SEARCH

SEARCH BY CITATION

Summary

  1. Top of page
  2. Summary
  3. Introduction
  4. The N-end rule: what is an N-degron?
  5. Bacterial N-degrons: how are they generated?
  6. Modification of the N-terminus: new enzymes and activities
  7. N-degron recognition by ClpS, the bacterial N-recognin
  8. ClpAP, an ATP-fuelled machine of destruction (mechanism of action)
  9. N-degron delivery and degradation: more than recognition of the N-terminal residue
  10. Substrates and the physiological function of the N-end rule pathway
  11. Compartmental protein turnover by the N-end rule pathway?
  12. Acknowledgements
  13. References

The N-end rule pathway is a highly conserved process that operates in many different organisms. It relates the metabolic stability of a protein to its N-terminal amino acid. Consequently, amino acids are described as either ‘stabilizing’ or ‘destabilizing’. Destabilizing residues are organized into three hierarchical levels: primary, secondary, and in eukaryotes – tertiary. Secondary and tertiary destabilizing residues act as signals for the post-translational modification of the target protein, ultimately resulting in the attachment of a primary destabilizing residue to the N-terminus of the protein. Regardless of their origin, proteins containing N-terminal primary destabilizing residues are recognized by a key component of the pathway. In prokaryotes, the recognition component is a specialized adaptor protein, known as ClpS, which delivers target proteins directly to the ClpAP protease for degradation. In contrast, eukaryotes use a family of E3 ligases, known as UBRs, to recognize and ubiquitylate their substrates resulting in their turnover by the 26S proteasome. While the physiological role of the N-end rule pathway is largely understood in eukaryotes, progress on the bacterial pathway has been slow. However, new interest in this area of research has invigorated several recent advances, unlocking some of the secrets of this unique proteolytic pathway in prokaryotes.


Introduction

  1. Top of page
  2. Summary
  3. Introduction
  4. The N-end rule: what is an N-degron?
  5. Bacterial N-degrons: how are they generated?
  6. Modification of the N-terminus: new enzymes and activities
  7. N-degron recognition by ClpS, the bacterial N-recognin
  8. ClpAP, an ATP-fuelled machine of destruction (mechanism of action)
  9. N-degron delivery and degradation: more than recognition of the N-terminal residue
  10. Substrates and the physiological function of the N-end rule pathway
  11. Compartmental protein turnover by the N-end rule pathway?
  12. Acknowledgements
  13. References

Intracellular proteolysis is a fundamental process performed by all living organisms. For many years it was simply seen as a cellular strategy to remove damaged proteins and recycle amino acids. Currently, however, it is viewed as a sophisticated method to regulate important cellular processes. In eukaryotes, intracellular proteolysis involves the covalent attachment of a small protein molecule, ubiquitin (Ub), to an internal Lys on the substrate resulting in the ATP-dependent degradation of the ubiquitylated protein by the 26S proteasome. Interestingly, two functional homologues of ubiquitin were recently identified in prokaryotes. The prokaryotic ubiquitin-like protein (Pup) is responsible for modification of protein substrates in Mycobacterium sp. (for a recent review see Burns and Darwin, 2010), while two small archaeal modifier proteins (SAMPs) were shown to modify substrate proteins in Haloferax volcanii (Humbard et al., 2010). In both cases, the modifying proteins (Pup or SAMP) are attached to internal Lys residues in the substrate and the resulting pupylated or SAMPylated protein is degraded by the proteasome. Although this recent evidence indicates that prokaryotes can use Ub-like modifications to tag and degrade substrate proteins, most prokaryotes do not use this system, rather they use specific adaptor proteins to recognize and deliver protein substrates for degradation (see later).

Independent of the tagging or delivery mechanism used, intracellular proteolysis contributes to a variety of diverse functions in a range of bacterial species, from general protein quality control (Dougan et al., 2002a) (through the removal of misfolded, damaged or incompletely synthesized proteins), to pathogenesis (Butler et al., 2006) and development (Jenal, 2009; Moliere and Turgay, 2009). Similarly, it contributes to the regulated turnover of key signalling molecules or transcription factors in Escherichia coli thereby controlling a number of cellular pathways such as stress responses (Ades, 2008; Hengge, 2009). Interestingly, although we have developed a comprehensive understanding of several of these proteolytic pathways in E. coli, the physiological function of one highly conserved pathway – the N-end rule pathway – remains poorly understood. Despite this, many of the components of this pathway have been identified using model substrates, and ‘rules’ have been proposed. The central rule of the N-end rule pathway relates the metabolic stability (or half-life) of a protein to the identity of the N-terminal amino acid (Varshavsky, 1996). Like most proteolytic cascades, the N-end rule pathway involves a number of key steps: the conditional exposure of a degradation signal (in this case an N-terminal degradation signal, N-degron) on the target protein (substrate), timely and specific degron recognition, and delivery of the tagged protein to the appropriate proteolytic machine, where the substrate is unfolded, translocated and degraded. In this review, we highlight recent advances in our understanding of the bacterial N-end rule pathway, from N-degron generation, including the identification of new enzymes and specificities, to the molecular basis of N-degron recognition, and the mechanisms of substrate delivery. While some of the recent findings challenge our understanding of the pathway, most findings complement the existing rules.

The N-end rule: what is an N-degron?

  1. Top of page
  2. Summary
  3. Introduction
  4. The N-end rule: what is an N-degron?
  5. Bacterial N-degrons: how are they generated?
  6. Modification of the N-terminus: new enzymes and activities
  7. N-degron recognition by ClpS, the bacterial N-recognin
  8. ClpAP, an ATP-fuelled machine of destruction (mechanism of action)
  9. N-degron delivery and degradation: more than recognition of the N-terminal residue
  10. Substrates and the physiological function of the N-end rule pathway
  11. Compartmental protein turnover by the N-end rule pathway?
  12. Acknowledgements
  13. References

Protein turnover is invariably initiated by the presence of a degradation signal (degron). Usually this degron is located within the protein destined for proteolysis; however, protein turnover occasionally occurs in trans and the signal can be located on another subunit of the same protein complex (Schrader et al., 2009). These degradation signals are often hidden in the folded protein and only become exposed upon application of an external stimulus or stress. Degrons might be structurally defined but most are linear in nature and generally are located at or near the N- or C-terminus. In a seminal study more than 20 years ago, Alexander Varshavsky's laboratory discovered that, in Saccharomyces cerevisiae, the N-terminal residue of a protein is related to the metabolic stability of that protein and coined the term ‘The N-end rule’ (Bachmair et al., 1986). Varshavsky and colleagues discovered that some amino acids, when placed at the N-terminus of β-galactosidase, signalled its rapid degradation (half-life ∼3 min), while others did not (half-life > 20 h), giving rise to the terms ‘destabilizing’ and ‘stabilizing’ residues respectively. Interestingly, N-terminal acetylation of some proteins can also acts as a ‘destabilizing’ signal (Mayer et al., 1989) and recently the yeast ubiquitin ligase (Doa10) responsible for the recognition of this degron (Ac-N-degron) was identified, confirming that the turnover of these proteins occurs via a distinct branch of the N-end rule pathway (Hwang et al., 2010). The N-end rule has been identified in a broad range of organisms from E. coli (Tobias et al., 1991), to plants (Bachmair et al., 1993; Graciet et al., 2009a,b) and mammals (Gonda et al., 1989). Similarly, many of the cellular components responsible for the generation, recognition or degradation of these N-end rule substrates were identified in these organisms, indicating that the N-end rule pathway is evolutionarily conserved (Varshavsky, 1996). Invariably, the pathway is organized in a hierarchical manner, with destabilizing residues being classified into various levels (primary, secondary and tertiary); however, the type of destabilizing residue and the number of levels can vary (Varshavsky, 1996). For example, E. coli contains two levels of destabilizing residue (primary and secondary), while more complex eukaryotic organisms such as mammals contain an additional level of organization (tertiary destabilizing residues) (Fig. 1). Consistent with this organization, the enzymes responsible for generating these residues also vary in the different organisms (Fig. 1). For example, most tertiary destabilizing residues in mammals are converted into secondary destabilizing residues by N-terminal amidases, NTAN1 (Grigoryev et al., 1996) or NTAQ1 (Wang et al., 2009) and ultimately the primary destabilizing residue (R) is attached to a secondary destabilizing residue by arginyl-tRNA-protein transferase (ATE1) (Balzi et al., 1990). In contrast, primary destabilizing residues in E. coli are attached to secondary destabilizing residues by L/F-tRNA-protein transferase (LFTR).

image

Figure 1. The N-end rule pathway is hierarchical in both pro- and eukaryotes. In plants and mammals, tertiary destabilizing residues (N, Q and C) are converted into secondary destabilizing residues [D, E and oxidized C (C*)], by either N-terminal amidases NTAN1 and NTAQ1 or NO/O2 respectively. Ultimately a primary destabilizing residue (R) is conjugated to a secondary destabilizing residue by arginyl-tRNA-protein transferase (ATE1). In eukaryotes, primary destabilizing residues are recognized by separate binding pockets [type 1 (R, K and H) and type 2 (L, F, Y, W and I)] within the N-recognin. The bound substrate is ubiquitylated and ultimately degraded by the 26S proteasome. In contrast to eukaryotic cells, E. coli lack tertiary destabilizing residues and thus primary destabilizing residues (e.g. L or F) are attached to secondary destabilizing residues (R, K and M) by L/F-tRNA-protein transferase (LFTR). Regardless of their origin, all primary destabilizing residues (L, F, Y and W) are recognized by ClpS and delivered to ClpAP for degradation.

Download figure to PowerPoint

The specificity of E. coli LFTR (EcLFTR) was determined initially using model substrates and hence secondary destabilizing residues in E. coli were classically defined as arginine (a primary destabilizing residue in mammals) and lysine. Recently however, the first natural substrate of EcLFTR was identified (Ninnis et al., 2009) and surprisingly the N-terminal acceptor residue of this natural substrate was neither arginine nor lysine but methionine. This unexpected finding suggests that Met (together with Lys and Arg) can act as a secondary destabilizing residue in E. coli (Fig. 1), which raises several questions about the specificity of LFTR (discussed later). Finally, regardless of their origin, proteins containing N-terminal primary destabilizing residues are ultimately recognized by the substrate-binding component of the N-end rule pathway (ClpS in E. coli) and delivered to the appropriate protease (ClpAP in E. coli), where they are rapidly degraded in an ATP-dependent manner (Fig. 1).

Bacterial N-degrons: how are they generated?

  1. Top of page
  2. Summary
  3. Introduction
  4. The N-end rule: what is an N-degron?
  5. Bacterial N-degrons: how are they generated?
  6. Modification of the N-terminus: new enzymes and activities
  7. N-degron recognition by ClpS, the bacterial N-recognin
  8. ClpAP, an ATP-fuelled machine of destruction (mechanism of action)
  9. N-degron delivery and degradation: more than recognition of the N-terminal residue
  10. Substrates and the physiological function of the N-end rule pathway
  11. Compartmental protein turnover by the N-end rule pathway?
  12. Acknowledgements
  13. References

Protein synthesis in bacteria is initiated with formyl-Met (fMet). Therefore, given that fMet is not a destabilizing residue, the generation of an N-terminal degradation signal (N-degron) requires some form of post-translational modification. To date, at least three different methods of post-translational modification have been proposed for the generation of an N-degron in vivo (Fig. 2). However, since only two natural E. coli substrates have been identified, there is little to no evidence to validate most of the proposed pathways; hence, these pathways remain somewhat hypothetical.

image

Figure 2. Three pathways have been proposed for the generation of an N-degron in vivo. A. Following ribosomal synthesis, which in prokaryotes is initiated with formyl-methionine (fM), the formyl group is removed (step 1) by peptide deformylase (PDF). The processing of the initiating methionine (step 2a) by methionine aminopeptidase (MetAP) is dependent on the amino acid sequence adjacent to the initiating methionine. The initiating methionine is removed, when the second amino acid, 2 (blue) = A, G, S, C, P, and in some cases V or T. The initiating methionine is retained when the second amino acid, 2 (purple) = W, Y, F, L, R, K, M, D, E, I, H, Q, N. Primary (green) and secondary (yellow) destabilizing residues of the N-end rule pathway are incompatible with processing by MetAP (step 2b). However, in the case of PATase, L/F-tRNA-protein transferase (LFTR) attaches a primary destabilizing residue (N1) directly to the initiating Met. B and C. Proteolytic processing of a pre-N-degron containing a stabilizing N-terminal residue (NS) reveals either (B) a primary destabilizing residue, or (C) a secondary destabilizing residue (N2) onto which a primary destabilizing residue is attached.

Download figure to PowerPoint

The first proposed pathway involves the removal of the N-terminal Met to reveal an N-degron (Fig. 2A). In bacteria, this process involves two steps; first the formyl group of f-Met is removed by peptide deformylase (PDF) (Fig. 2A, step 1), which permits the removal of the N-terminal Met by methionine aminopeptidase (MetAP) (Fig. 2A, step 2a). However, the proposed specificity of MetAP is inconsistent with the generation of N-degrons, as proteins containing primary or secondary destabilizing residues adjacent to the N-terminal Met are generally not processed and, hence, retain their N-terminal Met (Fig. 2A, step 2b). Despite this general bias against bulky amino acids in the second position, removal of the initiating Met by MetAP has occasionally been observed in proteins containing Val or Thr in the second position (see Fig. 2A, step 2a) (Frottin et al., 2006). Therefore, although highly speculative, the removal of the N-terminal Met from some proteins containing other bulky amino acids in the second position (e.g. a primary destabilizing residue) might also be possible. Nevertheless, independent of this speculation, it is generally accepted that MetAP is unlikely to contribute to the biogenesis of an N-degron.

Surprisingly, a variation of this pathway (Fig. 2A, steps 1, 2b and 3) is responsible for the generation of the first bona fide substrate of the N-end rule pathway identified in bacteria (Ninnis et al., 2009). Indeed, the specificity of MetAP plays an important, albeit indirect, role in the metabolic stability of this natural substrate. In this case, the inability of MetAP to remove the N-terminal Met permits the attachment of a primary destabilizing residue, by LFTR, directly to the N-terminal Met of putrescine aminotransferase (PATase) (Fig. 2A, step 3). Although ∼20% of all E. coli proteins, such as PATase, are expected to retain their N-terminal Met (Frottin et al., 2006), this pathway is unlikely to represent a general method for the generation of N-degrons in vivo. Rather, it has been proposed that the N-terminal Met acts as a ‘conditional’ acceptor residue for LFTR (Ninnis et al., 2009) in which binding by LFTR is dependent on the biochemical nature of the residue adjacent to the N-terminal Met (discussed later).

The remaining model for the post-translational generation of a bacterial N-degron in vivo involves an initial processing event. Here a pre-N-degron (NS; Fig. 2B and C) is processed by an unknown endopeptidase, resulting in the exposure of a new N-terminus, revealing either a primary (N1; Fig. 2B) or secondary (N2; Fig. 2C) destabilizing residue. Interestingly, proteolytic processing by an endopeptidase (separase) has been observed in yeast, and is responsible for the cleavage of a subunit of cohesin, SCC1, resulting in its turnover by the N-end rule pathway (Rao et al., 2001). Furthermore, given that the specificity of MetAP largely precludes the generation of an N-degron (Fig. 2A), this type of ‘processing-dependent’ pathway remains a plausible method to generate an N-degron in E. coli. If a secondary destabilizing residue is exposed by a processing event, the LFTR-mediated attachment of a primary destabilizing residue is also predicted, in order for substrate recognition, by ClpS, to occur (Fig. 2C). To date, however, there is no evidence for the generation of a physiological substrate by this pathway. In contrast, and consistent with the second pathway (Fig. 2B), two recent studies showed that an N-terminally truncated fragment of the bacterial stress protein Dps (DNA protection during starvation) interacted with E. coli ClpS (Ninnis et al., 2009; Schmidt et al., 2009). Importantly, the binding of this Dps fragment to ClpS was dependent on the proteolytic removal of the first five residues of the protein, resulting in the exposure of a primary destabilizing residue at its new N-terminus (Ninnis et al., 2009; Schmidt et al., 2009). Although this fragment is recognized by ClpS and degraded by ClpAPS in vitro, it is currently unclear if processing of Dps is physiologically relevant or, indeed, if the fragment is degraded by the N-end rule pathway in vivo. In order to establish if the C-terminal fragment of Dps is a bona fide substrate of the N-end rule pathway, it will be important to identify the protease responsible for this processing step.

Modification of the N-terminus: new enzymes and activities

  1. Top of page
  2. Summary
  3. Introduction
  4. The N-end rule: what is an N-degron?
  5. Bacterial N-degrons: how are they generated?
  6. Modification of the N-terminus: new enzymes and activities
  7. N-degron recognition by ClpS, the bacterial N-recognin
  8. ClpAP, an ATP-fuelled machine of destruction (mechanism of action)
  9. N-degron delivery and degradation: more than recognition of the N-terminal residue
  10. Substrates and the physiological function of the N-end rule pathway
  11. Compartmental protein turnover by the N-end rule pathway?
  12. Acknowledgements
  13. References

The non-ribosomal addition of amino acids onto proteins was first identified in E. coli over 40 years ago (Kaji et al., 1965). Several years later, LFTR was identified as the protein responsible for the non-ribosomal transfer of Leu or Phe from tRNA to protein (Leibowitz and Soffer, 1969). However, it was a further 20 years before the activity of LFTR was linked to the metabolic stability of its protein substrates (Tobias et al., 1991). During this time, several detailed studies defined both the donor and acceptor specificity of LFTR using a variety of model proteins and peptides. These data demonstrated that LFTR is responsible for the attachment of Leu and Phe, and to a lesser extent Met and Trp, to an N-terminal Arg or Lys residue of a substrate. A structural explanation for this biochemical specificity was recently provided when the structure of EcLFTR was solved in complex with several different ligands (Suto et al., 2006; Watanabe et al., 2007). Despite the extensive biochemical and structural analysis of LFTR using model proteins and peptides, a recent study (Ninnis et al., 2009) identified the unforeseen observation that the N-terminal Met of the bacterial protein PATase is modified by LFTR (Fig. 3A). Importantly, the stability of PATase in vivo was dependent on the activity of LFTR. This unheralded specificity poses several important questions. For example, why does LFTR recognize and modify the N-terminal Met of PATase but not that of the model fusion protein, M-β-galactosidase? How can a broader specificity of LFTR be reconciled with the solved crystal structures of the complex? The structure of EcLFTR contains three deep binding pockets, one (d1) for the donor amino acid and two (a1 and a2) for the first two residues of the protein acceptor (Fig. 3B and C). The d1 binding site forms a hydrophobic pocket that is responsible for the recognition of the aminoacyl moiety of the aminoacyl-tRNA. This pocket can accommodate most unbranched hydrophobic amino acids but displays a preference for the binding of Leu and Phe (Abramochkin and Shrader, 1996; Suto et al., 2006). The first acceptor pocket (a1) recognizes the N-terminal acceptor residue in a sequence dependent manner, while the second acceptor pocket (a2) is responsible for the sequence independent recognition of the second amino acid of the acceptor polypeptide (Fig. 3B). Specifically, in the case of the FRYLG peptide bound complex of LFTR, the positively charged arginine residue (which represents the first residue of the acceptor protein) is bound in the negatively charged a1 pocket (Watanabe et al., 2007) and, consequently, the charge of the a1 binding pocket might be an important feature of enzyme specificity. Nevertheless, given that the negative charge of this pocket is largely incompatible with the binding of Met, it is interesting to speculate that, in some cases (e.g. where Met is the N-terminal acceptor residue), favourable recognition of the second amino acid of the acceptor protein (by the a2 pocket) compensates for unfavourable binding in the a1 pocket. Consistent with this notion (M)-PATase, but not M-β-galactosidase, is recognized by LFTR, suggesting that the binding of Asn (the second acceptor residue of PATase) in the a2 pocket is energetically more favourable than the binding of His (the second acceptor residue of the β-galactosidase chimeras) to the same pocket. Indeed, several amino acids, not only Asn, might bind favourably to the a2 pocket and overcome unfavourable binding to the a1 pocket. Therefore, determining the specificity of the a2 pocket might be crucial to further unravelling the activity of LFTR and better defining its role of the N-end rule pathway in bacteria.

image

Figure 3. Specificity of L/F-tRNA-protein transferase and other l-transferases. A. The natural substrate PATase is recognized and modified by LFTR. In this case a primary destabilizing residue (e.g. L) is attached to the N-terminal acceptor residue (M). B. Electrostatic potential of the binding pocket of EcLFTR (PDB: 2Z3L) in complex with a peptide (FRY, shown in yellow). C. Cartoon of the binding pocket of LFTR [showing the donor binding pocket (d1) and two acceptor binding pockets (a1 and a2)]. Experimentally the a1 pocket can bind to the secondary destabilizing residues (R, K and M). In the case of the basic secondary destabilizing residues (R and K), the a2 pocket is proposed to lack specificity (X = any amino acid). In contrast, when methionine (M) is bound to the a1 pocket the specificity of the a2 pocket is proposed to be limited (? = selected amino acids including N), although currently it is known to accommodate N. D. A sequelogue of LFTR, from P. falciparum, is responsible for the attachment of a type 1 primary destabilizing residue (R) to the acid acceptor residues (D and E). E. A sequelogue of the eukaryotic R-transferase, from the pathogen V. vulnificus, is responsible for the addition of a type 2 primary destabilizing residue (L) to the acidic acceptor residues (D and E).

Download figure to PowerPoint

Interestingly, two other aminoacyl-transferases have also been shown to have unusual specificity (Graciet et al., 2006). The first enzyme, PfATEL1 from the parasite Plasmodium falciparum, is a sequelogue of E. coli LFTR. Although the sequence of PfATEL1 is homologous to bacterial LFTRs, it exhibits the activity of mammalian ATE1 and, hence, mediates the addition of Arg to acidic secondary destabilizing residues (Fig. 3D). The second enzyme, bacterial protein transferase (Bpt), was identified in Vibrio vulnificus and, in contrast to PfATEL1, is a sequelogue of human ATE1 (Graciet et al., 2006). Nevertheless, Bpt displays a hybrid activity, exhibiting the acceptor specificity of mammalian ATE1 and the donor specificity of bacterial LFTR; hence, it catalyses the attachment of Leu (rather than Arg) to acidic N-terminal acceptor residues (Fig. 3E). This assortment of newly identified specificities suggests that the binding site of these enzymes is more malleable than first thought and begs the question of whether more aminoacyl-transferases specificities remain to be identified.

N-degron recognition by ClpS, the bacterial N-recognin

  1. Top of page
  2. Summary
  3. Introduction
  4. The N-end rule: what is an N-degron?
  5. Bacterial N-degrons: how are they generated?
  6. Modification of the N-terminus: new enzymes and activities
  7. N-degron recognition by ClpS, the bacterial N-recognin
  8. ClpAP, an ATP-fuelled machine of destruction (mechanism of action)
  9. N-degron delivery and degradation: more than recognition of the N-terminal residue
  10. Substrates and the physiological function of the N-end rule pathway
  11. Compartmental protein turnover by the N-end rule pathway?
  12. Acknowledgements
  13. References

The recognition of an N-degron is performed by a group of proteins collectively known as N-recognins (N-degron recognition component). In bacteria, a single N-recognin (the adaptor protein, ClpS) is responsible for the recognition of primary destabilizing residues, while in humans at least four distinct N-recognins (UBR1, UBR2, UBR4 and UBR5) have been identified (Tasaki et al., 2005; 2009; Erbse et al., 2006). The bacterial N-recognin, ClpS, is a small (12 kDa) ‘cone-shaped’ protein, composed of two distinct regions: a short N-terminal region that forms a largely extended structure, and a globular C-terminal domain that contains two conserved regions, one for interaction with ClpA and the other a hydrophobic pocket for recognition of type 2 (L, F, Y and W) primary destabilizing residues (Guo et al., 2002; Zeth et al., 2002; Erbse et al., 2006; Wang et al., 2008a; Schuenemann et al., 2009) (Fig. 4A). Indeed the recognition of type 2 primary destabilizing residues is an essential step in the degradation of these substrates, which is solely performed by the adaptor protein ClpS (Erbse et al., 2006; Ninnis et al., 2009; Schmidt et al., 2009). In contrast to ClpS, eukaryotic N-recognins are large (> 200 kDa) multi-domain proteins that invariably contain a UBR box motif for the recognition of type 1 (R, K and H) destabilizing residues and exhibit E3 ligase activity for the ubiquitylation of these substrates (Figs 1 and 4B). In some cases (e.g. UBR1 and UBR2), they also contain a ClpS-like domain for the recognition of type 2 (L, F, Y, W and I) destabilizing residues (Figs 1 and 4B). Although the specificity of these binding pockets (type 1 and type 2) has long been defined, only recently was the atomic detail of the type 2 binding pocket (from bacterial ClpS) determined.

image

Figure 4. ClpS, a molecular switch for N-degron recognition and delivery. A. ClpS is composed of two distinct regions, a C-terminal globular domain and an extended N-terminal region. The C-terminal domain is involved in two separate functions; region A is required for recognition of the ClpA N-domain and distal to region A there is a hydrophobic pocket for the recognition of an N-degron (LV, shown in green). B. Several, N-degron binding residues in E. coli ClpS (EcClpS) are conserved in other bacterial [Caulobater crescentus ClpS (CcClpS) Q9A5I0] and eukaryotic [Saccharomyces cerevisiae UBR1 (Sc UBR1), P19812; Human UBR1 (Hs UBR1), Q8IWV7 and Human UBR2 (Hs UBR2) Q8IWV8] N-recognins. Human UBR1 contains a UBR box motif for the recognition of type 1 primary destabilizing residues (blue), a ClpS-like domain (ClpS domain) for the recognition of type 2 primary destabilizing residues (pink), a basic residue rich (BRR) region (cyan) and a really interesting new gene (RING) domain (red) for ubiquitylation of the substrate protein. C. Superposition of ClpS in complex with various N-degrons: E. coli ClpS in complex with l-peptide (PDB: 2W9R, pink), E. coli ClpS in complex with F-peptide (PDB: 2W8A, orange), C. crescentus ClpS in complex with Y-peptide (PDB: 3DNJ, red) and C. crescentus ClpS in complex with W-peptide (PDB: 3GQ1, blue). D and E. The type 2 N-degron binding site of (D) E. coli ClpS (PDB: 2W9R) and (E) C. crescentus ClpS (PDB: 3GI9). Residues are shown in stick representation for EcClpS (yellow) and CcClpS (blue) and the N-degron peptide in green (EcClpS) and orange (CcClpS). Figures were prepared using the program pymol (http://www.pymol.org).

Download figure to PowerPoint

These structures of ClpS (from E. coli and Caulobacter crescentus) were solved in complex with a complete set of bacterial type 2 destabilizing residues (L, F, Y and W) (Wang et al., 2008a; Roman-Hernandez et al., 2009; Schuenemann et al., 2009). Although the three-dimensional structure of E. coli ClpS was already known and general aspects of substrate binding by ClpS had been proposed (Guo et al., 2002; Zeth et al., 2002; Lupas and Koretke, 2003; Erbse et al., 2006), these structures revealed the first atomic details of the interaction between an N-recognin and its substrate. The N-degron is bound in a small hydrophobic cavity located on the surface of ClpS, distal to the ClpA binding site (Fig. 4A). Although this pocket is relatively small in volume (approximately 200 Å3) it is large enough to accommodate the side-chain of Leu, Phe or Tyr and can expand to accommodate the side-chain of Trp (Fig. 4C). The mouth of the pocket is lined with negatively charged (D35, D36) and polar (N34, T38, H66) residues (Fig. 4A and D). These residues form a network of hydrogen bonds with the backbone atoms of the first two residues of the N-degron (Fig. 4D and E). The arrangement is consistent with an important role for the α-amino group of the N-degron, as was previously shown using peptide libraries (Erbse et al., 2006). Indeed, this network of hydrogen bonds also provides a molecular explanation for the relatively high affinity of N-degron binding (Kd∼0.2–4 µM, for an N-degron containing Phe at the N-terminus) given the size of the interface (∼200 Å2 buried surface area).

Interestingly, although the structures of E. coli and C. crescentus ClpS (2W9R and 3GI9 respectively) are essentially the same (r.m.s.d. = 0.9 Å), the details of N-degron binding vary. In the C. crescentus complexes, the α-amino group of the N-degron is fully protonated and forms three H-bonds, two with ClpS residues N47 (N34 in E. coli) and H79 (H66 in E. coli) and one with a water molecule (Wang et al., 2008a; Roman-Hernandez et al., 2009) (Fig. 4E). In contrast, the α-amino group of the N-degron in the E. coli complex forms only two H-bonds, one with ClpS (N34) and one with a water molecule (Schuenemann et al., 2009). In this case, rather than forming a third H-bond with the α-amino group of the N-degron, H66 forms an H-bond with the carbonyl oxygen of the N-degron (Fig. 4D). Interestingly, consistent with a central role for the α-amino group in N-degron binding, replacement of N34 by alanine abolishes the ClpS-mediated binding and degradation of a model N-end rule substrate (Wang et al., 2008a). Surprisingly, replacement of H66 to alanine was reported to have a less dramatic effect than the N34A substitution, but binding was clearly reduced (Wang et al., 2008a). This small but significant difference in the effect of these two proteins variants lends support to the idea that N34 and H66 participate in different atomic interactions, but further analysis of the interface is required to define a role for H66 in N-degron binding. Nevertheless, given that N-degron dipeptides (e.g. FR) but not primary destabilizing amino acids (e.g. F, L, Y or W) can inhibit N-degron binding (Baker and Varshavsky, 1991; Ninnis et al., 2009; Schmidt et al., 2009), the residue involved in co-ordinating the carbonyl oxygen is likely to play an important role in substrate recognition. Interestingly, although the majority of these polar residues are conserved between bacterial and eukaryotic N-recognin protein sequences, two residues (H66 and T38) are found only in bacterial ClpS sequences. In contrast the eukaryotic N-recognin residues equivalent to T38 and H66 in E. coli ClpS have been replaced by His and Asp respectively. These modifications to the binding pocket might either alter the orientation of the bound substrate or expand substrate specificity of eukaryotic N-recognins (Fig. 4B).

Although recognition of the N-degron backbone is a critical part of substrate binding, the ‘fit’ of the N-degron side-chain into the binding pocket also plays an important role in fine-tuning binding affinity and specificity. For example, an N-terminal Phe exhibits the best ‘fit’ to the pocket, the highest affinity and the most rapid degradation while, in contrast, an N-degron containing an N-terminal Trp residue exhibits the weakest affinity and the slowest degradation rate (Schuenemann et al., 2009). Consistent with the notion that the volume of the pocket is an important determinant of substrate binding, two hydrophobic residues (M40 and M62) that line the cavity modulate the volume of the cavity and, hence, its specificity. Substitution of one residue (M40A) was sufficient to expand the binding repertoire of ClpS to include Ile, albeit weakly, while replacement of both residues (M40A/M62A) improved the recognition, and hence degradation of a substrate bearing an N-terminal Trp (Wang et al., 2008a; Schuenemann et al., 2009; S. Kralik, K. Zeth and D.A. Dougan, unpublished). Nevertheless, despite this structural plasticity, the binding pocket is unable to accommodate the side-chain of Met or β-branched amino acids such as Ile and Val. Interestingly, neither M40 nor M62 is conserved in eukaryotic N-recognins and, hence, these residues might contribute to the subtle yet distinct specificity exhibited by the eukaryotic UBRs. Collectively, the biophysical measurements of substrate binding favour the recognition of proteins bearing an N-terminal Leu or Phe (and not amino acids containing a β-branched side-chain), which is entirely complementary with the biochemical and physiological role of LFTR.

ClpAP, an ATP-fuelled machine of destruction (mechanism of action)

  1. Top of page
  2. Summary
  3. Introduction
  4. The N-end rule: what is an N-degron?
  5. Bacterial N-degrons: how are they generated?
  6. Modification of the N-terminus: new enzymes and activities
  7. N-degron recognition by ClpS, the bacterial N-recognin
  8. ClpAP, an ATP-fuelled machine of destruction (mechanism of action)
  9. N-degron delivery and degradation: more than recognition of the N-terminal residue
  10. Substrates and the physiological function of the N-end rule pathway
  11. Compartmental protein turnover by the N-end rule pathway?
  12. Acknowledgements
  13. References

In E. coli, a single two-component proteolytic machine (ClpAP) is responsible for the degradation of proteins bearing an N-terminal destabilizing residue (Tobias et al., 1991). The peptidase (ClpP) is a barrel shaped oligomer composed of two heptameric rings, stacked back-to-back (Wang et al., 1997). The active sites are located within an internal chamber, near the interface between the two rings and substrate entry into this chamber is restricted to a narrow portal (10 Å in diameter) on top of the heptameric ring. Upon hydrolysis of the substrate, the peptide products are released through side pores in the cylinder at the interface between the two rings (Sprangers et al., 2005). Given that entry into the chamber is limited to a narrow portal, ‘folded’ proteins require assistance to enter the chamber. In the case of ClpP, this ‘unfolding’ assistance comes in two forms (ClpA or ClpX), with ClpA being the unfoldase of choice for proteins bearing an N-terminal destabilizing residue. Like most other peptidase-associated unfoldases, ClpA is a member of the AAA+ (ATPases associated with various cellular activities) superfamily of proteins (Neuwald et al., 1999) that form a ring shaped hexamer in the presence of ATP. The hexameric rings of ClpA interact with ClpP at one or both ends of the ClpP cylinder to form asymmetric (1:1) or symmetric (2:1) ClpAP complexes. Interestingly, symmetric complexes of ClpAP are most efficient at processing substrates (Maglica et al., 2009).

ClpA is composed of two AAA+ domains (often referred to as D1 and D2), both of which contain several important motifs and loops, including those for the binding (Walker A) and hydrolysis (Walker B) of ATP (Gottesman et al., 1990), the pore-1 loop for recognition and translocation of substrates (Hinnerwisch et al., 2005) and the IGF/L and pore-2 loops for interaction with ClpP (Martin et al., 2007). Despite high sequence conservation of these AAA+ domains, both appear to serve different functions. The first domain (D1) is crucial for oligomerization, while the second domain (D2) is primarily responsible for ATP hydrolysis (Singh and Maurizi, 1994). Interestingly, variants of ClpA lacking ATPase activity in either D1 or D2 are only able to process substrates with ‘intermediate’ or ‘low’ local stability, respectively, suggesting that each domain can function independently, at least to a limited extent (Kress et al., 2009). However, the ATPase activity of both domains is required for the efficient processing of substrates with ‘high’ local stability (Kress et al., 2009), indicating that both domains work together to unfold and translocate substrates into ClpP.

Like many other AAA+ proteins, ClpA also contains an additional domain that, in this case, is located at the N-terminus. This N-terminal domain (N-domain), although dispensable for some ClpA activities, plays a crucial role in substrate specificity. Not only is it essential for docking of the adaptor protein ClpS and, hence, the delivery of N-end rules substrates, but it also modulates the processing of other substrates (Lo et al., 2001; Zeth et al., 2002; Dougan et al., 2002b; Erbse et al., 2008). Indeed, flexibility of these domains has been proposed to regulate substrate access to the pore of ClpA and, hence, contribute to substrate processing (Cranz-Mileva et al., 2008). Interestingly, substitution of two highly conserved residues within the N-domain dramatically alters the processing of certain substrates without adversely affecting the binding (Erbse et al., 2008).

N-degron delivery and degradation: more than recognition of the N-terminal residue

  1. Top of page
  2. Summary
  3. Introduction
  4. The N-end rule: what is an N-degron?
  5. Bacterial N-degrons: how are they generated?
  6. Modification of the N-terminus: new enzymes and activities
  7. N-degron recognition by ClpS, the bacterial N-recognin
  8. ClpAP, an ATP-fuelled machine of destruction (mechanism of action)
  9. N-degron delivery and degradation: more than recognition of the N-terminal residue
  10. Substrates and the physiological function of the N-end rule pathway
  11. Compartmental protein turnover by the N-end rule pathway?
  12. Acknowledgements
  13. References

Although the recognition of an N-terminal destabilizing residue, by ClpS, is an essential step in the degradation of N-end rule substrates (Erbse et al., 2006; Ninnis et al., 2009; Schmidt et al., 2009; Schuenemann et al., 2009), this step alone is not sufficient to initiate substrate turnover. Indeed, the efficient degradation of a substrate bearing an N-terminal destabilizing residue is dependent on the presence of a short unstructured region between the folded domain and the N-degron (Erbse et al., 2006). A recent systematic analysis of this linker region demonstrated that at least four amino acids between the folded domain and the N-degron are required for efficient delivery of the substrate to ClpA (Wang et al., 2008b). To date, however, it remains unclear if the length requirement of the linker region is affected by the intrinsic stability of the folded protein or the local secondary structure (i.e. the region downstream of the linker). Interestingly, the composition of the linker affects substrate turnover (Erbse et al., 2006; Wang et al., 2008b; Ninnis et al., 2009). The presence of negatively charged residues near an N-terminal destabilizing residue disfavours the binding of ClpS (Erbse et al., 2006). Consistent with this finding, natural ClpS substrates lack negatively charged residues adjacent to the N-degron (Ninnis et al., 2009; Schmidt et al., 2009). Importantly, this specificity of ClpS is compatible with the activity of LFTR in E. coli although, surprisingly, it is somewhat inconsistent with the recently identified activity of Bpt in V. vulnificus. Hence, it will be interesting to see if the specificity of VvClpS reflects this subtle change in transferase activity.

Aside from ClpS binding, the composition of the linker region also contributes to other aspects of N-degron turnover. Detailed analysis of several substrates (both model and natural) eventually identified an additional feature (a hydrophobic element) located within this unstructured linker region (Ninnis et al., 2009). This element is generally located between 6 and 12 residues downstream of the N-degron, and is required for substrate delivery to the ClpAP complex, possibly through the direct interaction with the pore-1 loop of ClpA. Consistent with a role in substrate delivery, truncation of the linker region or deletion of the hydrophobic element did not affect ClpS binding but prevented substrate delivery to ClpA and degradation by ClpP. Moreover, in a competition assay, substitutions in the hydrophobic element in one substrate hindered the delivery of the other substrate, suggesting that substrate handover, to ClpA, might be an active process involving the hydrophobic element. In some ways, this hydrophobic element is analogous to the internal Lys residue present in eukaryotic N-end rule substrates, which is required for the ubiquitylation of the substrate and, hence, its downstream processing by the 26S proteasome (Varshavsky, 1996).

Collectively, the current data suggest a model for the delivery of an N-degron, by ClpS, to the ClpAP protease that involves a number of crucial steps (Fig. 5). First, a type 2 primary destabilizing residue is recognized by a hydrophobic pocket located on the surface of ClpS. Next, ClpS, together with its bound substrate, docks to ClpA in a two-step process. The initial docking phase occurs between the N-terminal domain of ClpA and the C-terminal globular domain of ClpS (Dougan et al., 2002b; Guo et al., 2002; Zeth et al., 2002). Although this initial docking step (which does not require the N-terminal region of ClpS) is sufficient to prevent the degradation of SsrA-tagged proteins by ClpAP, it is unable to mediate the degradation of N-end rule substrates (Dougan et al., 2002b; Guo et al., 2002; Hou et al., 2008; De Donatis et al., 2010). In this ‘initial’ complex, ClpS is likely to be bound to the N-domain, located on the periphery of the hexamer, distant from the hexameric pore of ClpA. Movement of the flexible ClpA N-domains places the bound ClpS (together with its substrate) in close proximity to the pore of hexameric ClpA. This step brings the N-terminal region of ClpS into contact with a second, as yet unknown site in ClpA, which is proposed to trigger a conformational change in ClpA and/or ClpS (Zeth et al., 2002; Hou et al., 2008) resulting in the formation of a high-affinity complex between a single molecule of ClpS and a hexamer of ClpA (De Donatis et al., 2010). The conformational change in ClpA is required for substrate delivery, and presumably for recognition of the hydrophobic element by ClpA. The ‘committed’ phase in ClpS binding (to ClpA) might trigger a conformational change in ClpS, which acts as a switch or timing mechanism to co-ordinate substrate release from ClpS, with substrate delivery to ClpA. When ClpS commits to ClpA binding, then N-degron delivery is turned on; however, prior to reaching the committed step, N-degron delivery is turned off. Interestingly, the ‘committed’ binding of RepA was recently shown to prevent the ClpS-dependent delivery of a model N-end rule substrate and visa versa (De Donatis et al., 2010). These data suggest that RepA and N-end rule substrate use similar elements in ClpA for substrate processing. Nevertheless, further experiments are needed to determine if this is indeed the case.

image

Figure 5. ClpS-mediated N-degron delivery to ClpAP. ClpS is critical for the recognition and delivery of all N-end rule substrates. To date all N-end rule substrates contain three important features: (i) an N-terminal primary destabilizing residue for recognition by ClpS, (ii) a hydrophobic element, downstream of the N-degron, for delivery to ClpA, and (iii) an unstructured linker region between the N-degron and hydrophobic element to facilitate substrate delivery. Following docking to the N-domain, the extended N-terminal region of ClpS ‘locks’ onto ClpA through a second, as yet unidentified binding site within ClpA. In the absence of this ‘high-affinity’ binding to ClpA, ClpS is unable to delivery N-end rule substrates. One proposal is that ClpS binding triggers a conformational change in ClpA resulting in the exposure of a new binding site in ClpA for the recognition of ClpS-mediated substrates. This change facilitates the transfer of N-end rule substrates to ClpA, possibly through a direct interaction between the hydrophobic element on the substrate and the pore loops of ClpA, resulting in the translocation of the substrate.

Download figure to PowerPoint

Substrates and the physiological function of the N-end rule pathway

  1. Top of page
  2. Summary
  3. Introduction
  4. The N-end rule: what is an N-degron?
  5. Bacterial N-degrons: how are they generated?
  6. Modification of the N-terminus: new enzymes and activities
  7. N-degron recognition by ClpS, the bacterial N-recognin
  8. ClpAP, an ATP-fuelled machine of destruction (mechanism of action)
  9. N-degron delivery and degradation: more than recognition of the N-terminal residue
  10. Substrates and the physiological function of the N-end rule pathway
  11. Compartmental protein turnover by the N-end rule pathway?
  12. Acknowledgements
  13. References

Despite many recent advances in N-degron recognition and delivery, the physiological role of the N-end rule pathway in bacteria remains elusive. However, two separate studies have recently made progress towards this goal (Ninnis et al., 2009; Schmidt et al., 2009). These studies identified a number of candidate substrates, three of which, PATase, Dps and PutA (Proline utilization protein A), were common to both studies. As mentioned the metabolic stability of PATase is dependent on the N-end rule pathway (Ninnis et al., 2009). The steady-state levels of this enzyme increase, and its turnover is inhibited, in E. coli strains lacking in components of the N-end rule pathway (e.g. LFTR or ClpS). In E. coli, PATase catalyses the conversion of putrescine to l-glutamate and 4-aminobutanal via an aminotransferase reaction using 2-oxoglutarate as an amino group acceptor (Samsonova et al., 2003) and, as such, might contribute to cellular polyamine levels. Given that the level of polyamines is critical for a variety of cellular functions, and that high concentrations of putrescine are toxic, the turnover of the PATase by the N-end rule pathway might play an important role in putrescine homeostasis.

In contrast to PATase, the metabolic stability of truncated Dps (Dps6–167), the second potential substrate of the N-end rule pathway is uncertain. Nevertheless, it is intriguing to consider the turnover of Dps6–167 by the N-end pathway, given that full-length Dps (Dps2–167) and not Dps6–167 is a substrate of the ClpXP machinery (Flynn et al., 2003). Dps levels increase dramatically during stationary phase and in response to oxidative stress. The primary role of Dps is to protect DNA; it does so through two activities, one involving the direct interaction with DNA to form a physical barrier and the other its Fe-chelating activity to protect DNA from the toxic by-products of the Fenton reaction. Interestingly, removal of the first five residues of Dps (Dps6–167) abolishes its interaction with DNA and its turnover by ClpXP, but not its degradation by the ClpAPS machinery (Flynn et al., 2003; Ninnis et al., 2009; Schmidt et al., 2009). Therefore, dissociation of Dps from the Dps–DNA complex might be required after a prolonged stationary phase in order for the cell to reinitiate growth. In this case, Dps is processed by an unknown endopeptidase that renders it insensitive to ClpXP-mediated degradation and, hence, is removed by the N-end rule pathway. Consistent with this idea, ΔclpP cells exhibit a defect in stationary phase adaptation (Weichart et al., 2003).

The final example, PutA, is not verified as a substrate of the N-end rule pathway but was identified as a ClpS-interacting protein in both studies (Ninnis et al., 2009; Schmidt et al., 2009). PutA is a multifunctional protein that, in the presence of proline, associates with the membrane and acts as a proline catabolic enzyme while the cytosolic form of PutA regulates proline catabolism by repressing transcription of proline utilization genes (putA and putP). Interestingly, a mutant E. coli strain lacking LFTR activity exhibited increased proline catabolism (Deutch and Soffer, 1975), consistent with the turnover of PutA by the N-end rule pathway. However, the same group was unable to demonstrate a direct role for LFTR in modification of PutA (Scarpulla and Soffer, 1979); therefore, it remains unclear whether the metabolic stability of PutA is controlled by this pathway. PutA is composed of three separate regions, the N-terminal region (residues 1–261) is responsible for DNA binding, while the middle (residues 261–612) and C-terminal (residues 650–1130) regions are responsible for enzyme activity. Interestingly, the ClpS-interacting protein identified as PutA by Ninnis et al. (2009) was equivalent to the size of a PutA fragment lacking the N-terminal DNA-binding domain (∼100 kDa). Hence, in conditions where proline catabolism must be shut down while maintaining the transcription repression of put genes, the cleavage of the N-terminal DNA-binding domain of PutA by an unknown endopeptidase would permit the turnover of the C-terminal enzymatic domains by the N-end rule pathway.

Currently however, there is no obvious feature that links the N-end rule pathway to a single metabolic or stress response pathway. As such, the bacterial (or at least E. coli) N-end rule pathway might act to modulate the levels of different enzymes acting in quite distinct pathways. Nevertheless, it will be interesting to see which, if any, of the remaining ClpS-interacting proteins identified to date (Ninnis et al., 2009; Schmidt et al., 2009) are bone fide substrates of the N-end rule pathway.

Compartmental protein turnover by the N-end rule pathway?

  1. Top of page
  2. Summary
  3. Introduction
  4. The N-end rule: what is an N-degron?
  5. Bacterial N-degrons: how are they generated?
  6. Modification of the N-terminus: new enzymes and activities
  7. N-degron recognition by ClpS, the bacterial N-recognin
  8. ClpAP, an ATP-fuelled machine of destruction (mechanism of action)
  9. N-degron delivery and degradation: more than recognition of the N-terminal residue
  10. Substrates and the physiological function of the N-end rule pathway
  11. Compartmental protein turnover by the N-end rule pathway?
  12. Acknowledgements
  13. References

Although there have been several recent advances in dissecting the bacterial N-end rule pathway and the equivalent eukaryotic pathway of the cytosol, one area of N-end rule research that has received little attention is the compartmental or organellar N-end rule pathway. Currently, there is little experimental evidence to support the idea of an ‘authentic’ version of the N-end rule pathway in the organelles of eukaryotic cells, but a seemingly unrelated equivalent of this pathway was suggested to operated in yeast mitochondria (Vogtle et al., 2009). In contrast to mitochondria, sequelogues of several N-end rule pathway components have been identified in several plant species and various eukaryotic pathogens including P. falciparum (Dougan et al., 2002b; Lupas and Koretke, 2003; Yu and Houry, 2007; E. Valente and D.A. Dougan, unpublished) which are expected to be located in the chloroplast or apicoplast respectively. Consistent with the idea of an N-end rule pathway in these organelles, the cyanobacterium Synechococcus sp. was recently shown to contain a functional homologue of ClpS (SyClpS) that is able to bind to a model N-end rule substrate (Andersson et al., 2009). Interestingly, in contrast to gamma-proteobacteria, cyanobacteria lack ClpA and, thus, in this organism ClpS functions with an alternative AAA+ protease ClpCP that is closely related to the ClpCP protease of Bacillus subtilis and the ClpAP protease of E. coli (Andersson et al., 2009). It will be interesting to see if this proteolytic machine (ClpCP/ClpS) functions in a similar manner to the ClpAPS complexes of E. coli or the ClpCP/MecA complexes of B. subtilis (Kirstein et al., 2006; 2007). Nevertheless, it is likely that a ‘bacterial’ version of the N-end rule pathway operates in the chloroplast of plants and the apicoplast of apicomplexan parasites.

References

  1. Top of page
  2. Summary
  3. Introduction
  4. The N-end rule: what is an N-degron?
  5. Bacterial N-degrons: how are they generated?
  6. Modification of the N-terminus: new enzymes and activities
  7. N-degron recognition by ClpS, the bacterial N-recognin
  8. ClpAP, an ATP-fuelled machine of destruction (mechanism of action)
  9. N-degron delivery and degradation: more than recognition of the N-terminal residue
  10. Substrates and the physiological function of the N-end rule pathway
  11. Compartmental protein turnover by the N-end rule pathway?
  12. Acknowledgements
  13. References