| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
,1
,
* Scientific Institute IRCCS E. Medea, Bosisio Parini (LC), Italy;
Molecular Genetics Program, Center for Medical Sciences, Wadsworth Center, Albany, New York, USA;
Dino Ferrari Centre, Department of Neurological Sciences, University of Milan, IRCCS Ospedale Maggiore Policlinico, Mangiagalli and Regina Elena Foundation, Milan, Italy; and
Department of Biomedical Sciences, School of Public Health, State University of New York, Albany, New York, USA
1Correspondence: IRCCS E. Medea, Via Don L. Monza 20, 23842 Bosisio Parini (LC), Italy. E-mail: msironi{at}bp.lnf.it
ABSTRACT
Deletions within the dystrophin gene (DMD) account for >70% of mutations leading to Duchenne and Becker muscular dystrophies (DMD and BMD). Deletion breakpoints were reported to be scattered within regions that also represent meiotic recombination hot spots. Recent studies indicates that deletion junctions arise from nonhomologous end joining (NHEJ), a major pathway for repairing DNA double-strand breaks (DSBs) in mammals. Here we show that a region in intron 47 (i.e., a major deletion hot spot in the DMD gene) generates DSBs during meiosis in yeast and harbors a cluster of previously sequenced deletion breaks. Mapping of breakpoints in 26 BMD/DMD patients indicated that the frequency of breakpoint occurrence around this region is 3-fold higher than expected by chance. These findings suggest that DSBs mediate deletion formation in intron 47 and possibly account for the high frequency of meiotic recombination in the region. Statistical analysis indicated the presence of at least one other breakpoint cluster in intron 47. Taken together, these results suggest that the primary events in deletion formation occur within discrete regions and that the scattered breakpoint distribution reflects both a variable degree of DSB end processing and the availability of a small (compared to the huge regions involved) deletion junction sample.—Sironi, M., Pozzoli, U., Comi, G. P., Riva, S., Bordoni, A., Bresolin, N., Nag, D. K. A region in the dystrophin gene major hot spot harbors a cluster of deletion breakpoints and generates double-strand breaks in yeast.
Key Words: DMD/BMD • yeast meiosis
MUTATIONS IN THE dystrophin gene (DMD) located on Xp21 are responsible for Duchenne and Becker muscular dystrophies (DMD and BMD, respectively). The gene, composed of 79 exons, spans >2.5 Mb and occupies 
0.1% of the genome (1
,2)
. Worldwide incidence of BMD is 1/30,000 male births whereas occurrence of DMD is 1 in 3500, with one-third of the cases arising from new mutations (3)
. Large intragenic deletions and duplications together account for >two-thirds of the mutations leading to DMD and BMD and, despite heterogeneity in both deletion size and location, two hot spots have been identified: a minor hot spot is located around intron 7, while the major one involves exons 40–55 (4
,5)
. These two large gene regions also represent major meiotic recombination hot spots (6)
. Sequencing of deletion junctions involving both hot spots revealed no clear clustering of chromosome breaks and failed to identify any element or feature that could account for deletion formation (7
8
9
10)
. Analysis of available deletion junctions also indicated that they arise from rejoining of broken ends via nonhomologous end joining (NHEJ) (8
9
10)
. NHEJ plays a central role in processes of illegitimate recombination in mammals (reviewed in ref. 11
) and represents a major pathway for the repair of DNA double-strand breaks (DSBs) (11). These observations suggest that the primary event leading to dystrophin deletions could be DSB formation within the DMD locus.
DSBs are initiators of several genetically programmed events, such as V(D)J recombination and a high level of meiotic exchange (12
13
14)
; however, DSBs also arise during DNA replication and due to other endogenous processes. In the yeast Saccharomyces cerevisiae, formation of meiosis-specific DSBs has been studied extensively both genetically and physically; DSBs generally occur at specific locations, the so-called recombination hot spots (reviewed in ref. 15
). Yeast artificial chromosomes (YACs) containing human genomic DNA, covering portions of the DMD gene, undergo DSB formation during meiosis in yeast, and the frequency of YAC double-strand breakage reflects the degree of meiotic recombination in humans (16)
. These observations suggest that sequence or structural features that contribute to DMD genomic instability and DSB formation are preserved in yeast and represent target sites for meiotic DSBs. We have suggested (10)
that intervals with higher breaking frequency within already described deletion-prone regions could exist in the dystrophin gene. In particular, upon analysis of deletion breakpoints in a population of BMD patients, we identified a 1250 bp interval in intron 47 (IVS47) where three deletion breakpoints were located; given the size of IVS47 (54 kb), the close proximity of these breaking sites can hardly be due to chance.
We therefore designed experiments to investigate whether the region encompassing the three IVS47 breakpoints also generates DSBs during yeast meiosis. Several other deletion breakpoints have been identified and sequenced within the large intron 44 (IVS44, 148 kb). To determine whether they also represent DSB sites, we included three intron 44 regions in our studies. Our data show that a region in IVS47 induces DSBs during meiosis in yeast and harbors a breakpoint cluster, suggesting that deletion events arise via a DSB repair pathway.
MATERIALS AND METHODS
Yeast strains, media, and genetic techniques
Saccharomyces cerevisiae strains were derived from AS4 (
trp1–1 arg4–17 tyr7–1 ura3 ade6) and AS13 ( leu2-Bst ura3–52 ade6) background (17)
. All his4-DMD mutant alleles were introduced into the AS13 chromosome by a two-step replacement procedure. The rad50S mutation was introduced into the chromosome by use of the plasmid pNKY349, as described previously (18)
. Haploid and diploid strains used in this study were as follows: DNY107 (AS4, rad50S); DNY115 (AS13, rad50S X DNY107); MSY11 (AS13 his4-DMD71 X DNY107); MSY12 (AS13 his4-DMD47–1 X DNY107); MSY13 (AS13 his4-DMD47–2 X DNY107); MSY14 (AS13 his4-DMD47–3 X DNY107); MSY16 (AS13 his4-DMD44–3 X DNY107); MSY27 (AS13 his4-DMD44–1 X DNY107); MSY28 (AS13 his4-DMD44–2 X DNY107); MSY35 (AS13 his4-DMD47–4 X DNY107); MSY36 (AS13 his4-DMD47–5 X DNY107); SRY17 (AS13 his4-DMD47–6 X DNY107); SRY19 (AS13 his4-DMD47–7 X DNY107); SRY20 (AS13 his4-DMD47–8 X DNY107).
Standard genetic methods and media were used (19)
. Diploid strains were sporulated in 1% potassium acetate as described previously (18)
.
Plasmid construction
All DMD insertions were made at the SalI site within the coding sequence of the HIS4 gene. DMD sequences were inserted into the unique SalI site present in the plasmid pBC2, which contains an XhoI/BglII fragment of the HIS4 coding region inserted into the SalI and BamHI-digested pRS306 (20)
. DMD fragments were polymerase chain reaction (PCR) -amplified from human genomic DNA using primer pairs carrying SalI linkers. Primer sequences are provided in the Supplemental Methods. PCR reactions were carried out in 50 µl volumes using JumpStart REDAccuTaq DNA polymerase (Sigma-Aldrich, St. Louis, MO, USA); the annealing temperature was 60°C in the first three cycles and 68°C for the remaining 32 cycles; extension was carried out at 68°C (following the manufacturer’s specifications). PCR products were then digested with SalI and ligated into SalI-digested pBC2. All constructs were confirmed by DNA sequencing. The unique SnaBI or ClaI sites within the HIS4 sequence, present in all of the constructs, were used for targeting.
Physical analysis of meiotic DNA
Meiotic DNA was isolated as described (18)
and digested with either PvuII or NsiI. The resulting fragments were separated on a 0.8% agarose gel. The DNA was transferred to a nylon membrane, which was then hybridized with a 32P-labeled XhoI-BglII fragment of HIS4 as a probe.
Patient selection and deletion junction sequencing
The "Telethon Bank of DNA, Nerve and Muscle Tissues" (no. GTF02008) located at the Department of Neurological Sciences, I .R.C.C.S. Ospedale Maggiore Policlinico, Mangiagalli and Regina Elena Foundation, Milan, Italy, was the source of the DNA samples used in this study. Written informed consent was obtained (and preserved in original) from all subjects or their caregivers at the time of primary diagnostic procedures, with explicit consent given to future use for research purposes, according to the Declaration of Helsinki. This protocol was approved by the Institutional Review Board of the I .R.C.C.S Ospaedale Maggiore Policlinico, Mangiagalli and Regina Elena, Milan, Italy. Dystrophin gene deletion analysis was performed using routine procedures (21
,22)
. Mapping of intron 47 breakpoints was performed as already described (10)
; in particular, markers IVS47 C3, W3, Z3, and A4 (Fig. 1
) were used to map breakpoint locations. Deletion junctions were obtained either by mapping of the other break site, followed by long-range direct PCR (described in ref. 10
) or by long-range inverse PCR (in the case of junction B4). PCR reactions were carried out using JumpStart REDAccuTaq DNA polymerase (Sigma-Aldrich) with 7 min elongation. For inverse PCR, 2 µg of genomic DNA was digested with BglII and purified using microcon YM-30 columns (Millipore, Bedford, MA, USA); 200 ng of digested DNA was then ligated in a final volume of 200 µl for 14 h at 16°C. The ligation mixture was column-purified and resuspended in 20 µl. PCR reactions were performed using 2 µl of ligated DNA and JumpStart REDAccuTaq DNA polymerase; the elongation time was set to 7 min. Two nested PCR reactions were carried out; primer sequences were as follows: InvC3-For, GAGATCAATTAAACATGGAG; InvC3-Rev, CTCCAAGGTTGTGCATGAAAC; InvC3-ForNes, GGATTAACATACGTAAGGTTG; InvC3-RevNes:CTACAGTTTTGTCTGAAGCAC. The PCR product was gel purified and sequenced with the same primers that had been used for amplification.
|
Cell lines, DNA extraction, and sodium bisulfite modification assay
An Epstein-Barr transformed lymphoblastoid cell line and SY-5Y neuroblastoma cells were grown in RPMI or DMED medium, respectively, supplemented with 10% calf serum. For DNA extraction, harvested cells were lysed in TE (10 mM Tris, pH 8, 1 mM EDTA) with 0.5% SDS, followed by proteinase K digestion. A saturated solution of NaCl was then added and DNA was precipitated using isopropanol. For the sodium bisulfite modification assay, 30 µl of 10 mM hydroquinone and 520 µl of 2M sodium metabisulfite were added to 50 µl of genomic DNA (5 µg). The reaction was covered with mineral oil and incubated for 18–20 h at 50°C. The DNA was then purified using Microcon YM-30 columns and desufonated by addition of NaOH to a final concentration of 0.3M for 10 min at room temperature. The DNA was then recovered by ethanol precipitation and resuspended in 50 µl of TE (10 mM Tris, pH 8, 1 mM EDTA). PCR reactions were carried out using 1.5 µl of treated DNA and JumpStart REDAccuTaq DNA polymerase; 35 amplification cycles were performed. Purified PCR fragments were then cloned in bacterial plasmids using the TOPO-TA cloning kit (Invitrogen, Carlsbad, CA, USA). At least 10 clones were analyzed for each amplicon.
Analysis of breakpoint distribution
Kernel density estimates were used to identify deletion breakpoint clusters in intron 47. A 29 deletion breakpoint sample was retrieved from previous studies (8
9
10
,23)
; 10,000 breakpoint samples of equal size to the original were independently drawn from a uniform distribution over the intron sequence. Kernel density estimates were computed for each sample using a gaussian kernel and a band width of 1000 bp. A pointwise envelope was obtained from the 95th percentile of the kernel estimate of simulated samples at each position. Deletion breakpoint clusters were identified as regions where the density estimate of the original sample was greater than the 95% envelope.
RESULTS
A region in intron 47 induces DSBs during meiosis
We previously hypothesized (10)
that DMD genomic regions containing multiple breakpoints may include sites of chromosomal breakage. Formation of DSBs has been extensively characterized in the yeast S. cerevisiae, using both physical and genetic approaches. We have used the yeast system to test our hypothesis that much of the deletion events in the DMD gene occur via DSB formation.
The ability to induce DBSs during yeast meiosis was initially assessed for seven fragments from the DMD gene major hot spot. All human sequences were inserted within the HIS4 locus, which exhibits a high level of meiotic recombination in our strain background. Three inserts derived from DMD IVS44, and each covered the region where a deletion breakpoint had been located (10)
. In particular, regions spanning 800 bp (from –200 to +600, where+1 indicates the breakpoint) surrounding each break site were cloned. Three other inserts derived from IVS47 and their relative locations are shown in Fig. 1
. As mentioned above, we earlier identified a 1250 bp region in intron 47 where three deletion breakpoints were located (Fig. 1)
; we thus speculated that some sequence elements within this region induce DSBs, ultimately leading to deletion events. The three tested inserts were designed to cover the regions surrounding the three breakpoints, with two of them partially overlapping (Fig. 1)
. Finally, a control insert was cloned within HIS4; this sequence derives from DMD IVS71, a region within the DMD gene where deletions are extremely rare (Leiden Muscular Dystrophy Pages, http://www.dmd.nl/). The rad50S mutation, which prevents processing of DSBs, was subsequently introduced into the yeast strains carrying HIS4-DMD alleles. Diploids were then obtained by mating each strain with DNY107.
All diploid strains were sporulated, and meiotic DNA was extracted from samples collected 0, 8, and 24 h after transfer to the sporulation medium. Meiotic DNA was physically analyzed in order to evaluate whether any of the inserted fragment was able to induce DSBs during yeast meiosis. Meiotic DNA was digested with either PvuII or NsiI (the latter was used for MSY28, which harbors a DMD insert containing a PvuII site) and subjected to Southern blot. Most meiotic recombination at the HIS4 locus occurs due to DSB formation at the HIS4 promoter (referred to as site I in Fig. 2
). As diagrammed in Fig. 2
, in all cases a PvuII enzymatic digestion and Southern hybridization with an XhoI-BglII fragment of the HIS4 coding region as a probe is expected to yield a band (2.4 kb) corresponding to the almost-complete wild-type (WT) HIS4 gene and a portion of BIK1 plus a fainter product corresponding to DSB formation at HIS4 promoter (site I, 1.9 kb). These are the only detectable products when DNY115 (HIS4/HIS4 rad50S/rad50S) DNA is digested with PvuII (Fig. 3
B). Strains carrying heterozygous his4-DMD alleles are expected to yield at least two additional bands: one corresponding to the inserted allele, and one due to DSB formation at site I on chromosome containing the DMD insert (Fig. 2)
. Similar patterns (although with differing sizes) are expected to derive from NsiI digestions (Fig. 2
and Fig. 3C
). A PvuII digestion of MSY14 (HIS4/his4-DMD47–3 rad50S/rad50S) meiotic DNA generated an extra band (1.5 kb) (Fig. 3A
). The size of this band suggests it originated through formation of DSBs within the DMD insert; moreover, the smeared appearance of the band suggests that breaks had occurred at multiple sites. No other strains showed any evidence of additional bands, or therefore of DSB formation (Fig. 3A-C
).
|
|
Refinement of the DSB-prone region
Since inserts DMD 47–3 and DMD 47–2 partially overlap (Fig. 1)
(although only the former induced DSBs in yeast), we speculated that the sequence responsible for DSB formation resides in the non-overlap region (186 bp). Also, we wished to determine whether the same 186 bp fragment was still able to induce DSB formation when located at the 3' end of an insert. To investigate this possibility, we constructed yeast strains carrying inserts DMD 47–4 and 47–5 (Fig. 1)
within the HIS4 locus as described above. In addition, the 186 bp non-overlapping region was split into three smaller fragments (47–6, -7, and -8, Fig. 1
), and these were used to construct three additional strains. The diploid strains were sporulated and meiotic DNA was analyzed physically (Fig. 4
). The 186 bp region induced DSBs, although with much lower frequency than the longer 47–3 fragment (Fig. 4A
). Conversely, no DSB was observed when the 186 bp region was located at the 3' end of an insert (DMD 47–5, MSY 36) or when split into three smaller fragments (Fig 4B
).
|
DSBs occur within the DMD-47 inserts
We next wished to determine whether DSBs occurred within the DMD inserts or in the flanking HIS4 sequences. To this aim, the 0 h sample of DNY115 was digested with PvuII and SalI, then compared to the 24 h samples of MSY14 and 35 digested with PvuII. In particular, digested DNA was blotted and hybridized to either the right or left probe (see Fig. 2
). In case DSBs occurred on one side of the DMD insert, one of the two DSB-derived fragments from MSY14 and MSY35 would be equal to or smaller than the control fragments derived from DNY115 DNA. The results shown in Fig. 5
indicate that both fragments generated by the DSB were larger than the control DNY115 fragments, suggesting that DSBs occur within the DMD inserts.
|
Deletion breakpoints are clustered in intron 47
We next wished to better investigate whether the DSB-prone sequence we identified in the above experiments also represented a region where deletion breakpoint locations are more frequent than expected. To further analyze breakpoint distribution in IVS47, we retrieved from published works (8
9
10
, 23)
all sequenced deletion breakpoints located in intron 47; a total of 29 distinct locations was obtained. Another described breakpoint was found to map within the IVS47C3-A4 region (B7 in Fig. 1
) (8)
. Analysis of breakpoint distribution was performed using kernel density estimates. As shown in Fig. 6
, two IVS47 regions displayed breakpoint density estimates greater than the 95% envelope value deriving from sample simulations; the first region is centered around the DSB-prone region; the second cluster is located further downstream and contains four breakpoints.
|
We consequently wished to verify whether in a population of DMD/BMD patients carrying deletions involving intron 47, the frequency of breakpoints around the DSB-prone region was higher than expected. We mapped breakpoint locations for 26 DMD/BMD patients who were predicted, on the basis of deletion analysis, to display one breakpoint within intron 47. Three patients (11.5%) were found to have a deletion breakpoint mapping between markers 47C3 and 47A4. In particular, the three breaks were located in a region shorter than 2 kb centered around marker IVS47W3 (breakpoints B4-B6 in Fig. 1
); deletion junctions were obtained for the three patients and are reported in Supplemental Fig. 1. Given the size of intron 47 (54.22 kb), breakpoint occurrence in this region is 3-fold higher than would be expected if breaks were randomly distributed (3.7%). Therefore, an unusually high deletion breakpoint density is observed around the DSB-prone region we have identified above.
Sequence analysis of the DSB/deletion-prone region
To identify any sequence or feature that might be responsible for DSB formation, we searched the 47–3 region of IVS 47 for previously identified motifs associated with recombination hot spots (as listed in ref. 24
); no recombinogenic motif was found to be over-represented within the DSB-prone region relative to either IVS47 flanking sequences (IVS47C3-A4 interval) or IVS71. Further, no CoHR consensus (25)
was retrieved in this IVS47 region. The presence of direct, inverted, complementary, and mirror repeats was assessed using an already described algorithm (26)
. No evident association between local complexity properties and DSB induction or breakpoint location was observed (Supplemental Fig. 2B).
We next wished to verify whether non-B DNA structure formation might have a role in DSB or deletion induction. Sodium bisulfite has been widely used to analyze DNA single-strandedness at the single-molecule level, since this compound converts unpaired cytosines to uracils, and therefore to thymines after PCR amplification. We subjected the region-spanning markers IVS47 C3 to A4 (4 kb) to bisulfite analysis by splitting it into overlapping fragments so as to obtain PCR amplicons in the range of 800–900 bp. Treated DNA deriving from two different cell lines (Epstein-Barr virus transformed lymphoblasts and SY-5Y neuroblastoma cells) was used as a template. Since the modification pattern was indistinguishable between the two (data not shown), only data corresponding to SY-5Y neuroblastoma cells were used for analysis. The same procedure was applied to the entire intron 71 (4.3 kb), which functioned as a control for background modification. In the IVS47 region and in IVS71, the overall cytosine conversion percentage was similar, amounting to 16.99 and 18.09, respectively. No evident association was observed between conversion occurrence and location of the DSB-prone region or breakpoints (Supplemental Fig. 2C). To determine whether long regions of single strandedness (possibly corresponding to non-B structures) might be masked by background conversion, we searched for regions where seven or more consecutive cytosines were modified in at least one clone, as previously suggested (27)
. The IVS47 region displays more regions of consecutive cytosine modification than does IVS71 (Supplemental Fig. 2D). The DSB-prone fragment contains regions of single strandedness, although such regions are also present in DMD inserts that did not generate DSBs. In the context of breakpoints, four (B1, B4, B5, B7) were located within a region where several stretches of consecutive cytosine conversion occur, while the remaining three breaks did not show close association with single-stranded regions. Bisulfite sensitivity was also assessed for meiotic DNA deriving from yeast strains carrying the DMD inserts; the pattern for each DMD region was highly similar between human and yeast (data not shown).
DISCUSSION
Duchenne muscular dystrophy, with an incidence of 1:3500 live male births, is one of the most common X-linked genetic diseases. Unlike many other recurrent chromosome rearrangements, no causative sequence or structure has been identified for induction of deletion events within the locus. Recent work suggested that formation of DSBs represents the primary event in deletion formation (8
9
10)
, and the occurrence of DSBs would also help explain the elevated meiotic recombination rate involving the locus in general and the two deletion hot spots in particular (6)
. The yeast S. cerevisiae has been shown to be a good model for the study of DSB-prone regions in the human genome, and the ability of DMD sequences to generate DSBs during yeast meiosis had been reported through the use of YACs containing various gene regions (16)
.
Our results showed that a good correlation exists between regions of the DMD locus that induce meiotic DSBs in yeast and the deletion breakpoints observed in human patients. If there is a hot spot for DSB formation, and if DSBs are responsible for deletion events, we would expect a high concentration of deletions breakpoints near the hot spot. Indeed, cluster analysis indicated that the frequency of deletion breakpoints around the DSB-prone region is significantly higher than would be expected on the hypothesis of random occurrence. Moreover, among 26 patients analyzed in this study, 11.5% of intron 47 deletion breakpoints were located within a 2 kb region centered around marker IVS47W3 (Fig. 1)
, resulting in a 3-fold higher breakpoint occurrence than expected by chance. The same region generates a high level of meiotic DSBs (Fig. 3A
and Fig. 5
), suggesting that DMD deletion events with breakpoints in this interval are likely to arise via DSB formation. Once a DSB is formed, several events can lead to the formation of a deletion. For example, the broken ends may be degraded by exonucleases, then sealed by NHEJ. Alternatively, the ends of two DSBs can be ligated, resulting in deletion of the intervening sequences. DSBs generated by the meiotic recombination machinery are normally repaired by homologous recombination. However, it is possible that when two breaks are formed relatively close to one another, they are channeled into the NHEJ pathway, resulting in the deletion of the intervening sequences. It is worth noting that although little is known about the extent of cellular processing of DNA ends resulting from DSBs, earlier studies (28
, 29)
indicate that DSB induction in rodent cell lines can result in variable DNA loss (up to 8 kb) at illegitimate recombination junctions. We therefore consider it likely that DSB formation within the region we have identified in IVS47 (i.e., in fragment 47–3) accounts for deletion breakpoints located further up- or downstream in the intron (including all breakpoints shown in Fig. 1
) through loss of genetic material at DSB sites during the NHEJ process.
In yeast, meiotic DSBs normally form at intergenic regions that contain open chromatin structures (30
31
32
33)
. However, in yeast and mammals, sequences that have the potential to form secondary structural motifs (hairpins, cruciforms, triplex structures, etc.) in vivo are known to stimulate recombination, and they represent preferred sites for chromosome breakage (27
, 34
35
36
37
38
39)
. For example, breakpoints of gross deletions in the PKD1 gene were shown to coincide with non-B DNA conformations (38)
. We were unable to identify any repeated sequences that could give rise to secondary structures in vivo or any previously identified sequence motifs within the 47–3 sequence, including the GC content (31.1% for 47–3), that are believed to generate meiotic DSBs. Sodium bisulfite modification analysis was also performed in an attempt to identify any sequence within the DSB-prone region that might form non-B DNA structures, since we speculated that such structures might not be identifiable by the use of existing algorithms. Indeed, the sodium bisulfite modification assay was recently applied to identify formation of secondary structural motifs within the Bcl-2 major breakpoint region, which is cleaved by the RAG complex and is involved in lymphoma translocations (27)
. The IVS47 region and intron 71 showed a cytosine conversion ratio (17–18%) that is higher than the conversion percentage seen in the Bcl-2 major hot spot. However, this difference may not be related to the high DSB frequency observed in our studies for the following reasons. First, data derived from different experimental systems may not be directly comparable. Second, no difference was observed between the IVS47 deletion-prone region and the intron 71 sequence. Also, the slightly higher frequency of consecutive cytosine conversion (
7 consecutive Cs) in the IVS47 region as compared to intron 71 does not warrant further speculation given that no association could be identified between conversion peaks and either DSB formation or breakpoint location. One possibility is that chromatin conformation—in particular, nucleosome positioning—competes with secondary structure formation, as previously indicated (40)
, and consequently favors some structures while hindering others. In this case, modification profiles obtained for naked DNA may not precisely reflect secondary structure formation in living cells. Otherwise, either additional or different mechanisms must be invoked to account for DSB/deletion frequency within the analyzed intron 47 region. In analogy with yeast recombination hot spots (30
31
32
33)
, the DMD 47–3 region may adopt an open chromatin structure, allowing recruitment of the recombination machinery. Another possible mechanism is that DMD regions that exhibit a high level of DSBs may fortuitously create some protein binding site. The binding of such proteins may recruit the meiotic recombination machinery at the DSB site via protein-protein interaction. This hypothesis gains support from the observation that Spo11 (which is believed to be the catalytic subunit of the enzyme that generates meiotic DSBs), when fused with the Gal4 DNA binding domain, not only catalyzes DSB formation at normal meiosis-specific sites but also stimulates DSB formation near the Gal4 binding sites (41)
. Alternatively, the binding of the protein may generate open chromatin structures, thus allowing the recombination machinery to generate DSBs.
Although several breakpoints have been detected in intron 44, no DSB was observed in our experiments. In our assay, we can detect DSBs that are produced at a level that is >0.1% of the total meiotic HIS4-specific DNA. It is possible that intron 44 has weak DSB sites that cannot be detected by our physical assay. A second possibility is that DSB sites reside outside the intron 44 sequence analyzed in our assay. Indeed, as mentioned above, while the presence of multiple localized breakpoints may represent numerous break sites, it is also possible that multiple breakpoints arise from a single DSB. In this latter case, we might have failed to test the location(s) where the DSB(s) is formed. Moreover, our results indicate that a 186 bp DNA fragment within the 47–3 insert is sufficient to generate DSBs during meiosis; however, the neighboring sequence also plays an important role, since the frequency of DSB formation was reduced when the context was changed. The absence of DSBs in the analyzed intron 44 sequence and in intron 47 sequences that failed to generate DSBs may also reflect the absence of correct sequence context that can induce the generation of meiotic DSBs. It is also possible that a portion of deletion events in intron 44, as well as in other DMD gene regions, arises via a mechanism different from DSB formation of such replication slippage between short direct repeats, which represent a common finding at DMD deletion junctions (see refs. 8
9
10
and junction B5 in the Supplemental online material).
One of the inserts in our studies (insert 44–1) contained a 30 bp TG.AC dinucleotide repeat, where a deletion breakpoint had been identified; these sequences are highly repetitive in the human genome and in other eukaryotic genomes as well. In a previous study, the evolutionarily conserved TG.CA dinucleotide repeats have been shown to promote reciprocal exchange during meiosis (42)
. Our results indicate that the increased frequency of reciprocal exchange is not due to recombination initiation within or in the immediate vicinity of the dinucleotide repeats, since the insert containing TG dinucleotide repeats failed to induce meiotic DSBs. It is possible that the repetitive sequences facilitate resolution of the Holliday junctions that are formed due to recombination initiation outside the repetitive sequences.
The major dystrophin deletion hot spot (exons 40–55) covers 715 kb and harbors 75% of deletion breakpoints involving the gene (Leiden Muscular Dystrophy Pages, http://www.dmd.nl). In the commonly held view, breakpoint locations form a continuum along this huge region. Our data indicate that clustering of deletion breakpoints occurs within intron 47 (which harbors >10% of deletion breakpoints in the whole gene) and that one of the two clusters we identified maps to a region where DSBs occur during yeast meiosis. This finding suggests that DSBs mediate deletion formation in intron 47 and may account for the high frequency of meiotic recombination in the region. These results could indicate that the primary events in deletion formation occur within discrete regions and that the scattered breakpoint distribution reflects the variable degree of DSB end processing as well as the availability of a small (compared to the huge regions involved) deletion junction sample. It has been shown that clustering of recombination hot spots occurs in mammals over long genomic regions (43
44
45)
; we speculate that the DMD major hot spot represents one such region, and that deletion and recombination hot spots can colocalize and have a common mode of origin (i.e., DSB formation).
ACKNOWLEDGMENTS
We are grateful to Dr. R. Giorda for providing cell lines and for useful discussion. We also thank Drs. R. Cagliani, M. C. Bonaglia, and G. Menozzi for their comments on the manuscript. Work in D.N.’s laboratory was partially supported by National Institutes of Health grant GM56266.
Received for publication December 22, 2005. Accepted for publication April 27, 2006.
REFERENCES
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
