|Home | About | Journals | Submit | Contact Us | Français|
We have adapted the techniques of DNA footprint analysis to an Applied Biosystems 3730 DNA Analyzer. The use of fluorescently labeled primers eliminates the need for radioactively labeled nucleotides, as well as slab gel electrophoresis, and takes advantage of commonly available automated fluorescent capillary electrophoresis instruments. With fluorescently labeled primers and dideoxynucleotide DNA sequencing, we have shown that the terminal base of each digested fragment may be accurately identified with a capillary-based instrument. Polymerase chain reaction (PCR) was performed with a 6FAM-labeled primer to amplify a typical target promoter region. This PCR product was then incubated with a transcriptional activator protein, or bovine serum albumin as a control, and then partially digested with DNase I. A clone of the promoter was sequenced with the Thermo Sequenase Dye Primer Manual Cycle Sequencing kit (USB) and the FAM-labeled primer. Through the use of Genemapper software, the Thermo sequenase and DNasei digestion products were accurately aligned, providing a ready means to assign correct nucleotides to each peak from the DNA footprint. This method was used to characterize the binding of two different transcriptional activator proteins to their respective promoter regions.
DNA footprinting was developed in 1977 to elucidate nucleotide bases that contact DNA-binding proteins in a specific and competitive manner.1 This technique employs gel electrophoresis to identify region(s) from labeled DNA fragments that become protected from enzymatic digestion by virtue of the protein binding to specific sequences. The resulting electropherogram appears to have a “footprint” that shows the disappearance of bands where the protein was bound. During DNA footprint analysis:2,3 (1) the DNA is first labeled with radioactive phosphate, (2) the protein is then added, (3) the DNA is fragmented (by enzymatic or chemical means), (4) the fragments are separated by slab-gel poly-acrylamide-urea electrophoresis, and (5) the resulting radiogram is analyzed manually.
The goal of this study was to utilize non-radioactive means to determine specific binding sites, at base-pair resolution, of transcription regulator proteins to specific promoter sequences. In the methods developed here, we have substituted the radioactive labels with a fluorescent dye (FAM) that is easily detected with an automated fluorescent DNA analysis instrument (Applied Biosystems 3730 DNA Analyzer, Foster City, CA) that resolves oligonucleotides by capillary electrophoresis. This instrument also eliminates the need for polyacrylamide gel electrophoresis. Previous studies had utilized fluorescent dyes,4–6 and a capillary-based automated DNA sequencer was shown to be sensitive enough to identify protected sites and hypersensitive regions. However, identification of the terminal base of each fragment was determined by comparison to a BigDye Terminator DNA sequencing reaction.6 This is not accurate due to the significantly different migration rates of BigDyes compared to FAM fluorescent labels, and the instrumentation and/or software are unable to compensate for these differences. Thus, in this study we have overcome the inherent inaccuracy of previous work using capillary-based electrophoresis instrumentation. In addition, in order to determine the exact bases to which the protein is bound, the target DNA was sequenced with dideoxynucleotide-based reactions7 utilizing a FAM-labeled primer8 followed by analysis with the 3730 automatic DNA analyzer. The resulting electropherograms, obtained from sequencing reactions of digested DNA, were then aligned using GeneMapper software to determine the DNA footprint pattern as well as identify the 5′ nucleotide of each digested fragment.
This modification of a previously described protocol was used to study two different bacterial transcription activators. The first protein, CbbR, binds to the promoter and activates expression of the cbbI operon of Rhodobacter sphaeroides.9 The cbbI operon contains genes that encode enzymes that are necessary for carbon dioxide fixation in this photosynthetic bacterium. Previously, the binding site for CbbR had been determined by DNaseI footprint analysis utilizing a slab-gel system and a radioactively labeled probe.9 The DNaseI protection pattern determined in this study via the modified automated procedure compared favorably with the pattern previously reported.9 The second protein employed is HrpY, which binds to the hrpS promoter from Pantoea stewarti subsp. stewartii.10 The Hrp (hypersensitivity response and pathogenicity) proteins are necessary for plant pathogenicity of Pantoea stewarti subsp. stewartii.11 HrpY is part of a two-component system that activates the hrpS gene and subsequently initiates a regulatory cascade that results in Hrp protein production. The HrpY protein was found to have two distinct binding regions within the promoter region of the hrpS gene.
The plasmid pJG336, which contains the cbbI operon, was used as a template to generate a 293-bp probe that encompasses bases −216 to +48 of cbbFI. The probe was generated by polymerase chain reaction (PCR) with the primers Fpcbb01-FAM (5′-(6-FAM)-ACGCC-GAAGGCTTCCTCCAAG-3′) and Fpcbb02-HEX (5′-(HEX)-GTCCTGCAACTCGGCCGGTAT-3′) from Operon Biotechnologies, Inc. The PCR was performed for 30 cycles under the following conditions: 94°C for 60 sec, 50°C for 60 sec, and 72°C for 60 sec. Varying amounts of CbbR protein ranging from 0 to 20 μg were incubated with 1.83 μg of Poly(dI-dC) for 10 min at room temperature in binding buffer (30 mM potassium glutamate, 1 mM dithiothreitol (DTT), 5 mM magnesium acetate, 2 mM CaCl2, 0.125 mg/mL bovine serum albumin (BSA), 30% glycerol in 10 mM Tris HCl, pH 8.5). After this, 500 ng of fluorescently labeled probe was added to the reaction mixture to a final volume of 50 μL and incubated for 20 min at room temperature. Following several DNase I digestion optimization experiments, 0.2 μg of DNase I (Worthington Biochemicals, Lakewood, NJ) was added to the reaction and incubated for 5 min at room temperature. The reaction was stopped by incubating at 75°C for 10 min. Control digestions with the probe were performed in the absence of protein. The DNA fragments were purified with the QIAquick PCR Purification kit (Qiagen, Valencia, CA) and eluted in 40 μL H2O to eliminate salts that can interfere with capillary electrophoresis. Digested DNA, 5.0 μL, was added to 4.9 μL HiDi formamide (Applied Biosystems, Foster City, CA) and 0.1 μL GeneScan-500 LIZ size standards (Applied Biosystems). The samples were analyzed with the 3730 DNA Analyzer, G5 dye set, running an altered default genotyping module that increased the injection time to 30 sec and the injection voltage to 3 kV.
Plasmid pMM5811, which contains the genes hrpL, hrpXY, and hrpS, was used as a template to generate target fragment A, a 330-bp fragment that encompasses bases −231 to +95 of the promoter region from hrpS11. Fragment A was generated by PCR with the primers SF3506-FAM (5′-(6-FAM)-GATTGCTCTTAATTTA-CAAAT-3′) and SR3835 (5′-GCATAAGAAATACCAT-GTCA-3′) from IDT-DNA, Inc. (Coralville, IA). PCR was performed over 30 cycles at the following conditions: 95°C for 60 sec, 50°C for 60 sec, 72°C for 60 sec. Labeled fragment A, 45 ng, was incubated with varying amounts of His6-HrpY protein ranging from 0 to 40 μM in binding buffer (150 mM KCl, 5 mM MgCl2, 0.1 mM EDTA, 1 mM DTT, 8% glycerol in 10 mM Tris HCl, pH 8.0). After several optimization experiments, the nuclease digestion was found to work best with 0.0025 Kunitz units of DNase I (Worthington Biochemicals, Lakewood, NJ) per 20-μL reaction for 5 min at 26°C. The reaction was stopped with 0.25 M EDTA and extracted with phenol-chloroformisoamyl alcohol (25:24:1). Control digestions with Fragment A were done with 20 μg of BSA instead of His6-HrpY. The DNA fragments were purified with the QIAquick PCR Purification kit (Qiagen, Valencia, CA) and eluted in 50 μL Tris buffer to eliminate salts that can interfere with capillary electrophoresis. Digested DNA, 5.0 μL, was added to 4.9 μL HiDi formamide (Applied Biosystems) and 0.1 μL GeneScan-500 LIZ size standards (Applied Biosystems). The samples were analyzed with the 3730 DNA Analyzer, G5 dye set, running an altered default genotyping module that increased the injection time to 30 sec and the injection voltage to 3 kV.
Plasmid pMM58 was sequenced with the primer SF3506-FAM and the Thermo Sequenase Dye Primer Manual Cycle Sequencing Kit (USB, Inc., Cleveland, OH) according to the manufacturer’s instructions, with the exception of the maximum amount of primer (2 pmol), template (400 ng), and number of cycles (60) used. Each reaction was diluted fivefold in water, and 1 μL was added to 8.9 μL HiDi and 0.1 μL GeneScan-500 LIZ size standards. Plasmid pJG336, which contains the cbbI operon, was sequenced the same as above with the labeled primers Fpcbb01-FAM and Fpcbb02-HEX. The positive control reactions with pUC19 and the TAMRA-40 forward primer from the Thermo Sequenase kit were performed as described in the manual. The samples were analyzed on the 3730 DNA Analyzer with the default genotyping module and the G5 dye set. All DNA patterns were analyzed with GeneMapper version 3.5 or 3.7. The electropherograms were horizontally aligned in the sample plot window through the use of the size standards, and the DNA sequence was read left to right (5′ to 3′) by observing the next peak from the trace of each nucleotide.
Plasmid pMM58 was sequenced with primer SR3506 (unlabeled) with the BigDye Terminator v3.1 Cycle Sequencing Kit (Applied Biosystems) according to the manufacturer’s protocol, with the exception that all volumes were reduced by one half. Immediately after cleanup with a Performa DTR V3 Plate (EdgeBioSystems, Gaithersburg, MD), the entire sequencing reaction was loaded onto the 3730 DNA Analyzer. In addition, the two PCR products used as probes were sequenced with unlabeled primers prior to DNaseI digestion the same as immediately above, to confirm that the PCR amplified the correct sequence.
The autoradiograms were imaged with a BioRad (Hercules, CA) VersaDoc 1000 Imaging System. The densitometry traces of the autoradiograms were generated with BioRad (Hercules, CA) PDQuest v.7.3 software.
To confirm that the peak patterns seen in the electropherogram are similar to patterns seen with traditional methods, i.e., radioactivity and gels, a previously published DNase I footprint analysis was repeated using fluorescent dyes and capillary electrophoresis (Figure 11).). The resulting DNA fragment pattern after digestion in the presence of the CbbR protein (Figure 1B1B)) was very similar to the densitometry trace of the autoradiogram. There are two protected regions similarly arranged, with several unprotected bases separating them. From the A+G cleavage standard in the autoradiogram and the corresponding DNA sequence (data not shown) for the electropherogram, the protected regions were determined to be at identical locations. The negative control reactions (Figure 1C and DD),), in which CbbR was replaced with BSA, show uniform peaks across the same fragment of DNA and therefore contain no protected regions.
We took advantage of the multiplex capability of the 3730 DNA Analyzer and the ability of the Genemapper software to superimpose electropherograms in order to do a more thorough comparison of the protection pattern generated by cbbR. The protected regions of both strands of DNA can be determined simultaneously from the same genotyping file (.fsa) through the use of two different fluorescent dyes, such as FAM and HEX, when each is attached to a different primer. The protected region of the sense strand was observed (Figure 2A2A),), with the black trace belonging to the protected sample and the red trace depicting the unprotected sample. The protected region was determined to be 58 bases in length, from base −69 to −11 (Figure 2B2B),), with three bases in the middle accessible to DNaseI whether CbbR was present or not, e.g., bases −38 and −39. For the anti-sense strand, a protected region was observed from base −16 to −71, with several obvious hypersensitive locations: bases −42, −43, −53, −54, and −71 (Figure 2B2B).). Hypersensitive sites are commonly seen in DNaseI footprint analysis. The two protected regions are adjacent but not overlapping, as expected with a LysR- type protein such as CbbR.9
In order to accurately identify the 5′ base of each peak in the electropherogram from the digested DNA fragment, four Thermo Sequenase Dye Primer Sequencing reactions were analyzed (Figures 3A and 3B3B).). These yielded the predicted sequence for the hrpS promoter.7 The cbbI operon promoter was also sequenced the same as the hrpS promoter, and the resulting data (not shown) were used to identify the peaks in Figure 22.. In order to provide additional confirmation that the correct sequence could be obtained with this approach, positive control reactions using the Thermo Sequenase kit were performed. The expected sequence for pUC19 (data not shown) was generated, which confirmed that DNA could be sequenced by analysis with Genemapper. In addition, pMM58 was sequenced and analyzed with a conventional kit, BigDye Terminator, to confirm the sequence determined from the Thermo Sequenase electropherograms (Figure 3C3C).
The DNA digestions were aligned (Figure 44)) through the use of the Genescan 500-LIZ size standard. The alignment is very accurate, with an R2 value of 0.98 or higher for each of the size standard curves (data not shown). The digested DNA patterns, from fragment size 40 to 330 bp, were very similar when comparing HrpY to BSA-treated fragments (Figure 44),), except for the region between 90 and 220 bp, which is identified as the region of interest (ROI). As expected, the BSA treatment did not produce any protected regions, i.e., a regular distribution of fragments 35 to 330 bases in length.
The ROI in Figure 44 was expanded (Figures 55 and 66)) in addition to being aligned with the sequencing reactions. GeneMapper v 3.7 has the ability to assign different colors and superimpose the traces, which significantly aids the visual analysis of the protected and hypersensitive regions (Figures 55 and 66).). The blue traces illustrate the HrpY treatment, whereas the red traces are from the BSA control. The binding region of Figure 55 corresponds to bases −107 to −84 in relation to the transcription start site of the hrpS gene.10 The protected bases in region I are 5′-TAGTCAG-GAAATCCTTACAATCCT-3′. There is a hypersensitive site immediately 5′ of the protected region at bases CAC. In Figure 66,, the binding region corresponds to bases −79 to −63, with hypersensitive sites at the bases −49 and −41 to 39. The protected bases in region II are 5′-AATTCCT-TACCCGATA-3′. The protected regions (Figures 55 and 66)) resemble other transcription activator binding sites3 that were determined by the traditional method (radioactive DNA and a slab gel electrophoresis) in that there are two distinct protection sites of about the same size and same location in regards to the −35 region of the bacterial promoter. The location of the HrpY binding sites agrees with the results of genetic experiments, deletion analysis, and gel-shift analysis as reported by Merighi et al.10
Our work represents a continuing evolution of DNA footprinting techniques that parallels the evolution of DNA sequencing instrumentation. DNA footprinting analysis protocols were introduced in 19782 following the introduction of dideoxynucleotide sequencing in 1977.7 An automated DNA sequencer that utilized fluorescent dye labels, and a slab gel system, were both used in 19944 to illustrate regions of DNA protected from nuclease digestion well after the introduction of fluorescent DNA sequencing in 1986.8 In 2000, a capillary electrophoresis instrument, Applied Biosystems 310 DNA Analyzer, was used to identify protected and hypersensitive regions of DNA.5 The 310 DNA Analyzer, introduced in 1995, was the first commercial capillary-based DNA analysis instrument. The use of the 310 DNA Analyzer and ROX size standards6 could align the traces precisely, similar to Figure 22,, but an accurate means to identify the last nucleotide of each fragment was not apparent from this earlier work. The dyes from the BigDye Terminator Cycle Sequencing Kit cause DNA to migrate significantly differently from digested FAM-attached DNA fragments. When comparing BigDye- to FAM-labeled DNA, the 310 DNA Analyzer cannot compensate for the different migration rates; therefore, the precise location of the DNA binding region could only be estimated based on the size of the fragments. In this investigation, we have developed a simple and reliable means to identify the base of each peak in the DNA digestion pattern via the use of a dye primer–based sequencing kit (Figure 55).). The accuracy problem is avoided by using the FAM-labeled primer in the DNA sequencing reaction, and subsequently determining the sequence after alignment of the four reactions (one for each nucleotide) with GeneMapper software, which is designed for genotyping applications. This approach has many advantages over traditional DNA footprinting techniques—in particular, the use of (i) automated capillary electrophoresis instrumentation and (ii) fluorescent labeling of the DNA.
First, capillary electrophoresis-based instruments, such as the Applied Biosystems 310, 3100, and 3730 DNA Analyzers, are now very common and can be found at most core facilities at universities and other laboratories. The software that is utilized for analysis, Genotyper or GeneMapper, typically is provided with the automated DNA-sequencing instrument, so it is as readily available. Moreover, for virtually all automated capillary electrophoresis instruments, the run-time and subsequent analysis can be completed in 3 to 4 h, as compared to 1 to 3 d for slab gels and subsequent autoradiography. Finally, if the signal is too low or too high, then the capillary electrophoresis conditions can be adjusted and the digestion products analyzed again the same day, whereas slab gels would require an additional 1–3 d if the resulting auto-radiogram is not found to be adequate.
Second, fluorescent dyes are more stable than 32P, i.e., years as compared to a few weeks. Therefore, the primers or PCR products do not need to be labeled immediately prior to use for every DNA footprint experiment. Moreover, fluorescent dye–DNA conjugates are conveniently available, since they can be readily obtained from a variety of manufacturers, and their purchase does not require a license. Additionally, fluorescent dyes are significantly easier to handle. For example, the dyes do not require the use of shielding or a separate area/room just for the use of radioactivity. Finally, the health risk of these dyes is much lower than that of radioactive phosphate, and there are no federal government agencies or institutional regulatory offices and radioactivity license issues to be concerned about. Fluorescent dyes do, however, have the disadvantage of being less sensitive than radioactive labels in general, but they are sufficiently sensitive for this technique without using extraordinary methods, as demonstrated (Figure 44).
The reagents employed for the DNA footprinting protocols presented here cost about the same as those used in the traditional method; however, considerable savings with regard to time and labor represents a real advantage of the new system. All of the biochemical reactions and methods, e.g., PCR, DNA digestion, etc., are the same except for the labeling. The radioactive nucleotide, 32P-αATP, costs about $40.00 for 0.25 mCi (MP Biomedicals, Irvine, CA), which is sufficient for many PCR reactions, as compared to $70.00 to bind a dye, such as FAM (Operon), to an oligonucleotide. However depending upon how many times the radioactive nucleotide needs to be purchased to complete the experiment(s), the fluorescent label may cost much less than radioactive labels when all is said and done. The cost of a manual sequencing gel and buffer, about $2.00 (BioRad, Hercules, CA), is less than the cost of POP7 and buffer, about $4.00 (Applied Biosystems). Although this cost is twice as much for the automated DNA analyzer, the difference is not significant when compared to the cost of the biochemical reagents and enzymes in their entirety. Certainly, at every step there is significant savings in labor. Radioactivity usage also entails regular monitoring, casting of gels, and scrupulous cleaning and removal of label from plates, not to mention subsequent drying of the gel, and imaging. All of these steps add considerable time and expense to a typical DNase protection experiment.
In conclusion, we have demonstrated that fluorescent-labeled automated capillary electrophoresis is comparable and, we believe, far superior to radioactive-labeled slab gel electrophoresis protocols. Finally, using the well-described CbbR-cbbI promoter system as a model,9 we have duplicated previous footprinting patterns obtained by the more tedious gel electrophoresis methods. In addition, we have shown the new methods to be useful for identifying two distinct binding regions of the HrpY promoter region of the hrps gene.