Site-Specific Binding of the 9.5 Kilodalton DNA-Binding Protein

Direct Visualization of the Binding of c-Myc/Max Heterodimeric b-HLH-LZ to E-Box Sequences on the hTERT Promoter. Réjean Lebel, François-Olivier McD...
0 downloads 0 Views 549KB Size
Download PDF
Biomacromolecules 2005, 6, 1252-1257

1252

Site-Specific Binding of the 9.5 Kilodalton DNA-Binding Protein ORF80 Visualized by Atomic Force Microscopy M. Lysetska,† H. Zettl,‡ I. Oka,† G. Lipps,‡ G. Krauss,‡ and G. Krausch*,† Lehrstuhl fu¨r Physikalische Chemie II and Lehrstuhl fu¨r Biochemie, Universita¨t Bayreuth, 95440 Bayreuth, Germany Received September 6, 2004; Revised Manuscript Received December 17, 2004

Atomic force microscopy (AFM) has been used to examine the binding properties of the DNA-binding protein ORF80 to DNA. ORF80 is a 9.5 kDa protein that binds site-specifically to double-stranded DNA of the sequence TTAA-N7-TTAA. Direct sizing of the protein complexes on DNA fragments from the plasmid pRN1 with AFM shows that the protein ORF80 binds preferentially to two positions. These positions agree well with the ORF80 binding sites determined by footprinting analysis. The measurements allow an estimate of the stoichiometry of the DNA-protein complexes. In contrast to previous results, the single-molecule experiments suggest that only a low number of ORF80 molecules bind to a DNA-binding site. Introduction pRN11

Plasmid from the acidothermophilic archaeon Sulfolobus islandicus shares three highly conserved open reading frames with the other members of the plasmid family pRN. One of the open reading frames, ORF80, encodes a 9.5 kDa protein. Recombinant ORF80 has been overexpressed in Escherichia coli and characterized by Lipps et al.2 ORF80 is a sequence-specific double-stranded DNAbinding protein. According to footprinting and gel shift analysis, ORF80 has two binding sites on the plasmid pRN1. These binding sites are separated by 60 bp and are both located upstream of its own gene. The consensus motif of the binding site is the palindromic sequence TTAA-N7TTAA. Fluorescence titrations and gel-shift experiments have indicated the formation of large protein-DNA complexes of distinct stoichiometry.2 The physiological function of the gene product ORF80 is not yet known. Within the plasmid family pRN the protein ORF80 is the most highly conserved protein, suggesting that ORF80 is an essential protein for the maintenance or replication of the plasmid. The lack of sequence similarity of ORF80 to characterized proteins does not allow us to predict the function of the protein. However, ORF80 exhibits a weak similarity in amino acid sequence and (predicted) secondary structure to several sequencespecific DNA-binding proteins with a so-called winged helix. Proteins with this DNA-binding motif are bacterial repressors and eukaryotic transcription factors as well as proteins involved in initiating DNA replication. To get an insight into the function of this small DNA-binding protein, we analyzed the architecture of the ORF80-DNA complexes by atomic force microscopy. Our measurements were performed under near-physiological buffer concentrations in liquids. Atomic force microscopy has proven to be a powerful tool for the study of protein-DNA complexes, providing infor* Corresponding author: e-mail [email protected]. † Lehrstuhl fu ¨ r Physikalische Chemie II. ‡ Lehrstuhl fu ¨ r Biochemie.

mation on the topology of the DNA and the stoichiometry of the complexes. Important insight into changes in the topology of DNA during the process of DNA repair8 could be obtained by high-resolution AFM. As examples, wrapping of DNA around the replication protein RPA,6 kinking of the DNA in complexes with the repair protein XPC,9 and cooperative formation of RecA nucleofilaments on DNA10 have been studied by AFM. Materials and Methods DNA Substrates. A 538 bp long DNA fragment was obtained by PCR with the forward primer 5′-CGCCACTTGGCGAGAAATTTGCTCAAAG-3′ and the reverse primer 5′-GGTTGAGCTCGAGTCACAGGAGTTCGTCACGGC3′ by use of the plasmid puc-pRN1 as template. The PCR reaction was purified by spin chromatography on Silica spin columns (Qiagen) and the concentration of the 538 bp fragment was determined by UV spectroscopy. The purity of the DNA preparation was checked by gel electrophoresis with subsequent ethidium bromide staining. The PCR product contained two full binding sites for ORF80 and had the following structure: (N)191-TTAA-N7-TTAA-(N)49-TTAAN7-TTAA-(N)268. Protein. ORF80 was overexpressed in E. coli and purified as described earlier.2 Because of the high aggregation tendency the protein was stored in 6 M urea at -20 °C. ORF80 retains its DNA-binding activity under these conditions. AFM Imaging and Analysis. AFM measurements were performed on a MultiMode AFM (Digital Instruments, Santa Barbara, CA) operated in tapping mode. Silicon nitride oxidation sharpened tips were used. The drive frequencies ranged between 3.4 and 34 kHz. All measurements were performed in a buffer containing 8 mM HEPES (pH 8), 2 mM KCl, 100 mM NaCl, and 2 mM MgCl2. Sample preparation for AFM measurements was improved by further

10.1021/bm0494489 CCC: $30.25 © 2005 American Chemical Society Published on Web 03/01/2005

Site-Specific Binding of ORF80 Visualized by AFM

Biomacromolecules, Vol. 6, No. 3, 2005 1253

Figure 2. Dimensions of the objects measured from AFM images. Distribution of the frequencies of (A) DNA contour length, (B) DNA contour length with one globular object bound to the DNA molecules, and (C) DNA that carries two distinct globular objects, and measurements of the heights of (D) pure DNA, (E) ORF80 molecules, and(F) DNA/ORF80 complexes are shown. Solid lines represent Gaussian fits of the frequency distribution.

Figure 1. AFM images of the biomolecules. (A) Immobilized 538 bp DNA molecules on mica. (B) ORF80 molecules immobilized on the mica surface without nickel addition. (C) ORF80 (100 nM) was incubated with 3 nM DNA. Single DNA molecules are visible that carry globular objects. (D) Same as in panel C but larger scan size. (E, F) Same as in panel D but 300 nM ORF80 (arrows point to the ORF80DNA complexes; bar is 200 nm; z range is 10 nm).

purification of all solutions with a 3 kDa microfilter (Millipore Corp.). Immobilization of the biomolecules on mica was achieved via 2.5-5 mM NiCl2 without further rinsing of the samples and omitting any drying. For all AFM investigations we kept the DNA concentration constant at 3 nM and varied the protein concentration. Image processing was performed with the Nanoscope III 4.43r8 software package along with some homemade software routines for image analysis. For quantitative analysis only molecules that were entirely imaged in the chosen scan area were used. Results For the AFM experiments on the ORF80-DNA interaction we choose a 538 bp DNA molecule containing two full binding sites for ORF80 (see Materials and Methods). One binding site is located approximately in the middle of the DNA molecule, while the other is located slightly asymmetrically (at 49.1% and 37.1% of the DNA total contour length, respectively). Figure 1A shows a typical image of the 538 bp DNA molecules immobilized on mica. The average contour length of the DNA molecules measured from

the AFM images is 177 ( 12 nm (Figure 2A). This value is in agreement with the expected length of 181.8 nm for a 538 bp long DNA fragment (0.338 nm/bp). The average height of the DNA molecules as measured from the AFM images is 1.6 ( 0.3 nm (Figure 2D). Since ORF80 is positively charged under the buffer conditions used, it is easily immobilized on the negatively charged mica surface without counterion addition. Figure 1B shows a typical AFM image of ORF80. Five microliters of 600 nM ORF80 solution was used for sample preparation. Under these conditions isolated ORF80 molecules adhere to the surface. From a set of images similar to the one shown in Figure 1B, one can measure the dimensions of the adsorbed protein. We find a height of 2.4 ( 0.6 nm (Figure 2E). DNA-protein complexes were obtained by preincubation of DNA and ORF80 in defined ratios for 10 min. Subsequently, the solution containing the complexes was deposited onto a freshly cleaved mica surface. Immobilization of the molecules in liquids for AFM investigation was performed as described earlier.6 As can be seen in Figure 1C-F, we observe small globular objects bound to the DNA. No such objects were visible on DNA immobilized in the absence of ORF80 (Figure 1A). To characterize the size of these complexes, we have calculated a height histogram (Figure 2F). The height of these objects is 2.8 ( 0.7 nm, which is slightly higher than the dimensions of the ORF80 molecules immobilized on mica (Figures 1B and 2E). Figure 1C-F demonstrates that despite its small size, ORF80 molecules complexed to DNA can be well resolved by the AFM measurements in the liquid environment. ORF80 binds to DNA rather weakly with an apparent dissociation constant of 140 nM.2 Therefore we assumed that a rather high concentration of ORF80 was required to obtain

1254

Biomacromolecules, Vol. 6, No. 3, 2005

Lysetska et al.

Figure 3. Gallery of 3D plots of the AFM images of the DNA ORF80 complexes: DNA molecules carrying (A) one or (B) two globular objects, (C) surface plot of the DNA that carries two globular objects of different heights, and (D) one elongated object sitting on the DNA molecule. Lateral data scale is 170 × 170 nm; z data scale is 10 nm.

ORF80-DNA complexes to be observed in the AFM measurements. Indeed, at a protein concentration of 100 nM only about 15% of the DNA molecules were found to be complexed with protein (Figure 1C,D). In contrast, at a protein concentration of 300 nM nearly 80% of DNA molecules were bound by ORF80 (Figure 1E,F). At this protein concentration we also observed larger aggregates of DNA molecules, possibly bound together via ORF80 molecules (Figure 1E,F). To investigate the specificity of the ORF80 DNA binding, we measured the ORF80 positions along the DNA molecules on more than 200 protein-DNA complexes. Examples of the high-resolution AFM images of the ORF80-DNA complexes are shown in Figure 3. We find complexes with one (Figure 3A) or with two globular objects along a DNA molecule (Figure 3B). Quantitative analysis of the ORF80 positions along the DNA clearly shows that the distribution of the ORF80 along the DNA molecule is not random (Figure 4A,B). In Figure 4A we plot the location of the ORF80 molecules along DNA molecules with a single binding site occupied by protein. Figure 4B presents the data for DNA molecules carrying two globular objects. We find that in the case of a single ORF80 binding event to the DNA (Figure 4A) the ORF80 molecules are preferentially located near the middle of the DNA molecule, while the asymmetric site is only rarely populated. For DNA molecules carrying two distinct globular objects, analysis of the position of these objects corresponded to the location of the ORF80 binding sites (Figure 4B). These experiments clearly show the sitespecific binding of ORF80. While site-specific binding of the proteins to the DNA molecules is convincingly seen in the AFM images, it is hardly possible to determine the exact number of proteins bound to each binding site (i.e., the stoichiometry of the complexes). Cross-sectional analysis of the AFM images

Figure 4. Distribution of the frequencies of the positions of the ORF80 along DNA molecule. (A, B) Analysis of the DNA molecules with (A) one and (B) two ORF80 molecules. When only one protein molecule is bound, the protein binds preferentially in the middle of the DNA molecule. When two globular objects are bound on the DNA, the protein molecules are located at around 35%, 37% (?), and 48% of the DNA contour length. Lines correspond to the Gaussian fitting of the frequencies distribution. (C) Comparison of the nucleic acid sequences of the two ORF80 binding sites. The core of the palindromic consensus motif is boxed with the center of 2-fold symmetry depicted with an oval. Identical bases are connected with vertical lines.

shows that the globular objects have an average height of 2.8 ( 0.7 nm (Figure 2F), which is only slightly higher than the size of uncomplexed ORF80 (Figure 2E), suggesting that these complexes contain one or two ORF80 molecules. In addition, we observe complexes with two globular objects of different heights located on the DNA (Figure 3C), probably corresponding to the binding of a larger number of proteins to a single binding site. For example, a crosssection taken along the line connecting the centers of the globular objects located on the DNA presented in Figure 3C

Site-Specific Binding of ORF80 Visualized by AFM

Biomacromolecules, Vol. 6, No. 3, 2005 1255

Figure 5. Interaction of the DNA with ORF80. (A)At 100 nM ORF80 concentration only rare cases of DNA-protein complexes are visible. Arrows indicate DNA molecules with only one protein molecule. (B) One of the complexes marked in panel A with an asterisk. Two DNA molecules can be connected via protein molecules that are located on their specific binding sites. (C) Higher magnification of the complex marked in panel A with a cross shows that binding sites are glued together via protein and only free ends of 82 and 61 nm are visible. (D) Overview of the numerous DNA-protein agglomerates. Higher ORF80 concentrations (1 µM) lead to an increase of the DNA-protein agglomerates. Objects containing different numbers of the DNA and protein molecules are seen.

shows that the larger object is 4.2 nm high while the smaller object is 3.2 nm high. The height of the DNA molecules in these images amounts to 1.9 nm. Sometimes rather long objects are observed up to 12 nm in length (Figure 3D), indicating the formation of large complexes. At an ORF80 concentration of 100 nM, most complexes contain one or two globular objects on a single DNA molecule. In addition, we also observe complexes involving two different DNA strands (Figure 5A,B). DNA molecules in these aggregates appear to be connected via proteins bound to the different strands. Only about 1.5% of all DNA molecules are present in such complexes. At yet higher ORF80 concentrations, large agglomerates involving many DNA molecules are formed (Figure 5D). These aggregates are of varying shape and dimensions, but from the AFM images it is obvious that they are composed of many DNA and protein molecules. A quantitative analysis of the contour length of the DNA molecules yields no difference between the DNA alone and DNA molecules bound to ORF80 (Figure 2A-C). This finding indicates that the DNA molecules do not “wrap around” the proteins. Only in the rare cases where big globular objects are observed on the DNA (Figure 5C) is the apparent DNA contour length shorter, by approximately 20 nm (marked with arrows in Figure 2B,C). This shortening may derive from looping out of the DNA between the proteins located on the two different ORF80 binding sites. The height of such objects is up to 4.5 nm (see bracket in Figure 2F). In summary, the AFM experiments show specific binding of ORF80 to a long DNA molecule exhibiting two ORF80 binding sites. At the protein concentrations studied here, we were not able to identify complexes containing large numbers of protein bound to a single DNA molecule, as was shown by Lipps et al.2 To critically test the results obtained by AFM, we have performed complementary experiments using fluorescence correlation spectroscopy (FCS) (not shown here).7 In line with the AFM results, FCS experiments using dye-

labeled DNA did not indicate a mass increase characteristic of the binding of considerably more than one or two ORF80 molecules. Discussion The most significant result of our AFM experiments is the direct observation of specific binding of the ORF80 molecules to the two binding sites on the dsDNA. To our knowledge, protein-DNA complexes of such small size have not been studied with AFM before. We have been able to show the simultaneous formation of distinct complexes at the two DNA binding sites, which are separated by only 60 base pairs, corresponding to a distance of 22.3 nm. Single complexes are also visible and in these cases the complex is preferentially located at the more centrally located binding site. Due to the finite lateral resolution of AFM and the broad length distribution of the complex it is, however, not possible to clearly distinguish both binding positions (Figure 4A). The two ORF80 binding sites are similar but not identical (see Figure 4C). The nucleobases between and outside the halfsites of the consensus motif TTAA-N7-TTAA can contribute to the binding of ORF80 to DNA. Most likely the neighboring sequences of the “middle” binding site lead to a stronger binding of ORF80, explaining the larger number of binding events at the center position. A 3-fold stronger binding in terms of dissociation constant requires only a marginal increase in binding free enthalpy of about 3 kJ/mol, which may be easily contributed by contacts to nucleobases outside the core consensus. When two protein complexes could be clearly identified on a single dsDNA molecule, their positions were found to coincide well with the theoretically expected positions. Analysis of the contour length of the DNA does not indicate a wrapping of the DNA around ORF80. Wrapping occurs often in larger protein-DNA complexes as, for example, in nucleosomes,11 and it has been also observed in smaller protein-DNA complexes as, for example, for RPA bound to damaged DNA6. For DNA complexes of transcrip-

1256

Biomacromolecules, Vol. 6, No. 3, 2005

tion factors,12 which often bind as dimers to symmetrically arranged half-sites of a DNA element, no wrapping has been described. ORF80 also appears to form this type of complexes. Our AFM measurements allow a crude estimate of the maximal number of protein molecules present in the complexes. This analysis is important since from fluorescence studies we have deduced a very high stoichiometry of up to 24 ORF80 molecules bound to a single 538 bp DNA molecule carrying two full ORF80 binding sites.2 ORF80 is a protein of molecular mass 9.5 kDa. Assuming a globular shape and a density of about 1.2 g/cm3, one would expect a diameter of about 3 nm for a single ORF80 molecule. This value compares well with the height of about 3 nm determined from AFM images of ORF80 in the absence of DNA (Figure 2E). The height of the majority of the DNA-protein complexes as measured from the AFM images is 2.8 ( 0.7 nm, suggesting that about two ORF80 molecules are bound in these cases to the palindromic TTAA-N7-TTAA motif. A significantly higher stoichiometry as indicated by the fluorescence measurements of Lipps et al. is clearly not seen for the majority of complexes investigated. However, up to 20% of the complexes show a larger height (peaks bracketed in Figure 2F), as expected from binding of a single ORF80 molecule to a half-site, suggesting the presence of multiple ORF80 molecules. Similarly high values are observed neither for pure DNA (Figure 2D) nor for pure protein (Figure 2E). Overall, the AFM images show some heterogeneity of the complexes (Figure 3C). The cross-section analysis of DNAprotein complex presented in Figure 2F reveals globular objects with different heights that may correspond to a variable number of bound proteins in comparison to Figure 2B. However, most of the complexes are compatible with the binding of a single ORF80 molecule to a halfsite of the ORF80 recognition motif. Possibly, complexes of lower stoichiometry in which the DNA is more easily accessible for ionic interactions with the Ni-coated surface, are deposited preferentially to the surface. Because of the tip broadening effect, the widths of the globular objects sitting on DNA cannot be determined unambiguously. For example, in Figure 3D we see a single globular object of a somewhat elongated shape. It is possible that several ORF80 molecules are located close to each other on the protein binding sites. Structures as depicted in Figure 5C probably derive from gluing together the two ORF80 binding sites by protein-protein interactions. One can often see a sharp bend in DNA at the location a globular object is located. Only short arms, for example, of 82 and 61 nm of the DNA molecule are visible in these cases (Figure 5C). Increasing the ORF80 concentration shows the presence of complexes containing more than one DNA molecule, and these DNA molecules are glued together at the positions of protein attachment (like in Figure 5A,B). The increase in ORF80 concentration appears to promote nonspecific binding, and proteins bound to the ends of the dsDNA are also observed (Figure 4A,B). The aggregation tendency of ORF80 has already been observed in our previous study, where a high concentration of nonspecific competitor DNA was used

Lysetska et al.

to analyze the binding to the radioactively labeled specific DNA. The inclusion of nonspecific competitor DNA was not possible for the AFM experiments reported here. We were forced to work in a narrow concentration range: at too low protein concentration, we are not able to observe enough binding events, and at too high protein concentration, aggregation takes place. At low protein concentrations the specific binding to the binding sites will predominate. At higher protein concentrations DNA-bound ORF80 apparently engages in protein-protein interactions leading to higher complexes. The AFM experiments had to be performed with a large stoichiometric excess of protein to DNA, which will accordingly promotes the aggregation of ORF80. Although we could clearly demonstrate the site-specific binding of ORF80, it remains uncertain whether ORF80 can form a larger specific nucleoprotein complex at a specific site of the plasmid. These larger nucleoprotein complexes are observed at replication origins; for example, the dnaA protein from E. coli assembles into a distinct nucleoprotein complex13 that prepares the DNA for strand opening and replication. Further studies will be required to establish whether ORF80 is also implicated in replication initiation. Conclusion An AFM study of the ORF80 interaction with dsDNA shows distinct peculiarities of the protein binding to the DNA. At low protein concentrations, specific binding of ORF80 to two binding sites that are separated by 60 nucleotides can be followed. We do not observe wrapping of the DNA around the bound protein. The specific complexes formed contain about two ORF80 molecules per TTAA-N7-TTAA motif. Higher ORF80 concentrations lead to the formation of big protein-DNA agglomerates that contain numerous DNA and protein molecules. The AFM results are in agreement with FCS measurements, which also indicate that only a small number of ORF80 molecules bind to each binding site on the DNA. Acknowledgment. We are grateful for financial support through the Deutsche Forschungsgemeinschaft (Kr1369/11). References and Notes (1) Keeling, P. J.; Klenk, H. P.; Singh, R. K.; Feeley, O.; Schleper, C.; Zillig, W.; Doolittle, W. F.; Sensen, C. W. Plasmid 1996, 35, 141144. (2) Lipps, G.; Ibanez, P.; Stroessenreuther, T.; Hekimian, K.; Krauss, G. Nucleic Acids Res. 2001, 29, 4973-4982. (3) Jankowski, T.; Janka, R. ConfoCor 2sThe Second Generation of Fluorescence Correlation Microscopies, 2001. (4) Schwille, P.; Meyer-Almes, F. J.; Rigler, R. Biophys. J. 1997, 72, 1878-1886. (5) Foldes-Papp, Z.; Rigler, R. Biol. Chem. 2001, 382, 473-478. (6) Lysetska, M.; Knoll, A.; Boehringer, D.; Hey, T.; Krauss, G.; Krausch, G. Nucleic Acids Res. 2002, 30, 2686-2691. (7) Lysetska, M. Intact and Damaged DNA and their Interaction with DNA Binding Proteins: a Single Molecule Approach; Dissertation, Universita¨t Bayreuth, Bayreuth, Germany, 2004. (8) Janicijevic, A.; Ristic, D.; Wyman, C. J. Microsc. 2003, 212, 264272. (9) Janicijevic, A.; Sugasawa, K.; Shimizu, Y.; Hanaoka. F.; Wijgers, N.; Djurica, M.; Hoeijmakers, J. H.; Wyman, C. DNA Repair (Amst.) 2003, 2, 325-36. (10) Sattin, B. D.; Goh, M. C. Biophys. J. 2004, 87, 3430-3436.

Site-Specific Binding of ORF80 Visualized by AFM (11) Luger, K.; Mader, A. W.; Richmond, R. K.; Sargent, D. F.; Richmond, T. J. Nature 1997, 389, 251-260. (12) Tan, S.; Richmond, T. J. Curr. Opin. Struct. Biol. 1998, 8, 41-48.

Biomacromolecules, Vol. 6, No. 3, 2005 1257 (13) Erzberger, J. P.; Pirruccello, M. M.; Berger, J. M. EMBO J. 2002, 21, 4763-4773

BM0494489