DNA affinity capture and protein profiling by SELDI-TOF mass spectrometry: effect of DNA methylation

Abstract

We report here that surface enhanced laser desorption/ionization–time of flight (SELDI-TOF) mass spectrometry, as performed on a Ciphergen Biosystems ProteinChip System, can be used in conjunction with DNA affinity capture (DACA) to study specific DNA–protein binding. Using DNA molecules bound to a surface, sequence-specific interactions can be detected as demonstrated by a mutation affecting the binding profile for TBP, a transcription factor. Also, a comparison between methylated and unmethylated promoter-containing DNA fragments shows numerous binding profile differences over a mass range extending to >60 kDa. The binding of several proteins is inhibited by methylation of the DNA.

INTRODUCTION

The purpose of this study was to demonstrate that surface enhanced laser desorption/ionization–time of flight (SELDI-TOF) mass spectrometry (MS) can be used in conjunction with DNA affinity capture (DACA) to study DNA–protein interactions. SELDI-TOF MS is a specialized form of matrix-assisted laser desorption/ionization–time of flight (MALDI-TOF) MS. In normal MALDI-TOF MS, dilute samples of analyte in a crystalline matrix are deposited on a sample probe and irradiated with a short UV laser pulse. Energy absorbed by the matrix desorbs both matrix and analyte molecules into the gas phase and produces ions primarily through proton transfer. The ions are accelerated into a time of flight analyzer that measures the ratio of mass to charge. MALDI produces primarily singly charged ions and thus a direct determination of the molecular mass (1).

In SELDI-TOF MS, as introduced by Hutchens and Yip (2), the probe surface plays an active role in the extraction, presentation, structural modification and/or amplification of the sample. The sample probe we used, the ProteinChip array from Ciphergen Biosystems, can be derivatized with a variety of biochemical moieties covalently linked to the surface. These include antibodies, receptors, enzymes, DNA, ligands and lectins, as well as other chemical components commonly used for chromatography, i.e. anion exchange, cation exchange, metal affinity and reverse phase (3). The ProteinChip array is derivatized, incubated with the cellular fluid of interest and washed to reduce non-specific interactions. Matrix is added to release the bound proteins for SELDI-TOF analysis (4). Proteins of considerable size, ≥60 kDa, can be detected as molecular ions without fragmentation. SELDI-TOF ‘on-chip’ handling and processing techniques have been employed to profile proteins in several biological systems, including temperature-sensitive proteins of Yersinia pestis (5), a virulence factor knockout in Neisseria meningitidis (6), tumor progression in colon, bladder, head and neck cancers (7–9) and soft tissue regeneration proteins in mice (10). To our knowledge, this is the first published example of SELDI-TOF MS used to study DNA–protein interactions.

A common, extremely useful technique for the study of DNA–protein interactions is the gel shift or electrophoretic mobility shift assay (EMSA), which uses a radioactively labeled, small DNA fragment and measures electrophoretic mobility in the presence and absence of DNA-bound protein(s) (11). Numerous studies, including the use of EMSA, have shown that 5-methylcytosine strongly affects DNA–protein interactions (12) and inhibits transcription when present in promoters (13). Large multiprotein complexes are known to form on transcriptionally active promoters (14) and in vivo footprinting studies have shown that the protein footprints at active promoters are absent on inactive, methylated promoters (15–17). Such studies have been done for the human XIST minimal promoter used in this study (16,18).

We investigated whether SELDI-TOF MS could complement EMSA, providing increased sensitivity, more flexibility in less time and accurate mass information, all without the need for radioactivity. We have combined DACA with SELDI-TOF MS and used the resulting DACA-TOF technique for the study of DNA–protein interactions. DACA-TOF MS detects the sequence-specific binding of TBP, a transcription factor, and the effect of cytosine methylation on proteins binding to the human XIST minimal gene promoter (19). Thus, DACA-TOF MS is a convenient, sensitive and relatively rapid technique that can be used to study specific DNA-binding proteins and protein complexes even using crude nuclear extracts.

MATERIALS AND METHODS

Cell culture

For the preparation of nuclear extracts, HeLa cells were grown in 5% CO₂ at 37°C. The cells were maintained in Dulbecco’s modified Eagles’s medium supplemented with 10% fetal calf serum. Cells were grown to 70% confluence in 150 cm² Petri dishes (2 × 10⁷ cells/dish) prior to harvest.

Nuclear extracts

Extracts were prepared as described by Dignam et al. (20) with some modifications. Briefly, cells were washed twice with phosphate-buffered saline (PBS) and incubated on ice for 20 min in lysis buffer I [10 mM HEPES, pH 7.9, 1.5 mM MgCl₂, 10 mM KCl, 0.3 M sucrose, 0.1 mM EGTA, 0.5 mM dithiothreitol (DTT), 0.5% Nonidet P-40 and one protease inhibitor tablet per 50 ml (Complete Protease Inhibitor Cocktail; Roche Applied Science)]. Cells were scraped off, collected in pre-cooled 50 ml Falcon tubes and spun for 25 min at 3500 r.p.m. and 4°C. The resulting pelleted nuclei were resuspended and gently extracted in 2.5 vol buffer II (20 mM HEPES, pH 7.9, 25% v/v glycerol, 0.42 M NaCl, 1.5 mM MgCl₂, 0.2 mM EDTA, 0.1 mM EGTA, 0.5 mM DTT and one protease inhibitor tablet per 50 ml). The mixture was Dounce homogenized with 10 strokes of the B pestle and centrifuged for 30 min in an Eppendorf centrifuge at 14 000 r.p.m. and 4°C. The supernatant was then ultracentrifuged for 120 min at 45 000 r.p.m. at 4°C. The final supernatant was dialyzed against 50 vol buffer III (20 mM HEPES, pH 7.9, 20% v/v glycerol, 0.2 mM EDTA, 100 mM KCl, 0.5 mM DTT and one protease inhibitor tablet per 50 ml). Nuclear extracts were stored at –80°C in aliquots. The protein concentration was measured by the Bradford method (21). The extracts used for this study had a protein concentration of 5 mg/ml.

Capture molecules

Figure 1 shows the DNA fragments used in these experiments. The following oligonucleotides were used to make the 60 bp, double-stranded DNA fragments, either wild-type or mutant (mut): X-T/Y-S, 5′-biotin-CCCCCCCTTCAGTTCTT`AAAGCGCTGCAATTCGCTGCTGCAGCCATATTTCTTACTCTCT-3′; X-T/Y-L, 5′-AGAGAGTAAGAAATATGGCTGCAGCAGCGAATTGCAGCGCTTTAAGAACTGAAGGGGGGG-3′; X-T/Y-MutS, 5′-biotin-CCCCCCCTTCAAAGTGAGGTGCGCTGCAATTCGCTGCTGCAGCATCTGCGCTTACTCTCT-3′; X-T/Y-MutL, 5′-AGAGAGTAAGCGCAGATGCTGCAGCAGCGAATTGCAGCGCACCTCACTTTGAAGGGGGGG-3′.

Figure 1.

DNA capture molecules. The sequence shown is the 180 bp capture DNA used in these experiments. Brackets enclose the XIST minimal promoter sequence. Parentheses enclose the 60 bp fragment used in the TBP experiments. hX-S1 and hX-L1 denote the PCR primers used to produce the full capture molecule. Sites of DNA methylation, which are CpG dinucleotides, are shown in bold. Sites of known DNA-binding proteins are boxed in gray. For the TBP factor, the mutant sequence is denoted below the wild-type sequence.

A 180 bp PCR fragment containing the human XIST minimal promoter (19) was amplified from genomic DNA of X8-GT2 DNA, a Chinese hamster hybrid cell containing an inactive human X chromosome (22). The primer set for the XIST minimal promoter was hX-S1 (5′-TGAACCAACCAAATCACAAAGATGTC-3′) and hX-L1 (5′-CCCAATGCAGAGAGATCTTACAGTC-3′).

One-half of the purified fragment (2–3 µg) was methylated at every CpG by treatment with 20 U SssI methylase (New England Biolabs) in the manufacturer’s Buffer II (150 mM NaCl, 10 mM Tris–HCl, 10 mM MgCl₂, 1 mM DTT, pH 7.9, supplemented with 160 µM S-adenosylmethionine) for 1 h at 37°C. The completeness of the reaction was tested by the use of the methylation-sensitive restriction endonuclease HpaII in NEB Buffer I (10 mM bis,Tris-propane–HCl, pH 7.0, 10 mM MgCl₂, 1 mM DTT) for 1 h at 37°C. No detectable cleavage was observed for the fragments used in this study.

Affinity capture and washing of DNA-binding proteins

DNA affinity capture was carried out as shown in Figure 2 using a Ciphergen Biosystems ProteinChip System (3,4). Briefly, 5.0 µl of monomeric streptavidin (0.1 mg/ml) in PBS was added to each spot on a PS1 series ProteinChip array containing carbonyl diimidazole (Ciphergen Biosystems), which interacts with primary amines to form stable crosslinks. The solution was mixed and incubated at 4°C for 1 h. Excess streptavidin was removed and 15 µl of the blocking agent, 1 M ethanolamine (pH 9.0), was added, mixed and evaporated for 30 min. The chip was washed three times in PBS containing 1% Triton X-100, once in PBS and copiously in water. Excess water on the chip was removed by wicking, but the spots were kept moist.

Figure 2.

DACA-TOF MS. A 5′-biotinylated, double-stranded DNA molecule of interest is bound to a streptavidin-coated surface to produce a ProteinChip array. DNA affinity chromatography (sample preparation, binding and washing) is carried out ‘on-chip’. Analysis of bound proteins is by SELDI-TOF MS.

Aliquots of 5 pmol of each 5′-biotinylated, double-stranded capture DNA were added to the streptavidin-coated spots and incubated for 1 h at room temperature and high humidity. Excess solution was drawn off. The ProteinChip array was washed twice in PBS containing 1% Triton X-100, once in PBS, copiously with water and equilibrated in 1× protein binding buffer (10 mM Tris–HCl pH 7.4, 0.5 mM EDTA pH 8.0, 50 mM NaCl, 0.5 mM DTT, 2.5 mM MgCl₂, 0.25% Triton X-100) (23). Excess binding buffer was removed by wicking, but the spots were kept moist.

An aliquot of 1–2 µl of HeLa nuclear extract (5 mg/ml) was added to each spot in 1× protein binding buffer supplemented with 0.25 µg/µl poly(dI·dC) and the array was incubated for 1 h at room temperature and high humidity. The excess binding reaction mixture was drawn off and the ProteinChip washed (see below and figure legends) and kept moist prior to the addition of two 0.5 µl aliquots of the MALDI matrix, a saturated solution of sinapinic acid in 50% acetonitrile and 0.5% trifluoroacetic acid.

An amount of 4–5 pmol of capture DNA was sufficient for reproducible binding; additional capture DNA made no apparent difference, in agreement with Ciphergen Biosystems specifications as to array binding capacity. The optimum concentration of nuclear extract was found to be between 5 and 10 µg protein/reaction. The array surface alone has some non-specific binding activity and overloading resulted in a general broadening of peaks.

Preliminary experiments established loading and wash conditions, treating each spot on the ProteinChip array differently. Samples were washed for varying times in PBS containing 0.1% Triton X-100 and once in PBS prior to addition of the MALDI matrix. Shortening the wash time while increasing the number and stringency of washes improved spectra clarity and decreased non-specific binding. For example, extra washes led to elimination of the non-specific peak at 24.7 kDa seen in Figure 5. For a given wash condition the spectra such as shown in Figure 5 are reproducible for laser intensities between 120 and 150 µJ.

Figure 5.

Effect of DNA methylation on DNA-binding proteins. DACA-TOF MS spectra of HeLa nuclear extract proteins using the 180 bp XIST promoter capture DNA. (A) Non-methylated DNA. (B) DNA methylated at each CpG (5′mCpG-XIST). Insets show the expanded mass ranges 10 000–15 000 and 30 000–60 000 m/z. An asterisk indicates a protein peak of both absolute and relative intensity at least 3-fold greater for the unmethylated XIST promoter capture DNA. The ProteinChip array–binding protein complexes were washed once in PBS containing 0.1% Triton X-100 and once in PBS for 30 min each. The laser setting was 150 µJ.

SELDI-TOF mass spectrometry

Analysis of the ProteinChip array was carried out in a PBS-II ProteinChip Array Reader (Ciphergen Biosystems) according to an automated data collection protocol. The apparatus was operated in positive ion mode with source and detector voltages of 20 and 1.8 kV, respectively. Time lag focusing was used with a pulse voltage and pulse lag time of 3000 V and 673 ns, respectively. In various experiments laser intensity varied between 120 and 150 µJ from a nitrogen laser emitting at 337 nm. The digitizer operated at 250 MHz. The average of 65 collected laser shots is presented for each spectrum. Data interpretation was aided by use of the ProteinChip software v.2.1b.

RESULTS AND DISCUSSION

In vitro and in vivo studies have established that XIST is a methylation silenceable promoter that binds the TATA-binding protein (TBP) (16–18). This promoter was used to investigate binding of HeLa nuclear extract proteins to DNA capture molecules. When a 60 bp fragment that included a portion of the XIST minimal promoter (Fig. 1) was used as the capture DNA, a protein of appropriate size for TBP (36.3 kDa) (24) bound preferentially to the wild-type sequence (Fig. 3). Binding was greatly reduced when the capture fragment contained a mutated TBP-binding site. A protein at 32.7 kDa was observed as well. The intensity of this peak was also reduced by the mutation, consistent with it being a component of the TBP-associated complex.

Figure 3.

Effect of mutation on protein binding. Mass spectra of HeLa nuclear extract proteins in the mass range 32 000–37 000 m/z obtained using the 60 bp, truncated XIST capture DNA. The upper trace was obtained with the DNA containing the consensus binding site for TBP and the lower trace was obtained using the mutated DNA (see Fig. 1). The peak consistent with singly charged TBP is indicated. The ProteinChip array–binding protein complexes were washed once for 5 min in 1× protein binding buffer minus DTT, six times for 5 min each in PBS containing 1% Triton X-100, 100 mM NaCl and copiously with water for 5 min. The laser setting was 150 µJ.

HeLa nuclear extracts were also incubated with 180 bp capture fragments containing the entire human XIST minimal promoter sequence (Fig. 1) and a comparison was made with a fragment methylated at all CpG residues. The 180 bp unmethylated XIST capture DNA binds a 36.3 kDa protein as well as several others (Fig. 4). Although positive protein identification cannot be made based solely on the molecular mass information in DACA-TOF MS, several of the peaks observed are consistent with human TBP-associated factors (TAFs) comprising components of the RNA polymerase II–TFIID transcriptional complex that is known to associate with promoters. For example, the ions corresponding to 28.5 and 31.3 kDa are consistent with known TAFs, histone-like hTAF_II28 and hTAF_II31 (25). Also, a protein subunit possibly corresponding to the basal transcription initiation machinery, TFIIF (30 kDa), was observed (26). It is also clear that methylation of CpG sites in the capture DNA inhibits the binding of nearly all of the proteins in the 28–36 kDa mass range (Fig. 4). Conclusive identification of the bound proteins will require additional studies, but it is clear that DACA-TOF MS provides a sensitive, convenient assay to guide purification or the study of specific DNA–protein and protein–protein complexes.

Figure 4.

Effect of DNA methylation on TBP binding. Mass spectra of HeLa nuclear extract proteins in the mass range 28 000–37 000 m/z using the 180 bp XIST promoter capture DNA with (lower trace) and without (upper trace) 5-methylcytosine at all CpG residues (see Fig. 1). Protein peaks that are consistently observed using the unmethylated DNA but not observed with the methylated DNA are indicated with an asterisk. The ProteinChip array–binding protein complexes were washed as in Figure 3. The laser setting was 150 µJ.

A comparison of binding over the entire mass range of observable proteins is shown in Figure 5. Wash conditions for this experiment were purposefully less stringent in order to preserve weaker interactions. Overall the profiles for methylated and unmethylated DNA are remarkably similar and clearly many proteins bind similarly to both capture DNAs. However, several proteins show absolute and relative intensities that are much greater for the unmethylated XIST promoter DNA. This result is, of course, consistent with the assembly of multiprotein complexes at unmethylated promoters (14). Ions corresponding to proteins that show clear preferential binding to the unmethylated DNA are marked with an asterix (Fig. 5, insets). No proteins were detected that bind with clear preference to methylated DNA.

CONCLUSIONS

DACA-TOF MS, the combination of DACA and SELDI-TOF MS, can be used to study DNA–protein interactions, showing sequence-specific binding differences and differences due to modification of the DNA. The ProteinChip array format has several advantages over EMSA (11), including increased speed, convenience, flexibility and the amount of information generated per experiment. A ProteinChip array contains eight spots and each spot can be, in effect, a separate experiment. Thus, binding conditions, substrate competition and mutation effects can be queried on a single chip without the need for a radioisotope. The ProteinChip array can provide accurate protein mass information for DNA–protein and protein–protein interactions as well, thus providing initial insight as to the proteins involved. Analysis software allows direct comparison of results, spot to spot and array to array. Acquisition of the mass spectrometry data takes <30 min and turnaround time between ProteinChip arrays is only a few minutes. In comparison to standard MALDI-TOF MS techniques, SELDI-TOF MS represents some important enhancements. The ‘on-chip’ affinity capture and reduction of excess salt and non-specific binding contaminants through on-probe washing improves analyte signal intensity and resolution, as well as eliminating ‘offline’ handling losses typical of many other MALDI-TOF MS sample preparation techniques.

In summary, SELDI-TOF MS works well in conjunction with DNA affinity capture to provide DNA-binding profiles which show sequence-specific binding differences, binding of proteins to the human XIST promoter and differences due to cytosine methylation at CpG sites in the promoter.