Mechanical properties of BiP protein determined by nano‐rheology

Abstract

Immunoglobulin Binding Protein (BiP) is a chaperone and molecular motor belonging to the Hsp70 family, involved in the regulation of important biological processes such as synthesis, folding and translocation of proteins in the Endoplasmic Reticulum. BiP has two highly conserved domains: the N‐terminal Nucleotide‐Binding Domain (NBD), and the C‐terminal Substrate‐Binding Domain (SBD), connected by a hydrophobic linker. ATP binds and it is hydrolyzed to ADP in the NBD, and BiP's extended polypeptide substrates bind in the SBD. Like many molecular motors, BiP function depends on both structural and catalytic properties that may contribute to its performance. One novel approach to study the mechanical properties of BiP considers exploring the changes in the viscoelastic behavior upon ligand binding, using a technique called nano‐rheology. This technique is essentially a traditional rheology experiment, in which an oscillatory force is directly applied to the protein under study, and the resulting average deformation is measured. Our results show that the folded state of the protein behaves like a viscoelastic material, getting softer when it binds nucleotides‐ ATP, ADP, and AMP‐PNP‐, but stiffer when binding HTFPAVL peptide substrate. Also, we observed that peptide binding dramatically increases the affinity for ADP, decreasing it dissociation constant (K _D) around 1000 times, demonstrating allosteric coupling between SBD and NBD domains.

Abbreviations

BiP

Immunoglobulin Binding Protein

ER

endoplasmic reticulum

Hsp70

heat shock protein 70

NBD

nucleotide binding domain

SBD

substrate binding domain

GNPs

Gold Nanoparticles

Introduction

Immunoglobulin Binding Protein (BiP) is a ∼75 kDa, ∼ 8–10 nm sized, monomeric protein, member of a highly conserved and ubiquitous group of chaperones, the Hsp70 family, constituting one of the principal actors maintaining proteostasis inside the cell. BiP plays crucial roles in different processes and functions, such as protein synthesis, folding, assembly, post‐translational translocation across the Endoplasmic Reticulum (ER) membrane,1, 2 and it is known as the master regulator of the ER. Structurally, BiP has two highly conserved domains: the N‐terminal Nucleotide‐Binding Domain (NBD), and the C‐terminal Substrate‐Binding Domain (SBD), connected by a hydrophobic linker. ATP binds and it is hydrolyzed to ADP in the NBD, and BiP's extended polypeptide substrates bind in the SBD. The NBD is formed by two lobes, between which is the nucleotide‐binding site. The SBD, on the other hand, can be further divided into a compact β‐sandwich domain harboring a pocket for substrate binding and an α‐helical domain at its C‐terminal end, the so‐called “lid”, covering this pocket.2, 3

To be functional, BiP couples the Mg²⁺ dependent hydrolysis of ATP to large conformational changes, involving both of its structural domains, leading to a dynamic ATP dependent cycle transiting between high and low affinity states for its substrate binding.4, 5 Thus, lid opening and closing over the bound substrate, and distance fluctuations between the SBD and NBD domains, which become farther upon substrate binding and ATP hydrolysis, have been associated with BiP functionality. Moreover, allosteric communication between both domains has been suggested.3, 6 Surprisingly, the precise mechanism involved in the coupling between nucleotide and substrate binding with conformational dynamics in BiP ‐and in most chaperones‐, is not fully understood. In general, this can be explained as the experimental techniques employed in the past to study Hsp70 dynamics, have been mainly based on high resolution crystal structures, obtained by X‐Ray crystallography or NMR.3, 7 Nevertheless, it has been difficult to crystallize a whole Hsp70 protein (at present day there is just one complete Hsp70 crystallized), obtaining mainly lot of partial Hsp70 crystal structures, leading to an incomplete dynamical whole picture.3, 8 Moreover, most substrate binding studies have been performed using small peptides instead of entire unfolded proteins.9, 10 Previously work done with DnaK (a close homologue of yeast BiP, around 40% sequence identity) shows that it binds and stabilizes partially folded protein structures.11 Interesting and novel approaches to study dynamics in chaperones have arised combining spectroscopic techniques, like Heteronuclear Single Quantum Coherence Spectroscopy (HSQC) with computational simulations.12, 13, 14 A different and novel approach to study mechanical properties of proteins has been the manipulation of single molecules subjected to force application.15 Recently, our group has been studying BiP binding properties, unfolding completely a protein substrate using single molecule optical tweezers. Our results showed that BiP binds to the unfolded state of MJ0366 substrate protein reversibly, preventing its refolding, and that this effect depends on both the type and concentration of nucleotides.15

Taking all this in consideration, and assuming that BiP mechanical properties are beginning to be understood, our group has been trying to find new molecular descriptors about BiP acting as a chaperone and as a molecular motor during ER translocation.16 To achieve this, in this study, we have focused on studying how BiP's mechanical properties are correlated with nucleotides and substrate binding, considering modifications in the internal viscosity of the folded state of BiP, and how these properties modulate its function.

Dynamic and mechanical properties of proteins are correlated, as the latter depend on the forces and energies involved in maintaining proteins own structure and folding, as well as on the conformational changes coupled to ligand binding. Moreover, dynamic properties of the enzymes are determined by physical‐structural changes that they experience as a function of the different transitions that they visit within a viscoelastic regime, as a consequence of ligand binding or the action of a force.17, 18, 19 In this way, when an enzyme is subjected to a force it can behave as an elastic or viscous material (viscoelastic regime), getting stiffer or more flexible (softer), depending on the force applied on it. Protein's viscoelastic transition is a universal mechanical property of the folded state, and it is relevant for the large conformational changes, which often accompany substrate binding in proteins.20 Nano‐rheology is a novel ensemble technique that allows studying the mechanical properties of enzymes in their folded state, determining changes in their mechanical behavior as a function of an applied force.20, 21, 22 In this study, we used nano‐rheology to study the mechanical properties of the folded state of BiP protein, trying to determine changes in mechanical properties as consequence of force application and ligand binding (nucleotides and peptide substrate). In particular with this technique, we have studied the viscoelastic nature of BiP mechanics by directly tethering BiP between a gold surface and gold nanoparticles (GNPs), through BiP's two exposed cysteines (cys) residues. With the nano‐rheology setup, we applied a force to an ensemble of molecules, measuring the protein's averaged deformation. Interestingly, our results show that the folded state of the protein behaves like a viscoelastic material, getting softer when it binds nucleotides‐ ATP and ADP‐, but stiffer when binding HTFPAVL peptide substrate. Also, we observed that peptide binding highly improves the affinity for ADP, decreasing it dissociation constant (K _D) around 1000 times, demonstrating allosteric coupling between SBD and NBD domains. To our knowledge, this is the first study focusing on the mechanical properties of a folded state chaperone, considering how changes in its internal viscosity can modulate its function, performed with nano‐rheological determinations.

Results

In order to rule out that cys insertion would cause any alteration in BiP functionality, activity measurements23 and circular dichroism assays were performed for wild type (wt) and mutant cys BiP, before performing nano‐rheology studies (see Supplementary Material). The results suggest that cys insertions affect mildly BiP functionality and structure as cys BiP is active and possess a similar secondary structure to wt (Fig. S2).

Mechanical properties of BiP folded protein

We studied the mechanical properties of an active BiP exploring changes in the viscoelastic behavior of its folded state, measuring protein deformation when subjected to force, in presence of different ligands (nano‐rheology setup Fig. 1). Particularly, we focused on BiP mechanical response upon ADP binding, because it has been described that BiP enters in a high affinity state for its protein substrate when it is bound to ADP. Therefore, this enabled us to shift BiP to a nucleotide‐dependent conformational state in which substrate binding is promoted. Nevertheless, we could not use a fully unfolded protein as a substrate, because any unfolding agent used to extend the substrate protein in bulk (e.g., guanidine chloride, temperature, among others), will also unfold BiP. Therefore, the small and well characterized HTFPAVL peptide18 was utilized as BiP protein substrate.

Figure 1.

BiP nano‐rheology setup. This illustration shows the flow chamber with BiP attached to both gold surfaces, the parallel plates capacitor geometry used for mechanical excitation, and the evanescent wave scattering optics used for read out. BiP was directly tethered between a gold film surface evaporated on a glass slide and 20 nm diameter GNPs, constituting the lower part of a thick flow chamber. To complete it, a gold‐coated cover slip was arranged at 200 µm distance, being the upper part of the chamber, closing it and acting like a capacitor arrangement. BiP attachment proceeds via two exposed cys residues at positions 185 and 537, located in NBD and SBD, respectively. An oscillatory force is exerted over BiP by placing it between two conducting plates and a voltage is applied across them. GNPs are covered with ssDNAs on the surface to negatively charge them.

BiP mechanical response to nucleotides and peptide binding

In absence of ADP, our results show that BiP has a viscoelastic behavior because at low frequencies it assumes viscous properties (dissipation parameters γ have a bigger contribution to the mechanical response than the elastic parameter κ), whereas at high frequencies it acts more elastic. We defined this curve in practical terms as the “base curve”. Thus, the ''base curve'' corresponds to BIP mechanical response only in presence of the experimental buffer, without any ligand.

In order to quantify changes in the mechanical behavior of the protein upon addition of ligands, we fit the experimental amplitude versus frequency data with a simple model of viscoelasticity, Eq. (1), see materials and methods. This form describes the actual experimental curves only approximately, but it has the advantage of summarizing the mechanics of the molecule in just two parameters, κ and γ (see Eq. (1), where ω_c = κ/γ). Because viscoelasticity is a situation of distributed elasticity and viscosity, an increase in either the elastic parameter κ or the viscous parameter γ, or both, signifies an overall “stiffening” of the molecule, while a decrease signifies “softening.” Our procedure is to measure the amplitude versus frequency response for a given sample first in the absence of ligands, and then again in the presence of a ligand. To quantitatively compare the two responses we use the fit Eq. (1), where the parameter F₀, which is the amplitude of the applied force, is kept the same for the two cases. Thus, we quantify this comparison in terms of the change in the parameters κ and γ upon addition of the ligand.

Comparing with the base curve, changes in BiP viscoelastic behavior were observed when ADP was added. The internal dissipation parameter γ and the elastic parameter κ decreased 32% and 25%, respectively. We summarize this observation thus: ADP binding makes BiP less rigid, or in other words, “softer”. As shown in Figure 2, the fitted curve corresponding to the viscoelastic behavior of BiP with ADP “shifted up” as compared with the BiP without ADP fitted curve.

Figure 2.

Amplitude of the response versus forcing frequency for BIP in presence and absence of ADP substrate. Circles represent the mechanical response of the protein in experimental buffer. In the presence of 2 mM ADP, the mechanical properties of the protein change, (triangles). Since the deformation amplitude has increased for the same applied force, the protein has become softer. The dissipation parameter γ and elastic parameter κ decrease up to ∼25% in presence of ADP. Error bars represent the standard deviation of four measurements.

To explore BiP mechanical response dependence on the nucleotide type bound, we added ATP and the same effect obtained with ADP was observed: there is an upper shift in the curve of BiP viscoelastic response in presence of ATP, compared with the base curve (Fig. S3). So when adding ATP, BiP gets softer: we observed a γ and κ decrease, up to 38% and up to 24%, respectively, as compared with the base curve. Then, we assesed BiP mechanical response to the peptide substrate binding. Surprisingly, we observed a “shift down” of the curve corresponding to BiP viscoelastic behavior with bound substrate, as compared with the curves of BiP with either ADP or ATP (Fig. S3). With the bound substrate peptide the internal dissipation parameter γ decreased up to ∼16% and κ increased up to ∼29%, the overall effect making BiP stiffer (Fig. 3).

Figure 3.

Amplitude of the response versus forcing frequency for BIP in presence and absence of peptide and ATP. The initial mechanical response of protein in experimental buffer is represented by the circles. In presence of 70 μM peptide, the protein becomes stiffer, dissipation parameter γ decreases up to ∼16% and elastic parameter κ increases up to ∼29% (diamonds). In presence of 2 mM ATP after removing the peptide, BiP becomes softer (squares), the variation with respect to the original condition showing an increase in γ and κ of about ∼52% and ∼13%, respectively. Finally, when the substrate is removed the viscoelastic behavior returns to the original values (crosses). Error bars represent the standard deviation of four measurements.

Reversibility in BiP mechanical response after nucleotides and substrate binding

We explored the reversibility in the mechanical response of BiP after substrate binding. To achieve this, the substrate peptide was removed by washing the nano‐rheology chamber where BiP was attached. The chamber was washed with 3 mL of experimental buffer (EB), removing only the peptide in solution and not BiP (chamber volume between the gold slides was 20 μL). The results obtained show that after peptide removal BiP returns to a viscoelastic behavior similar to the base curve, evidencing that BiP mechanical response to nucleotides and peptide binding is reversible (γ decrease up to ∼3% and κ increase up to ∼3%, very close to zero; Fig. 3).

As a control, the mechanical response for wt BiP (without exposed cys residues) was explored, in absence and presence of both ATP and ADP, not simultaneously. In this case, the viscoelastic response is rather feature‐less, and comparable to the response of heat‐denatured proteins in the apparatus.24 When nucleotides were added (ATP or ADP), no change in the mechanical response was observed. This is in contraposition with mutant BiP response to nucleotide binding, where a shift in the response curve was shown before (Fig. S4).

K _D's determination for BiP nucleotides

To corroborate that BiP was in a functional folded state when testing its mechanical response to force, K _D's determinations25 were performed for ADP, ATP, and AMP‐PNP in presence or absence of substrate peptide, as shown in Table 1. The K _D determined for ADP without peptide was 0.60 ± 0.19 μM. However, the amplitude response of BiP as a function of ADP concentration in presence of 1.8 μM peptide showed two different behaviors. Therefore, two different K _D's for ADP were obtained in presence of the peptide, applying Eq. (4) to fit the curve. First, we determined a small K _D1 about 0.0026 ± 0.00008 μM, and a bigger second K _D2 about 0.56 ± 0.14 μM similar to the K _D for ADP previously referenced in literature,26 and measured by ourselves without the peptide (Fig. 4). Moreover, we determined a K _D for ATP about 0.96 ± 0.21 μM (as shown in Fig. S5) and K _D for AMP‐PNP of 0.48 ± 0.26 μM, as shown in Figure S6.

Table 1.

Binding Parameters for BiP in Different Conditions

Substrate		K D Experimental Values (µM)	K D Reference Values (µM)
ADP		0.60 ± 0.19	0.29 ± 0.0226
ADP + 1.8 µM peptide	K D1	0.0026 ± 0.00008
	K D2	0.56 ± 0.14
ATP		0.96 ± 0.21	0.20 ± 0.0226
AMP‐PNP		0.48 ± 0.26

Figure 4.

K_D determination for ADP in absence and presence of 1.8 μM peptide substrate using nano‐rheology. (A) Amplitude response of BiP as a function of ADP concentration. The obtained K _D for ADP was 0.60 ± 0.19 μM. (B) Amplitude response of BiP as a function of ADP concentration in presence of 1.8 μM peptide. Two different binding affinities for ADP were observed working at non‐saturating conditions of peptide substrate. We obtained first a K _D1 = 0.0026 ± 0.00008 μM, we assumed that it corresponds to BiP bound to ADP in absence of the peptide, and a second K _D2 = 0.56 ± 0.14 μM corresponding to BiP binding to ADP in presence of the peptide.

Discussion

Wt BiP mechanical response to force

Wt BiP does not have exposed cys residues, so this protein adheres non‐specifically and in multiple orientations to the nano‐rheological fluid chamber. Unlike the mechanical response of mutant BiP, when adding ATP and ADP no changes in the initial viscoelastic behavior (base curve) were observed. This may be due to multiple orientations of the molecules in the chamber, or, more probably, to surface denaturation of the adsorbed molecules.

BiP mechanical response to force upon ADP and ATP binding

When observing the mechanical response of the mutated protein, where a type of orientation of the bound protein is favored (since the exposed cys residues adhere to the plates and to GNPs), it was possible to observe changes in the mechanical response of the protein in presence of different ligands. Specifically, we observed that with bound ADP the protein becomes softer. The reason of this behavior can be explained with previous studies and is related with the ATPase cycle of BiP. When the protein is in the ADP state, the NBD and SBD seem to be in a dynamic distance distribution with a general trend toward domain separation,1, 8 as both domains act more independently because they are making only transient contacts between them. Complementary, in other studies, it has been shown that BiP's hydrophobic linker connecting both domains can act as an allosteric switch,3, 6 and in particular, that in the ADP bound state the linker is more flexibly than in the ATP bound state11, 27 [Fig. 5(C)].

Figure 5.

Model for mechanical response of BiP in presence of different ligands. (A) BiP unbound state. (B) The structure is stiffer in presence of peptide because the lid of BiP is closed, then generating a compact state. (C) In contrast, the structure is softer in two cases: first, in presence of ADP, the domains are separated by the linker elongation. SBD seems to be in a dynamic distance distribution with a general trend toward domain separation. Second, in presence of ATP, the lid is more flexible and the domains are closer leading to an important rigidity decrease.

Also the protein becomes softer in presence of ATP (Fig. S3 and Fig. 3), but in this case the responsible for this behavior might be only BiP's SBD. In the ATP bound state the NBD transmits this status to the SBD, leading to an opening of the substrate binding cavity by increasing SBD flexibility and lid opening,6, 13 carrying to an important rigidity decrease [Fig. 5(C)]. Basically, It has been suggested that nucleotides binding may trigger helices and hydrogen bridges restructuration, stabilizing or destabilizing the protein, modifying its flexibility or its stiffness. Therefore, communication between both binding domains and the linker dynamics may be implied in this mechanical response.3, 6, 13, 14 Also, novel studies have related the changes in protein elasticity with some local unfolding/refolding events called cracking,28, 29 could being another way in which BiP could respond to ligand binding connecting different dynamic states.

BiP mechanical response to force upon HTFPAVL peptide substrate binding

In contrast to the ADP and ATP behavior, BiP was stiffened in presence of HTFPAVL substrate peptide (Fig. 3). This can be explained considering that it has been documented that when peptide binds to the SBD domain, the lid closes, apparently generating a compact state, forming more interactions between the lid and the substrate.1 This would result in a remarkable rigidity increase in BiP [Fig. 5(B)].

K _D's determination for BiP nucleotides and allosterism between SBD and NBD

Finally, when determining K _D for ADP, a similar value from previous study18 was obtained, confirming that the protein was in a folded functional state when nano‐rheological studies were performed under our experimental conditions. Additionally, BiP's affinity for ADP in presence of 1.8 µM HTFPAVL substrate peptide was determined, measuring two different K _D's for the nucleotide. We worked with a non‐saturating peptide concentration (10 times smaller to the dissociation constant previously reported, about 11.6 μM 18), just to observe the behavior of two different species present in solution: BiP bound to ADP with bound peptide, and BiP bound to ADP with unbound peptide, and test if we were able to observe their own different mechanical responses in our experimental conditions: For the higher K _D obtained (K _D2 = 0.56 ± 0.14 µM) we assumed that it corresponds to BiP bound to ADP in absence of the peptide, since the K _D2 is similar to the one we obtained previously in the assay without peptide (K _D = 0.60 ± 0.19 µM). Complementary, the first K _D1 (K _D1 = 0.0026 ± 0.00008 µM) determined was two orders of magnitude smaller than K _D2, and we think it could correspond to BiP´s binding affinity for ADP in presence of the peptide.

Another interesting point to note is the allosteric communication between both BiP structural domains, which it has been described as an important feature that modulates BiP functionality.3, 6 In our group we have studied allosterism in both directions: first, from the NBD to the SBD,15 studying BiP mechanical properties at the single molecule level with optical tweezer manipulation, and now in this work from the SBD to the NBD with ensemble measurements with nano‐rheological experimental setup. This study has shown that there exists a communication between both domains, as when peptide is bound in the SBD, the affinity for ADP bound in the NBD is modified (evidenced by the drop in K _D value), demonstrating that SBD is allosterically coupled to NBD, as described in previous studies.30 This was expectable, considering that it was reported also that peptide binding accelerates BiP's ATPase activity.31 However, it has not been tested before that the mechanical response of BiP upon ligand binding, can modulate the function of both binding domains particularly operating via changes in its internal viscosity.

Considering our results, we believe that these changes in the stiffness and flexibility in BiP could be related with its role, acting as a chaperone or as a molecular motor.

Materials and Methods

Cys site‐directed mutagenesis of BiP

Sequence alignments using Basic Local Alignment Search Tool (BLAST) and ClustalW were performed to determine identity between yeast and murine's BiP aminoacidic sequences. These results enabled us to know the positions in the yeast BiP sequence to insert two exposed cys residues, considering the work performed by Marcinowski et al.,18 in order to attach the protein to the coverslip to perform nano‐rheology experiments. We found that murine BiP V166C and G518C, were equivalent to yeast V185 and G537. Site‐directed mutagenesis of residues V185 and G537 were carried out using the QuikChange II Site‐Directed Mutagenesis kit (Agilent Technologies) using as a template the pMR2623 plasmid carrying the yeast wild type (wt) BiP gene and an annealing temperature of 72°C. Two pairs of oligonucleotides were used to construct the mutants, each pair complementary to opposite strands of the template. The underlined bases indicate the codon for the new amino acid. Only forward oligonucleotides are shown:

V185C: gaagattatttaggcactaag tgc acccatgctgtcgttactgttc

G537C: tgaaggtgtctgccacagataag tgc actggtaaatccgaatctatcac

BiP expression and purification

BiP gene with a N‐terminal (His)₆ tag (without signal sequence) from Saccharomyces cerevisiae was kindly provided by Dr. Jeffrey Brodsky (Pittsburgh University),32 and was overexpressed in the RR1 strain transformed with the expression vector pMR2623. BiP protein was expressed and purified using a 1 mL His Trap HP column prepacked with nickel (General Electric, USA), following the expression and purification protocol of McClellan et al.,19 with some modifications, regarding the sequential washing steps, where 5% glycerol was not added in the fifth step (more details in supplementary material). Protein was eluted with 250 mM imidazole in 50 mM HEPES pH 7.4, 300 mM NaCl, 5 mM mercaptoethanol obtaining an over 95% pure protein (Fig. S7). Protein concentration was determined using Bradford assay with BSA as standard.33

Nano‐rheology experimental setup

Chamber construction

To study BiP mechanical properties upon ligand binding, we employed nano‐rheology experimental setup (Fig. 1). BiP was directly tethered between a gold film surface evaporated on a glass slide and 20 nm diameter GNPs, constituting the lower part of a thick flow chamber. Gold coated slides and coverslips are prepared by evaporating 3 nm Cr followed by 30 nm Au on glass slides and coverslips using an e‐beam evaporator machine (CHA MARK 40). Preparing the flow chamber takes 3 days. On the first day, protein is first diluted to a final concentration of 2 μM in KH₂PO₄ 1M pH 7. This is the optimum pH which minimizes the nonspecific binding of BiP to gold. Then 500 μL of the protein solution is inserted into a bordered area in the slide for overnight at room temperature. On the second day the slide is first rinsed with a large amount of distilled water in order to remove the unbound proteins and the imidazole from the solution where BiP was kept. After that the slide is immersed in a buffer solution containing 0.5 mM TCEP (3,3’,3”‐Phosphanetriyltripropanoic acid from Thermo Fisher Scientific) in KH₂PO₄ 1 M pH 7 for ∼30 min in order to break the dimer bonds and reduce the protein. The slide is then rinsed once again and immersed in 20 nm GNPs (from Nanocs) for ∼80 min. When GNPs were bound the slide turns slightly red. To negatively charge the GNPs, they were immersed in KH₂PO₄ 1 M pH 4 and incubated with 2 μM thiol‐modified single stranded (ss) DNA (/5ThioMC6‐D/AAAAAAAAAAAAAAAAAAAACGCATTCAGGAT), overnight. Thiol‐modification enables the DNA to bind to gold because of the S‐H group. On the third day, first the unbound DNA is washed away with a large amount of distilled water and then the slide is immersed in 500 μL of experimental buffer, containing: 20 mM Trizma Base (Sigma) pH 7; 2 mM MgCl₂ (Sigma); 25 mM KCl (Sigma). Finally to close the chamber, a gold‐coated cover slip was arranged at 200 µm distance, constituting the upper part of the chamber, closing it and acting like a capacitor arrangement (Fig. S1).34 After this a fresh chamber that last 8 h is ready to be used.

Nano‐rheology experimental measurements

Only fresh chambers must be used for nano‐rheology experiments. Applying an alternating current (AC) potential difference between the two gold films, generated a mechanical response in the protein, measured as an averaged deformation (Amplitude).22 BiP was attached to both gold surfaces through two exposed cys residues inserted by site‐directed mutagenesis (as described before), at positions 185 and 537, located in NBD and SBD, respectively. Using this experimental setup, two types of determinations were performed: BiP mechanical response to an applied force in absence and presence of ligands, and nucleotides dissociation constants (K _D) determinations. All the nano‐rheogical determinations were performed using the experimental buffer, working at room temperature. BiP concentration utilized in each experiment was 2 μM.

BiP mechanical response to an applied force

In these experiments, we studied BiP viscoelastic behavior varying the oscillatory force applied over the protein, in absence and presence of different ligands: ATP, ADP, and HTFPAVL substrate peptide.18 To modify the oscillatory force, the frequency was varied in a range between 10 and 200 Hz, fixing the amplitude of the applied voltage. The ligand concentrations used were 2 mM ATP (Sigma), 2 mM ADP (Sigma) and 70 μM HTFPAVL substrate peptide (gently provided by Dr. David S. King, Howard Hughes Medical Institute and University of California at Berkeley; Figs. 3 and 4). Each curve in the graph corresponds to the average of four experimental runs.

To describe BiP mechanical response, the averaged deformation of the protein subjected to the oscillatory force was measured, and compared to the Maxwell model of viscoelasticity:22

$$\left|z\right|=F_0/\gamma \omega \sqrt{1+{\left(\omega /{\omega }_c\right)}^2}$$ (1)

In the context of our experiments, the corresponding deformation amplitude is |z, frequency is ω, ω_C parameter is κ/γ where κ is an elastic parameter (dimensions of force/length) and γ is a dissipation parameter (dimensions of mass/time). Finally, F ₀ is the amplitude of the applied force. In our case in the graphs ω and |z| correspond to frequency and amplitude axis, respectively, and then the data was fitted using the following expression:

$$Amplitude=A/Frequency\sqrt{1+{\left(Frequency/B\right)}^2}$$ (2)

where A = F ₀/γ, then the dissipation parameter γ is proportional to 1/A and B = ω_c = κ/γ ∝ κA, then the elastic parameter κ is proportional to B/A

Nucleotide dissociation constant (K _D) determination

With our experimental setup, the other assay performed was K _D determination for BiP nucleotidic ligands. AMP‐PNP (Roche, a non‐hydrolizable ATP analogue), and ATP (Sigma) K _D determinations were performed both in absence of the HTFPAVL substrate peptide. ADP (Sigma) K _D determinations was performed in absence and presence of the fixed substrate peptide concentration (1.8 μM). To achieve this, force was maintained constant in amplitude and frequency at 10 Hz, and ADP concentration was varied in a range between 0 and 500 μM (Fig. 4). Each point in the graph corresponds to the average of four measurements. For the binding isotherm measurements, the following expression was considered:

$$\left|z\right|=\frac{\alpha }{1+K_D/\left[ligand\right]}$$ (3)

where [ligand] is the ligand concentration used, K _D is a dissociation constant determined, and α is a normalization factor. In each corresponding case, assuming the presence of more than one ligand‐protein species, the following expression was used for fitting the curve:

$$\left|z\right|=\frac{\alpha 1}{1+K_{D1}/\left[ligand\right]}+\frac{\alpha 2}{1+K_{D2}/\left[ligand\right]}$$ (4)

Conclusion

With nano‐rheological setup we were able to measure directly the viscoelastic behavior of BiP protein in the presence of different substrates and perform direct and mechanical binding measurements.

Our study shows that BiP is stiffer in the presence of substrate peptide, and softer upon nucleotide binding. Also, we observed that peptide binding highly improves ADP affinity, demonstrating allosteric coupling between SBD and NBD domains.

Supporting information

Supporting Information