Terminal transferase-dependent PCR: A versatile and sensitive method for in vivo footprinting and detection of DNA adducts

 1998 Oxford University Press Nucleic Acids Research, 1998, Vol. 26, No. 7 1807–1811 Terminal transferase-dependent PCR: a versatile and sensitive method for in vivo footprinting and detection of DNA adducts Jun-ichiro Komura and Arthur D. Riggs* Biology Department, Beckman Research Institute of the City of Hope, Duarte, CA 91010, USA Received October 30, 1997; Revised and Accepted February 11, 1998 ABSTRACT INTRODUCTION Only a few methods are useful for quantitatively displaying DNA lesions and chromatin structure in mammalian cells at singlenucleotide resolution (1,2). One such method is ligation-mediated PCR (LMPCR) (3–5), which is commonly used because it combines nucleotide-level resolution with the sensitivity of PCR. LMPCR has been used successfully in this laboratory and others for numerous in vivo studies of mammalian cells, especially for the detection of protein–DNA interactions (footprints) (1), for the analysis of cytosine methylation (3) and for the mapping of DNA damage (6,7). In the conventional method of LMPCR (Fig. 1, left), substrate genomic DNA is cut at the sites of altered bases either with a specific enzyme or by use of chemical reagents (e.g. Maxam–Gilbert cleavage). The resulting single-strand breaks are converted to blunt-ended termini by extension from a gene-specific primer and are ligated to a double-stranded linker. The sequences between the linker and a second (nested) gene-specific primer are amplified by PCR and the products are visualized as radioactive sequence ladders. LMPCR is very sensitive but has some limitations. First, it measures directly only nicks or breaks in DNA. Secondly, the 5′ end MATERIALS AND METHODS Cells and DNA preparation Culture of BML-2 cells (9), DNA isolation, UV irradiation of cells or DNA and separation of the expressed Xist allele from the silent allele were performed as described by Komura et al. (10). Cleavage of UV-irradiated DNA at the sites of cyclobutane pyrimidine dimers using photolyase and T4 endonuclease V, or at the sites of pyrimidine–pyrimidone (6–4) photoproducts by treatment with piperidine, was carried out as described (11,12). Maxam–Gilbert cleavage of DNA was done as described (13), using 80 µg of genomic DNA. *To whom correspondence should be addressed. Tel: +1 626 301 8352; Fax: +1 626 358 7703; Email: We report here a new, sensitive and versatile genomic sequencing method, which can be used for in vivo footprinting and studies of DNA adducts. Starting with mammalian genomic DNA, single-stranded products are made by repeated primer extension; these products are subjected to homopolymeric ribonucleotide tailing at the 3′ termini with terminal deoxynucleotidyl transferase and then ligated to a double-stranded linker having a complementary 3′ overhang, and used for PCR. This terminal transferase-dependent PCR (TDPCR) method can generate band signals many-fold stronger than conventional ligation-mediated PCR (LMPCR). A UV photofootprint in the mouse Xist gene promoter can be easily detected using TDPCR. No special enzymes or chemical reagents are needed to convert DNA adducts into strand breaks. Any lesion that blocks primer extension should be detectable. of the template molecule must be phosphorylated (or phosphorylatable) and ligatable, because it must undergo blunt-end ligation after primer extension. Thirdly, only molecules in which primer extension has continued to the end of the template strand are able to participate in blunt-end ligation. Thus, prematurely terminated molecules are invisible to LMPCR. Finally, several hundred bands usually are seen in every LMPCR ladder, and each band must start from at least one genomic template molecule. Therefore, in practice, thousands of molecules of genomic DNA are needed to avoid poor quantitation and missing bands due only to statistical sampling fluctuations. The Terminal transferase-Dependent PCR (TDPCR) method reported here (Fig. 1, right) depends on cohesive-end ligation to the 3′ ends of DNA molecules resulting from primer extension, followed by controlled ribonucleotide tailing by terminal deoxynucleotidyl transferase (8). TDPCR provides an alternative to LMPCR that should eliminate or lessen each of the limitations described above. TDPCR does not require ligatable 5′ ends, detects prematurely terminated molecules, can sample each template molecule many times and thus can be more sensitive than LMPCR. The method should aid footprint and chromatin structure experiments that need high sensitivity and should enable use of new footprinting or DNA-damaging agents whose products on DNA cannot easily be converted to ligatable 5′ termini. A photofootprint in the mouse Xist gene promoter is shown to be clearly detected by TDPCR. 1808 Nucleic Acids Research, 1998, Vol. 26, No. 7 Primers and linkers The gene-specific primers A1, A2 and A3 for the mouse Xist promoter region are the same as previously described (10). Linker α and the linker-primer (upper strand of α; Fig. 2) are the standard oligonucleotides usually used for LMPCR (5). Linkers β, γ, δ and ε are related to α as shown in Figure 2. Linkers γ, δ and ε were synthesized by the DNA synthesis shared resource facility of the City of Hope with an aminopentyl blocking group at their 3′ termini (14). For TDPCR the 5′ terminus of the lower oligonucleotide of each linker was phosphorylated by incubation in a 100 µl reaction consisting of 22.2 µM oligonucleotide, 50 U T4 polynucleotide kinase (New England Biolabs), 1
buffer supplied by the manufacturer and 1 mM ATP, at 37
C for 2 h. After inactivation of the enzyme by incubation at 65
C for 20 min, 11 µl of a 200 µM solution of the upper oligonucleotide was added. The mixture was heated to 95
C and allowed to cool gradually. Terminal deoxynucleotidyl transferase-dependent PCR (TDPCR) During the set up of reactions and between heating steps, manipulations were performed on ice unless stated otherwise. Genomic DNA (0.3–0.5 µg) was linearly amplified in a 30 µl reaction consisting of 1.2 U of Vent (exo–) DNA polymerase (New England Biolabs), 1
ThermoPol Buffer (New England Biolabs), extra 4 mM MgSO4, 3.3 mM Tris (pH 7.5), 0.3 mM EDTA, 250 µM each dNTP and 20 nM primer A1. Temperature cycles, which varied from 1 to 30 in frequency, were for 1 min at 95
C (5 min at 95
C for the first cycle), 3 min at 47
C, and 2 min Figure 2. Linkers used in this study. Linker α was used for blunt-end ligation in conventional LMPCR. Linkers β, γ, δ and ε were used for cohesive-end ligation in TDPCR; they are identical in sequence but have different 3′-blocking patterns. A solid circle indicates the presence of a blocking amine. at 72
C. After thermal cycling, each sample was incubated at 95
C for 2 min, then 20 µl was transferred into another tube containing 80 µl of a solution composed of 2.5 M ammonium acetate, 2.5 mM EDTA and 40 µg glycogen, and precipitated with 250 µl of ethanol. The precipitate was dissolved in 10 µl of 1/10 TE (1 mM Tris pH (...truncated)