescue clone transgene inserts from the tobacco genome. Arrowheads labeled L and R indicate the left and right T-DNA borders, respectively. Enzymes used to digest DNA for the blot analyses are indicated; bars show the probes used. Abbreviations are as follows: S, SacII; P, PstI; M, MspI or HpaII; E, EcoRI; N, NsiI; T, NOS transcription terminator; t, transcription terminator from the pea gene encoding ribulose bisphosphate carboxylase 3C gene (accession no. emb/XO4334/PSRBCS3C). (B) Transvection strategy: heteroalleles lacking either the 35S enhancer or the minimal promoter (P)-GUS gene were produced by Cre or FLP activity, respectively. Because neither heteroallele can produce GUS activity on its own, a positive GUS result can only be obtained in the heteroallelic combination if the 35SS enhancer on one allele trans-activates the minimal promoter on the other allele and this should require pairing of alleles. Diagonally hatched flanking regions represent sequence homology contributed by pEMBL and the KAN gene on the left and the NOS gene on the right.

The construct was introduced into the tobacco genome by leaf disk transformation. Four individual transformed lines (13, 14, 17, and 19) that expressed all three reporter genes (Fig 2 and data not shown) and that contained single independently segregating transgene loci were recovered. Cre-mediated removal of the 35S enhancer was achieved in all four lines (Fig 3A); successful excision of P-GUS by the FLP recombinase was obtained only with lines 14 and 17 (Fig 3B). As discussed below, the transgene locus in line 13 contained multiple FRT sites, which probably interfered with P-GUS excision. It is not known why P-GUS in line 19 was refractory to FLP activity. No GUS activity was detectable in plants containing either the 35Senh allele or the GUS allele (Fig 2). These plants were backcrossed with untransformed tobacco to remove the Cre or FLP loci and then selfed until homozygous for the 35Senh or GUS alleles. Plants homozygous for the respective unaltered 35SGUS allele were obtained by selfing each of the four lines.

View larger version (22K): In this window In a new window Download PPT slide   Figure 2. Quantification of GUS activity in transgenic lines and their Cre and FLP derivatives. GUS activity in leaves of adult plants of homozygous transformed lines 14, 17, and 19, as well as line 13 S2 progeny 1–4 (all are homozygous except 4) is shown. In contrast to the original lines, the Cre-derived 35Senh alleles (obtained for all four lines) and the FLP-derived GUS alleles (obtained for lines 14 and 17 only) produced no GUS activity. The averages of at least three separate determinations are shown. Data for the parental 13 plant are not available. NT, normal tobacco.

View larger version (74K): In this window In a new window Download PPT slide   Figure 3. DNA blot analysis of heteroalleles produced by Cre and FLP-mediated site-specific recombination. DNA isolated from representatives of the indicated plant lines was digested with either PstI (A) or EcoRI (B). The first lanes show the presence of the 35Senh (A) or GUS gene (B) in two or more of the original lines (13, 14, 17, and 19). The second lanes show absence of the sequence following excision by the Cre or FLP recombinase, respectively. Deletion of the 35S enhancer by the Cre recombinase was accomplished for all four transgenic lines. Excision of the P-GUS gene by the FLP recombinase was achieved only with lines 14 and 17. The probes used are shown as solid bars under the construct map in Fig 1A and consisted of 35S enhancer sequences (A) and the entire GUS coding sequence (B).

To test for A-TV in lines 14 and 17, intercrosses were made between plants homozygous for the respective 35Senh alleles and GUS alleles. Progeny, which all have a genotype of 35Senh/GUS (Fig 4D), were tested for GUS activity. None was detected in either histochemical or fluorometric assays carried out on at least 25 seedlings from each cross (data not shown). GUS activity was also not observed in progeny obtained by crossing lines 14 and 17 GUS plants with lines 13 and 19 35Senh plants (data not shown), indicating that E-TV did not occur.

View larger version (22K): In this window In a new window Download PPT slide   Figure 4. Possible allelic combinations. (A) Homozygous for unaltered allele—present in 100% of selfed progeny of all lines (exception: line 13 S2 no. 4, which produced 25% homozygous progeny in selfings); (B) hemizygous for unaltered allele—present in 100% of backcross (BC) progeny from all lines (exception: line 13 S2 no. 4, where 50% of selfed and 50% of BC progeny were hemizygous); (C) unaltered allele combined with 35Senh allele; (D) combination of 35Senh and GUS alleles. The combination in D did not produce any GUS activity with lines 14 and 17 for which both 35Senh and GUS alleles were available. Therefore, the allelic transvection (A-TV) and ectopic transvection (E-TV) discussed in the text refer to the combination shown in C. In the case of A-TV, the 35SGUS gene and 35S variant were from the same line; for E-TV, they were each from a different line. The relative amounts of GUS activity obtained for each of the four lines (13, 14, 17, and 19) with these allelic combinations are shown to the right. The amount measured in hemizygotes (B) was taken as 1x. Line 13 produced variable GUS activity. Average results from the A and C combinations in the line 13 parent, 13 S2 nos. 2 and 3, approximated unbracketed values; average results from A and C combinations in line 13 S2 nos. 1 and 4 approximated bracketed values (Fig 5 and Fig 6). n.d., not determined.

To improve the chances of homologous pairing, which might have been disrupted by the deletions in the 35Senh and GUS alleles, crosses were made to combine the 35Senh allele with the respective 35SGUS unaltered allele from each of the four lines (Fig 4C). Progeny of these crosses were tested for GUS to see whether the activity exceeded the amount observed in backcross (BC) progeny containing the unaltered allele in the hemizygous condition alone (Fig 4B). Increased GUS activity with the heteroallelic combination would provide evidence for trans-activation of the 35Senh allele, possibly through the action of 35S enhancer present on the unaltered allele. GUS activity was also measured in selfed progeny of plants homozygous for the unaltered allele (Fig 4A).

For lines 14, 17, and 19, GUS activity was dependent on gene dosage (homozygous selfed progeny contained 2 times as much as hemizygous BC progeny) and there was no detectable trans-activation (GUS activity in progeny containing the heteroallelic A-TV combination was comparable to that observed in hemizygous BC progeny; Fig 5, lines 14, 17, and 19). In line 13, a different pattern was obtained. Two observations are relevant. First, unlike the relatively uniform GUS values seen with homozygous selfed progeny in the other three lines, selfed progeny of line 13 exhibited a wide range of GUS activity, with some values falling to those observed in hemizygous BC progeny. Second, enhanced GUS activity (on average 1.4 times the hemizygous level in BC progeny) was observed with the combination of heteroalleles designed to test for A-TV (Fig 5, line 13). These results suggested that GUS activity was unstable in homozygous progeny of line 13 and that transvection was taking place.

View larger version (6K): In this window In a new window Download PPT slide   Figure 5. Relative GUS activity with different allelic combinations from the four original transgenic lines. Homozygous seedlings obtained from selfing, hemizygous seedlings from backcrossing (BC), and seedlings containing the heteroallelic combination designed to test for A-TV (Fig 4C) were assayed for GUS activity. Each point represents GUS activity in an extract prepared from an individual seedling. The average hemizygous levels in BC progeny were set at 1x. In line 13, seedlings containing the 35Senh allele alone are shown to demonstrate that this allele contributed no GUS activity. Some selfed progeny showed more than 2x GUS activity, which can be considered an additional reflection of the variability of this locus. Tests for ectopic transvection (E-TV) are shown for lines 17 and 19. These seedlings contained the unaltered 35SGUS allele from lines 17 or 19 and the 35Senh allele from line 13. No E-TV was observed. Statistically significant differences between the BC and A-TV values were only observed with line 13.

To examine this phenomenon in line 13 further, the homozygous parent was selfed and four progeny (S₂ nos. 1–4) that contained varying amounts of GUS activity (Fig 2) were selected for analysis. Seeds from selfing, backcrossing, and intercrossing to produce the heteroallelic A-TV combination were obtained to test the three allelic pairs shown in Fig 4A–C, respectively. The variability evident in the original line 13 parental plant was also manifested in the S₂ progeny, where different patterns of GUS activity were observed for each line. In the case of S₂ no. 3, a distribution similar to the parental line was observed, i.e., a wide range of GUS values in homozygous selfed progeny, some reaching hemizygous BC levels or slightly below, and increased GUS activity in the heteroallelic A-TV combination (on average, 1.6 times the hemizygous level measured in BC offspring; Fig 6, S₂ no. 3). GUS activity was less variable in selfed progeny of line 13 S₂ no. 2, where a relatively good dosage effect was observed together with a generalized enhancement of GUS activity in the A-TV combination (on average 1.3 times the hemizygous level observed in BC offspring; Fig 6, S₂ no. 2).

View larger version (11K): In this window In a new window Download PPT slide   Figure 6. Relative GUS activity with different allelic combinations in S2 progeny of line 13. Abbreviations and explanations are as in Fig 5. E-TV crosses are shown for S2 nos. 1 and 2. These seedlings contained the unaltered 35SGUS allele from lines 13 S2 nos. 1 and 2, respectively, and the 35Senh alleles from lines 14, 17, or 19 as indicated. In all cases, the differences between the A-TV values and BC values were statistically significant, whereas differences between BC and E-TV values were not statistically significant. In S2 nos. 1 and 4, differences between the homozygous selfed and hemizygous BC seedlings were not statistically significant. Similar profiles were obtained with populations of seedlings from at least two different pods of each cross. Although derived from selfing a homozygous parent, S2 no.4 behaved as a hemizygote, producing 25 and 50% GUS-negative progeny in the selfing and crosses, respectively.

The most dramatic results were obtained with line 13, S₂ nos. 1 and 4 (Fig 6, S₂ nos. 1 and 4). In both cases, the dosage effect disappeared: homozygous selfed progeny produced amounts of GUS activity similar to hemizygous BC progeny. Moreover, in the heteroallelic A-TV combination, average GUS activity was up to two times that observed in the homozygous and hemizygous offspring. Similar results were obtained in reciprocal crosses, indicating no parental effect. An additional peculiarity of line 13, S₂ no. 4 was that it behaved as a hemizygote, even though it was produced by selfing the homozygous parental line 13 plant. With line 13 S₂ no. 4, 25% of the selfed progeny and 50% of the progeny obtained from the crosses were GUS-negative (Fig 6, S₂ no. 4). DNA blot analysis demonstrated that transgene sequences were missing in GUS-negative progeny (data not shown). This was apparently due to deletion of the transgene sequences, probably during meiosis, and not to chromosome loss because line 13 S₂ no. 4 contained the normal number of chromosomes (data not shown).

In contrast to lines 14, 17, and 19, the transgene locus in line 13 thus appeared to be susceptible to transvection: GUS expression in homozygous progeny could be highly variable and in some cases (S₂ nos. 1 and 4) a gene silencing phenomenon akin to dosage compensation, where two copies of the 35SGUS gene were expressed at a level comparable to one copy, was regularly observed. Moreover, in addition to this partial silencing of unaltered 35SGUS alleles in the homozygous state, line 13 exhibited trans-activation, as revealed by enhanced GUS activity when the unaltered 35SGUS allele and the 35Senh allele, which contributed no GUS activity, were together in the same genome.

To determine whether structural variation could account for these differences in behavior, the four transgene loci were cloned out of the tobacco genome and the nucleotide sequences were determined. Lines 14, 17, and 19 contained single copies of the transgene construct, although an internal rearrangement was present in line 14 (Fig 7). In contrast, line 13 contained a large scrambled locus comprising reiterated transgene sequences and substantial binary vector DNA. Despite this complexity, line 13 contained only one complete 35SGUS gene, which was present in the spacer region between two NOS gene sequences arranged as an inverted repeat (IR). The lone 35S enhancer at this locus could thus be present in the loop of a cruciform. Consistent with the simple structures of the transgene loci in lines 14, 17, and 19, the KAN and NOS genes were expressed strongly in progeny. In contrast, progeny of line 13 displayed only weak kanamycin resistance and negligible amounts of nopaline (data not shown), presumably owing to progressive silencing of the multiple copies of the NOS promoter in advanced generations.

View larger version (25K): In this window In a new window Download PPT slide   Figure 7. Structures of the transgene loci in lines 13, 14, 17, and 19. Genomic -clones (lines 13 and 14) or rescue clones (lines 17 and 19) containing the respective transgene inserts were isolated and sequenced. Solid regions indicate transgene sequences (described in Fig 1A); shaded blocks denote binary vector sequences; open flanking regions represent plant DNA. Locus 13 is large and scrambled (arrows below maps indicate the relative orientations of different fragments); two contiguous regions of this locus are connected by the dashed diagonal line. Portions arranged as an inverted repeat (IR) are shown. The single 35S enhancer in the locus is present in the spacer of the IR and is highlighted as a white block. Note the two FRT sites (top half) in addition to the two surrounding the intact GUS gene (bottom half). Lines 14, 17, and 19 contain essentially single copies of the transgene construct. Right and left borders are missing in line 19, as is the NOS promoter driving the NOS gene. Nevertheless this plant is nopaline positive, indicating that the flanking plant DNA, part of which is homologous to a portion of an extensin gene (ext) promoter (accession no. L38908), has promoter activity. A left T-DNA border is missing in line 17; plant DNA to the right of transgene sequences in line 17 was not recovered in the rescue clone. In line 14, the region containing the NOS and 35SGUS genes was inverted relative to the KAN gene and some transgene sequences were largely deleted (e.g., pEMBL), but otherwise no additional scrambling was observed. Some flanking binary vector DNA was present at the left in locus 14. Abbreviations are as in Fig 1.

Although structural differences provided an explanation for the distinct behaviors of the four individual transgene loci, it was more difficult to account for variations in the four line 13 S₂ lines. DNA blot analysis using probes throughout the transgene construct revealed no major differences in structure or methylation among line 13 S₂ nos. 1–3 (Fig 8, A–D) and their chromosome numbers were normal (data not shown). The only differences among the line 13 S₂ lines were observed with the hemizygote S₂ no. 4, in which methylation in the NOS promoter of the KAN gene (Fig 8B, lane 4) was detected. The significance of this difference is not known.

View larger version (71K): In this window In a new window Download PPT slide   Figure 8. DNA blot analysis of line 13 S2 progeny. DNA isolated from line 13 S2 nos.1–4 was digested with the appropriate restriction enzymes and probed with sequences throughout the transgene construct (Fig 1A) as follows: (A) HpaII or MspI, 35S probe; (B) EcoRI/PstI plus or minus SacII, KAN probe; (C) EcoRI, GUS coding region probe; and (D) NsiI plus or minus SacII, NOS probe (Fig 1). The blots in A, B, and D examine methylation in the four lines. Two lanes representing digests minus and plus a methylation-sensitive restriction enzyme are shown for each line. Digestion with either HpaII (CpG methylation-sensitive) or MspI (CpG methylation-insensitive) produced two 35S bands of the expected size (Fig 1A), indicating no methylation at a HpaII site in this single-copy region (A). For S2 nos. 1–3, an EcoRI/PstI fragment shifted to the smaller fragment following the addition of methylation-sensitive SacII, indicating no methylation of this site in the NOS promoter of the single KAN gene (B). In all four plant lines, two GUS bands, corresponding to the intact 35SGUS gene and a truncated GUS coding region, were observed (C). Conversion of several NsiI bands to smaller fragments in all four lines after addition of methylation-sensitive SacII revealed little or no methylation in the NOS sequences, which are present in at least five partial or complete copies at locus 13 (D). The only visible difference in structure or methylation was in S2 no. 4, where a SacII site in the NOS promoter of the KAN gene was methylated, as indicated by the maintenance of the large fragment after addition of SacII (arrow in B, lane 4). S2 no. 4 signals appear fainter because this plant is hemizygous for the transgene locus, whereas other lines are homozygous. The fragments observed with all probes can be accounted for in the locus 13 structure determined by sequencing (Fig 7). All fragments cosegregated in backcross progeny demonstrating that they are linked. The line 13 35Senh allele appeared identical to S2 nos. 1–3 with these probes (data not shown) except for the absence of the 35S enhancer (Fig 3A).

Because dsRNAs that are processed to small (21–25 nucleotide) RNAs have been implicated in both RNA-mediated TGS and PTGS (MATZKE et al. 2001 ), we tested whether the gene silencing phenomenon observed in line 13 might be due to the unintentional production of dsRNAs or small RNAs containing GUS or 35Senh sequences. GUS mRNA was present in plants from lines 13, 14, 17, and 19 (Fig 9A) at steady-state levels, approximating the relative amounts of GUS activity in these plants (Fig 2). Although it is difficult to predict the length of dsRNAs that might be synthesized from locus 13, RNase protection experiments failed to reveal dsRNAs derived from either GUS or 35Senh sequences (data not shown). Moreover, no detectable amounts of 23-nucleotide RNAs containing GUS or 35Senh sequences were observed in plants 13 S₂ nos. 1 and 2 (Fig 9B and data not shown), which differed approximately twofold in GUS activity (Fig 2) and steady-state GUS RNA levels (Fig 9A). Plants containing the 35Senh allele did not produce GUS mRNA (Fig 9A), as expected from the absence of GUS activity in these plants (Fig 2), nor did these plants contain dsRNA or small RNAs derived from GUS or 35Senh sequences (Fig 9B and data not shown). The absence of small or dsRNAs in plants showing reduced or no GUS activity argues against RNA-mediated silencing of the 35SGUS gene in plant 13 S₂ no. 1, supporting the involvement of DNA-DNA pairing in the silencing phenomenon observed.

View larger version (63K): In this window In a new window Download PPT slide   Figure 9. RNA analyses. (A) Northern blot analysis of GUS mRNA. Total RNA preparations isolated from leaves of adult plants of transgenic lines 14, 17, 19, 13 S2 no. 1, 13 S2 no. 2, 13 35Senh, and untransformed tobacco (NT) were electrophoresed, blotted onto nitrocellulose, and probed with the GUS-coding sequence (top) or an actin probe (bottom) as a loading control. The relative levels of GUS RNA correlate roughly with the average GUS activities measured in plants from these lines (Fig 2): line 17 is the highest; lines 14 and 19 are somewhat less; 13 S2 no. 2 contains at least twice as much GUS RNA as does 13 S2 no. 1. Both GUS RNA and GUS enzyme activity are absent in line 13 35S. (B) Attempt to detect small RNAs. RNA preparations enriched for small (21–25 nucleotide) RNAs were isolated from the indicated plants, electrophoresed on 15% acrylamide gels, electroblotted onto a nylon membrane, and hybridized to a GUS antisense RNA probe. Two tobacco lines, 6b5 and T4, in which a 35SGUS gene is silenced by PTGS, were used as positive controls for GUS small RNAs (HAMILTON and BAULCOMBE 1999 ); line 14, in which no GUS silencing occurred, and untransformed tobacco (NT) were used as negative controls. Although signals are clearly visible in the 6b5 and T4 lanes, no signals were observed in line 13 plants, including S2 no. 1, which contains approximately one-half the GUS activity and GUS RNA as S2 no. 2, or the line 13 35Senh derivative, which augments GUS activity when combined with the unaltered allele in lines 13 S2 no. 1. Similar results were obtained with a GUS sense RNA probe (data not shown). Under identical conditions, no small RNAs derived from the 35Senh were detected (data not shown). The position of a 23-nucleotide marker is indicated to the left.

To determine whether the transgene locus in line 13 was able to trans-activate the ectopic loci in lines 14, 17, and 19 (E-TV), crosses were made to combine either the line 13 35Senh allele with the unaltered lines 14, 17, and 19 35SGUS alleles or the line 13 unaltered 35SGUS allele with the lines 14, 17, and 19 35Senh alleles. No significant enhancement of GUS activity above the hemizygous BC level was observed, indicating no appreciable E-TV (Fig 5, lines –17 and 19; Fig 6, S₂ nos. 1 and 2). These findings further substantiate a role for allelic pairing in the transvection effects observed in line 13.

	DISCUSSION

TOP ABSTRACT MATERIALS AND METHODS RESULTS DISCUSSION LITERATURE CITED

In this study, a system designed to mimic transvection was established to attempt to detect somatic pairing of transgene alleles in tobacco. This system assessed the ability of an enhancer on one allele to trans-activate an enhancerless promoter on the second allele. The required heteroalleles were generated at four different locations in the tobacco genome by Cre-lox-mediated excision. No evidence for trans-activation was obtained for three loci containing simple transgene inserts, indicating that transvection does not normally occur with single-copy, unrearranged transgenes. One complex transgene locus, however, exhibited behavior that can be interpreted in terms of allelic pairing. Depending on which alleles were present, two distinct transvection effects—resulting in either trans-activation or trans-silencing—were observed. Because trans-activation presumably required close physical association of transgene heteroalleles, it can be inferred that the trans-silencing phenomenon also involved allelic pairing. Moreover, the frequency and degree of silencing correlated roughly with the strength of trans-activation, providing a further connection between the two phenomena. RNA-mediated silencing could be ruled out because the expected hallmarks of RNA silencing—dsRNAs and small RNAs containing GUS or 35Senh sequences—were not detected in plants with diminished GUS activity. These results thus provide molecular evidence suggesting that complex alleles can physically associate in somatic plant cells in a manner that influences gene expression.

The large size and repetitiveness of transgene locus 13 possibly induced the formation of heterochromatin capable of forming homologous associations (HENIKOFF 1997 ; MARAHRENS 1999 ). Despite the complex character of locus 13, however, the intact 35SGUS gene undergoing transvection was present as a single copy in the original T-DNA sequence context, similar to the 35SGUS gene at other three loci. Moreover, the 35SGUS gene in line 13 was never fully silenced, as might be expected from wholesale heterochromatinization of the locus. These observations can be reconciled with the two types of transvection effects observed in line 13 by considering the possible importance of an IR that could make up the stem of a cruciform containing the 35S enhancer in the loop. This structural arrangement could facilitate allelic interactions and determine whether trans-activation or silencing occurred.

Silencing took place when two unaltered alleles that each possessed a 35S enhancer were present in homozygotes. Conversely, trans-activation occurred with the heteroallelic combination in which the 35S enhancer was deleted from one allele (Fig 10). These results suggest that allelic pairing affected the activity of enhancer sequences. Indeed, a close link between transcriptional regulatory elements and homology effects has been suggested by a number of other observations (MORRIS et al. 1999 ; WU and MORRIS 1999 ). It has also been noted that unpaired loops might be particularly prone to initiating a homology search (WU and MORRIS 1999 ). The presence of the 35S enhancer in the loop of a potential cruciform might thus account for its ability to locate and to interact with allelic partners in a way that modifies gene expression.

View larger version (23K): In this window In a new window Download PPT slide   Figure 10. Two types of transvection effects observed with the 35SGUS transgene in line 13. trans-activation, in which the heteroallelic 35SGUS/GUS combination produced GUS activity in excess of the 35SGUS unaltered allele in the hemizygous condition alone (>1x GUS), provided evidence for allelic pairing. One explanation for this effect is that the 35S enhancer on the unaltered allele acted in cis and in trans to drive expression from both P-GUS genes (A). An alternative possibility is that pairing of two structurally dissimilar alleles deforms one allele, improving the accessibility of the single 35Senh to transcription factors (solid spheres) and increasing expression from one allele (B). When the unaltered 35SGUS allele was present in the homozygous state, GUS activity could be reduced to that observed in hemizygotes (1x GUS). This silencing phenomenon, which resembled dosage compensation, was presumably a consequence of allelic pairing. It is not yet known whether one allele was expressed and one was silent (C) or whether expression from both alleles was halved (D).

While IRs have been repeatedly implicated in trans-silencing effects in plants (MUSKENS et al. 2000 ), this is the first case potentially linking IRs and trans-activation. Whether IRs are involved in silencing or activation probably depends on their gene content and length; the size and composition of any intervening spacer; whether the IR is transcribed; and whether any DNA secondary structures, such as hairpins or cruciforms, are produced in vivo. In the effects described here, the importance of the IR possibly rested in its potential to form the stem of a cruciform that placed the 35S enhancer in a loop, a position favoring trans-interactions. Sequences present in the repeated regions of an IR can lead to different types of HDGS. For example, IRs containing promoter sequences or transcribed regions can trigger trans-silencing at the transcriptional (METTE et al. 2000 ) or post-transcriptional (STAM et al. 1998 ) level, respectively. It has been uncertain whether IRs act as trans-silencers through DNA-DNA interactions or by means of dsRNA (SELKER 1999 ). While some cases of trans-silencing involving IRs do appear to rely on dsRNA (METTE et al. 2000 ; MUSKENS et al. 2000 ), we were unable to detect dsRNA or 21–25 nucleotide RNAs containing either GUS or 35Senh sequences in line 13 plants. Moreover, transvection was restricted to the heteroallelic pair and did not extend to homologous transgenes at ectopic loci as might be expected if a diffusible dsRNA were involved. These data provide support for the idea that IRs have trans-sensing ability and can modify gene expression in certain circumstances via DNA-DNA pairing. Given that lines 14, 17, and 19, which did not exhibit transvection, contained only single copies of the transgene construct, it still remains possible that ectopic pairing interactions could occur between two loci that each contain an IR. The information on transgene structural requirements obtained from the tobacco transvection system is being used to reproduce the effects in Arabidopsis with the hope of obtaining multiple lines that can be used to examine the phenomenon in more detail.

The mechanism of trans-activation is not yet known. Conceivably, pairing could permit the enhancer to function both in cis and in trans to activate expression from two alleles (Fig 10A). Alternatively, pairing of structurally dissimilar heteroalleles could lead to deformation of one allele, causing the 35S enhancer to become more accessible to transcription factors, and boosting expression from one allele (Fig 10B). These possibilities, both of which require physical interactions between alleles, can be tested when similar experiments are performed with marked alleles whose individual expression can be distinguished.

The magnitude of trans-activation is notable. In line 13 S₂ no. 4, average GUS activity resulting from the heteroallelic combination was nearly double that observed in hemizygotes, and thus approached a level normally expected in homozygotes. This indicates that certain transgene alleles can locate each other in a large genome containing much repetitive DNA. [The tobacco haploid genome contains 4.4 pg of DNA (BENNETT and LEITCH 1998 ), 45% of which consists of repeated sequences (ZIMMERMAN and GOLDBERG 1977 )]. Moreover, despite the apparent lack of stable pairing of homologous chromosomes in most somatic plant cells, the trans-activation we observed suggests that relatively long-lived associations can form between specific regions, in particular those that have one or more special features noted here, such as repetitiveness or the potential to form secondary structures. It is unclear why the strengths of the transvection effects were variable within line 13. If a cruciform is required to promote homology searching and allelic pairing, then perhaps the locus 13 cruciform, with its relatively long spacer (12 kb) separating IRs that are 2 kb in length, might not have formed frequently. Many IRs are present in higher eukaroytic genomes and they must be considered sources of not only physical instability but also epigenetic variability. In addition to the transvection effects observed with the transgene locus in line 13, the apparent meiotic loss of one allele in line 13 S₂ no. 4 is consistent with the general lability of IRs.

The silencing phenomenon seen in homozygous line 13 progeny differs from those reported previously in plants by resembling dosage compensation, where two copies of a gene are expressed at a level comparable to one copy. Normally, homozygous transgene loci that undergo HDGS are suppressed considerably more than the 50% observed here (VAN HOUDT et al. 2000 ). It is not known whether the silencing observed in homozygous line 13 plants resulted from inactivation of just one GUS allele (Fig 10C) or whether activity of each allele was reduced by a fractional amount (Fig 10D). We assume that silencing occurred at the transcriptional level because of the apparent involvement of enhancer sequences and the absence of dsRNA or small RNAs containing GUS sequences, which would be indicative of PTGS. Dosage compensation suggests some sort of counting mechanism. The IR, which can conceivably pair in cis, might already be counted as two copies, the normal diploid number. This could result in the inactivation of one allele in homozygotes where the IR would be perceived as four copies. In mammals, dosage compensation is exemplified by X-chromosome inactivation. Similar to the results reported here, physical interactions between Xist alleles have been proposed as a means to silence and heterochromatize one allele but not the other (MARAHRENS 1999 ). Other cases of monoallelic expression in mammals, either with (LASALLE and LALANDE 1996 ; DUVILLIE et al. 1998 ) or without (HOLLANDER et al. 1998 ) parental imprints, could also possibly result from homologous pairing, particularly at regions of heterochromatic repeats (MARAHRENS 1999 ) or IRs.

None of the three single-copy loci exhibited either trans-activation or silencing. This suggests that unlike the examples from filamentous fungi, where simple sequence duplications are sufficient to trigger DNA modifications, single-copy loci do not appreciably pair in somatic plant cells. Alternatively, pairing could occur but be inconsequential for gene expression. The stable activity of the three single-copy loci, all of which expressed the 35SGUS gene at levels higher than locus 13, does not support threshold models for gene silencing. Complex loci, particularly those containing IRs, continue to be most frequently implicated in silencing effects in plants (STAM et al. 1998 ; LUFF et al. 1999 ).

In summary, the results presented here provide evidence that alleles of a large and repetitive transgene locus can interact in trans—most likely via DNA-DNA pairing—to modify gene expression either positively or negatively in somatic cells. Given the abundance of various types of repeats in most plant and mammalian genomes and the proximity of repeats to many genes, it is possible that pairing interactions involving these repeats affect the activity of endogenous genes more often than currently believed. Transgenes can be useful tools for probing the types of effects that are possible and their sequence and structural requirements.

	FOOTNOTES

¹ Present address: Istituto Multidisciplinario de Biología Vegetal (IMBIV), 5000 Córdoba, Argentina.

	ACKNOWLEDGMENTS

We thank N. Kilby for providing the FLP-expressing tobacco line; J. Odell for the Cre-expressing tobacco line; N.H Chua for pUC XGUS-46 (pEMBL-GUS-46); I. Papp for pFJS166, which contained the 166 bp FRT site; A. Hamilton for tobacco line T4; and H. Vaucheret for tobacco line 6b5. We are grateful to C.-T. Wu, J. Kooter, D. Baulcombe, and Y. Marahrens for helpful discussions and advice. This work was supported by the Austrian Fonds zur Förderung der wissenschaftlichen Forschung (grant no. Z21-MED) and the European Union (contract no. BIO4-CT96-0253).

Manuscript received November 27, 2000; Accepted for publication February 15, 2001.

	LITERATURE CITED

TOP ABSTRACT MATERIALS AND METHODS RESULTS DISCUSSION LITERATURE CITED

BAULCOMBE, D. C. and J. J. ENGLISH, 1996  Ectopic pairing of homologous DNA and post-transcriptional gene silencing in transgenic plants. Curr. Opin. Biotechnol. 7:173-180.

BENDER, J., 1998  Cytosine methylation of repeated sequences in eukaryotes: the role of DNA pairing. Trends Biochem. Sci. 23:252-256[Medline].

BENFEY, P. N., L. REN, and N.-H. CHUA, 1989  The CaMV 35S enhancer contains at least two domains which can confer different developmental and tissue-specific expession patterns. EMBO J. 8:2195-2202[Medline].

BENNETT, M. D. and I. J. LEITCH, 1998  Nuclear DNA amounts in angiosperms—583 new estimates. Ann. Bot. 80:169-196[Abstract/Free Full Text].

BIRCHLER, J. A., M. PAL BHADRA, and U. BHADRA, 2000  Making noise about silence: repression of repeated genes in animals. Curr. Opin. Genet. Dev. 10:211-216[Medline].

BRADFORD, M. M., 1976  A rapid and sensitive method for the quantitation of microgram quantities of protein utilizing the principle of protein-dye binding. Anal. Biochem. 72:248-254[Medline].

COGONI, C. and G. MACINO, 2000  Post-transcriptional gene silencing across kingdoms. Curr. Opin. Genet. Dev. 10:638-643[Medline].

COLOT, V., L. MALOISEL, and J.-L. ROSSIGNOL, 1997  Interchromosomal transfer of epigenetic states in Ascobolus: transfer of DNA methylation is mechanistically related to homologous recombination. Cell 86:855-864.

DALE, E. C. and D. W. OW, 1991  Gene transfer with subsequent removal of the selection gene from the host genome. Proc. Natl. Acad. Sci. USA 88:10558-10562[Abstract/Free Full Text].

DUVILLIE, B., D. BUCCHINI, T. TANG, J. JAMI, and A. PALDI, 1998  Imprinting at the mouse Ins2 locus: evidence for cis- and trans-allelic interactions. Genomics 47:52-57[Medline].

FAGARD, M. and H. VAUCHERET, 2000  Trans)gene silencing in plants: how many mechanisms? Annu. Rev. Plant Physiol. Plant Mol. Biol. 51:167-194.

FAUGERON, G., 2000  Diversity of homology-dependent gene silencing strategies in fungi. Curr. Opin. Microbiol. 3:144-148[Medline].

GARRICK, D., S. FIERING, D. I. K. MARTIN, and E. WHITELAW, 1998  Repeat-induced gene silencing in mammals. Nat. Genet. 18:56-59[Medline].

GOLDSBOROUGH, A. S. and T. B. KORNBERG, 1996  Reduction of transcription by homologue asynapsis in Drosophila imaginal discs. Nature 381:807-810[Medline].

HAMILTON, A. and D. C. BAULCOMBE, 1999  A novel species of small antisense RNA in post-transcriptional gene silencing. Science 286:950-952[Abstract/Free Full Text].

HENIKOFF, S., 1997  Nuclear organization and gene expression: homologous pairing and long-range interactions. Curr. Opin. Cell Biol. 9:388-395[Medline].

HENIKOFF, S. and L. COMAI, 1998  Trans-sensing effects: the ups and downs of being together. Cell 93:329-332[Medline].

HENIKOFF, S. and T. D. DREESEN, 1989  Trans-inactivation of the Drosophila brown gene: evidence for transcriptional repression and somatic pairing dependence. Proc. Natl. Acad. Sci. USA 86:6704-6708[Abstract/Free Full Text].

HOLLANDER, G. A., S. ZUKLYS, C. MOREL, E. MIZOGUCHI, and K. MOBISSON et al., 1998  Monoallelic expression of the interleukin-2 locus. Science 279:2118-2121[Abstract/Free Full Text].

HOLLICK, J. B., J. E. DORWEILER, and V. L. CHANDLER, 1997  Paramutation and related allelic interactions. Trends Genet. 13:302-307[Medline].

JEFFERSON, R. A., 1987  Assaying chimeric genes in plants: the GUS gene fusion system. Plant Mol. Biol. Rep. 5:387-405.

JORGENSEN, R.A., 1992  Silencing of plant genes by homologous transgenes. Agric. Biotechnol. News Inf. 4:265-273.

KILBY, N. J., G. J. DAVIES, M. R. SNAITH, and J. A. H. MURRAY, 1995  FLP recombinase in transgenic plants: constitutive activity in stably transformed tobacco and generation of marked cell clones in Arabidopsis. Plant J. 8:637-652[Medline].

LASALLE, J. M. and M. LALANDE, 1996  Homologous association of oppositely imprinted chromosomal domains. Science 272:725-728[Abstract].

LUFF, B., L. PAWLOWSKI, and J. BENDER, 1999  An inverted repeat triggers cytosine methylation of identical sequences in Arabidopsis. Mol. Cell 3:505-511[Medline].

MARAHRENS, Y., 1999  X-inactivation by chromosomal pairing events. Genes Dev. 13:2624-2632[Free Full Text].

MATZKE, A. J. M. and M. A. MATZKE, 1986  A set of novel Ti-plasmid-derived vectors for the production of transgenic plants. Plant Mol. Biol. 7:357-365.

MATZKE, M. A. and A. J. M. MATZKE, 1995  How and why do plants inactivate homologous (trans)genes? Plant Physiol. 107:679-685[Medline].

MATZKE, M. A., M. PRIMIG, J. TRNOVSKY, and A. J. M. MATZKE, 1989  Reversible methylation and inactivation of marker genes in sequentially transformed tobacco plants. EMBO J. 8:643-649[Medline].

MATZKE, M. A., A. J. M. MATZKE, G. J. PRUSS, and V. B. VANCE, 2001  RNA-based silencing strategies in plants. Curr. Opin. Genet. Dev. 11:221-227[Medline].

MEINS, F., 2000  RNA degradation and models for posttranscriptional gene silencing. Plant Mol. Biol. 43:261-273[Medline].

METTE, M. F., J. VAN DER WINDEN, M. A. MATZKE, and A. J. M. MATZKE, 1999  Production of aberrant promoter transcripts contributes to methylation and silencing of unlinked homologous promoters in trans. EMBO J. 18:241-248[Medline].

METTE, M. F., W. AUFSATZ, J. VAN DER WINDEN, M. A. MATZKE, and A. J. M. MATZKE, 2000  Transcriptional gene silencing and promoter methylation triggered by double-stranded RNA. EMBO J. 19:5194-5201[Medline].

MORRIS, J. R., P. K. GEYER, and C.-T. WU, 1999  Core promoter elements can regulate transcription on a separate chromosome in trans. Genes Dev. 13:253-258[Abstract/Free Full Text].

MUSKENS, M. W., A. P. VISSERS, J. N. M. MOL, and J. M. KOOTER, 2000  Role of inverted DNA repeats in transcriptional and post-transcriptional gene silencing. Plant Mol. Biol. 43:243-260[Medline].

ODELL, J. T., and S. H. RUSSELL, 1994 Use of site-specific recombination systems in plants, pp. 219–270 in Homologous Recombination and Gene Silencing in Plants, edited by J. PASZKOWSKI. Kluwer Academic, Dordrecht/Norwell, MA.

PAPP, I., V. A. IGLESIAS, E. A. MOSCONE, S. MICHALOWSKI, and S. SPIKER et al., 1996  Structural instability of a transgene locus in tobacco is associated with aneuploidy. Plant J. 10:469-478[Medline].

PIETRZAK, M., R. D. SHILLITO, T. HOHN, and I. POTRYKUS, 1986  Expression in plants of two bacterial antibiotic resistance genes after protoplast transformation with a new plant expression vector. Nucleic Acids Res. 14:5857-5868[Abstract/Free Full Text].

PIRROTTA, V., 1999  Transvection and chromosomal trans-interaction effects. Biochim. Biophys. Acta 1424:M1-M8[Medline].

PLASTERK, R. H. A. and R. F. KETTING, 2000  The silence of the genes. Curr. Opin. Genet. Dev. 10:562-567[Medline].

SCHERNTHANER, J. P., M. A. MATZKE, and A. J. M. MATZKE, 1988  Endosperm-specific activity of a zein gene promoter in transgenic tobacco plants. EMBO J. 7:1249-1255[Medline].

SELKER, E. U., 1997  Epigenetic phenomena in filamentous fungi: useful paradigms or repeat-induced confusion? Trends Genet. 13:296-301[Medline].

SELKER, E. U., 1999  Gene silencing: repeats that count. Cell 97:157-160[Medline].

STAM, M., A. VITERBO, J. N. M. MOL, and J. M. KOOTER, 1998  Position dependent methylation and transcriptional silencing of transgenes in inverted repeat T-DNA repeats: implications for posttranscriptional silencing of homologous host genes in plants. Mol. Cell. Biol. 18:6165-6177[Abstract/Free Full Text].

TARTOF, K. D. and S. HENIKOFF, 1991  Trans-sensing effects from Drosophila to humans.. Cell 65:201-203[Medline].

VAN HOUDT, H., M. VAN MONTAGU, and A. DEPICKER, 2000  Both sense and antisense RNAs are targets for the sense transgene-induced posttranscriptional silencing mechanisms. Mol. Gen. Genet. 263:995-1002[Medline].

WASSENEGGER, M., 2000  RNA-directed DNA methylation. Plant Mol. Biol. 43:203-220[Medline].

WIANNY, F. and M. ZERNICKA-GOETZ, 2000  Specific interference with gene function by double-stranded RNA in early mouse development. Nat. Cell Biol. 2:70-75[Medline].

WOLFFE, A. P. and M. A. MATZKE, 1999  Epigenetics: regulation through repression. Science 286:481-486[Abstract/Free Full Text].

WU, C-T. and J. R. MORRIS, 1999  Transvection and other homology effects. Curr. Opin. Genet. Dev. 9:237-246[Medline].

ZIMMERMAN, J. L. and R. B. GOLDBERG, 1977  DNA sequence organization in the genome of Nicotiana tabacum. Chromosoma 59:227-252.

This article has been cited by other articles:

				R. S. Hawley and W. D. Gilliland Sometimes the Result Is Not the Answer: The Truths and the Lies That Come From Using the Complementation Test Genetics, September 1, 2006; 174(1): 5 - 15. [Abstract] [Full Text] [PDF]

				M. Fojtova, A. Bleys, J. Bedrichova, H. Van Houdt, K. Krizova, A. Depicker, and A. Kovarik The trans-silencing capacity of invertedly repeated transgenes depends on their epigenetic state in tobacco. Nucleic Acids Res., January 1, 2006; 34(8): 2280 - 2293. [Abstract] [Full Text] [PDF]

				W. Yang, R. A. Jefferson, E. Huttner, J. M. Moore, W. B. Gagliano, and U. Grossniklaus An Egg Apparatus-Specific Enhancer of Arabidopsis, Identified by Enhancer Detection Plant Physiology, November 1, 2005; 139(3): 1421 - 1432. [Abstract] [Full Text] [PDF]

				R. J. Pratt, D. W. Lee, and R. Aramayo DNA Methylation Affects Meiotic trans-sensing, Not Meiotic Silencing, in Neurospora Genetics, December 1, 2004; 168(4): 1925 - 1935. [Abstract] [Full Text] [PDF]

				P. Fransz, J. H. de Jong, M. Lysak, M. R. Castiglione, and I. Schubert Interphase chromosomes in Arabidopsis are organized as well defined chromocenters from which euchromatin loops emanate PNAS, October 29, 2002; 99(22): 14584 - 14589. [Abstract] [Full Text] [PDF]

				S. Liepelt, R. Bialozyt, and B. Ziegenhagen Wind-dispersed pollen mediates postglacial gene flow among refugia PNAS, October 29, 2002; 99(22): 14590 - 14594. [Abstract] [Full Text] [PDF]

				A. Kloti, X. He, I. Potrykus, T. Hohn, and J. Futterer Tissue-specific silencing of a transgene in rice PNAS, August 6, 2002; 99(16): 10881 - 10886. [Abstract] [Full Text] [PDF]

				J.-L. Chen, K. L. Huisinga, M. M. Viering, S. A. Ou, C.-t. Wu, and P. K. Geyer Enhancer action in trans is permitted throughout the Drosophila genome PNAS, March 19, 2002; 99(6): 3723 - 3728. [Abstract] [Full Text] [PDF]

				M. Kusaba, C.-W. Tung, M. E. Nasrallah, and J. B. Nasrallah Monoallelic Expression and Dominance Interactions in Anthers of Self-Incompatible Arabidopsis lyrata Plant Physiology, January 1, 2002; 128(1): 17 - 20. [Full Text] [PDF]

				B.-r. Jeong, D. Wu-Scharf, C. Zhang, and H. Cerutti From the Cover: Suppressors of transcriptional transgenic silencing in Chlamydomonas are sensitive to DNA-damaging agents and reactivate transposable elements PNAS, January 22, 2002; 99(2): 1076 - 1081. [Abstract] [Full Text] [PDF]