Advertisement
JBC

HOME HELP FEEDBACK SUBSCRIPTIONS ARCHIVE SEARCH TABLE OF CONTENTS
QUICK SEARCH:   [advanced]


     

Originally published In Press as doi:10.1074/jbc.M502952200 on April 18, 2005

J. Biol. Chem., Vol. 280, Issue 24, 22749-22760, June 17, 2005
This Article
Right arrow Abstract Freely available
Right arrow Full Text (PDF)
Right arrow Supplemental Data
Right arrow All Versions of this Article:
280/24/22749    most recent
M502952200v1
Right arrow Submit a Letter to Editor
Right arrow Alert me when this article is cited
Right arrow Alert me when eLetters are posted
Right arrow Alert me if a correction is posted
Right arrow Citation Map
Services
Right arrow Email this article to a friend
Right arrow Similar articles in this journal
Right arrow Similar articles in PubMed
Right arrow Alert me to new issues of the journal
Right arrow Download to citation manager
Right arrowRequest Permissions
Citing Articles
Right arrow Citing Articles via HighWire
Right arrow Citing Articles via Google Scholar
Google Scholar
Right arrow Articles by Raghavan, S. C.
Right arrow Articles by Lieber, M. R.
Right arrow Search for Related Content
PubMed
Right arrow PubMed Citation
Right arrow Articles by Raghavan, S. C.
Right arrow Articles by Lieber, M. R.
Social Bookmarking
Add to CiteULike   Add to Complore   Add to Connotea   Add to Del.icio.us   Add to Digg   Add to Reddit   Add to Technorati  
What's this?

Evidence for a Triplex DNA Conformation at the bcl-2 Major Breakpoint Region of the t(14;18) Translocation*{boxs}

Sathees C. Raghavan,abcde Paul Chastain,f Jeremy S. Lee,g Balachandra G. Hegde,ch Sabrina Houston,abcde Ralf Langen,ch Chih-Lin Hsieh,ci Ian S. Haworth,cj and Michael R. Lieberabcdek

From the aNorris Comprehensive Cancer Center, Departments of bPathology, cBiochemistry and Molecular Biology, dBiological Sciences, eMolecular Microbiology and Immunology, and iUrology, hZilka Neurogenetics Institute, and jDepartment of Pharmaceutical Sciences, University of Southern California Keck School of Medicine, Los Angeles, California 90033, the fDepartments of Biochemistry and of Pathology, University of North Carolina, Chapel Hill, North Carolina 27599, and the gDepartment of Biochemistry, University of Saskatchewan, Saskatoon, Saskatchewan S7N5E5, Canada

Received for publication, March 17, 2005 , and in revised form, April 14, 2005.


   ABSTRACT
TOP
 ABSTRACT
INTRODUCTION
 EXPERIMENTAL PROCEDURES
 RESULTS
 DISCUSSION
 REFERENCES
 
The most common chromosomal translocation in cancer, t(14;18), occurs at the bcl-2 major breakpoint region (Mbr) in follicular lymphomas. The 150-bp bcl-2 Mbr, which contains three breakage hotspots (peaks), has a single-stranded character and, hence, a non-B DNA conformation both in vivo and in vitro. Here, we use gel assays and electron microscopy to show that a triplex-specific antibody binds to the bcl-2 Mbr in vitro. Bisulfite reactivity shows that the non-B DNA structure is favored by, but not dependent upon, supercoiling and suggests a possible triplex conformation at one portion of the Mbr (peak I). We have used circular dichroism to test whether the predicted third strand of that suggested structure can indeed form a triplex with the duplex at peak I, and it does so with 1:1 stoichiometry. Using an intracellular minichromosomal assay, we show that the non-B DNA structure formation is critical for the breakage at the bcl-2 Mbr, because a 3-bp mutation that disrupts the putative peak I triplex also markedly reduces the recombination of the Mbr. A three-dimensional model of such a triplex is consistent with bond length, bond angle, and energetic restrictions (stacking and hydrogen bonding). We infer that an imperfect purine/purine/pyrimidine (R.R.Y) triplex likely forms at the bcl-2 Mbr in vitro, and in vivo recombination data favor this as the major DNA conformation in vivo as well.


   INTRODUCTION
TOP
 ABSTRACT
 INTRODUCTION
EXPERIMENTAL PROCEDURES
 RESULTS
 DISCUSSION
 REFERENCES
 
The t(14;18) translocation is the most common chromosomal translocation in human cancer and occurs in nearly all follicular lymphomas (1–4). It is the most important among lymphoid translocations, because follicular lymphomas account for nearly half of all non-Hodgkin lymphomas (5). The t(14;18) translocation is assumed to result from recombination between the bcl-2 gene and V(D)J subexons during V(D)J recombination in pre-B cells (Supplemental Fig. 1). This translocation juxtaposes the bcl-2 locus on chromosome 18 to the intron enhancer of the immunoglobulin heavy chain locus on chromosome 14, leading to the overexpression of the antiapoptotic protein Bcl-2 and thereby to follicular lymphoma (1, 2, 6–9).

Although the bcl-2 gene is more than 200 kb long, in most follicular lymphoma patients, the break at chromosome 18 occurs within a small 150-bp region located in the untranslated portion of the terminal (third) exon, designated as the major breakpoint region (Mbr)1 (see Supplemental Fig. 1 for orientation) (2, 3). Within the 150-bp Mbr, there are three peaks (hotspots) of breakage, each about 15–20 bp in size (Supplemental Fig. 1) (3). The characteristic precision and distribution profile within the bcl-2 Mbr strongly suggest some strong local regional specificity to the bcl-2 Mbr (5).

Analysis of patient breakpoint junctions suggested that a V(D)J recombination-mediated mechanism generates a break on chromosome 14, in which DH and JH subexons are cleaved at a pair of signal sequences (12- or 23-signal) by the RAG complex (RAG1, RAG2, and HMG1). However, the mechanism of breakage at the bcl-2 Mbr was completely unknown, although it had been speculated that the Mbr sequence may act like a cryptic signal sequence. Recently, we and others showed that whereas the RAG complex misrecognizes some breakpoint sites as cryptic heptamer/nonamer signals (10–12), the bcl-2 Mbr does not function as a cryptic signal (12, 13).

We recently reported that the bcl-2 Mbr adopts a non-B form structure within the mammalian chromosome (13). This altered DNA structure can be recapitulated when the bcl-2 Mbr is present on a human minichromosome or on an E. coli episome (13). The structure at the 150-bp bcl-2 Mbr contains distinctive regions of single-strandedness that correspond well to the regions of translocation frequency among patients. Based on an intracellular transfection system, we could reconstitute the t(14;18) translocation process inside mammalian cells and show that it is dependent on the RAG complex (13). In addition, we showed that the structure formation could be recapitulated on DNA fragments of various lengths containing the bcl-2 Mbr region (14). The non-B structure formation is favored in the presence of Mg2+ at neutral pH and is stable for days. The structure formation also requires Hoogsteen hydrogen bonding (14). These features rule out many non-B DNA conformations but do not permit us to determine the specific non-B conformation that does exist at the Mbr.

Triple helical DNA (triplex DNA) structures can form inter- or intramolecularly and were first reported nearly 50 years ago (15–18). Triplexes are formed at polypurine-polypyrimidine (R.Y) tracts (19–22). Triplexes formed within the same duplex DNA region are known as intramolecular triplexes, or H-DNA (23). The third strand (R or Y) occupies the major groove of the duplex DNA (R.Y). There are two general types of triplexes. Depending on whether the third strand of the triplex is rich in purines (R) or pyrimidines (Y), they are called R.R.Y triplexes or Y.R.Y triplexes (20, 21). Each of these triplexes is known to form under specific reaction conditions. Y.R.Y triplexes are favored in acidic pH, whereas R.R.Y triplexes prefer neutral pH and require the presence of divalent metal cations, making the latter more likely under physiologic conditions (20, 21). R.R.Y triplexes tolerate more versatile pairing schemes and form more rapidly (24). H-DNA is formed by a fold-back mechanism, stabilized by supercoiling, and half of the length of the R or Y strand (fourth strand) is unpaired (20, 21).

Here we show that the bcl-2 Mbr is capable of adopting two different non-B conformations at peak I (termed I-{alpha} and I-{beta}), in addition to the one non-B conformation at peak III, within the mammalian cells. The Mbr structure formation is facilitated by supercoiling, although it is not dependent upon it. Full molecule bisulfite sequencing experiments reveal strand asymmetry at the single-stranded regions of the bcl-2 Mbr in mammalian cells or when the DNA is harvested from Escherichia coli. Gel shift assays and electron microscopy show that a triplex-specific antibody binds to the bcl-2 Mbr. Circular dichroism studies show that the structure is an R.R.Y triplex. Based on these studies, we provide both two- and three-dimensional models for an R.R.Y structure at peak I of the bcl-2 Mbr. In sum, there are 14 lines of experimentation using five different methods indicating an R.R.Y triplex conformation at the bcl-2 Mbr. In addition, we find that such triplex formation is the basis for the bcl-2 translocation.


   EXPERIMENTAL PROCEDURES
TOP
 ABSTRACT
 INTRODUCTION
 EXPERIMENTAL PROCEDURES
RESULTS
 DISCUSSION
 REFERENCES
 
Enzymes, Chemicals, and Reagents—Sodium bisulfite and other chemical reagents are from Sigma. Restriction enzymes and DNA-modifying enzymes are from New England Biolabs (Beverly, MA). Components for cell culture are from Irvine Scientific (Santa Ana, CA). Radioisotope-labeled nucleotides were purchased from PerkinElmer Life Sciences.

Plasmid Construction—The plasmid constructs were made by modifying the SV40-based plasmid, pGG51 (25). A SalI-digested, 300-bp PCR fragment of the bcl-2 Mbr was cloned into the SalI site of pGG51 after removing the 12-signal normally at this site. The resulting plasmid, pXW5, contains the bcl-2 fragment in the physiologic transcriptional orientation relative to the lac transcriptional promoter. In the case of pSCR33, after deleting the 12-signal from pGG51 by SalI digestion, the linear DNA was self-ligated. The pSCR41 was constructed by cloning a mutated 300-bp PCR product of bcl-2 Mbr (see "Mutagenesis at the Bcl-2 Mbr") into the SalI site of pSCR33. The plasmid, pSCR1, was constructed by cloning the 300-bp SalI fragment into the SalI site of pBSS3, a derivative of Bluescript plasmid in which the lacZ promoter has been deleted (26).

Bisulfite Modification Assay—The bisulfite modification assay was carried out as described earlier (13, 14). See Supplemental Material for details.

Mutagenesis at the bcl-2 Mbr—Mutagenesis was carried out using standard PCR protocol with mutant sequence primers (27). See Supplemental Material for details.

Monoclonal Antibody Detection of Triplex DNA—0, 25, 50, 75, and 100 ng each of two triplex antibodies (Jel466 or Jel318) were individually incubated with 1 µg of the supercoiled plasmid DNA (pXW5) in 1x buffer containing 20 mM Tris acetate (pH 7.5), 10 mM magnesium acetate, 50 mM potassium acetate, and 1 mM dithiothreitol in a 20-µl volume. After incubation at 37 °C for 1 h, a 700-bp bcl-2 Mbr fragment was released by AatII/BamHI digestion for 2 h at 37 °C. Digestion products were resolved by 1% agarose gel electrophoresis and transferred to a nylon membrane by capillary blotting. Southern hybridization was carried out by using the same labeled AatII/BamHI fragment as the probe (labeling with [{alpha}-32P]dCTP). Hybridization was carried out at 65 °C overnight. Radioactive blots were washed twice in 2x SSC, 0.5% SDS and twice in 0.2x SSC, 0.5% SDS and exposed to a phosphor imager screen. Radioactive signals were detected by scanning.

For control experiments, pSCR33 (the 300-bp bcl-2 Mbr is absent in this plasmid) was incubated with 0, 50, and 100 ng of Jel466 in a 20-µl volume. Here, the same AatII/BamHI digestion was used to release the fragment, which is 300 bp smaller. In another control, after incubation with the antibody, a 1.6-kb fragment (AvrII/AatII) was released from the backbone of the plasmid and resolved on an agarose gel. For Southern hybridization, the same 1.6-kb fragment was used as the probe after labeling. Other steps were the same as above.

Electron Microscopy—DNA fragments (488 and 425 bp) for electron microscopy were prepared by PCR amplification. The DNA (1 µg) was then gel-purified and incubated in triplex-forming buffer B (50 mM Tris-HCl (pH 7.2), 10 mM MgCl2, 100 mM NaCl, and 100 µM spermine) at 37 °C overnight. The products were then purified and resuspended in buffer containing 10 mM Tris-HCl (pH 7.2), 10 mM MgCl2. About 300 ng of DNA was then resuspended in 10 µl of buffer containing 20 mM Tris acetate (pH 7.5), 10 mM magnesium acetate, 50 mM potassium acetate, and 1 mM dithiothreitol and incubated with 100 ng of triplex-specific antibody (Jel466 or Jel318) or without antibody at 37 °C for 1 h. Following the 1-h incubation, the samples were fixed for 5 min at 22 °C by adding an equal amount of 1.2% glutaraldehyde and then diluted with 280 µl of TE(10 mM Tris, pH 7.5, 1 mM EDTA). The fixed samples were absorbed to thin carbon foils, washed, air-dried, and rotary shadow-cast with tungsten at high vacuum (28). Micrographs were taken on a Phillips CM12 electron microscope at 40 kV. Images for publication were captured using a Gatan CCD camera and then using DigitalMicrograph 3.3 (Gatan Inc.).

Circular Dichroism—The CD spectra were recorded on a Jasco J-810 spectropolarimeter fitted with a JASCO PFD-425S Peltier temperature control system at 22 or 80 °C in a wavelength range of 210–310 nm. The DNA samples containing duplex DNA (1.6 µM) and third strand oligomer (1.6 µM) were incubated together in a buffer containing 20 mM PIPES (pH 7.0), 140 mM KCl, 5 mM MgCl2, and 1 mM spermidine for 1 h at 22 °C. DNA samples were placed on a 0.1-cm path length square quartz cuvette and used for the scanning. 30 scans were performed per spectrum. Spectra of buffer solution were subtracted to correct for base-line artifacts. CD spectra were normalized in units of molar ellipticity (degrees cm2 dmol–1).

Human Recombination Assay—The human lymphoid cell line, Reh, was grown logarithmically, transfected with the appropriate plasmid substrates using the electroporation/DEAE-dextran method as described previously (29), and cultured for 48 h at 37 °C. The various steps involved in the human V(D)J recombinase assay described previously (25, 30, 31). See Supplemental Material for details.


   RESULTS
TOP
 ABSTRACT
 INTRODUCTION
 EXPERIMENTAL PROCEDURES
 RESULTS
DISCUSSION
 REFERENCES
 
More than One Non-B Conformation Exists at the bcl-2 Mbr within Mammalian Cells—A significant amount of single-strandedness exists at the bcl-2 Mbr based on probing using a bisulfite modification assay on genomic DNA, sensitivity to KMnO4 and OsO4 in vivo, a P1 nuclease assay in vitro, and a gel shift assay in vitro (13, 14). KMnO4 and OsO4 in vivo analysis are useful methods but suffer from the fact that they do not represent minority molecular conformations well. For the bcl-2 Mbr, there is the additional problem that there is a paucity of Ts where these chemicals react, and this same limitation applies to psoralen (20). Here we have used the bisulfite modification assay to further understand the nature of the non-B DNA structure within the mammalian cells. The chromosomal DNA was extracted from mammalian cells (Reh, a pre-B cell line) using a nondenaturing method to minimize disturbances to any non-B DNA conformation. The DNA was then treated with bisulfite. (Similar results were obtained regardless of whether the bisulfite is used at pH 5.2 or 6.5 (13).) The molecules with non-B DNA structure were enriched with restriction enzyme digestion and detected by PCR, cloning, and sequencing (see "Experimental Procedures"). These molecules showed an increase in bisulfite conversion of cytosine to thymine (suggesting single-strandedness) at or around peaks I and III (Fig. 1). There is only background bisulfite reactivity at peak II. Detailed analysis of sequenced molecules shows that in many of the molecules, the single-stranded region exists at or upstream of peak I and/or at peak III at the same time (Fig. 1A, labeled I-{alpha}; III; or I-{alpha}, III). This suggests that two local, independent non-B DNA structures may exist, sometimes within the same molecule. However, in other molecules, the single-strandedness is downstream of peak I and at peak III (Fig. 1A, labeled I-{beta} and III). This finding suggests that at peak I, there may be more than one non-B DNA structure. In addition, there are molecules with less common structures, such as single-strandedness at I-{gamma} alone or at I-{gamma} plus III (Fig. 1A). Comparable classifications can be made for the bottom strand as well, though these are less obvious (Fig. 1B). Therefore, the present study shows the precise location and nature of the non-B DNA structure formation at the level of the individual molecule within the mammalian cells. There is also an asymmetry in the bisulfite conversion between the top and bottom strands, and this asymmetry is more obvious at peak I (see below).



View larger version (70K):
[in this window]
[in a new window]
  		FIG. 1.
The nature of bisulfite reactivity at the bcl-2 Mbr on chromosomal DNA. Bisulfite sensitivity is restricted to peaks I and III. A, bisulfite sensitivity of molecules from the top strand; B, corresponding bottom strand data. The population of molecules shown have substantial focal and recurring regions of single-strandedness (non-B form). In both A and B, each row of circles represents one DNA molecule. Each dark circle represents an instance of a bisulfite-converted cytosine (single-stranded), whereas open circles represent cytosines resistant to bisulfite (i.e. double-stranded DNA). Note that only the C residues are depicted, but a scale to indicate length in nucleotides along the DNA is shown along the bottom. The designations I, II, and III correspond to the three translocation frequency peaks within the Mbr. Molecules are grouped according to similar structural configurations existing at peaks I, II, and III. In order to select the molecules with structure, we have enriched by SacII digestion after bisulfite treatment. This provides a 2–3-fold enrichment for molecules with the non-B structure at peak I, which is where SacII cleaves.

 
Non-B DNA Structure Formation Is Favored by Supercoiling—We find that molecules with continuous stretches of bisulfite conversion (i.e. molecules with the non-B form DNA structure) are highest for minichromosomes harvested from human cells (47%) and supercoiled plasmid DNA harvested from E. coli (45%), both by nondenaturing methods. However, in the case of chromosomal DNA, the frequency of molecules with structure was only about 28% (Fig. 2A). This is consistent with the large amount of superhelical tension that it is possible to achieve on circular molecules (prokaryotic and eukaryotic plasmids). In the case of chromosomal DNA, where there is a documented (32), but low level of supercoiling compared with circular DNA, a relatively lower fraction of molecules form the structure (Fig. 2A). When there is no supercoiling (on short linear DNA fragments containing the bcl-2 Mbr), we see only 10–15% of the molecules with the non-B structure (14). Similarly, we find only 18% of the molecules with the non-B structure when the plasmid DNA is linearized (Fig. 2A). Therefore, our results suggest that structure formation is facilitated when supercoiling is present. Although the frequency of molecules with the non-B structure varies in proportion to the superhelical tension, the overall bisulfite conversion frequency is comparable among the molecules with the non-B structure in both eukaryotic and prokaryotic cells (Fig. 2B).



View larger version (36K):
[in this window]
[in a new window]
  		FIG. 2.
Histograms showing frequency of non-B DNA structure formation on the basis of bisulfite sensitivity at the bcl-2 Mbr from various systems. A, frequency of molecules with seven or more conversions in a string. B, frequency of bisulfite conversion in molecules with the non-B DNA conformation. In both A and B, abbreviations are as follows. chr, chromosomal DNA; mini, minichromosome from human cells; sc, pXW5 supercoiled DNA; lin, pXW5 linear DNA; mut sc, mutant supercoiled plasmid; mut lin, mutant linearized plasmid. The mutation is the GGG to CCC change (see "Experimental Procedures"). The number of molecules analyzed in each category are as follows: chromosomal DNA, 55; minichromosomal DNA, 19; supercoiled plasmid DNA, 20; linear plasmid DNA, 34; mutant supercoiled plasmid DNA, 13; and linear mutant plasmid DNA, 12. These frequencies are a compilation of all molecules of a given source across several experiments, and error bars cannot be derived. The major differences in A are between supercoiled and linearized and between mutant supercoiled and mutant linearized. The major differences in B are between mutant supercoiled and mutant linearized.

 
Structural Analysis of Both Strands of Individual DNA Molecules—The bisulfite modification assay described thus far provides information about only one strand of a duplex DNA molecule at a time. We were interested in gaining information about both strands of the same molecule, wherever feasible, since that would dramatically help us to predict the nature of the structure. One may think of this as "full-molecule" information rather than the "half-molecule" information obtained when the two strands must be separated. Full-molecule analysis is possible in the case of E. coli plasmids bearing the bcl-2 Mbr region but on a very limited scale (see "Experimental Procedures"). We have fully characterized a small number of molecules on both strands (Fig. 3, A and B). We find that the single-stranded regions for full molecules correspond extremely well to the collective information obtained for populations of half-molecules. In particular, the regions of single-strandedness on the top and bottom portions match very well with the population information for the chromosomal bcl-2 Mbr (Fig. 1). This information confirms that the pooled half-molecule information is very similar to, if not indistinguishable from, full-molecule information. This finding also indicates that there is asymmetry between the top and bottom strands, particularly at peak I and to some extent at peak III.

Multiple Factors That Affect Triplex Formation, but Not B-DNA, Also Affect the bcl-2 Mbr Structure—Although clustering of patient breakpoints has been observed in the 150-bp region of the bcl-2 Mbr region, we could not find any canonical sequences known to form structures such as symmetric bubbles, heterologous loops, cruciforms, Z-DNA, tetraplexes (quadraplexes or G-quartets), or triplexes. However, a relatively purine-rich bottom strand just upstream of peak I of the Mbr suggested the possibility of triplex formation, albeit with obvious mismatches. In earlier studies, we have seen that the non-B structure formation at the bcl-2 Mbr is favored in the presence of Mg2+ and spermidine at neutral pH (14). Above, we have noted that the structure formation is favored when supercoiling is present, and supercoiling is known to favor triplex formation (20–22). In a putative triplex, the third strand would most likely be Hoogsteen paired with another duplex DNA portion, and the fourth strand would remain unpaired. In order to further investigate this possibility, we studied factors known to affect triplex formation. We have seen that 7-deazaadenine and 7-deazaguanine (analogues of adenine and guanine, respectively) interfere with Hoogsteen hydrogen bond formation and also abolish the non-B structure formation (14). This further supports a triplex structure, because B-form DNA does not rely on base pairing at the N7 position.



View larger version (10K):
[in this window]
[in a new window]
  		FIG. 3.
Bisulfite modification of both strands of the same DNA molecule. Plasmid containing bcl-2 Mbr (pSCR1) was subjected to bisulfite modification (see "Experimental Procedures"). The bcl-2 Mbr was then excised by restriction enzyme digestion and cloned into a fresh vector backbone. The ligated products were transformed into an ung E. coli strain (uracil glycosylase minus) to prevent repair of the uracil. Plasmid DNA was purified from these colonies and retransformed to separate the top and bottom strands of DNA. The plasmid DNA was extracted from randomly selected colonies and sequenced. Bisulfite modified thymines are indicated by dark filled circles above the top or below the bottom strands. A and B are two independent "full molecules" sequenced. The three peaks are underlined.

 
Because the maximal single-strandedness is at peaks I and III of the Mbr, mutations within these regions might affect the non-B structure. We mutated CCC -> GGG on the top strand and vice versa on the bottom strand by using a standard mutagenesis protocol (see "Experimental Procedures"); such a change would not alter B-form DNA. The 300-bp PCR fragment containing the mutated bcl-2 Mbr was then cloned into pXW5 to replace the bcl-2 Mbr. The resulting plasmid (pSCR41) was examined using the bisulfite modification assay. The mutant Mbr shows a 3-fold reduction in the fraction of supercoiled plasmid molecules that adopt significant stretches of single-strandedness at the Mbr (Fig. 2A). Interestingly, when pSCR41 is linearized before the bisulfite modification assay, we cannot detect any molecules with significant single-strandedness (Fig. 2A). These results show that the CCC -> GGG change causes a reduction in bisulfite conversion frequency in the Mbr (Fig. 2B). P1 nuclease sensitivity analyses confirm these findings (data not shown). The CCC -> GGG alteration would make a triplex structure weaker, since three Hoogsteen pairing GGG residues on the purine-rich second strand have been altered to "CCC" (see "Discussion").

Polyvalent cations are known to favor triplexes at neutral pH (21). We find a significant enhancement in single-stranded character (15% increase in the number of molecules with the non-B structure) when the bcl-2 Mbr-containing plasmid is incubated with spermine or spermidine (data not shown).

A Triplex-specific Antibody Binds to bcl-2 Mbr—To independently test for triplex character at the bcl-2 Mbr, we used monoclonal antibodies that are known to bind triplex DNA but do not bind single-stranded or double-stranded DNA (33, 34). The Jel466 antibody has a binding preference for GC-rich triplexes (33), whereas the Jel318 antibody has been shown to bind to AT-rich triplexes (34). Solid phase radioimmunoassay and competition experiments have shown that the Jel466 antibody prefers the triplex form of poly[d(TC)]·poly[d(GA)]. This antibody binds weakly to the triplex derived from poly[d(G)]·poly[d(C)], but there was no interaction with poly(T.A.T) triplexes. In contrast, Jel318 binds with T.A.T triplexes (33, 34).

For the studies here, after incubation of the plasmid, pXW5, with either of the two different triplex-specific antibodies, a 708-bp fragment containing the bcl-2 Mbr was liberated by restriction enzyme digestion. The DNA was resolved on an agarose gel, and the antibody binding was detected by Southern blotting. There is a discrete shift in the mobility of the 708-bp fragment when it is incubated with the Jel466 antibody (Fig. 4A). The intensity of the shifted bands increases according to the increase in Jel466 antibody concentration. A maximum of three shifted bands are visible (Fig. 4A). However, we do not find any specific shift in the 708-bp fragment when it is incubated with the Jel318 antibody. Moreover, control fragments lacking the Mbr do not bind the Jel466 monoclonal antibody (Fig. 4, B and C; see "Experimental Procedures"). These findings strongly argue for a triplex conformation.

To visualize the structure, electron microscopy (EM) was carried out on a DNA fragment containing the bcl-2 Mbr. A 488-bp PCR fragment containing the 150-bp Mbr region was incubated in the triplex-forming buffer (see "Experimental Procedures"). The DNA containing the altered structure was then incubated with or without antibody (either Jel466 or Jel318) and used for electron microscopy (see "Experimental Procedures"). A 425-bp control PCR fragment was amplified from 300 bp upstream of the bcl-2 Mbr and analyzed in parallel.

In most (80–85%) of the EM images, the DNA molecules show no structure formation (Fig. 5, A and G, and Supplemental Table I). However, about 15% of the molecules (19 of 133 molecules) reveal a darkening along the DNA within the Mbr, suggesting the existence of an altered DNA structure at this location (Fig. 5, B and G). The precise location of these structures match peak III in most of the cases and peak I in some other cases. The fraction of molecules obtained with such a structure by this method (15%) matched well with results of the biochemical experiments described previously (14).

After antibody incubation, in the case of the 488-bp Mbr fragment, the Jel466 antibody (which binds GC-rich triplex DNA) is bound at the Mbr region in about 17% of molecules (113 molecules of 679) (Fig. 5H and Supplemental Table I). The antibody appears to bind about one-third of the way into the DNA fragment, which is where the Mbr is located (Fig. 5, C–F, and H). An additional 5% of molecules showed binding at both the very end of the DNA fragment and internally at a position consistent with the Mbr. Thus, Jel466 binds essentially to all of the DNA molecules that contain the non-B DNA structure. (The end binding may be due to breathing at DNA ends, consistent with the binding activity in the gel shift assay for control DNA at high antibody concentrations (Fig. 4C, lane 3, and data not shown).)

When the same DNA is treated with the Jel318 (which binds AT-rich triplex DNA), no significant amount of antibody binding can be detected internal to the DNA ends. In this case, 95% of the DNA molecules are free of antibody (Fig. 5H and Supplemental Table I). Hence, the Mbr is recognized by a GC-rich triplex-specific antibody but not by an AT-rich triplex-specific antibody.

When a control DNA fragment (425 bp) lacking the bcl-2 Mbr is treated with Jel466 and examined by EM, only 5 molecules of 137 (4%) showed antibody binding internally (Fig. 5H and Supplemental Table I). The Jel318 binding pattern for the 425-bp DNA (no Mbr region) and the 488-bp Mbr DNA is indistinguishable. Hence, the GC-rich triplex-specific antibody binding is specific to the Mbr.



View larger version (41K):
[in this window]
[in a new window]
  		FIG. 4.
Triplex-specific antibody binds to the bcl-2 Mbr. A, two different triplex-specific antibodies (Jel466 or Jel318) were incubated with pXW5 in a volume of 20 µl. After restriction enzyme digestion (AatII/BamHI), bands were resolved on a 1% agarose gel, and Southern hybridization was carried out. Jel466 is specific for GC-rich triplex DNA, and Jel318 is specific for AT-rich triplex DNA. Amounts of the antibodies are indicated across the top of the gel. The large asterisk indicates the shifted bands due to the triplex-specific antibody binding to the bcl-2 Mbr. B, to test whether or not the shift that we observe is specific to the bcl-2 Mbr, Jel466 was incubated with pSCR33 (a plasmid generated after removal of the bcl-2 fragment from pXW5). Southern hybridization with the 453-bp (AatII/BamHI) probe does not show any shift in the mobility of the 453-bp fragment. C, a second control consisting of a 1.6-kb backbone region of pXW5 was tested for Jel466 binding. This region also lacks the bcl-2 Mbr. When this 1.6-kb probe is used to hybridize with the backbone of pSCR33 after Jel466 treatment, we do not observe any specific shift in the mobility. We note only nonspecific binding of the antibody when high antibody concentration is used (lane 3 and data not shown).

 
Computational Evaluation of Candidate Triplex Conformations at the bcl-2 Mbr—Based on the various lines of evidence described thus far, we have inferred some form of triplex conformation at peak I of the Mbr. Based on the extensive bisulfite sequencing studies and other experiments, it is evident that there are at least two non-B conformations of the peak I triplex structure that exist (I-{alpha} and I-{beta}). Candidate structures for peak I are shown in Fig. 6A (I-{alpha}) and Supplemental Fig. 2 (I-{beta}). Because the structure is formed at neutral pH, is dependent on divalent cations, and is favored by polycations, the most likely form of triplex is the purine-purine-pyrimidine (R.R.Y) type.

How does this putative structure fit with the observed data? The bottom strand of the duplex DNA at peak I is purine-rich. In the case of structure I-{alpha}, the bottom strand upstream of peak I (R) folds back to form an intramolecular triplex (H-DNA); hence, Hoogsteen hydrogen bonding occurs with the duplex (R.Y) from the region immediately downstream of peak I, resulting in an R.R.Y triplex. Here the R strands will be in antiparallel directions (Fig. 6A; see "Discussion" for more details). Such a triplex formation will result in the pyrimidinerich (Y') top strand remaining single-stranded (Fig. 6, A-C). This model explains the molecules with single-strandedness observed upstream of peak I by different probing methods. Similarly, the alternative triplex structure configuration of peak I, I-{beta} also can be drawn, which is again an R.R.Y triplex (Supplemental Fig. 2). This structure explains the molecules with single-strandedness at and downstream of peak I.

We have further studied the validity of the proposed structure by constructing the three-dimensional structure for the I-{alpha} structure (Fig. 6, B and C). The model was built using in-house software, NASDAC (35), and then energy-minimized using AMBER6 (36). In the model, the displaced single strand rotates around the triplex with a displacement of about 20 Å from the triplex axis and a twist of about one-third the twist of the triplex itself, but the exact position of the displaced strand is difficult to predict (Fig. 6, B and C). The H-DNA conformer minimizes to an AMBER6 energy that is comparable with, although not as favorable as, a B-DNA duplex formed by the same strands. This indicates that indeed the I-{alpha} structure is geometrically plausible.

Circular Dichroism (CD) Studies of the bcl-2 Mbr—CD spectroscopy is able to distinguish between duplex and triplex conformations because of different optical properties (37, 38), and here we use CD to test the triplex conformation in three ways. First, we test the above triplex models using component oligonucleotides. Second, we test the conditions and stability of the triplex formation. Finally, we test the molar ratio requirements of the duplex strand versus the third strand.

In order to test the I-{alpha} structure, CD spectroscopy has been carried out by mixing the 42-bp duplex DNA from downstream of peak I with the predicted segment of single-stranded DNA (called the third strand of the triplex) derived from upstream of peak I (either the purine-rich third strand or a control pyrimidine-rich third strand). The duplex DNA and third strand were incubated together in a buffer containing 20 mM PIPES, 140 mM KCl, 5 mM MgCl2, and 1 mM spermidine at pH 7.0 for 1 h at 24 °C. This particular composition and concentration of cations were used because they have been shown to mimic the eukaryotic cellular environment (39, 40). CD spectra were taken either using duplex DNA itself, the third strand alone, or duplex DNA with the third strand. In each case, spectra derived from the buffer alone were subtracted to correct for the background. CD spectra obtained from duplex DNA shows a maximum molar ellipticity of 15,000 at a wavelength of 275 nm (Fig. 7A). When we add the third strand (purine-rich, SCR114) that would form a triplex according to the model in Fig. 6A, we see the maximum absorption at or around 265 nm, and the maximum ellipticity is dramatically different (Fig. 7A), suggesting triplex formation, as has been seen for other triplexes (41). When we use a control third strand (SCR170) of the same length that is supposed to remain single-stranded, according to the model in Fig. 6A, we do not see a dramatic difference in the absorption spectra (Fig. 7A). Comparable results were obtained when we incubated the same duplex DNA with a different length purine-rich oligonucleotide (SCR163), which could form the triplex according to our model. Again, when we use a control oligonucleotide (SCR162, pyrimidine-rich) containing the complementary sequence of SCR163, we do not see any dramatic difference in the absorption spectra (Fig. 7B). These results provide strong support for the triplex region shown in Fig. 6.



View larger version (90K):
[in this window]
[in a new window]
  		FIG. 5.
Electron microscopy of the triplex-specific antibody binding to the bcl-2 Mbr. A 488-bp PCR DNA fragment containing the bcl-2 Mbr was resuspended in 10 µl of buffer containing 20 mM Tris acetate (pH 7.5), 10 mM magnesium acetate, 50 mM potassium acetate, and 1 mM dithiothreitol and incubated with 100 ng of the triplex-specific antibody, Jel466, or without antibody at 37 °C for 1 h (see "Experimental Procedures"). A and B, the 488-bp DNA fragment without the addition of the Jel466 antibody. C–F, the 488-bp DNA fragment after incubation with the Jel466 antibody. The arrows indicate locations of non-B structure (B) or of antibody bound to the structure (C–F). The scale bar is 50 nm. G, the histogram shows the frequency of non-B DNA structure formation at the bcl-2 Mbr. In G, non-B form denotes the non-B DNA structure-containing fraction of DNA molecules, and B form denotes normal duplex DNA (B-DNA). H, the histogram shows the detection of triplex antibody binding to the bcl-2 Mbr by EM. In H, 1 and 2 represent a PCR fragment that is 488 bp in length and contains the bcl-2 Mbr; 3 and 4 represent a PCR fragment that is a 425-bp control duplex DNA. In 1 and 3, these DNA fragments were incubated with Jel466 before EM. In 2 and 4, these DNA fragments were incubated with Jel318 before EM.

 
Triplex DNA is known to be destroyed upon incubation at high temperatures (21). When we formed the structure but then incubated at 80 °C for 5 min (and measured the spectrum at 80 °C), the triplex conformation appeared to be destroyed, and there is little difference in the spectra (with respect to the ellipticity or the shift in maximum ellipticity) resulting from duplex DNA alone, duplex with the triplex-forming oligonucleotide, or duplex with a control oligonucleotide (Fig. 7, compare A with C; also compare B with D).

The findings were confirmed by subtracting the spectra derived from the duplex DNA alone plus the spectra derived from the oligonucleotides alone from the corresponding spectra obtained after incubation of the third strand with duplex DNA. Only the triplex-forming oligonucleotide plus the duplex give a clear signal above background, whereas the control third strands do not (Fig. 7, E and F).

The specific structure proposed (Fig. 6) was further tested by mutating nucleotides in the third strand (SCR114) while maintaining the sequence of the duplex DNA constant. An alteration of GGG to CCC (SCR171) results in a loss of the dramatic difference in the CD spectra (Fig. 8A). In this case, the maximum amplitude is much lower than the spectra derived from the duplex DNA at SCR114. When we use a third strand with eight Gs altered to eight Cs (SCR172), the CD spectra showed comparable results as SCR171, indicating no structure formation. These results further emphasize the requirement of Hoogsteen bonding in the structure formation.



View larger version (47K):
[in this window]
[in a new window]
  		FIG. 6.
Model for the peak I R.R.Y triplex conformation. The model illustrates the interactions of two strands, which are colored 5'-red-orange-pink-3' and 5'-cyan-gray-blue-3', respectively, and shows the positions of the Watson-Crick paired strands (red and blue), the Hoogsteen-paired third strand of the triplex (pink), the displaced single strand (cyan), and the connecting loop regions (orange and gray). A, two-dimensional model of proposed H-DNA conformation of peak I. Single-stranded regions are represented by unpaired sequences. Hoogsteen base pairing is indicated by a filled oval (if there are two hydrogen bonds), by an open oval (if there is one hydrogen bond), or by a vertical line (if there is the potential for weak hydrogen bonding). The position of translocation at peak I is indicated by the long curved line between the strands in the loop region. B and C, two views of a three-dimensional model of the proposed H-DNA conformation of peak I. The model was built using in-house software, NASDAC (35), and then energy-minimized using AMBER6 (36). The B-DNA duplex regions that emerge from the H-DNA are indicated by the arrows in C. Associated sodium counterions are shown in white. In the model, the displaced single strand rotates around the triplex with a displacement of about 20 Å from the triplex axis and a twist of about one-third the twist of the triplex itself, but the exact position of the displaced strand is difficult to determine. The H-DNA conformer minimizes to an AMBER6 energy that is comparable with, although not as favorable as, a B-DNA duplex formed by the same strands.

 
When the ratios of the duplex DNA relative to the third strand are varied 1:1, 1:2, 1:3, and 1:4 and then as the concentration of the triplex-forming third strand increases, the maximum amplitude also increases significantly (Fig. 8B). However, when the control third strand is used, such an increase is only attributable to the absorbance of that single-stranded oligonucleotide (Fig. 8C). We also find a concentration-dependent shift in wavelength corresponding to the maximum amplitude in the case of the triplex-forming oligonucleotide (Supplemental Fig. 5). This was comparable with the results obtained when we used a control DNA triplex for CD formed from oligonucleotides previously shown to form such a triplex (SCR213 and SCR214 incubated with third strand, SCR213) (Supplemental Fig. 6) (40).

In order to further determine the molar interaction between the duplex DNA and the third strand, we have constructed a Job plot (42, 43), where a third strand is added to the duplex DNA in increasing molar fraction. The CD spectra were taken, and the maximum amplitude positions were plotted as a function of the increasing molar fraction of the third strand (Fig. 8D). We find maximum interaction at 40–50% of the third strand oligomer concentration. This is at or close to 1:1 stoichiometry and is consistent with the R.R.Y triplex model in Fig. 6 for peak I of the bcl-2 Mbr.

These results indicate that the bcl-2 Mbr is able to form a triplex DNA structure and supports the existence of the peak I-{alpha} structure proposed (Fig. 6).

Mutation at bcl-2 Mbr Reduces the Breakage at the Mbr in Vivo—We have seen above that a 3-nucleotide mutation (CCC -> GGG change) at peak I causes a reduction in the triplex formation at the bcl-2 Mbr. Previously, we have used episomal DNA substrates in a human recombination assay system to reconstitute the t(14;18) translocation process (13). In the present study, we used the same assay system to study the effect of triplex formation on the bcl-2 translocation. We transfected episome, pSCR41, the 3-bp mutant version of bcl-2 Mbr, or the bcl-2 Mbr containing substrate, pXW5, into Reh pre-B cells (see "Experimental Procedures"). We find that the recombination frequency of pXW5 is 0.041%. The recombination frequency of pSCR41 is 0.0035%. This represents an 11.7-fold reduction in the recombination frequency of the bcl-2 Mbr when the Mbr has the 3-bp mutation (Table I). Equally important, the pattern of the recombinant breakpoint is dramatically changed when the Mbr is mutated. In the case of pXW5, 76.3% of cases break within the Mbr (Table I). However, in the case of the Mbr mutant (pSCR41), only 51.7% of cases break at the Mbr, and 48.3% of cases break outside of the Mbr, which is indistinguishable from the expected random break distribution (Table I). Moreover, we do not find any breakpoints at the peak I, in the case of pSCR41, when the mutation is created. These results show that mutation at the bcl-2 Mbr reduces the recombination frequency of the bcl-2 Mbr. These results provide strong functional data indicating that the triplex formation is essential for the bcl-2 Mbr breakage in vivo.


View this table:
[in this window]
[in a new window]
  		TABLE I
A 3-bp exchange at the bcl-2 Mbr affects the frequency of breakage at the bcl-2 Mbr in vivo

The episomes (pSCR41 and pXW5) were transfected into Reh cells. DNA was recovered 48 h later and analyzed. The number of substrate molecules that replicated in the Reh cells is indicated as DA, which stands for DpnI-resistant, ampicillin-resistant. The recombinants that are ampicillin-chloramphenicol (double)-resistant are sequenced and analyzed. DACMbr are the number of ampicillin-chloramphenicol double resistant transformants (recombinants) obtained with a breakpoint at bcl-2 Mbr. DACother is the recombinants having breakpoints at a 160-bp region that is outside the Mbr. The recombination efficiency of Mbr is calculated by dividing DACMbr by DA. The frequency of recombinants obtained is indicated as a percentage, under the "Events" headings, and actual numbers of recombinant clones of a given type (divided by the total recombinants) are indicated in parentheses. Bcl-2 indicates recombinants obtained from Reh cells after transfecting pXW5 (which has the wild type bcl-2 Mbr sequence). "Bcl-2 mutant" indicates that cells were transfected with the mutant version of the bcl-2 Mbr episome, pSCR41.

 



View larger version (25K):
[in this window]
[in a new window]
  		FIG. 7.
Circular dichroism studies indicate triplex DNA formation at the bcl-2 Mbr. A 42-bp duplex DNA was created by annealing oligonucleotides designed from downstream of peak I of the bcl-2 Mbr. In order to test the triplex DNA formation, different oligonucleotides spanning upstream of peak I, either from the top strand (SCR162 or SCR170) or bottom strand (SCR114 or SCR163), are incubated with the duplex DNA in a buffer containing 20 mM PIPES, 140 mM KCl, 5 mM MgCl2, and 1 mM spermidine at pH 7.0 for 1 h at 22 °C. The CD spectroscopy was carried out over a wavelength range of 210–310 nm at 22 °C. In all panels, the same duplex DNA is used. There is a 4-nt length difference in the third strand; otherwise, sequences are the same between A, C, and E and B, D, and F. A, the green curve is a CD spectrum resulting from a 42-bp duplex in the above triplex buffer (pH 7.0). The black curve is the same but with the addition of a third strand oligomer (SCR114), which is purine-rich (the experimental strand, O114). The red curve is a spectrum derived from the control oligonucleotide (SCR170) represented as O170, where the added third strand is pyrimidine-rich. B, the green curve is a CD spectrum resulting from a 42-bp duplex in the above triplex buffer (pH 7.0). The black curve is the same but with the addition of a third strand oligonucleotide (SCR163), which is purine-rich and represented by O163. The red curve is a spectrum derived from a control strand represented as O162, where the added third strand is pyrimidine-rich. C and D are exactly the same as A and B, respectively, except that after the incubation in triplex buffer, the CDs were taken at 80 °C. C, the green curve is a 42-bp duplex DNA. The black curve is the same, but with the addition of a third strand (experimental, O114). The red curve is a control (O170). D, the green curve is duplex DNA. The black curve is the same but with the addition of a third strand, which is purine-rich (experimental, O163). The red curve is a control (O162). E and F are the spectra resulting after subtracting the duplex DNA and oligonucleotide spectra from the "triplex spectra." E, the black curve is the subtraction spectrum resulting from the experimental oligonucleotide (labeled O114) in standard buffer (20 mM PIPES, 140 mM KCl, 5 mM MgCl2, and 1 mM spermidine). The red curve is a control spectrum (labeled O170). F, the black curve is the spectrum derived from experimental oligonucleotides (labeled as O163) in the above buffer (pH 7.0). The red curve is a spectrum from the control oligonucleotide (labeled as O162)

 

   DISCUSSION
TOP
 ABSTRACT
 INTRODUCTION
 EXPERIMENTAL PROCEDURES
 RESULTS
 DISCUSSION
REFERENCES
 
Evidence for a Triplex DNA Structure at the bcl-2 Mbr—The analysis of individual DNA molecules from chromosomal DNA containing bcl-2 Mbr indicates two different non-B DNA structures existing simultaneously at peaks I and III in some DNA molecules (Fig. 1). In other molecules, the peak I structure and the peak III structure are present individually. In the case of peak I, two distinct non-B conformations exist (Fig. 1). These observations from chromosomal DNA are consistent with our earlier results using in vitro gel shift assays on shorter DNA fragments containing the bcl-2 Mbr (14).

The structure formation is influenced significantly by superhelical tension, as would be expected for a triplex. The E. coli-derived plasmid has the highest superhelical tension, but there is also some superhelical tension that is measurable in mammalian chromosomal DNA (32). Mutation of three G residues to C residues on the bottom strand of peak I of the bcl-2 Mbr favors the B-form configuration on a plasmid DNA containing bcl-2 Mbr. This is the strand that is expected to form the Hoogsteen hydrogen bonding in a triplex configuration (see below). The 3-bp alteration would be expected to reduce the number of Hoogsteen interactions and, hence, the stability of a triplex structure.



View larger version (25K):
[in this window]
[in a new window]
  		FIG. 8.
Structural characterization of the triplex DNA at the bcl-2 Mbr by circular dichroism. A, sequence alteration in the purine-rich third strand affects spectra of the bcl-2 Mbr. The green curve is a 42-bp duplex in the triplex buffer (20 mM PIPES, 140 mM KCl, 5 mM MgCl2, and 1 mM spermidine). The black curve is the same but with the addition of a 42-nt third strand, which is purine-rich, called O114. Hence, the structure can be referred to as R.R.Y. The red curve is a control where the added third strand is pyrimidine-rich. The turquoise curve is the spectrum for an 8-nt randomization of sequence in the middle of SCR114 represented as O172. For the blue curve, there is a 3-nt randomization of sequence in the middle of SCR114, represented as O171. B, titration of the experimental purine-rich third strand oligomer with the 42-bp duplex DNA. The turquoise curve is the duplex DNA. The blue, green, red, and black curves are the duplex + third strand oligomer (O114) added in 1:1, 1:2, 1:3, and 1:4 molar excess relative to the duplex DNA. C, same as in B, except that the third strand is a pyrimidine-rich oligonucleotide (O96). D, Job plot showing the interaction of duplex DNA with a triplex-forming third strand at peak I of the bcl-2 Mbr.

 
The full-molecule bisulfite sequencing data further support the triplex conformation. The peak I data on both molecules of full-molecule sequencing (Fig. 3) are exactly at the expected locations for the proposed peak I triplex (Fig. 6). We do find asymmetric single-stranded regions in peak III as well; however, additional studies are needed to permit proposal of an actual structure for peak III.

A triplex-specific antibody binds to the bcl-2 Mbr directly, indicating the existence of a triplex structure at the Mbr. CD spectroscopy provides a strong independent test of a specific triplex conformation at peak I.

In summary, there are 14 lines of experimentation using five completely different methods (chemical probing (bisulfite, KMnO4, OsO4); gel mobility shift (di- and polyvalent cation dependence, 3-bp mutation, and deazaadenine and deazaguanine incorporation); triplex-specific monoclonal antibody binding (EM and gel shift); circular dichroism (strand dependence, lithium independence, heat dependence, and sensitivity to mutation); and an intracellular recombination substrate assay) that all point to an R.R.Y triplex conformation.

Specific Nature of the Triplex Conformation at the bcl-2 Mbr— There are two general categories of triplexes as defined by the types of nucleotides composing the third strand. Depending on whether the third strand of the triplex is purine (R)- or pyrimidine (Y)-rich, an R.R.Y triplex or a Y.R.Y triplex may form (20, 21). Y.R.Y triplexes favor acidic pH, whereas R.R.Y triplexes prefer neutral pH and may therefore form under physiologic conditions (20, 21). R.R.Y triplexes are also known to tolerate more versatile pairing schemes and form more rapidly (24).

We observe long stretches of single-strandedness at peak I and peak III, independent of one another. This raises the possibility of two independent triplexes at these two locations. Among many possible triplexes that one could draw for peak I, we have shown two possibilities (Fig. 6A and Supplemental Fig. 2). The two structures shown for peak I are both R.R.Y configurations. Any Y.R.Y conformation appears to be unlikely, based on the polarity of the strands, Hoogsteen hydrogen bonding, CD spectroscopy, chemical probing, and oligonucleotide annealing data (Figs. 1, 3, 7, and 8) (13, 14). The full-molecule bisulfite modification assay results (derived from a plasmid bearing the bcl-2 Mbr) are quite consistent with the R.R.Y structure for peak I drawn in Fig. 6, where the middle purine (R) strand (third strand) of the H-DNA includes most of peak I and the adjacent upstream region of the Mbr. Here the duplex (R.Y) region of the triplex is derived from the region immediately downstream of peak I. Based on a three-dimensional model of the structure (Fig. 6, B and C), the bisulfite modification assay, oligonucleotide annealing, and CD spectroscopy, we determined that 28 nucleotides can undergo Hoogsteen base pairing in this structure. Of the 28 Hoogsteen triads, 8 are G.G.C, 2 are A.A.T, 1 is T.A.T, 2 are C.A.T, 8 are G.A.T, 2 are A.G.C, 3 are T.C.G, 1 is A.C.G, and 1 is C.T.A (the third strand is designated in italic type). Recently, formation of G.A.T or A.G.C Hoogsteen pairings has been reported, although these mismatched triads are less stable, because they can only form one Hoogsteen hydrogen bond (44). The T.C.G mismatched triplet also can be formed in antiparallel triplexes, but it is also weaker (only one hydrogen bond) (45). C.A.T is a novel triplet, which, based on the three-dimensional structure, contains two hydrogen bonds and is equally stable as G.G.C or A.A.T. In this structure, only 4 nucleotides are unable to find Hoogsteen partners on the third strand. The large degree of sensitivity observed on the top strand matches very well with the long single-stranded regions identified by chemical probing. Similarly, increased sensitivity noted at five cytosines on the bottom strand also can be explained because they fall exactly at the positions where the DNA must fold (loop region in Fig. 6A) for the structure to form at peak I.

Although the majority of the bisulfite sensitivity and oligomer annealing observed by the gel shift assay at peak I (I-{alpha}) of bcl-2 Mbr can be explained by the first model (Fig. 6), an alternative model (I-{beta}) is required to explain bisulfite sensitivity in a fraction of molecules (Fig. 1, compare the single-strandedness at peak I of molecules under I-{alpha} versus I-{beta}) as well as in the oligomer annealing observed on shorter DNA fragments (14). In this alternate configuration, the triplex formation also takes place using exactly the same sequences as in Fig. 6 and comparable Hoogsteen pairings (21 pairs), but the third strand (R) comes from downstream of peak I (Supplemental Fig. 2). The most compelling finding supporting the structure in Supplemental Fig. 2 (for a subset of molecules) is that the oligonucleotide SCR113 can pair completely with the single-stranded region of the triplex (14).

How do we reconcile that some of the data are most consistent with the triplex in Fig. 6, yet other data are most consistent with a very similar triplex covering the same region in Supplemental Fig. 2? We suspect that these two triplex isomers can both form and that the observed population of non-B molecules is a mixture of the two (Fig. 1). Inside the eukaryotic nucleus, each triplex form may exist in individual genomic DNA molecules, as indicated by chemical probing; however, under the conditions of the PAGE assay, either structure may predominate, based on the length of the surrounding sequence (Fig. 1) (13, 14). Triplex structures can be drawn for peak III as well (not shown), but additional experiments will be needed before there is a sufficient basis for suggesting candidate structures at peak III.

We do not observe any significant single-strandedness at peak II by any of our methods, yet peak II is no less frequently a site of chromosomal translocation. One possibility is that once structure formation occurs at peaks I and III, then the single-stranded region of the peak I triplex may pair with the single-stranded region of the peak III triplex to provide additional stabilization. The resulting higher order structure may result in sufficient distortion at peak II to favor breakage. Preliminary attempts to model conformations in which peak I and III interact by hydrogen bonding show maximal bending at peak II, consistent with this possibility.

Are there any other instances where triplex DNA results in a disease? The recently reported sticky DNA is an example of triplex DNA that is thought to occur in Friedreich ataxia (46). The authors provide evidence for two R.R.Y triplexes interacting together to form the sticky DNA on a plasmid. In another study, an Y.R.Y triplex was proposed to form at the same Friedreich ataxia sequences (47). Both of the studies show that GAA repeats can form the triplex DNA and, hence, provide a potential explanation for Friedreich ataxia. One major difference between those studies and the current one is that the Friedreich ataxia triplex is an ideal repeat sequence. In contrast, the bcl-2 Mbr sequence is not any type of repeat sequence.

Intramolecular triplexes (H-DNA) are known to form on ideal homopurine-homopyrimidine mirror repeats in vitro (19, 20, 22, 23). Ideal sequences do not, however, appear to be an absolute requirement, because intramolecular triplexes have been documented within sequences that are neither homopurine-homopyrimidine nor mirror repeats (48). Other studies also show that some mismatches are allowed during H-DNA formation (23, 49–52). In addition to mismatches, one study showed that nonmirror repeated sequences can form intramolecular triplexes (53). In the case of the bcl-2 Mbr, although the sequences are purine-pyrimidine rich, there are mismatches as well. This means that the triplex DNA formed at the bcl-2 Mbr is almost certain to have some unpaired/mismatched positions. Does this affect the stability of the triplex? The total number of hydrogen bonds and stacking (proposed in our model) help provide stabilization (Fig. 6).

How Does the RAG Complex Recognize the Triplex DNA Structure?—We have previously shown that the RAG complex is the nuclease responsible for the t(14;18) translocation (13). We also have shown that the RAG complex induces nicks at the bcl-2 Mbr. More recently, we have shown that the RAG complex induces double strand breaks at the Mbr (54). These double strand breaks occur by inducing two independent nicks on two closely located regions of the Mbr on the two antiparallel strands. We also have direct evidence that the RAG complex cleavage occurs only if the altered DNA structure is present at the bcl-2 Mbr site and not on the duplex form of the Mbr (54). The RAGs bind to and nick duplex DNA-single-stranded DNA transition points of the triplex and other non-B DNA structures, such as bubbles and heterologous loops (54). The nicking by RAGs at other non-B DNA sites in the genome is a possible explanation for other fragile sites among the lymphoid translocations.

Proposed Triplex DNA Structure at the bcl-2 Mbr May Lead to bcl-2 Translocation—Our studies thus far have suggested that triplex DNA is formed at the bcl-2 Mbr and that a 3-nucleotide exchange at peak I reduces the structure formation. By using an intracellular human recombination assay, we also were able to reconstitute the t(14;18) translocation process (13). By using the same recombination system, we now show that minichromosomes carrying the 3-nucleotide exchange at peak I significantly reduce the recombination frequency. The mutation also causes the pattern of breakpoints within the Mbr to shift away from peak I. These results suggest that breaks at the Mbr are influenced by the triplex structure. The 3-nucleotide exchange reduces the triplex structure formation and is also capable of reducing the bcl-2 Mbr recombination within the cells. Therefore, the present study provides evidence for the role of triplex DNA in the formation of the bcl-2 translocation.


   FOOTNOTES
 
* These studies were supported by grants from the National Institutes of Health (to M. R. L.). The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked "advertisement" in accordance with 18 U.S.C. Section 1734 solely to indicate this fact. Back

{boxs} The on-line version of this article (available at http://www.jbc.org) contains supplemental material, including an additional six figures and one table. Back

k To whom correspondence should be addressed: Norris Cancer Center, Rm. 5428, 1441 Eastlake Ave., MC9176, Los Angeles, CA 90033. Tel.: 323-865-0568; Fax: 323-865-3019; E-mail: lieber{at}usc.edu.

1 The abbreviations used are: Mbr, major breakpoint region; bcl-2 Mbr, major breakpoint region of the bcl-2 gene; PIPES, 1,4-piperazinediethanesulfonic acid; EM, electron microscopy; nt, nucleotide(s). Back


   ACKNOWLEDGMENTS
 
We thank Dr. D. M. Crothers (New Haven, CT) for advice on the CD experiments. We thank Dr. J. D. Griffith (Chapel Hill, NC) for advice on the EM and for providing EM support (National Institutes of Health Grant GM31819). We thank M. Bayramyan for computational assistance; Dr. S. Jayasinghe for CD spectroscopy assistance; and Drs. R. Anderson (National Institutes of Health), P. Qin (University of Southern California), and D. Shibata (University of Southern California) for discussions.



   REFERENCES
TOP
 ABSTRACT
 INTRODUCTION
 EXPERIMENTAL PROCEDURES
 RESULTS
 DISCUSSION
 REFERENCES
 

  1. Bakhshi, A., Wright, J. J., Graninger, W., Seto, M., Owens, J., Cossman, J., Jensen, J. P., Goldman, P., and Korsmeyer, S. J. (1987) Proc. Natl. Acad. Sci. U. S. A. 84, 2396–2400[Abstract/Free Full Text]
  2. Wyatt, R. T., Rudders, R. A., Zelenetz, A., Delellis, R. A., and Krontiris, T. G. (1992) J. Exp. Med. 175, 1575–1588[Abstract/Free Full Text]
  3. Jager, U., Bocskor, S., Le, T., Mitterbauer, G., Bolz, I., Chott, A., Kneba, A., Mannhalter, C., and Nadel, B. (2000) Blood 95, 3520–3529[Abstract/Free Full Text]
  4. Buchonnet, G., Jardin, F., Jean, N., Bertrand, P., Parmentier, F., Tison, S., Leprete, S., Contenin, N., Lenain, P., Stamatoullas-Bastard, A., Tilly, H., and Bastard, C. (2002) Leukemia 16, 1852–1856[CrossRef][Medline] [Order article via Infotrieve]
  5. Raghavan, S. C., and Lieber, M. R. (2004) Cell Cycle 3, 762–768[Medline] [Order article via Infotrieve]
  6. Tsujimoto, Y., Gorham, J., Cossman, F., Jaffe, E., and Croce, C. M. (1985) Science 229, 1390–1393[Abstract/Free Full Text]
  7. Bakhshi, A., Jensen, J. P., Goldman, P., Wright, J. J., McBride, O. W., Epstein, A. L., and Korsmeyer, S. J. (1985) Cell 41, 899–906[CrossRef][Medline] [Order article via Infotrieve]
  8. Cleary, M. L., Meeker, T. C., Levy, S., Lee, E., Trela, M., Sklar, J., and Levy, R. (1986) Cell 44, 97–106[CrossRef][Medline] [Order article via Infotrieve]
  9. Cotter, F., Price, C., Zucca, E., and Young, B. D. (1990) Blood 76, 131–135[Abstract/Free Full Text]
  10. Lewis, S. M., Agard, E., Suh, S., and Czyzyk, L. (1997) Mol. Cell. Biol. 17, 3125–3136[Abstract]
  11. Raghavan, S. C., Kirsch, I. R., and Lieber, M. R. (2001) J. Biol. Chem. 276, 29126–29133[Abstract/Free Full Text]
  12. Marculescu, R., Le, T., Simon, P., Jaeger, U., and Nadel, B. (2002) J. Exp. Med. 195, 85–98[Abstract/Free Full Text]
  13. Raghavan, S. C., Swanson, P. C., Wu, X., Hsieh, C.-L., and Lieber, M. R. (2004) Nature 428, 88–93[CrossRef][Medline] [Order article via Infotrieve]
  14. Raghavan, S. C., Houston, S., Hegde, B. G., Langen, R., Haworth, I. S., and Lieber, M. R. (2004) J. Biol. Chem. 279, 46213–46225[Abstract/Free Full Text]
  15. Felsenfeld, G., Davies, D. R., and Rich, A. (1957) J. Am. Chem. Soc. 79, 2023–2024[CrossRef]
  16. Chamberlin, M. J., and Patterson, D. L. (1965) J. Mol. Biol. 12, 410–428[Medline] [Order article via Infotrieve]
  17. Morgan, A. R., and Wells, R. D. (1968) J. Mol. Biol. 37, 63–80[CrossRef][Medline] [Order article via Infotrieve]
  18. Arnott, S., and Selsing, E. (1974) J. Mol. Biol. 88, 509–521[CrossRef][Medline] [Order article via Infotrieve]
  19. Frank-Kamenetskii, M. D., and Mirkin, S. M. (1995) Annu. Rev. Biochem. 64, 65–95[Medline] [Order article via Infotrieve]
  20. Sinden, R. R. (1994) DNA Structure and Function, Academic Press, Inc., San Diego
  21. Soyfer, V. N., and Potaman, V. N. (1996) Triple-helical Nucleic Acids, Springer-Verlag, New York
  22. Wells, R. D., Collier, D. A., Hanvey, J. C., Shimizu, M., and Wohlrab, F. (1988) FASEB J. 2, 2939–2949[Abstract]
  23. Mirkin, S. M., Lyamichev, V. I., Drushlyak, K. N., Dobrynin, V. N., Filippov, S. A., and Frank-Kamenetskii, M. D. (1987) Nature 330, 495–497[CrossRef][Medline] [Order article via Infotrieve]
  24. Faucon, B., Mergny, J. L., and Helene, C. (1996) Nucleic Acids Res. 24, 3181–3188[Abstract/Free Full Text]
  25. Gauss, G. H., and Lieber, M. R. (1996) Mol. Cell. Biol. 16, 258–269[Abstract]
  26. Yu, K., Chedin, F., Hsieh, C.-L., Wilson, T. E., and Lieber, M. R. (2003) Nature Immunol. 4, 442–451[CrossRef][Medline] [Order article via Infotrieve]
  27. Ausubel, F. M., Brent, R., Kingston, R. E., Moore, D. D., Seidman, J. G., Smith, J. A., and Struhl, K. (eds) (1996) Current Protocols in Molecular Biology, John Wiley & Sons, Inc., New York
  28. Griffith, J. D., and Christiansen, G. (1978) Annu. Rev. Biophys. Bioeng. 7, 19–35[CrossRef][Medline] [Order article via Infotrieve]
  29. Gauss, G., and Lieber, M. R. (1992) Nucleic Acids Res. 20, 6739–6740[Free Full Text]
  30. Gauss, G. H., and Lieber, M. R. (1993) Mol. Cell. Biol. 13, 3900–3906[Abstract/Free Full Text]
  31. Gauss, G. H., Domain, I., Hsieh, C.-L., and Lieber, M. R. (1998) Eur. J. Immunol. 28, 351–358[CrossRef][Medline] [Order article via Infotrieve]
  32. Kramer, P. R., and Sinden, R. R. (1997) Biochemistry 36, 3151–3158[CrossRef][Medline] [Order article via Infotrieve]
  33. Agazie, Y. M., Lee, J. S., and Burkholder, G. D. (1994) J. Biol. Chem. 269, 7019–7023[Abstract/Free Full Text]
  34. Lee, J. S., Burkholder, G. D., Latimer, L., Haug, B., and Braun, R. P. (1987) Nucleic Acids Res. 15, 1046–1061
  35. Chambers, E. J., Price, E. A., Bayramyan, M. Z., and Haworth., I. S. (2003) J. Biomol. Struct. Dyn. 21, 111–125[Medline] [Order article via Infotrieve]
  36. Case, D. A., Pearlman, D. A., Caldwell, J. W., III, T. E. C., Ross, W. S., Simmerling, C. L., Darden, T. A., Merz, K. M., Stanton, R. V., Cheng, A. L., Vincent, J. J., Crowley, M., Tsui, V., Radmer, R. J., Duan, Y., Pitera, J., Massova, I., Seibel, G. L., Singh, U. C., Weiner, P. K., and Kollman, P. A. (1999) AMBER6, University of California, San Francisco
  37. Wu, P., Kawamoto, Y., Hara, H., and Sugimoto, N. (2002) J. Inorg. Biochem. 91, 277–285[CrossRef][Medline] [Order article via Infotrieve]
  38. Bergethon, P. R. (1998) The Physical Basis of Biochemistry, pp. 249–281, Springer-Verlag, New York
  39. Singleton, S. F., and Dervan, P. B. (1993) Biochemistry 32, 13171–13179[CrossRef][Medline] [Order article via Infotrieve]
  40. Goobes, R., Cohen, O., and Minsky, A. (2002) Nucleic Acids Res. 30, 2154–2161[Abstract/Free Full Text]
  41. Lee, J. S., Johnson, D. A., and Morgan, A. R. (1979) Nucleic Acids Res. 6, 3073–3091[Abstract/Free Full Text]
  42. Bloomfield, D. A., Crothers, D. M., and Tinoco, I. (1974) Physical Chemistry of Nucleic Acids, pp. 322–328, Harper and Row, New York
  43. Polak, M., and Hud, N. V. (2002) Nucleic Acids Res. 30, 983–992[Abstract/Free Full Text]
  44. Sakamoto, N., Larson, J. E., Iyer, R. R., Montermini, L., Pandolfo, M., and Wells, R. D. (2001) J. Biol. Chem. 276, 27178–27187[Abstract/Free Full Text]
  45. Gowers, D. M., and Fox, K. R. (1999) Nucleic Acids Res. 27, 1569–1577[Abstract/Free Full Text]
  46. Sakamoto, N., Chastian, P. D., Parniewski, P., Ohshima, K., Pandolfo, M., Griffith, J. D., and Wells, R. D. (1999) Mol. Cell 3, 465–475[CrossRef][Medline] [Order article via Infotrieve]
  47. Gacy, A. M., Goellner, G. M., Spiro, C., Chen, X., Gupta, G., Bradbury, E. M., Dyer, R. B., Mikesell, M. J., Yao, J. Z., Johnson, A. J., Richter, A., Melancon, S. B., and McMurray, C. T. (1998) Mol. Cell 1, 583–593[CrossRef][Medline] [Order article via Infotrieve]
  48. Dayn, A., Samadashwily, G. M., and Mirkin, S. M. (1992) Proc. Natl. Acad. Sci. U. S. A. 89, 11406–11410[Abstract/Free Full Text]
  49. Macaya, R. F., Gilbert, D. E., Malek, S., Sinsheimer, J. S., and Feigon, J. (1991) Science 254, 270–274[Abstract/Free Full Text]
  50. Gilbert, D. E., and Feigon, J. (1999) Curr. Opin. Struct. Biol. 9, 305–314[CrossRef][Medline] [Order article via Infotrieve]
  51. Hanvey, J. C., Klysik, J., and Wells, R. D. (1988) J. Biol. Chem. 263, 7386–7396[Abstract/Free Full Text]
  52. Belotserkovskii, B. P., Veselkov, A. G., Filippov, S. A., Dobrynin, V. N., Mirkin, S. M., and Frank-Kamenetskii, M. D. (1990) Nucleic Acids Res. 18, 6621–6624[Abstract/Free Full Text]
  53. Klysik, J. (1995) J. Mol. Biol. 245, 499–507[CrossRef][Medline] [Order article via Infotrieve]
  54. Raghavan, S. C., Swanson, P. C., Ma, Y., and Lieber, M. R. (2005) Mol. Cell. Biol., in press

Add to CiteULike CiteULike   Add to Complore Complore   Add to Connotea Connotea   Add to Del.icio.us Del.icio.us   Add to Digg Digg   Add to Reddit Reddit   Add to Technorati Technorati    What's this?


This article has been cited by other articles:


Home page
 Genome ResHome page
H. Inagaki, T. Ohye, H. Kogo, T. Kato, H. Bolor, M. Taniguchi, T. H. Shaikh, B. S. Emanuel, and H. Kurahashi
Chromosomal instability mediated by non-B DNA: Cruciform conformation and not DNA sequence is responsible for recurrent translocation in humans
Genome Res., February 1, 2009; 19(2): 191 - 198.
[Abstract] [Full Text] [PDF]

Home page
 J Natl Cancer Inst MonogrHome page
M. R. Lieber, S. C. Raghavan, and K. Yu
Mechanistic Aspects of Lymphoid Chromosomal Translocations
J Natl Cancer Inst Monographs, July 1, 2008; 2008(39): 8 - 11.
[Abstract] [Full Text] [PDF]

Home page
 J. Cell Sci.Home page
R. Piergentili and C. Mencarelli
Drosophila melanogaster kl-3 and kl-5 Y-loops harbor triple-stranded nucleic acids
J. Cell Sci., May 15, 2008; 121(10): 1605 - 1612.
[Abstract] [Full Text] [PDF]

Home page
 MutagenesisHome page
W. Wang, J. Xu, L. Xu, B. Yue, and F. Zou
The instability of (GpT)n and (ApC)n microsatellites induced by formaldehyde in Escherichia coli
Mutagenesis, September 1, 2007; 22(5): 353 - 357.
[Abstract] [Full Text] [PDF]

Home page
 Genome ResHome page
H. Kurahashi, H. Inagaki, E. Hosoba, T. Kato, T. Ohye, H. Kogo, and B. S. Emanuel
Molecular cloning of a translocation breakpoint hotspot in 22q11
Genome Res., April 1, 2007; 17(4): 461 - 469.
[Abstract] [Full Text] [PDF]

Home page
 Nucleic Acids ResHome page
A. Bacolla, J. R. Collins, B. Gold, N. Chuzhanova, M. Yi, R. M. Stephens, S. Stefanov, A. Olsh, J. P. Jakupciak, M. Dean, et al.
Long homopurine*homopyrimidine sequences are characteristic of genes expressed in brain and the pseudoautosomal region.
Nucleic Acids Res., January 1, 2006; 34(9): 2663 - 2675.
[Abstract] [Full Text] [PDF]

This Article
Right arrow Abstract Freely available
Right arrow Full Text (PDF)
Right arrow Supplemental Data
Right arrow All Versions of this Article:
280/24/22749    most recent
M502952200v1
Right arrow Submit a Letter to Editor
Right arrow Alert me when this article is cited
Right arrow Alert me when eLetters are posted
Right arrow Alert me if a correction is posted
Right arrow Citation Map
Services
Right arrow Email this article to a friend
Right arrow Similar articles in this journal
Right arrow Similar articles in PubMed
Right arrow Alert me to new issues of the journal
Right arrow Download to citation manager
Right arrowRequest Permissions
Citing Articles
Right arrow Citing Articles via HighWire
Right arrow Citing Articles via Google Scholar
Google Scholar
Right arrow Articles by Raghavan, S. C.
Right arrow Articles by Lieber, M. R.
Right arrow Search for Related Content
PubMed
Right arrow PubMed Citation
Right arrow Articles by Raghavan, S. C.
Right arrow Articles by Lieber, M. R.
Social Bookmarking
Add to CiteULike   Add to Complore   Add to Connotea   Add to Del.icio.us   Add to Digg   Add to Reddit   Add to Technorati  
What's this?

HOME HELP FEEDBACK SUBSCRIPTIONS ARCHIVE SEARCH TABLE OF CONTENTS
 All ASBMB Journals   Molecular and Cellular Proteomics 
 Journal of Lipid Research   ASBMB Today 
Copyright © 2005 by the American Society for Biochemistry and Molecular Biology.
Advertisement
spacer
Advertisement
Advertisement