Article Information

• PDF (293 kb)
• Full Text with Large Figures
• Cited by in Scopus (24)

PubMed

• Articles by B. Bonanni
• Articles by S. Cannistraro

Related Articles

• LexA-DNA Bond Strength by Single Molecule Force Spectroscopy...

  LexA-DNA Bond Strength by Single Molecule Force Spectroscopy Biophysical Journal, Volume 87, Issue 4, 1 October 2004, Pages 2683-2690 F. Kühner, L.T. Costa, P.M. Bisch, S. Thalhammer, W.M. Heckl and H.E. Gaub Abstract The SOS system of is coordinated by two proteins: LexA, a repressor protein of several unlinked genes, and the coprotease RecA. As known to date LexA controls 31 genes with slightly different DNA binding motifs allowing for a variable degree of repression from one gene to the other. Besides the SOS system LexA plays an important role in the regulation of transcription. The protein regulates transcription by using particular motifs to bind DNA, the helix-turn-helix motif. Here, we employed AFM-based single molecule force spectroscopy to characterize the interaction of LexA protein with two different DNA motifs: and . We measured the dissociation rates to be 0.045s for and 0.13s for respectively, which is in accordance with the predicted higher affinity between LexA- compared to LexA- The widths of the binding potentials were determined to be 5.4±1Å and 4.9±0.5Å, respectively. This short-ranged potential is characteristic for a stiff hydrogen-bonding network between protein and DNA. The unbinding occurs in a breakup rather than a gradual sliding. Abstract | Full Text | PDF (276 kb)

• Extending Bell's Model: How Force Transducer Stiffness Alter...

  Extending Bell's Model: How Force Transducer Stiffness Alters Measured Unbinding Forces and Kinetics of Molecular Complexes Biophysical Journal, Volume 94, Issue 7, 1 April 2008, Pages 2621-2630 Emily B. Walton, Sunyoung Lee and Krystyn J. Van Vliet Abstract Forced unbinding of complementary macromolecules such as ligand-receptor complexes can reveal energetic and kinetic details governing physiological processes ranging from cellular adhesion to drug metabolism. Although molecular-level experiments have enabled sampling of individual ligand-receptor complex dissociation events, disparities in measured unbinding force among these methods lead to marked variation in inferred binding energetics and kinetics at equilibrium. These discrepancies are documented for even the ubiquitous ligand-receptor pair, biotin-streptavidin. We investigated these disparities and examined atomic-level unbinding trajectories via steered molecular dynamics simulations, as well as via molecular force spectroscopy experiments on biotin-streptavidin. In addition to the well-known loading rate dependence of predicted by Bell's model, we find that experimentally accessible parameters such as the effective stiffness of the force transducer can significantly perturb the energy landscape and the apparent unbinding force of the complex for sufficiently stiff force transducers. Additionally, at least 20% variation in unbinding force can be attributed to minute differences in initial atomic positions among energetically and structurally comparable complexes. For force transducers typical of molecular force spectroscopy experiments and atomistic simulations, this energy barrier perturbation results in extrapolated energetic and kinetic parameters of the complex that depend strongly on . We present a model that explicitly includes the effect of on apparent unbinding force of the ligand-receptor complex, and demonstrate that this correction enables prediction of unbinding distances and dissociation rates that are decoupled from the stiffness of actual or simulated molecular linkers. Abstract | Full Text | PDF (300 kb)

• Affinity-Matured Recombinant Antibody Fragments Analyzed by ...

  Affinity-Matured Recombinant Antibody Fragments Analyzed by Single-Molecule Force Spectroscopy Biophysical Journal, Volume 93, Issue 10, 15 November 2007, Pages 3583-3590 Julia Morfill, Kerstin Blank, Christian Zahnd, Beatrice Luginbühl, Ferdinand Kühner, Kay-E. Gottschalk, Andreas Plückthun and Hermann E. Gaub Abstract For many applications, antibodies need to be engineered toward maximum affinity. Strategies are in demand to especially optimize this process toward slower dissociation rates, which correlate with the (un)binding forces. Using single-molecule force spectroscopy, we have characterized three variants of a recombinant antibody single-chain Fv fragment. These variants were taken from different steps of an affinity maturation process. Therefore, they are closely related and differ from each other by a few mutations only. The dissociation rates determined with the atomic force microscope differ by one order of magnitude and agree well with the values obtained from surface plasmon resonance measurements. However, the effective potential width of the binding complexes, which was derived from the dynamic force spectroscopy measurements, was found to be the same for the different mutants. The large potential width of 0.9nm indicates that both the binding pocket and the peptide deform significantly during the unbinding process. Abstract | Full Text | PDF (388 kb)

• …more

Copyright © 2005 The Biophysical Society. All rights reserved.
Biophysical Journal, Volume 89, Issue 4, 2783-2791, 1 October 2005

doi:10.1529/biophysj.105.064097

Spectroscopy, Imaging, Other Techniques

Single Molecule Recognition between Cytochrome C 551 and Gold-Immobilized Azurin by Force Spectroscopy

B. Bonanni^*, ,  , A.S.M. Kamruzzahan^†, A.R. Bizzarri^*, C. Rankl^†, H.J. Gruber^†, P. Hinterdorfer^† and S. Cannistraro^*

^* Biophysics and Nanoscience Centre, Istituto Nazionale Fisica della Materia-Consorzio Nazionale Interuniversitario per le Scienze Fisiche della Materia (INFM-CNISM), Dipartimento di Scienze Ambientali, Università della Tuscia, Viterbo, Italy
^† Institute for Biophysics, Johannes Kepler University of Linz, Linz, Austria

Address reprint requests to Beatrice Bonanni, Biophysics and Nanoscience Centre, INFM-CNISM, Dipartimento di Scienze Ambientali, Università della Tuscia, Largo dell’Università, I-01100 Viterbo, Italy. Tel.: 39-0761-357027; Fax: 39-0761-357179.

Over the last decade, redox metalloproteins have gained particular interest for possible applications in the challenging field of molecular bio-nanoelectronics 1,2,3,4. Their nanoscale dimension, as well as the variety of exploitable biological functions—such as recognition capability, catalysis, and electron transfer (ET)—may be combined with the progressing power of modern microelectronics for implementation of hybrid nanodevices. The possible integration of metalloproteins with electronic transducers has been recently explored with the aim to “wire them up” in efficient ET chains for biosensing applications at the level of the single molecule 1,5.

Planar gold surfaces appear as attractive solid supports for such applications, thanks to the ability to stably bind various kinds of proteins without affecting their biological activity 5 as well as for enabling both optical and electrical transduction schemes 6,7. Several groups have addressed this issue exploiting the high affinity of protein disulphides and thiols for gold, resulting in efficient conduction through the molecule toward the electrode 8,9,10,11,12,13.

Once the metalloprotein is successfully coupled with the electrode, a further step toward the applications in nano-biosensing devices may be represented by the recognition of the biomolecule physiological partner. Indeed, intermetalloprotein ET occurs through several steps 14,15 including protein recognition, which implies discriminating between reaction partners and other proteins, and after protein-specific binding with the formation of a (transient) complex 16. Once the specific complex is formed, ET can take place with optimal efficiency; the complex subsequently dissociates to yield the products 14,17.

Even if the ET process between two metalloproteins has been subject of investigation by many groups, most experiments were focused on the study of the free proteins in liquid. In this article we focus on the individual recognition between two redox partners after one of them has been immobilized on a gold electrode. The recognition of a metalloprotein by its redox partner directly linked to a metal electrode may, in principle, be detected as an electric signal, with potential applications for ultrasensitive biosensing at the single-molecule level.

The coupling of the redox protein Azurin (AZ) (from Pseudomonas aeruginosa) to gold electrodes has been extensively studied by our group 9,18. This small metalloprotein (MW=14.6) bears a single copper ion that switches between Cu(I) and Cu (II) oxidation states during the ET process. Opposite to the active site there is an exposed disulfide moiety (Cys³-Cys²⁶) which is capable to provide a stable binding of the protein to a gold substrate 19,20, resulting in efficient conduction through the molecule toward the electrode 9,11. The retention of AZ redox functionality upon chemisorption on gold has been clearly assessed by differential pulse voltammetry experiments 12 as well as by cyclic voltammetry measurements and electrochemical scanning tunneling microscopy 9,11,18.

P. aeruginosa cytochrome c 551 (C551) is a redox partner of AZ. This small protein (MW=9) has a heme-iron redox center, the iron switching its oxidation state between +2 and +3 in biological function. Both C551 and AZ are involved in the anaerobic respiratory chain of P. aeruginosa, and several experimental evidences of ET between these two metalloproteins have been provided 21,22,23,24,25. In particular, the observed efficiency of this ET process—in both directions—is consistent with the formation of an ET complex 17.

Even if, to the best of our knowledge, crystallographic evidence of the AZ/C551 complex is not yet available, protein residues involved in the interaction have been explored experimentally by means of site-directed mutagenesis strategy by Cutruzzolà et al. 24: these authors established that the interface between the two partners, during mutual interaction, comprises the hydrophobic patches surrounding residues Val²³, Ile⁵⁹ for C551 and Met⁴⁴, Met⁶⁴ for AZ.

A possible configuration of the AZ/C551 complex has been provided by computational docking methods (Brunori, E., A. R. Bizzarri, B. Bonanni, and S. Cannistraro, unpublished). Indeed, starting from the structures of component proteins and applying the docking method known as GRAMM 26,27, we were able to determine a configuration for the AZ/C551 complex which is consistent with experimental data about residues involved in AZ/C551 interaction 24. The best complex provided by our docking procedure (Brunori, E., A. R. Bizzarri, B. Bonanni, and S. Cannistraro, unpublished) has been selected by optimizing a variety of complex characteristic quantities (distance Fe-Cu, accessible surface area, H-bond, etc.) and is shown in Fig. 1.

Display large version of this figure

Figure 1

Schematic representation of best C551/AZ complex as obtained by GRAMM computational docking method (Brunori, E., A. R. Bizzarri, B. Bonanni, and S. Cannistraro, unpublished). The active sites (Cu for AZ, Fe for C551) are represented with small spheres. The disulfide group Cys³-Cys²⁶ of AZ is represented in golden-yellow. Residues involved in protein interaction (Met⁴⁴ and Met⁶⁴ for AZ, Val²³ and Ile⁵⁹ for C551, as from Ref. 24 are evidenced as well.

The molecular interaction between AZ and C551 can, in principle, be investigated by atomic force microscopy (AFM). Undoubtedly, AFM has been proven to be an ideal tool for measuring intermolecular forces, thanks to force sensitivity of the order of no more than a few picoNewtons, contact areas as small as 10nm², displacement sensitivity of 0.1nm, and the ability to operate in physiological conditions. By using a measuring tip with ligands bound and recording force versus distance cycles on surface-bound receptors, unbinding forces of ligand-receptor pairs can be probed at the level of single-molecule 28,29,30,31,32. Ligands have been often covalently coupled to the tip via long flexible spacer molecules 33,34,35. By this linkage method, the ligand on the tip is free to move and orient for unconditioned recognition of the surface-bound receptor, thus favoring complex formation. The characteristic stretching of the spacer 36, preceding complex dissociation, allows us to better discriminate specific unbinding events from unspecific adhesion, with great advantages for the study of low-affinity physiological partners 37.

In this article the molecular interaction between C551 and gold-immobilized AZ has been investigated by force spectroscopy. For optimal interaction between the two partners, proteins have been linked to substrate and tip with proper orientation for mutual interaction. More specifically, AZ has been immobilized on gold via Cys³ and Cys²⁶ residues, thus directing the protein to a configuration in which the hydrophobic patch, involved in interaction with C551 24, is oriented up 38—away from the electrode surface—i.e., facing the tip of an AFM (see Fig. 1). Conversely, C551 molecules have not been adsorbed directly on the AFM tip, but by employing a few-nm-long cross-linker that allows protein reorientation over the AZ sample, thus facilitating mutual interaction.

The two redox partners here investigated display a relative low affinity (the estimated value of equilibrium constant for complex dissociation being in the range 10⁻⁶–10⁻⁴ M; Ref. 22—that is, lower than values for ligand-receptor pairs commonly studied by force spectroscopy, but still comparable to that for cadherin dimers, which have been successfully investigated by this technique 37. It is also important to keep in mind that the likely transient nature of the ET complex here investigated 16 renders such ligand-receptor pairs very different from more stable complexes commonly studied by force spectroscopy experiments. To the best of our knowledge, this article represents the first study of molecular recognition of two redox partners by means of force spectroscopy. Here, the specificity of molecular recognition has been clearly demonstrated at the level of single molecule even after immobilization on gold of one of the partners. Additionally, by exploring the unbinding force of the complex for different values of loading rate, the dissociation kinetics has been investigated.

The AZ was purchased from Sigma-Aldrich (St. Louis, MO) and used without further purification, after dissolving in 50mM sodium phosphate buffer solution, pH 7. The substrates used in our experiments consist of vacuum-evaporated thin gold films (thickness 250nm) on borosilicate glass (using the product arrandee, Dr. Dirk Schröer, Werther, Germany). Before incubation with 35μM AZ solution, the gold substrates have been flame-annealed, to obtain recrystallized Au(111) terraces. The annealed Au(111) substrates were then incubated overnight at 4°C. Samples were gently rinsed with buffer solution to remove any unadsorbed material and immediately immersed in buffer for fluid imaging and force spectroscopy.

Cytochrome c 551 from P. aeruginosa (C551) was purchased from Sigma-Aldrich and used without further purification. The cross-linker used to bind C551 to the AFM tip is a heterobifunctional polyethylene-glycol derivative (PEG) which has been synthesized by some of the authors 33. The linker is 10-nm-long and bears an aldehyde (ALD) moiety at one end and n-hydroxy-succinimide (NHS) at the other end. A schematic representation of tip functionalization (with a success rate of ∼40%) is shown in Figure 2a. The Si₃N₄ cantilevers (Veeco Instruments, Santa Barbara, CA) were contact microlevers with backside gold coating and an oxide-sharpened tip. Probes have been first cleaned in three changes of chloroform and then treated with amine groups by overnight incubation with ethanolamine-HCl dissolved in dimethylsulfoxide. Subsequent incubation of tips with PEG solution results in stable binding of the cross-linker NHS-end with the amino-treated tip surface (being that the amine-NHS reaction is faster than the amine aldehyde reaction). PEG solution has been adjusted to ensure low density of cross-linkers on the Si₃N₄ surface, and therefore single-molecule detection by the tip. Tips functionalized with the PEG were finally incubated with the C551 solution. The cytochrome molecules stably bind to the ALD-end of the flexible PEG via one of its eight lysine residues, homogeneously distributed on the molecule surface (see Figure 2b). Before tip incubation with cytochrome, the C551 solution was passed through a Sephadex G-25 desalting column (PD10, Pharmacia, Peapack, NJ), to change the initial ammonium acetate pH 4.7 buffer (which may interfere with final aldehyde-lysine binding) to 50mM sodium phosphate buffer solution, pH 7. Functionalized tips were stored in buffer solution at 4°C until use. The cantilever spring constant of the functionalized tips (nominal value 0.030N/m) was experimentally determined by thermal noise analysis 39 with uncertainty of ∼7% on each cantilever. Measured values for used cantilevers ranged from 0.022 to 0.038N/m.

Display large version of this figure

Figure 2

(a) Schematic representation of the AFM tip functionalized with C551 via the ALD-PEG-NHS cross-linker. (b) Schematic representation of C551, with the heme group in red. Binding of C551 to the ALD-PEG is achieved via one of the eight lysine residues (highlighted in light blue) homogeneously distributed on the molecule surface. Residues Val²³ and Ile⁵⁹, belonging to the protein surface involved in interaction with AZ (as from Ref. 24, are represented as well.

Measurements were carried out by using alternatively a PicoSPM (Molecular Imaging, Phoenix, AZ) or a Nanoscope IIIa/Multimode AFM (Digital Instruments, Santa Barbara, CA).

For the characterization of Au(111) substrates, images were taken in air in contact mode configuration, by using silicon nitride contact microlevers (Veeco Instruments) with backside gold coating and an oxide-sharpened tip. The nominal spring constant was 0.3N/m.

Imaging of AZ/Au(111) samples was performed in buffer solution by intermittent contact AFM, either by magnetic AC mode (MAC mode by PicoSPM) or by conventional Tapping Mode (TMAFM by Nanoscope IIIa). In this case, the silicon nitride cantilevers (Veeco Instruments) had nominal spring constant of 0.1–0.5N/m and typical correspondent resonant frequency of 8–30KHz were used. The amplitude set point was selected to be ∼85% of free amplitude value.

To perform force spectroscopy experiments, the imaging tip was then replaced with the PEG-C551 functionalized tip. With such tips no images were taken, but only force-distance cycles were recorded over the AZ sample in buffer solution (50mM sodium phosphate, pH 7) at constant lateral position. To limit the force that the tip applied to the surface, as well as to minimize deviations between experiments, a relative trigger of 15–30nm was used in all deflection-distance curves. Approach and pulling velocity were identical in all experiments and varied from 30nm/s to 3000nm/s. All force-distance cycles here presented have been recorded setting the encounter time equal to zero.

The quality of the annealed gold surface was assessed by contact-mode AFM measurements, which showed the presence of atomically flat (111) terraces over hundreds of nanometers, with a typical roughness of ∼0.05–0.1nm.

The AZ/Au(111) sample was characterized in buffer solution by intermittent contact AFM. A typical MAC-Mode image of such a sample is reported in Fig. 3, which shows a stable and dense layer of molecules over the gold surface. The presence of a stable AZ monolayer on the gold surface after overnight incubation is also assessed by further AFM and scanning tunneling microscopy analysis performed by our group on analogous samples 9,18. In particular, samples prepared following such a protocol display reproducible good-quality scanning tunneling microscopy imaging, even after repetitive scans, indicative of full coverage with molecules stably bound to gold, as for covalent binding 9. The functionality of adsorbed AZ has been demonstrated by cyclic voltammetry experiments, shown elsewhere 9,11. The high concentration of molecules on the substrate, the stable linkage to gold and more importantly the preserved AZ functionality, are characteristics which make the sample suitable for force spectroscopy analysis. In particular, the retention of ET capability is indicative of redox center, and residues close to it (likely involved in interaction with C551; Ref. 24 being not significantly affected by immobilization.

Display large version of this figure

Figure 3

MAC-Mode AFM image of AZ monolayer over Au(111) substrate. The image has been recorded in buffer solution. Scale bar: 100nm.

The imaging tip was then replaced with the PEG-C551 functionalized tip for force spectroscopy experiments in contact mode. A representative force versus distance cycle recorded with the functionalized tip is plotted in Fig. 4. The retrace curve displays the characteristic stretching of the flexible PEG-linker, which shows a typical nonlinear behavior of the force versus distance retraction curve 36, and following C551-AZ recognition and unbinding. Similar curves have been recorded, at a scan rate of 3000nm/s, over different sites on the sample surface. Many specific recognition events (characterized by the delayed nonlinear course in the retraction curve) were observed in every lateral position probed with same tip, indicative of reproducible behavior during force-distance cycles, whereas very few adhesion events (characterized by linear retrace force slope) were observed. Reproducibility of unbinding events with the same tip is indicative of molecules not being pulled off the surfaces of the probe or the substrate during the process, since if AZ (or C551) molecules had remained attached to the probe (sample) surface, no repeatability would have been observed. Thousands of curves have been recorded and analyzed to evaluate the unbinding probability as well as the most probable unbinding force and typical unbinding length (see Fig. 4 for their definition). The unbinding force was measured from the pull-off jump of the retraction curve; in the case of multi-event retraces, only the last pull-off jump was considered. Values below the noise level, namely the standard deviation of the flat portion of sampled force-distance curve (for instance, 10 pN for the force-distance curve plotted in Fig. 4), were not taken into account. Statistical distribution of unbinding forces was found to be reproducible when hundreds of curves were recorded over different sites on the sample surface. In Figure 5a, the histogram of pull-off forces measured in AFM force-distance curves between C551-PEG tip and AZ substrate (scan rate=3000nm/s) is represented. Data come from the analysis of 268 unbinding events over 1500 force-distance cycles, which correspond to a binding probability of ∼18%. From the force histogram we estimate for the most probable unbinding force a value of 74 pN. To ensure that the most probable values are independent of the histogram arrangement, different classes of variable width were compared, and no significant differences in the maxima were found.

Display large version of this figure

Figure 4

Typical force-distance curve recorded with a PEG-C551 functionalized tip over an AZ/Au(111) sample showing a rupture event during retraction. A graphical representation of unbinding length (L) and unbinding force (F) is provided.

Display large version of this figure

Figure 5

(a) Histogram of the unbinding forces as measured from 268 rupture events observed over 1500 force-distance cycles. (b) Unbinding length statistical distribution evaluated for same series of curves providing the force distribution shown in a.

Regarding the binding probability, 18% is a bit lower than binding probabilities usually experienced by other complexes. For instance, many antigen-antibody investigated by force spectroscopy experiments show a binding probability close to 50% (40% for Human Serum Albumin and its antibody, Ref. 34; 60% for single-chain Fv antibody and antigen, Ref. 40; and 50% for γ-glutamyltransferase and anti-γ-glutamyltransferase AFM tip, Ref. 41. On the other hand, a binding probability as low as 15% is also found for some complexes such as P-selectin/ligand complexes (10–15% at a loading rate close to 10nN/s; Ref. 42, or in some experiments on streptavidin-biotin complexes (25%; Ref. 43, as well as for cadherin dimers 44. In the latter case, the binding probability is found to be ∼10%, even if such a value significantly increases when the tip is allowed to rest on the sample hundreds of milliseconds before retracting. As a matter of fact, a similar increase in the binding probability is also observed for our complex, the frequency of recognition events almost doubling for encounter time varying between 0 and 500ms, even if in the latter case multiple interactions cannot be ruled out 37. Undoubtedly, the steric hindrance of the binding process, due to the densely packed AZ-monolayer as well as to the low flexibility of the molecules on the gold substrate, is likely to have an effect in lowering binding probability 45 and, on the other hand, in increasing it for longer tip-surface contact times. In this respect, a different immobilization strategy (namely with lower AZ concentration and/or with the introduction of a spacer between AZ and gold) might somehow help in protein recognition. Nevertheless, it is also worth noticing that the relative low binding probability we observe may be a particular aspect of AZ/C551 interaction, since the not-high-specificity is thought to be a characteristic of ET complexes 16.

From the unbinding curves, a distribution of unbinding length was calculated, by evaluating the extension corresponding to the nonlinear portion of the force versus distance retraction curve (see Fig. 4). The resulting length statistical distribution, correspondent to force distribution of Figure 5a, is reported in the histogram of Figure 5b. The most probable unbinding length is ∼18nm, a value which well compares to typical values found in other force spectroscopy experiments using similar PEG cross-linkers 34. Moreover, this value is in good agreement with the extension expected for similar PEG upon stretching 36. Such finding provides further evidence that the corresponding pull-off events, selected as specific unbinding events, derive from the stretching of the PEG linker and following protein unbinding 46, whereas a possible unfolding of one of the interacting proteins is unlikely.

To test the reproducibility of the experiment, thousands of force-versus-distance curves were also recorded with different C551-PEG tips over AZ/Au(111) samples (data not shown here). The statistical distributions found for unbinding force and unbinding length provided similar peak values to those shown in Figure 5ab, previously discussed, indicative of the good reproducibility of our results.

As already mentioned, the selection of curves displaying specific unbinding events is facilitated by the introduction of the PEG cross-linker, since its stretching preceding the pull-off and characterized by nonlinear behavior of force versus distance curve allows us to discard aspecific adhesion curves. Nevertheless, to sustain the specificity of rupture events, control experiments were performed to test inhibition of AZ/C551-tip recognition. To this aim, force-distance curves were recorded over the AZ/Au(111) sample using the same C551 functionalized tip which provided the 18% probability of unbinding events, but after the addition of free AZ solution in the fluid cell. To ensure that most of the C551 on the tip was in complex with free AZ, the added AZ solution concentration was 15μM, that is, comparable to K_d for C551/AZ complex (K_d ≈ 10⁻⁶–10⁻⁴M; see Ref. 22. After free AZ addition, the tip and substrate were incubated 30min in the AZ solution before starting to record force-versus-distance cycles. A corresponding increase of number of curves displaying no rupture events was observed, the unbinding probability progressively changing from the initial average value of 18% down to a final stable value of 5.5%, therefore resulting in 70% overall decrease of number of unbinding events. The significant decrease of binding probability is indicative of blocking recognition between C551 on the tip and the AZ adsorbed on the Au(111) surface, likely due to binding of free AZ to the C551 proteins on the tip (i.e., tip-blocking).

The unbinding force distribution before and during tip-blocking is represented in Fig. 6. The similarity of the two histograms seems to indicate that the observed 5.5% pertaining unbinding events are related to C551/AZ specific recognition not-blocked. The 70% decrease in unbinding events is consistent with choosing blocking agent concentration close to K_d (as also observed for other complexes; Ref. 37, K_d representing the concentration of ligand that—at equilibrium—causes binding to half receptors 47. It is worth noting that the persistence of a 3–6% residual unbinding activity is, however, observed in most blocking experiments, even in very effective blocking conditions 34,41, and that such residual activity is likely to be induced by the forced interaction between the two molecules, due to the experimental setup. Nevertheless, the lack of full blocking may be ascribed, to some extent, also to the transient nature of the complex, which may determine the occurrence of a sort of dynamic exchange of binding between the cytochrome on the tip and the free AZ in solution.

Display large version of this figure

Figure 6

Comparison of unbinding force statistical distribution before (dashed) and after (solid) tip-blocking. After tip-blocking, the total number of unbinding events significantly decreases, corresponding to a final 70% reduction of unbinding probability.

After the tip-blocking experiment, the measuring tip was washed out in copious buffer solution. By recording force-distance cycles in temporal sequence, a progressive recovering in the number of unbinding events was observed, up to a final stable unbinding probability of 17%. Unbinding probability during tip-wash is shown in Fig. 7, together with values before and during tip-blocking (note that reported values refer to statistical analysis of force-distance curve series recorded in temporal sequence). The recovering of unbinding events during the washout is indicative of a reversible blocking mechanism, and reminds us of the weak nature of the ET complex here investigated 16.

Display large version of this figure

Figure 7

Unbinding probability before (shaded) and during (solid) tip-blocking, as well as after tip-washing (dashed shaded line). The reported values refer to a statistical analysis of a force-distance curve series recorded in a temporal sequence.

Additional experiments have been performed to investigate the unbinding kinetics of C551/AZ complex. In principle, from loading rate-dependent measurements of the unbinding force (dynamic force spectroscopy), it is possible to determine the dissociation rate of a bond 48,49. More specifically, the unbinding process of a ligand-receptor pair under the influence of an external loading force can be treated, in terms of the Bell model 50, as a kinetic problem of escape from a potential well: the effect of the external load is then to tilt the interaction potential and facilitate thermally activated escape from the bound state 48,51,52. As a consequence, the dissociation rate constant k_off of a bond is amplified by the application of an external force F as

(1)

(2)

Display large version of this figure

Figure 8

(a) Statistical distributions of unbinding force for maximum and minimum values of applied loading rate. At every loading rate, 1000 force-distance curves have been recorded. (b) Loading-rate dependence of the most probable unbinding force F* resulting from a Gaussian fit to the histogram distribution. Force statistical errors are given by standard deviation (2σ/N^1/2 for 95.4% confidence level). Error on loading rate is evaluated from uncertainty in the effective spring constant of the cantilever. The solid line is a numerical fit of experimental data to the Bell model (see Eq. (2) in the text). Best-fitting parameters are k_off (0)=(14±2) s⁻¹ and X_β=(0.14±0.01) nm.

It is commonly accepted that ET complexes, such as AZ/C551, have high off-rates (as well as on-rates) with typical k_off up to 10³s⁻¹ (whereas k_on is the range 10⁷–10⁹M⁻¹s⁻¹) 16. Such peculiarity renders the AZ/C551 complex very different from most complexes usually studied by force spectroscopy experiments, which are more stable. As a matter of fact, for the well-known biotin-avidin complex, the typical timescale of the binding process is of the order of months (k_off=9–10⁻⁸s⁻¹ and k_on=7–10⁷M⁻¹s⁻¹; see Ref. 53, whereas antigen-antibody complexes have higher off-rates, but are still in the range of 0.01s⁻¹.

Even if the value of 14s⁻¹ we estimate is not as high as typical k_off expected for ET transient complexes 16, it is, however, much higher than off-rates typically estimated by dynamic force spectroscopy for “stable” ligand-receptor pairs. For these, conversely, dynamic force spectroscopy provided k_off (0) values in the range of 10⁻³–10⁻¹s⁻¹ (as for antigen-antibody; Ref. 54 or even lower (for the well-known biotin-avidin, k_off (0) ∼10⁻⁵s⁻¹; Ref. 43; still, for the low-affinity complex of cadherin dimers a value not larger than 1.8s⁻¹ has been provided experimentally 37.

The best-fitting value for X_β, describing the difference in separation of C551 and AZ between the bound and the transition state, is (0.14±0.01) nm—which is comparable to values found for other ligand-receptor pairs.

As widely discussed, the unbinding force may strongly depend on loading rate; thus the comparison of the AZ/C551 unbinding forces with values from literature for other complexes can be made only if considering the data recorded at same loading rate. Therefore, to complete the comparison of data from our experiments with similar studies—available from literature—on other complexes, we focus our attention on a particular value of the real loading rate, namely 10nN/s. As a matter of fact, many force spectroscopy experiments on several ligand-receptor pairs have been performed applying such loading-rate value. From the plot shown in Figure 8b, we can provide an estimate for the most probable unbinding force F* of ∼95 pN when a real loading rate of 10nN/s is applied. This value can be now compared to unbinding forces of other ligand-receptor pairs, as summarized in Table 1 from literature. From the table, we may conclude that the force we measure for AZ/C551 is in the range of forces typically measured for many ligand-receptor complexes, varying between 20 and 240 pN.

The study of individual recognition between C551 and gold-immobilized AZ has been, for the first time, accomplished by atomic force spectroscopy and the specificity of this molecular interaction has been clearly demonstrated by control block experiments. The dissociation kinetics of the complex has been explored and an estimate of the complex off-rate has been provided, which (as expected) results much higher than typical values for more stable complexes usually investigated by this technique.

Importantly, the effectiveness of molecular recognition between two redox partners when one of them is directly linked to a metal electrode is a good starting point for possible detection of single recognition events as a recordable electric signal. Our results thus confirm the great potentiality of redox proteins for implementation of ultrasensitive bio-nanodevices designed for biological screening at the single-molecule level.