Thank you for visiting nature.com. You are using a browser version with limited support for CSS. To obtain the best experience, we recommend you use a more up to date browser (or turn off compatibility mode in Internet Explorer). In the meantime, to ensure continued support, we are displaying the site without styles and JavaScript.

# Functional cooperativity between the trigger factor chaperone and the ClpXP proteolytic complex

### Subjects

An Author Correction to this article was published on 06 May 2021

## Abstract

A functional association is uncovered between the ribosome-associated trigger factor (TF) chaperone and the ClpXP degradation complex. Bioinformatic analyses demonstrate conservation of the close proximity of tig, the gene coding for TF, and genes coding for ClpXP, suggesting a functional interaction. The effect of TF on ClpXP-dependent degradation varies based on the nature of substrate. While degradation of some substrates are slowed down or are unaffected by TF, surprisingly, TF increases the degradation rate of a third class of substrates. These include λ phage replication protein λO, master regulator of stationary phase RpoS, and SsrA-tagged proteins. Globally, TF acts to enhance the degradation of about 2% of newly synthesized proteins. TF is found to interact through multiple sites with ClpX in a highly dynamic fashion to promote protein degradation. This chaperone–protease cooperation constitutes a unique and likely ancestral aspect of cellular protein homeostasis in which TF acts as an adaptor for ClpXP.

## Introduction

In the very crowded environment of the cell, molecular chaperones and proteases function to ensure proper protein homeostasis either by promoting the folding of newly synthesized proteins and maintaining the correct conformation of pre-existing proteins or by degrading misfolded and unfolded proteins, respectively. Generally, some chaperones such as trigger factor (TF) and the Hsp70 system (DnaK/DnaJ/GrpE) act early during the folding process, while chaperonins (GroEL/GroES), and Hsp90 (HtpG) act at later stages of folding1. In addition, there is redundancy in the function of folding chaperones such as that found for TF and DnaK in folding newly synthesized proteins2,3.

Escherichia coli TF is a 48 kDa protein containing an N-terminal ribosome-binding domain (TFN), a peptidyl-prolyl cis/trans isomerase (PPIase) domain (TFP), and a discontinuous domain with two arm motifs (TFA) (Fig. 1a)4. TF is highly abundant with an estimated concentration of ~50 µM in the cell5; it exists as a monomer and as a dimer. As a monomer, TF engages the ribosome in a 1:1 stoichiometry binding next to the exit tunnel of the 50 S large ribosomal subunit. This allows TF to interact with the nascent chain as it emerges from the ribosomal exit tunnel6,7. As a dimer, TF assists in various folding processes and has anti-aggregation activity8.

ClpXP is one of the main ATP-dependent protein degradation complexes in bacteria. It plays an essential role in protein quality control by removing misfolded, damaged, and regulatory proteins. The proteolytic complex consists of the tetradecameric serine protease ClpP9 and the hexameric ClpX unfoldase belonging to the AAA+ superfamily10,11. Substrate recognition by the ClpXP system occurs via specific recognition motifs, known as degrons, which are typically located at the N- or C-terminal ends of protein substrates. Some degrons are introduced by a specialized SsrA-tagging system which rescues stalled ribosomes12. Alternatively, adaptor proteins aid in the recruitment of substrates by tethering selected substrates to ClpX for degradation13. ClpX has two domains (Fig. 1a), an N-terminal zinc-binding domain (ZBD) and a C-terminal AAA+ domain that contains the characteristic Walker A and Walker B nucleotide binding and recognition motifs. ClpX binds to target substrates and uses the energy derived from ATP hydrolysis to unfold and thread substrates into the ClpP chamber through axial pores present at opposite ends of the ClpP cylinder. Substrates are generally cleaved into peptides of about 7–8 residues that exit the ClpP chamber through transient equatorial side pores14.

In phylogenetic studies of chaperones, we were intrigued to find that tig, the gene encoding the TF chaperone, was frequently found clustered with the clpPX operon in many species throughout the bacterial kingdom (Fig. 1a, Supplementary Fig. 1). This observation prompted us to seek out the functional link between TF and the ClpXP system. Unexpectedly, TF was found to enhance the ClpXP-dependent degradation of several substrates including the phage protein λO, the stationary phase sigma factor RpoS, and SsrA-tagged proteins both in vitro and in vivo. Based on results from biochemical and structural analyses, we propose that TF binds to ClpX by anchoring onto the AAA+ domain of the chaperone complex, thereby engaging the ZBD domain and modulating its movement towards the ClpP chamber. In doing so, TF accelerates the degradation of substrates. Our results suggest that TF facilitates ClpXP-mediated degradation of a select group of proteins and, hence, acts as an adaptor of the ClpXP system. This interplay between a folding chaperone and a major proteolysis complex was unexpected and provides new insights into bacterial proteostasis.

## Results

### Phylogenetic and genome analyses suggest a functional association between TF and ClpXP

In E. coli, clpP and clpX genes form the clpPX operon, while the tig gene constitutes its own operon that is located immediately upstream of the clpPX operon (Fig. 1a, Supplementary Fig. 1). In certain bacteria such as Bacillaceae, Staphylococcaceae, and Lactobacillaceae within the Firmicutes phyla and Gloeobacteraceae within the Cyanobacteria phyla, clpX and clpP are located at different loci but the tig gene is directly upstream of clpP (Supplementary Fig. 1). This initial observation prompted us to explore the genomic positioning of tig, clpP, and clpX in all available bacterial genomes. NCBI representative species and each selected genus were mapped onto the tree of life using the interactive tree of life tool15 (Supplementary Fig. 1). While all three genes are highly conserved and present throughout the bacterial kingdom; clpP and clpX were the only ones missing in Tenericutes. In Dehalococcoides sp., clpP is present without clpX, but clpA is found instead. In addition, we noticed that the lon gene encoding the Lon protease was also typically found close to clpP and clpX and always positioned downstream of them; so, we included lon in our analysis as well (Supplementary Fig. 1).

In general, the four genes tig, clpP, clpX and lon are almost always located next to each other in the order, tig-clpP-clpX-lon, and this cluster is conserved in Proteobacteria especially in species belonging to the Delta-, Beta-, and Gamma-classes (Supplementary Fig. 1). Some bacteria have tig next to clpX (tig-clpX), while the majority feature tig next to clpP (tig-clpP). Only in a small number of bacteria, the three genes are located far apart from each other, namely in Chlorobiaceae and Rickettsiales families, and in some members of Firmicute phyla. lon is also conserved but absent in the Staphylococcaceae, Streptococcaceae, and Mycobacteriaceae families16. Most notably, lon is clustered next to tig (tig-lon) in the ClpXP-lacking Mycoplasmataceae family.

Genes that are clustered together on the genome tend to have similar functions, are targeted to the same pathways, or are expressed in the same compartment of the cell17. Furthermore, the evolutionary conservation of such clustering is strong evidence for functional association18. Hence, the conserved genomic proximity led us to investigate potential functional and physical interactions between TF and ClpXP.

### TF modulates the ClpXP-dependent degradation rates of substrates

In order to test whether there is a functional relationship between TF and the ClpXP system, we conducted an in vitro degradation assay screen looking for the effect of TF on the rates of ClpXP-mediated degradation of three well-established ClpXP substrates: λO phage replication protein19, MuA phage transposase20, and GFP-SsrA21. In addition to TF, we also used the major E. coli chaperones DnaK/DnaJ/GrpE together, DnaK, DnaJ, GroEL/GroES and HtpG. DnaK and TF are known to have overlapping functions and recognize similar substrates2,3,22. Hence, DnaK serves as a control for the effect of TF on ClpXP-mediated degradation.

We observed three outcomes in our assays. First, for GFP-ssrA and λO substrates with DnaK/DnaJ/GrpE, DnaK, DnaJ, GroEL/GroES, and HtpG chaperones, no significant effect of the chaperone(s) on the degradation rates was observed (Fig. 1b, Supplementary Fig. 2a). TF also had no effect on FlhDC and TnaA (Supplementary Fig. 2b, c), two known substrates of ClpXP23,24. Second, the presence of chaperones slowed down the degradation rate as generally observed for reactions involving MuA, suggesting that chaperones interact with MuA and protect it from ClpXP degradation (Fig. 1b, Supplementary Fig. 2a). The same effect was seen with TF and the antitermination λ phage protein, λN-His (Supplementary Fig. 2d), and the fumarate nitrate reduction regulatory protein, Fnr (Supplementary Fig. 2e). Third, TF caused a significant increase in the rate of ClpXP-mediated degradation of λO. In addition to λO, we noticed a slight increase in the rate of GFP-ssrA degradation upon TF addition (Fig. 1b, Supplementary Fig. 2a). It should be noted that TF itself is not a substrate of ClpXP or ClpX∆ZBDP in the absence or presence of other ClpXP substrates (Supplementary Fig. 2f). ClpXΔZBD (ClpX AAA+) lacks the ZBD domain that is required for λO degradation25.

If TF strictly acts as a folding or holding chaperone in these assays, we would expect it to either have no effect or to protect the substrate from degradation by ClpXP, hence reducing the degradation rate. Alternatively, it is conceivable that TF might prevent the formation of small aggregates making substrates more accessible to ClpX and accelerating their degradation by ClpXP. However, there is no evidence of λO or GFP-SsrA aggregation under the conditions of our experiments. Hence, the enhancement of λO and GFP-SsrA ClpXP-dependent degradation is a unique result and suggests a functional association between the folding chaperone and the degradation complex for specific substrates.

To obtain a quantitative measure of the effect of TF on ClpXP-mediated degradation of λO, we took advantage of the higher number of intrinsic tryptophan residues within the λO protein sequence compared to other proteins in the reaction (λO: nine tryptophans, ClpX: none, ClpP: none and TF: one) and used this to measure the intrinsic tryptophan fluorescence change as a proxy for λO degradation. λO fluorescence decreased upon its degradation by ClpXP + ATP or the unspecific protease Proteinase K (Supplementary Fig. 3a). No significant change in fluorescence was observed for λO in the presence of TF alone plus or minus ATP, ClpXP without ATP, ClpXΔZBDP + ATP, or ClpX(E185Q)P + ATP (Supplementary Fig. 3a). ClpX(E185Q) has a mutation in the Walker B motif rendering it ATPase inactive26.

Subsequently, the degradation of λO by ClpXP + ATP was monitored by fluorescence in the presence of increasing concentrations of TF. As shown in Fig. S3B, the initial degradation rate of λO was enhanced in a TF-dependent manner. The titration curves yielded a Kd of 49 µM. In addition, by using increasing concentrations of the λO substrate in the presence of 1 µM ClpX, 1.2 µM ClpP, and 5 µM TF, we were able to determine the KM and kcat values for the degradation reaction (Fig. 1c, Supplementary Fig. 3c). TF reduced the KM for this substrate by a factor of about 2 and increased the kcat 1.3-fold. Hence, TF was able to increase the catalytic efficiency (kcat/KM) of ClpXP for λO by a factor of 2.6 indicating that it increases ClpXP selectivity toward λO (Fig. 1c).

Based on the above results, it is reasonable to expect that TF might modulate the ATPase activity of ClpX. It is well established that the ATPase activity of the chaperone is affected by ClpX adaptor proteins, substrates, and ClpP binding25,27. The presence of 5 μM TF stimulated the ATPase of 1 μM ClpX by about 20% compared to ClpX alone, while the control protein, bovine serum albumin, had no significant effect on the ClpX ATPase. In the presence of 3.9 µM of the λO substrate, we observe a 1.5- to 2-fold increase of ATPase activity with the highest activity produced when TF is present (Fig. 1d). In titration experiments, TF enhanced the ATPase activity of ClpX and ClpXP in the absence and presence of substrate and to a lesser extent that of ClpX∆ZBD, and ClpX∆ZBDP (Fig. 1e). Importantly, TF did not affect the ATPase activity of the related AAA + protein ClpA (Fig. 1e). These results show that TF stimulates the ATPase activity of ClpX mainly via the ZBD and this effect is independent of the ATPase increase produced by the λO substrate.

The TF single-domain (TFN, TFP, and TFA) and two-domain constructs (TFNP, TFNA, and TFPA) did not have a significant effect on the ClpXP-dependent degradation of λO (Supplementary Fig. 4a). Furthermore, they did not enhance the ATPase activity of ClpX, ClpX∆ZBD, or ClpA (Supplementary Fig. 4b), indicating that full-length TF is required for such an effect on ClpX.

### λO degradation is compromised in tig deletion mutant strain

In parallel experiments, λO stability was analyzed under more physiological conditions in WT, Δtig, and ΔclpX lysogens carrying λcI857Sam7 prophage28. Phage replication was induced by heat shock. Cells were then pulse-labeled with radioactive methionine and chased with nonradioactive methionine. λO protein level was monitored by immunoprecipitating the protein with anti-λO antibody from samples at the indicated time points during the chase period and then visualized by autoradiography (Fig. 2b). Consistent with the results for λO overexpressed from a plasmid, λO levels in the lysogen were lower in WT compared to Δtig lysogens at the early time points. λO was not degraded in the ∆clpX lysogen (Fig. 2b).

### The effect of TF on λO degradation is dependent on the presence of an exposed N-terminal ClpX-recognition sequence

It is known that the N-terminus of λO is required for the protein to be recognized by ClpXP19. Consistent with this, λO expressed from a plasmid with an N-terminal His-tag followed by a tobacco etch virus (TEV) protease cut site (HV-tag) is stable against degradation by ClpXP in vivo in the absence or presence of TF (Fig. 2c). As a result, TF cannot induce the ClpXP-dependent degradation of a substrate not exposing a ClpX recognition motif. Furthermore, in contrast to what was observed for λO, the levels of HV-λO at the zero-time point were the same in WT and Δtig strains, which strongly suggests that TF promotes the rapid degradation of newly synthesized λO protein. We address this observation below.

We further tested the requirement of the λO N-terminus for its ClpXP-dependent degradation in vitro. Consistent with in vivo results, there was no degradation of the HV-λO substrate in the absence or presence of the TF chaperone (Fig. 2d). To establish that the TF effect is exerted through the N-terminus of λO (NλO, residues 1–12), we fused NλO to the N-terminus of a poor ClpXP substrate, the 30S ribosomal protein S7. The degradation of S7 by ClpXP is slowed down in the presence of TF both in vitro and in vivo (Fig. 2d, Supplementary Fig. 5c). Similarly, a fusion of NλO with S7 (NλO-S7) was still a poor ClpXP substrate and was degraded at a similar rate as S7 alone; however, NλO-S7 degradation in vitro was enhanced by about tenfold in the presence of TF (Fig. 2d). This demonstrates that an exposed ClpX-recognition sequence in the substrate is critical for enhancement of its degradation by TF.

### TF enhances the ClpXP-dependent degradation of λO by promoting the downward movement of the ZBD dimer (ZBD2) towards ClpP

We had earlier demonstrated that ZBD2 undergoes a nucleotide-dependent movement between two conformational states, one of which positions ZBD2 away from ClpP and on top of the AAA+ ring, while the other state positions ZBD2 closer to ClpP (Fig. 3a)29. We named these two states distal and proximal. This movement of ZBD2 was also modulated by the adaptor SspB. We argued that this ZBD2 movement promotes substrate translocation and subsequent degradation by ClpP. Therefore, we tested whether TF modulates this movement.

We took advantage of an HV-tagged ClpX construct (refer to Fig. 3b for sequence) that we had previously used as a proxy for ZBD2 movements29. In these assays, a lower molecular weight ClpX is produced when HV-ClpX is clipped by bound ClpP in active ClpXP complexes in the presence of ATP29. Interestingly, this clipping of HV-ClpX was drastically enhanced by TF (Fig. 3b). With the addition of ClpP and ATP, the rate of disappearance of HV-ClpX was >2× faster in the presence of TF than in its absence (Fig. 3b). Interestingly, the enhanced clipping of the HV-ClpX by TF was seen even in the absence of ATP and was further confirmed using the ATPase-deficient HV-ClpX(E185Q) Walker B mutant26 (Fig. 3b, c). A ClpX mutant unable to bind to ClpP, HV-ClpX(F269W)30, was not significantly degraded or clipped in the presence of TF and wild-type ClpXP (Fig. 3d), demonstrating that clipping is not from free HV-ClpX29. In our previous study, we found that mutations in two conserved proline residues (P64A and P66A) located between the ZBD and the AAA+ domains of ClpX affect nucleotide-dependent changes transmitted from the AAA+ to the ZBD29. Hence, we checked if the enhancement of clipping by TF was through these proline residues (Supplementary Fig. 5d). Our results show that TF was still able to enhance HV-ClpX(P64A/P66A) clipping, suggesting a mechanism that does not act through these residues and, importantly, that the PPIase activity of TF is not needed for this activity. Together, our results show that TF promotes the movement of ZBD2 from the distal to the proximal state even in the absence of ATP (Fig. 3a).

We hypothesize that TF binds an individual ZBD2 and allows the unbound ZBD2s to switch to a proximal state and that this would lead to the enhancement of ClpXP-dependent λO degradation. To test this, we prepared complexes of ClpXP with different combinations of WT ClpX and ClpX∆ZBD protomers whose ZBD2 composition per complex can be predicted using combinatorial statistics (see “Methods”). ClpX molecules were obtained by mixing subunits of ClpX and ClpX∆ZBD at different ratios. The initial rates of λO degradation in the absence of TF were obtained (Fig. 3e). We observed an increase in the rate of λO degradation upon increasing ClpX∆ZBD in the complex with maximal degradation rate obtained when the population of complexes consisted of 17% without ZBD2, 13% with one ZBD2, 13% with two ZBD2, and 17% three ZBD2 (Fig. 3e, Supplementary Fig. 5e). The degradation rate then decreased upon further increasing the ClpX∆ZBD concentration, and no λO degradation was observed in the absence of ZBD (Fig. 3e, Supplementary Fig. 5e) consistent with the fact that ZBD is needed for λO degradation25. These results demonstrate that reducing steric hindrance among the three ZBD2 allows for faster translocation of λO through ClpX and into ClpP. Hence, we propose that TF acts by keeping one of the ZBD2s in a distal conformation and, as a result, promoting the proximal conformation of the other ZBD2s resulting in a faster λO degradation rate (Fig. 3a).

### Direct binding of TF to ClpX

Subsequently, we investigated whether TF directly binds to ClpX, ClpX∆ZBD (AAA+ domain), ZBD, ClpP, and the λO substrate using surface plasmon resonance (SPR) experiments. In these assays, TF was immobilized on the sensor chip and different proteins were injected over the chip (Supplementary Fig. 6). TF was found to interact with ClpX but not ClpP and, within ClpX, it interacted with both the ZBD and AAA+ domains (Supplementary Fig. 6). TF has a higher affinity to the AAA+ domain compared to ZBD (2 µM vs. 12.8 µM, respectively; Supplementary Fig. 6b). It should be noted that, given the oligomeric state of ClpX, the Kd obtained is an apparent dissociation constant, but it provides a qualitative assessment of the binding between TF and ClpX. TF bound λO with a slightly better affinity than to ClpX (3.8 µM vs. 9.3 µM, respectively; Supplementary Fig. 6b); however, this suggests that λO can partition or be transferred between the two systems.

To further investigate the multiple interactions between TF and ClpX or λO, nuclear magnetic resonance (NMR) titration experiments were performed. For higher sensitivity and resolution of the NMR spectra, TF was divided into two segments: TFPA (113–432) and TFN (1–117). Although full-length TF exists in a monomer–dimer equilibrium, TFPA and TFN were shown to exist as monomers in solution31. Thus, the perturbations are expected to be derived from direct interaction with the titrant and are distinct from the monomer–dimer transition. NMR spectra for isotopically labeled TFPA and TFN in the absence and presence of λO, ZBD, or ClpX∆ZBD (AAA+ domain) were acquired and the chemical shift perturbation (CSP) and intensity reductions (IR) were monitored (Fig. 4a–c, Supplementary Figs. 711). For TFPA, the resonances from backbone amide groups as well as methyl groups were monitored, whereas only those from backbone amide groups were monitored for TFN.

First, the interaction between TF and λO was investigated. The addition of λO induced significant CSP and IR to the resonances of TFPA (Supplementary Fig. 7). On the other hand, much less significant perturbations were observed for the resonances of TFN upon the addition of λO (Supplementary Fig. 8), indicating that λO is mainly recognized by P and A domains of TF. Mapping of the perturbations onto the structure of TF shows that the affected amino acid residues are located at multiple sites, especially at the P domain and Arm 1 of the A domain (Supplementary Fig. 7b, c), which are known to be involved in the recognition of client proteins31.

Next, we investigated the interaction between TF and ZBD or ClpX∆ZBD (Fig. 4a–c). The addition of ZBD induced significant perturbations to several resonances of TFPA (Fig. 4a, Supplementary Fig. 9). The mapping of the perturbed resonances on the structure of TF indicated that multiple regions of TF, including arm 2 and the hinge region connecting the TFP and TFA domains are involved in the recognition of ZBD (Fig. 4a). On the other hand, the resonances from TFN indicated very small perturbations (Supplementary Fig. 10). Thus, the NMR data indicate that TFA is responsible for the recognition of ZBD in ClpX.

The addition of ClpX∆ZBD induced significant perturbations to the resonances from TFPA (Fig. 4b, Supplementary Fig. 11a, b). More specifically, significant chemical shift changes and/or peak IR were observed for the resonances derived from the tip of the TFP, the hinge, arm 1, and arm 2, indicating that these regions of TF are responsible for the recognition of ClpX∆ZBD. On the other hand, the addition of ClpX∆ZBD to TFN induced minor chemical shift changes for the resonances located around the loop region, indicating that the loop region of TFN may weakly interact with ClpX∆ZBD (Fig. 4c, Supplementary Fig. 11c).

The NMR data show that the perturbed resonances are derived from multiple regions of TF. Although some of these perturbations can be derived from allosteric effects and not always indicate direct contacts, the data suggests that multiple regions of TF are involved in the binding of ZBD and AAA+ domains of ClpX. It is important to note that more regions of TF are perturbed by the addition of ClpX∆ZBD compared to ZBD, suggesting that the greater affinity of TF for the AAA+ domain obtained by SPR (Supplementary Fig. 6) can be attributed to more residues involved in the interaction. Notably, several hydrophobic amino acid residues of TF, which were identified to be involved in the interaction with ClpX, are also responsible for the recognition of the client protein λO (Supplementary Fig. 7), as well as, other unfolded proteins as shown in the solution structure of TF in complex with PhoA31 (Supplementary Fig. 12a, b). The overlap between ClpX-binding sites and the substrate-binding sites on TF suggests competition between the substrate protein and ClpX for binding to TF. Furthermore, some of the ClpX-binding sites on TF overlap with the interface of the TF dimer8 (Supplementary Fig. 12c), which suggests that TF binds to ClpX as a monomer.

To identify the TF-binding site on ZBD dimer, we measured the NMR for isotopically labeled ZBD in the absence and presence of TFPA (Fig. 4d, Supplementary Fig. 11d). Our data show that the addition of TFPA induced significant intensity reduction for the resonances originating from the residues located on the hydrophobic surface of ZBD (Supplementary Fig. 12d). This observation highlights yet again, the importance of the hydrophobic interaction between TF and ClpX.

### Peptide array analysis indicates binding between TF and the AAA+ domain of ClpX

Given the difficulty in obtaining and assigning the NMR spectra of ClpX∆ZBD, we further characterized the interacting motifs between ClpX AAA+ and TF using peptide array technology. Miniaturized libraries of overlapping 12mer peptides were synthesized by walking through the amino acid sequence of ClpX∆ZBD or TF, advancing two (ClpX∆ZBD) or three (TF) amino acids at a time (Supplementary Fig. 13). The TF peptide array was first probed with His-ClpX∆ZBD. Binding of His-tagged protein was detected using 6×His mAB-HRP. Peptides that were recognized directly by the antibody probe were eliminated from the analysis. TF peptides recognized by His-ClpX∆ZBD and containing at least two consecutive spots on the membrane were grouped, resulting in six consensus motifs (motifs 1–6) (Supplementary Fig. 13a, c). The sequence of the four isolated peptides, which did not have neighboring spots on the membrane, are also shown (Supplementary Fig. 13a). Consensus motifs and binding peptides were mapped onto the crystal structure of TF32 for visualization (Fig. 5a). The ClpX∆ZBD peptide array was probed in a similar manner using purified His-TF resulting in 12 consensus motifs (motifs 1–12) in ClpX∆ZBD (Supplementary Fig. 13B, D), which were mapped onto the cryoEM structure of ClpX∆ZBD33 (Fig. 5a). Together, our peptide array analyses are consistent with the NMR results (Fig. 4) and support a model whereby TFP and TFN predominantly bind the AAA+ domain of ClpX, and TFA mainly binds the ZBD.

### Mutagenesis confirms the binding sites between TF and ZBD of ClpX

To better understand the mechanism by which TF enhances λO degradation by ClpXP, we performed a mutagenesis screen to identify sites within TF that abolish this effect. A total of 38 individual or combined mutations of TF were generated mainly by replacing the target residue with mostly Ala (Supplementary Fig. 14a), and each mutant was tested in degradation assays. Eight TF mutants were found to significantly prevent the enhancement of ClpXP-dependent degradation of λO (Fig. 5b, Supplementary Fig. 14b). Consistent with our previous results from NMR and peptide array analysis, the location of these mutations are found at the tip of the P-domain, the hydrophobic pocket of arm 2, and the residues surrounding the hinge region connecting the P- and the A-domains (Fig. 5c). Interestingly, the TF(F198A) mutation known to abolish peptidyl-prolyl cis/trans isomerase (PPIase) activity34 behaved like wild-type TF suggesting that the PPIase function of TF is not required for enhancement of ClpXP activity (Supplementary Fig. 14b). This is consistent with our previous results (Supplementary Fig. 5d).

To test whether these residues affect chaperone function of TF, we performed anti-aggregation assays using GAPDH as a substrate. GAPDH denatured in guanidine was diluted into a solution lacking guanidine and the formation of GAPDH aggregates was monitored by light scattering. In the absence of TF, GAPDH forms aggregates as indicated by the rapid increase in light scattering, while the presence of TF significantly suppressed the formation of aggregates demonstrating the anti-aggregation activity of TF (Fig. 5d). No effect on anti-aggregation activity was observed for TF mutants K239A and L226A, while TF mutant E241A had a small change in anti-aggregation activity (Fig. 5d). Taken together, these results suggest that these three residues are important in the functional interaction between TF and ClpX (Fig. 5b, c), rather than the interaction between TF and its substrate.

Our previous results indicated that ZBD of ClpX is a critical binding site for TF (Figs. 1e, 3, and 4a, d); hence, we performed degradation assays of λΟ using ClpX variants with mutations in ZBD in the absence and presence of wild type TF. Out of the ten residues mutated, Phe16 and Ile46 located within the hydrophobic side pocket of the ZBD dimer were found to completely abolish the TF-enhancement of λO degradation by ClpXP (Fig. 5e, f, Supplementary Fig. 14d). This is also consistent with our NMR results (Fig. 4d). Importantly, these mutations did not affect ClpX activity as ClpX was still able to recognize and unfold λO for degradation by ClpP (Fig. 5e) in agreement with our previous work using these mutants35.

This observation highlights the importance of the hydrophobic interactions between TF and ClpX. Furthermore, the hydrophobic region encompassing F16 and I46 overlaps with the binding site of the λO substrate35 (Fig. 5f), suggesting that: (1) TF and λO might alternate in binding to the same ZBD2 or (2) TF and λO bind simultaneously to ClpX but use different ZBD2s within the hexameric complex.

### TF and ClpX adaptors use a similar mechanism to enhance λO degradation by ClpXP

Based on our previous data, we hypothesize that TF restricts the conformation of some but not all ZBD dimers upon binding and this, in turn, allows access of other ZBD2s into the pore of ClpXP (Fig. 3a). Some adaptor proteins, like SspB and RssB, have also been shown to bind the ZBD of ClpX and to regulate the ClpXP-mediated degradation of their cognate substrates36,37,38. Indeed, they bind to the same hydrophobic region bound by TF (Fig. 5f).

### TF facilitates the degradation of newly synthesized proteins in E. coli

To determine if TF has a global effect on the degradation of newly synthesized proteins in E. coli, we performed pulse-chase experiments in WT and Δtig cells under normal growth conditions (37 °C in LB media). The overall proteolysis of newly synthesized proteins was slower in the Δtig strain compared to WT (Fig. 6a). One hour after the chase, about 7.7% and 6.0% of newly synthesized proteins were degraded in WT and Δtig cells, respectively. The lower degradation capacity in Δtig strain was complemented by placing the tig gene under its own promoter on a plasmid back into Δtig cells (Δtig + ptig, Fig. 6a left panel). Evidently, TF is responsible for facilitating the degradation (rather than folding) of about 2% of newly synthesized proteins under regular growth conditions.

Under heat shock conditions (42 °C), there was an overall increase in the total amount of proteins degraded compared to 37 °C but the difference between WT and Δtig at 42 °C was similar to that at 37 °C (12.0% in WT and 10.0% in Δtig cells after 1 hr, Fig. 6a middle panel). This could possibly be related to the fact that TF is cold shock-induced rather than heat shock induced39. When protein misfolding was induced in cells by adding the proline analogue azetidine-2-carboxylate (AZC), the overall proteolysis of newly synthesized proteins after 1 h of the chase was 15.6% in WT and 11.8% in Δtig cells (Fig. 6a right panel), higher than that for normal growth conditions. Hence, under this condition, TF promoted the degradation of twice the number of newly synthesized proteins (about 4%) compared to the regular growth condition.

### Degradation of SsrA-tagged proteins is affected in tig deletion mutant strain

To test if TF has a global effect on the ClpXP-mediated degradation of SsrA-tagged proteins, we grew WT, ∆tig, ∆clpPX, and ∆smpB cells at 37 °C to mid-log phase (OD600 ~0.6); cells were then harvested, and total proteins were isolated using trichloroacetic acid (TCA) precipitation. The smpB gene encodes for the RNA chaperone essential for SsrA-tagging on stalled ribosomes40. Subsequently, western blot analysis was performed of equally loaded samples using an anti-SsrA antibody. As expected, the ∆smpB strain showed no accumulation of SsrA-tagged proteins. In contrast, ∆clpPX strain exhibited a significant accumulation of such proteins highlighting the fact that SsrA-tagged proteins are predominantly degraded by ClpXP under our experimental conditions (Fig. 6b). Importantly, the ∆tig strain exhibited about threefold higher accumulation of SsrA-tagged proteins compared to WT (Fig. 6b).

Since the ClpXP-dependent degradation of GFP-SsrA is also modulated by TF (Fig. 1b), we tested if that holds true in vivo. Overexpressed GFP-ssrA was degraded at a faster rate in WT compared to ∆tig strain (Fig. 6c). It is known that the SspB adaptor can enhance GFP-SsrA degradation by about tenfold41,42,43. In vivo, we found that the absence of sspB slowed GFP-ssrA degradation more than in the strain lacking tig (Fig. 6c). However, the double mutant (∆tigsspB) significantly accumulated more GFP-ssrA. These results strongly suggest that TF has a similar role as SspB for the degradation of SsrA-tagged proteins.

We confirmed the effect of TF on GFP-ssrA by performing degradation assays measuring fluorescence and, for comparison, in the presence of SspB (Fig. 6d). We noticed that, when TF and SspB were added together, the rate further increased suggesting they have an additive effect on ClpX in enhancing GFP-ssrA degradation. Taken together, our data suggest two possible scenarios: (1) TF may work together with the SspB adaptor in the delivery or processing of SsrA-tagged proteins by ClpXP, or (2) TF might solely be responsible for the ClpXP-dependent degradation of a small subset of SsrA-tagged proteins. In this regard, it is important to note that SspB activity was initially identified in partially purified ribosomes44.

The above results imply that the lack of both TF and SspB could lead to the accumulation of incomplete polypeptides in the cell, which should cause fitness defects. To test this, we checked the growth of MC4100 ∆tig, ∆sspB, and ∆tigsspB strains at 30 °C. There were no major differences in growth between the single mutants ∆tig and ∆sspB compared to WT cells (Supplementary Fig. 16a), suggesting that TF and SspB are not essential for cell growth at these temperatures. In contrast, the double mutant ∆tigsspB showed growth impairment that was restored by overexpressing tig under the control of its native promoter from a plasmid (Supplementary Fig. 16a). These results suggest that tig and sspB have a negative genetic interaction and, hence, likely function in related pathways.

### The ClpXP-dependent degradation of RpoS is also modulated by TF

To discover other ClpXP substrate proteins whose degradation is modulated by the TF chaperone, we used a catalytically inactive ClpP protease that would act as a trap (ClpPtrap) for substrates that are translocated into the protease chamber. The ClpPtrap was C-terminally tagged by a SPA-tag, which can be used for pulldown experiments45, and was expressed in E. coli MC4100 cells lacking the endogenous clpP as well as the clpA genes since ClpA can also act as a cap for ClpP. The smpB gene was deleted to prevent the formation of SsrA-tagged proteins that would also be trapped. Proteins that co-purified with ClpPtrap-SPA in ΔclpPΔclpAΔsmpB strain were compared to those identified in the ΔclpPΔclpAΔsmpBΔtig strain using mass spectrometry approaches. Control experiments were carried out using the WT ClpP-SPA construct.

Identified proteins from each background were filtered at a peptide count greater or equal to three and of high confidence (>95%) (Supplementary Fig. 16b). In total, 43 proteins were identified from the strain lacking TF and 42 from the strain expressing TF (Fig. 6e). From these, 15 were trapped only in the presence of TF including RpoS (also known as σS; Fig. 6e, Supplementary Fig. 16b), the master regulator of stationary phase46. RpoS is a well-established ClpXP substrate47.

### Degradation of ribosome-stalled λO is enhanced by TF

We showed above (Figs. 2a and 6a) that TF promotes the degradation of newly translated proteins. This could occur either cotranslationally or shortly after release from the ribosome. To further establish that TF stimulation of ClpXP-dependent degradation of λO can occur cotranslationally we performed the following experiment. The C-terminus of λO was fused to the SecM motif composed of 17 amino acids, which has been found to interact with the ribosomal polypeptide exit tunnel and to cause translational arrest51. We verified that λO-SecM is indeed arrested on ribosomes by fractionation on sucrose cushion. The degradation of the fusion protein was visualized by western blot analysis after inhibiting translation by tetracycline. Unlike the case for λO (Fig. 2a–c), the levels of λO-SecM were similar in WT, Δtig, and ΔclpX cells at the zero time point (Fig. 7a). However, like λO, the degradation of λO-SecM was slower in the absence of TF compared to WT (Fig. 7a).

To ensure that the degradation of λO-SecM occurred while the protein was attached to the ribosome, cell lysates at different time points of the experiment were separated on sucrose cushions (see “Methods”) and the ribosomal fractions were blotted for λO-SecM. Ribosome bound λO-SecM was degraded at a slower rate in Δtig compared to WT strain (Fig. 7b). Overall, these results indicate that the ClpXP-dependent degradation of λO-SecM is predominantly cotranslational and is facilitated by TF.

### TF complements for the lack of endogenous ClpX-adaptor function in other bacterial species

We noticed that the sspB gene is present in most Proteobacteria but absent in the rest of the bacterial kingdom (Supplementary Figs. 1 and 17a). In contrast, the tig gene is conserved. For example, the Gram-positive dental pathogen Streptococcus mutans does not contain sspB or SspB-like adaptors of ClpX but has tig, clpX, clpP, and smpB genes. Therefore, to demonstrate that TF could act as an ancestral adaptor of ClpX, we tested for any TF-dependent effects on the degradation of GFP-ssrA in S. mutans by performing in vitro degradation assays. The S. mutans SsrA tag contains a valine residue (VAA) instead of leucine in the last three residues of the ClpX recognition motif (Supplementary Fig. 17b); so, we used GFP with the S. mutans SsrA-tag fused to it at the C-terminus as a substrate. Our experiments show that in the absence of TFsm, degradation of 5 µM GFP-ssrAsm was slow (2 µM ClpXsm and 2.4 µM ClpPsm), but surprisingly when TFsm (15 µM) was added to the reaction, GFP-ssrAsm was degraded >2x faster (Supplementary Fig. 17b). GFP-ssrA containing the E. coli sequence was not recognized by ClpXPsm. This suggests that TF has an ancestral SspB-like function in S. mutans capable of modulating the ClpXP-mediated degradation of SsrA-tagged proteins.

## Discussion

In this work, we have demonstrated a functional interaction between TF and ClpXP. The presence of such a functional interaction was initially suggested by the close proximity of the tig gene to the clpX and clpP genes as well as lon across bacteria. Using purified proteins, we found that, while different chaperones had no effect or slowed down the degradation of ClpXP substrates, TF was the only chaperone capable of enhancing the degradation of λO, GFP-ssrA, and RpoS substrates. Indeed, we estimate that about 2% of newly synthesized proteins are degraded in a TF-dependent manner. Based on a kinetic analysis, TF acts to enhance the enzymatic selectivity of ClpXP for certain substrates such as λO and to increase the catalytic efficiency of the protease almost threefold. TF also enhances the ATPase activity of ClpX.

The presence of an exposed degron sequence on the substrate protein is important as well. The results of the in vitro and in vivo experiments of Fig. 2 demonstrate that the N-terminus of λO is required for the enhancement of its ClpXP-dependent degradation by TF. Fusing the N-terminus of λO to other proteins such as S7 also allowed for such enhancement to be observed. This requirement is essential for selective TF-mediated enhancement of ClpXP degradation as TF is highly abundant in the cell.

Interestingly, the experiments of Fig. 3 indicate that TF interacts with a ZBD dimer, thus allowing the other ZBD2s, which perhaps are bound to a substrate, to move freely towards the ClpP chamber. In doing so, we proposed that TF enhances the degradation of these substrates. Thus, the binding of TF to the ZBD primes ClpXP to degrade ZBD-bound substrates. Indeed, a similar effect is seen with the Caulobacter crescentus adaptor CpdR, where its binding to the ZBD primes ClpX to recruit downstream components52. Our results highlight the critical role of ZBD in the enhancement effect TF has on substrate degradation by ClpXP. ZBD is a well-known substrate and adaptor protein recognition hotspot on ClpX, and, hence, this further suggests that TF functions as an adaptor protein of ClpX.

To gain an understanding of the molecular basis by which TF does this, we mapped the binding interfaces between TF and ClpX using SPR, NMR, peptide array analysis, and mutational analysis. These data show that the binding interface between TF and ClpX involves all three domains of TF (P, A, and N) and both domains of ClpX (ZBD and AAA+). The TFP and the TFN domains interact mainly with the AAA+ domain of ClpX and more specifically to regions located between the large and small subunits of the hexameric ring. The TFA domain interacts with the ZBD of ClpX in a region that partially overlaps with the λO binding site on ZBD. Notably, several hydrophobic amino acid residues of TF, which were identified to be involved in the interaction with ClpX, are also responsible for recognizing λO and other unfolded protein substrates as shown in the solution structure of TF in complex with PhoA31. Our data show that some of the single point mutations of TF have no effect on substrate recognition, but induce significant reduction in the interaction with ClpX, which suggests a specific interaction between TF and ClpX. Overlap between ClpX-binding sites and the substrate-binding sites on TF implies competition between the substrate protein and ClpX for binding to TF. A similar mechanism was observed recently between the RssB adaptor, ClpX and its cognate substrate RpoS38. Furthermore, some of the ClpX-binding sites on TF overlap with the interface of the TF dimer8, suggesting that TF binds to ClpX as a monomer.

TF also enhanced the ClpXP-dependent degradation of SsrA-tagged proteins both in vitro and in vivo. This observation leads us to propose that ribosome-bound TF is also important for mediating the degradation of SsrA-tagged proteins soon after their release from the ribosome. This might be a distinguishing feature of TF compared to SspB, as SspB might act to promote the degradation of proteins at later stages after their release from ribosomes. In this regard, it was very striking to find that the impaired growth of cells lacking both TF and SspB can be restored upon overexpression of TF, implying an overlap in the function of these two proteins.

Interestingly, the N-terminus of RpoS is similar to that of λO. However, the effect of TF on RpoS degradation required the presence of the RssB adaptor. We speculate that RssB exposes the N-terminus of RpoS to allow for its degradation by ClpXP. This leads us to propose that the ClpXP-dependent degradation of proteins having an unstructured N-terminal sequence similar to that of λO can be modulated by TF.

Indeed, as shown in Supplementary Fig. 17, bacterial species lacking SspB likely rely on TF to promote the ClpXP-dependent degradation of their SsrA-tagged proteins. Hence, while TF might act as a chaperone to fold proteins in some instances, it can also act as a chaperone to promote protein degradation rendering TF a central hub in modulating the flux of proteins between folding and degradation.

Based on our work, we propose the presence of both a (1) co-translational and a (2) post-translational model of protein degradation in prokaryotes (Fig. 7c). In the first scenario, TF acts as a bridge between the ribosome and ClpX. It is possible that TFN binds the ribosome, TFP binds the AAA+ domain of ClpX and TFA binds ZBD. A nascent polypeptide chain exiting the ribosomal exit tunnel would first interact with TFA and then be transferred to the ZBD or AAA+ domain of ClpX via their respective recognition sequences (Fig. 7c, left panel). The transfer of the nascent chain from TF to ClpX would also lead to the enhancement of ClpX ATPase activity. This model assumes a static interaction of TF with the ribosome. Alternatively, since TF interaction with the ribosome is dynamic53, the second scenario would involve the formation of a complex of TF-ClpXP that is capable of binding ZBD-bound substrates. Here, TFN and TFp anchor to the AAA+ domain of ClpX and the TFA partially binds to the ZBD. It is important to note that for each ClpX ring, there are six AAA+ domains and three ZBD2 that have potential binding sites for TF and for substrates. TF-bound to a ZBD2 would then allow another ZBD2 bound to a substrate to move towards the ClpXP pore (Fig. 7c; right panel).

The association between a highly promiscuous and abundant chaperone-like TF with a main protease like ClpXP implies a system in which many diverse proteins are sampled promptly for folding or degradation in the cell. The role of TF as an adaptor protein of ClpXP is not only a new function of the chaperone but also describes a novel mechanism of proteostasis in bacteria.

## Methods

### Phylogenetic analysis

The nucleotide sequences of tig, smpB, sspB, rssB, clpP, clpX, and lon were retrieved from the NCBI nonredundant protein database using the blastx function of BLAST v. 2.6.0+54. The resulting sequences were aligned, and the generated multiple sequence alignments were employed to create the phylogenetic tree using the CLC Genomics Workbench v. 11.0.1 (Aarhus, Denmark). The phylogenies were calculated using the Neighbor-Joining method with BLOSSUM62 distance matrix55, and the confidence of the topologies were estimated using a bootstrap approach with 500 iterations.

### Bacterial strains

Gene deletions were performed with the P1 phage transduction method using the E. coli Keio library56. Clean knockouts were obtained by using the temperature-sensitive pCP20 plasmid encoding the FLP recombinase to remove the resistance cassette. The generated strains had no resistance markers and were verified by western blotting using α-TF (1:5000 dilution), α-ClpX (1:5000), α-ClpP (1:3000), and α-SsrA (1:3000) antibodies (Supplementary Fig. 18). ClpA and SspB KO were confirmed by PCR. Refer to Supplementary Table 1 for strains used in this study.

### Plasmids

The gene for the catalytically inactive ClpP mutant S111A was fused to a SPA affinity purification tag at its 3′ end and cloned into a modified pPAL7 (Bio-Rad) vector to generate a ClpPtrap-SPA construct. As a control, the same construct was made using WT ClpP. S7 was cloned into p11 vector57 with an N-terminal His-tag followed by a TEV protease cleavage site for purification of S7 protein. To monitor S7 degradation in vivo, untagged S7 was cloned into pET3a vector under the control of the T7 promoter and transformed into WT, tig::cat, and clpX::cat MC4100 strains. To express S7 in MC4100 strains, the pT7pol26 plasmid expressing T7 RNA polymerase58 was also transformed into these strains. The fusion construct NλO-S7 was made by fusing the NλO: MTNTAKILNFGR encoding sequence to the N-terminus of S7. the pBAT-His-Sumo-Barnase-SecM plasmid was a gift from Dr. Bernd Bukau (University of Heidelberg, Germany). Barnase was removed and λO was inserted between the AflII and BamH1 sites of pBAT plasmid to generate pBAT-His-Sumo-λO-SecM. To generate a construct without His-Sumo, the sequence of λO-SecM was PCR amplified from pBAT-His-Sumo-λO-SecM and inserted into pET15b plasmid between the NcoI and HindIII cut sites. For endogenous λO protein turnover experiments, WT, tig::cat, and clpX::cat MC4100 strains were lysogenized with λcI857Sam7 phage, which was provided by Dr. Alan Davidson (University of Toronto, Canada). Streptococcus mutans UA159 was used to obtain TFS. mutans, ClpPS. mutans and ClpXS. mutans, that were inserted into pProEXHTb. The SsrAS. mutans sequence -AKNTNSYAVAA was used to make GFP-ssrAS. mutans.

Plasmid pPK823 (pET11a) overexpressing Fnr was from Dr. Patricia K. Kiley (University of Wisconsin, USA). Plasmid pUHE21‐2 expressing the FlhDC complex was from Dr. Tomoko Yamamoto (Chiba University, Japan)59. Plasmid pHF010 expressing λN-His protein was from Dr. Irene Lee (Case Western Reserve University, USA). RssB was cloned into a pET-SUMO vector with a N-terminal SUMO-tag followed by a His-tag and a SUMO protease cleavage site for purification of RssB protein. Mutagenesis was carried out using the QuickChange kit (Stratagene).

Primers used for cloning several of the plasmids and mutants are given in Supplementary Table 2.

### Protein purification

All recombinant proteins were overexpressed in E. coli BL21(DE3) cells. The plasmid pProEXHTa TF expressing trigger factor was from Dr. José M. Barral (University of Texas, USA). TF was induced at mid-log phase with 1.0 mM IPTG for 3.5 h at 37 °C. His-tagged TF was purified on Ni-NTA column (Qiagen) and washed with 50 mM imidazole in buffer A (25 mM Tris-HCl, pH 7.5, 300 mM NaCl, 10% glycerol, and 1 mM DTT) to remove ATPase contaminants before elution. Eluted TF fractions were tested and only fractions without ATPase activity were collected. The His-tag was removed using His-tagged TEV protease during dialysis in buffer B (25 mM Tris-HCl, pH 7.5, 200 mM KCl, 10% glycerol, and 1 mM DTT). Purification of TF was performed in one day and the protein was flash-frozen in liquid nitrogen after removing TEV. Purification of the ClpX protein using the plasmid pProEXHTb ClpX was performed the same way as TF. Purification of RssB expressed from the pET-SUMO plasmid was carried out as for TF but using SUMO protease to remove the SUMO-His tag.

ClpP60, ClpA61, λO62, λN63, MuA64, S765, Fnr66, RpoS23, and SspB25 were expressed and purified according to published protocols. HtpG was precipitated in 70% ammonium sulfate and further purified using an anion exchange Q Sepharose column followed by size exclusion chromatography in 50 mM NaH2PO4, 300 mM NaCl, 10% glycerol. FlhDC complex proteins were purified as described59.

### Peptide array synthesis and analysis

Miniaturized peptide libraries of E. coli TF and E. coli ClpX were synthesized by Fmoc chemistry on a cellulose membrane using an ASP222 Autospot robot (Intavis AG) essentially as described67. The membrane was blocked with 1× Superblock (Pierce) in TBS buffer (20 mM Tris-HCl pH 7.5, 137 mM NaCl) containing 0.05% Tween-20 and probed for 1 h at 22 °C with 0.5 µM of purified His-ClpXΔZBD and 1 mM ATPγS or 0.5 µM of His-TF in TBS buffer containing 0.05% Tween-20, 10% Superblock, and 5% sucrose. The membrane was washed three times for 10 min in TBS buffer containing 0.05% Tween-20 and 50 µM ATPγS (His-ClpXΔZBD). His-protein binding was detected by probing the membrane directly with an anti-6×His monoclonal antibody conjugated to horseradish peroxidase used at a dilution of 1:5000 (mAB-HRP; BD Bioscience) and visualized by chemiluminescence. Non-specifically bound peptides were eliminated by subtracting spots that were also detected in a control experiment using non-tagged protein and mAB-HRP. Protein binding motifs were mapped onto the available crystal structure for TF (PDB: 1W26)32 and the cryoEM structure of ClpX∆ZBD (PDB: 6PP5)33 for visualization.

### NMR experiment

For NMR titration experiments, [U-15N; Ala-13CH3; Met-13CH3; Ile-δ1-13CH3; Leu/Val-13CH3/13CH3]-labeled TFPA (TF 113-432) and [U-15N]-labeled TFN (TF 1-117) were expressed in E. coli BL21(DE3) cells and purified as described previously8,31. TFPA (113-432) was cloned into the pCold vector (Takara Bio) and fused to His6-tag. TFN (1–117) was cloned into pET16b vector (Novagen) and fused to His6-MBP and a TEV protease cleavage site. Cells were grown in a medium containing 15NH4Cl (1 gL−1) and ampicillin (100 mg L−1). For the preparation of 1H–13C methyl-labeled samples, α-ketobutyric acid (50 mg L−1) and α-ketoisovaleric acid (85 mg L−1), [13CH3] methionine (50 mg L−1), [2-2H, 13CH3] alanine (50 mg L−1) were added to the culture 1 h before the addition of IPTG. Protein expression was induced by the addition of 0.5 mM IPTG at OD600 ~0.6, followed by ~16 h of incubation at 18 °C. Cells were harvested and resuspended in the lysis buffer containing 50 mM Tris-HCl pH 8.0, 500 mM NaCl. Cells were disrupted by a sonicator and centrifuged at 50,000×g for 45 min. Proteins were purified using Ni Sepharose 6 Fast Flow resin (GE Healthcare). In the case of TFN that contains TEV cleavage site, the His6-MBP tag was removed by TEV protease at 4 °C (incubation for 16 h). Proteins were further purified by gel filtration using Superdex 75 16/60 or 200 16/60 columns (GE Healthcare) in 25 mM 50 mM Tris-HCl (pH 7.5), 400 mM NaCl, 10% glycerol, 5 mM β-mercaptoethanol.

ClpX ZBD and AAA+ were cloned into pProEXHTa and expressed in E. coli BL21(DE3) cells. Protein expression was induced by the addition of 1 mM IPTG at OD600 ~0.6, followed by ~16 h of incubation at 18 °C. Proteins were purified by the same procedure as TFN as described above, but 100 μM ZnCl2 was supplemented to ZBD prior to size exclusion chromatography.

All of the proteins were prepared in the NMR buffer containing 50 mM Tris-HCl (pH 7.5), 200 mM KCl, 25 mM MgCl2, 2% glycerol, 5 mM β-mercaptoethanol. TF domains were concentrated to ~0.2 mM. The AAA+ domain was added to the TF domains in the presence of ADP. NMR experiments were performed on a Bruker 600 MHz NMR at 22 °C. Perturbation of the backbone and side chain methyl resonances were monitored using 2D 1H-15N transverse relaxation-optimized spectroscopy (TROSY) spectra and 1H-13C heteronuclear multiple quantum coherence (HMQC), respectively. Spectra were processed using the NMRPipe software68. Data analysis was performed by Olivia software (http://fermi.pharm.hokudai.ac.jp/olivia/). The resonance assignments were obtained from the BMRB Entries 19835 and 5665 for TF and ClpX ZBD, respectively. The perturbations of the resonances were evaluated by chemical shift change or intensity change. CSPs for backbone amide groups were calculated using the equation

$${\Delta}\delta = \sqrt {{\Delta}\delta _{\mathrm{H}}^2 + \left( {0.2{\Delta}\delta _{\mathrm{N}}} \right)^2},$$
(1)

where ∆δH and ∆δN are the chemical shift changes of 1H and 15N, respectively. CSPs for methyl groups were calculated by the equation

$${\Delta}\delta = \sqrt {\left( {\frac{{{\Delta}\delta _{\mathrm{H}}}}{\alpha }} \right)^2 + \left( {\frac{{{\Delta}\delta _{\mathrm{C}}}}{\beta }} \right)^2},$$
(2)

where ∆δH and ∆δC are the chemical shift changes of 1H and 13C, respectively, upon the addition of the ligand, and α and β are the chemical shift distribution of 1H and 13C, respectively, of methyl groups as reported in the Biological Resonance Data Bank (http://www.bmrb.wisc.edu). The perturbed resonances were mapped onto the structure of TF or ClpX ZBD. The cutoff of the CSP mapping was determined by ∆δave + ∆δSD, where ∆δave and ∆δSD are average and standard deviation, respectively, of the chemical shift change by the addition of the ligand. The cutoff of the intensity perturbation mapping was determined by rint,ave − rint,SD, where rint,ave and rint,SD are average and standard deviation, respectively, of the intensity ratio.

### GAPDH anti-aggregation assay

Aggregation of denatured GAPDH from rabbit muscle (Sigma; G-2267) was measured as described previously8,31. Totally, 125 µM GAPDH was denatured by 3 M guanidine-HCl in 20 mM potassium phosphate (pH 7.0) 100 mM KCl, 4 mM 2-mercaptoethanol, 0.5 mM EDTA, 0.05% NaN3 for overnight at 4 °C. Upon 50-fold dilution of the denatured enzyme into the buffer without guanidine-HCl, aggregation was monitored by 90° light scattering at 620 nm in a spectrofluorometer (FP-8500, JASCO Corporation) in the absence or presence of 1 µM of TF or TF mutant. The measurement was carried out at 20 °C.

Degradation assays were carried out in the absence or presence of TF. Degradation reaction mixture contained 1.2 μM ClpP, 3.9 μM substrate, 0 or 5.0 μM TF, and an ATP regeneration system (13 units/mL of creatine kinase and 16 mM creatine phosphate) in PD buffer (25 mM HEPES, pH 7.5, 5 mM MgCl2, 5 mM KCl, 0.03% (w/v) Tween 20, and 10% glycerol). All concentrations are those of protomers. Components were incubated at 37 °C for 3 minutes before adding 1.0 μM ClpX to start the reaction. At given time points, aliquots were withdrawn and mixed with 4× Laemmli buffer to stop the reaction. Proteins were then resolved on SDS-PAGE gels and visualized by Coomassie blue staining. Protein bands on the gels were quantified using GelEval (FrogDance). Anti-MuA (dilution 1:3000) antiserum was kindly provided by Dr. George Chaconas (University of Calgary, Canada). GFP-SsrA degradation was monitored by fluorescence using a Fluorolog spectrofluorometer (JobinYvon) with excitation wavelength set at 395 nm and emission wavelength set at 509 nm. RpoS degradation assays were performed using 3.9 µM of RpoS, 0.005 µM RssB, and 20 µM lithium potassium acetyl phosphate as phosphate donor50.

### ATPase activity measurement

The effect of TF on ClpX ATPase activity was measured using a coupled assay69 in a 96-well microtiter plate read using the Molecular Devices SpectraMax 340 PC 384 Microplate Reader. The reaction volume was 150 µL and assays were performed by preincubating 1.0 µM ClpX, various concentrations of TF, 0.2 mM NADH, 3.0 mM phosphoenolpyruvate, 4.7 units of pyruvate kinase, and 7.4 units of lactate dehydrogenase in PD buffer at 37 °C for 3 min. ClpP was added at 1.2 µM, and λO at 3.9 µM if needed. After preincubation for 3 min, 5 mM ATP was added to start the reaction and the change in absorbance at 340 nm was measured for 20 min at 37 °C. ATPase assays using ClpA were performed under the same concentration of the different components but in a high salt buffer (25 mM HEPES, pH 7.5, 20 mM MgCl2, 300 mM KCl, 0.03% (w/v) Tween 20, and 10% glycerol). ATP hydrolysis was calculated by measuring the disappearance of NADH at 340 nm, assuming 1:1 correspondence between ADP formation and NADH oxidation. The extinction coefficient used for NADH is 6220 M−1 cm−1 (ref. 69). Each reaction was performed at least three times.

### SPR assay

SPR measurements were carried out on a Biacore X from GE Healthcare; all experiments were done at 22 °C. TF was immobilized on a CM5 chip (GE Healthcare) using the Biacore Amine Coupling kit. The analyte was injected at a flow rate of 20 μL/min in running buffer (10 mM HEPES, pH 7.5, 150 mM NaCl, 3 mM EDTA, and 0.005% (w/v) P20 surfactant), and adequate injection time (90 s) was allowed for the interaction to reach steady-state if possible. To regenerate the surface between binding experiments, 2 M NaCl was injected for 1 min and washed with running buffer. From the results of the sensorgrams, response units at steady state were plotted against the corresponding analyte concentrations to generate the binding curve, and the Kd was calculated using WinCurveFit program assuming one-site Langmuir binding model for most binding experiments.

### Calculation of protomer distributions in mixed ClpX and ClpX∆ZBD complexes

In order to calculate the subunit composition upon mixing ClpX and ClpX∆ZBD in defined ratios (Fig. 3e), the following combinatorial statistics were used. We use p to represent the fractional concentration of ClpX (containing a trimer-of-dimers of ZBD)

$$p = \frac{{{\mathrm{[ClpX]}}}}{{\left[ {{\mathrm{ClpX}}} \right]{\mathrm{ + [ClpX}}{\Delta}{\mathrm{ZBD]}}}}.$$
(3)

The concentrations are those of protomers. The mixed ClpX:ClpX∆ZBD hexameric complexes formed have full-length ClpX composition (i.e., ZBD2 composition) that can be calculated, assuming random mixing, according to the following formula.

$$C_k^3p^k\left( {1 - p} \right)^{3 - k},$$
(4)

where k is the number of full-length ClpX subunits in the complex and

$$C_k^3 = \frac{{3^!}}{{k!\left( {3 - k} \right)!}}.$$
(5)

### Pulldowns using ClpPtrap-SPA

The ClpPtrap-SPA and the T7 polymerase containing plasmid pT7pol26 were cotransformed into different MC4100 mutant strains. For control experiments, a vector expressing WT ClpP was transformed into ∆clpPclpAsmpB strain. 11-mL cultures were incubated overnight at 37 °C. The next day, these cultures were used to inoculate 4 L cultures, which were then grown to mid-log phase (OD600 = 0.4). Cultures were induced for 4 h using 1 mM IPTG at 30 °C. Cell pellets were then resuspended in lysis buffer (25 mM Tris-HCl pH 7.5, 300 mM NaCl, 10% glycerol, and 1 mM DTT containing protease inhibitor cocktail (P8849; Sigma-Aldrich) and lysed using French press followed by centrifugation at 13,000×g for 30 min. The supernatants were collected for pull-down assays.

Totally, 200 μL of pre-washed FLAG antibody resin used at a dilution of 1:5000 (A2220; Sigma-Aldrich) was added to 20 mL supernatant and incubated with gentle rocking overnight at 4 °C. Beads were then transferred into flow columns and washed extensively with buffer (25 mM TrisHCl pH 7.5, 300 mM NaCl, 10% Triton X-100, 0.5 mM DTT, protease inhibitor cocktail 25×). Proteins were eluted from the FLAG beads using 1 mg/ml FLAG peptide (Sigma-Aldrich). Eluted proteins were subsequently digested with trypsin and identified using an Orbitrap mass spectrometer (ThermoFisher Scientific). The resulting MS/MS spectra were then queried against the protein-coding sequences of E. coli using the SEQUEST search engine. Matches with high confidence (>95% probability) were then evaluated using spectral counts and probability scores computed by the STATQUEST algorithm. Only proteins with three or more unique peptides were considered. To quantify the data, the normalized spectral abundance factor method was used70, which employs spectral counts normalized to protein length and to the total spectral count in each sample.

### Monitoring λO and GFP-ssrA degradation in cells

Plasmid pRLM266 expressing untagged λO under the T7 promoter was transformed into WT, tig::cat, and clpX::cat MC4100 strains. The strains were cotransformed with the pT7pol26 plasmid expressing the T7 polymerase58. Cells were grown in LB broth at 37 °C until mid-log phase (OD600 = 0.6); IPTG was then added at a final concentration of 0.4 mM to induce λO expression for 2 h. Tetracycline was added at a final concentration of 30 μg/mL to inhibit translation (time 0), and 100 µL of cells were withdrawn after the addition of tetracycline at 0, 10, 13, 30, 45, 60, and 90 min. Cell density was measured using OD600 for each time point. Cells were pelleted and resuspended in 1× SDS loading buffer with equal cell density. Whole-cell lysates were subjected to western blot analysis using rabbit antisera specific for λO, ClpX, ClpP, and TF proteins. About 105 cells were loaded for each time point to monitor the degradation of λO proteins. The same procedure was performed to monitor GFP-ssrA levels.

For monitoring λO-SecM degradation, pET15b λO-SecM was transformed into MC4100 strains also containing pT7pol26 plasmid, and the same procedures were followed as above except that 10 times more cells (106) were loaded for western blot analysis to monitor λO-SecM due to lower expression level of the fusion protein.

For ribosome fractionation, experiments were carried out as described above. For each time point, 10 mL of cells were taken, pelleted, and resuspended in buffer R [10 mM Tris-HCl, pH 7.5, 10 mM MgCl2, 60 mM NH4Cl, 400 mM potassium acetate, 1:1000 dilution of RNAse inhibitor (Promega), 1 U/mL RNAse free DNase I, and EDTA-free protease inhibitor cocktail from Sigma (1 tablet in 50 mL buffer)] to the same cell density. Cells were lysed by sonication and debris was spun down at 20,000g for 30 min. Equal amounts of supernatant were loaded on a 3-fold or greater volume of sucrose cushion (0.7 M sucrose, 10 mM TrisHCl, pH 7.5, 10 mM MgCl2, 60 mM NH4Cl, 400 mM KOAc, and 1:1000 dilution of RNAse inhibitor) and centrifuged at 148,379×g (average) for 5 h using Beckman MLS-50 ultracentrifuge. Pelleted ribosomes were resuspended in buffer S (20 mM TrisHCl, pH 7.5, 10 mM KCl, 10 mM MgCl2, and 10% sucrose). Samples containing equal amounts of ribosomal material were separated on SDS-PAGE gels and analyzed by western blot.

### Monitoring λO degradation in lysogens

WT and tig::cat MC4100 strains were lysogenized with the heat-inducible prophage λcI857Sam7. The lysogens were grown in M63 minimal media supplemented with 0.2% glucose at 30 °C until mid-log phase (OD600 = 0.6) and then heat shocked at 42 °C for 8 min to induce phage replication. After heat shock, cells were pulse labeled with 200 µCi of [35S] methionine (PerkinElmer, 1000 Ci/mmol) for 2 min, and chased with 200 µg/mL nonradioactive methionine at 37 °C for 30 min. At different time points, samples were withdrawn during the chase period, and λO protein was immunoprecipitated using α-λO polyclonal rabbit antiserum (dilution of 1:3000) and 300 µg Protein A-Acrylic beads from Staphylococcus aureus (Sigma).

### Monitoring degradation of newly synthesized proteins

WT, tig::cat, and tig::cat MC4100 cells complemented with a plasmid expressing tig under its own promoter were grown at 37 °C in M9 minimal medium supplemented with 0.4% glucose (MG medium) until mid-log phase, OD600 = 0.6. Cells were then pulse-labeled with 60 µCi/mL [35S] methionine (PerkinElmer, 1000 Ci/mmol) for 2 min and chased with 0.5 mg/mL non-radioactive methionine for 3 min. To remove excess [35S] methionine, cells were washed three times in MG medium containing non-radioactive methionine and resuspended in MG medium with non-radioactive methionine and 30 µg/mL tetracycline to inhibit translation. Aliquots of cells were taken at different time points, and samples were precipitated with TCA. To determine the relative amount of degraded proteins, radioactivity in the TCA-soluble fraction, containing small peptides, and TCA-insoluble fraction, containing larger polypeptides, was measured by liquid scintillation counting (Beckman Coulter LS6500 Multipurpose Scintillation Counter). TCA-insoluble fractions were resolubilized in 8 M urea before measuring radioactivity. The percentage of degraded proteins was calculated by taking the ratio of the counts from TCA-soluble fraction over total scintillation counts (TCA-soluble plus TCA-insoluble). For heat shock and AZC treatment conditions, the procedures were the same except that cells were heated at 42 °C for 10 min or treated with 100 µg/mL AZC before pulse labeling with [35S] methionine, respectively.

### Reporting summary

Further information on experimental design is available in the Nature Research Reporting Summary linked to this paper.

## Data availability

The data that support this study are present in the paper and Supplementary information, and are available from the corresponding author upon request. Source data are provided with this paper.

## References

1. 1.

Balchin, D., Hayer-Hartl, M. & Hartl, F. U. In vivo aspects of protein folding and quality control. Science 353, aac4354 (2016).

2. 2.

Deuerling, E., Schulze-Specking, A., Tomoyasu, T., Mogk, A. & Bukau, B. Trigger factor and DnaK cooperate in folding of newly synthesized proteins. Nature 400, 693–696 (1999).

3. 3.

Teter, S. A. et al. Polypeptide flux through bacterial Hsp70: DnaK cooperates with trigger factor in chaperoning nascent chains. Cell 97, 755–765 (1999).

4. 4.

Martinez-Hackert, E. & Hendrickson, W. A. Promiscuous substrate recognition in folding and assembly activities of the trigger factor chaperone. Cell 138, 923–934 (2009).

5. 5.

Crooke, E., Guthrie, B., Lecker, S., Lill, R. & Wickner, W. ProOmpA is stabilized for membrane translocation by either purified E. coli trigger factor or canine signal recognition particle. Cell 54, 1003–1011 (1988).

6. 6.

Patzelt, H. et al. Binding specificity of Escherichia coli trigger factor. Proc. Natl Acad. Sci. USA 98, 14244–14249 (2001).

7. 7.

Bhandari, V. & Houry, W. A. Substrate interaction networks of the Escherichia coli chaperones: trigger factor, DnaK and GroEL. Adv. Exp. Med. Biol. 883, 271–294 (2015).

8. 8.

Saio, T., Kawagoe, S., Ishimori, K., Kalodimos, C. G. Oligomerization of a molecular chaperone modulates its activity. Elife 7, e35731 (2018).

9. 9.

Yu, A. Y. & Houry, W. A. ClpP: a distinctive family of cylindrical energy-dependent serine proteases. FEBS Lett. 581, 3749–3757 (2007).

10. 10.

Neuwald, A. F., Aravind, L., Spouge, J. L. & Koonin, E. V. AAA+: a class of chaperone-like ATPases associated with the assembly, operation, and disassembly of protein complexes. Genome Res. 9, 27–43 (1999).

11. 11.

Snider, J., Thibault, G. & Houry, W. A. The AAA+ superfamily of functionally diverse proteins. Genome Biol. 9, 216 (2008).

12. 12.

Keiler, K. C., Waller, P. R. & Sauer, R. T. Role of a peptide tagging system in degradation of proteins synthesized from damaged messenger RNA. Science 271, 990–993 (1996).

13. 13.

Kirstein, J., Moliere, N., Dougan, D. A. & Turgay, K. Adapting the machine: adaptor proteins for Hsp100/Clp and AAA+ proteases. Nat. Rev. Microbiol. 7, 589–599 (2009).

14. 14.

Liu, K., Ologbenla, A. & Houry, W. A. Dynamics of the ClpP serine protease: a model for self-compartmentalized proteases. Crit. Rev. Biochem. Mol. Biol. 49, 400–412 (2014).

15. 15.

Ciccarelli, F. D. et al. Toward automatic reconstruction of a highly resolved tree of life. Science 311, 1283–1287 (2006).

16. 16.

Knipfer, N., Seth, A., Roudiak, S. G. & Shrader, T. E. Species variation in ATP-dependent protein degradation: protease profiles differ between mycobacteria and protease functions differ between Mycobacterium smegmatis and Escherichia coli. Gene 231, 95–104 (1999).

17. 17.

Zampieri, M., Soranzo, N., Bianchini, D. & Altafini, C. Origin of co-expression patterns in E. coli and S. cerevisiae emerging from reverse engineering algorithms. PLoS ONE 3, e2981 (2008).

18. 18.

Overbeek, R., Fonstein, M., D’Souza, M., Pusch, G. D. & Maltsev, N. The use of gene clusters to infer functional coupling. Proc. Natl Acad. Sci. USA 96, 2896–2901 (1999).

19. 19.

Gonciarz-Swiatek, M. et al. Recognition, targeting, and hydrolysis of the lambda O replication protein by the ClpP/ClpX protease. J. Biol. Chem. 274, 13999–14005 (1999).

20. 20.

Mhammedi-Alaoui, A., Pato, M., Gama, M. J. & Toussaint, A. A new component of bacteriophage Mu replicative transposition machinery: the Escherichia coli ClpX protein. Mol. Microbiol. 11, 1109–1116 (1994).

21. 21.

Karzai, A. W., Roche, E. D. & Sauer, R. T. The SsrA-SmpB system for protein tagging, directed degradation and ribosome rescue. Nat. Struct. Biol. 7, 449–455 (2000).

22. 22.

Deuerling, E. et al. Trigger factor and DnaK possess overlapping substrate pools and binding specificities. Mol. Microbiol. 47, 1317–1328 (2003).

23. 23.

Tanaka, K., Takayanagi, Y., Fujita, N., Ishihama, A. & Takahashi, H. Heterogeneity of the principal sigma factor in Escherichia coli: the rpoS gene product, sigma 38, is a second principal sigma factor of RNA polymerase in stationary-phase Escherichia coli. Proc. Natl Acad. Sci. USA 90, 3511–3515 (1993).

24. 24.

Flynn, J. M., Neher, S. B., Kim, Y. I., Sauer, R. T. & Baker, T. A. Proteomic discovery of cellular substrates of the ClpXP protease reveals five classes of ClpX-recognition signals. Mol. Cell 11, 671–683 (2003).

25. 25.

Wojtyra, U. A., Thibault, G., Tuite, A. & Houry, W. A. The N-terminal zinc binding domain of ClpX is a dimerization domain that modulates the chaperone function. J. Biol. Chem. 278, 48981–48990 (2003).

26. 26.

Hersch, G. L., Burton, R. E., Bolon, D. N., Baker, T. A. & Sauer, R. T. Asymmetric interactions of ATP with the AAA+ ClpX6 unfoldase: allosteric control of a protein machine. Cell 121, 1017–1027 (2005).

27. 27.

Wah, D. A., Levchenko, I., Baker, T. A. & Sauer, R. T. Characterization of a specificity factor for an AAA+ ATPase: assembly of SspB dimers with ssrA-tagged proteins and the ClpX hexamer. Chem. Biol. 9, 1237–1245 (2002).

28. 28.

Gottesman, S., Clark, W. P., de Crecy-Lagard, V. & Maurizi, M. R. ClpX, an alternative subunit for the ATP-dependent Clp protease of Escherichia coli. Sequence and in vivo activities. J. Biol. Chem. 268, 22618–22626 (1993).

29. 29.

Thibault, G., Tsitrin, Y., Davidson, T., Gribun, A. & Houry, W. A. Large nucleotide-dependent movement of the N-terminal domain of the ClpX chaperone. EMBO J. 25, 3367–3376 (2006).

30. 30.

Kim, Y. I. et al. Molecular determinants of complex formation between Clp/Hsp100 ATPases and the ClpP peptidase. Nat. Struct. Biol. 8, 230–233 (2001).

31. 31.

Saio, T., Guan, X., Rossi, P., Economou, A. & Kalodimos, C. G. Structural basis for protein antiaggregation activity of the trigger factor chaperone. Science 344, 1250494 (2014).

32. 32.

Ferbitz, L. et al. Trigger factor in complex with the ribosome forms a molecular cradle for nascent proteins. Nature 431, 590–596 (2004).

33. 33.

Fei X., et al. Structures of the ATP-fueled ClpXP proteolytic machine bound to protein substrate. Elife. 9, e52774 (2020).

34. 34.

Kramer, G. et al. Trigger factor peptidyl-prolyl cis/trans isomerase activity is not essential for the folding of cytosolic proteins in Escherichia coli. J. Biol. Chem. 279, 14165–14170 (2004).

35. 35.

Thibault, G. & Houry, W. A. Role of the N-terminal domain of the chaperone ClpX in the recognition and degradation of lambda phage protein O. J. Phys. Chem. B 116, 6717–6724 (2012).

36. 36.

Dougan, D. A., Weber-Ban, E. & Bukau, B. Targeted delivery of an ssrA-tagged substrate by the adaptor protein SspB to its cognate AAA+ protein ClpX. Mol. Cell 12, 373–380 (2003).

37. 37.

Park, E. Y. et al. Structural basis of SspB-tail recognition by the zinc binding domain of ClpX. J. Mol. Biol. 367, 514–526 (2007).

38. 38.

Micevski D., et al. Insight into the RssB-mediated recognition and delivery of sigma(s) to the AAA+ protease, ClpXP. Biomolecules 10, 615 (2020).

39. 39.

Kandror, O. & Goldberg, A. L. Trigger factor is induced upon cold shock and enhances viability of Escherichia coli at low temperatures. Proc. Natl Acad. Sci. USA 94, 4978–4981 (1997).

40. 40.

Karzai, A. W., Susskind, M. M. & Sauer, R. T. SmpB, a unique RNA-binding protein essential for the peptide-tagging activity of SsrA (tmRNA). EMBO J. 18, 3793–3799 (1999).

41. 41.

Levchenko, I., Grant, R. A., Flynn, J. M., Sauer, R. T. & Baker, T. A. Versatile modes of peptide recognition by the AAA+ adaptor protein SspB. Nat. Struct. Mol. Biol. 12, 520–525 (2005).

42. 42.

Bolon, D. N., Wah, D. A., Hersch, G. L., Baker, T. A. & Sauer, R. T. Bivalent tethering of SspB to ClpXP is required for efficient substrate delivery: a protein-design study. Mol. Cell 13, 443–449 (2004).

43. 43.

McGinness, K. E., Bolon, D. N., Kaganovich, M., Baker, T. A. & Sauer, R. T. Altered tethering of the SspB adaptor to the ClpXP protease causes changes in substrate delivery. J. Biol. Chem. 282, 11465–11473 (2007).

44. 44.

Levchenko, I., Seidel, M., Sauer, R. T. & Baker, T. A. A specificity-enhancing factor for the ClpXP degradation machine. Science 289, 2354–2356 (2000).

45. 45.

Babu, M. et al. Sequential peptide affinity purification system for the systematic isolation and identification of protein complexes from Escherichia coli. Methods Mol. Biol. 564, 373–400 (2009).

46. 46.

Schweder, T., Lee, K. H., Lomovskaya, O. & Matin, A. Regulation of Escherichia coli starvation sigma factor (sigma s) by ClpXP protease. J. Bacteriol. 178, 470–476 (1996).

47. 47.

Zhou, Y. & Gottesman, S. Regulation of proteolysis of the stationary-phase sigma factor RpoS. J. Bacteriol. 180, 1154–1158 (1998).

48. 48.

Studemann, A. et al. Sequential recognition of two distinct sites in sigma(S) by the proteolytic targeting factor RssB and ClpX. EMBO J. 22, 4111–4120 (2003).

49. 49.

Becker, G., Klauck, E. & Hengge-Aronis, R. Regulation of RpoS proteolysis in Escherichia coli: the response regulator RssB is a recognition factor that interacts with the turnover element in RpoS. Proc. Natl Acad. Sci. USA 96, 6439–6444 (1999).

50. 50.

Zhou, Y., Gottesman, S., Hoskins, J. R., Maurizi, M. R. & Wickner, S. The RssB response regulator directly targets sigma(S) for degradation by ClpXP. Genes Dev. 15, 627–637 (2001).

51. 51.

Rutkowska, A. et al. Large-scale purification of ribosome-nascent chain complexes for biochemical and structural studies. FEBS Lett. 583, 2407–2413 (2009).

52. 52.

Lau, J., Hernandez-Alicea, L., Vass, R. H. & Chien, P. A phosphosignaling adaptor primes the AAA+ Protease ClpXP to drive cell cycle-regulated proteolysis. Mol. Cell 59, 104–116 (2015).

53. 53.

Kaiser, C. M. et al. Real-time observation of trigger factor function on translating ribosomes. Nature 444, 455–460 (2006).

54. 54.

Altschul, S. F., Gish, W., Miller, W., Myers, E. W. & Lipman, D. J. Basic local alignment search tool. J. Mol. Biol. 215, 403–410 (1990).

55. 55.

Henikoff, S. & Henikoff, J. G. Amino acid substitution matrices from protein blocks. Proc. Natl Acad. Sci. USA 89, 10915–10919 (1992).

56. 56.

Baba, T. et al. Construction of Escherichia coli K-12 in-frame, single-gene knockout mutants: the Keio collection. Mol. Syst. Biol. 2, 2006 0008 (2006).

57. 57.

Zhang, R. G. et al. Structure of Thermotoga maritima stationary phase survival protein SurE: a novel acid phosphatase. Structure 9, 1095–1106 (2001).

58. 58.

Mertens, N., Remaut, E. & Fiers, W. Tight transcriptional control mechanism ensures stable high-level expression from T7 promoter-based expression plasmids. Biotechnology 13, 175–179 (1995).

59. 59.

Takaya, A. et al. YdiV: a dual function protein that targets FlhDC for ClpXP-dependent degradation by promoting release of DNA-bound FlhDC complex. Mol. Microbiol. 83, 1268–1284 (2012).

60. 60.

Flanagan, J. M., Wall, J. S., Capel, M. S., Schneider, D. K. & Shanklin, J. Scanning transmission electron microscopy and small-angle scattering provide evidence that native Escherichia coli ClpP is a tetradecamer with an axial pore. Biochemistry 34, 10910–10917 (1995).

61. 61.

Leung, E. et al. Activators of cylindrical proteases as antimicrobials: identification and development of small molecule activators of ClpP protease. Chem. Biol. 18, 1167–1178 (2011).

62. 62.

Roberts, A. B. et al. Purification and properties of a type beta transforming growth factor from bovine kidney. Biochemistry 22, 5692–5698 (1983).

63. 63.

Patterson-Ward, J. et al. Utilization of synthetic peptides to evaluate the importance of substrate interaction at the proteolytic site of Escherichia coli Lon protease. Biochim. Biophys. Acta 1794, 1355–1363 (2009).

64. 64.

Baker, T. A., Mizuuchi, M., Savilahti, H. & Mizuuchi, K. Division of labor among monomers within the Mu transposase tetramer. Cell 74, 723–733 (1993).

65. 65.

Wimberly, B. T., White, S. W. & Ramakrishnan, V. The structure of ribosomal protein S7 at 1.9 A resolution reveals a beta-hairpin motif that binds double-stranded nucleic acids. Structure 5, 1187–1198 (1997).

66. 66.

Lazazzera, B. A., Bates, D. M. & Kiley, P. J. The activity of the Escherichia coli transcription factor FNR is regulated by a change in oligomeric state. Genes Dev. 7, 1993–2005 (1993).

67. 67.

Lee, J. et al. Heat shock protein (Hsp) 70 is an activator of the Hsp104 motor. Proc. Natl Acad. Sci. USA 110, 8513–8518 (2013).

68. 68.

Delaglio, F. et al. NMRPipe: a multidimensional spectral processing system based on UNIX pipes. J. Biomol. NMR 6, 277–293 (1995).

69. 69.

Norby, J. G. Coupled assay of Na+,K+-ATPase activity. Methods Enzymol. 156, 116–119 (1988).

70. 70.

Zybailov, B. et al. Statistical analysis of membrane proteome expression changes in Saccharomyces cerevisiae. J. Proteome Res. 5, 2339–2347 (2006).

71. 71.

Donaldson, L. W., Wojtyra, U. & Houry, W. A. Solution structure of the dimeric zinc binding domain of the chaperone ClpX. J. Biol. Chem. 278, 48991–48996 (2003).

72. 72.

Letunic, I. & Bork, P. Interactive tree of life (iTOL) v3: an online tool for the display and annotation of phylogenetic and other trees. Nucleic Acids Res. 44, W242–W245 (2016).

## Acknowledgements

We thank Ms. Hyesuk Kim for help with peptide array synthesis and Dr. Gursonika Binepal for help with generating some of the mutant strains. KR was supported by a Canadian Institutes of Health Research (CIHR) Training Program in Protein Folding and Interaction Dynamics: Principles and Diseases fellowship (TGF-53910), a University of Toronto Fellowship from the Department of Biochemistry, and an Ontario Graduate Scholarship. This work was supported by a CIHR Project grant (PJT-148564) to WAH, by the Natural Sciences and Engineering Research Council of Canada (DG-20234) to MB, by JSPS KAKENHI (17H05657, 17H05867, 18H05229, 19K06504), AMED Practical Research Project for Rare/Intractable Diseases (JP20ek0109437), and PRESTO JST to TS, JSPS KAKENHI (19H05769) to KI, and by the National Institutes of Health (GM111084) to FTFT.

## Author information

Authors

### Contributions

WAH, KR, and AYHY conceived the project and wrote the manuscript. KR and AYHY carried out most of the experiments in the manuscript. AO, SRK, GT, EL, YWZ, MT, MH, and NM helped with cloning, mutagenesis, protein degradation assays, and SPR experiments. HZ was involved in carrying out and analyzing the NMR experiments under the supervision of KI and TS. SP and ZM were involved in carrying out and analyzing the mass spectrometry experiments under the supervision of MB. SL and FTFT performed the peptide array analysis. WAH coordinated the project. All authors were involved in editing the manuscript.

### Corresponding author

Correspondence to Walid A. Houry.

## Ethics declarations

### Competing interests

The authors declare no competing interests.

Peer review information Nature Communications thanks Kürsad Turgay, and the other, anonymous, reviewer(s) for their contribution to the peer review of this work. Peer reviewer reports are available.

Publisher’s note Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.

## Rights and permissions

Reprints and Permissions

Rizzolo, K., Yu, A.Y.H., Ologbenla, A. et al. Functional cooperativity between the trigger factor chaperone and the ClpXP proteolytic complex. Nat Commun 12, 281 (2021). https://doi.org/10.1038/s41467-020-20553-x

• Accepted:

• Published:

• ### The cryo-EM structure of the chloroplast ClpP complex

• Ning Wang
• Yifan Wang
• Cuimin Liu

Nature Plants (2021)

By submitting a comment you agree to abide by our Terms and Community Guidelines. If you find something abusive or that does not comply with our terms or guidelines please flag it as inappropriate.