The Nature of the Force-Induced Conformation Transition of dsDNA

Publication Date (Web): February 23, 2010. Copyright © 2010 American Chemical Society. *Corresponding author. Prof. Wenke Zhang, State Key Laboratory...
0 downloads 10 Views 2MB Size
Download PDF
pubs.acs.org/Langmuir © 2010 American Chemical Society

The Nature of the Force-Induced Conformation Transition of dsDNA Studied by Using Single Molecule Force Spectroscopy Ningning Liu, Tianjia Bu, Yu Song, Wei Zhang, Jinjing Li, Wenke Zhang,* and Jiacong Shen State Key Laboratory of Supramolecular Structure and Materials, College of Chemistry, Jilin University, Changchun, 130012, P. R. China

Hongbin Li Department of Chemistry, University of British Columbia, Vancouver, British Columbia, Canada V6T 1Z1 Received January 4, 2010. Revised Manuscript Received February 15, 2010 Single-stranded DNA binding proteins (SSB) interact with single-stranded DNA (ssDNA) specifically. Taking advantage of this character, we have employed Bacillus subtilis SSB protein to investigate the nature of force-induced conformation transition of double-stranded DNA (dsDNA) by using AFM-based single molecule force spectroscopy (SMFS) technique. Our results show that, when a dsDNA is stretched beyond its contour length, the dsDNA is partially melted, producing some ssDNA segments which can be captured by SSB proteins. We have also systematically investigated the effects of stretching length, waiting time, and salt concentration on the conformation transition of dsDNA and SSB-ssDNA interactions, respectively. Furthermore, the effect of proflavine, a DNA intercalator, on the SSB-DNA interactions has been investigated, and the results indicate that the proflavine-saturated dsDNA can be stabilized to the extent that the dsDNA will no longer melt into ssDNA under the mechanical force even up to 150 pN, and no SSB-DNA interactions are detectable.

Introduction An understanding of the mechanical stability of a dsDNA, and how it is influenced by the binding of other molecules, is of crucial importance for a deep understanding of numerous important biological processes, including DNA transcription and replication. In the past years, direct mechanical measurements on single molecules have been reported with a number of different techniques, including atomic force microscopy (AFM), optical tweezers, magnetic beads, glass microneedles, and the biomembrane force probe.1 Among these techniques, AFM-based SMFS is a versatile platform for the study of inter- or intramolecular interactions in both (natural) biological and synthetic systems. This technique has offered novel perspectives, revealing structural and mechanical properties of synthetic and biopolymers.2-12 DNA is among the most intensively studied polymeric molecules *Corresponding author. Prof. Wenke Zhang, State Key Laboratory of Supramolecular Structure and Materials, College of Chemistry, Jilin University. E-mail: [email protected]. Telephone number: þ86-431-85159203.

(1) Merkel, R. Phys. Rep. 2001, 346, 343–385. (2) Florin, E.-L.; Moy, V. T.; Gaub, H. E. Science 1994, 264, 415–417. (3) Rief, M.; Gautel, M.; Oesterhelt, F.; Fernandez, J. M.; Gaub, H. E. Science 1997, 276, 1109–1112. (4) Janshoff, A.; Neitzert, M.; Oberd€orfer, Y.; Fuchs, H. Angew. Chem., Int. Ed. 2000, 39, 3212–3237. (5) Clausen-Schaumann, H.; Seitz, M.; Krautbauer, R.; Gaub, H. E. Curr. Opin. Chem. Biol. 2000, 4, 524–530. (6) Hugel, T.; Seitz, M. Macromol. Rapid Commun. 2001, 22, 989–1016. (7) Zhang, W. K.; Zhang, X. Prog. Polym. Sci. 2003, 28, 1271–1295. (8) Liu, C. J.; Shi, W. Q.; Cui, S. X.; Wang, Z. Q.; Zhang, X. Curr. Opin. Solid State Mater. Sci. 2005, 9, 140–148. (9) M€uller, D. J.; Dufr^ene, Y. F. Nat. Nanotechnol 2008, 3, 261–269. (10) Schwartz, D. K.; Whitten, D. G. Langmuir 2008, 24, 1109 (special issue on Molecular and Surface Forces). (11) (a) Zhang, X.; Liu, C. J.; Wang, Z. Q. Polymer 2008, 49, 3353–3361. (b) Zhang, Y. H.; Yu, Y.; Jiang, Z. H.; Xu, H. P.; Wang, Z. Q.; Zhang, X.; Oda, M.; Ishizuka, T.; Jiang, D. L.; Chi, L. F.; Fuchs, H. Langmuir 2009, 25, 6627–6632. (c) Yu, Y.; Zhang, Y. H.; Jiang, Z. H.; Zhang, X.; Zhang, H. M.; Wang, X. H. Langmuir 2009, 25, 10002–10006. (12) Giannotti, M. I.; Vancso, G. J. ChemPhysChem 2007, 8, 2290–2307.

Langmuir 2010, 26(12), 9491–9496

by using SMFS. SMFS studies on dsDNA reveal the complex conformational changes in its double-helix secondary structures during stretching. It was previously found that the dsDNA reveals a highly cooperative overstretching transition at 65 pN followed by a nonequilibrium melting transition at 150 pN.13,14 However, the subtle conformation during the overstretching transition is still under debate, and there are different opinions about this process. Some predicted that a majority of the DNA base pairs melted or dissociated as the DNA was stretched during the cooperative transition,15-18 while different opinions presented by others indicated that DNA was unwound to form a ladder-like structure and the base pairs remained hybridized during this transition.13,14,19-24 The dsDNA used in the experiments above is (13) Rief, M.; Clausen-Schaumann, H.; Gaub, H. E. Nat. Struct. Biol. 1999, 6, 346–349. (14) Clausen-Schaumann, H.; Rief, M.; Tolksdorf, C.; Gaub, H. E. Biophys. J. 2000, 78, 1997–2007. (15) Williams, M. C.; Wenner, J. R.; Rouzina, I.; Bloomfield, V. A. Biophys. J. 2001, 80, 874–881. (16) Rouzina, I.; Bloomfield, V. A. Biophys. J. 2001, 80, 882–893. (17) Williams, M. C.; Rouzina, I.; Bloomfield, V. A. Acc. Chem. Res. 2002, 35, 159–166. (18) (a) McCauley, M. J.; Williams, M. C. Biopolymers 2007, 85, 154–168. (b) Shokri, L.; Rouzina, I.; Williams, M. C. Phys. Biol. 2009, 6, 025002. (c) Shokri, L.; Marintcheva, B.; Eldib, M.; Hanke, A.; Rouzina, I.; Williams, M. C. Nucleic. Acids. Res. 2008, 36, 5668–5677. (d) Shokri, L.; Marintcheva, B.; Richardson, C. C.; Rouzina, I.; Williams, M. C. J. Biol. Chem. 2006, 281, 38689–38696. (e) Pant, K.; Karpel, R. L.; Rouzina, I.; Williams, M. C. J. Mol. Biol. 2005, 349, 317–330. (f) Pant, K.; Karpel, R. L.; Rouzina, I.; Williams, M. C. J. Mol. Biol. 2004, 336, 851–870. (g) Pant, K.; Karpel, R. L.; Rouzina, I.; Williams, M. C. J. Mol. Biol. 2003, 327, 571–578. (h) Shokri, L.; McCauley, M. J.; Rouzina, I.; Williams, M. C. Biophys. J. 2008, 95, 1248–1255. (19) Cluzel, P.; Lebrun, A.; Heller, C.; Lavery, R.; Viovy, J.-L.; Chatenay, D.; Caron, F. Science 1996, 271, 792–794. (20) Smith, S. B.; Cui, Y. J.; Bustamante, C. Science 1996, 271, 795–799. (21) Cocco, S.; Yan, J.; Leger, J.-F.; Chatenay, D.; Marko, J. F. Phys. Rev. E 2004, 70, 011910 (1-18). (22) Zhou, H. J.; Zhang, Y.; Ou-Yang, Z. Phys. Rev. Lett. 1999, 82, 4560–4563. (23) Chiang, K. N.; Yuan, C. A.; Han, C. N.; Chou, C. Y.; Cui, Y. J. Appl. Phys. Lett. 2006, 88, 023902 (1-3). (24) Lebrun, A.; Lavery, R. Nucleic Acids Res. 1996, 24, 2260–2267.

Published on Web 02/23/2010

DOI: 10.1021/la100037z

9491

Article

Liu et al.

very long (several micrometers in length), and along the molecules there may be some nicks which can labilize and split the dsDNA into two single strands more easily when the dsDNA is stretched by mechanical force, complicating the force-induced conformation transition. In addition, although researchers have tried to utilize some additives (e.g., bacteriophage SSB and glyoxal), which can grab the ssDNA that might be produced during the overstretching transition, the arguments on force-induced melting transition were not so convincing due to the fact that in those systems the additives can initially destabilize the dsDNA.18 So, the true conformation of dsDNA during the overstretching transition is still not clear. In an effort to understand the nature of such force-induced conformation transition experimentally, we have employed AFM-based SMFS together with a carefully prepared dsDNA fragment to test those different models. To do that, dual-labeled dsDNA (biotin and thiol) fragments (∼3000 bp, ∼1 μm in length) were prepared by using PCR reaction, in which 50 end labeled primers were used. The labeled dsDNA was bridged by their 50 ends between a streptavidin-functionalized AFM tip and a gold substrate via specific molecular recognition (biotin-streptavidin) and gold-thiol chemistry and was then stretched and relaxed subsequently under different buffer conditions. The SSB protein, which exists widely in a variety of organisms, is a kind of protein which can interact with ssDNA specifically. After introducing Bacillus subtilis SSB proteins into the DNA system, a marked hysteresis appeared between the stretching and relaxation force curves when the dsDNA was stretched to the overstretching transition region and then relaxed. This phenomenon has been ascribed to the fact that the dsDNA molecule is partially melted into single strands and the SSB proteins can capture these ssDNA fragments. We have also investigated the effects of stretching length, waiting time, and salt concentration on the force-induced conformation transition of dsDNA and SSB-ssDNA interactions. Using the similar method described above, we have also found that, in the presence of proflavine, a kind of intercalator which can stabilize dsDNA, the dsDNA cannot be melted into single-stranded DNA, and no SSB-DNA interactions are detectable.

Materials and Methods Materials. High-purity deionized water (>18 MΩ cm), produced from a Millipore system, was used for the preparation of all experimental buffers. The PBS (from Sigma) buffer was composed of 10 mM phosphate buffer, 2.7 mM KCl, and 137 mM NaCl, pH 7.4. Another two kinds of Tris-HCl buffer (pH 7.8) were both composed of 50 mM Tris-HCl and 1 mM EDTA but different concentrations of NaCl, 50 mM and 200 mM. KOD DNA polymerase was purchased from Toyobo. The concentration of proflavine (from Sigma) used in this study was 20 μM. DTT (from Sigma) solution (0.2 mM) was prepared by diluting 0.2 μL DTT (1 M in dH2O) in 1 mL PBS buffer solution. The Bacillus subtilis SSB protein was overexpressed and purified as described.25,26 The concentration of SSB protein was determined spectrophotometrically. Dual-labeled linear DNA fragments (3058 bp) were obtained using PCR amplification with pCERoriD plasmid25 as a template (about 44 ng/reaction). HPLC-pure primers, 50 -biotin-GAGTCAGTGAGCGAGGAAGC-30 and 50 -thiol-AGCTCACTCAAAGGCGGTAA-30 , were purchased from Sigma-Genosys (Suffolk, UK) to perform all these preparations. The primers (25) Zhang, W. K.; Dillingham, M. S.; Thomas, C. D.; Allen, S.; Roberts, C. J.; Soultanas, P. J. Mol. Biol. 2007, 371, 336–348. (26) Zhang, W. K.; Machon, C.; Orta, A.; Phillips, N.; Roberts, C. J.; Allen, S.; Soultanas, P. J. Mol. Biol. 2008, 377, 706–714.

9492 DOI: 10.1021/la100037z

were purchased already modified with the relevant end group, such as thiol or biotin. The reactions were carried out in 200 μL PCR tubes; each contains 50 μL of reaction mixture as the final volume following the standard procedure of KOD Plus PCR Kit in an Eppendorf thermal cycler (Applied Biosystems, Eppendorf AG, Germany). Initially before 30 cycles, there was a hot start by heating the solution to 94 °C for 2 min. Each cycle comprised three steps: first, a denaturation step in which the solution was heated to 94 °C for 15 s; second, an annealing step lasted for 30 s (the annealing of the template and the primers was carried out at a temperature of 10 °C below the melting temperature which is the average of both primers); third, an extension phase lasted 3 min at 68 °C. The amplification obtained was checked by running the PCR reaction on a 0.5% agarose-TAE gel, and the solutions in the wells that showed amplification were pooled together. The DNA was then purified by using EasyPure PCR Purification Kit (TransGen Biotech, Beijing) and quantified by using Helios UV Visible spectrophotometer (Thermo Electronics) at 260 nm to give a concentration of 120 ng/μL.

Immobilization of Labeled Linear DNA Fragments onto Gold Substrate. The purified 50 -thiolated and biotin-labeled linear DNA was absorbed from its solution onto a clean gold surface for 12 h, followed by immersing in DTT solution (0.2 mM) for 30 min in order to desorb the DNA, which was immobilized on the surface by physisorption. DNA Stretching Experiments. The force spectroscopy experiments were carried out on a NanoWizardII BioAFM (JPK instrument AG, Berlin, Germany) in contact mode by using streptavidin (Promega) functionalized Si3N4 AFM tips (Veeco Metrology, CA) as described before.26,27 The spring constants of the employed AFM cantilevers were calibrated using the thermal noise method.28,29 The measured values ranged from 20 to 25 pN/ nm. During the stretching experiment, streptavidin-coated AFM tip was brought to contact with the dsDNA immobilized on the gold substrate. The dsDNA could then be picked up and manipulated by the AFM tip via biotin-streptavidin interactions. As long as a single dsDNA is attached between the AFM tip and the solid substrate, it can be stretched and relaxed in different buffer solutions and the corresponding stretching and relaxation traces can be recorded accordingly. The pulling speed was 2 μm/s for all the stretching-relaxation experiments unless stated otherwise.

Results and Discussion Comparison of Force Curves for dsDNA in the Absence and Presence of SSB Protein. Figure 1A shows some typical force curves of pure dsDNA obtained by repeatedly manipulating the same DNA molecule. As can be seen from the figure, each pair of stretching and relaxation curves can be superimposed very well, even though the stretching force has reached as high as 200 pN. This indicates that the dsDNA fragments were happy on the surface and no damage (such as multiple site nicking) had happened during the sample preparation. In addition, this immobilization method has increased the chance of successful manipulation, namely, it is easier for the AFM tip to pick up a DNA molecule, and the same molecule can be stretched and relaxed many times without rupture.27 In addition, due to the fact that the DNA-AFM tip interactions are based on the reversible specific interactions between biotin and streptavidin, even if the dsDNA molecule gets detached from the AFM tip, it can be picked up again under relatively mild condition compared with that of the physisorption-based immobilization methods, in which harsh contact between the AFM tip and the sample is needed. Frequent harsh contact can cause damage to the DNA (27) Zhang, W. K.; Barbagallo, R.; Madden, C.; Roberts, C. J.; Woolford, A.; Allen, S. Nanotechnology 2005, 16, 2325–2333. (28) Hutter, J. L.; Bechhoefer, J. Rev. Sci. Instrum. 1993, 64, 1868–1873. (29) Butt, H.-J.; Jaschke, M. Nanotechnology 1995, 6, 1–7.

Langmuir 2010, 26(12), 9491–9496

Liu et al.

Figure 1. Typical force spectra of dsDNA obtained by repeatedly manipulating of the same DNA molecule in the absence (A) and presence (B) of SSB protein. There is a marked hysteresis between each pair of stretching and relaxation force curve after SSB proteins are introduced.

fragment, for example, nicking or breaking of dsDNA, complicating the force-induced conformation transition. After introducing SSB proteins (6.5 μM) into the system, a marked hysteresis appeared between each pair of stretching and relaxation force curves as shown in Figure 1B. In view of SSB protein interacting with ssDNA specifically rather than with dsDNA, these hysteresis can be attributed to the fact that the dsDNA is partially melted into ssDNA fragments during overstretching transition and the SSB proteins capture these ssDNA fragments, slowing down the rehybridization process of complementary strands during relaxation. Effect of Stretching Length on the Conformation Transition of dsDNA and SSB-ssDNA Interactions. After a dsDNA molecule was picked up by an AFM tip and was then stretched and relaxed repeatedly, we increased the stretching length gradually and found that the hysteresis increased accordingly, as shown in Figure 2A. A simple explanation for this phenomenon is that as the DNA molecule is stretched beyond its contour length, the helix structure is unwound, and some base pairs are broken (e.g., the AT rich region), producing ssDNA fragments. Applying an external force onto a dsDNA molecule is, in a way, equivalent to increasing the temperature of the system. It is reported that, when the DNA is heated up, there are some transient single-stranded breathing intermediates within dsDNA,30 causing a lower melting temperature. The action of dsDNA being partially melted by mechanical force in our experiment is similar to such a thermal melting process. As the molecule is stretched even further, more and more ssDNA fragments are produced along the dsDNA; as a result, more SSB-ssDNA complexes will form, which slow down the rehybridization process even further. What will happen after the dsDNA is relaxed, then? To answer this question, we superimposed these stretching curves in Figure 2A, and found that they could be superposed very well, producing a similar height as that (30) Cubeddu, L.; White, M. F. J. Mol. Biol. 2005, 353, 507–516.

Langmuir 2010, 26(12), 9491–9496

Article

Figure 2. Effect of stretching length on the conformation transition of dsDNA and SSB-ssDNA interactions. With the increase of stretching length, the hysteresis between stretching-relaxation traces increases as well (A). The good superposition of the stretching curves indicates that after relaxation the DNA being stretched goes back to its original state (without SSB binding) (B).

of pure dsDNA at ∼65 pN (see Figure 2B). This indicates that after relaxation the dsDNA can go back to its original state without any SSB proteins staying on it, and also indicates that the presence of Bacillus subtilis SSB protein does not affect the stability of a relaxed dsDNA. This is crucial, since if the additives (the SSB protein in our case) themselves can affect the stability of dsDNA, it will be difficult to determine the nature of forceinduced conformation transition of a pure dsDNA. Williams et al., who are active supporters of the melting model, have done systematic work on SSB-DNA interactions using two other types of SSB proteins, bacteriophage T4 gp32 and T7 gp2.5.18 However, due to the fact that these two types of proteins can destabilize dsDNA, their conclusion on the melting model was not that convincing. Our results here can further prove their arguments in that, even if the dsDNA is not initially destabilized by Bacillus subtilis SSB protein, ssDNA fragments can still appear and be captured by the SSB during the overstretching transition. Scheme 1 shows a possible model for the process of forceinduced conformation transition of dsDNA and SSB-ssDNA interactions. When the dsDNA molecule is stretched between the AFM tip and the gold substrate by mechanical force of about 65 pN, the double helix structure is unwound along the dsDNA, and in some region, e.g., the AT rich region, the hydrogen bonds of base pairs are broken, producing some ssDNA fragments. If the melting happens in the middle of dsDNA, a melting “bubble”30 will form, while if this happens in two ends of dsDNA, free ssDNA will be produced, as shown in Scheme 1 II. In the absence of SSB protein, the separated complementary strands can rehybridize very quickly during relaxation. As a result, no hysteresis will appear on the time scale of our SMFS experiments. In the presence of SSB protein, however, the transient ssDNA fragments are captured by SSB proteins, which inhibit the rehybridization of ssDNA during relaxation, causing the appearance of hysteresis between the stretching and relaxation curves (Scheme1 III). As in our experiment, the ssDNA fragments (at least half of them) are DOI: 10.1021/la100037z

9493

Article

Liu et al.

Scheme 1. Possible Model for Force-Induced Conformation Transition of a dsDNA and SSB-ssDNA Interactions

Figure 3. Four typical stretching and relaxation cycles on dsDNA in the presence of SSB protein with different waiting times at the same extended state. The hysteresis increases gradually as the waiting time is increased from 0 to 3 s.

produced between the AFM tip and the gold substrate under tension; this may prevent the ssDNA from wrapping around the SSB proteins efficiently by the binding modes reported previously.31,32 So, the SSB proteins can only partially interact with ssDNA fragments. As a result, the interactions may not be firm enough comparing to the complete wrapping modes, so the SSB proteins are easier to strip off from the ssDNA during relaxation. Finally, the dsDNA returns to its original state and no SSB proteins stay on the DNA molecule (Scheme1 IV). Our model on the force-induced conformation transition of a dsDNA in the overstretching transition region support the partial melting argument and agrees quite well with the recent theoretical simulations.33,34 Although Mameren et al. have shown that the conversions (dsDNA to ssDNA) prefers to start from a nick or free DNA ends, small-scale melting “bubbles” (e.g., in the AT rich region) cannot be totally excluded due to the limited optical resolution of single-molecule fluorescence imaging.35 In addition, our results show that, at the beginning of overstretching transition, only a small portion of dsDNA is melted the rest (bound segments) are still kept base-paired, although the DNA already exists in a very extended form. With the increase of stretching length, more bound segments are converted to melted form, as shown in Figure 2A. Effect of Waiting Time on the Conformation Transition of dsDNA and SSB-ssDNA Interactions. If our assumption on the nature of force-induced conformation transition of dsDNA is true, then the increase of waiting time during the stretched state should affect the overstretching transition and SSB-ssDNA interactions as well. To test that, we performed the stretchingrelaxation experiment with different waiting times. That is, we stretched the DNA molecule to a designed length (e.g., halfway of the overstretching transition plateau), then kept the molecule under such extension length for certain amount of time before the molecule was relaxed. We find that, as the waiting time is increased, the hysteresis becomes stronger accordingly, as shown in Figure 3. This is reasonable, since longer waiting time during the stretched state will first increase the chance of SSB binding to (31) Bujalowski, W.; Lohman, T. M. Biochemistry 1986, 25, 7799–7802. (32) Kozlov, A. G.; Lohman, T. M. Biochemistry 2002, 41, 6032–6044. (33) Hanke, A.; Ochoa, M. G.; Metzler, R. Phys. Rev. Lett. 2008, 100, 018106 (1-4). (34) Whitelam, S.; Pronk, S.; Geissler, P. L. J. Chem. Phys. 2008, 129, 205101 (1-8). (35) Mameren, J. V.; Gross, P.; Farge, G.; Hooijman, P.; Modesti, M.; Falkenberg, M.; Wuite, G. J. L.; Peterman, E. J. G. Proc. Natl. Acad. Sci. 2009, 106, 18231–18236.

9494 DOI: 10.1021/la100037z

the exposed ssDNA. Although the rate of our SSB proteins binding to ssDNA may be as rapid as that of E. coli SSB,32 there might still exist some naked gaps on the ssDNA fragments. As the waiting time is increased, the proteins far away from the DNA will come to occlude ssDNA, or rearrangement of SSB along the ssDNA fragments may happen. In addition, a prolonged external load will weaken the bound base pairs converting more overstretched dsDNA fragments into ssDNA. Either of these two situations can produce more SSB-ssDNA complexes, causing more hysteresis between stretching and relaxation curves. Effect of Salt Concentration on the Rehybridization and SSB-ssDNA Interactions. From the discussion above, we know that in the presence of enough SSB the number of broken base pairs during stretching will determine how big the hysteresis is. In addition, factors that can affect the binding behavior (binding strength and mode) of SSB protein with ssDNA will also affect the rehybridization and the hysteresis eventually. Although the Bacillus subtilis SSB protein has not been adequately studied,25,26 it has a similar sequence to several other kinds of prokaryotic ssDNA-binding proteins, such as E. coli SSB. This indicates that extensive homology, especially in the Cterminal DNA binding domains, exists among these proteins. For E. coli SSB, electrostatic interaction and stacking between base pairs and protein play major roles in SSB-ssDNA interactions.36 It has been shown recently that salt concentration strongly affects the rezipping of two unzipped cDNA strands in the presence of E. coli SSB protein.37 This has been attributed to different binding modes as well as the binding affinity under different salt concentrations. In our current system, the reannealing of the separated complementary strands should in nature be the same as that of rezipping mentioned above, although the ssDNA fragments in our case are produced by force-induced melting under higher external load. To see the effect of salt concentration on the interactions between Bacillus subtilis SSB protein and ssDNA and the conformation transition of dsDNA, we have performed the stretching-relaxation cycles in the presence of Bacillus subtilis SSB protein with different NaCl concentrations, 50 mM and 200 mM, respectively. To exclude effects, such as the stretching length/ratio, on the hysteresis, we have tried to keep the molecule stretched to the same stretching ratio in these two different NaCl buffers. After superimposing two pairs of typical normalized stretching and relaxation force curves obtained under (36) Raghunathan, S.; Kozlov, A. G.; Lohman, T. M.; Waksman, G. Nat. Struct. Biol. 2000, 7, 648–652. (37) Hatch, K.; Danilowicz, C.; Coljee, V.; Prentiss, M. Nucleic Acids Res. 2008, 36, 294–299.

Langmuir 2010, 26(12), 9491–9496

Liu et al.

Article

Figure 4. Two pairs of typical normalized stretching-relaxation curves obtained in the presence of SSB protein in 50 mM and 200 mM NaCl buffers with both stretching curves marked in red, whereas the relaxation curves obtained in 50 mM and 200 mM NaCl buffers are marked in blue and black, respectively.

each experimental condition, we find that the hysteresis between the stretching and the relaxation curve is stronger in 50 mM NaCl buffer than in 200 mM NaCl, as shown in Figure 4, while no such effect has been observed for stretching-relaxation traces in the absence of SSB protein (in other words, no hysteresis is observed during the stretching and relaxation of a pure dsDNA under these two salt concentrations). This indicates that, similar to the E. coli SSB-ssDNA interactions,37 the Bacillus subtilis SSB-ssDNA interactions also show salt dependence. Consider the following two facts: first, the salt concentrations in our experiment is not so high; second, the SSB proteins bind to the extended ssDNA. We believe that it is the higher binding cooperativity under lower salt (50 mM NaCl) concentration that mainly contributes to the relatively stronger hysteresis. It is very likely that, in low salt buffer solutions, the Bacillus subtilis SSB proteins adopt a cooperative binding mode with ssDNA, for example, the (SSB)35 (a SSB tetramer bind with 35 DNA bases) mode, causing the formation of more SSB-ssDNA complexes slowing down the reannealing process even further. However, in relatively higher salt buffers (e.g., 200 mM NaCl), the binding is less cooperative, causing the incomplete coverage of SSB proteins on ssDNA. As a result, SSB protein can be removed more quickly/easily from the melted ssDNA during relaxation, resulting in the decrease of hysteresis. Proflavine Binding Prevent the Melting of dsDNA by External Force. Proflavine is an acridine dye that binds to DNA by simple intercalation between base pairs. Intercalation occurs by the insertion of planar ligands between neighboring base pairs involving interaction with the π-orbitals of the stacked DNA bases. It is known to bind without sequence specificity and unwinds the double helix by 11° per molecule bound.38 Increasing proflavine concentration shortens the overstretching plateau and decreases the cooperativity of the overstretching process.39 In our experiment, we used the proflavine with a concentration of 20 μM, and the cooperativity of the conformation transition was already reduced significantly, as shown in Figure 5A. After introducing SSB proteins (6.5 μM) into the DNA-proflavine system, the profiles of the stretching-relaxation force curves are found to be similar to that of the curves in the absence of SSB protein; i.e., no hysteresis appeared during subsequent stretching and relaxation cycles when the DNA-proflavine complexes were stretched by force even up to 150 pN (Figure 5). When (38) Graves, D. E.; Velea, L. M. Curr. Org. Chem. 2000, 4, 915–929. (39) Krautbauer, R.; Fischerl€ander, S.; Allen, S.; Gaub, H. E. Single Mol. 2002, 3, 97–103.

Langmuir 2010, 26(12), 9491–9496

Figure 5. Typical force curves of proflavine-treated dsDNA in the absence (A) and presence of SSB (B,C) with different waiting times, 0 s (B), and 2 s (C). There is no hysteresis between stretching and relaxation curves under the mechanical force up to 150 pN.

the waiting time at the stretched state was increased to 2 s, which is more likely to cause strand separation of dsDNA (as discussed above, Figure 3), there was no hysteresis when the DNA-proflavine complexes were stretched by force up to 120 pN (Figure 5C). Due to the limitation of biotin-streptavidin interaction which is the weakest link in the bridge structure, it is difficult to apply larger force on the dsDNA in our current system. These results indicate that, in the presence of this concentration proflavine, no ssDNA fragments are produced by stretching. Then, a plausible explanation for this phenomenon is that when proflavine binds to dsDNA preferentially, it stabilizes the helix by planar interaction. As more proflavine is added, the fractions of proflavine-free dsDNA decrease, requiring more energy to melt dsDNA. When the concentration of proflavine reaches a critical value, dsDNA is saturated by the proflavine and no regions can be melted by force, so no hysteresis appears during our experiment.27 From the discussion above, we know that electrostatic interactions play important roles in SSB binding to ssDNA. In addition, during our SMFS experiment, ssDNA fragments were produced under external force. In other words, SSB proteins bound mainly to the extended ssDNA instead of free relaxed ones. For proflavine-saturated dsDNA, it will also exist in a very extended conformation (such as the ladder-like conformation). However, no obvious binding between SSB proteins and such dsDNA were detectable, as indicated by the good superposition of the stretching and relaxation traces (see Figure 5B,C). Of course, the relative high rigidity of dsDNA as compared with ssDNA will somehow affect the effective binding (especially the wrapping) of SSB proteins with proflavine-extended DNA. Apart from that, this also means that other interactions, such as intercalation and hydrophobic interactions between exposed DNA bases and the proteins, also contribute quite a lot to SSB-ssDNA interactions. In proflavine-dsDNA complexes, the DNA bases are very well paired and their freedom is restricted (not as free as that of ssDNA). As a result, these paired bases cannot interact with SSB proteins.

Conclusions By using AFM-based SMFS together with the ssDNA binding properties of Bacillus subtilis SSB protein, we have successfully revealed the nature of force-induced conformation transition of the dsDNA. Our results indicate that dsDNA is partially melted into ssDNA during the overstretching transition (i.e., dsDNA exist as a mixture of the dsDNA and molten ssDNA) at the DOI: 10.1021/la100037z

9495

Article

mechanical force of about 65 pN, and the SSB proteins are able to capture the transient ssDNA fragments slowing down the rehybridization process, causing the hysteresis between stretching and relaxation traces. It is also found that, in the overstretching transition region, the longer the dsDNA is stretched, the more the dsDNA is melted; the longer the waiting time at the extended state, the more ssDNA fragments will be produced. A comparative study on the interactions of SSB proteins with proflavinesaturated dsDNA show that both the electrostatic interaction and intercalating (as well as hydrophobic) interaction are the main driving forces for the formation of Bacillus subtilis SSB-ssDNA complexes which has not been reported previously. This study

9496 DOI: 10.1021/la100037z

Liu et al.

has deepened our understanding on the mechanical stability of dsDNA, and may shed light on the mechanism of many basic biology processes, where SSB protein has participated, in vivo, including DNA replication, recombination, and repair. Further investigations on the direct quantitative measurement of Bacillus subtilis SSB and ssDNA interactions are underway. Acknowledgment. This work was funded by the NSFC Special Fund Program (20844003), NSFC Key Program (20834003), and NSFC International Cooperation and Exchange Program (20640420622). W. Z. would like to thank Prof. Panos Soultanas (Nottingham University) for continuous help.

Langmuir 2010, 26(12), 9491–9496