Skip to main content
NCBI home page
NIHPA Author Manuscripts logoLink to NIHPA Author Manuscripts
. Author manuscript; available in PMC: 2012 May 1.
Published in final edited form as: Mol Microbiol. 2011 Mar 28;80(3):588–595. doi: 10.1111/j.1365-2958.2011.07614.x

The unfolding story of anthrax toxin translocation

Katie L Thoren 1, Bryan A Krantz 1,2,3
PMCID: PMC3094749  NIHMSID: NIHMS286998  PMID: 21443527

Summary

The essential cellular functions of secretion and protein degradation require a molecular machine to unfold and translocate proteins either across a membrane or into a proteolytic complex. Protein translocation is also critical for microbial pathogenesis, namely bacteria can use translocase channels to deliver toxic proteins into a target cell. Anthrax toxin (Atx), a key virulence factor secreted by Bacillus anthracis, provides a robust biophysical model to characterize transmembrane protein translocation. Atx is comprised of three proteins: the translocase component, protective antigen (PA); and two enzyme components, lethal factor (LF) and edema factor (EF). Atx forms an active holotoxin complex containing a ring-shaped PA oligomer bound to multiple copies of LF and EF. These complexes are endocytosed into mammalian host cells, where PA forms a protein-conducting translocase channel. The proton motive force unfolds and translocates LF and EF through the channel. Recent structure and function studies have shown that LF unfolds during translocation in a force-dependent manner via a series of metastable intermediates. Polypeptide-binding clamps located throughout the PA channel catalyze substrate unfolding and translocation by stabilizing unfolding intermediates through the formation of a series of interactions with various chemical groups and α-helical structure presented by the unfolding polypeptide during translocation.

Keywords: protein translocation, protein unfolding, Brownian ratchet, nonspecific binding, proton gradient

Introduction

Transmembrane protein translocases, protein degradation machinery, disaggregases, bacterial toxins and various forms of bacterial secretion apparatuses are often comprised of multiprotein complexes that catalyze a series of unfolding and translocation reactions. These reactions are central to many cellular processes, including protein trafficking, organelle biogenesis, protein quality control, cell-cycle regulation, and microbial pathogenesis processes, such as toxin internalization (Fig. 1A). Generally, folded proteins are thermodynamically stable under normal cellular conditions; therefore, energy-consuming molecular machines are required to unfold, translocate and further process these protein substrates (Fig. 1B). The processes of protein unfolding and translocation pose numerous challenges for the translocase machine (Fig. 1C). Investigating how translocases overcome these challenges touches upon exciting questions in structural and molecular biology. How are the available driving forces harnessed? How does a molecular machine mechanically unfold different multi-domain substrates? How does the machine process and interact with unfolded polypeptide chains, which not only possess combinatorially complex chemistry at the residue level but also occupy a large configurational space? How are the counterproductive diffusive forces mitigated and/or harnessed by the transporter during translocation? In this Microreview, we address these questions by describing the major advances in understanding protein translocation using the anthrax toxin model system.

Figure 1. Protein translocation challenges.

Figure 1

Open in a new tab

A. A possible scheme for Atx assembly and entry into host cells. Proteolytically-activated PA monomers oligomerize into ring-shaped heptameric or octameric pre-channels, which can bind 3 or 4 LF/EF substrates, respectively. These complexes are endocytosed. Upon acidification of the endosome, the PA pre-channel converts to the channel state, ultimately allowing LF and EF to translocate into the cytosol. B. A possible protein unfolding and translocation pathway for Atx depicted in three successive steps: docking, protein unfolding, and translocation of the unfolded chain. C. Challenges encountered during translocation. During translocation substrates are mechanically unfolded by the driving force; the mechanical resistance, however, can vary significantly depending on the relative orientation of the substrate to the force vector. Combinatorial chemical complexity arises as the unfolded chain presents a wide array of changing combinations of side chain chemistries to the channel. Conformational heterogeneity is also present in the unfolded substrate polypeptides. Combinatorial chemical complexity and conformational heterogeneity present significant challenges for substrate recognition. Finally, counterproductive diffusive forces must also be overcome during translocation.

Anthrax toxin

Anthrax toxin (Atx), reviewed extensively elsewhere (Young & Collier, 2007), is a three-protein virulence factor secreted by B. anthracis. The three protein components include the 83-kDa protective antigen (PA) and two ~90-kDa enzymatic factors, called lethal factor (LF) and edema factor (EF). In order for Atx to function, it must assemble into an active holotoxin complex, which contains multiple copies of LF and EF bound to a ring-shaped PA oligomer, called the PA pre-channel (Fig. 1A). Once assembled, the PA-LF/EF complex is endocytosed by the host cell, and transferred to an acidic compartment. Under these acidic conditions, the PA pre-channel undergoes a conformational change and inserts into the membrane, forming a cation-selective channel. Using the proton gradient (ΔpH) that develops across the endosomal membrane, the PA channel unfolds and translocates LF and EF into the cytosol of the host cell (Krantz et al., 2005, Krantz et al., 2006, Thoren et al., 2009). Once inside the cytosol, LF (a zinc-metalloprotease) and EF (a Ca2+- and calmodulin-activated adenylcyclase) disrupt a variety of cell-signaling pathways, manifesting ultimately in general immune system dysfunction and potentially death.

Structure and assembly of the PA translocase

Recent structural studies on the soluble PA pre-channel and the membrane-inserted channel state have provided insight into the mechanism of toxin assembly and translocation. While it was previously believed that PA forms a uniform population of heptameric oligomers (Milne et al., 1994, Petosa et al., 1997, Mogridge et al., 2002, Lacy et al., 2004, Katayama et al., 2008), Kintzer et al. have shown that PA assembly is quite complex, and PA forms mixtures of heptamers and octamers in solution and on cell surfaces (Kintzer et al., 2009). Detailed characterization of the two purified oligomers demonstrated that the octameric form is more stable than the heptameric form under neutral pH and physiological temperature (37 °C); the heptameric form is more prone to insolubility, extensive aggregation, and inactivation than its octameric counterpart since it more readily converts to the channel form at higher pH and temperatures (Kintzer et al., 2010a). This difference in stability may result from increases in surface area burial and intradomain interactions within PA (Kintzer et al., 2009), as determined from the crystal structures of the octamer (Kintzer et al., 2009) and heptamer (Lacy et al., 2004, Petosa et al., 1997). This stability difference, however, is not observed when either PA oligomer is bound to the extracellular domain of its host cell receptor, and the receptor domain effectively stabilizes heptameric and octameric PA through a similar mechanism (Kintzer et al., 2010b, Lacy et al., 2004). Aside from its potential importance in understanding anthrax pathogenesis, the novel octameric architecture has been successfully exploited in structural studies due to its increased thermostability and higher level of symmetry (Kintzer et al., 2009, Feld et al., 2010).

The structure of the heptameric PA channel was recently imaged using electron microscopy (EM) (Katayama et al., 2008). The study found that the heptameric PA channel is mushroom-shaped and approximately 170 Å long, and 125 Å wide at its maximum dimensions. The wider, cap-shaped part of the structure is about 70 Å long and likely contains the LF/EF binding sites. The thinner, stem-shaped part of the structure is 100 Å long and inserts into and spans the membrane bilayer. The channel's stem can, in fact, insert into a model membrane bilayer disk (Katayama et al., 2010). Overall this EM model corresponds with earlier work that suggested the stem is an extended β-barrel structure (Benson et al., 1998, Nassi et al., 2002). From these studies and other modeling studies (Krantz et al., 2004), it is inferred that the β barrel stem is unable to accommodate structures larger than an ~10 to15 Å-wide α helix, which means that LF and EF must unfold during translocation.

Translocation models

The broad applicability of the translocation problem in numerous biological systems has led to the proposal of several general translocation models. In many of these systems, the substrate must be first unfolded and then transported as an unfolded chain across a membrane or into a proteolytic complex. These unfolding and translocation steps require some sort of chemo-mechanical coupling of an energy source to the physical unwinding of the substrate polypeptide, namely through ATP hydrolysis or dissipation of a chemical gradient, such as a proton gradient. Two types of mechanisms have emerged for protein translocases: (i) an active pushing/pulling mechanism and (ii) a passive Brownian-ratchet mechanism.

In the active pushing/pulling mechanism (Glick, 1995), central pore loops or other protein domain structures within the protein translocase, unfolding, or processing machine can contain critical substrate binding sites, which engage the substrate polypeptide (Wang et al., 2001, Hinnerwisch et al., 2005, Lum et al., 2008, Martin et al., 2008, Glynn et al., 2009). Upon cycles of ATP hydrolysis, these pore loops move like actuators to push/pull the substrate polypeptide (Glick, 1995, Wang et al., 2001, Hinnerwisch et al., 2005, Lum et al., 2008, Martin et al., 2008, Glynn et al., 2009, Zimmer et al., 2008). Thus unfolding forces in this mechanism can be generated rather directly by these loop movements via ATP hydrolysis, and translocation directionality can be enforced by allowing substrate interactions to occur in only the power-stroke direction and not during the resetting of these loops.

In the Brownian-ratchet mechanism, the thermal diffusive motion of the translocating polypeptide chain is biased in a directional manner by means of an external energy gradient. In a theoretical proposal, Oster and colleagues suggested that a chemical gradient of heat shock protein (differing in concentration across the membrane) could prevent backward/retrograde diffusion by binding the substrate chain as it emerged from a translocase channel, thereby acting as a steric clamp (Simon et al., 1992). Thus the key distinguishing feature between this mechanism and the active push/pull mechanism is how forces are developed on the substrate: for the Brownian-ratchet mechanism, Brownian-motion itself becomes rectified; and for the active mechanism, the ATP-dependent power stroke is directly coupled to ATP hydrolysis. The Brownian-ratchet mechanism has, however, drawn criticism principally on the issue of substrate unfolding, because it is not expected to generate enough force to denature a folded protein (Glick, 1995). This criticism assumes that channels generally do not assist in protein unfolding or bind to unfolded structure. Thus it has been postulated that the Brownian-ratchet mechanism would only be able to act upon pre-unfolded substrates.

The proton-motive driving force

In the Atx system, there is emerging evidence in support of a proton gradient (ΔpH)-driven Brownian-ratchet mechanism. Atx has proven to be a useful model system to characterize transmembrane protein translocation in part because the three proteins can be expressed recombinantly and studied independently. Also, translocation can be monitored directly using electrophysiology, where the applied driving force can be externally controlled and continuously adjusted (Zhang et al., 2004, Krantz et al., 2006). Control over the driving force is critical in order to identify barriers in the mechanism, determine their driving force dependencies, and ultimately distinguish between translocation models (Feld et al., 2010, Krantz et al., 2006, Krantz et al., 2005, Thoren et al., 2009). Using planar lipid bilayer electrophysiology, the PA channel can be inserted into model membrane bilayers and the pH and membrane potential conditions can be precisely controlled. Protein translocation can be observed by monitoring the restoration of ion conductance once the protein completely traverses the channel (Feld et al., 2010, Krantz et al., 2006, Krantz et al., 2005, Thoren et al., 2009, Zhang et al., 2004). While LF's amino-terminal PA-binding domain (LFN) can be driven through the PA channel under a pure membrane potential (Δψ) (Zhang et al., 2004), a one-to-two unit ΔpH resembling that expected across the endosomal membrane is better able to translocate full-length LF and EF than a pure Δψ (Krantz et al., 2006). Thus since the translocation of full-length LF and EF depends more on a ΔpH than a Δψ, the mechanism of ΔpH-driven translocation has emerged as an active area in Atx research.

A Brownian-ratchet model (also referred to as the charge-state ratchet) has been invoked to explain how a ΔpH is harnessed by the channel during translocation (Fig. 2) (Krantz et al., 2006). In this mechanism, acidic residues in the translocating chain protonate upon entering the PA channel, because the channel is cation-selective (and, therefore, anion repulsive). Once these protonated acidic groups emerge from the PA channel into the higher pH solution on the cytosolic side of the membrane, they can deprotonate, thereby allowing an electrostatic repulsion to develop between the negatively-charged channel and the exiting polypeptide chain. This electrostatic repulsion may effectively capture Brownian-motion, further driving translocation and enforcing directionality. Some recent tests of this Brownian-ratchet hypothesis have lent support to the model. Using a semisynthesis approach, Pentelute et al. show that charged residues are important for ΔpH-driven translocation (Pentelute et al., 2010). Also, it has been shown that negative charges from sulfonic acid groups attached artificially to the substrate protein inhibit translocation (Basilio et al., 2009). Since these groups cannot be protonated and their charge cannot be neutralized, it was proposed that the channel's cation-selective filter was rejecting these strong anions. Finally, the substrate unfolding step occurs in the most ΔpH-dependent step of translocation, demonstrating that the ΔpH mechanism can generate fairly substantial forces, since proteins generally require tens of piconewtons to unfold (Thoren et al., 2009).

Figure 2. Molecular solutions to the challenges of protein translocation.

Figure 2

Open in a new tab

Multiple clamping sites in the PA channel facilitate substrate unfolding by stabilizing partially unfolded intermediates. The α clamp elegantly mitigates the combinatorial chemical complexity and conformational heterogeneity of the substrate by recognizing uniformly shaped helical structures with broad sequence specificity (Feld et al., 2010). The α-clamp panel was rendered by morphing the protein databank (PDB) coordinates, 3KWV (Feld et al., 2010), with a model of the folded form of LFN docked on PA using the PDB coordinates, 1J7N (Pannifer et al., 2001). The aromatic residues in the Φ clamp also provide a means to interact with many types of substrate chemistries by taking advantage of hydrophobic effect and allowing for broad specificity (Krantz et al., 2005). The ΔpH driving force can harness diffusive Brownian motion by changing the protonation state of anionic residues in the substrate during translocation (Krantz et al., 2006).

Protein unfolding

Driving forces are, of course, critical to protein translocation, but how are they coupled to the unfolding step of the mechanism? Part of the answer to this question, in the case of Atx, may be inferred by its endocytic trafficking mechanism. Since Atx complexes enter cells via the acidified endocytic route, it seemed reasonable that LF and EF may be destabilized under these acidic pH conditions. Solution studies indicate, in fact, that LF and EF are destabilized under the mildly acidic pH conditions likely encountered in the endosomal compartment (pH 5 to 6) (Krantz et al., 2004). Specifically, it was found that under endosomal pH conditions LFN and EF's respective amino-terminal domain populate a fairly compact molten-globule intermediate with a large amount of secondary structure and disrupted tertiary packing interactions (Krantz et al., 2004).

While acidic pH conditions similar to those found inside the endosomal compartment can destabilize the substrate protein, it was uncertain how the substrate protein would actually unfold on the surface of the PA channel. This unfolding step of the translocation mechanism was recently interrogated by site-directed mutagenesis, thermodynamic stability studies, and planar lipid bilayer electrophysiology (Thoren et al., 2009). These studies reveal that there are two distinct barriers in LFN's translocation mechanism: an unfolding barrier and a translocation barrier. Thoren et al. demonstrated through mutational studies that the unfolding barrier is 10-fold more force-dependent than the translocation barrier, irrespective of the type of driving force (whether ΔpH or ΔΨ) (Thoren et al., 2009). Furthermore, by examining the effects of destabilizing mutations across the protein's structure, the precise location of the folded substructure, which is rate-limiting to the unfolding step of translocation, was identified. This structure in LFN corresponds to a β-sheet subdomain. From optical force microscopy studies, it is well known that β-sheet regions can form the rate-limiting mechanical breakpoint in a force-dependent unfolding mechanism, where the β-sheet orientation and topology are key determinants of the forces required for unfolding (Fig. 1C) (Brockwell et al., 2003). As little unfolding appears to occur in the less force-dependent step, Thoren et al. suggest that this step likely involves the translocation of the unfolded chain. Interestingly, this translocation barrier appears to impose an effective speed limit on the translocation of unfolded protein substrates (on the order of ~10 seconds for ~100-700 residue proteins) for the Atx system (Krantz et al., 2006, Thoren et al., 2009) and others (Burton et al., 2001, Huang et al., 1999, Kenniston et al., 2003), where this limit is invariant with the type of driving force applied (whether ATP, ΔΨ, or ΔpH).

The Φ clamp is a translocase active site in the PA channel

In an effort to understand how a molten-globular, partially-folded substrate would be linearized and fully unfolded during translocation, Krantz et al. identified potential active-site residues lining the PA channel (Krantz et al., 2005). While the lumen of the channel (in both the cap and the β-barrel stem region (Nassi et al., 2002)) is mainly lined with small hydrophilic residues, one prominent hydrophobic residue, Phe-427, was identified within the channel's cap (Krantz et al., 2005). Electron paramagnetic resonance spectroscopy (EPR) studies revealed that these Phe residues converge within the channel, forming a radially-symmetrical aromatic clamp site, called the “Φ clamp” (Krantz et al., 2005) (Fig. 2). Electrophysiology studies showed that the Φ-clamp structure is required for protein translocation. The Φ-clamp site forms an ion-conductance bottleneck in the channel, suggesting these Phe-427 residues make a narrow approach and could form a polypeptide binding site in the channel. This model was confirmed when it was observed that mutations to the Φ-clamp site allow the substrate to backslide or retrotranslocate in an unproductive manner, inhibiting efficient translocation. Model-compound-binding studies revealed that the Φ-clamp site possesses broad substrate specificity, where the multifaceted aromatic surfaces of the Phe residues preferred cationic, aromatic and hydrophobic substrates, consistent with the π-cloud electrons of the Phe residues making π-π, cation-π, and π-dipole interactions (Krantz et al., 2005). A more recent report shows that replacing a single Phe residue with a charged residue in the Φ-clamp site is strongly deactivating (Janowiak et al., 2010). This result suggests that the site may act cooperatively—which is anticipated for hydrophobic interaction sites. Therefore, the Φ-clamp site serves a chaperone-like function, where it may interact with a broad spectrum of sequences presented by the protein substrate as it translocates.

More recent studies have demonstrated that the Φ-clamp site actively unfolds LFN during translocation (Thoren et al., 2009). This result is not entirely surprising given the fact that the central hydrophobic clamp structure binds to unfolded protein and may thereby stabilize partially unfolded states (Krantz et al., 2005). Thus by binding to the translocating chain, the channel may help to unfold the substrate. The Φ clamp does not work alone, however, as has been shown in studies of ΔpH-driven translocation (Krantz et al., 2006). The ability of the channel to couple the ΔpH driving force to translocation requires a functioning Φ-clamp site (Krantz et al., 2006). This synergy between the proposed ΔpH-driven charge-state ratchet and the Φ clamp suggests that changes in the substrate's protonation state are coupled in some way with the activity of the Φ-clamp site. In conclusion, these findings are somewhat paradoxical and challenge the conventional wisdom about protein-conducting channels, because they suggest that the channel binds to and releases substrate in a continuous manner during translocation.

The α clamp is an unfoldase active site on the PA channel

Recently, Feld et al. solved the crystal structure of a core of a lethal toxin complex. The complex contains four LFN moieties bound to a PA octamer (Feld et al., 2010). The most interesting feature of the structure reveals that the first α helix and β strand of each LFN unfold and dock into a deep amphipathic cleft on the surface of the PA. This cleft, termed the “α clamp”, is framed by the conserved twin Ca2+-ion-binding sites found in each PA monomer (Feld et al., 2010) (Fig. 2). These twin Ca2+-ion-binding sites appear analogous to the peptide-binding groove found on a calmodulin-peptide complex (Meador et al., 1992). Because the side chains that structure the cleft walls directly complex with the Ca2+ ions, the majority of the surface in the cleft is composed of backbone atoms, thus few specific interactions in the groove are possible for the substrate helix. Moreover, the few side chains within the cleft are from Phe residues and the non-charged atoms in a nearby Arg residue. The other part of the recognition motif is a small β-sheet docking site found near the α-clamp site; this site recognizes the β strand from LFN using a pair of backbone hydrogen bonds only. Overall, the cleft forms a nearly perfect fit for a generic protein helix, where all the interactions appear nonspecific.

The crystal structure of the interaction of the partially unfolded LFN in complex with the PA oligomer was confirmed by a number of functional studies. To test whether LFN must unfolds upon binding PA, Feld et al. found that a disulfide cross-linked version (which prevented the α1 helix from unfolding) bound 104-fold less tightly to the PA channel than the non-cross-linked version (Feld et al., 2010). EPR studies also report that LFN undergoes a similar unfolding transition upon binding PA (Jennings-Antipov et al., 2011). Through extensive mutagenesis studies on both PA and the substrate LFN, the α-clamp site was shown to be able to bind a broad array of polypeptide substrates (Feld et al., 2010). Additional studies have revealed that mutations in Phe and Pro residues in the α-clamp site inhibit both full length LF and LFN translocation; however, some mutants affected LF more so than LFN, leading to the hypothesis that additional folded domains in LF are also unfolded and processed by the α-clamp site in an analogous manner. These results further implicate aromatic residues as key players in the molecular mechanism of nonspecific protein binding and mechanical unfolding. In conclusion, the PA channel appears to have the ability to unfold LF and EF by recognizing a generic α helix and a small portion of proximal β strand. These results imply a means for the PA channel to denature and unfold the multidomain LF, where the process may be repeated on each folded portion of the substrate during translocation.

Chaperone-assisted translocation

PA's translocase activity is sufficient for translocation in planar bilayer electrophysiology assays (Feld et al., 2010, Kintzer et al., 2009, Krantz et al., 2006, Krantz et al., 2005, Thoren et al., 2009, Zhang et al., 2004). Under modest ΔpH conditions at room temperature, LF requires about 100 s to translocate via either the heptameric or octameric PA channel (Kintzer et al., 2009). Nonetheless, auxilliary factors from the host cell also may assist translocation. For example, ATP and cytosolic factors (including heat shock protein 90, thioredoxin reductase and the β subunit of the coat protein complex) enhance the translocation of diphtheria toxin A domain (DTA) (Ratts et al., 2003) and the LFN-DTA fusion protein substrate (Tamayo et al., 2008), as measured by the establishment of DTA ADP-ribosyltransferase activity outside of the endosomal lumen. However, the role of these chaperones in the translocation pathway remains unsettled; the enhancement of enzymatic activity could either indicate that the chaperones assist during unfolding and translocation or they function post-translocationally to properly fold the translocated substrate. A more recent study (Dmochewitz et al., 2010), reports that drug compounds, which inhibit the chaperones, heat shock protein 90 and cyclophilin, block the cellular uptake of LFN-DTA, but not LF. It is unlikely that either the channel or the cellular chaperones operate in pure isolation during Atx translocation, because a portion of LF or EF must translocate in order to allow a cellular chaperone to engage a partially translocated protein and resume the translocation process. The protein-processing mechanisms elucidated for the PA channel (Feld et al., 2010, Krantz et al., 2005, Thoren et al., 2009) and discussed within this Microreview also should apply broadly to how host-cell chaperones facilitate protein translocation and other downstream refolding processes. In order to bind to an unfolded, translocating protein, chaperones would have to recognize common features on the unfolded chain and mitigate similar challenges presented in Figure 1C.

Concluding remarks

Molecular machines face a number of challenges in transporting a protein either across a membrane or into a proteolytic complex (Fig. 1C). Initially, the substrate protein is unfolded, requiring the transduction of a source of free energy into a mechanical unfolding force. Subsequently, the unfolded polypeptide chain is translocated through a narrow channel. These two steps are coupled processes, of course, and pose several challenges beyond mechanical unfolding itself: (i) the unfolded chain presents an ever-changing array of possible chemistries and possible backbone configurations as it is translocated; and (ii) diffusive motion could result in counter-productive backsliding of the polypeptide (Fig. 1C). How does a molecular machine overcome these challenges during translocation? Current models (Krantz et al., 2005, Krantz et al., 2006, Feld et al., 2010) propose that multiple nonspecific binding sites are critical features of unfolding machines. By recognizing general features of an unfolded chain, like uniformly shaped α-helical secondary structure, backbone hydrogen bonds, or hydrophobic groups, these nonspecific binding sites provide an elegant solution (Fig. 2) to the configurational and chemical variability of the unfolded state (Fig. 1C). In addition, binding the polypeptide in multiple locations helps stabilize unfolded intermediates, productively resists backsliding (retro-translocation), and allows the available free energy to be more efficiently converted into a productive mechanical force.

In the PA channel, several discrete clamping sites are observed that address these challenging aspects of translocation, namely the α clamp, the Φ clamp, and an anion-charge-repulsion site implicated in ΔpH-driven translocation (Fig. 2). The α clamp is not only important to unfolding the first helix of LFN, but it may also play a role in unfolding later domains of a translocating substrate. The next clamping site encountered in the channel is the Φ clamp, which can stabilize further unfolding by binding to sequences nonspecifically (albeit favoring regions rich in aromatic, hydrophobic and cationic groups), while also limiting diffusive motions that result in retro-translocation. Finally, under appropriate ΔpH-driving-force conditions, Brownian motion may be captured or rectified through the changing protonation state of acidic residues in the translocating chain.

In the Atx system, we find that the electrical potential and the ΔpH generate sufficient driving forces to unfold LF and EF, where the latter driving force is more effective. Nonspecific binding sites provide a means to engage folded substrate, creating points of resistance, or fulcrums, which allow the driving force to be more efficiently applied to a folded domain. Once partially unfolded, the nonspecific clamp sites bind and stabilize partially unfolded states of the substrate, further lowering the unfolding barrier of the translocation mechanism. Moreover, these critical peptide binding clamps do not function in isolation. Increasingly, it has become apparent that these sites work cooperatively: the α clamp works together with the Φ clamp to unfold LF, and the ΔpH-driven charge-state ratchet works in conjunction with the Φ clamp to drive translocation.

On one hand, the idea that numerous nonspecific binding sites can form nonspecific interactions with the substrate appears to be quite problematic, as extensive binding would create thermodynamic traps and impair translocation. On the other hand, the penalty of having numerous clamping sites may be offset by some of the following benefits, notably orienting the substrate toward the central lumen, reducing the stability of the substrate, and minimizing the diffusional mobility and backward translocation during translocation. Thus as perceived from the kinetic benefits observed in the Atx system, multiple clamp sites do not simply act as thermodynamic traps, rather they are able to catalyze translocation by populating a series of partially unfolded intermediates, which lower rate-limiting unfolding barriers and minimize diffusive forces stemming from unconstrained backbone configurational entropy. Analogous clamping sites are broadly observed in many other transporters and unfoldases (Wang et al., 2001, Hinnerwisch et al., 2005, Lum et al., 2008, Martin et al., 2008, Glynn et al., 2009, Levchenko et al., 2003, Van den Berg et al., 2004, Zimmer et al., 2008); and therefore, we expect aspects of this model to be generally applicable to protein translocation mechanisms. Future work should continue to address the structures of substrate proteins in complex with these clamping sites, probe the dynamic interplay between the clamps, examine force-dependent unfolding mechanism of LF and EF, and further examine the validity of the Brownian-ratchet mechanism.

Acknowledgements

The authors thank Geoffrey Feld for preparing a figure illustration. We thank other members of the Krantz Lab for thoughtful discussions. This work was supported by University of California start-up funds; an NIH research grant, R01-AI077703 (B.A.K.); and an NIH training grant, T32GM08295 (K.L.T.)

References

  1. Basilio D, Juris SJ, Collier RJ, Finkelstein A. Evidence for a proton-protein symport mechanism in the anthrax toxin channel. J Gen Physiol. 2009;133:307–314. doi: 10.1085/jgp.200810170. [DOI] [PMC free article] [PubMed] [Google Scholar]
  2. Benson EL, Huynh PD, Finkelstein A, Collier RJ. Identification of residues lining the anthrax protective antigen channel. Biochemistry. 1998;37:3941–3948. doi: 10.1021/bi972657b. [DOI] [PubMed] [Google Scholar]
  3. Brockwell DJ, Paci E, Zinober RC, Beddard GS, Olmsted PD, Smith DA, Perham RN, Radford SE. Pulling geometry defines the mechanical resistance of a beta-sheet protein. Nat Struct Biol. 2003;10:731–737. doi: 10.1038/nsb968. [DOI] [PubMed] [Google Scholar]
  4. Burton RE, Siddiqui SM, Kim YI, Baker TA, Sauer RT. Effects of protein stability and structure on substrate processing by the ClpXP unfolding and degradation machine. Embo J. 2001;20:3092–3100. doi: 10.1093/emboj/20.12.3092. [DOI] [PMC free article] [PubMed] [Google Scholar]
  5. Dmochewitz L, Lillich M, Kaiser E, Jennings LD, Lang AE, Buchner J, Fischer G, Aktories K, Collier RJ, Barth H. Role of CypA and Hsp90 in membrane translocation mediated by anthrax protective antigen. Cell Microbiol. 2010;13:359–373. doi: 10.1111/j.1462-5822.2010.01539.x. [DOI] [PMC free article] [PubMed] [Google Scholar]
  6. Feld GK, Thoren KL, Kintzer AF, Sterling HJ, Tang II, Greenberg SG, Williams ER, Krantz BA. Structural basis for the unfolding of anthrax lethal factor by protective antigen oligomers. Nature Struct Mol Biol. 2010;17:1383–1390. doi: 10.1038/nsmb.1923. [DOI] [PMC free article] [PubMed] [Google Scholar]
  7. Glick BS. Can Hsp70 proteins act as force-generating motors? Cell. 1995;80:11–14. doi: 10.1016/0092-8674(95)90444-1. [DOI] [PubMed] [Google Scholar]
  8. Glynn SE, Martin A, Nager AR, Baker TA, Sauer RT. Structures of asymmetric ClpX hexamers reveal nucleotide-dependent motions in a AAA+ protein-unfolding machine. Cell. 2009;139:744–756. doi: 10.1016/j.cell.2009.09.034. [DOI] [PMC free article] [PubMed] [Google Scholar]
  9. Hinnerwisch J, Fenton WA, Furtak KJ, Farr GW, Horwich AL. Loops in the central channel of ClpA chaperone mediate protein binding, unfolding, and translocation. Cell. 2005;121:1029–1041. doi: 10.1016/j.cell.2005.04.012. [DOI] [PubMed] [Google Scholar]
  10. Huang S, Ratliff KS, Schwartz MP, Spenner JM, Matouschek A. Mitochondria unfold precursor proteins by unraveling them from their N-termini. Nature Struct Biol. 1999;6:1132–1138. doi: 10.1038/70073. [DOI] [PubMed] [Google Scholar]
  11. Janowiak BE, Fischer A, Collier RJ. Effects of introducing a single charged residue into the phenylalanine clamp of multimeric anthrax protective antigen. J Biol Chem. 2010;285:8130–8137. doi: 10.1074/jbc.M109.093195. [DOI] [PMC free article] [PubMed] [Google Scholar]
  12. Jennings-Antipov LD, Song L, Collier RJ. Interactions of anthrax lethal factor with protective antigen defined by site-directed spin labeling. Proc Natl Acad Sci U S A. 2011;108:1868–1873. doi: 10.1073/pnas.1018965108. [DOI] [PMC free article] [PubMed] [Google Scholar]
  13. Katayama H, Janowiak BE, Brzozowski M, Juryck J, Falke S, Gogol EP, Collier RJ, Fisher MT. GroEL as a molecular scaffold for structural analysis of the anthrax toxin pore. Nature Struct Mol Biol. 2008;15:754–760. doi: 10.1038/nsmb.1442. [DOI] [PMC free article] [PubMed] [Google Scholar]
  14. Katayama H, Wang J, Tama F, Chollet L, Gogol EP, Collier RJ, Fisher MT. Three-dimensional structure of the anthrax toxin pore inserted into lipid nanodiscs and lipid vesicles. Proc Natl Acad Sci USA. 2010;107:3453–3457. doi: 10.1073/pnas.1000100107. [DOI] [PMC free article] [PubMed] [Google Scholar]
  15. Kenniston JA, Baker TA, Fernandez JM, Sauer RT. Linkage between ATP consumption and mechanical unfolding during the protein processing reactions of an AAA+ degradation machine. Cell. 2003;114:511–520. doi: 10.1016/s0092-8674(03)00612-3. [DOI] [PubMed] [Google Scholar]
  16. Kintzer AF, Sterling HJ, Tang II, Abdul-Gader A, Miles AJ, Wallace BA, Williams ER, Krantz BA. Role of the protective antigen octamer in the molecular mechanism of anthrax lethal toxin stabilization in plasma. J Mol Biol. 2010a;399:741–758. doi: 10.1016/j.jmb.2010.04.041. [DOI] [PMC free article] [PubMed] [Google Scholar]
  17. Kintzer AF, Sterling HJ, Tang II, Williams ER, Krantz BA. Anthrax toxin receptor drives protective antigen oligomerization and stabilizes the heptameric and octameric oligomer by a similar mechanism. PLoS ONE. 2010b;5:e13888. doi: 10.1371/journal.pone.0013888. [DOI] [PMC free article] [PubMed] [Google Scholar]
  18. Kintzer AF, Thoren KL, Sterling HJ, Dong KC, Feld GK, Tang II, Zhang TT, Williams ER, Berger JM, Krantz BA. The protective antigen component of anthrax toxin forms functional octameric complexes. J Mol Biol. 2009;392:614–629. doi: 10.1016/j.jmb.2009.07.037. [DOI] [PMC free article] [PubMed] [Google Scholar]
  19. Krantz BA, Finkelstein A, Collier RJ. Protein translocation through the anthrax toxin transmembrane pore is driven by a proton gradient. J Mol Biol. 2006;355:968–979. doi: 10.1016/j.jmb.2005.11.030. [DOI] [PubMed] [Google Scholar]
  20. Krantz BA, Melnyk RA, Zhang S, Juris SJ, Lacy DB, Wu Z, Finkelstein A, Collier RJ. A phenylalanine clamp catalyzes protein translocation through the anthrax toxin pore. Science. 2005;309:777–781. doi: 10.1126/science.1113380. [DOI] [PMC free article] [PubMed] [Google Scholar]
  21. Krantz BA, Trivedi AD, Cunningham K, Christensen KA, Collier RJ. Acid-induced unfolding of the amino-terminal domains of the lethal and edema factors of anthrax toxin. J Mol Biol. 2004;344:739–756. doi: 10.1016/j.jmb.2004.09.067. [DOI] [PubMed] [Google Scholar]
  22. Lacy DB, Wigelsworth DJ, Melnyk RA, Harrison SC, Collier RJ. Structure of heptameric protective antigen bound to an anthrax toxin receptor: a role for receptor in pH-dependent pore formation. Proc Natl Acad Sci USA. 2004;101:13147–13151. doi: 10.1073/pnas.0405405101. [DOI] [PMC free article] [PubMed] [Google Scholar]
  23. Levchenko I, Grant RA, Wah DA, Sauer RT, Baker TA. Structure of a delivery protein for an AAA+ protease in complex with a peptide degradation tag. Mol Cell. 2003;12:365–372. doi: 10.1016/j.molcel.2003.08.014. [DOI] [PubMed] [Google Scholar]
  24. Lum R, Niggemann M, Glover JR. Peptide and protein binding in the axial channel of Hsp104. Insights into the mechanism of protein unfolding. J Biol Chem. 2008;283:30139–30150. doi: 10.1074/jbc.M804849200. [DOI] [PMC free article] [PubMed] [Google Scholar]
  25. Martin A, Baker TA, Sauer RT. Pore loops of the AAA+ ClpX machine grip substrates to drive translocation and unfolding. Nature Struct Mol Biol. 2008;15:1147–1151. doi: 10.1038/nsmb.1503. [DOI] [PMC free article] [PubMed] [Google Scholar]
  26. Meador WE, Means AR, Quiocho FA. Target enzyme recognition by calmodulin: 2.4 A structure of a calmodulin-peptide complex. Science. 1992;257:1251–1255. doi: 10.1126/science.1519061. [DOI] [PubMed] [Google Scholar]
  27. Milne JC, Furlong D, Hanna PC, Wall JS, Collier RJ. Anthrax protective antigen forms oligomers during intoxication of mammalian cells. J Biol Chem. 1994;269:20607–20612. [PubMed] [Google Scholar]
  28. Mogridge J, Cunningham K, Collier RJ. Stoichiometry of anthrax toxin complexes. Biochemistry. 2002;41:1079–1082. doi: 10.1021/bi015860m. [DOI] [PubMed] [Google Scholar]
  29. Nassi S, Collier RJ, Finkelstein A. PA63 channel of anthrax toxin: an extended β-barrel. Biochemistry. 2002;41:1445–1450. doi: 10.1021/bi0119518. [DOI] [PubMed] [Google Scholar]
  30. Pannifer AD, Wong TY, Schwarzenbacher R, Renatus M, Petosa C, Bienkowska J, Lacy DB, Collier RJ, Park S, Leppla SH, Hanna P, Liddington RC. Crystal structure of the anthrax lethal factor. Nature. 2001;414:229–233. doi: 10.1038/n35101998. [DOI] [PubMed] [Google Scholar]
  31. Pentelute BL, Barker AP, Janowiak BE, Kent SB, Collier RJ. A semisynthesis platform for investigating structure-function relationships in the N-terminal domain of the anthrax Lethal Factor. ACS Chem Biol. 2010;5:359–364. doi: 10.1021/cb100003r. [DOI] [PMC free article] [PubMed] [Google Scholar]
  32. Petosa C, Collier RJ, Klimpel KR, Leppla SH, Liddington RC. Crystal structure of the anthrax toxin protective antigen. Nature. 1997;385:833–838. doi: 10.1038/385833a0. [DOI] [PubMed] [Google Scholar]
  33. Ratts R, Zeng H, Berg EA, Blue C, McComb ME, Costello CE, vanderSpek JC, Murphy JR. The cytosolic entry of diphtheria toxin catalytic domain requires a host cell cytosolic translocation factor complex. J Cell Biol. 2003;160:1139–1150. doi: 10.1083/jcb.200210028. [DOI] [PMC free article] [PubMed] [Google Scholar]
  34. Simon SM, Peskin CS, Oster GF. What drives the translocation of proteins? Proc Natl Acad Sci USA. 1992;89:3770–3774. doi: 10.1073/pnas.89.9.3770. [DOI] [PMC free article] [PubMed] [Google Scholar]
  35. Tamayo AG, Bharti A, Trujillo C, Harrison R, Murphy JR. COPI coatomer complex proteins facilitate the translocation of anthrax lethal factor across vesicular membranes in vitro. Proc Natl Acad Sci USA. 2008;105:5254–5259. doi: 10.1073/pnas.0710100105. [DOI] [PMC free article] [PubMed] [Google Scholar]
  36. Thoren KL, Worden EJ, Yassif JM, Krantz BA. Lethal factor unfolding is the most force-dependent step of anthrax toxin translocation. Proc Natl Acad Sci USA. 2009;106:21555–21560. doi: 10.1073/pnas.0905880106. [DOI] [PMC free article] [PubMed] [Google Scholar]
  37. Van den Berg B, Clemons WM, Jr., Collinson I, Modis Y, Hartmann E, Harrison SC, Rapoport TA. X-ray structure of a protein-conducting channel. Nature. 2004;427:36–44. doi: 10.1038/nature02218. [DOI] [PubMed] [Google Scholar]
  38. Wang J, Song JJ, Franklin MC, Kamtekar S, Im YJ, Rho SH, Seong IS, Lee CS, Chung CH, Eom SH. Crystal structures of the HslVU peptidase-ATPase complex reveal an ATP-dependent proteolysis mechanism. Structure. 2001;9:177–184. doi: 10.1016/s0969-2126(01)00570-6. [DOI] [PubMed] [Google Scholar]
  39. Young JA, Collier RJ. Anthrax toxin: receptor binding, internalization, pore formation, and translocation. Annu Rev Biochem. 2007;76:243–265. doi: 10.1146/annurev.biochem.75.103004.142728. [DOI] [PubMed] [Google Scholar]
  40. Zhang S, Udho E, Wu Z, Collier RJ, Finkelstein A. Protein translocation through anthrax toxin channels formed in planar lipid bilayers. Biophys J. 2004;87:3842–3849. doi: 10.1529/biophysj.104.050864. [DOI] [PMC free article] [PubMed] [Google Scholar]
  41. Zimmer J, Nam Y, Rapoport TA. Structure of a complex of the ATPase SecA and the protein-translocation channel. Nature. 2008;455:936–943. doi: 10.1038/nature07335. [DOI] [PMC free article] [PubMed] [Google Scholar]

RESOURCES