An efficient chemical method to generate repetitive sequences depleted DNA probes

Abstract

We describe an efficient method to remove repetitive sequences from DNA of microdissected whole chromosomes, chromosome arms, locus specific probes, and specific bands. The chemical approach described removes human repetitive DNA sequences from a source DNA, eliminating the need for blocking such sequences when the source DNA used as a probe is hybridized with a target DNA. It employs subtracting hybridized biotin-labeled repetitive-sequence DNA complex with phenol and chloroform after incubation of hybridized products with avidin. The method produces unique products that are formed after such repetitive sequences have been removed from the DNA. We have applied our newly designed subtraction strategy to microdissected chromosomes in generating whole chromosome painting (e.g., chromosome 4), chromosome arm (e.g., 1q and 15q), and band (e.g., 3q26) specific probes, and we have demonstrated its utility using flow-sorted chromosome. FISH analyses using these probes show consistent strong signals with no cross-hybridization, and optimal hybridization results can be obtained relatively quickly. © 2006 Wiley-Liss, Inc.

INTRODUCTION

It has been known for decades that chromosome rearrangements exist in most, if not all, human tumors [Mitelman et al., 1991] and certain human hereditary diseases [Frezal and Schinzel, 1991]. Distinct chromosomal abnormalities in tumors lead to the activation of proto-oncogene products, creation of tumor-specific fusion proteins, or inactivation of tumor suppressor genes. Since chromosomal banding techniques were developed, cytogenetic study of nonrandom chromosome abnormalities in malignant cells has become an integral part of the diagnostic and prognostic workup of many human cancers [Sandberg, 1990]. Additionally, cytogenetic studies followed by molecular analysis of recurring chromosomal rearrangements have greatly facilitated the identification of genes related to the pathogenesis of both hereditary diseases and cancer. For example, the tumor suppresser gene RB-1 was identified based on the observation of deletion of chromosome 13q14 in retinoblastoma [Yunis and Ramsay, 1978] and the proto-oncogene C-MYC was shown to be involved in the chromosome translocation t(8;14) in human Burkitt's lymphomas [Zech et al., 1976]. However, not all cytogenetically visible chromosome rearrangements can be determined by conventional cytogenetic-banding analysis. This technique's limitation prevents complete karyotypic analysis in many human cancers, particularly solid tumors. However, this technical limitation has been complemented by the development of fluorescence in situ hybridization (FISH) technique [Pinkel et al., 1988]. After a decade's effort, a variety of fluorescent probes, such as human whole chromosome painting probes [Collins et al., 1992], chromosome arm-specific probes [Guan et al., 1996], and chromosome band-specific probes [Meltzer et al., 1992; Guan et al., 1993, 1995], have been developed and widely applied in both research and clinical diagnosis.

The major problem with the vast majority of commercially available fluorescent DNA probes is that they contain many different kinds of repetitive sequences. The most common is a 300-bp Alu repeat, occurring on average once every 3 kb and accounting for approximately 9% of the genome [Singer, 1982]. A long repeat sequence L1 appears in genomic DNA approximately every few kilo-bases. Because of these repetitive sequences, Cot-1 DNA is widely used as a blocking agent to inhibit hybridization of repeats present within DNA probes with those scattered throughout the target genome [Lichter et al., 1988; Pinkel et al., 1988], thus reducing background signals. Cot-1 DNA has a double-stranded structure and represents sequences repeated more than 10,000 times per haploid human genome.

The disadvantages of using these conventional background blocking methods include: (1) the pre-hybridization process tends to decrease the fluorescent signals due to self hybridization of the unique sequences in the probe before hybridization to the target sequences; (2) technical experience is required to effectively block the repetitive sequences in the probe; (3) human Cot-1 DNA is expensive; and (4) the pre-hybridization process is time-consuming.

In an attempt to overcome these disadvantages, removal of repetitive sequences from probes has been previously reported by magnetic purification and PCR-assisted affinity chromatography [Craig et al., 1997]; but this method has not been widely applied in laboratories. Their method requires magnetic beads and a magnetic separator and technical knowledge in its use. A more practical and efficient method to remove repetitive sequences is needed. We have designed a novel method of removing repetitive sequences from DNA generated by chromosome microdissection. Our method can generate probes that are more specific. We have improved upon and employed a method first used by Clapp [1996], that is, removing biotinilated hybridized repetitive sequences by incubating the mixture with avidin and subtracting hybridized repetitive sequences with phenol and chloroform. The reference mixture consists of hybridization of the DNA with biotinilated repetitive DNA sequences. The remaining DNA post-subtraction is then recovered from solution by formation of a precipitate using acetate and alcohol.

In this report we describe our successful removal of repetitive sequences from microdissected DNA for chromosome 4, 1q, 15q, and 3q26 using this new subtraction strategy. The repetitive sequence-depleted (ReSeD) DNA as probes produced strong, uniform, and specific fluorescent signals after hybridization with little to no background staining when FISH was performed on metaphase chromosomes after 3 hr of hybridization.

METHODS

Summary of Procedures

Repetitive sequences are removed from a source DNA (or a DNA probe) by hybridizing the source DNA containing both unique and repetitive sequences with a driver DNA containing predominately repetitive sequences that hybridize with the repetitive sequences of the source DNA. In the process undesirable repetitive sequences of the source DNA and the driver DNA hybridize to form a product. The repetitive sequences of the source DNA, the driver DNA, or both are attached to a protein-based label moiety that is transferred via hybridization to the hybridized product. The hybridized product containing the repetitive sequences is then extracted with a solution that dissolves or separates proteins from nucleic acid molecules to remove the repetitive sequences from the product. The remaining portions of the source DNA form a nucleic acid probe having a substantial portion of the repetitive sequences removed therefrom.

The human driver DNA that is part of the mixture may be human Cot-1 DNA or other DNA fragments that contain primarily repetitive sequences. These DNA fragments are biotin-labeled.

The reaction that takes place is hybridization between the source DNA and the driver DNA that has biotin-labeled repetitive sequences. After the reaction has been completed, the hybridized repetitive sequences are removed by incubating the resultant product with avidin and subtracting the hybridized repetitive sequences with phenol. Such hybridized repetitive sequences remain in solution after the addition of a salt of a weak acid, for example, sodium acetate, and the remaining source DNA is recovered as a precipitate to serve as a ReSeD probe.

Microdissection and PCR Amplification of Source DNA

Whole chromosome painting, chromosome arm, and band specific probes used in this research were obtained by microdissection [Guan et al., 1996]. Briefly, microdissection was performed with glass microneedles controlled by a micromanipulator attached to an inverted microscope. The target region of a chromosome was cut and transferred to a 20 µl of collecting solution containing pepsin. Five copies of microdissected DNA fragments from the target chromosome region were pooled in 5 µl collecting solution containing pepsin. The dissected DNA was incubated at 37°C for 1 hr and served as a source DNA for the experiments. Two microliters of each source DNA were added to PCR reaction mix (50 µl) which contained 10 mM Tris-HCl, pH 8.4, 2 mM MgCl₂, 50 mM KCl, 200 µM each dNTP (Bioline Inc., Randolph, MA) 2 µM primer and 2 units Taq DNA polymerase (New England Biolabs, Inc., Ipswich, MA). The reaction was heated to 96°C for 2 min, followed by 25 cycles at 94°C for 1 min, 1 min at 56°C, and 1 min at 72°C, with a 5 min final extension at 72°C. The degenerate primer UN1 (5′CGGGAGATCCGACTCGAGNNNNNNA TGTGG-3′) was used to amplify the source DNA.

Preparation of Digoxigenin (Dig)-Labeled Source DNA by PCR

The source DNA at approximate 4 ng/µl was used for the labeling process. In brief, 6 µl of source DNA was added to the PCR reaction mix (100 µl) containing 4 µM UN1 primer, 0.4 mM of each dNTP (Bioline), 4 mM MgSO₄, 10 units of Taq DNA polymerase (Biolabs, Inc.) and 0.04 mM dig-11-dUTP (Roche Applied Science, Indianapolis, IN). The reaction was heated to 96°C for 2 min, followed by 28 cycles at 94°C for 1 min, 1 min at 56°C, and 1 min at 72°C, with a 5 min final extension at 72°C.

Preparation of Biotin-Labeled Human Repetitive Sequences (Driver DNA)

Driver DNA is the DNA that contains representative genomic fragments of repetitive DNA sequences. Normally, two types of driver DNA exist, including (1) human DNA fragments of various repetitive sequences and (2) Cot-1 DNA. Cot-1 DNA was used as the driver DNA in our experiments. Ten micrograms of Cot-1 DNA (Invitrogen Corporation, Carlsbad, CA) was biotin-labeled following the procedures as indicated in the nick translation kit (Invitrogen).

Hybridization of Driver DNA With Source DNA

Two hundred forty nanograms (in 60 µl of labeling product) of dig-labeled source DNA were mixed with 10 µg (in 55 µl of labeling product) of biotin-labeled Cot-1 DNA in 20 µl of hybridization buffer (55% formamide, 10% dextran sulfate, 1× SSC). The mixture was denatured at 100°C for 10 min and hybridized at 55°C overnight.

Subtraction of Repetitive Sequences from Source DNA

Ten microliters of avidin (2.0 mg/ml) (Vector Laboratories, Inc., Burlingame, CA) were added to 135 µl of the source DNA/driver DNA hybridization mixture, which was incubated at 37°C for 20 min. Subsequently, 240 µl of ddH₂O and 300 µl of buffer saturated phenol being added to the hybridization mixture was vortexed for 30 sec, and centrifuged at 14,000 rpm for 5 min. The supernatant was transferred to a clean tube with 300 µl of phenol:chloroform:isoamyl alcohol (25:24:1), vortexed for 30 sec, and centrifuged at 14,000 rpm for 5 min. The supernatant was transferred to a clean tube with 144 µl of chloroform, vortexed for 30 sec and centrifuged at 14,000 rpm for 5 min again. The supernatant was transferred to a clean tube with 1/10 volume of 3 M sodium acetate and three volume 100% ethanol, mixed, and precipitated at −20°C overnight. The tube was centrifuged at 14,000 rpm for 30 min; the supernatant was discarded, and the pellet was air dried and resuspended in 10 µl of ddH₂O. Concentration of the remaining source DNA after the subtraction was measured at about 40 µg/ml in the10 µl ddH₂O. The dig-labeled ReSeD source DNA was used as the probes for our FISH analysis.

FISH Analysis on Metaphase Chromosomes

Hybridization to metaphase chromosomes was carried out as follows: 3 µl of a ReSeD probe or a regular microdissected probe (both at about 40 µg/ml) and 7 µl of hybridization buffer (55% formamide, 10% dextran sulfate, 1× SSC) were mixed in a 0.5 ml tube to make 10 µl of hybridization mixture without Cot-1 DNA added. The hybridization mixture was denatured at 70°C for 5 min. Meanwhile, chromosomal DNA was denatured on slides by immersing the slides in denaturing solution (70% formamide, 2× SSC, pH 7) for 2 min at 70°C, dehydrating the slides through a 70%, 85%, and 100% ethanol series for 1 min, respectively, air-drying the slides, and pre-warming the denatured slides on a 37°C slide warmer before hybridization. Subsequently, for each slide the 10 µl of hybridization mixture was added onto the slide and covered with a 22 × 22 mm² coverslip that was sealed with rubber cement. Following 3 hr of hybridization at 37°C, the slide was washed in the buffer (50% formamide, 2× SSC) at 45°C for three times (5 min each time) and re-washed with 4× SSC/0.05% Tween 20 for three times (2 min each time) followed by washing with 4× SSC for another 2 min at room temperature. Finally, 50 µl of diluted anti-dig-rhodamine (2 µg/ml, Roche) at 1:100 dilution with PNM buffer were added onto the slide and covered with a 22 × 22 mm² coverslip or parafilm. The slide was kept in a moisture chamber in dark for 40 min at room temperature. The slide was then washed with 4× SSC/0.05% Tween 20 for three times (2 min each time) followed by washing with 4× SSC for another 2 min at room temperature. The 4,6-diamino-2-phenyl-indole (DAPI, 1 µg/ml) was used as a counterstain.

RESULTS AND DISCUSSION

Using our method described here, we have produced chromosome probes by extracting repetitive DNA sequences from microdissected whole chromosomes, chromosome arms, locus specific probes, and specific bands.

Demonstration of the Subtraction Efficacy

To assure that repetitive DNA sequences were removed significantly, we tested our subtraction procedures on the source DNA of the long arm of chromosome 15 (15q). Two hundred forty nanograms of microdissected 15q DNA were used for the subtraction. The DNA fragments obtained before and after the subtraction were run on a 1% agarose gel and the results are shown in Figure 1. When microdissected DNA (Fig. 1, lane 1) was used in the subtraction, the added avidin bound to the source DNA/biotin-labeled driver DNA complex, and the majority of repetitive sequences were removed after subtraction (lane 2). These findings demonstrate that our methods are optimal for the subtraction of repetitive sequences from a source DNA.

Figure 1. Subtraction of microdissected chromosome 15q DNA with phenol. When microdissected chromosome 15q DNA (lane 1) was used in the subtraction procedures, the majority of the repetitive DNA sequences was removed after the subtraction (lane 2).

Evaluation of the Quality of Individual ReSeD Probes

FISH was performed as described above on metaphase chromosomes obtained from two normal controls by using dig-labeled ReSeD probes specific for whole chromosome 4, 1q, and 3q26, along with their dig-labeled regular microdissected probes for direct comparison. Pretreatment with Cot-1 DNA in the hybridization procedures was not performed for both the ReSeD and regular probes. Ten metaphase cells were evaluated for each probe per specimen. All photographs were taken through the microscope with no computer capturing or enhancement involved.

In all evaluations, generally the ReSeD probe signals were more visible, uniform, and brighter, with little to no background staining as compared with the signals generated with the regular probes without Cot-1 blocking. Furthermore, the ReSeD probes were more specific to each region and did not show any cross-hybridizations. Described below are results representative of each probe tested.

Application to Whole Chromosome Painting Probes

Figure 2A and 2B shows comparative results of the regular and ReSeD probes for whole chromosome 4. Figure 2A shows hybridization of whole chromosomes 4 with the regular microdissected DNA probe. The signals are not very uniform along the target chromosomes. The same probe is shown in Figure 2B after having its repetitive sequences removed. Two bright signals uniformly cover the length of the chromosomes. Except for auto-florescence, no cross-hybridization is apparent.

Figure 2. FISH results obtained with both the microdissected regular and ReSeD probes without Cot-1 blocking in hybridization, including a regular whole chromosome 4 painting probe (A) and its ReSeD probe (B), a regular chromosome 1q-specific probe (C), and its ReSeD probe (D), and a regular 3q26 band-specific probe (E) and its ReSeD probe (F). In general, the signals of ReSeD probes are more visible, uniform, specific, and brighter, with little to no background staining.

Application to Chromosome Arm-Specific Probes

Figure 2C and 2D shows an example of our method applied to the chromosome-1 long-arm specific probe, Ch 1q. Hybridization with the regular microdissected 1q probe without Cot-1 blocking is depicted in Figure 2C. In that figure one can see two painted long arms of chromosome 1 (short arrows). The two bright spots on the painted arms are the heterochromatic regions of chromosome 1. The signals are not uniform along the chromosome arms. Importantly, cross-hybridization signals are evident among several other chromosome centromeric regions (long arrows). The results of removing repetitive sequences from this probe were dramatic. Shown in Figure 2D is the result of hybridization with the chromosome 1q probe after having its repetitive sequences removed by our method. In that figure two bright uniformly painted chromosome-1 long arms are distinct (arrows). As can be seen, the signals are bright and uniform along the chromosome length, with a clean background and no cross-hybridization to other chromosomes.

Application to Chromosome Band-Specific Probes

The results of applying our method to a band-specific probe is shown in Figure 2E and 2F. The probe is for chromosome band 3q26. Hybridization with the regular 3q26 microdissected probe without Cot-1 blocking is shown in Figure 2E. Although the targeted bands show sufficient intensity as to be distinguished (arrows), there is significant cross-reactivity. The result of removing repetitive DNA sequences from this probe is shown in Figure 2F. By contrast to the regular probe, the ReSeD probe shows much greater intensity in signals and more specificity with no cross-reactivity (arrows).

Application to Flow Sorted Human Chromosomes

We recently demonstrated the utility of our method to generate repetitive DNA sequence depleted probes from flow-sorted chromosomes 1, 2, 3, 4, and 5. FISH analyses using these probes show consistent strong signals with no cross-hybridization. Thus, our method works well on flow sorted chromosomes as well.

In summary, we have demonstrated that repetitive DNA sequences can be removed from a source DNA using the efficient chemical method of phenol subtraction to generate probes that produce signals brighter and more specific as compared to the regular probes when hybridized without Cot-1 blocking. These probes have the advantage to simplify the FISH protocol as no Cot-1 or pre-annealing is necessary during hybridization. Notably, in our experiments all ReSeD probes were able to generate optimal results following 3 hr of hybridization, further demonstrating the efficiency of using these probes for FISH analysis.