All bacterial strains and plasmids used in this study are listed in Supplementary Table 1. Cultures were grown and maintained in LB broth or M9 minimal medium (Difco) at 37°C with shaking at 225 rpm, with the exception of strains bearing the cpxA24 mutation, which were grown at 30°C in the presence of amikacin (3 μg/ml) to prevent accumulation of suppressors as previously described (Raivio et al., 1999). Isopropyl-β-D-thiogalactopyranoside (IPTG, Invitrogen) was added to a concentration of 0.1 mM to induce gene expression from pCA24N- and pMPM-K3-based vectors. Antibiotics (Sigma) were added as necessary at the following concentrations: chloramphenicol (Cam), 25 μg/mL; kanamycin (Kan), 30 or 50 μg/mL; ampicillin (Amp), 100 μg/mL; spectinomycin (Spc), 25 μg/ml.

All BW25113 mutants were taken from the Keio collection (Baba et al., 2006). An MC4100 ΔcpxR knockout mutant was generated using P1 transduction to move the desired mutant allele from the Keio collection (Baba et al., 2006) into wildtype MC4100 as previously described (Silhavy et al., 1984). The inducible pCA24N-based plasmids used in this study were obtained from the ASKA collection (Kitagawa et al., 2005). Transcriptional luminescent reporters containing the promoter regions of cpxP and nuoA were constructed as previously described (Price and Raivio, 2009; Wong et al., 2013; Guest et al., 2017). The psdhC::lux reporter was constructed similarly. Briefly, the promoter region of sdhC gene was amplified by PCR, using PsdhCFwdCln and PsdhCRevCln primers (Supplementary Table 2). Purified PCR products and the pJW15 vector (MacRitchie et al., 2008) were digested with EcoRI and BamHI (Invitrogen), and the insert was ligated upstream of the luxABCDE operon in the pJW15 plasmid. Correct insertion of the promoter sequence was verified by PCR and sequencing using pNLP10F and pNLP10R primers (Supplementary Table 2).

For the construction of pTrc-nlpE vector, nlpE was amplified via PCR with recombinant Taq polymerase (Invitrogen) using nlpE_NcoI_F and nlpE_WT_HindIII_R primers (Supplementary Table 2). PCR products were purified using a QIAGEN QIAQuick PCR purification kit according to the manufacturers protocol. Amplified nlpE was digested along with purified pTrc99A using Fast Digest EcoRI and HindIII (Thermo Scientific). Digests were purified using the aforementioned PCR purification kit and were then ligated together using T4 DNA ligase. Ligations were then transformed into One Shot TOP10 chemically competent cells (Thermo Fisher). Correct insertion of the nlpE sequence was verified by PCR and sequencing using pTrc99A_F and pTrc99A_R primers (Supplementary Table 2).

The pMPM-NuoA-3 × FLAG plasmid was constructed by amplifying nuoA from the E2348/69 chromosome via PCR using primers nuoAFLAGFwd and nuoAFLAGrev (Supplementary Table 2). Primer nuoAFLAGrev contains the nucleotide sequence to insert a triple FLAG-tag directly upstream of the nuoA stop codon. PCR was performed using high-fidelity phusion DNA polymerase (ThermoFisher) according to the manufacturers protocol with the addition of 10% betaine. The DNA band corresponding to nuoA-3 × FLAG DNA was gel extracted and cleaned using the GeneJet gel purification kit (Fermentas). Both nuoA-3 × FLAG and pMPM-K3 DNA were digested with the HindIII and XbaI restriction endonucleases (Invitrogen) according to the manufacturers protocol. nuoA-3 × FLAG DNA was then ligated downstream of an IPTG inducible Plac promoter in the pMPM-K3 vector and transformed into One Shot TOP10 chemically competent E. coli (Invitrogen) as per the manufacturer’s protocol. PCR and DNA sequencing were used to confirm the presence of nuoA-3 × FLAG fragment within pMPM-K3 using M13F and M13R primers (Supplementary Table 2). All DNA sequencing was performed by The University of Alberta Molecular Biology Services Unit (MBSU). All plasmids in this study were transformed into E. coli strains via chemical competency (Silhavy et al., 1984).

Strains containing pcpxP::lux, psdhC::lux, or pnuoA::lux reporter plasmids were grown overnight in LB broth with shaking at 37°C. Cells were pelleted by centrifugation and washed twice with phosphate-buffered saline (PBS). The cell density was standardized to OD₆₀₀ 1.0, pelleted, and resuspended in 1 mL of 1 × PBS. Standardized cultures were serially diluted 10-fold and 10 μL of each dilution was spotted onto M9 minimal agar containing 0.4% glucose, malic acid, or succinic acid (Sigma). Glucose-succinate gradient agar plates were created by pouring M9 containing 0.4% glucose into an angled plate. The layer was allowed to solidify in the inclined position and the second portion of agar containing 0.4% succinic acid was poured into the plate, placed on a level surface and allowed to solidify (Weinberg, 1959). Agar pH was adjusted to 7.0 with sodium hydroxide. Bacteria were grown for 24–48 h at 37°C statically. Luminescence was determined by imaging the light produced by strains using the UVP Colony Doc-It Imaging Station (Biorad). Luminescence was quantified using Fiji (ImageJ).

For the assay performed in liquid medium, bacteria were grown overnight as described above, subcultured at a dilution factor of 1:100 into 5 ml of fresh LB broth and incubated for 2 h at 37°C with aeration. 200 μL of each subculture was aliquoted into a black-bottomed 96-well plate, and luminescence in counts per second (CPS) and OD₆₀₀ were measured every 1 h for 7 h post-subculture using a PerkinElmer VICTOR™ X3 multilabel reader. Luminescence and OD₆₀₀ values measured from a blank well containing uncultured LB were subtracted from each sample. CPS was standardized to the OD₆₀₀ to correct for differences in cell numbers between samples. All experiments contained three technical replicates and were carried out three times.

Samples used for Western blot analysis were prepared by diluting overnight cultures 1:100 into 25 ml fresh LB containing appropriate concentrations of antibiotics. Bacteria were grown at 37°C with shaking to an OD₆₀₀ of ∼ 0.35. IPTG was added to a concentration of 0.1 mM and bacteria were grown for an additional 30 mins as before. 2 × 1 mL samples were collected. Cells were pelleted by centrifugation at 21,130 × g for 1 min. One sample was resuspended in 50 μL 2 × Laemmli sample buffer (Sigma) and the other sample was resuspended in 50 μL 1 × PBS. Protein concentration was determined from the sample resuspended in phosphate-buffered saline using the Pierce BCA protein assay kit (ThermoFisher) according to the manufacturers’ protocol. Samples resuspended in 50 μL 2 × Laemmli sample buffer were denatured by boiling for 5 mins. Sample volumes were standardized according to their determined protein concentrations and separated on a 12% SDS gel at 110 V for 1.5 h in Tris–glycine running buffer (10% SDS, 250 mM Tris, 1.2 M glycine).

Proteins were transferred onto a nitrocellulose membrane via the trans-blot semi-dry transfer system (Bio-Rad) at 15 V for 22 mins using semi-dry Towbin transfer buffer (78 mM glycine, 1.3 mM SDS, 20% methanol). Membranes were blocked in 5% MTS (2.5% skim milk powder, 154 mM NaCl, 1 mM Tris) or 2% BSA (154 mM NaCl, 1 mM Tris, 1% Tween 20, 2% bovine serum albumin) for 1 h at room temperature with shaking at 10 rpm. Primary α-FLAG (Sigma, BioLegend), α-PhoA (Abcam) and α-RNAPα (BioLegend) antibodies were diluted by a factor of 1:5,000 into 5% MTS or 2% BSA. Membranes were incubated with the primary antibody for 1 h at either room temperature with shaking at approximately 10 rpm or overnight at 4°C with rocking. Following incubation with the primary antibody, membranes were washed for 5 mins in wash solution (154 mM NaCl, 1 mM Tris, 1% Tween 20) four times. Alkaline-phosphatase (AP) anti-rabbit secondary antibodies (Sigma) were diluted at a factor of 1:10,000 in 5% MTS and were used to detect the α-PhoA (Abcam) and α-FLAG (Sigma) primary antibodies. IRDye®680RD Goat anti-Mouse 925-68070 and IRDye®800CW Goat anti-Rabbit 925-32211 secondary antibodies (LI-COR) were diluted at a factor of 1:15,000 in 5% MTS and were used to detect the α-RNAPα (BioLegend) and α-FLAG (BioLegend) primary antibodies, respectively.

Membranes were then incubated with the secondary antibody for 1 h at room temperature with shaking at approximately 10 rpm. Membranes were washed following incubation with the secondary antibody as before. Proteins from the membranes incubated with AP secondary antibodies were detected using the Immun-Star alkaline phosphatase chemiluminescence kit (Bio-Rad). All membranes were imaged with the Bio-Rad ChemiDoc MP imaging system. Quantification of each band compared to the wildtype was performed using band intensity analysis in Fiji (ImageJ). Experiments were performed in biological triplicates, and a representative blot is shown in each case.

Bacteria were grown overnight in 5 mL LB at 37°C with shaking. The following day, bacteria were subcultured at a 1:100 dilution into 25 mL fresh LB and grown at 37°C with shaking to an OD₆₀₀ ∼ 0.5. IPTG was then added to a final concentration of 0.1 mM and bacteria were grown to an OD₆₀₀ of 1.0. 1 mL samples were collected and cells were pelleted by centrifugation at 21,130 × g for 1 min. After the supernatant was removed, cells were resuspended in 50 μL 2 × Laemmli sample buffer (Sigma). Immediately after the sample was removed, the protein synthesis inhibitor chloramphenicol was added to the remaining culture at a concentration of 100 μg/mL. The culture was incubated at 37°C and shaken at 225 rpm. 1 mL of culture was collected right before the addition of chloramphenicol and at 1, 5, 10, 20, 30, 45, 90, and 120 min(s) after the addition of chloramphenicol. Sample collection at each timepoint was performed as described above. 10 μL of each sample was loaded onto a 12% SDS polyacrylamide gel. Incubation with primary and secondary antibodies, detection and band quantification were performed as described above. Each experiment was repeated three times, and a representative blot is shown in each case.

Succinate dehydrogenase activity was measured using a kit (Abcam). Samples were prepared following the manufacturer’s protocol by diluting overnight cultures 1:50 into 5 mL fresh LB and growing them to OD₆₀₀ ∼0.5 at 37°C with shaking. 1 mL of culture was taken, pelleted by centrifugation, washed with 1 × PBS and pelleted again. Cultures were then standardized to the same optical density OD₆₀₀ of 1.0 in the final volume of 1 mL by adding an appropriate volume of 1 × PBS, pelleting, and resuspending in 200 mL of ice-cold SDH Assay Buffer (Abcam). Cell lysis was performed by sonication. Samples were centrifuged at 10,000 × g for 5 mins and the supernatant was transferred into a fresh tube. 50 μL of each sample was loaded into a 96-well plate, and 50 μL of SDH reaction mix containing the SDH substrate mix, and the probe (Abcam) was added to them. Absorbance at 600 nm was measured for 30 mins at 25°C in kinetic mode using the Cytation5™ Cell Imaging Multi-Mode Reader (BioTek). The succinate dehydrogenase activity of the samples was calculated according to the manufacturers’ directions. Data is representative of the means and standard deviations of three biological replicates.

Wildtype BW25113 or knockout mutants were grown overnight in 2 mL LB at 37°C with shaking. The following day, cells were pelleted by centrifugation and washed twice with 1 × PBS. The density was standardized to OD₆₀₀ 1.0 by suspending an appropriate volume of cells in 1 ml of 1 × PBS. 20 μL of each sample was loaded into a 96-well plate, each well containing 180 μL of either 0.4% glucose (Sigma), 0.4% malic acid (Sigma) or 0.4% succinic acid (Sigma) M9 minimal medium (Difco), pH 7.0. The plate was incubated in the Cytation5™ Cell Imaging Multi-Mode Reader with 330 rpm shaking at 37°C for 48 h.

Recent studies have demonstrated that the Cpx envelope stress response regulates the expression and activity of NDH-I and cytochrome bo₃ complexes in EPEC (Guest et al., 2017). In addition, Cpx-dependent downregulation of genes encoding the succinate dehydrogenase complex was observed in a large microarray dataset; however, the exact mechanism of regulation has not been revealed (Raivio et al., 2013). To investigate the impact of the Cpx pathway on the SDH, we grew strains carrying either the vector control or a reporter plasmid in which 500 bp upstream of the sdhCDAB operon were fused to the luxCDABE gene cluster in liquid LB medium aerobically for 24 h. Luminescence activity of the psdhC::lux reporter in the presence or absence of the Cpx pathway was measured every hour until the stationary phase was reached. We chose to use the luminescence activity produced by the pcpxP::lux reporter gene as a positive control, since CpxP is one of the most upregulated members of the Cpx regulon (Price and Raivio, 2009).

In agreement with previous observations, the expression of the pcpxP::lux reporter in the wildtype BW25113 was higher than in a cpxA null background at every stage of growth (Figure 1A). Along with cpxP, we assayed the activity of the pnuoA::lux reporter and validated its Cpx-dependent downregulation as reported in the past (Figure 1B; Guest et al., 2017). Loss of Cpx response resulted in a ∼2-fold increase in the pnuoA::lux activity compared to that of the wildtype. When we assayed the luminescence activity of the psdhC::lux reporter, we found similar increases in the absence of a functional Cpx response, while the wildtype activity of the Cpx TCS lead to transcriptional repression of the sdhC promoter (Figure 1C).

To corroborate the luminescence profile of the control and ETC reporters in liquid LB medium, we also examined expression of the reporter genes during growth on solid LB medium (Figure 1D). The patterns of pnuo::lux, psdhC::lux and pcpxP::lux reporter activity observed on solid medium resembled those in the liquid medium assay. In the wildtype strain, the pcpxP::lux reporter was strongly activated, while light was not detectable in a cpxA null mutant background. This pattern was reversed for the pnuo::lux and psdhC::lux reporters which demonstrated low activity in the presence of the intact Cpx pathway (Figure 1D).

Our results suggest that the Cpx TCS could directly inhibit the expression from the sdhC promoter. We thus decided to examine the region upstream of sdhCDAB for the presence of any putative CpxR binding sites and identified one approximately 150 bp downstream of the predicted sdhC transcription start site (TSS) by using Virtual Footprint¹ (Münch et al., 2005; Supplementary Figure 1).

Our experimental results with the psdhC::lux reporter predict that the enzymatic activity of SDH will be lower in the presence of an active Cpx TCS relative to a mutant lacking the Cpx response. To test this hypothesis, we assayed the rates of succinate oxidation in the wildtype E. coli strain, as well as in cpxR::kan and cpxA24 mutants, reflecting a deactivated and constitutively activated Cpx pathway, respectively (Raivio and Silhavy, 1997). We also included a strain overexpressing NlpE as an alternative way to induce the Cpx response (Figure 2). In this assay, the oxidation of succinate is accompanied by the transfer of electrons to an artificial electron acceptor (probe), which changes color depending on the enzymatic activity of the sample. When in the oxidized state, the probe is blue with maximal absorption at 600 nm, whereas when reduced it becomes colorless. We found that the activity of the SDH complex decreased upon activation of the Cpx system in both the cpxA24 mutant and the NlpE overexpression backgrounds. This result was predictable based on the repressive effect of the Cpx response on the sdhC promoter observed in Figures 1C,D.

Surprisingly, the loss of the Cpx response did not result in higher SDH activity, as might be anticipated with the relieved inhibition of sdhCDAB transcription. Instead, knocking out cpxR resulted in decreased succinate oxidation rates, equivalent to those observed in the presence of the strongly Cpx-activating cpxA24 allele (Figure 2B). This result is reminiscent of a previously observed decrease in the oxygen consumption rate in the ΔcpxR mutant, despite the fact that transcription of the operons encoding NDH-I and cytochrome bo₃ are upregulated in this background (Guest et al., 2017). Together, these findings suggest that the Cpx TCS not only transcriptionally regulates the SDH protein complex, but also regulates factors responsible for its proper activity and biogenesis.

Escherichia coli is capable of growing on a number of different sugar substrates, generating the majority of its ATP via the process of oxidative phosphorylation (Ammar et al., 2018). Carbon sources that cannot be utilized through the process of substrate-level phosphorylation, or fermentation, are called non-fermentable and require a functional ETC for sufficient energy generation. Such carbon sources include succinate, malate and glutamate (Prüss et al., 1994; Chang et al., 2004; McNeil et al., 2012). Poor growth of NDH-I mutants on either malate or succinate has been proposed to reflect low energy conservation efficiency due to a low level of ATP inside the cells (Kervinen et al., 2004). Previously reported data and our observations suggest that the Cpx TCS affects transcription and function of membrane complexes required for growth on non-fermentable carbon sources, including those associated with electron transport, the TCA cycle and oxidative phosphorylation (Figures 1B–D; Raivio et al., 2013; Guest et al., 2017). We thus hypothesized that the activity of the Cpx pathway would be required under conditions that create increased demand for respiration To test this, we spotted strains bearing luminescent reporter plasmids containing pnuoA::lux, psdhC::lux, or pcpxP::lux promoter fusions along a glucose-succinate carbon source concentration gradient on minimal medium (Figure 3). We found that expression of the cpxP promoter increases as the concentration of succinate increases, whereas in the presence of glucose pcpxP::lux demonstrates background levels of luminescence comparable to the empty vector control. None of the reporters were activated to high levels in the presence of glucose, which supports our hypothesis that growth under conditions that increase demand for respiration leads to Cpx pathway activation. Due to high pcpxP::lux activity, nuoA and sdhC promoter activities were not detected by the imaging system on short exposure times, therefore we spotted the strains containing ETC reporters on a different plate and assayed them separately to image the lower luminescence activity (Figure 3). As expected, activity of the pnuoA::lux and psdhC::lux reporters increased commensurate with the succinate concentration gradient. Notably, the expression of these transcriptional reporters was much higher in the absence of the Cpx response, which supports our previous findings in Figures 1B–D (data not shown). Based on the results of this experiment, we suggest that the cellular need for Cpx-mediated stress adaptation increases when respiratory (and possibly other) protein complexes are in increased demand.

The Cpx response directly represses nuo operon transcription to ensure envelope integrity during stress and reduce excessive protein traffic within the IM (Guest et al., 2017; Figure 1B). The possibility of additional levels of regulation has been hypothesized because the oxygen consumption in EPEC was found to be reduced in cells with either an active or inactive Cpx pathway (Guest et al., 2017). To determine whether the Cpx response regulates respiratory proteins beyond transcription, we constructed a plasmid that expresses a triple FLAG-tagged NuoA subunit of the NDH-I complex from an exogenous, IPTG inducible promoter. We chose NuoA protein due to its scaffolding role in the assembly of the NDH-I membrane arm (Friedrich et al., 2016). Given that the NuoA-3 × FLAG construct is expressed from a CpxR-independent IPTG-inducible promoter, any effects of Cpx pathway on NuoA protein levels should be independent of transcription. To confirm that the NuoA-3 × FLAG fusion protein is functional, we complemented a chromosomal ΔnuoA mutant that has a growth defect on minimal medium supplemented with malate, with either the empty vector pMPM-K3 or the plasmid carrying NuoA-3 × FLAG. The exogenously expressed NuoA restored growth of the ΔnuoA mutant, demonstrating that the protein was expressed properly and was functional, whereas complementation with an empty vector resulted in the absence of growth (Supplementary Figure 2). All strains demonstrated similar growth levels on minimal medium supplemented with glucose, which is a fermentable carbon source.

We assayed the amount of NuoA-3 × FLAG protein in wildtype, ΔcpxR and cpxA24 strains of E. coli via Western blotting and observed a decrease in NuoA-3 × FLAG levels when the Cpx response was constitutively activated. A slight accumulation of the protein was detected in the absence of Cpx response (Figure 4A). Expression of NuoA-3 × FLAG in the cpxA24 mutant was 8% that of the wildtype strain 30 mins after induction, whereas NuoA-3 × FLAG levels were increased by 14% in the ΔcpxR mutant compared to the wildtype (Figure 4A).

To gain more insight into how the activity of the Cpx response affects protein turnover over time, we performed a protein stability assay by utilizing the protein synthesis inhibitor chloramphenicol and monitoring the rate of NuoA-3 × FLAG degradation over 120 mins after translation has been halted (Figure 4B). Relative quantification of the bands showed that in the absence of the Cpx response, approximately half of the starting amount of NuoA-3 × FLAG protein was degraded in 45 mins, which was 25 mins longer than in the wildtype strain, indicating slower protein turnover (Figure 4C and Supplementary Table 3). Unexpectedly, the rates of NuoA-3 × FLAG degradation over the course of the experiment were comparable between the cpxA24 and the wildtype strains, although the total amount of protein degraded by the end of the experiment was larger in the cpxA24 (Figures 4B,C and Supplementary Table 3). Together, our results suggest that Cpx-regulated protein degrading factors are responsible for faster and more efficient turnover of NuoA-3 × FLAG proteins in the wildtype strain. Counterintuitively, an increased rate of degradation does not seem to be the only reason for the diminished amount of NuoA observed in cpxA24 strain backgrounds (Figure 4A).

We found that the rate of NuoA-3 × FLAG proteolysis was reduced when the Cpx response was inactivated, suggesting that the Cpx pathway assists protein turnover (Figures 4B,C). Given that bacteria with a compromised ETC are not able to generate sufficient amounts of ATP to support growth on non-fermentable carbon sources, we hypothesized that Cpx-regulated protein folding and degrading factors may impact the biogenesis of the ETC, together with the ability of the bacterial cells to generate proton motive force (PMF) and maintain their ATP levels. The candidate gene list for this experiment was derived from previous publications (Price and Raivio, 2009; Raivio et al., 2013) and unpublished RNASeq data. To identify genes of interest, we performed a preliminary screening of several of these envelope-associated protein folding and degrading factors (unpublished data). Cpx-regulated genes were overexpressed in a ΔnuoA strain containing the NuoA-3 × FLAG expression vector and NuoA-3 × FLAG protein levels were analyzed by dot blot. Overexpression of the DegP, HtpX, PpiD, and YccA proteins had the largest impact on the abundance of NuoA-3 × FLAG protein, surprisingly resulting in a greater than 2-fold increase in the amount of NuoA-3 × FLAG in comparison to the vector control. Furthermore, the protease/chaperone DegP, the IM protease HtpX, proteolytic modulating factor YccA, and periplasmic chaperone PpiD, have all been implicated in maintaining the integrity of the envelope and responding to stress generated by protein misfolding (Danese et al., 1995; Shimohata et al., 2002; Yamamoto and Ishihama, 2006; Matern et al., 2010; Raivio et al., 2013).

To investigate the role these proteins may have in the quality control of the ETC complexes, we analyzed growth of their knockout mutants in minimal media supplemented with glucose, succinate or malate in BW25113 E. coli (Figure 5; Kitagawa et al., 2005). As shown in Figure 5, deleting cpxR, ppiD, yccA, htpX, or degP resulted in growth defects of differing severity in comparison to the wildtype. All strains were able to grow on media supplemented with glucose, where their growth would not solely depend on energy generated via respiration. In addition, we noticed that the growth phenotypes of some mutants differed depending on whether the media contained malate or succinate as a carbon source.

Our data demonstrate that knocking out cpxR, ppiD, or yccA severely attenuates growth in media supplemented with malate, whereas these defects are less pronounced in succinate. Notably, deletion of genes encoding IM protease HtpX and periplasmic protease/chaperone DegP resulted in more modest growth defects (Figure 5). This result was somewhat surprising, as Cpx-mediated regulation of these proteolytic factors is an important element of envelope quality control (Jones et al., 2002; Shimohata et al., 2002), although it is possible their functions overlap with other proteolytic factors. To further demonstrate that these factors are required for respiratory growth, we performed a complementation assay where the mutants were transformed with an empty vector or a plasmid expressing the gene of interest. We were partially able to complement growth defects of every mutant in media where succinate or malate were the sole carbon source (Supplementary Figure 3). Together, these data demonstrate that the absence of a functional Cpx system and several associated protein folding and degrading factors compromises proper biogenesis of the ETC, resulting in decreased viability in media that demand respiration.

Our results implicate the Cpx-regulated envelope quality control factors YccA and PpiD, and to a lesser extent the proteases DegP and HtpX, in the biogenesis of the ETC (Figures 4, 5). We therefore hypothesized that knocking out ppiD, yccA, degP, or htpX may impact NuoA-3 × FLAG protein abundance. Here, we demonstrate that deletion of these factors alters NuoA-3 × FLAG protein levels compared to the wildtype (Figure 6). Deletion of degP, ppiD, yccA, or htpX increased NuoA-3 × FLAG abundance by a factor of 5.56, 4.83, 3.95, and 1.49, respectively, in comparison to wildtype. Although the specific fold changes in NuoA-3 × FLAG levels varied between biological replicates within this experiment, we consistently detected an increased amount in the mutant strains relative to the wildtype. In agreement with our previous findings, deletion of CpxR also resulted in accumulation of NuoA-3 × FLAG (Figures 4A, 6). Overall, our results suggest that several Cpx-regulated protein folding and degrading factors affect abundance of the NuoA protein and support the hypothesis of the Cpx-dependent regulation of the NDH-I complex at the post-translational level.