The life of proteins under mechanical force

Jörg Schönfelder ab, Alvaro Alonso-Caballero a, David De Sancho ac and Raul Perez-Jimenez *ac
aCIC nanoGUNE, Tolosa Hibilbidea 76, 20018 San Sebastián, Spain. E-mail:; Tel: (+34) 943 57 4009
bIMDEA Nanosciences, 28049 Madrid, Spain
cIKERBASQUE, Basque Foundation for Science, 48013 Bilbao, Spain

Received 30th November 2017

First published on 23rd February 2018

Although much of our understanding of protein folding comes from studies of isolated protein domains in bulk, in the cellular environment the intervention of external molecular machines is essential during the protein life cycle. During the past decade single molecule force spectroscopy techniques have been extremely useful to deepen our understanding of these interventional molecular processes, as they allow for monitoring and manipulating mechanochemical events in individual protein molecules. Here, we review some of the critical steps in the protein life cycle, starting with the biosynthesis of the nascent polypeptide chain in the ribosome, continuing with the folding supported by chaperones and the translocation into different cell compartments, and ending with proteolysis in the proteasome. Along these steps, proteins experience molecular forces often combined with chemical transformations, affecting their folding and structure, which are measured or mimicked in the laboratory by the application of force with a single molecule apparatus. These mechanochemical reactions can potentially be used as targets for fighting against diseases. Inspired by these insightful experiments, we devise an outlook on the emerging field of mechanopharmacology, which reflects an alternative paradigm for drug design.

image file: c7cs00820a-p1.tif

Jörg Schönfelder

Jörg Schönfelder is a postdoctoral researcher at the CIC nanoGUNE institute. He studied physics (Diploma 2008) at the Friedrich-Schiller University Jena, Germany, and biomedical engineering (BSc 2009) at the Cork Institute of Technology, Cork, Ireland. Afterwards he completed his doctoral studies at the IMDEA Nanoscience Institute in the research group of Prof. Victor Muñoz in Madrid (PhD 2015), before he joined the laboratory of Dr Raúl Pérez-Jiménez at CIC nanoGUNE, San Sebastián, Spain. His current research interests include (i) protein folding using single molecule force spectroscopy and (ii) single cell biophysics combining fluorescence microscopy and force spectroscopy techniques.

image file: c7cs00820a-p2.tif

Alvaro Alonso-Caballero

Alvaro Alonso-Caballero is a post-doctoral researcher who has recently joined Prof. J. M. Fernandez laboratory at Columbia University, New York. He studied biology (BSc 2011) and biophysics (MSc 2012) at the Autonomous University of Madrid, and then he joined Prof. Marisela Velez laboratory at the Institute of Catalysis and Petrochemistry, Madrid (2012). Later on he completed his doctoral studies at CIC nanoGUNE in San Sebastián (PhD 2017) in the Nanobiomechanics group led by Prof. Raúl Pérez-Jiménez. His current research interests include the mechanical study of viral and bacterial adhesion proteins using single-molecule force spectroscopy techniques.

image file: c7cs00820a-p3.tif

David De Sancho

David De Sancho received his PhD from Universidad Complutense in Madrid under the supervision of Prof. Antonio Rey. Then he did postdoctoral work with Prof. Victor Muñoz at the Spanish Research Council and Dr Robert B. Best at the University of Cambridge. In the last few years, David has obtained a “tenure track” research fellowship that he currently holds at CIC nanoGUNE, collaborating with Prof. Pérez-Jiménez's laboratory. David's main research focus is in the study of protein conformational dynamics using computational approaches. His main achievements are in the development of statistical mechanics models of protein folding experiments, the study of internal friction of peptides and proteins, the development of models of molecular kinetics and the analysis and interpretation of single molecule experiments.

image file: c7cs00820a-p4.tif

Raul Perez-Jimenez

Raul Perez-Jimenez received his PhD in 2005 from the University of Granada under the supervision of Prof. Jose M. Sanchez-Ruiz. Then he moved to New York to do postdoctoral training in the Department of Biological Sciences at Columbia University under the supervision of Prof. Julio Fernandez, where he started his research on protein mechanics. In 2013, he was appointed Ikerbasque Professor at CIC nanoGUNE in San Sebastián, Spain, where we leads the Nanobiomechanics group. His main interest focusses on the study of mechanical forces in biological systems. In his lab he also studies how mechanical forces are involved in diseases and disorders, developing research on mechanomedicine and mechanopharmacology.


Proteins are the most complex and versatile macromolecules, responsible for undertaking many of the functions in living organisms.1 Key to this enormous complexity is the ability of proteins to correctly fold into well-defined three dimensional structures, which enable proteins to fulfill their designated functions, ranging from catalysis to transport or signaling. Even in the case of proteins that do not acquire an accurately defined three dimensional structure, protein function is encoded in the amino acid sequence, which is determined by the information stored in the genes. Failures in the correct folding of proteins can provoke protein misfolding and aggregation, which can end in severe neurodegenerative diseases like Alzheimer's and Parkinson's.2 Nature has evolved a response to this problem involving molecular chaperones and other factors that, in the context of the cell, guarantee efficient folding and function.3,4

Since the seminal finding of reversible folding of proteins in vitro over 50 years ago,5 an immense body of work has investigated protein folding with a focus on the biochemical and physical aspects behind this process.6,7 Most of this work has a clear biophysics focus, and the consensus view derived not only from experiments, but also from theory and computation is that proteins fold on a multidimensional energy landscape via the exploration of a myriad of possible pathways leading to the native state.8 However, this view emerging from controlled experiments on single domains in test tubes and simulations of small proteins in water boxes usually ignores elements that are essential in the protein folding process in vivo. In order to investigate the protein folding process in closer connection with how it occurs in the cell cytoplasm, additional chemical and biological components have to be considered, turning the underlying energy landscape in vitro into a much more complex scenario in vivo. In other words, the energy landscape view in vivo must encompass the whole protein folding life cycle from the initial step of protein synthesis in the ribosome to its destination point in the cell and its final disposal in the proteasome system. In the past few years, researchers have increasingly headed in this direction, enabling a comparison of the protein behavior in vitro and in vivo.9,10 Interestingly, it has been found that the kinetics and thermodynamics of proteins can be indeed very similar in vitro and in vivo, like in the reported case of the amyotrophic lateral sclerosis (ALS) associated protein SOD1.11 However, in other cases the surrounding cytoplasm can have a stabilizing or destabilizing effect on the protein depending on its sequence.12

The development of single molecule techniques has made an important contribution in the experimental study of protein folding inside the cell, removing the ensemble averaging which is inherent to measurements in bulk and adding an unprecedented ability to manipulate the individual steps in the folding life cycle. Especially two techniques revealed enormous potential: on one hand single molecule fluorescence spectroscopic techniques, like Förster Resonance Energy Transfer (FRET) or Fluorescence Lifetime Correlation Spectroscopy (FLCS),13 and on the other hand single molecule force spectroscopy (smFS).14 Using single-molecule FRET the folding of proteins was recently investigated inside the cell, whose results indicated a modest change compared to their folding behavior in vitro.15,16 In single molecule fluorescence spectroscopy measurements a chemical denaturant or temperature are often used to measure protein stability, which acts globally to perturb the protein energy landscape. In contrast, in smFS measurements a single protein or a specific reaction can be directly manipulated with a mechanical force with direction and amplitude, enabling unique control in the investigation of the life cycle of proteins17 and single chemical bond reactions.18 It is important to note that although this external perturbation may seem unnatural, in reality mechanical force is conceivably more biologically relevant than using chemical denaturants or temperature changes. Indeed, mechanical force plays a key role in a good number of cellular processes,19 including cell membrane responses to physical stimuli.20 Significant cases are the protein folding/unfolding at the ribosome,21 or supported by mechanical foldases,22 or during the protein disposal process at the proteasome.23 The exertion of a mechanical force on membrane embedded proteins has been recently described to induce conformational changes altering the function of the corresponding protein.24 This force enables, for instance, the regulation of mechanosensing ion channels.20,25 Furthermore, smFS allows probing possible targets of diseases caused by bacterial or viral infections, which often use mechanical changes in surface proteins to invade the cell. This is the case of the long filamentous structures used for infection by pathogenic bacteria called pili,26 or the case of the CD4 surface protein,27 which plays a key role during the infection by the human immunodeficiency virus (HIV). In this way, smFS measurements enable the emergence of the new field of mechanopharmacology. This discipline focuses on the study of biochemical reactions under force that can be potentially mediated by mechano-active compounds that affect the mechanical stability of proteins involved in diseases.

Our review will address these recent developments in smFS to investigate the biological protein life cycle and how chaperones and enzymes change the proteins’ mechanical stability due to single chemical bond reactions. Our main focus is on the context of how biochemical reactions trigger protein folding inside the cytoplasm, whereas the physics of protein folding change just moderately. These single chemical bond reactions play a critical role in the stabilization and folding of proteins in vivo. We end with an outlook on the new field of mechanopharmacology with a focus on infectious diseases.

The start of the protein life cycle in the ribosome

In all living organisms proteins are synthesized in the ribosome, the molecular machine where the life cycle of proteins begins (Fig. 1). This large macromolecular RNA–protein complex accurately translates the mRNA into a polypeptide chain in a process that has been widely studied by biochemical approaches, high resolution crystal structures and single particle electron cryo-electron microscopy.28,29 The large subunit of the ribosome contains the peptidyl-transferase center (PTC) and catalyzes peptide bond formation, while its small subunit mediates the translation of the mRNA into the correct polypeptide chain. During the elongation phase of protein synthesis, nascent polypeptides are elongated from the N to the C terminus direction one amino acid at a time. It is exactly at this stage where the protein folding process starts in vivo (Fig. 1). While in the test tube the protein folds from the full-length polypeptide chain in order to acquire its designated three-dimensional structure, in the ribosome the translated part of the polypeptide chain may already start to fold into elements of secondary and tertiary structures.
image file: c7cs00820a-f1.tif
Fig. 1 The protein life cycle. The protein life cycle is illustrated starting from the nascent protein in the ribosome (PDB code 5UYL) and ending in the protein degradation process. The protein folding process can already start in the ribosome exit channel by the formation of alpha helices or small tertiary elements. After leaving the ribosome the protein may encounter different chaperone assisted pathways depicted in the figure. Ribosome bound Trigger Factor (TF) is shown, which can help the protein to fold (PDB code 1W2B). Assisted folding under cytoplasmic/cytosolic chaperones is depicted in the presence of Hsp70 (PDB code 4PO2). The possible translocation to the membrane pore complex occurring either co-translationally (ribosome binding domain Sec61 PDB code 3J7Q) or post-translationally (SECYEG PDB code 5CH4 and chaperone SecB PDB code 5JTL) is illustrated. The oxidative folding pathway is shown in the case of the SECYEG (PDB code 5CH4) pore complex assisted by the oxidoreductase enzyme DsbA (PDB code 1A2J). One possibility of the protein degradation is depicted with the proteasome 26S (PDB code 1FNT) and the ubiquitinated protein substrate (PDB code 1UBQ).

Since the first observation of the ribosome exit tunnel in the large subunit (Fig. 1), it was speculated that the tunnel could play an active role in protein folding. Its dimensions of ∼100 Å length and 10 to 20 Å width at its narrowest and widest point suggest that just alpha helices or small tertiary elements are able to fold within the tunnel.30 However, small proteins and protein domains (<50 aa) were recently shown to fold already in the exit tunnel.31,32 Interestingly, in these studies a force sensor composed of the translational arrest peptide (AP) from bacterial secretion monitor (SecM) proteins was used, which typically arrests the protein synthesis in the ribosome at the beginning of the polypeptide exit tunnel. SecM is a protein with the capacity of stalling during the translation elongation process in the ribosome, which causes the blocking of the protein synthesis process.30 Just by the exerted pulling force of the folding protein inside the polypeptide tunnel, the AP was able to unblock the ribosome. However, a direct estimation of the mechanical force was not possible in these investigations. Further smFS by Goldman et al. using fluorescence experiments confirmed that the cotranslational folding of the protein Top7 (at the end of the polypeptide exit tunnel) exerts a mechanical force on the nascent protein chain, which is still inside the ribosome.21 The exerted force was estimated to be around 12 pN from measurements with optical tweezers (OT), which is the force at which the folding and unfolding rates of Top7 are equal. Thus, mechanical force seems to play a role at the beginning of the protein folding life cycle at the ribosome, as it is able to regulate elongation to translation by rescuing the ribosome from the stalled state.21 The experimental setup and the proof of principle experiment are shown in Fig. 2a and b. Therein the stalling effect of SecM during translation elongation is shown using calmodulin as a protein substrate and its release using the antibiotic puromycin. In another study, this setup enabled measuring the effect of the electrostatic repulsion of the ribosome on the folding rate of the nascent folding protein T4-lysozyme.33 More recently, the synthesis of a nascent polypeptide protein chain was observed in real time under constant forces applied with OT.34 These measurements enabled correlating the rates of synthesis and cotranslational folding with specific features in the amino acid sequences of proteins, like the presence of positively charged amino acids, proline residues or hydrophobic stretches.

image file: c7cs00820a-f2.tif
Fig. 2 Single molecule techniques used to study the protein life cycle. (a) Experimental setup for optical tweezers experiments. When the nascent chain is transferred to puromycin, the assembly breaks. The structure of calmodulin was obtained from PDB code 1CLL. Reproduced with permission from ref. 21. The American Association for the Advancement of Science, copyright © 2015. (b) Example trace for the restart experiment. After the “hopping” signature of calmodulin is observed (inset) at 7 pN, the force is increased to 20 pN. Red arrow: The tether breaks after ∼3 min at 20 pN. Reproduced with permission from ref. 21. The American Association for the Advancement of Science, copyright © 2015. (c) Schematic of the magnetic tweezers experiments, showing the octamer of protein L tethered between a glass coverslip and the paramagnetic bead. The force is applied by changing the separation of the permanent magnets and the bead. Reproduced with permission from ref. 22. Nature Publishing Group, copyright © 2017. (d) Dynamics of the protein L octamer at three different forces with (red) and without (black) TF. First, the protein is fully unfolded by a fingerprint pulse, where the eight unfolding events are identified as eight length steps. Second, a refolding pulse is set at a lower force (4.3, 7.4 and 11.9 pN, from bottom to top). At 4.3 pN (faint color, lowest length) all domains are able to fold by themselves, leading to a maximum probability of folding. Therefore, TF does not have any significant effect. At 11.9 pN (faint color, highest length), the protein is not able to refold (0 probability of folding) and TF does not affect the probability of folding. In the intermediate force range (7.4 pN, solid colors) TF greatly increases the probability of folding, reflected in a higher number of folded domains. Reproduced with permission from ref. 22. Nature Publishing Group, copyright © 2017. (e) A schematic illustration of the AFM experiment is shown. In order to simplify the sketch just a single I27 domain with cryptic disulfide bonds is shown. The first force is applied with the AFM cantilever, which induces unfolding of the I27 domain where the disulfide bond becomes solvent-exposed. At this point it can be reduced by DsbA, yielding a mixed disulfide complex. Then, the force is relaxed, allowing the substrate to collapse to enable folding and oxidation, and the initial redox states of the enzyme and substrate are recovered. (f) The image shows a representative experimental trace of oxidative folding with DsbA of the eight domain polyprotein I27 substrate. The insets show the unfolding steps of the I27 domains until the disulfide bonds are exposed at a force of 165 pN. Afterwards the corresponding reduction events of the disulfide bonds by DsbA are shown at a force of 75 pN. Then the force is reduced to zero and the substrate is allowed to refold. In the second pulse the amount of correctly folded I27 domains is revealed by the amount of reduction events.

The protein life cycle in the cytoplasm

Depending on the organism, bacteria or eukaryotic cells, and on the type of the synthesized protein including the sequence and functional role, right after leaving the ribosome the nascent folded protein may encounter different pathways (Fig. 1). We focus here on four possible pathways that have been successfully studied using smFS measurements, namely folding assisted by ribosome-bound chaperones, folding under cytoplasmic chaperones, secretory protein folding assisted by chaperones and oxidative folding. The advantage of applying smFS experiments is that the protein substrate can be manipulated directly and held in the unfolded or folded state, and under these conditions the interaction with chaperones or other binding partners can be detected in real time at a single bond level.

(1) Folding assisted by ribosome-bound chaperones

At the end of the polypeptide exit tunnel, ribosome binding chaperones can interact with the nascent polypeptide in order to assist the correct folding of the protein and also to prevent protein aggregation. This occurs for proteins that are too large for folding autonomously after protein synthesis in the ribosome. The chaperone in action in bacteria is the Trigger Factor (TF), while in eukaryotic cells its equivalents are the ribosome-associated complex (RAC) and the nascent-chain-associated complex (NAC).

The Trigger Factor (TF) has been the one receiving most attention in experimental studies with smFS. These include the work of Mashaghi et al. using OT35 and by Haldar et al. using magnetic tweezers (MT)22 (Fig. 2c and d). Mashaghi et al.35 concentrated on the effect of TF on the maltose binding domain protein (MBP) substrate, either in the single domain structure or as a polyprotein with four domains. With the single MBP, TF was found to have a stabilizing effect as it binds partially folded structures of MBP at zero force. In other words, TF reshapes the protein energy landscape of the protein substrate, actively guiding it to the folded structure. A remarkable effect of TF on a misfolded substrate was shown when the MBP4 polyprotein, a substrate prone to aggregation, was used. After successive folding/unfolding cycles of MBP4, the individual MBP domains were found in the majority in their native folded states. In the second work by Haldar et al.,22 the chaperone activity kinetics of TF were measured in correspondence to an applied mechanical force on a polyprotein chain substrate made of eight B1 domains of protein L (Fig. 2b). The activity of TF on the unfolded substrate was found to be force dependent, suggesting a link of force with the biological function of the chaperone. In addition, the activity of TF to refold the polyprotein L8 was at its highest under a force of 7.4 pN, while at zero force TF stabilizes most of the unfolded or partially folded states of protein L.

(2) Folding under cytoplasmic chaperones

When in the cytoplasm, proteins that have not folded autonomously often interact with chaperones like those from the heat shock protein (Hsp) family. The members of this family have been classified by their molecular weight (Hsp40, Hsp60, Hsp70, Hsp90, Hsp100 and the small Hsps). The first chaperone system assisting folding is called DnaK/DnaJ in bacteria and Hsp70/Hsp40 in eukaryotes, and their function is to help in the completion of the folding of the protein in vivo. Afterwards folding can be still accomplished by the so-called chaperonins (GroEL/GroES in bacteria and TRiC in eukaryotes) or with further Hsp systems, like Hsp90 in eukaryotes and HtpG in bacteria. Different chaperone systems either interact directly with each other or use specific adapter proteins for the protein substrate exchange.

Using atomic force microscopy (AFM), Nunes et al. studied the effect of the DnaJ–DnaK–GrpE on a polyprotein made of eight titin I27 domains, a system that is prone to aggregate.36 DnaJ was found to decrease misfolding, and Nunes et al. found an optimal DnaJ/DnaK ratio that prevents the misfolding of I27 almost entirely. Furthermore, when the nucleotide exchange factor GrpE was added, a further decrease of the unfolding state of I27 was observed. The broad functional plasticity of the single homologue chaperone DnaK from the Hsp70 system on the maltose binding protein (MBP) polypeptide substrate was revealed in much detail using OT.37 First, it was found that DnaK can bind to the near native structure of MBP and stabilize it against forced unfolding. Secondly, DnaK was found to act also on the unfolded form of MBP, preventing it from refolding. Furthermore, DnaK was also shown to stabilize partially folded structures of MBP, suggesting its role in suppressing protein aggregation. Additionally, important structural elements like the helical lid and the binding groove of the chaperone were identified. This was achieved by constructing structurally different mutants of DnaK, which consecutively changed the described stabilization effects.

The function of stabilizing and preventing protein aggregates of the small Hsp (sHsp) chaperones (Hsp42 and Hsp26) has also been investigated recently in detail in another study using OT38 using the maltose binding protein (MBP) as a substrate. Here Hsp42 was shown to decrease the unfolding force of single MBP and to prevent agglomeration of the polyprotein made of four MBP domains.

(3) Secretory protein folding assisted by chaperones

Around 30% of all proteins need to be translocated to their corresponding extracytoplasmic locations, mainly the periplasm in bacteria and the endoplasmic reticulum (ER) in eukaryotes.39 This is the case of membrane embedded, secretory or cell surface/plasma membrane proteins. The secretory (SEC) pathway is finely tuned in order to guarantee the arrival of the protein to the designated location. Therein, several obstacles must be overcome, such as avoiding cytoplasmic folding and maintaining the protein in an unfolded/loosely folded but soluble state. Furthermore, they have to be exactly targeted to export channels in order to reach the trans side of the membrane. Afterwards, they need to be sorted depending on their final destination, and finally they have to fold. Typically, nascent unfolded polypeptides are targeted via a hydrophobic amino acid sequence tag (signal peptides) and via a single carboxyl-terminal signal sequence in order to reach a heterotrimeric complex of membrane proteins used for protein translocation through the membrane. This is the SecYEG complex in bacteria and the Sec61 complex in eukaryotes.40,41 The translocation to the corresponding membrane pore complex can occur co-translationally (protein synthesis at the membrane) or post-translationally (with the help of chaperones, see Fig. 1). In bacteria, there are many different post-translational pathways, where export specific chaperones like SecA and SecB or general chaperones like TF help in first sorting and then targeting the nascent polypeptide chain in the unfolded state to the SecYEG complex.42 After trespassing the SecYEG pore the signal peptides are cleaved from the nascent polypeptide chain and several further chaperones and oxidoreductase enzymes like DsbA promote protein folding. In eukaryotes, chaperones like Hsp70, Hsp40 and ASNA1 help the nascent polypeptide chain to target the Sec61 complex post-translationally.39 After trespassing, Sec61 complex chaperones like PDI help further in the folding of the protein.

The interaction of chaperone SecB with the maltose binding protein (MBP) polypeptide before it reaches the SecYEG channel has been investigated with OT.43 It was shown that SecB binds to the unfolded MBP core and keeps it in the unfolded state. Furthermore, when the core of MBP was not unfolded, SecB could not bind to MBP. Hence, SecB was shown to prevent the formation of stable tertiary interactions, keeping MBP in a molten globule like state in order to retain its extended conformation during translocation through the SecYEG channel. Similar results were found for the immunoglobulin binding protein (BiP), which plays an important role in assisting protein translocation into the endoplasmic reticulum (ER).44 In this case, OT experiments discovered that BiP binds to the unfolded state of its substrate protein MJ0366, which is therefore prevented from refolding.

The effect of periplasmic chaperones Skp and SurA on the folding of outer transmembrane proteins (Omps) (in particular, the β-barrel ferric hydroxamate uptake receptor FhuA) in Gram-negative bacteria was recently investigated by smFS using AFM.45 In these experiments, multiple full length FhuAs were reconstituted in lipid membranes from E. coli. Single FhuAs were then unfolded with the AFM tip and the effect of the presence of the periplasmic chaperones was estimated. SurA and Skp have been found to act differently on the FhuA substrate. While Skp maintains FhuA mostly in the unfolded state, SurA favors equally its unfolded and folded states and this balance is shifted to the folded state when the interaction time with the mechanically unfolded FhuA substrate increases. However, when Skp and SurA were used together, FhuA was mostly kept in the unfolded state after mechanical unfolding.

(4) Oxidative folding

Forming a disulfide bond involves connecting two cysteine residues. This involves that the connecting cysteine residues must find each other in the unfolded state and hence change the covalent structure of the protein before it folds into its three-dimensional structure. Most of the disulfide bond carrying proteins are secretory and extracellular proteins. Oxidative folding describes the process of inserting disulfide bonds helping the protein to fold, which typically occurs after translocation through the corresponding membrane pore into an oxidative environment. This task is fulfilled by an enzymatic system made of chaperones and catalysts inside the eukaryotic endoplasmic reticulum and in the bacterial periplasm. The oxidoreductase enzyme DsbA46 in bacteria and the protein disulfide isomerase (PDI)47 in eukaryotes are the main catalysts of the oxidative folding of proteins (Fig. 1). Other famous oxidoreductases exist like thioredoxin (TRX) and glutaredoxin, forming altogether a ubiquitous family of enzymes that catalyze thiol–disulfide exchange.48

The two main oxidoreductase chaperones PDI49 and DsbA50 have been investigated using AFM. In the case of PDI, a polyprotein chain of eight repeats of the titin I27 domain was used as a substrate, with each of the I27 domains containing a disulfide bond. By unfolding the I27 substrate until it exposed its disulfide bond to the surrounding buffer, the reduced PDI was able to react with the I27 substrate to create a mixed-disulfide between PDI and I27. The chemical reaction between PDI and the I27 substrate can break the cryptic disulfide of the I27 domain, which results in the total unfolding of the I27 substrate. After reduction, the I27 substrate was led to refold again. An additional unfolding step could then clarify if a disulfide bond in the I27 substrate was formed or not. With this experimental setup, the attachment and release kinetics of the PDI I27 substrate were estimated. Furthermore, the oxidation and folding kinetics of the I27 substrate were compared to the presence of PDI. It was shown how reduced and oxidized PDI catalyzed disulfide formation in a folding protein. Therein PDI revealed a passive effect in oxidative folding, leaving the folding rate of the I27 substrate unaffected.49 Interestingly, also non-native inter-domain disulfide formation could be observed, which corresponds to a different length in unfolding the I27 domain substrate and therefore precludes proper folding.

In the case of DsbA, a very similar kinetic behavior to that of PDI was detected using the same eight repeat I27 polyprotein substrate and experimental approach50 (Fig. 2e and f). This approach is especially significant in the case of DsbA since proteins that are exported to the periplasm need to be translocated through a channel such as the SECYeg translocon, in an unfolded extended state. Once in the periplasm, the protein folds in the presence of DsbA, which assists in the oxidation of disulfide bonds needed for proper folding. Thus, the AFM experiments resemble this process by first extending the protein and triggering disulfide bond reduction and then refolding the protein under oxidative conditions against DsbA. DsbA oxidised its substrate slightly more rapidly than PDI and fewer non-native disulfide bonds were detected. In summary, DsbA was found to follow the same passive placeholder model as in the case of PDI but is better in preventing non-native disulfides. This higher efficiency may be related to the fact that many of the proteins that are folded in the periplasm are crucial for the survival of the bacterium and no errors can be permitted. Among these proteins we find the Fim domains (FimA, FimG, FimF and FimH) that make the pilus type-1 in E. coli.51 The pilus is essential for the motility and adhesion of bacteria. All the Fim domains are disulfide bonded and the lack of disulfides prevents proper adhesion of the bacterium, thus impairing the pathogenicity of the strain.52

The termination of the protein life cycle in the cytoplasm

The protein life cycle ends primarily in either the ubiquitin–proteasome system (UPS) or the autophagy–lysosome pathway (ALP), which removes misfolded proteins and aggregates.53 Ubiquitination of proteins is used in both pathways as a signal for degradation. Short lived, misfolded and damaged proteins are primarily degenerated by the UPS pathway, while the ALP recognizes larger protein aggregates and potentially dangerous cellular components. E3 ubiquitin ligase CHIP plays a key role in ubiquitinating chaperone-bound protein substrates whose refolding has failed. This stimulates the UPS pathway, which ends in the degradation of the ubiquitinated substrate in the proteasome. However, as the UPS and the ALP mainly degrade free proteins in the proteasome, a third pathway also exists, which is the so-called ATP-dependent AAA proteases responsible for degrading membrane proteins.54

Fascinating smFS investigations exist in which the whole degradation pathway of a protein substrate within an AAA family protease molecular machine (ClpXP) could be followed at a single molecule level using OT.23,55 These molecular machines were of special interest, as it was suggested that the AAA proteases could provide a pulling force on the membrane proteins which need to be degraded. Within the ClpXP protease machine, the ClpX recognizes specific short peptide sequences, unfolds the protein substrate and translocates into the chamber of ClpP. In the work of Aubin-Tam23 the mechanics of the unfolding of a filamin polyprotein substrate and its translocation could be tracked and investigated (Fig. 3). It was found that ClpXP unfolding was less than 1 ms, while translocation cost more time. Interestingly, the dwell time between consecutive unfolding events varied. This indicates that a successful unfolding of an entire or partial domain with one ATP-fuelled power stroke of the ClpX is stochastic and that several power strokes are often needed. Furthermore, it has been estimated that ClpXP performs at least 5 kBT of work per translocation step. Independent work conducted in parallel by Maillard et al.55 also showed that ClpXP indeed applies a mechanical force to unfold the protein substrate, in this case GFP. This also confirmed the described ATP-fuelled power stroke mechanism of ClpX. This experimental setup has now enabled investigating, e.g., the mechanochemical coupling between the ATP hydrolysis mechanism of ClpXP and its energetic conversion into mechanical force.56 Further OT experiments with ClpXP using a polyprotein substrate supported a stochastic mechanism of ClpXP protein denaturation independent of the substrate stabilities.57

image file: c7cs00820a-f3.tif
Fig. 3 Protein degradation within the ClpXP protease machine. (a) Cartoon of the ClpXP protease midway through degradation of a multidomain substrate. The substrate is threaded through the pore of the ClpX hexamer (yellow-brown). Loops in the pore grip the substrate, and downward loop movements, powered by ATP binding and hydrolysis, unfold each domain and translocate the denatured polypeptide into the lumen of ClpP (green), where degradation to small fragments occurs. On the right a representation of an optical double-trap in a passive force-clamp geometry is shown. ClpXP was attached to one polystyrene bead, a multidomain substrate was tethered to a second bead by double-stranded DNA, and connectivity between the beads was maintained by interactions between ClpXP and the substrate. Reproduced with permission from ref. 23. Cell Press, copyright © 2011. (b) Representative extension-versus-time traces of unfolding and translocation events, accompanying ClpXP degradation of the multidomain substrate at different forces using 2 mM ATP. Traces are offset on the y axis to avoid overlap. Gray lines represent raw data following decimation; colored lines represent averages over a 10-point sliding window. Red segments represent FLN domains, blue segments represent the HaloTag domain, and green segments represent translocation of the 47-residue linker between the HaloTag domain and FLN1. Dwells between FLN subunits unfolding in the top 8 pN trace were spaced at 4 nm (inset), which is the length of a folded FLN domain. No traces contained eight FLN unfolding events, presumably because C-terminal FLN domains were unfolded and translocated before measurements began. Reproduced with permission from ref. 23. Cell Press, copyright © 2011.


Apart from studying biological processes like the protein folding cycle inside the cell, smFS is also an ideal experimental setup for investigating single bond chemistry reactions using, for example, the disulfide bond as a substrate for a covalent bond.18 During the last decade, the chemical kinetics of disulfide bond cleavage within the protein substrate under different forces and with different reducing agents like small nucleophiles (1,4-DL-dithiothreitol (DTT),58 tris(2-carboxyethyl)phosphine (TCEP),59 hydroxide/hydrosulfide anions,60,61L-cysteine,62N-acetylcysteine (NAC),63 methylamine or histamine64) or small enzymes (thioredoxin (TRX)65 or glutaredoxin66) were investigated using AFM. In the following, we will summarize the main findings in investigating single chemical bond reactions; deeper insight can be found in an excellent recent review by Garcia-Manyes.67

In a typical mechanochemical experiment the disulfide bond containing protein substrate is unfolded with a mechanical force until it exposes its disulfide bond to the surrounding buffer. Afterwards the buffer containing chemical reducing agents cleaves the disulfide bond and the corresponding reduction rate can be measured under different forces and under different concentrations. These experiments allow parameters like the bond elongation parameter Δxr to be estimated, which describes the transition state of disulfide bond reduction. Herein, disulfide bond reduction using small nucleophiles shows a single exponential dependence on the applied mechanical force.68 There are nevertheless a few exceptions, where the reducing agent shows a biphasic force dependency like in the case of the enzyme TRX.65 TRX shows two separate pathways of disulfide bond reduction, depending on the applied force. This was explained by the different binding configurations of TRX onto the disulfide bond containing protein substrate. Using hydroxide anions,60 it has been found that a force-activated reactivity switch exists. At this point the dependence between the reduction rate and applied force changed. It was suggested that at a certain applied force the conformation of the disulfide bond alters (from a cis to trans conformation), which leads to a different energy landscape and hence a different bond elongation parameter Δxr. Not only disulfide bond reduction, but also disulfide isomerization could be measured using AFM. In this experimental setup a polyprotein substrate made of titin I27 domains carrying two disulfide bonds was unfolded in the presence of the reducing agent L-cysteine.62 As a result, a complex reduction network was found, which could be explained by intramolecular isomerization of new disulfide bonds. AFM experiments also enable the investigation of the oxidative folding of the disulfide bond and hence its reformation. In this context, the reactivity of cysteine sulfenic acid (Cys-SOH), which is the first product of disulfide bond oxidation, has been followed at high pH in the absence of catalytic enzymes to investigate the principle of disulfide reformation.69 Interestingly, a higher refolding rate of the disulfide bond containing I27 domains including the polyprotein substrate was found when sulfenylated. Therefore the reformation of the disulfide bond drastically increases the folding efficiency. A further investigation showed how smFS allows predicting whether small nucleophiles are capable of triggering non-enzymatic oxidative protein folding63 depending on their thermodynamic reactivity. Different kinds of chemical bonds were also investigated by applying a mechanical force on the so-called metalloproteins, the ion-binding protein rubredoxin70 and the blue-copper proteins azurin and plastocyanin.71 Both experiments displayed a complex releasing behavior of the metal from the corresponding active site of the protein and the suitability of smFS experiments to investigate metal–ligand bonds.

Apart from mechanochemical studies involving disulfide bond formation and cleavage, the impact of disulfide bonds on the overall mechanical stability of a protein has been discussed. Interestingly, the presence of a disulfide bond can affect the mechanical unfolding of a protein in different ways as it was investigated in comparing the oxidized and reduced variants of FimG and titin I91, the new nomenclature of the former I27 domain.72 Depending on the orientation and surrounding of the disulfide bond within the folded protein, the mechanical stability of the protein can increase (FimG) or decrease (titin I91). In a recent investigation on the evolution of disulfide bonds within a small eight domain size fragment (I65–I72) of the giant muscle protein titin using the AFM revealed a correlation between the amount of disulfide bonds inside the titin domains and its mechanical stability.73 By bringing back to life the corresponding titin fragments from four extinct species using ancestral sequence reconstruction (ASR), it could be shown that ancient titin fragments contained more disulfide bonds than the corresponding titin fragments in modern amniote species (Fig. 4).

image file: c7cs00820a-f4.tif
Fig. 4 Mechanochemistry of disulfide bond reduction using smFS. (a) Schematics of the experiment. Disulfide bonded domains (red) show a two-step unfolding pattern. The first step (∼12 nm) corresponds to the unfolding of the beta sheets that are not trapped in the disulfide bond, while the second step (∼15 nm) shows the unfolding of the rest of the protein after the reduction of the disulfide bond cause by thioredoxin. Note that non-disulfide bonded domains have a single-step (∼27 nm) unfolding pattern that represents the stretching of the whole domain. Reproduced and adapted with permission from ref. 73. Nature Publishing Group, copyright © 2017. (b) Above an experimental force-clamp trace of LSCA titin is shown. First, a pulse of force of 135 pN during 2 s is applied that triggers unfolding of non-disulfide-bonded domains (inset arrows) and disulfide-bonded domains up to their disulfide bond (asterisks in the inset). The disulfide bonds can be reduced by Trx enzymes present at 10 μM concentration. Three reduction events are monitored at a force of 80 pN (green line). An experimental force-clamp trace for human titin is shown below, which indicates one reduction event. Reproduced and adapted with permission from ref. 73. Nature Publishing Group, copyright © 2017.

A different biochemical aspect which can be followed by smFS experiments is the protein–ligand interaction, as it has been observed that this process can alter the mechanical stability of a pure protein substrate.74 Many examples have been described by which a protein–ligand interaction increases the mechanical stability of the protein. Experiments with the small protein GB1 showed an increase in its unfolding force when bound to IgG antibody fragments,75 and the combination together with nickel ions had an additive effect on the enhancement of protein mechanical stability.76 In the case of calmodulin (CaM) the presence of a target peptide called mastropan (Mas) slowed down its unfolding rate, whereas Ca2+ ions were found to increase its folding rate and its unfolding force detected with equilibrium smFS measurements using AFM.77 Both ligands were shown to act completely different on the CaM energy landscape. How the direction of the applied force affects the protein–ligand interaction between the Maltose Binding Protein (MBP) and its substrate maltose has been described by smFS measurements using AFM.78 When a mechanical force is applied in the direction of the direct interaction between MBP and maltose, the mechanical stability of MBP increases upon substrate binding. Hence, a mechanical stabilizing effect of protein–ligand binding is only expected if the ligand stabilizes the protein along the chosen pulling direction. Using mutagenesis, the key residues of the protein–substrate interaction could be found. In a recent work the mechanical stability regulation upon nucleotide binding of the nucleotide binding domain (NBD) of chaperone DnaK was detected using OT79 and insertion of structural mutants. In this study MgADP/MgATP was found not to change the general unfolding force of the NBD, but its unfolding pattern and therefore its relative mechanical stability. A sophisticated smFS OT experimental setup was used to reveal the effect of inhibitor binding on the conformational dynamics of the protein substrate of adenylate kinase in the presence of ATP.80 A different type of chemical bond involving the so-called zinc-finger structural motif containing four Zn–S bonds (ZnS4) was measured in a recent work, where the complex unfolding pathway of chaperone DnaJ was investigated using AFM.81 Additionally, the binding to hydrophobic short peptides APPY revealed an increase in the mechanical stability of Domain I of DnaJ.

In addition to investigating the change in mechanical stability introduced by protein ligand binding, smFS experiments can offer the possibility of determining in detail the corresponding binding energies as has been shown in a recent study for nucleic acids and peptides.82 Therein, using OT the ligand-binding interaction of the restriction enzyme endonuclease Eco RI to a 30-base pair DNA hairpin with the included Eco RI recognition site was chosen as proof of principle. As the mechanical stability increases upon binding, bound and unfolded states could be distinguished using far from equilibrium pulling experiments. Using the fluctuation theorem83 for ligand binding, it is then possible to directly determine the binding free energy from pulling experiments. Furthermore, the dependence of binding energies on the concentration of the ligand could be followed, confirming the law of mass action on a single molecule level. The general power of the method was later on confirmed for different ligand binding systems to investigate selectivity, allosteric effects and misfolding, using the ligand echinomycin, which can bind specifically and unspecifically to a DNA hairpin. Here the binding energies for each mode could be calculated and compared.

Future perspective on mechanochemistry in drug design

In the past few years, the question of how physical forces affect cells and tissues and how these forces are related to diseases and disorders has attracted the interest of scientists. Cells mechanically sense their microenvironment and respond to the mechanical information with changes at the molecular level, resulting in processes known as mechanotransduction84 and mechanosensing.20 Here smFS has also been instrumental in this context, in particular in the case of the extracellular matrix protein talin.85 When talin is stretched, it activates the binding to the protein vinculin and therefore starts the process of force transduction. Many human diseases originate as a consequence of alterations in the mechanotransduction of cells and tissues and the alteration of the mechanical properties of the effector proteins responsible for these processes. Changes in the composition, stiffness and architecture of the extracellular matrix are found in cancer tissues, increasing the proliferation and survival of tumor cells. Other diseases in which these changes are relevant are fibrosis, asthma, atherosclerosis, and heart disease, among others.86–88 Several examples have been demonstrated recently where the mechanical and chemical regulation has been studied at the single-molecule level on proteins known to be related to several diseases, like neurodegenerative (Alzheimer's, Parkinson's),2 viral or bacterial pathologies.27,89 As a result, the possibility has emerged to rationally design drugs capable of altering the mechanical behavior of proteins involved in these pathologies, developing the burgeoning field of mechanopharmacology.

As has been mentioned before, ligand–protein binding such as antibody binding or metal chelation alters the mechanical stability of the target protein.74 A similar trend was observed for dihydrofolate reductase (DHFR), an essential enzyme whose inhibition blocks DNA synthesis and leads to cell death. Methotrexate is an anticancer drug that strongly binds to DHFR, inducing conformational changes and affecting its thermodynamic stability. SmFS experiments conducted on DHFR in the presence of this drug showed an increment in the mechanical unfolding force of ∼60 pN. This enzyme has been used as a model for studying protein translocation through mitochondrial pores and degradation on ATP-fueled proteasomes. These smFS experiments demonstrated that ligand binding mechanically stabilized DHFR, increasing the force required to unfold this enzyme, thus hindering both its translocation and degradation.90 Ligand binding could also be interesting for treating diseases where protein aggregation leads to neurodegenerative pathologies. An interesting case is the development of diseases such as bovine spongiform encephalopathy resulting from the aggregation of the prion protein PrP. Gupta et al.91 studied with smFS OT measurements the ligand binding effect of the anti-prion pharmacological chaperone Fe-TMPyP on PrP, demonstrating that this molecule stabilized the PrP native state mechanically and kinetically, increasing the unfolding energy barrier due to native state stabilization. Besides this role in the native state, Fe-TMPyP bound to the misfolded states of PrP. Therefore it prevents aggregation seeding for both correctly and incorrectly folded PrP, thus giving a chance to the misfolded states to reach the native state. In another study, the mechanical stability of other misfolding or aggregation prone proteins like Protein S from the βγ-crystallin protein family was shown to be altered by calcium binding.92 Misfolding or aggregation of Protein S is related to pathologies like cataracts and melanoma. In the absence of calcium the unfolding force of the N-terminal domain of Protein S follows a bimodal distribution, showing two conformations of different stabilities. In the presence of calcium, the lower mechanically stable conformation disappears and only the stronger one is registered, which is less prone to protein aggregation.

The stiffening of proteins upon ligand binding could also be used against the adhesion structures utilized by bacteria to colonize human tissues, such as the urinary or the gastrointestinal tract. These bacterial adhesive structures made of hundreds of proteins linearly arranged are called fimbriae or pili. Here, mechanopharmacological targeting of these structures could be an alternative strategy to the increasingly less effective current antibiotic treatments. Weakening these structures or impairing their biomechanical properties would help in preventing bacterial infections. This premise was demonstrated on the uncoiling–coiling ability of enterotoxigenic E. coli (ETEC) CS20 fimbriae. It is known that the passive oral administration of anti-CS20 antibodies prevents ETEC diarrhea; however, the specific mechanism of their action was unknown. Using force spectroscopy, Singh et al.93 demonstrated that antibody binding increased the force required to uncoil CS20 fimbriae, which limits its extension under force. This extra mechanical load required to uncoil the fimbriae results in a higher force transmitted to the tip-end bond between the fimbriae and the host epithelium, influencing the durability of this interaction. But besides affecting the uncoiling cycle, the recoiling ability of the fimbriae was also impaired due to the antibody binding, preventing the correct stacking of fimbrial proteins to reform the helical configuration.93 A further work made on the same ETEC model but with the CFA/1 and CS2 fimbriae revealed again that anti-fimbrial antibody binding damages their mechanical resilience, probably by clamping stacked layers from the helical configuration and also linking pairs of fimbriae through the action of bivalent antibody binding.94 Unlike the Gram negative bacterium E. coli, where the pilus structure is held through non-covalent interactions between the proteins and the pilus adopts a helical quaternary configuration, in Gram positive bacteria like diphtheria etiological agent Corynebacterium diphtheriae the pilus proteins are linked through covalent isopeptide bonds and the pilus lacks a higher order organization.95 Echelman et al.89 demonstrated the recovery of the mechanical stability of the C. diphtheriae pilus shaft protein SpaA in the presence of calcium after mechanical unfolding. It has been proposed that the presence of a coordinated calcium ion in the structure of SpaA is suggested to be important for the mechanical stability of the protein.96 Once again ligand–protein binding proves to be a crucial factor in strengthening bacterial adhesive organelles, and is therefore a promising target for designing drug analogs able to disrupt these coordination places.

A possible biomechanical target for mechanopharmaceutical treatment found during viral infection is the attachment of Human Immunodeficiency Virus (HIV) to CD4+ lymphocyte T cells. CD4 is a cell-surface receptor made of four extracellular domains (D1–D4), one transmembrane domain, and a cytoplasmic domain. This receptor is involved in the adaptive immune response and is used by HIV-surface glycoprotein gp120 for attachment to the cell as a first step leading to infection.97 After the binding of gp120 to the D1 domain of CD4, it is suggested that certain flexibility in the extracellular domains of CD4 enables the virus to approach the cell membrane in order to locate a chemokine co-receptor, an interaction required for downstream infection events.98 One previous study used CD4 variants of different lengths showing that HIV infectivity was positively correlated with the length of the receptor.99 In a study performed by us, smFS AFM experiments were carried out with the first two domains of CD4 (CD4D1D2) in order to characterize the nanomechanics of these two most distal domains.27 Therein the existence of a hierarchical unfolding behavior of CD4D1D2 was demonstrated, where D2 unfolded before D1 although D2 showed higher mechanical stability. The D2 domain protects the D1 domain from mechanical unfolding and only after D2 extension D1 can be extended, due to the destabilization transmitted through a shared interdomain β-strand. The study proved that CD4D1D2 hierarchical mechanical unfolding takes place also at low constant forces (20 pN) in the 15–20 s time window. This low force regime resembles the expected forces that a CD4-bound viral particle could exert by Brownian motion,100 cellular uptake dynamics,101 viral surface movements, and cell–cell interactions.102 This could cause a mechanical extension of the CD4 domains, providing higher flexibility. Hence, this extra flexibility could potentially provide some benefits to the virus. For instance, a flexible linker could facilitate co-receptor binding onto the surface. Also, this flexibility could increase the ability of the gp120 trimer to bind up to three CD4 molecules, as it has been suggested that binding multiple CD4 enhances infectivity.103 From a mechanical point of view, unfolding of domains in multimeric proteins acts as a shock absorber, allowing the virus to remain attached.104

In the same study the effect of Ibalizumab, a broadly neutralizing antibody with a potent inhibitory effect on HIV infection,99,105 on the nanomechanics of CD4D1D2 was tested. Force experiments showed a remarkable stabilization effect of the CD4D1D2 tandem with Ibalizumab, which increased their mechanical stability and therefore delayed the extension of CD4D1D2 (see Fig. 5a–c). Hence CD4 stiffening could be related to the inhibitory effect of Ibalizumab and therefore reveals a mechanopharmacological target, which might prevent the encounter of the HIV with the coreceptor. Furthermore, this study also tested the mechanochemical regulation of the CD4D1D2 disulfide bond by the thioredoxin enzyme. During HIV infection several oxidoreductase enzymes regulate the oxidation state of disulfide bonds present both in CD4 domains and in the viral glycoprotein gp120. Apparently, the thioredoxin-induced reduction of the CD4D2 domain disulfide bond seems to facilitate infection.106–108 Here, disulfide bond enzymatic reduction and mechanical extension of the domains provide an extra length to CD4, which therefore increases the probability of co-receptor binding to HIV (see Fig. 5d). The smFS experiments showed that CD4D1D2 disulfide bond reduction took place upon mechanical unfolding, indicating that these buried disulfides are only solvent exposed and accessible to thioredoxin after mechanical extension. This suggested that the HIV bound particle could induce D2 unfolding, exposing its disulfide for enzymatic reduction.27 Hence, the mechanopharmacological treatment could be oriented to stiffen CD4 and to prevent its disulfide bond reduction. Furthermore, the mechanical targeting of CD4 appears to be a less disruptive and a more precise strategy, because interfering with thioredoxin could have more pernicious effects as it regulates not only CD4.108

image file: c7cs00820a-f5.tif
Fig. 5 Mechanical stability modulation of CD4 with Ibalizumab using smFS. (a) Schematic representation of an AFM experiment of the polyprotein (I91)2-CD4D1D2-(I91)2 (with I91 being the new nomenclature of the former I27 domain of titin). The I91 modules are used as molecular fingerprints. The polyprotein is attached to a gold-covered coverslide in one end and a cantilever tip in the other end. On the right a typical AFM force-ramp trace at a speed of 33 pN s−1 is shown in the upper panel, indicating the unfolding of CD4D1D2 before the titin I91 domains. Below the corresponding length a histogram for the unfolding of CD4D1D2 is given. Two dominant peaks at 8.3 ± 1.1 nm (red) for CD4D1 and 13.7 ± 1 nm (blue) for CD4D2 are observed. A less prominent peak at 22.5 ± 1.0 nm (green) corresponds to the simultaneous unfolding of CD4D1 and CD4D1 (n = 72). Reproduced and adapted with permission from ref. 27. American Chemical Society, copyright © 2014. (b) Schematic representation of an AFM experiment of the polyprotein (I91)2-CD4D1D2-(I91)2 in the presence and binding of Ibalizumab. On the right a force-ramp trace of the (I91)2-CD4D1D2-(I91)2 polyprotein in the presence of Ibalizumab is illustrated. The mechanical extension of CD4D1D2 is detected after all I91 domains at a force of ∼250 pN. Histogram of step size for the unfolding of CD4D1D2 in force-ramp mode and in the presence of Ibalizumab (n = 76). The unfolding of CD4D1 and that of CD4D2 are measured at 8.6 ± 0.7 nm (red) and 13.3 ± 1.1 nm (blue), respectively. We monitored a dominant peak at around 22.2 ± 0.9 nm (green) corresponding to the simultaneous unfolding of CD4D1 and CD4D2. Reproduced and adapted with permission from ref. 27. American Chemical Society, copyright © 2014. (c) Histogram of the initial unfolding force of CD4D1D2 and in the absence (red, n = 37) and presence (blue, n = 54) of Ibalizumab. A displacement toward a higher force is observed when Ibalizumab is present. Reproduced and adapted with permission from ref. 27. American Chemical Society, copyright © 2014. (d) A schematic pathway of HIV infection is depicted. The first picture shows the CD4 cell surface receptor in detail with its 4 domains D1–D4 (PDB code 1WIP). The inset shows the disulfide bond in the D2 domain. Then during the first step the HIV binds to the CD4 receptor with its surface glycoprotein gp120 (figure was obtained by combining PDB codes 3J70 and 1WIP). Then depending on the presence and binding of Ibalizumab (PDB code 3O2D) to the CD4 cell receptor, the HIV might be able to unfold the D1D2 domain. At stage three the disulfide bond of the unraveled D2 domain can be reduced by thioredoxin (PDB code 1ERT), enabling higher flexibility to CD4, helping the HIV particle to bind the co-receptor CCR5 (PDB code 4MBS). This higher flexibility can potentially help the gp120 trimer to bind another CD4 molecule or even to act as a shock absorber to prevent viral detachment. All these are plausible scenarios that would occur in the presence of mechanical force.


The newest results from smFS experiments on the protein folding, translocation and denaturation processes during the protein life cycle highlight that mechanical force plays a pivotal role in vivo. In addition, they indicate that biochemical reactions between the protein substrate and chaperones or small molecules are key mechanisms for the cell in order to regulate the protein life cycle. Further experiments should go the next step in mimicking the cell environment. Here, single molecule fluorescence experiments have been already begun to go into this direction in measuring the effect of crowding on the protein15 and RNA109 folding inside the cell. Current developments to perform smFS experiments in the cell are recently reviewed by Norregaard et al.,110 which offer very promising perspectives. Nevertheless, it must be noted that we are still far from reaching, in the context of the cell, the resolution that smFS techniques provide. Therefore, conducting smFS measurements in vitro is still a necessary prerequisite before performing experiments in the cellular environment.

Mechanochemistry, the study of biochemical processes under mechanical force, has been revolutionized as smFS experimental setups allow the observation of chemical bond reactions at the single molecule level. Furthermore, it presents the prerequisite for mechanopharmacology, a new research perspective which uses the results from ligand–protein binding investigations and predictions in order to control the mechanical stability of proteins associated with diseases, such as surface proteins used by viral and bacterial pathogens. The presented smFS studies of the HIV antibody Ibalizumab27 and the anti-prion pharmacological chaperone Fe-TMPyP on PrP91 pave the way towards new methods, where drugs for both pathogenic and non-pathogenic diseases could be approached under a mechanochemical viewpoint. The mechanopharmacology idea may develop as the future strategy for the designing of fine-tuned drugs able to alter the mechanical stability of proteins involved in human diseases.

Conflicts of interest

There are no conflicts to declare.


A. A.-C. was funded by the predoctoral program of the Basque Government. We acknowledge financial support from the Spanish Ministry of Economy, Industry and Competitiveness grant BIO2016-77390-R to R. P.-J. and CTQ2015-65320-R to D. D. S.; and Marie Curie Career Integration Grants (CIG) FP7-PEOPLE-2013-CIG from the European Commission to R. P.-J. This work was also supported by the Spanish Ministry of Economy, Industry and Competitiveness under the Maria de Maeztu Units of Excellence Programme – MDM-2016-0618.


  1. B. Alberts, A. Johnson, J. Lewis, D. Morgan, M. Raff, K. Roberts and P. Walter, Molecular Biology of the Cell, Sixth Edition, Taylor & Francis Group, 2014 Search PubMed.
  2. C. M. Dobson, Nature, 2003, 426, 884–890 CrossRef CAS PubMed.
  3. D. Balchin, M. Hayer-Hartl and F. U. Hartl, Science, 2016, 353, aac4354–aac4354 CrossRef PubMed.
  4. F. U. Hartl, Annu. Rev. Biochem., 2017, 86, 21–26 CrossRef CAS PubMed.
  5. C. B. Anfinsen, Science, 1973, 181, 223–230 CAS.
  6. M. Oliveberg and P. G. Wolynes, Q. Rev. Biophys., 2005, 38, 245 CrossRef CAS PubMed.
  7. A. R. Fersht, Nat. Rev. Mol. Cell Biol., 2008, 9, 650–654 CrossRef CAS PubMed.
  8. J. D. Bryngelson, J. N. Onuchic, N. D. Socci and P. G. Wolynes, Proteins, 1995, 21, 167–195 CrossRef CAS PubMed.
  9. A. Gershenson and L. M. Gierasch, Curr. Opin. Struct. Biol., 2011, 21, 32–41 CrossRef CAS PubMed.
  10. A. J. Wirth and M. Gruebele, Bioessays, 2013, 35, 984–993 CrossRef CAS PubMed.
  11. J. Danielsson and M. Oliveberg, Curr. Opin. Struct. Biol., 2017, 42, 129–135 CrossRef CAS PubMed.
  12. S. Mittal, R. K. Chowhan and L. R. Singh, Biochim. Biophys. Acta, 2015, 1850, 1822–1831 CrossRef CAS PubMed.
  13. P. R. Banerjee and A. A. Deniz, Chem. Soc. Rev., 2014, 43, 1172–1188 RSC.
  14. J. Schönfelder, D. De Sancho and R. Perez-Jimenez, J. Mol. Biol., 2016, 428, 4245–4257 CrossRef PubMed.
  15. I. König, A. Zarrine-Afsar, M. Aznauryan, A. Soranno, B. Wunderlich, F. Dingfelder, J. C. Stüber, A. Plückthun, D. Nettels and B. Schuler, Nat. Methods, 2015, 12, 773–779 CrossRef PubMed.
  16. H. Gelman, A. J. Wirth and M. Gruebele, Biochemistry, 2016, 55, 1968–1976 CrossRef CAS PubMed.
  17. C. J. Bustamante, C. M. Kaiser, R. A. Maillard, D. H. Goldman and C. A. M. Wilson, Annu. Rev. Biophys., 2014, 43, 119–140 CrossRef PubMed.
  18. J. Liang and J. M. Fernández, ACS Nano, 2009, 3, 1628–1645 CrossRef CAS PubMed.
  19. Y. Javadi, J. M. Fernandez and R. Perez-Jimenez, Physiology, 2013, 28, 9–17 CrossRef CAS PubMed.
  20. A. Anishkin, S. H. Loukin, J. Teng and C. Kung, Proc. Natl. Acad. Sci. U. S. A., 2014, 111, 7898–7905 CrossRef CAS PubMed.
  21. D. H. Goldman, C. M. Kaiser, A. Milin, M. Righini, I. Tinoco and C. Bustamante, Science, 2015, 348, 457–460 CrossRef CAS PubMed.
  22. S. Haldar, R. Tapia-Rojo, E. C. Eckels, J. Valle-Orero and J. M. Fernandez, Nat. Commun., 2017, 8, 668 CrossRef PubMed.
  23. M.-E. Aubin-Tam, A. O. Olivares, R. T. Sauer, T. A. Baker and M. J. Lang, Cell, 2011, 145, 257–267 CrossRef CAS PubMed.
  24. C. Pliotas and J. H. Naismith, Curr. Opin. Struct. Biol., 2017, 45, 59–66 CrossRef CAS PubMed.
  25. J. Teng, S. Loukin, A. Anishkin and C. Kung, Pflugers Arch., 2015, 467, 27–37 CrossRef CAS PubMed.
  26. J. Alegre-Cebollada, C. L. Badilla and J. M. Fernández, J. Biol. Chem., 2010, 285, 11235–11242 CrossRef CAS PubMed.
  27. R. Perez-Jimenez, A. Alonso-Caballero, R. Berkovich, D. Franco, M. W. Chen, P. Richard, C. L. Badilla and J. M. Fernandez, ACS Nano, 2014, 8, 10313–10320 CrossRef CAS PubMed.
  28. T. A. Steitz, Nat. Rev. Mol. Cell Biol., 2008, 9, 242–253 CrossRef CAS PubMed.
  29. N. Fischer, A. L. Konevega, W. Wintermeyer, M. V. Rodnina and H. Stark, Nature, 2010, 466, 329–333 CrossRef CAS PubMed.
  30. D. N. Wilson and R. Beckmann, Curr. Opin. Struct. Biol., 2011, 21, 274–282 CrossRef CAS PubMed.
  31. O. B. Nilsson, R. Hedman, J. Marino, S. Wickles, L. Bischoff, M. Johansson, A. Müller-Lucks, F. Trovato, J. D. Puglisi, E. P. O'Brien, R. Beckmann and G. von Heijne, Cell Rep., 2015, 12, 1533–1540 CrossRef CAS PubMed.
  32. J. Marino, G. Von Heijne and R. Beckmann, FEBS Lett., 2016, 590, 655–660 CrossRef CAS PubMed.
  33. C. M. Kaiser, D. H. Goldman, J. D. Chodera, I. Tinoco and C. Bustamante, Science, 2011, 334, 1723–1727 CrossRef CAS PubMed.
  34. F. Wruck, A. Katranidis, K. H. Nierhaus, G. Büldt and M. Hegner, Proc. Natl. Acad. Sci. U. S. A., 2017, 114, E4399–E4407 CrossRef CAS PubMed.
  35. A. Mashaghi, G. Kramer, P. Bechtluft, B. Zachmann-Brand, A. J. M. Driessen, B. Bukau and S. J. Tans, Nature, 2013, 500, 98–101 CrossRef CAS PubMed.
  36. J. M. Nunes, M. Mayer-Hartl, F. U. Hartl and D. J. Müller, Nat. Commun., 2015, 6, 6307 CrossRef CAS PubMed.
  37. A. Mashaghi, S. Bezrukavnikov, D. P. Minde, A. S. Wentink, R. Kityk, B. Zachmann-Brand, M. P. Mayer, G. Kramer, B. Bukau and S. J. Tans, Nature, 2016, 539, 448–451 CrossRef CAS PubMed.
  38. S. Ungelenk, F. Moayed, C.-T. Ho, T. Grousl, A. Scharf, A. Mashaghi, S. Tans, M. P. Mayer, A. Mogk and B. Bukau, Nat. Commun., 2016, 7, 13673 CrossRef PubMed.
  39. B. C. S. Cross, I. Sinning, J. Luirink and S. High, Nat. Rev. Mol. Cell Biol., 2009, 10, 255–264 CrossRef CAS PubMed.
  40. B. van de Berg, W. M. Clemons, I. Collinson, Y. Modis, E. Hartmann, S. C. Harrison and T. A. Rapoport, Nature, 2004, 427, 36–44 CrossRef PubMed.
  41. S. Shao and R. S. Hegde, Annu. Rev. Cell Dev. Biol., 2011, 27, 25–56 CrossRef CAS PubMed.
  42. A. Tsirigotaki, J. De Geyter, N. Šoštaric, A. Economou and S. Karamanou, Nat. Rev. Microbiol., 2016, 15, 21–36 CrossRef PubMed.
  43. P. Bechtluft, R. G. van Leeuwen, M. Tyreman, D. Tomkiewicz, N. Nouwen, H. L. Tepper, A. J. Driessen and S. J. Tans, Science, 2007, 318, 1458–1461 CrossRef CAS PubMed.
  44. M. P. Ramírez, M. Rivera, D. Quiroga-Roger, A. Bustamante, M. Vega, M. Baez, E. M. Puchner and C. A. M. Wilson, Protein Sci., 2017, 26, 1404–1412 CrossRef PubMed.
  45. J. Thoma, B. M. Burmann, S. Hiller and D. J. Müller, Nat. Struct. Mol. Biol., 2015, 22, 795–802 CAS.
  46. H. Kadokura, F. Katzen and J. Beckwith, Annu. Rev. Biochem., 2003, 72, 111–135 CrossRef CAS PubMed.
  47. B. Wilkinson and H. F. Gilbert, Biochim. Biophys. Acta, 2004, 1699, 35–44 CrossRef CAS.
  48. J. L. Martin, Structure, 1995, 3, 245–250 CrossRef CAS PubMed.
  49. P. Kosuri, J. Alegre-Cebollada, J. Feng, A. Kaplan, A. Inglés-Prieto, C. L. Badilla, B. R. Stockwell, J. M. Sanchez-Ruiz, A. Holmgren and J. M. Fernández, Cell, 2012, 151, 794–806 CrossRef CAS PubMed.
  50. T. B. Kahn, J. M. Fernández and R. Perez-Jimenez, J. Biol. Chem., 2015, 290, 14518–14527 CrossRef CAS PubMed.
  51. S. Geibel and G. Waksman, Biochim. Biophys. Acta, Mol. Cell Res., 2014, 1843, 1559–1567 CrossRef CAS PubMed.
  52. F. Jacob-Dubuisson, J. Pinkner, Z. Xu, R. Striker, A. Padmanhaban and S. J. Hultgren, Proc. Natl. Acad. Sci. U. S. A., 1994, 91, 11552–11556 CrossRef CAS.
  53. I. Dikic, Annu. Rev. Biochem., 2017, 86, 193–224 CrossRef CAS PubMed.
  54. T. Langer, Trends Biochem. Sci., 2000, 25, 247–251 CrossRef CAS PubMed.
  55. R. A. Maillard, G. Chistol, M. Sen, M. Righini, J. Tan, C. M. Kaiser, C. Hodges, A. Martin and C. Bustamante, Cell, 2011, 145, 459–469 CrossRef CAS PubMed.
  56. M. Sen, R. A. Maillard, K. Nyquist, P. Rodriguez-Aliaga, S. Pressé, A. Martin and C. Bustamante, Cell, 2013, 155, 636–646 CrossRef CAS PubMed.
  57. J. C. Cordova, A. O. Olivares, Y. Shin, B. M. Stinson, S. Calmat, K. R. Schmitz, M. E. Aubin-Tam, T. A. Baker, M. J. Lang and R. T. Sauer, Cell, 2014, 158, 647–658 CrossRef CAS PubMed.
  58. A. P. Wiita, S. R. K. Ainavarapu, H. H. Huang and J. M. Fernandez, Proc. Natl. Acad. Sci. U. S. A., 2006, 103, 7222–7227 CrossRef CAS PubMed.
  59. S. Rama, K. Ainavarapu, A. P. Wiita, L. Dougan, E. Uggerud, J. M. Fernandez, S. R. Koti Ainavarapu, A. P. Wiita, L. Dougan, E. Uggerud and J. M. Fernandez, J. Am. Chem. Soc., 2008, 130, 6479–6487 CrossRef PubMed.
  60. S. Garcia-Manyes, J. Liang, R. Szoszkiewicz, T.-L. Kuo and J. M. Fernández, Nat. Chem., 2009, 1, 236–242 CrossRef CAS PubMed.
  61. J. Liang and J. M. Fernández, J. Am. Chem. Soc., 2011, 133, 3528–3534 CrossRef CAS PubMed.
  62. J. Alegre-Cebollada, P. Kosuri, J. A. Rivas-Pardo and J. M. Fernández, Nat. Chem., 2011, 3, 882–887 CrossRef CAS PubMed.
  63. A. E. M. Beedle, M. Mora, S. Lynham, G. Stirnemann and S. Garcia-Manyes, Nat. Commun., 2017, 8, 15658 CrossRef CAS PubMed.
  64. D. J. Echelman, A. Q. Lee and J. M. Fernández, J. Biol. Chem., 2017, 292, 8988–8997 CrossRef CAS PubMed.
  65. A. P. Wiita, R. Perez-Jimenez, K. A. Walther, F. Gräter, B. J. Berne, A. Holmgren, J. M. Sanchez-Ruiz and J. M. Fernandez, Nature, 2007, 450, 124–127 CrossRef CAS PubMed.
  66. J. Alegre-Cebollada, P. Kosuri, D. Giganti, E. Eckels, J. A. Rivas-Pardo, N. Hamdani, C. M. Warren, R. J. Solaro, W. A. Linke and J. M. Fernández, Cell, 2014, 156, 1235–1246 CrossRef CAS PubMed.
  67. S. Garcia-Manyes and A. E. M. Beedle, Nat. Rev. Chem., 2017, 1, 0083 CrossRef.
  68. S. Garcia-Manyes, T.-L. Kuo and J. M. Fernández, J. Am. Chem. Soc., 2011, 133, 3104–3113 CrossRef CAS PubMed.
  69. A. E. M. Beedle, S. Lynham and S. Garcia-Manyes, Nat. Commun., 2016, 7, 12490 CrossRef CAS PubMed.
  70. P. Zheng, S.-i. J. Takayama, A. G. Mauk and H. Li, J. Am. Chem. Soc., 2013, 135, 7992–8000 CrossRef CAS PubMed.
  71. A. E. M. Beedle, A. Lezamiz, G. Stirnemann and S. Garcia-Manyes, Nat. Commun., 2015, 6, 7894 CrossRef CAS PubMed.
  72. A. Manteca, Á. Alonso-Caballero, M. Fertin, S. Poly, D. De Sancho and R. Perez-Jimenez, J. Biol. Chem., 2017, 292, 13374–13380 CrossRef CAS PubMed.
  73. A. Manteca, J. Schonfelder, A. Alonso-Caballero, M. J. Fertin, N. Barruetabena, B. F. Faria, E. Herrero-Galan, J. Alegre-Cebollada, D. De Sancho and R. Perez-Jimenez, Nat. Struct. Mol. Biol., 2017, 24, 652–657 CAS.
  74. X. Hu and H. Li, FEBS Lett., 2014, 588, 3613–3620 CrossRef CAS PubMed.
  75. Y. Cao, T. Yoo, S. Zhuang and H. Li, J. Mol. Biol., 2008, 378, 1132–1141 CrossRef CAS PubMed.
  76. Y. Cao, Y. D. Li and H. Li, Biophys. J., 2011, 100, 1794–1799 CrossRef CAS PubMed.
  77. J. P. Junker, F. Ziegler and M. Rief, Science, 2009, 323, 633–637 CrossRef CAS PubMed.
  78. M. Bertz and M. Rief, J. Mol. Biol., 2009, 393, 1097–1105 CrossRef CAS PubMed.
  79. D. Bauer, D. R. Merz, B. Pelz, K. E. Theisen, G. Yacyshyn, D. Mokranjac, R. I. Dima, M. Rief and G. Žoldák, Proc. Natl. Acad. Sci. U. S. A., 2015, 112, 10389–10394 CrossRef CAS PubMed.
  80. B. Pelz, G. Žoldák, F. Zeller, M. Zacharias and M. Rief, Nat. Commun., 2016, 7, 10848 CrossRef CAS PubMed.
  81. J. Perales-Calvo, A. Lezamiz and S. Garcia-Manyes, J. Phys. Chem. Lett., 2015, 6, 3335–3340 CrossRef CAS PubMed.
  82. J. Camunas-Soler, A. Alemany and F. Ritort, Science, 2017, 355, 412–415 CrossRef CAS PubMed.
  83. I. Junier, A. Mossa, M. Manosas and F. Ritort, Phys. Rev. Lett., 2009, 102, 070602 CrossRef PubMed.
  84. T. Iskratsch, H. Wolfenson and M. P. Sheetz, Nat. Rev. Mol. Cell Biol., 2014, 15, 825 CrossRef CAS PubMed.
  85. A. del Rio, R. Perez-Jimenez, R. Liu, P. Roca-Cusachs, J. M. Fernandez and M. P. Sheetz, Science, 2009, 323, 638–641 CrossRef CAS PubMed.
  86. K. A. Jansen, D. M. Donato, H. E. Balcioglu, T. Schmidt, E. H. J. Danen and G. H. Koenderink, Biochim. Biophys. Acta, Mol. Cell Res., 2015, 1853, 3043–3052 CrossRef CAS PubMed.
  87. D. Ingber, Ann. Med., 2003, 35, 564–577 CrossRef PubMed.
  88. A. W. M. Haining, T. J. Lieberthal and A. del Rio Hernandez, FASEB J., 2016, 30, 2073–2085 CrossRef CAS PubMed.
  89. D. J. Echelman, J. Alegre-Cebollada, C. L. Badilla, C. Chang, H. Ton-That and J. M. Fernández, Proc. Natl. Acad. Sci. U. S. A., 2016, 113, 2490–2495 CrossRef CAS PubMed.
  90. S. R. K. Ainavarapu, L. Li, C. L. Badilla and J. M. Fernandez, Biophys. J., 2005, 89, 3337–3344 CrossRef CAS PubMed.
  91. A. N. Gupta, K. Neupane, N. Rezajooei, L. M. Cortez, V. L. Sim and M. T. Woodside, Nat. Commun., 2016, 7, 12058 CrossRef CAS PubMed.
  92. Z. N. Scholl, Q. Li, W. Yang and P. E. Marszalek, J. Biol. Chem., 2016, 291, 18263–18275 CrossRef CAS PubMed.
  93. B. Singh, N. Mortezaei, B. E. Uhlin, S. J. Savarino, E. Bullitt and M. Andersson, Sci. Rep., 2015, 5, 13678 CrossRef CAS PubMed.
  94. B. Singh, N. Mortezaei, S. J. Savarino, B. E. Uhlin, E. Bullitt and M. Andersson, J. Bacteriol., 2017, 199, e00665-16 CrossRef PubMed.
  95. A. P. A. Hendrickx, J. M. Budzik, S.-Y. Oh and O. Schneewind, Nat. Rev. Microbiol., 2011, 9, 166–176 CrossRef CAS PubMed.
  96. H. J. Kang, N. G. Paterson, A. H. Gaspar, H. Ton-That and E. N. Baker, Proc. Natl. Acad. Sci. U. S. A., 2009, 106, 16967–16971 CrossRef CAS PubMed.
  97. R. J. Harris, S. M. Chamow, T. J. Gregory and M. W. Spellman, Eur. J. Biochem., 1990, 188, 291–300 CrossRef CAS PubMed.
  98. Ashish, I. J. Juncadella, R. Garg, C. D. Boone, J. Anguita and J. K. Krueger, J. Biol. Chem., 2008, 283, 2761–2772 CrossRef CAS PubMed.
  99. M. M. Freeman, M. S. Seaman, S. Rits-Volloch, X. Hong, C. Y. Kao, D. D. Ho and B. Chen, Structure, 2010, 18, 1632–1641 CrossRef CAS PubMed.
  100. T. J. English and D. A. Hammer, Biophys. J., 2004, 86, 3359–3372 CrossRef CAS PubMed.
  101. K. Welsher and H. Yang, Nat. Nanotechnol., 2014, 9, 198–203 CrossRef CAS PubMed.
  102. C. J. Burckhardt and U. F. Greber, PLoS Pathog., 2009, 5, e1000621 Search PubMed.
  103. P. D. Kwong, M. L. Doyle, D. J. Casper, C. Cicala, S. A. Leavitt, S. Majeed, T. D. Steenbeke, M. Venturi, I. Chaiken, M. Fung, H. Katinger, P. W. I. H. Parren, J. Robinson, D. Van Ryk, L. Wang, D. R. Burton, E. Freire, R. Wyatt, J. Sodroski, W. A. Hendrickson and J. Arthos, Nature, 2002, 420, 678–682 CrossRef CAS PubMed.
  104. A. F. Oberhauser, P. E. Marszalek, H. P. Erickson and J. M. Fernandez, Nature, 1998, 393, 181–185 CrossRef CAS PubMed.
  105. A. A. Haqqani and J. C. Tilton, Antiviral Res., 2013, 98, 158–170 CrossRef CAS PubMed.
  106. N. Cerutti, M. Killick, V. Jugnarain, M. Papathanasopoulos and A. Capovilla, J. Biol. Chem., 2014, 289, 10455–10465 CrossRef CAS PubMed.
  107. A. Gallina, T. M. Hanley, R. Mandel, M. Trahey, C. C. Broder, G. A. Viglianti and H. J. P. Ryser, J. Biol. Chem., 2002, 277, 50579–50588 CrossRef CAS PubMed.
  108. T. S. Stantchev, M. Paciga, C. R. Lankford, F. Schwartzkopff, C. C. Broder and K. A. Clouse, Retrovirology, 2012, 9, 97 CrossRef CAS PubMed.
  109. M. Gao, D. Gnutt, A. Orban, B. Appel, F. Righetti, R. Winter, F. Narberhaus, S. Müller and S. Ebbinghaus, Angew. Chem., Int. Ed., 2016, 55, 3224–3228 CrossRef CAS PubMed.
  110. K. Norregaard, R. Metzler, C. M. Ritter, K. Berg-Sørensen and L. B. Oddershede, Chem. Rev., 2017, 117, 4342–4375 CrossRef CAS PubMed.

This journal is © The Royal Society of Chemistry 2018