Functional pleiotropy of an intramolecular triplex-forming fragment from the 3'-UTR of the rat Pigr gene

ISABEL FABREGAT1, KATHERINE S. KOCH2, TAKASHI AOKI4, ALLAN E. ATKINSON2, HUONG DANG2, OLGA AMOSOVA5, JACQUES R. FRESCO5, CARL L. SCHILDKRAUT6 and HYAM L. LEFFERT2,3

1 Department of Bioquimica y Biologia Molecular, Instituto de Bioquimica, Centro Mixto CSIC/UCM, Facultad de Farmacia, Universidad Complutense, 28040 Madrid, Spain
2 Department of Pharmacology
3 Center For Molecular Genetics, School of Medicine, University of California, San Diego, La Jolla, California 92093
4 Department of Biochemistry, Faculty of Pharmaceutical Sciences, Health Sciences University of Hokkaido, Ishikari-Tobetsu, Hokkaido 061-0293, Japan
5 Department of Molecular Biology, Princeton University, Princeton, New Jersey 08544
6 Department of Cell Biology, Albert Einstein College of Medicine, Bronx, New York 10461


    ABSTRACT
 TOP
 ABSTRACT
 INTRODUCTION
 MATERIALS AND METHODS
 RESULTS
 DISCUSSION
 REFERENCES
 
A microsatellite-containing 359-bp restriction fragment, isolated from the rat Pigr gene (murine polymeric immunoglobulin receptor gene) 3'-untranslated region (3'-UTR) and inserted into 3'-UTR or 3' flanking positions in transcription units of supercoiled plasmids, attenuates luciferase reporter gene expression in orientation- and position-dependent ways following transient transfection of human 293 cells. The same fragment stimulates orientation-dependent gene expression in a 5' flanking position. Plasmid linearization abrogates both orientation- and position-dependent responses. Cell-free translation reveals that 5' and 3' flanking expression responses are proportional to increased and decreased luciferase mRNA levels, whereas 3'-UTR expression is associated with control mRNA levels. Hypersensitivity to nucleases S1 and P1, gel mobility differences between supercoiled plasmids carrying opposing microsatellite orientations, and anomalous melting profiles of this fragment are also observed. These results suggest that functional pleiotropy of this fragment depends on the DNA context of its purine-rich microsatellite strand and on DNA supercoiling. Intramolecular triplexes stabilized by supercoiling and secondary structures of purine repeat-rich mRNAs may also confer regulatory properties to similar genomic elements.

genome regulators; polypurine-polypyrimidine elements; d(GGA)n and d(GAA)n triplet repeats


    INTRODUCTION
 TOP
 ABSTRACT
 INTRODUCTION
 MATERIALS AND METHODS
 RESULTS
 DISCUSSION
 REFERENCES
 
MICROSATELLITES [poly-dispersed simple tandemly repeated microsatellites (STMSs)] are widely dispersed throughout eukaryotic genomes (10). Recent genetic evidence indicates that unstable STMSs exert deleterious biological effects. Notably, CGG, CTG, and GAA repeat expansions are implicated causally in 13 inherited neurological diseases (59). Expanded tracts are found both within and outside of transcription units; they are structurally diverse, and may disrupt gene expression in different ways (64).

Increasing evidence suggests that several STMSs constitutively regulate gene expression without expansion (17, 30, 60, 61). Some of these STMSs include nuclease S1-sensitive [TCC]n=3 repeats in the (+)strand (i.e., the native or forward orientation) of the epidermal growth factor promoter (24); intramolecular triplex (H-DNA) forming elements in the (+)strand of the c-MYC promoter (11); DNase I hypersensitive CT repeats in chromatin of the Drosophila hsp26 promoter (18, 39); a bifunctional S1-sensitive (+)strand [CT]n=7 repeat motif in the chicken malic enzyme promoter (66); and a diverse group of purine-rich (+)strand polypurine-polypyrimidine (R-Y) elements that share requirements of (+)strand orientation to exert their genetic effects (1, 6, 19, 53).

One such R-Y element, a transcribed triplet repeat-rich tract from the 3'-untranslated region (3'-UTR) of the rat polymeric immunoglobulin receptor gene (Pigr; Refs. 1, 29, 30), may be a site of RNA polymerase slippage during liver regeneration. During this proliferative transition, six Pigr cDNAs have been observed with STMS sequences that do not reflect predicted genomic coding sequence ([GGA]n=15[GAA]n=5AA[GAA]n=34); instead, expanded GGA (n = 24–26) or truncated GAA repeat motifs (n = 9–37) are revealed, and two different classes of polymeric immunoglobulin receptor protein (PIgR) mRNA have been postulated (30). The nature and purpose of this unusual RNA processing is unclear.

To investigate further the regulatory properties of this unusual Pigr tract, comparative biological and physical studies were performed in model systems using 359-bp and 400-bp restriction fragments containing this STMS. The results show that this genomic fragment exerts orientation- and position-dependent effects pleiotropically at DNA and RNA levels. These effects depend on DNA supercoiling and, as suggested from physical findings, the formation of different kinds of intramolecular triplexes along the STMS.


    MATERIALS AND METHODS
 TOP
 ABSTRACT
 INTRODUCTION
 MATERIALS AND METHODS
 RESULTS
 DISCUSSION
 REFERENCES
 

Construction, configuration, characterization, and linearization of DNA plasmids.
Supercoiled pSP72-based expression plasmids were constructed, purified, and characterized by standard procedures (1). Control plasmids contained transcription units composed of a 0.85-kb human cFOS 5' flanking tract (-711 to +45), a 1.6-kb firefly luciferase cDNA reporter, and a 0.8-kb SV40 polyadenylation cassette ("pA1" or "pA2," differing in flanking restriction sites). Experimental plasmids were constructed by subcloning 359-bp HaeIII-TaqI fragments into control plasmids. The resulting plasmids carried the STMS either in forward (MS) or reverse (SM) orientations at one of three positions: 5' flanking (upstream of the promoter), 3'-UTR (upstream of the SV40 pA2 cassette), and 3' flanking (downstream of the SV40 pA1 cassette). To construct the 5' flanking set, the HaeIII-TaqI fragment from pSPMS (1) was subcloned into the 2.8-kb pUC18 cloning vector (GenBank accession no. X70275). HindIII fragments from the pUC18 intermediate were inserted into pSPFOSLucpA1 (Ref. 1), upstream of the promoter in MS (->) or SM (<-) orientation. Construction strategies for the 3'-UTR and 3' flanking sets, and a plasmid set carrying STMSs without pA cassettes (pA-), were described elsewhere (1). Supercoiled plasmids (10 µg DNA) were linearized by overnight digestion at 37°C with 20 U BglI (GIBCO-BRL), purified by phenol/CHCl3 extraction, and analyzed by gel electrophoresis (30).

Transient transfection and reporter gene expression assays.
Human 293 cells were cultured in Dulbecco and Vogt’s modified Eagle’s medium (DVME) containing heat inactivated 10% bovine calf serum (BCS; 56°C, 30 min) under an atmosphere of 90% air and 10% CO2 at 37°C. Cells were plated into 3-cm Falcon plastic tissue culture dishes (1 x 105 cells/dish). Culture fluids were aspirated 2–3 days later (log phase, 50% confluence). The cells were shifted without washing into 2 ml fresh media and transfected by standard procedures (1). In each individual experiment, separate groups of cells were mock-transfected (vehicle only), or transfected with control plasmids pA1 or pA2 lacking the STMS, or transfected with experimental MS or SM plasmids. To normalize functional measurements in case of variable levels of plasmid uptake, all transfected groups received equal amounts of a second cotransfected plasmid, pRSVßGal, expressing an unrelated marker [ß-galactosidase (ßGal)]. After 15 h, the cultures were fluid changed; luciferase, ßGal activity, and percentage of ßGal+ cells/dish were measured 22 h later (1). The normalized results were pooled from the average absolute mean values (±SE; N = no. of culture dishes or pooled RNA samples) of 3–13 individual experiments for each treatment group. Tests of statistical significance (*P << 0.01) or nonsignificance (NS = P >> 0.05) were determined by paired (equal numbers of samples per treatment group) or unpaired (unequal numbers of samples per treatment group) Student’s t-tests using a SigmaPlot v.4.0 software program (SPSS, San Rafael, CA).

Primary fetal rat hepatocytes were plated (3.75 x 105 cells/3-cm dish) by standard procedures (36). Two or 6 days later, the cultures were shifted into 2 ml plating media containing 4% dialyzed fetal bovine serum (dFBS) and transfected for 5 h at 3% CO2 (Ref. 54). DNA CaPi coprecipitates (150 µl/dish) contained 0.5 µg and 2.0 µg of test plasmid and pRSVßGal, respectively (1). Following transfection, the cultures were shifted into fresh media containing 15% dFBS diluted 50% with conditioned media from 2- or 6-day-old sister cultures (27). Cell extracts were prepared 22 h later, and enzyme assays were performed as described above.

Extraction of supercoiled DNA from nuclei and Southern blotting.
Plasmid DNA molecules were extracted from sister cultures of cotransfected cells when reporter gene expression assays were made (see Hirt, Ref. 20). RNA-free cellular and control plasmid DNA samples (excluding undigested pRSVßGal) were linearized with BglI, purified, resuspended in TE buffer (10 mM Tris and 1 mM EDTA, pH 8) and quantified by A260/A280 ratios. Hirt extracts or DNA pooled from two mock-transfected cultures were loaded into lanes of 0.8% agarose gels (TBE buffer: 89 mM Tris base, 89 mM boric acid, 2 mM EDTA, pH 8.0; 0.5 µg EtBr/ml, where EtBr is ethidium bromide) and electrophoresed for 3 h at 60 V (see Ref. 28). The gels were soaked consecutively (twice for 30 min at 21°C) in 1.5 M NaCl and 0.5 M NaOH, then in 1.5 M NaCl and 0.5 M Tris·HCl, pH 7.5; DNA was blotted by capillary action onto Nylon membranes (GeneScreen). Prehybridization and hybridization buffers contained 50% vol/vol deionized formamide, 0.025 M K3PO4, 5x SSC, 5x Denhardt’s solution, 0.1% SDS, 50 ng salmon sperm DNA/ml, and 10% wt/vol dextran sulfate. DNA hybridization probes were generated by PCR in 50 µl reaction buffers containing 2 U Taq polymerase (GIBCO-BRL) and 1 mM MgCl2 as follows: 179-bp human cFOS probe [10 ng pUC19FOSFRTLuc DNA substrate (Ref. 31), 0.5 µM upper 5'-gttcccgtcaatc and lower 5'-gcgtgtcctaatc primers, 1 min at 95°C, 1 min at 94°C, anneal at 49°C, extend at 72°C for 15 s for 30 cycles, hold at 4°C]; and 629-bp ßGal probe [10 ng pLZRNL DNA substrate (Ref. 31), 0.5 µM upper 5'-tggcagggtgaaac and lower 5'-gcgtcgtgattagc primers, 1 min at 95°C, 1 min at 94°C, anneal at 52.7°C, extend at 72°C for 1 min for 30 cycles, hold at 4°C]. PCR products were isolated on low-melting point (LMP) agarose gels (GIBCO-BRL) and purified on Qiaex II columns as described by the vendor (Qiagen, Santa Clarita, CA). The products were labeled with [{alpha}-32P]dCTP by Ready-To-Go random priming kits (Pharmacia) to 5–10 x 108 cpm/µg DNA. Blotted membranes were incubated for 24 h at 42°C in the presence of 2–3 x 106 cpm 32P-labeled probe/ml hybridization solution. The membranes were washed twice for 10 min at 21°C with 2x SSC, and for 30 min at 42°C with 1x SSC, 0.1% SDS. Autoradiography was performed at -70°C with Kodak screens and X-OMAT film (Rochester, NY). Additional washing (0.2x SSC, 0.5% SDS) at 50–55°C was performed to reduce background. Prior to second rounds of hybridization (ßGal probe), membranes were incubated in stripping solution (10 mM Tris, pH 8.0, 1 mM EDTA, 1% SDS) for 30 min at 100°C and autoradiographed to verify removal of bound radioactivity.

Cell-free RNA translation assays.
Total and cytoplasmic RNAs were extracted from 293 cells using Qiagen RNeasy Mini kits and procedures supplied by the vendor. RNA samples were pooled from duplicate dishes (20–30 µg RNA/dish) into 60 µl eluates, quantified (A260), characterized by gel electrophoresis, and stored in 1.5-ml Eppendorf tubes at -70°C. Functional luciferase transcripts were quantified by in vitro cell-free translation assays (56). Working solutions, prepared from kits immediately before use (Promega, catalog no. L4380), contained wheat germ extract (200 µl), 0.15 mM amino acid stocks (14 µl), and 0.11 M potassium acetate (30 µl). Translation reactions in 1.5-ml Eppendorf tubes were started by adding 17 µl working solution to 8 µl of a solution containing 2 µg RNA (adjusted with RNase-free H2O) and, when noted, 2.5 µl of aqueous stock solutions of 0.1 mg RNase/ml or 0.1 mg cycloheximide/ml. Translation assays with Promega TNT kits were performed using the following components: 2 µg RNA, 12.5 µl wheat germ extract, 1.0 µl reaction buffer, 0.5 µl amino acid stock (minus Leu), 0.5 µl amino acid stock (minus Met), and diethyl pyrocarbonate-treated H2O (to bring the final solution to 25 µl). After incubations of 1.5 h at 30°C, 100 µl of Promega Luciferase Assay Reagent were added to each tube, and luciferase activity (LU/2 µg total RNA, where LU is light units) was quantified in triplicate by standard luminescence measurements at 10 s. Expression levels were corrected for ßGal staining frequencies: 5' flanking (7.6–9.1%), 3'-UTR (12.6–12.8%), and 3' flanking (5.7–5.8%), the differences between which were not statistically significant.

Digestion of plasmid DNAs with nucleases S1 and P1.
DNA plasmids (2.5 µg/5 µl TE) were incubated without or with 10 U nuclease S1 (Aspergillus oryzae, EC 3.1.30.1; Roche Molecular Biochemicals, Indianapolis, IN) in 15 µl reaction buffer (0.2 M NaCl, 0.05 M sodium acetate, and 1 mM ZnSO4, pH 4.5) for 30 min at 37°C. In separate assays, 2.0 µg of plasmid DNA were incubated without or with 10 µg of nuclease P1 (Penicillium citrinum, EC 3.1.30.1; Roche) in 25 µl of P1 reaction buffer (10 mM Tris·HCl, 10 mM MgCl2, 50 mM NaCl, pH 7.6; plus deionized H2O to bring the final solution to 50 µl) for 10 min at 37°C. The reactions were stopped by addition of EDTA to a final concentration of 37 mM. Reaction products were purified by standard procedures, and digested with 10 U of BglI for 3 h at 37°C. In control experiments, plasmid DNAs were digested first with BglI followed by nuclease S1 or P1, electrophoresed on 0.8% agarose gels (containing 0.5 µg EtBr/ml, pH 8) for 2.5–3 h at 50–60 V, visualized by transillumination, and sized by extrapolation from semilog plots of apparent Mr vs. migration distance. Gel photographs were scanned (HP Scanjet 4c) and collated using Microsoft Paintbrush and Paint Shop Pro (JASC, Minnetonka, MN).

Preparation of gel-purified DNA fragments for DNA melting assays.
Plasmid pSPFOSLucMSpA2 (Ref. 1) was used as the source of the STMS fragment. The plasmid was purified twice through CsCl, then digested with HindIII and BamHI to generate a 400-bp product carrying the HaeIII-TaqI fragment. The 0.4-kb product was isolated on 1% LMP agarose gels, purified over Qiagen Qiaquick columns using procedures supplied by the vendor, and further characterized by gel electrophoresis. Additional column washing steps were employed to minimize LMP agarose contamination of the final material (A260/A280 > 1.80). A 747-bp HindIII-EcoRI Aequorea green fluorescent protein (GFP) cDNA insert devoid of STMS motifs was prepared from plasmid pRAY (Invitrogen, La Jolla, CA) using identical procedures. Purified DNA fragments were dried with a speed vacuum. Absorbance-temperature profiles were obtained in 0.15 M NaCl, 10 mM cacodylate(Na+) at the desired pH, using an Aviv UV-VIS model 14DS computer-driven spectrophotometer equipped with a thermoelectric unit programmed to increase temperature stepwise every 2°C. The temperature was raised rapidly, and the mixture was allowed to equilibrate for 9 min before making absorbance measurements. Data were collected at set wavelengths and computer plotted (35). Melting profiles were reproduced three times.


    RESULTS
 TOP
 ABSTRACT
 INTRODUCTION
 MATERIALS AND METHODS
 RESULTS
 DISCUSSION
 REFERENCES
 

Genomic sequence of the STMS.
The sequence of the 359-bp HaeIII-TaqI STMS restriction fragment derived from exon 11 of the rat Pigr 3'-UTR (29, 30) is shown in Fig. 1. The microsatellite in this fragment is composed of an R-Y element that consists of a 60-nt G-rich tract, followed by two neighboring GGA and GAA triplet repeat motifs ([GGA]n=15[GAA]n=5AA[GAA]n=34). This R-Y element is conserved in the mouse Pigr gene (34); as shown in Fig. 1, alignments of both of its (+)strands are significantly homologous (P < 1 x 10-67). Additional polypurine and pyrimidine repeat motifs in the rat fragment, described in detail elsewhere (1), also occur in the mouse STMS.



View larger version (60K):
[in this window]
[in a new window]
 
Fig. 1. Genomic alignments of murine polymeric immunoglobulin receptor gene (Pigr) 3'-untranslated region (3'-UTR) poly-dispersed simple tandemly repeated microsatellites (STMSs). Rat (U08273) and mouse (AB001489) alignments were made using NCBI MACAW software as described elsewhere (30). Shaded regions define sites of perfect homology. GA-rich, [GGA]n=11–15, and [GAA]n=39–60 repeat domain start sites are indicated by arrows ({downarrow}, rat; {uparrow}, mouse). CHI-like, G5 and A5 motifs, and bHLH E-boxes are enclosed by rectangular bars. Histone-like stem loop and CT-rich tracts are underlined.

 
Reporter gene expression dependence on orientation and position of the STMS.
To investigate further the functional properties of the 359-bp fragment, by varying its position with respect to a transcription unit, human 293 cells were transfected with three different sets of supercoiled expression plasmids. Three functional response patterns were observed: the results are shown in Fig. 2 and tabulated in Fig. 3. With respect to the 5' flanking set, the MS plasmid significantly stimulated luciferase reporter gene expression ~2.2-fold (P < 10-5) compared with equivalent responses produced by the SM and control plasmids. As reported previously (1) for the 3'-UTR set (the STMS position in the Pigr gene), the MS plasmid significantly attenuated luciferase expression1 to 47% of the control plasmid level (P < 10-10), whereas gene expression fell only slightly to 85% of the control level (P ~ 0.04) after transfection with the SM plasmid. Similarly, with regard to the 3' flanking set (1), the MS plasmid significantly attenuated gene expression to 43% of control levels (P < 0.01); in contrast, no statistically significant attenuation of gene expression was produced by the SM plasmid (Fig. 2). Taken together, these observations indicate that when its purine-rich strand is in a forward orientation, the same STMS-carrying fragment is genetically pleiotropic: depending on its transcription unit position, it either stimulates or attenuates gene expression when it is situated upstream or downstream of a promoter, respectively.



View larger version (20K):
[in this window]
[in a new window]
 
Fig. 2. Luciferase reporter gene expression in 293 cells: in vivo cotransfection with supercoiled plasmids. The experiments were performed as described in MATERIALS AND METHODS and the legend to Fig. 3. The numbers of individual culture dishes (N) in each treatment group are indicated along the baseline. Statistical significance was determined as described in MATERIALS AND METHODS and the legend to Fig. 3. The calculated P values in each plasmid set are as follows: 5' flanking (control {leftrightarrow} MS, 1 x 10-5; MS {leftrightarrow} SM, 8 x 10-5; control {leftrightarrow} SM, 0.29), 3'-UTR (control {leftrightarrow} MS, 1 x 10-10; MS {leftrightarrow} SM, 1 x 10-8; control {leftrightarrow} SM, 0.035), and 3' flanking (control {leftrightarrow} MS, 0.01; MS {leftrightarrow} SM, 0.005; control {leftrightarrow} SM, 0.123). "MS" indicates the 359-bp rat Pigr HaeIII-TaqI genomic fragment, forward orientation; "SM" is the reverse orientation.

 


View larger version (20K):
[in this window]
[in a new window]
 
Fig. 3. Tabulated summary of luciferase reporter gene expression in 293 cells. Key transcription unit elements in the plasmid diagrams are defined in the graphic legend at bottom. Left column: in vivo cotransfection with supercoiled plasmids. Control plasmid levels (100% maximal) were 5' flanking, 36,300 ± 4,800 (N = 20); 3'-UTR, 32,900 ± 2,900 (N = 37); and 3' flanking, 53,400 ± 9,400 (N = 9) LU/mU ßGal. The experimental data and P values are provided in Fig. 2. Middle column: in vivo cotransfection with linearized plasmids. Control plasmid levels (100% maximal) were 5' flanking, 4,479 ± 763 (N = 10); 3'-UTR, 2,283 ± 393 (N = 26); and 3' flanking, 6,170 ± 739 (N = 6) LU/mU ßGal. No statistically significant differences were observed among or between the individual measurements within each set. The experimental data and P values are provided in Fig. 5. Right column: in vitro cell-free translation. Translation-specific control levels (100% maximal) were 5' flanking, 2,252 ± 434 (N = 10); 3'-UTR, 10,252 ± 4,957 (N = 11); and 3' flanking, 1,680 ± 315 (N = 15) cycloheximide-sensitive LU/2 µg total RNA. Luminometer and buffer control backgrounds were 80–100 light units (LU). The experimental data and P values are provided in Fig. 6.

 
To investigate the possibilities that differential cellular uptake or stability of supercoiled plasmid DNA contributed to differential responses, intracellular plasmids were isolated from nuclei of cotransfected cells and, following linearization by digestion with BglI, quantified by Southern blots. As shown in Fig. 4, despite some degradation, similar levels of plasmid DNAs were detected in cellular DNA samples harvested from cultures transfected by each member of the three plasmid sets. The levels of plasmid degradation were limited, variable, and unrelated to structural differences among the plasmids. Thus the bulk of the hybridizable material was observed as full-length plasmid (in all cases); almost no degradation was observed in the 3'-UTR and 3' flanking cases; and, when present, degradation was observed with control plasmids, MS and SM plasmids, as well as with ßGal-expressing control plasmids extracted from cotransfected cells.



View larger version (81K):
[in this window]
[in a new window]
 
Fig. 4. Plasmid DNA stability following cotransfection of 293 cells with supercoiled plasmids. Nuclear levels of plasmid DNAs were quantified by Southern blots, as described in MATERIALS AND METHODS. Cellular DNAs from 5' flanking (top), 3'-UTR (middle), and 3' flanking sets (bottom) were obtained from mock-transfected cells or from cells transfected with control plasmids or plasmids carrying HaeIII-TaqI fragments in MS or SM orientation. Blotting controls included three expression plasmids from each set, pSP72 and pRSVßGal. Numbers along y-axes are size (in kb); "s.c." denotes supercoiled pRSVßGal. Films were exposed for 10 min (cFOS probe) and 3 h (ßGal probe) at -70°C (top) or for 45 and 90 min (both probes) at 21°C (middle and bottom, respectively).

 
To investigate the possibility that soluble luciferase inhibitors or activators contributed to differential responses, mixing experiments were performed with freshly harvested extracts from transfected cultures, and luciferase activities were reanalyzed. No evidence for such effectors was detected (not shown).

To investigate cellular and species specificity of the effects of the STMS-carrying fragment, cotransfections were performed with fetal rat hepatocytes (Table 1). Absolute reporter gene expression levels in this primary system between 2–3 days (lag phase) and 6–7 days postplating (log phase) were less than the levels observed in the human 293 cells. However, the relative orientation- and position-dependent differences were greater than those observed in human cells, and all three plasmid sets produced responses quantitatively (5' flanking) and directionally similar to those observed with human 293 cells. In addition, the extent of the MS plasmid responses was related to hepatocyte growth cycle state (37). Thus the MS-containing plasmid (5' flanking) stimulated luciferase expression more during the lag (~3.1-fold) than during the log phase (~1.7-fold). In contrast to the levels of ~40–50% attenuation of expression in the 293 cells, the MS-containing plasmid markedly attenuated hepatocyte luciferase expression to 3% of the control during lag phase (3' flanking), and it completely extinguished luciferase expression during log phase (3'-UTR); SM-containing plasmids generated intermediate responses in both 3' flanking and 3'-UTR positions at both growth cycle times.


View this table:
[in this window]
[in a new window]
 
Table 1. Effects of the 359-bp HaeIII-TaqI DNA fragment in primary cultures of fetal rat hepatocytes

 
Reporter gene expression dependence on DNA supercoiling.
To investigate the role of DNA structure on the pleiotropic effects of this fragment, supercoiled expression vectors were converted to linearized molecules prior to transfection. When cotransfections were performed with such linearized plasmids and the 293 cells, pleiotropic reporter gene response patterns were completely disrupted. The results are shown in Fig. 5 and tabulated in Fig. 3. Under these conditions, no statistically significant stimulated, attenuated, or STMS orientation-dependent responses were observed. The possibilities that transfected plasmids in these experiments were recircularized inside nuclei or taken up differentially by the cells are both unlikely, since recircularization has been shown not to occur (62), nor does plasmid DNA uptake depend on plasmid topology (51, 62, 65).



View larger version (20K):
[in this window]
[in a new window]
 
Fig. 5. Luciferase reporter gene expression in 293 cells: in vivo cotransfection with linearized plasmids. The experiments were performed as described in MATERIALS AND METHODS and the legend to Fig. 3. Equal numbers of individual culture dishes (N) were used in each treatment group of each plasmid set as indicated along the baseline. Statistical significance was determined as described in the legend to Fig. 3. The calculated P values in each plasmid set are as follows: 5' flanking (control {leftrightarrow} MS, 0.25; MS {leftrightarrow} SM, 0.24; control {leftrightarrow} SM, 0.20), 3'-UTR (control {leftrightarrow} MS, 0.14; MS {leftrightarrow} SM, 0.057; control {leftrightarrow} SM, 0.47), and 3' flanking (control {leftrightarrow} MS, 0.79; MS {leftrightarrow} SM, 0.40; control {leftrightarrow} SM, 0.27).

 
Luciferase mRNA expression in cell-free translation systems.
Because of the potential for RNA polymerase II (Pol II) slippage across the STMS and mRNA transcript complexity (30), the lack of knowledge of the biologically active mRNA species encoded by this element (30), as well as the potential for low levels of reporter gene mRNAs following transient transfection (56), cell-free wheat germ systems were used to estimate the levels of translatable RNAs in extracts harvested from three sets of treated 293 cells.

Three different translation patterns were observed; the results are shown in Fig. 6 and tabulated in Fig. 3. Translatable mRNAs were present in total RNA samples in ratios (control:MS:SM plasmid) of 1:2.72:0.91 (5' flanking), 1:1:1 (3'-UTR), and 1:0.27:0.41 (3' flanking). The specificity of these results was validated by control experiments, which showed that luciferase-specific RNA-dependent light unit signals were abolished by protein and RNA synthesis inhibitors and absent in RNA samples derived from mock-transfected cells (Table 2). Although the causes of high nonspecific wheat germ extract background signals are unknown, these did not affect assay results, since similar expression trends were obtained with the TNT system, the wheat germ extracts of which produced low background signals of 80 LU. Equivalent ratios (1:1:1) were obtained when cytoplasmic RNA samples from the 3'-UTR set were translated using the low background system: 870 ±524 (control), 1,160 ±492 (MS), and 685 ±83 (SM) LU/2 µg RNA (±SD, N = 3; P > 0.05 [control {leftrightarrow} MS {leftrightarrow} SM; control {leftrightarrow} SM]).



View larger version (16K):
[in this window]
[in a new window]
 
Fig. 6. Luciferase reporter gene expression in 293 cells: in vitro cell-free translation. The experiments were performed as described in MATERIALS AND METHODS and the legend to Fig. 3. Equal numbers of individual RNA samples from two pooled cultures (N) were used in each treatment group of each plasmid set as indicated along the baseline. Statistical significance was determined as described in the legend to Fig. 3. The calculated P values in each plasmid set are as follows: 5' flanking (control {leftrightarrow} MS, 0.008; MS {leftrightarrow} SM, 0.003; control {leftrightarrow} SM, 0.75), 3'-UTR (control {leftrightarrow} MS, 0.93; MS {leftrightarrow} SM, 0.44; control {leftrightarrow} SM, 0.52), and 3' flanking (control {leftrightarrow} MS, 1 x 10-4; MS {leftrightarrow} SM, 0.25; control {leftrightarrow} SM, 0.007).

 

View this table:
[in this window]
[in a new window]
 
Table 2. Validation of translation of exogenous RNAs using a wheat germ extract system

 
Structural properties of STMS-containing supercoiled plasmids and a 400-bp HaeIII-BamHI fragment.
To identify structural differences between control and STMS-containing supercoiled plasmids which might be correlated with functional pleiotropy observed in the transfection and translation experiments summarized in Fig. 3, physical studies were performed with all three sets of plasmids and a purified STMS-containing 400-bp fragment.

As expected from the STMS sequence (Fig. 1) and non-B DNA structure proposed for this R-Y element (1, 30), supercoiled expression plasmids containing 359-bp inserts showed specific hypersensitivity to nuclease S1 at pH 4.5 (Fig. 7). Thus, depending on the transcription unit position of the STMS, two DNA fragments smaller than the full-length 6-kb plasmid were generated with sizes of either 4.6 and 1.4-kb (5' flanking, Fig. 7A), 3.8 and 2.2-kb (3'-UTR, Fig. 7B), or 4.5 and 1.5-kb (3' flanking, Fig. 7C).2 Virtually identical results were obtained with nuclease P1 at pH 7.6 (not shown). As deduced from the size patterns of these fragments, hypersensitive sites mapped into the GAA triplet repeat motif in each plasmid set. Under these conditions, specific hypersensitivities were not seen in control plasmids or when linearization with BglI preceded S1 or P1 digestion (not shown).



View larger version (26K):
[in this window]
[in a new window]
 
Fig. 7. Hypersensitivity of plasmids carrying HaeIII-TaqI fragments to nuclease S1. Control and MS or SM plasmids from 5' flanking (A), 3'-UTR (B), and 3' flanking (C) sets were treated and analyzed as described in MATERIALS AND METHODS. Each ethidium bromide (EtBr)-stained gel shows one lane used for sizing (1-kb+ GIBCO-BRL ladder) and four contiguous lanes carrying plasmid DNA samples, which were subjected (left to right) to mock digestion (uncut), or digestion with BglI, or digestion with nuclease S1, or digestion with nuclease S1 followed by BglI. Numbers along the right margins of the gels indicate the sizes of the control (5.7-kb) and STMS-containing plasmids (6.0-kb) linearized with BglI, and the sizes of the S1-hypersensitive fragments, as described in RESULTS. Digestion patterns were identical for both control plasmids; therefore, only the results obtained with pSPFOSLucpA1 are shown (A). Annotated diagrams of each plasmid set (not shown), which define pSP72 backbones and MCS cassettes containing chimeric inserts (1), and proposed sites of S1 digestion, can be obtained by E-mail to hleffert{at}ucsd.edu. The approximate sizes (nucleotides, nt) of the individual elements are: STMS, 360 nt; FOS, 850 nt; Luc, 1,600 nt; and pA, 850 nt.

 
It is well established that the binding of EtBr to supercoiled DNA causes the concomitant unwinding of duplex and superhelical structures (3). Therefore, to investigate the possibility that qualitative differences in plasmid supercoiling might be conferred by the STMS fragment, purified supercoiled plasmids from all three sets were incubated in solutions of increasing concentrations of EtBr and subjected to gel electrophoresis (Fig. 8). Under these conditions, differences in supercoiling among the treated plasmids should be reflected by reduced mobilities of the supercoiled molecules following electrophoresis. The results shown in Fig. 8 reveal that both 6-kb MS and SM plasmids migrated slower than the 5.7-kb control. Notably, however, in each set, the SM plasmid migrated slower than its MS counterpart between 25 and 100 µg EtBr/ml. Although the putative relaxed bands higher on the gels also showed lesser mobility decreases, under the electrophoresis conditions employed, these changes would not be due to changes in superhelicity. They would more likely be due to charge differences and increases in length and stiffness of such molecules (25).



View larger version (59K):
[in this window]
[in a new window]
 
Fig. 8. Effects of EtBr on the electrophoretic mobility of supercoiled DNA plasmids. Each of three different DNA samples (1 µg per reaction tube) from 5' flanking, 3'-UTR, and 3' flanking sets were incubated in TE (pH 8.0) with 0, 25, or 100 µg EtBr/ml for 2 min at 37°C. After incubation, the samples were electrophoresed through 0.75% agarose minigels for 5 h at 50 V (pH 8.0). Gels were stained with 500 ng EtBr/ml for 20 min at 21°C and visualized under transillumination. Each gel shows one lane used for sizing (1-kb+ GIBCO-BRL ladder; far left); the mobility of a 4.0-kb marker is indicated.

 
Thermal denaturation studies with the 400-bp linear fragment were motivated by the possibility that one or more segments of the duplex sequence might be capable of forming intramolecular triple helices, i.e., what has been called H-DNA. In fact, the melting profiles revealed a cooperative pre-duplex transition at 260 nm (Fig. 9A) and 280 nm (not shown) in the native material at pH 7. The fraction of the total hyperchromic change for that transition at pH 7 was ~7–8%. This transition was not readily reversed on initial slow cooling, so that remelting of the duplex did not show the triplex -> duplex transition (Fig. 9B). Generally, triplexes with homopyrimidine third strands are stabilized at lower pH because third strand C residues must be protonated to form two H-bonds to the G·C base pairs of the duplex (35). In the present case, while the triplex transition at pH 5 was still in evidence, the Tm was shifted from ~30°C at pH 7.0 to ~20°C (Fig. 9C). Thus, at pH 5, intramolecular triplex melting (no extraneous third strand was present) occurred at sufficiently low temperature that re-incorporation of the third strand segment into duplex was observed as an hypochromic change which preceded duplex melting. In addition, the hyperchromic change for the triplex -> duplex transition at pH 5 (Fig. 9C) was 2.5 times greater than at pH 7 (Fig. 9A). No pre-duplex transitions indicative of intramolecular triplex segments were observed for the 747-bp control DNA sample (38.69% [G+C]), which, devoid of an STMS, produced a single sigmoidal inflection curve near ~85.0°C, in agreement with what is shown in Fig. 9B.



View larger version (17K):
[in this window]
[in a new window]
 
Fig. 9. Thermal melting profiles of the STMS-containing 400-bp DNA duplex fragment. Note that the scales are different in each panel. A: native duplex between 0–60°C at pH 7 (solid line, solid circles, left y-axis scale). This shows the melting of the intramolecular triplex component within the duplex sequence. For comparison, the relevant part of the profile in B is shown as a dotted line with open circles on the same scale along the right y-axis in A. Note that the observed hyperchromic change for the triplex component is well above the spectrophotometer resolution. This transition was not readily reversible on cooling. B: melting profile of the recooled duplex after the profile in A was obtained. This profile shows the melting of the duplex above 80°C. C: native duplex between 0–60°C at pH 5. The melting of this intramolecular triplex component occurs at a lower temperature than at pH 7; this is followed by an hypochromic change reflecting duplex reformation.

 
R-Y elements in genes and genomic regulatory elements with homologies to the rat Pigr STMS.
To better document the qualitative and quantitative distribution of genomic R-Y elements with homologies to the Pigr 3'-UTR STMS, National Center for Biotechnology Information (NCBI) BLAST searches were performed using the rat HaeIII-TaqI fragment as the query sequence (1, 30). Hundreds of homologous sequences were identified from at least 15 species. Samples of statistically high scoring genes and the reported (or putative) functions of their R-Y elements have been cataloged (for the complete catalog, please refer to the Supplemental Material3 for this article, published online at the Physiological Genomics web site); the loca-tion of a lengthy Xq15 R-Y element has not been assigned. Eighteen genomic R-Y elements with known or proposed regulatory properties are listed in Table 3. In both pooled samples, the R-Y elements are distributed along transcription units as follows: 5' flanking, 27.1%; 5'-UTR, 3.4%; open reading frame (ORF), 22%; joining, 1.6%; intron, 10.2%; 3'-UTR, 17%; centromere and telomere, 3.4%; spacer, 6.8%; undefined, 8.5%.


View this table:
[in this window]
[in a new window]
 
Table 3. Partial list and proposed functions of genomic regulatory elements containing R-Y repeats

 

    DISCUSSION
 TOP
 ABSTRACT
 INTRODUCTION
 MATERIALS AND METHODS
 RESULTS
 DISCUSSION
 REFERENCES
 
Biological and physical studies of an STMS-containing 359-bp fragment from the rat Pigr gene suggest that its pleiotropic effects depend on (+)strand orientation of purine-rich triplet repeat tracts ([GGA]n=15[GAA]n=5AA[GAA]n=34), transcription unit position, and DNA supercoiling. Intracellular factors and cellular specificity might be contributory, since animal cell extracts contain R-Y element binding proteins (6, 42);4 and STMS-regulated 3'-UTR responses are nearly extinguished in transfected rat hepatocytes compared with human 293 cells. The physiological significance of these observations on Pigr expression is unclear; attenuator functions of the native 3'-UTR STMS are suggested from the inverse relationship between PIgR expression and rat liver development (23). Orientation-dependent effects have also been reported for transfected R-Y elements derived from 5' flanking tracts in the rat Ncam-c gene (6) and also from intron 1 of the human FRDA gene (48), but only attenuated responses were observed. In contrast, functional pleiotropy of a single genomic fragment has not been described previously.

Since the 359-bp fragment contains many putative transcription factor binding elements,5 its pleiotropic effects might depend on the sequences which flank the 5' (nt 1–29) and 3' ends (nt 284–359)6 of the fragment. This possibility has not been eliminated, but is unlikely since most of the proteins that recognize these elements do not require sequence polarity. Moreover, the reported behavior of such flanking elements, including elements situated in the STMS core (nt 30–283), often does not match the regulatory behavior of the Pigr fragment. For example, in contrast to orientation- and position-dependent stimulation observed with the 5' flanking MS plasmid, 5' flanking E-boxes in other genes behave like silencers (5, 46) or exert no effects (12); and in contrast to orientation- and position-dependent attenuation of gene expression observed with the 3'-UTR MS plasmid, 3'-UTR E-boxes in other genes behave like silencers (50) or activators (55). Furthermore, if flanking tracts were deleted, then the core fragment would still contain a PU box (GAGGAA, Ref. 26) and a JCV-repeat element (GGGNGGRR, Ref. 41) on the (+)strand, and an S1-HS element ([TCC]3; Ref. 24) on the (-)strand. Yet, in contrast to orientation- and position-dependent effects observed with 5' flanking, 3'-UTR, and 3' flanking plasmids, both PU and JCV elements behave like enhancers, since they exert orientation- and position-independent effects; the behavior of the [TCC]3 element is unknown. Further work is needed to resolve this issue definitively.

The Pigr STMS sequence is predictive of possible intramolecular triplex (32, 40, 47), tetraplex (21, 63), nodular (from H/H* hybrids) and tethered loop (15), or locally curved conformations (1, 9). Two lines of evidence suggest that an intramolecular triplex structure is formed. First, the 400-bp STMS-containing fragment displays anomalous UV melting profiles at pH 5 and pH 7 in the form of preduplex melting transitions that are consistent with formation of intramolecular triplex(es). Generally, such triplexes are intrinsically less stable than the duplexes in their progenitors, just as observed in this work. The fact that the triplex transition at pH 5 occurs at lower temperature than at pH 7 suggests that the triplexes formed at the two pH values have different third strands. This is consistent with the presence in the duplex of two types of triplet repeat, [GGA]n and [GAA]n in the homopurine strand segment of the duplex. Thus, at neutrality, the third strand is presumed to be [GAA]n and/or [GGA]n, whereas at pH 5 it must be [CCT]n and/or [CTT]n. The presence of at least one third G·C base pairs in the repeating sequence precludes the possibility of an intramolecular pyrimidine/triplex motif stable to 40°C at pH 7.0 or above (35). Since the hyperchromic change for the triplex -> duplex transition at pH 5 is much greater than at pH 7, the pH 5 triplex probably involves all the available homopyrimidine repeat sequences ([CCT]n and [CTT]n). In contrast, at neutrality the triplex formed with homopurine third strands must involve only the [GGA]n repeats. This is because a triplex with a high proportion of A residues, as in [GAA]n, is unfavored when the homopurine third strand runs anti-parallel to the homopurine strand of the target duplex, as must be the case for intramolecular triplex formation.

Second, STMS-containing supercoiled plasmids are hypersensitive to digestion with single-strand-specific nucleases. Scission occurs within or near GAA repeats, as expected for intramolecular triplexes formed at pH 4.5 or pH 7.6, at the lower pH suggestive of an homopyrimidine third strand and at the higher pH suggestive of an homopurine third strand. The presence of intramolecular triplexes in the linear 400-bp fragment is not inconsistent with the failure to detect nuclease-hypersensitive products after STMS-containing plasmids are linearized. Because of superhelical stress, the intramolecular helix is expected to be more thermostable than its linear counterpart (14, 15). Conversely, triplexes in linear DNA should be less stable than their supercoiled counterparts; in fact, the intramolecular triplexes in the linear 400-bp DNA sample were almost completely melted by 37°C. Thus triplexes in supercoiled DNA should be better enzyme substrates than triplexes in linear DNA. Consequently, under the conditions employed, S1 and P1 digestions performed at 37°C would be expected to show the effects of nicking in the regions of intramolecular triplexes in supercoiled plasmids but not in linear forms.

DNA supercoiling requirements of functional pleiotropy are unlikely to be caused by topological alterations in STMS-containing plasmids resulting from artifactually nicked single strands in triplexes. On the contrary, supercoiled plasmids are not simply nicked following transfection, but are instead linearized intracellularly, regardless of their structure, as are all nonlinear intermediates (62). Yet, despite topological homogenization, differential gene expression effects were observed between transfected MS and SM supercoiled plasmids in all three sets. Much higher levels of gene expression are also generated by initially supercoiled plasmids compared with nicked or linearized forms (62); in fact, gene expression ratios of ~10:1 ([supercoiled]:[linear]) were observed with all three plasmid sets throughout the nine individual treatment groups. Such quantitative differences occur because chromatin-associated supercoiled DNA templates, or regions of constrained plasmid supercoiling, increase the efficiency of utilization and the binding of regulatory factors prior to intracellular linearization (4, 62).

Electrophoretic mobility studies with each set of EtBr-treated plasmids suggest that STMS orientation affects DNA supercoiling. This behavior is consistent with forward and reverse STMS orientations producing different triplex structures (2, 44), but further investigation of comparative plasmid superhelical structure is needed to test this prediction. However, it is well-established that negative and positive DNA supercoiling are associated with efficient transcription by Pol II upstream and downstream of transcription complexes, respectively (8, 38). Since DNA supercoiling stabilizes intramolecular triplex formation and might be altered by STMS orientation, the triplexes formed by MS and SM plasmids might be structurally distinct. If, following transfection, such structural differences persist, then they might differentially perturb subsequent transcriptional and posttranscriptional processes.

According to this hypothesis, perturbations of negative supercoiling by the R-rich (+)strand STMS might augment transcription at 5' flanking positions, whereas different perturbations of negative supercoiling by the Y-rich (+)strand at similar 5' positions might restore transcription and gene expression to control levels. Coordinate increases in luciferase activity and translatable transcripts stimulated by the 5' flanking MS plasmid fit predictions of this hypothesis. Enhanced mRNA stability is unlikely, since 5' flanking sequences were not transcribed.

The hypothesis also predicts that MS plasmids should attenuate transcription when the STMS is positioned 3' to transcription complexes, by perturbing positive supercoiling upstream of such complexes (3'-UTR) or by perturbing positive supercoiling downstream of pA cassettes (3' flanking). In the former instance, the available evidence does not fit predictions, since all plasmid members of the 3'-UTR set encode detectable and equivalent levels of translatable transcripts.7 These findings implicate inefficient in vivo translation of mature R-rich (+)strand mRNAs rather than reduced nuclear export or selective mRNA instability, but transcriptional slippage (30) and differential processing are not eliminated. Prominent theoretical differences between the STMS (+) and (-) RNA strands, manifested by giant unpaired loops and hairpin stems with different secondary structures (1) and by noncanonical base pairing in the internal loops and bulges of the hairpin stems in both strands (16; and K. S. Koch, F. Major, and H. L. Leffert, unpublished observations), suggest that (+)strand loops or hairpin stems might sterically impede message access to intracellular translation machinery in vivo. However, this explanation assumes that cell-free translation systems operate less stringently than intracellular machinery (43), and it does not account for expression requirements of DNA supercoiling; thus additional processes must be involved.

In the latter instance, the available evidence fits predictions, since the control, attenuated, and restored levels of translatable luciferase transcripts encoded by the 3' flanking plasmid set were directly proportional to cellular luciferase activities. However, selective mRNA instability, formation of intermolecular triple helical complexes between third strand RNA transcripts and DNA targets in the STMS (48, 53), and differential RNA processing are not excluded.

Experiments in this and prior reports (1, 30) draw further attention to regulatory roles for constitutive R-Y elements that can form intramolecular DNA triplexes. More than 50 genes and regulatory elements across 12 different species are known to carry such elements in their (+)strands.8 Further investigations are needed to determine how these elements are distributed throughout genomes and to correlate systematically structural and functional pleiotropy.


    ACKNOWLEDGMENTS
 
We thank R. Reudigger and G. Walter for luciferase mRNA and discussion; F. Chisari and T. Hughes for plasmids pHcGAP and pRAY; T. Brown (MARC Summer Student Fellow) for help with P1 nuclease studies; and E. Bouhassira and B. Birshtein for thoughtful comments.

This work was supported by funds from the American Cancer Society (IN93R, to K. S. Koch); by National Institutes of Health Grants CA-71390 and 1P42-ES10337 (to H. L. Leffert) and GM-45751 (to C. L. Schildkraut); the Univ. of California Academic Senate (RW5M, to H. L. Leffert); Fundacion del Amo, Universidad Complutense de Madrid, Spain (to I. Fabregat); US Dept. of Energy (DE-FG02-96ER62202, to J. R. Fresco); and by a Grant-in-Aid for High Technology Research Program from the Ministry of Education, Sciences, Sports and Culture of Japan (to T. Aoki).


    FOOTNOTES
 
Article published online before print. See web site for date of publication (http://physiolgenomics.physiology.org).

Address for reprint requests and other correspondence: H. L. Leffert, Dept. of Pharmacology and Center for Molecular Genetics, School of Medicine, Univ. of California, San Diego, La Jolla, CA 92093 (E-mail: hleffert{at}ucsd.edu).

1 Similar patterns of attenuated reporter gene expression are observed with plasmids carrying GAA repeats from the mouse Pigr 3'-UTR, examined at 3'-UTR and 3' flanking transcription unit positions using neomycin phosphotransferase reporters and mouse NIH3T3 cells (T. Sato, personal communication, takashi-sato{at}yakult.co.jp). Back

2 Although highly purified DNA samples and standard experimental conditions were employed at each digestion step, complete digestion of the 6.0-kb STMS-containing plasmids by either S1 or P1 nuclease was not observed. To clarify the causes of incomplete digestion, further work is needed. Back

3 Supplemental material to this article is available online at http://physiolgenomics.physiology.org/cgi/content/full/5/2/53/DC1, or this catalog can be obtained directly by E-mailing H. L. Leffert (hleffert{at}ucsd.edu) or K. S. Koch (kkoch{at}ucsd.edu). Back

4 Preliminary results of gel shift experiments using end-labeled deoxynucleotide ligands, including [GAA]n=11, [GGA]n=11, and the double-stranded 400-bp HaeIII-BamHI fragment, reveal the presence of specific binding proteins in cell extracts (nuclear > cytoplasmic) isolated from human 293 cells and fetal rat hepatocytes (K. S. Koch, I. Fabregat, and H. L. Leffert, unpublished observations). Back

5 Putative transcription factor protein binding sites were determined with SIGSCAN4 (Ref. 52). Back

6 {gamma}-IRE-CS, TCF-1, GR, CF1, GMCSF-CS, LBP-1, hsp70-US, {alpha}-INF.2, PEA3, LF-A1, H4TF-2, H-2RIIBP/T3R-a1, NF-GMb, GATA-1, and bHLH E-boxes (at nt 284–288, and nt 327–332 in the 359-bp fragment) as determined with SIGSCAN4. Back

7 Luciferase mRNAs encoded by control and 3'-UTR MS and SM plasmids are undetectable on Northern blots of total or cytoplasmic RNA (I. Fabregat, A. E. Atkinson, H. L. Leffert, and K. S. Koch, unpublished observations). Back

8 Current NCBI BLAST searches suggest that more than 4,000 sites in the human genome are homologous to the Pigr STMS [P(N) < 10-10]. Back


    REFERENCES
 TOP
 ABSTRACT
 INTRODUCTION
 MATERIALS AND METHODS
 RESULTS
 DISCUSSION
 REFERENCES
 

  1. Aoki T, Koch KS, and Leffert HL. Attenuation of gene expression by a trinucleotide repeat-rich tract from the terminal exon of the rat hepatic polymeric immunoglobulin gene. J Mol Biol 267: 229–236, 1997.[ISI][Medline]
  2. Bates AD and Maxwell A. DNA Topology. Oxford, UK: IRL, 1993.
  3. Bauer W and Vinograd J. The interaction of closed circular DNA with intercalative dyes. I. The superhelix density of SV40 DNA in the presence and absence of dye. J Mol Biol 33: 141–171, 1968.[ISI][Medline]
  4. Bonilla PJ, Freytag SO, and Lutter C. Enhancer-activated plasmid transcription complexes contain constrained supercoiling. Nucleic Acids Res 19: 3965–3971, 1991.[Abstract]
  5. Brabletz T, Jung A, Hlubek F, Lohberg C, Meiler J, Suchy U, and Kirchner T. Negative regulation of CD4 expression in T cells by the transcriptional repressor ZEB. Int Immunol 11: 1701–1708, 1999.[Abstract/Free Full Text]
  6. Chen AS, Reyes A, and Akeson R. A homopurine:homopyrimidine sequence derived from the rat neuronal cell adhesion molecule-encoding gene alters expression in transient transfections. Gene 128: 211–218, 1993.[ISI][Medline]
  7. Chou SH, Zhu L, and Reid BR. Sheared purine purine pairing in biology. J Mol Biol 267: 1055–1067, 1997.[ISI][Medline]
  8. Cozzarelli NR, Boles TC, and White JH. DNA Topology and Its Biological Effects. Cold Spring Harbor, NY: Cold Spring Harbor Laboratory, 1990, p. 139–184.
  9. Crothers DM. Upsetting the balance of forces in DNA. Science 266: 1819–1820, 1994.[ISI][Medline]
  10. Davidson EH and Britten RJ. Regulation of gene expression: possible role of repetitive sequences. Science 204: 1052–1059, 1979.[ISI][Medline]
  11. Davis TL, Firulli AB, and Kinniburgh AJ. Ribonucleoprotein and protein factors bind to an H-DNA-forming c-myc DNA element: possible regulators of the c-myc gene. Proc Natl Acad Sci USA 86: 9682–9686, 1989.[Abstract]
  12. Feo S, Antona V, Barbieri G, Passantino R, Cali L, and Giallongo A. Transcription of the human beta enolase gene (ENO-3) is regulated by an intronic muscle-specific enhancer that binds myocyte-specific enhancer factor 2 proteins and ubiquitous G-rich-box binding factors. Mol Cell Biol 15: 5991–6002, 1995.[Abstract]
  13. Fernandes M, Xiao H, and Lis JT. Fine structure analyses of the Drosophila and Saccharomyces heat shock factor: heat shock element interactions. Nucleic Acids Res 22: 167–173, 1994.[Abstract]
  14. Frank-Kamenetskii MD. DNA Topology and its Biological Effects. Cold Spring Harbor, NY: Cold Spring Harbor Laboratory, 1990, p. 185–216.
  15. Frank-Kamenetskii MD and Mirkin SM. Triplex DNA structures. Annu Rev Biochem 64: 65–95, 1995.[ISI][Medline]
  16. Gautheret D, Major F, and Cedergren R. Modeling the three-dimensional structure of RNA using discrete nucleotide conformational sets. J Mol Biol 229: 1049–1064, 1993.[ISI][Medline]
  17. Gee JE, Revankar GR, Rao TS, and Hogan ME. Triplex formation at the rat neu gene utilizing imidazole and 2'-deoxy-6-thioguanosine base substitutions. Biochemistry 34: 2042–2048, 1995.[ISI][Medline]
  18. Gilmour DS, Thomas GH, and Elgin SC. Drosophila nuclear proteins bind to regions of alternating C and T residues in gene promoters. Science 245: 1487–1490, 1989.[ISI][Medline]
  19. Grabczyk E and Fishman MC. A long purine-pyrimidine homopolymer acts as a transcriptional diode. J Biol Chem 270: 1791–1797, 1995.[Abstract/Free Full Text]
  20. Hirt B. Selective extraction of polyoma DNA from infected mouse cell cultures. J Mol Biol 26: 365–369, 1967.[ISI][Medline]
  21. Howell R and Usdin K. The ability to form intrastrand tetraplexes is an evolutionarily conserved feature of the 3' end of L1 retrotransposons. Mol Biol Evol 14: 144–155, 1997.[Abstract]
  22. Huertas D and Azorin F. Structural polymorphism of homopurine DNA sequences. d(GGA)n and d(GGGA)n repeats form intramolecular hairpins stabilized by different base-pairing interactions. Biochemistry 35: 13125–13135, 1996.[ISI][Medline]
  23. Huling S, Fournier GR, Feren A, Chuntharapai A, and Jones AL. Ontogeny of the secretory immune system: maturation of a functional pIg-R regulated by gene expression. Proc Natl Acad Sci USA 89: 4260–4264, 1992.[Abstract]
  24. Johnson AC, Jinno Y, and Merlino GT. Modulation of epidermal growth factor receptor proto-oncogene transcription by a promoter site sensitive to S1 nuclease. Mol Cell Biol 8: 4174–4184, 1988.[ISI][Medline]
  25. Johnson PH and Grossman LI. Electrophoresis of DNA in agarose gels. Optimizing separations of conformational isomers of double- and single-stranded DNA. Biochemistry 16: 4217–4225, 1977.[ISI][Medline]
  26. Klemsz MJ, McKercher SR, Celada A, Van Beveren C, and Maki RA. The macrophage and B cell-specific transcription factor PU.1 is related to the ets oncogene. Cell 61: 113–124, 1990.[ISI][Medline]
  27. Koch K and Leffert HL. Growth control of differentiated fetal rat hepatocytes in primary monolayer culture. VI. Studies with conditioned medium and its functional interactions with serum factors. J Cell Biol 62: 780–791, 1974.[Abstract/Free Full Text]
  28. Koch KS, Fletcher RG, Grond M, Inyang AI, Lu XP, Brenner DA, and Leffert HL. Inactivation of plasmid reporter gene expression by one benzo(a)pyrene diol-epoxide DNA adduct in adult rat hepatocytes. Cancer Res 53: 2279–2286, 1993.[Abstract]
  29. Koch KS, Fletcher RG, Grond M, and Leffert HL. A 43 bp cDNA sequence homology and triplet repeat motif among putative polymeric immunoglobulin receptor messenger RNAs in regenerating rat liver. Hepatology 18: 226–228, 1993.[ISI][Medline]
  30. Koch KS, Gleiberman AG, Aoki T, Leffert HL, Feren A, Jones AL, and Fodor EJ. Discordant expression and variable numbers of neighboring GGA- and GAA-rich triplet repeats in the 3' untranslated regions of two groups of messenger RNAs encoded by the rat polymeric immunoglobulin receptor gene. Nucleic Acids Res 23: 1098–1112, 1995.[Abstract]
  31. Koch KS, Aoki T, Wang Y, Atkinson AE, Gleiberman AD, Glebov OK, and Leffert HL. Site-specific integration of targeted DNA into animal cell genomes. Gene 249: 135–144, 2000.[ISI][Medline]
  32. Krasilnikov AS, Panyutin IG, Samadashwily GM, Cox R, Lazurkin YS, and Mirkin SM. Mechanisms of triplex-caused polymerization arrest. Nucleic Acids Res 25: 1339–1346, 1997.[Abstract/Free Full Text]
  33. Krowczynska AM, Rudders RA, and Krontiris TG. The human minisatellite consensus at breakpoints of oncogene translocations. Nucleic Acids Res 18: 1121–1127, 1990.[Abstract]
  34. Kushiro A and Sato T. Polymeric immunoglobulin gene of mouse: sequence, structure and chromosomal location. Gene 204: 277–282, 1997.[ISI][Medline]
  35. Lavelle L and Fresco JR. UV spectroscopic identification and thermodynamic analysis of protonated third strand deoxycytidine residues at neutrality in the triplex d(C(+)-T)6:[d(A-G)6.d(C-T)6]; evidence for a proton switch. Nucleic Acids Res 23: 2692–2705, 1995.[Abstract]
  36. Leffert HL and Paul D. Studies on primary cultures of differentiated fetal liver cells. J Cell Biol 52: 559–568, 1972.[Abstract/Free Full Text]
  37. Leffert HL. Growth control of differentiated fetal rat hepatocytes in primary monolayer culture. V. Occurrence in dialyzed fetal bovine serum of macro-molecules having both positive and negative growth regulatory functions. J Cell Biol 62: 767–779, 1974.[Abstract/Free Full Text]
  38. Liu LF and Wang JC. Supercoiling of the DNA template during transcription. Proc Natl Acad Sci USA 84: 7024–7027, 1987.[Abstract]
  39. Lu Q, Wallrath LL, Granok H, and Elgin SC. (CT)n(GA)n repeats and heat shock elements have distinct roles in chromatin structure and transcriptional activation of the Drosophila hsp26 gene. Mol Cell Biol 13: 2802–2814, 1993.[Abstract]
  40. Mariappan SVS, Catasti P, Silks LA III, Bradbury EM, and Gupta G. The high-resolution structure of the triplex formed by the GAA/TTC triplet repeat associated with Friedreich’s ataxia. J Mol Biol 285: 2035–2052, 1999.[ISI][Medline]
  41. Martin JD, King DM, Slauch JM, and Frisque RJ. Differences in regulatory sequences of naturally occurring JC virus variants. J Virol 53: 306–311, 1985.[ISI][Medline]
  42. Maueler W, Kyas A, Keyl HG, and Epplen JT. A genome-derived (gaa.ttc)24 trinucleotide block binds nuclear protein(s) specifically and forms triple helices. Gene 215: 389–403, 1998.[ISI][Medline]
  43. Mendez E, Rocher A, Calero M, Girbes T, Citores L, and Soriano F. Primary structure of omega-hordothionin, a member of a novel family of thionins from barley endosperm, and its inhibition of protein synthesis in eukaryotic and prokaryotic cell-free systems. Eur J Biochem 239: 67–73, 1996.[Abstract]
  44. Mirkin SM and Frank-Kamenetskii MD. H-DNA and related structures. Annu Rev Biophys Biomol Struct 23: 541–576, 1994.[ISI][Medline]
  45. Mizuguchi G, Kanei-Ishii C, Sawazaki T, Horikoshi M, Roeder RG, Yamamoto T, and Ishii S. Independent control of transcription initiations from two sites by an initiator-like element and TATA box in the human c-erbB-2 promoter. FEBS Lett 348: 80–88, 1994.[ISI][Medline]
  46. Molkentin JD, Jobe SM, and Markham BE. Alpha-myosin heavy chain gene regulation: delineation and characterization of the cardiac muscle-specific enhancer and muscle-specific promoter. J Mol Cell Cardiol 28: 1211–1225, 1996.[ISI][Medline]
  47. Ohshima K, Kang S, Larson JE, and Wells RD. Cloning, characterization, and properties of seven triplet repeat sequences. J Biol Chem 271: 16773–16783, 1996.[Abstract/Free Full Text]
  48. Ohshima K, Montermini L, Wells RD, and Pandolfo M. Inhibitory effects of expanded GAA TTC triplet repeats from intron I of the Friedreich ataxia gene on transcription and replication in vivo. J Biol Chem 273: 14588–14595, 1998.[Abstract/Free Full Text]
  49. Palla F, Melfi R, Anello L, Di Bernardo M, and Spinelli G. Enhancer blocking activity located near the 3' end of the sea urchin early H2A histone gene. Proc Natl Acad Sci USA 94: 2272–2277, 1997.[Abstract/Free Full Text]
  50. Paul C, Simar-Blanchet AE, Ro HS, and Le Cam A. Characterization of three transcriptional repressor sites within the 3' untranslated region of the rat serine protease inhibitor 23 gene. Eur J Biochem 254: 538–546, 1998.[Abstract]
  51. Pina B, Hache RJ, Arnemann J, Chalepakis G, Slater EP, and Beato M. Hormonal induction of transfected genes depends on DNA topology. Mol Cell Biol 10: 625–633, 1990.[ISI][Medline]
  52. Prestridge DS. SIGNAL SCAN: a computer program that scans DNA sequences for eucaryotic transcription elements. Comput Appl Biosci 7: 203–206, 1991.[Abstract]
  53. Reaban ME and Griffin JA. Induction of RNA-stabilized DNA conformers by transcription of an immunoglobulin switch region. Nature 348: 342–344, 1990.[ISI][Medline]
  54. Rippe RA, Brenner DA, and Leffert HL. DNA-mediated gene transfer into adult rat hepatocytes in primary culture. Mol Cell Biol 10: 689–695, 1990.[ISI][Medline]
  55. Rippe RA, Umezawa A, Kimball JP, Breindl M, and Brenner DA. Binding of upstream stimulatory factor to an E-box in the 3'-flanking region stimulates alpha1(I) collagen gene transcription. J Biol Chem 272: 1753–1760, 1997.[Abstract/Free Full Text]
  56. Ruediger R and Walter G. Quantitation of luciferase reporter mRNA by in vitro translation of total cellular RNA is faster and more sensitive than northern blotting and primer extension. Anal Biochem 253: 267–269, 1997.[ISI][Medline]
  57. San Miguel P, Tikhonov A, Jin YK, Motchoulskaia N, Zakharov D, Melake-Berhan A, Springer PS, Edwards KJ, Lee M, Avramova Z, and Bennetzen JL. Nested retrotransposons in the intergenic regions of the maize genome. Science 274: 765–768, 1996.[Abstract/Free Full Text]
  58. Suda T, Mishima Y, Takayanagi K, Asakura H, Odani S, and Kominami R. A novel activity of HMG domains: promotion of the triple-stranded complex formation between DNA containing (GGA/TCC)11 and d(GGA)11 oligonucleotides. Nucleic Acids Res 24: 4733–4740, 1996.[Abstract/Free Full Text]
  59. Sutherland GR and Richards RI. Simple tandem DNA repeats and human genetic disease. Proc Natl Acad Sci USA 92: 3636–3641, 1995.[Abstract/Free Full Text]
  60. Tacke R, Tohyama M, Ogawa S, and Manley JL. Human Tra2 proteins are sequence-specific activators of pre-mRNA splicing. Cell 93: 139–148, 1998.[ISI][Medline]
  61. Tanaka K, Watakabe A, and Shimura Y. Polypurine sequences within a downstream exon function as a splicing enhancer. Mol Cell Biol 14: 1347–1354, 1994.[Abstract]
  62. Weintraub H, Cheng PF, and Conrad K. Expression of transfected DNA depends on DNA topology. Cell 46: 115–122, 1986.[ISI][Medline]
  63. Weitzmann MN, Woodford KJ, and Usdin K. The development and use of a DNA polymerase arrest assay for the evaluation of parameters affecting intrastrand tetraplex formation. J Biol Chem 271: 20598–20964, 1996.
  64. Wells RD and Warren ST. Genetic Instabilities and Hereditary Neurological Diseases. San Diego, CA: Academic, 1998.
  65. Xie TD, Sun L, Zhao HG, Fuchs JA, and Tsong TY. Study of mechanisms of electric field-induced DNA transfection. IV. Effects of DNA topology on cell uptake and transfection efficiency. Biophys J 63: 1026–1031, 1992.[Abstract]
  66. Xu G and Goodridge AG. Function of a C-rich sequence in the polypyrimidine/polypurine tract of the promoter of the chicken malic enzyme gene depends on promoter context. Arch Biochem Biophys 363: 202–212, 1999.[ISI][Medline]