Plant Cell Reports

, Volume 28, Issue 7, pp 1095–1104

A transformation booster sequence (TBS) from Petunia hybrida functions as an enhancer-blocking insulator in Arabidopsis thaliana


  • Jean-Michel Hily
    • USDA-ARS Appalachian Fruit Research Station
    • Department of Plant Pathology, New York State Agricultural Experiment StationCornell University
  • Stacy D. Singer
    • USDA-ARS Appalachian Fruit Research Station
  • Yazhou Yang
    • USDA-ARS Appalachian Fruit Research Station
    • USDA-ARS Appalachian Fruit Research Station
Genetics and Genomics

DOI: 10.1007/s00299-009-0700-8

Cite this article as:
Hily, J., Singer, S.D., Yang, Y. et al. Plant Cell Rep (2009) 28: 1095. doi:10.1007/s00299-009-0700-8


Several matrix-attachment regions (MARs) from animals have been shown to block interactions between an enhancer and promoter when situated between the two. Since a similar function for plant MARs has not been discerned, we tested the Zea maysADH1 5′ MAR, Nicotiana tabacumRb7 3′ MAR and a transformation booster sequence (TBS) MAR from Petunia hybrida for their ability to impede enhancer–promoter interactions in Arabidopsis thaliana. Stable transgenic lines containing vectors in which one of the three MAR elements or a 4 kb control sequence were interposed between the cauliflower mosaic virus35S enhancer and a flower-specific AGAMOUS second intron-derived promoter (AGIP)::β-glucuronidase (GUS) fusion were assayed for GUS expression in vegetative tissues. We demonstrate that the TBS MAR element, but not the ADH1 or Rb7 MARs, is able to block interactions between the 35S enhancer and AGIP without compromising the function of either with elements from which they are not insulated.


MARTransformation booster sequence35S promoter/enhancerAGAMOUS second intronArabidopsis thalianaEnhancer-blocking insulator



Matrix attachment region


Transformation booster sequence




AGAMOUS second intron-derived promoter


35S cauliflower mosaic virus promoter/enhancer


Green fluorescent protein


Genetic insulators are present in a wide variety of eukaryotic genomes and are defined by their ability to block interactions between enhancer and promoter when positioned between them, without compromising the ability of either to communicate with elements from which they are not insulated (Kellum and Schedl 1992; Conte et al. 2002), and/or their capacity to prevent the advancement of nearby condensed chromatin, thereby protecting gene expression from positive or negative chromatin effects (Kellum and Schedl 1991; Saitoh et al. 2000). In animals, numerous enhancer-blocking insulators have been identified, the most well-characterized of which include the gypsy retrotransposon (Geyer et al. 1986) and scs/scs′ paired elements (Kellum and Schedl 1991; 1992) from Drosophila, as well as the cHS4 insulator from the chicken β-globin locus (Chung et al. 1993; Hebbes et al. 1994). While the mechanism by which these enhancer-blocking insulators function remains a mystery, it appears that the proteins that bind the respective insulators come together to form clusters that are localized at the nuclear periphery, resulting in loops of DNA (Kellum and Schedl 1991; 1992; Gerasimova et al. 2000; Yusufzai et al. 2004).

Matrix attachment regions (MARs), which are characterized by their ability to bind a network of nonhistone proteins designated the nuclear matrix (Laemmli et al. 1992), are another type of DNA sequence that has been proposed to organize chromatin into loop domains (reviewed by Bode et al. 1996). While the exact in vivo function of these sequences is not known, it has been suggested that like insulators, they may delimit the boundaries of discrete domains (Avramova et al. 1995; Tikhonov et al. 2000) and play important regulatory roles in establishing the correct expression patterns of endogenous genes (Kohwi-Shigematsu et al. 1997; Liu et al. 1997). In animals, a small number of MAR elements have been shown to possess enhancer-blocking properties (Stief et al. 1989; Nabirochkin et al. 1998), and since proteins required for enhancer-blocking in vertebrates also associate with the nuclear matrix (Dunn et al. 2003), it is feasible that the functions of MARs and enhancer-blocking insulators may be related.

To date, numerous MARs have also been isolated from various plant species (Breyne et al. 1992; van der Geest et al. 1994; Avramova et al. 1995; Allen et al. 1996) and a great deal of research has been conducted to determine their effects as boundary elements when flanking transgene cassettes. The results of a large number of such studies have indicated that flanking a transgene with MAR sequences increases the level of transgene expression and/or reduces plant-to-plant variations in transgene expression (Allen et al. 1993; 1996; van der Geest et al. 1994; Mlynárová et al. 1996; Cheng et al. 2001; Abranches et al. 2005; Halweg et al. 2005; Verma et al. 2005). However, these reports have been highly variable and in some cases, MAR sequences have been found to have the exact opposite effects (Breyne et al. 1992; Torney et al. 2004) or to have no discernable consequence on transgene expression at all (Ülker et al. 1999; van Leeuwen et al. 2001; Petersen et al. 2002; Sidorenko et al. 2003). Moreover, MARs have also been shown to increase transformation efficiencies (Meyer et al. 1988; Buising and Benbow 1994; Petersen et al. 2002; Zhang et al. 2007), reduce the occurrence of transgene silencing (Abranches et al. 2005; Verma et al. 2005), and silence a nearby promoter (Torney et al. 2004). While it seems that plant MARs possess a wide range of functions, the most common of which is the prevention of position-dependent chromatin effects, whether these elements function as enhancer-blocking insulators remains to be determined.

To investigate the role of MARs as enhancer-blocking insulators in plants, we assayed three well-characterized MARs, including the 1.2 kb Nicotiana tabacumRb7 3′ MAR (Conkling et al. 1990), the 0.9 kb Zea maysADH1 5′ MAR (Avramova and Bennetzen 1993), and a 2 kb Petunia hybridaTBS MAR (Meyer et al. 1988; Buising and Benbow 1994), for their ability to impede enhancer–promoter interactions. These particular MAR elements were chosen for our assay because they exhibit distinct functions when flanking a transgene cassette. While the Rb7 MAR generally functions as a typical plant MAR element by increasing the levels of transgene expression and/or reducing variability in transgene expression between independent lines (Cheng et al. 2001; Verma et al. 2005), the ADH1 MAR often exhibits a repressing effect (Brouwer et al. 2002; Torney et al. 2004) and the TBS enhances transformation efficiencies and promotes extra-chromosomal recombination (Meyer et al. 1988; Engels and Meyer 1992; Buising and Benbow 1994). In our study, the three MAR elements, as well as a 4 kb bacteriophage λ fragment to control for possible length effects, were inserted between the 35S enhancer, which is located in the region of −343 to −90 nucleotides upstream of the 35S promoter (Kay et al. 1987) and promotes strong and constitutive expression from nearby promoters (Kay et al. 1987; Jagannath et al. 2001; Yoo et al. 2005; Zheng et al. 2007), and AGIP, which directs GUS reporter gene expression specifically in third and fourth whorl organs (Sieburth and Meyerowitz 1997; Liu and Liu 2008). Subsequently, the leaves of Arabidopsis plants bearing the resulting transgene cassettes were assayed for GUS activity. Our results indicate that the petunia TBS fragment, but not the maize or tobacco MARs, is able to impede 35S enhancer activation of the AGIP without compromising the performance of either with elements from which they are not insulated.

Materials and methods

Gene cloning and plasmid constructs

Representations of the transformation vectors used in this study can be found in Fig. 1. All vectors were generated using standard protocols and are present in a pBIN19 (Bevan 1984) background. A 2.6 kb HindIII-EcoRI 35S::GUS::nopaline synthase (nos) terminator fragment was isolated from the pBI505 vector (Datla et al. 1990) and inserted into the binary vector, pBIN19, to generate pR2059. To produce the remaining control plasmids, the pPZP-RSC1 multiple cloning site (Goderis et al. 2002; Hajdukiewicz et al. 1994) was initially isolated via digestion with EcoRI and BamHI and inserted into pBIN19. The resulting vector was linearized with PI-PspI, and a fragment comprising the partially duplicated 35S promoter, enhanced green fluorescent protein (eGFP) coding sequence and nos terminator was introduced to create JM70. To generate JM69, the approximately 3 kb AGIP was isolated as described previously (Liu and Liu 2008) and inserted in the forward orientation immediately upstream of the GUS coding region and pAG7 transcriptional terminator. Subsequently, JM70 was linearized with AscI and the AGIP::GUS::pAG7 cassette was introduced in an orientation which allowed the AGIP to be adjacent to the 35S promoter from the 35S::eGFP::nos terminator cassette. To generate the JM79 vector, JM69 was digested with PI-PspI to remove the 35S::eGFP::nos terminator fusion.
Fig. 1

Schematic representation of constructs utilized in this study. All constructs shown are in a pBIN19 background. Vectors pR2059, JM79, and JM70 are control plasmids for GUS and eGFP expression driven by either the 35S promoter or AGIP. Construct JM69 is the base of all remaining vectors and contains the 35S::eGFP and AGIP::GUS cassettes in opposite orientations. Constructs JM81, JM82, JM83, JM85, and JM87 are derivatives of JM69, each bearing a different DNA sequence between the 35S promoter and AGIP. 35S: 35S promoter of the cauliflower mosaic virus; AGIP: AGAMOUS second intron-derived promoter; GUS: β-glucuronidase coding sequence; pAG7: procyclin-associated gene 7 terminator; nos-t: nopaline synthase transcriptional terminator; np: nopaline synthase promoter; eGFP: enhanced green fluorescent protein coding sequence; λ: lambda phage spacer sequence; LB: left T-DNA border sequence; RB: right T-DNA border sequence

A 920 bp fragment containing the ADH1 5′ MAR region (GenBank accession number AF123535) was cloned from the Z. mays genome using primers H15′MARU (5′-ACG GAT CCA AAC AGT CAC TTA GGA TAT GT-3′) and H15′MARL (5′-TGG TCC GTG GAC GTG GTT TTC GCT-3′). A 1,167 bp fragment harboring the Rb7 MAR element (GenBank accession number U67919) was isolated from the N. tabacum genome using primers RB7MARU (5′-TCG ATT AAA GCT TCC AAT TAT ATT TGG TCT-3′) and RB7MARL (5′-ACT ATT TTC AGA AGA AGT TCC CAA TAG T-3′). A 2,036 bp fragment comprising the TBS was isolated from P. hybrida cultivar V26 (GenBank accession number EU864306) using primers PetMARU (5′-TTC CTA ACA CCT GGA GAA CCT TTT ATG T-3′) and PetMARL (5′-AAG TTG TAA TGA GTT GCT GGC CTC TCT-3′). All isolated fragments were verified by DNA sequencing and inserted into the pAUX3132 vector (Goderis et al. 2002). Resulting clones were digested with I-CeuI and the desired fragments were ligated into the I-CeuI site within the pPZP-RSC1 multiple cloning region of JM69 (located between the 35S enhancer and AGIP) to generate JM81 (maize ADH1 5′ MAR), JM82 (petunia TBS MAR) and JM83 (tobacco Rb7 MAR) vectors.

To produce a vector containing a spacer DNA sequence in lieu of a MAR element to control for potential length effects, bacteriophage λ (Genbank accession J02459) was initially digested with NcoI to generate a 3,967 bp fragment (23901–27868), which was isolated and cloned into pAUX3131 (Goderis et al. 2002). The resulting plasmid was digested with I-SceI and the λ spacer was inserted into the pPZP-RSC1 multiple cloning site of JM69 to yield JM85.

The presence of the desired putative insulator between enhancer and promoter in each of the final vectors was confirmed by sequencing.

Plant growth, transformation, and regeneration

Each vector was introduced into Agrobacterium tumefaciens strain GV3101 by electroporation, and the resulting recombinant bacteria was used for the transformation of Arabidopsis thaliana ecotype C10 (a gift from David Mount, University of Arizona) via the floral dip method (Clough and Bent 1998). First generation transformants were selected by plating surface-sterilized seeds on media containing half-strength Murashige and Skoog basal salt mix, 1X Gamborg’s B5 vitamins, 1% (w/v) sucrose, 0.8% (w/v) agar and 60 μg/ml kanamycin. Plates were incubated in the dark at 4°C for 96 h prior to their transfer to a regime of 16 h light [30 μmol photons/(m2 s)] and 8 h dark at 22–24°C. Antibiotic resistant plantlets were transferred to soil after approximately 2 weeks of growth. All lines utilized in this study were phenotypically normal.

PCR analysis of transgenic plants

Genomic DNA was extracted from young leaves of various transgenic lines as described by Kobayashi et al. (1998). PCR analyzes were performed using approximately 100 ng DNA as template and GoTaq DNA Polymerase (Promega, Madison, WI, USA) according to the manufacturer’s instructions. To test for the presence of the putative insulator and spacer sequences in an assortment of transformed plants, primers pPZP-RCS1infwd (5′-TGT TTG CCA TCG CTA CCT TAG-3′) and pPZPRCS1inrev (5′-TCT CTT AAG GTA GCG AGC TC-3′) were utilized to amplify a 1,001 bp fragment from JM81 lines (including the maize ADH1 5′ MAR), a 2,117 bp fragment from JM82 lines (including the petunia TBS MAR), a 1,248 bp fragment from JM83 lines (including the tobacco Rb7 MAR) and a 4,048 bp fragment from JM85 lines (including the λ NcoI sequence). In each case, DNA from four independent lines was tested, while DNA from a single JM69 line was utilized as a negative control. The amplification regime consisted of 94°C for 5 min, 30 cycles of 94°C for 45 s, 55°C for 1 min, and 72°C for 3 min, followed by a single iteration of 72°C for 7 min. Resulting DNA fragments from one line transformed with each construct were cloned and sequenced to verify their identities.

To ascertain that the 35S promoter, GUS and AGIP sequences were intact in plants bearing constructs with a functional insulator, PCR tests were conducted on DNA from three independent lines of JM69 and JM82, respectively, as well as an untransformed control. Primers PZPR1 (5′-AAG GTA GCG ATG GCA AAC AGC-3′) and eGFPR2 (5′-CGT CCT TGA AGA AGA TGG TGC-3′) were utilized to amplify the entire 35S promoter sequence, AGI-IIF11 (5′-TTA AGG TTT CGT ATT AGA ATC AC-3′) and pBINF1 (5′-CAC ACA GGA AAC AGC TAT GAC C-3′) were employed to amplify the entire GUS coding sequence, and PZPF2 (5′-GCT ACC TTA AGA GAG GAT ATC-3′) and GUSR1 (5′-CCA CCA ACG CTG ATC AAT TCC-3′) were used to amplify the entire AGIP. In each case, amplification conditions were as above but extension times of 1.5 min (for the 35S promoter), 2.5 min (for the GUS coding region) and 3.5 min (for the AGIP) were utilized. In each case, a fragment of the expected size was obtained.

Histochemical staining and fluoremetric assay of GUS activity

Histochemical staining for GUS activity was carried out basically as described by Jefferson et al. (1987). Leaf and floral tissues from each transgenic line generated, as well as an untransformed control, were incubated in 1 mM 5-bromo-4-chloro-3-indolyl-β-d-glucuronide (X-gluc) (in 100 mM phosphate buffer, pH 7.0, 10 mM EDTA, 0.5 mM potassium ferrocyanide and 0.1% Triton X-100) at 37°C for 24–48 h (depending on the tissue analyzed). Following staining, tissue was depigmented in a series of 70% ethanol washes.

Fluorometric assays of GUS activity were measured essentially as described by Jefferson et al. (1987). Three 3-week-old leaves from a minimum of 12 independent lines containing each vector, as well as an untransformed control, were analyzed. Leaf tissue was ground in liquid nitrogen and lysed in extraction buffer (50 mM sodium phosphate, pH 7.0, 10 mM EDTA, 0.1% SDS and 10 mM β-mercaptoethanol). The resulting lysates were cleared by centrifugation and 20 μl of the resulting supernatant was added to 50 μl assay buffer (extraction buffer containing 1.4 mM 4-methylumbelliferone β-d-galactopyranoside (MUG)). Samples were incubated at 37°C in a moist chamber for 1 h, after which time 130 μl stop buffer (200 mM sodium carbonate) was added to each sample. Fluorescence was measured using a CytoFluor 4,000 multi-well plate reader (Applied Biosystems, Foster City, CA). Concentrations of methylumbelliferone (MU) generated were established from the linear regression slopes of fluorescence emitted by MU standards. Protein concentrations were determined using the Bio-Rad protein assay (Bio-Rad, Hercules, CA, USA) with bovine serum albumin as a standard. Each sample was assayed in duplicate. GUS activity was expressed as the mean value of pmol MU generated per minute per milligram of protein from the independent lines analyzed. Statistical analyzes were executed using the Mann–Whitney test for nonparametric data.

Quantitative real-time RT-PCR

Total RNA was isolated from leaf tissues of three lines transformed with pR2059, JM79, JM69, and JM82, as well as an untransformed control, using the RNeasy Plant Mini kit according to the manufacturer’s instructions (Qiagen, Valencia, CA). Contaminating DNA was removed from RNA samples using the DNA-free system (Ambion, Austin, TX, USA). For each line, 500 ng of RNA was reverse transcribed with the Superscript VILO cDNA synthesis kit (Invitrogen, San Diego, CA). cDNA products were diluted 1/100 and 1.5 μl were utilized in each quantitative real-time PCR assay, which were performed twice in triplicate on an ABI Prism 7900HT detection system (Applied Biosystems) using the SYBR Green Master Mix (Applied Biosystems). PCR quantifications were carried out with primers GUSRTF2 (5′-AGT GAA GGG CGA ACA GTT CCT GAT-3′) and GUSRTR2 (5′- TTC AGC GTA AGG GTA ATG CGA GGT -3′) to amplify a 183 bp fragment of the GUS transcript and actin-specific primers (Charrier et al. 2002) to amplify a 108 bp fragment of the internal standard transcript, Actin2. The conditions for PCR amplification were 95°C for 10 min, followed by 40 cycles of 95°C for 15 s and 60°C for 1 min. Assays including no template or template from RT reactions lacking reverse transcriptase were included in each trial. Dissociation curves were generated to ascertain only a single product was produced in each case. Relative gene expression was determined from standard curves produced using serial cDNA dilutions. All GUS expression data was normalized to that of Actin2.

GFP detection

A Typhoon Trio fluorescence scanner with the control v5.0 software (Amersham Bioscience, Piscataway, NJ, USA) was used to scan protein lysates and/or floral tissue for visualization of fluorescence emitted from either GFP or chlorophyll as described by Hily and Liu (2009). The resulting images were overlaid for the simultaneous visualization of both GFP and chlorophyll fluorescence. Acquired images were extracted using the Image Quant TL v2005 software.


Generation and confirmation of stable transformants

Control vectors (Fig. 1) included pR2059 (containing the 35S::GUS fusion), JM70 (containing the 35S::eGFP fusion), JM79 (containing the AGIP::GUS fusion) and JM69 (containing the 35S::eGFP and AGIP::GUS fusions in a head-to-head orientation). All remaining experimental vectors (Fig. 1) were adaptations of JM69 with various sequences inserted between the 35S enhancer and AGIP. To determine the effects of MARs on 35S enhancer-mediated interference with AGIP-driven expression of the GUS transgene, vectors containing either the maize ADH1 5′ MAR (JM81), petunia TBS MAR (JM82) or tobacco Rb7 3′ MAR (JM83) as inserts were generated. Plasmid JM85 contains a 4 kb λ fragment to control for sequence length. Each vector was subsequently transformed into Arabidopsis, and stable transformants were isolated. Although the number of transgene copies per genome was not determined in these lines, the floral dip method utilized for transformation has been found previously to result in the introduction of a single T-DNA insert in over 50% of transgenic lines with an average of 1.5 transgene inserts per line (Alonso et al. 2003; Rosso et al. 2003). Transformation efficiencies were comparable to those routinely observed in our laboratory and were similar for all constructs tested (data not shown).

In order to ascertain that stable transformants possessed the correct inserts, PCR analyzes were carried out using DNA extracted from at least four randomly chosen independent JM81, JM82, JM83 (containing MAR-derived inserts), and JM85 (containing the λ spacer sequence) lines respectively, as well as JM69 (lacking any insert) as a negative control. In each case, an insert of the correct size was amplified and confirmed through sequencing.

The petunia TBS MAR, but not the ADH1 5′ MAR or Rb7 3′ MAR, diminishes 35S enhancer-mediated activation of GUS expression in leaf tissues

Leaves from transgenic lines containing each of the vectors, respectively, as well as an untransformed control, were analyzed both histochemically and fluorometrically for GUS activity to determine whether MARs inserted between the 35S enhancer and AGIP could prevent enhancer-mediated interference of AGIP-driven GUS expression (Fig. 2). In each case, average GUS activities from populations including at least 12 independent lines were calculated to compensate for possible variations in positional effects and copy number. As expected, lines containing JM70 (which does not possess the GUS transgene), as well as the untransformed control, exhibited average GUS activities at very low background levels (6.5 ± 1.0 and 1.6 ± 1.9 pmol MU/(min mg) of protein, respectively), with no observable GUS staining in the leaves. Lines containing JM79 (which bears the stamen- and carpel-specific AGIP::GUS fusion) also yielded an average GUS activity near zero [0.7 ± 2.7 pmol MU/(min mg)] in leaf tissues. While the majority of JM79 lines displayed no GUS staining in the leaves, 13.8% exhibited very weak staining. It has been found previously that the 35S promoter sequence functions as an enhancer in either orientation (Odell et al. 1988) and is capable of overriding the control of transgene expression when present in the same construct (Jagannath et al. 2001; Yoo et al. 2005; Zheng et al. 2007). This also appears to be the case in our study, as lines bearing the 35S::GUS fusion (pR2059) or adjacent 35S::eGFP and AGIP::GUS fusions (JM69) both exhibited very high frequencies of GUS staining in the leaves (98.1 and 94.4%, respectively), as well as elevated average GUS activities [2,523 ± 635 and 4,411 ± 897 pmol MU/(min mg), respectively] indicating that the 35S enhancer is able to effectively activate the stamen- and carpel-specific AGIP promoter in nontargeted leaf tissue.
Fig. 2

Analysis of GUS expression in various transgenic lines and an untransformed control. Leaves from lines transformed with vectors pR2059, JM69, JM79, JM70, JM85, JM81, JM82, and JM83, respectively, as well as an untransformed control (wt), were assayed both histochemically and fluorometrically for GUS activity. Images depict GUS staining in leaves from representative lines transformed with each construct. Graphs display the mean GUS activities in pmol MU per minute per milligram of protein from at least twelve independent lines (n = number of lines tested). Bars indicate standard errors. Asterisks denote populations with average GUS activities that were significantly reduced (P ≤ 0.01) compared to lines containing the parent vector, JM69

Similarly, 96.7% of lines containing the 4 kb λ control spacer sequence (JM85) displayed a GUS activity [2,659 ± 587 pmol MU/(min mg)] comparable to that of the JM69 population, indicating that sequences of at least 4 kb do not impede enhancer-mediated activation of AGIP-driven GUS expression as a result of their size. Neither the maize ADH1 5′ MAR (JM81) nor the tobacco Rb7 3′ MAR (JM83) were able to prevent enhancer–promoter interference as evidenced by the high percentages of lines exhibiting strong GUS expression in the leaves (100 and 96.3%, respectively), and average GUS activities [4,296 ± 613 and 3,317 ± 574 pmol MU/(min mg), respectively] in the same range as JM69 lines. Conversely, the majority of lines containing the TBS (JM82) lacked any detectable GUS expression in vegetative tissues, while only 31% of plants displayed very weak levels of GUS in the leaves, resembling JM79 lines. Furthermore, the average GUS activity for JM82 lines [58 ± 17 pmol MU/(min mg)] was significantly reduced compared to that of JM69 at P ≤ 0.01.

Quantitative real-time RT-PCR analyzes of GUS expression levels in the leaves of three randomly chosen lines containing the vectors pR2059, JM69, JM79, and JM82, respectively, as well as an untransformed control, confirmed histochemical and fluorometric GUS expression data (Fig. 3). These results indicate that the TBS is able to prevent 35S enhancer-mediated activation of GUS transcription in the leaves.
Fig. 3

Quantitative real-time RT-PCR analysis of GUS expression in a selection of transgenic lines and an untransformed control. Total RNA extracted from the leaves of three independent lines transformed with pR2059, JM69, JM79, and JM82, respectively, as well as an untransformed control, was utilized as template for real-time PCR analysis using either GUS-specific or actin-specific primers. Levels of GUS and actin mRNA were obtained using standard curves derived from serial dilutions of cDNA. All real-time PCR tests were performed in triplicate and the median values were used for normalization of GUS mRNA levels with the endogenous control mRNA, actin. Each block represents the mean normalized value of GUS mRNA for each construct. Bars denote standard errors

The TBS does not function as a silencer or interfere with transgene stability

It is possible that the reduction in GUS expression in the leaves of JM82 lines was caused by an effect of the TBS on the stability of the surrounding transgene sequence, leading to the deletion or truncation of the 35S enhancer, AGIP or GUS coding region. Therefore, we tested three independent JM69 and JM82 lines by PCR to analyze the integrity of the 35S enhancer, GUS coding region and AGIP (Fig. 4a). For every line assayed, fragments of identical size were amplified for the 35S enhancer, GUS coding region, and AGIP, respectively, indicating that intra-fragment rearrangement, recombination or deletion had not occurred in any of the plants tested.
Fig. 4

Analysis of transgene integrity and promoter activity in JM82 lines. a Three lines of JM69 and JM82 (containing the insulating petunia TBS MAR), respectively, as well as an untransformed control, were analyzed by PCR for the presence of the 35S promoter, GUS coding sequence and AGIP. In all transgenic lines tested, an appropriately sized fragment was amplified. M molecular weight ladder. (b and c) Native activities of the 35S promoter and AGIP were tested in JM69, JM79, JM70 and JM82 transgenic lines. Representative lines are shown in each case. Histochemical GUS assays were performed on floral tissue (b) and GFP fluorescence was scanned in inflorescences with associated cauline leaves (c) to ensure that the functions of the AGIP and 35S promoter, respectively, were not inhibited in JM82 lines. Results indicate that the presence of the TBS fragment does not destabilize transgene integrity or silence nearby promoters

To rule out the possibility that the TBS represses GUS expression through an inherent silencer function, histochemical GUS assays of floral tissue were carried out to ascertain that the activity of the AGIP was not affected in JM82 lines (Fig. 4b). As expected, no GUS staining was observed in flowers from lines containing JM70 or the untransformed control. Flowers from lines bearing the JM79 vector (which contains the AGIP::GUS cassette) displayed staining in a carpel- and stamen-specific pattern. JM69 lines displayed GUS expression in carpels and stamens as well as sepals and petals, again confirming that the 35S enhancer overrides AGIP-driven GUS expression when present in the same construct. Lines containing the TBS MAR element (JM82) exhibited staining only in stamens and carpels, resembling lines transformed with the JM79 construct. These findings indicate that the promoter activity of the AGIP is not compromised by the presence of the TBS fragment in any of the lines analyzed.

To further preclude a possible silencing effect of the TBS, a random subset of the lines tested for floral GUS expression were also analyzed for 35S promoter activity via GFP fluorescence (Fig. 4c). As anticipated, lines transformed with JM79 (n = 15), which does not possess the eGFP transgene, and the untransformed control, did not exhibit any GFP fluorescence. Conversely, 46.2 and 16.7% of lines containing the positive control constructs JM69 (n = 13) and JM70 (n = 30), respectively, and 37.5% of JM82 lines (n = 24) displayed GFP fluorescence with comparable intensities, indicating that 35S promoter activity was not affected by the presence of the TBS fragment. Although a relatively large proportion of plants exhibited silencing of eGFP, the phenomenon is not related to the presence of the TBS and instead appears to be an inherent property of the GFP sequence itself, as reported previously (for example Pang et al. 1997; Andika et al. 2005; de Folter et al. 2007; Hily and Liu 2009).

Taken together, these results provide evidence that the TBS MAR element does not interfere with transgene stability or have a silencing effect on either the 35S promoter or the AGIP. This implies that the TBS functions as a true enhancer-blocking insulator by impeding 35S enhancer activation only when positioned between the enhancer and its adjacent promoter.


We have addressed the question of whether MAR sequences have enhancer-blocking properties in plants by analyzing GUS activity in stable transformants containing vectors with the maize ADH1 5′ MAR, petunia TBS MAR, or tobacco Rb7 3′ MAR interposed between a 35S enhancer and AGIP::GUS fusion. To control for insert sequence length effects, stable transformants bearing constructs with a 4 kb λ NcoI fragment in place of a MAR element were also analyzed. While the ADH1 MAR, Rb7 MAR and λ spacer were unable to prevent enhancer–promoter interference, the TBS MAR effectively blocked interactions between the 35S enhancer and AGIP (Figs. 2, 3).

Previous studies have demonstrated that the P. hybrida TBS element (Meyer et al. 1988) possesses a different suite of functions than the typical plant MAR. Unlike the majority of other MARs, this sequence does not have any influence on transgene expression when included in the transformation vector (Petersen et al. 2002), but instead increases transformation efficiencies (Meyer et al. 1988; Buising and Benbow 1994; Galliano et al. 1995; Petersen et al. 2002). Although the exact mechanism by which this sequence functions is not known, it has been suggested that it may enhance recombination of transgenic vectors into the genome (Meyer et al. 1988; Buising and Benbow 1994) since it has also been shown to facilitate extra-chromosomal recombination in plant cells (Engels and Meyer 1992; Galliano et al. 1995).

Because of these recombination-related functions, it is possible that the TBS fragment utilized in this study inhibited GUS expression by increasing the probability of transgene rearrangements or truncations. However, our analyzes confirmed that the various components of the transgene were intact in every insulated line analyzed (Fig. 4a). While it is possible that the TBS catalyzes rearrangements at precise locations resulting in translocations that unlink the respective transgene cassettes, no such site-specific recombination function has been reported for the TBS as of yet. Hence, our results indicate that the enhancer-blocking function of the TBS fragment is mechanistically distinct from its role in the enhancement of recombination. Moreover, we also demonstrated that the diminishment of GUS expression in JM82 lines was not the result of an inherent capability of the TBS to silence nearby regulatory sequences since the flower-specific activity of the AGIP promoter was not compromised in these plants (Fig. 4b, c). Thus, our findings offer compelling evidence that the TBS functions as a true enhancer-blocking insulator in Arabidopsis, providing yet another novel function for this MAR element.

To date, only a very small number of sequences with enhancer-blocking activity have been identified in plants (van der Geest and Hall 1997; Jagannath et al. 2001). Interestingly, one such sequence comprised the β-phaseolin 3′ MAR from Phaseolus vulgaris (van der Geest and Hall 1997). However, like the ADH1 and Rb7 MAR elements utilized in this study, the β-phaseolin 5′ MAR was not able to elicit the same insulating effect. While the authors attributed the enhancer-blocking activity of the 1.3 kb β-phaseolin 3′ MAR exclusively to its length and suggested an enhancer facilitating function for the 1.0 kb 5′ MAR, there is evidence that the 35S enhancer can exert effects over distances as large as 78 kb (reviewed by Yoo et al. 2005), which indicates that relatively small lengths of spacer sequence between the 35S enhancer and target promoter would not necessarily block interactions. Our results provide additional evidence that the length of a spacer sequence (up to 4 kb) is not correlated with its enhancer-blocking function. Although the 2 kb TBS MAR element, which was the longest of the three MAR fragments analyzed in our study, was the only insert capable of functioning as an enhancer-blocking insulator, the fact that a 4 kb λ fragment could not induce the same effect suggests that its ability to block enhancer–promoter interference is not simply a result of its length. Instead, the TBS MAR, like some metazoan MARs (Stief et al. 1989; Nabirochkin et al. 1998) and possibly also the β-phaseolin 3′ MAR, likely possesses inherent sequence motifs that prevent interactions between an enhancer and promoter when situated between them.

Because MAR elements are generally relatively large, independent elements with various functions could be concealed within them (Holmes-Davis and Comai 1998). One such cis-acting element that may contribute to insulator activity is a promoter, since several insulators have been found to contain promoter sequence motifs as well as promoter activities (Smith and Corces 1995; Hogga and Karch 2002), which potentially interfere with communication between an enhancer and a promoter when placed between them (Kellum and Schedl 1992). While the possibility that the TBS element exhibits promoter activity cannot be ruled out at this point, promoter-like regions could not be identified within this sequence using bioinformatic tools (data not shown). It is also feasible that the TBS contains unknown protein-binding sites that enable its enhancer-blocking function, since several DNA-binding proteins that are sufficient to impede enhancer–promoter communication have been identified in animals (Parkhurst et al. 1988; Bell et al. 1999; Gaszner et al. 1999). Unfortunately, no such characterizations have been carried out in plants due to the paucity of information concerning enhancer-blocking insulators in this system.

Our finding that the TBS, but not the ADH1 or Rb7 MARs, functions as an enhancer-blocking insulator in Arabidopsis is not all that surprising since the effects of MARs on transgene expression in plants, as well as the enhancer-blocking effects of MARs in animals, are generally highly variable (for example Stief et al. 1989; Kellum and Schedl 1992; Petersen et al. 2002; Majumder and Cai 2003; Torney et al. 2004). It has been suggested that the broad range of conflicting results regarding the effects of MARs may be due to the use of different experimental parameters (reviewed by Torney et al. 2004). However, it is also possible that MARs are a heterogeneous group of elements that share only the capacity to bind the nuclear matrix (Holmes-Davis and Comai 1998). Our results imply that the faculty of a DNA fragment to hinder enhancer–promoter interactions cannot be predicted solely by its ability to interact with the nuclear matrix and that the enhancer-blocking function of the P. hybrida TBS MAR is almost certainly the result of alternative, as of yet unidentified, sequence motifs contained within its sequence.


The authors wish to thank Mr. Dennis Bennett (USDA-ARS, Kearneysville, WV) for his technical assistance, and Dr. Ann Callahan (USDA-ARS, Kearneysville, WV) for critical reading of the manuscript. This study was funded by the USDA-ARS Headquarter 2005 and 2007 classes of postdoctoral grants and a USDA CSREES BRAG grant (2006-03701).

Copyright information

© Springer-Verlag 2009