https://orcid.org/0000-0002-1469-4424
Figures
Abstract
The ATP-dependent cytoplasmic protease Lon has critical functions in protein quality control and cellular regulation in organisms across the three domains of life. In the opportunistic pathogen Pseudomonas aeruginosa, lon loss-of-function mutants exhibit multiple phenotypic defects in motility, virulence, antibiotic tolerance and biofilm formation. However, only a couple of native substrate proteins of Lon are described in P. aeruginosa until now and most of the phenotypes associated with Lon remain unexplained. Here, we searched for novel Lon substrates in P. aeruginosa by analyzing proteome-wide changes in protein levels and stabilities following lon overexpression. Our search yielded a large number of putative Lon substrates with diverse cellular functions, including metabolic enzymes, stress proteins and a significant fraction of motility-related proteins. In vitro degradation assays confirmed the metabolic protein SpeH, the heat shock protein IbpA as well as seven proteins involved in flagella- and type IV pilus-mediated motility as novel substrates of Lon. The new motility-associated substrates include both key regulators of motility (FliA, RpoN, AmrZ) as well as structural flagellar components (FliG, FliS and FlgE). Further, by isolating suppressor mutations bypassing the motility defect of lon- cells, we reveal that Lon-dependent degradation of the specific substrate SulA, a cell division inhibitor, is crucial for ensuring proper cell division and motility under optimal conditions. In sum, our work highlights Lon’s regulatory role in degrading functional proteins involved in critical cellular processes and contributes to a better molecular understanding of the pathways underlying Pseudomonas pathogenicity.
Author summary
Proteases specifically degrade other proteins inside cells and constitute promising drug targets due to their central roles in cell physiology. In the human pathogen Pseudomonas aeruginosa, the protease Lon has been implicated in the regulation of various processes that are linked to the ability of P. aeruginosa to tolerate antibiotic exposure and to establish infections. Here, we have determined the group of proteins that are directly degraded by Lon in P. aeruginosa. Among these proteins, we found several proteins known to be involved in swimming, swarming and twitching motility, which are critical behaviors of P. aeruginosa during host colonization. Further, we found that by degrading a cell division inhibitor protein, Lon ensures proper cell division, growth and motility under optimal growth conditions. Our work sheds light onto the molecular mechanisms linking the Lon protease to bacterial pathogenicity, which is critical for exploring the potential of Lon and other cellular proteases as antibiotic drug target.
Citation: Kallazhi A, Rahman A, Römling U, Jonas K (2025) Direct and indirect pathways linking the Lon protease to motility behaviors in the pathogen Pseudomonas aeruginosa. PLoS Pathog 21(6): e1013288. https://doi.org/10.1371/journal.ppat.1013288
Editor: Vincent T. Lee, University of Maryland, UNITED STATES OF AMERICA
Received: January 8, 2025; Accepted: June 10, 2025; Published: June 25, 2025
Copyright: © 2025 Kallazhi et al. This is an open access article distributed under the terms of the Creative Commons Attribution License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original author and source are credited.
Data Availability: All data generated or analysed during this study are included in the manuscript and supporting files. The mass spectrometry proteomics data have been deposited to the ProteomeXchange Consortium via the PRIDE partner repository with the dataset identifier PXD063095.
Funding: The study was funded by project grants from the Swedish Research Council (2016-03300, 2020-03545 and 2024-04942 to KJ), a future leaders grant from the Swedish Foundation for Strategic Research (SSF, FFL15-0005 to KJ), as well as funding from the Strategic Research Area (SFO) program distributed through Stockholm University (to KJ), Karolinska Institute (to UR) and the Lillian Sagens and Curt Ericssons Research Foundation (to UR). The funders had no role in the study design, data collection and analysis, decision to publish, or preparation of the manuscript.
Competing interests: The authors have declared that no competing interests exist.
Introduction
All cellular proteins are subject to strict regulation and quality control via the proteostasis network, which involves a set of protein machines that precisely control the synthesis, maintenance and degradation of proteins. One of the major components of this network are proteases, which are tasked with the degradation of proteins. This is vital, not only to remove misfolded or aggregated proteins, but also to precisely adjust the cellular concentrations of specific proteins according to the physiological and metabolic context of the cell [1]. One major group of proteases are the AAA+ (ATPases Associated with diverse cellular Activities) proteases that unfold and degrade substrate proteins in an ATP-dependent manner [2]. In order to precisely regulate the composition of the proteome, each of these proteases degrades a distinct panel of substrates, which are recognized through short amino acid recognition sequences, called degrons, and in some cases aided by adaptor proteins [3]. Most of these proteases exhibit a high degree of specificity and there are specific sets of proteins whose stability is exclusively controlled by each protease [4–6]. However, recent studies suggest that there is some overlap between the substrate pools of different proteases, especially under stress conditions [7–9].
Lon is one of the major cytoplasmic AAA+ proteases and is highly conserved across the three domains of life. It is comprised of a homo-hexameric ring, in which each monomer consists of an N-terminal domain for substrate recognition, an AAA+ domain for substrate unfolding using ATP hydrolysis and a C-terminal protease domain containing a Ser-Lys catalytic dyad required for proteolytic cleavage [10]. Lon was first described as a protein quality control protease that degrades un- or misfolded proteins as well as incomplete proteins [11,12]. A large body of subsequent work showed that it also has critical regulatory functions, thereby impacting a variety of important cellular processes [13]. In the model organisms Escherichia coli, Caulobacter crescentus and Bacillus subtilis, previously identified Lon substrates include regulators of stress responses, cell differentiation, and cell cycle progression, metabolic and structural proteins as well as antitoxin proteins of toxin-antitoxin systems [14–20]. For many years, relatively few native Lon substrates were known; however, recent developments in mass spectrometry-based proteomics approaches aided the search for protease substrates and resulted in the identification of hundreds of putative Lon substrates in selected model organisms [4,15,20]. In spite of these advances, the nature of Lon substrates still remains poorly studied in many other species, including bacterial pathogens.
In the opportunistic pathogen Pseudomonas aeruginosa, proteases including Lon have been shown to affect a variety of pathways that are linked to its pathogenicity [21–26]. P. aeruginosa is a common cause of community and hospital-acquired acute and chronic infections of the urinary tract, the airways and other bodily sites [27]. Due to its low nutritional needs, its high adaptability and the ability to develop antibiotic resistance, P. aeruginosa remains difficult to control. It is an ESKAPE pathogen listed by the WHO in the Bacterial Priority Pathogens List (WHO BPPL) [28,29]. Besides other virulence factors, a crucial contributor to the pathogenicity of P. aeruginosa is its ability to form biofilms [30]. Although biofilm formation is a predominantly sessile lifestyle, different forms of motility are required to successively initiate biofilm formation and shape the biofilm structure. It uses three major forms of motility called swimming, swarming and twitching, where the first two rely on rotations of its polar flagellum, while the latter relies on extensions and retractions of a type IV pilus. All three forms of motility have been shown to be critical for host colonization and cell invasion [31].
Based on phenotypic analyses of a lon loss-of-function mutant in P. aeruginosa, Lon has been reported to affect the three major types of motility as well as biofilm formation, virulence in amoeba and mouse lung infection models and antibiotic susceptibility [21–23,26]. Despite these wide-reaching effects, only a couple of proteins have been verified as direct Lon substrates in P. aeruginosa to date, namely the SOS-induced cell division inhibitor SulA that is also a well-studied substrate of E. coli Lon [18,32] and the RNA-binding protein Hfq [32]. Additionally, Lon has also been shown to be involved in the LasI/LasR quorum sensing pathway, where LasI degradation was shown to be dependent on Lon in vivo [33]. Given the multiple phenotypes exhibited by lon mutants and the wide range of Lon substrates that have been described in other model bacteria, Lon is expected to contribute to the regulation of a substantial number of proteins in P. aeruginosa.
To uncover novel substrates of Lon in P. aeruginosa and to unravel the molecular basis of previously reported phenotypes of lon loss-of-function, we used mass spectrometry-based quantitative proteomics to search for proteins with altered steady-state levels and stabilities in response to lon overexpression. Among a large group of putative substrates, we verified a metabolic protein SpeH, the heat shock protein IbpA and seven proteins associated with flagella- or type IV pili-mediated motility as direct Lon substrates. Furthermore, we demonstrate that Lon is critical for proper cell division and swimming motility in P. aeruginosa under optimal conditions by maintaining low levels of the cell division inhibitor SulA.
Results
Quantitative proteomics experiment reveals putative substrates of Lon
Previous work has shown that Lon substrates can be identified based on changes in their steady-state levels and protein stabilities in strains that either lack or overexpress lon [15,20]. Hence, we sought to find novel Lon substrates in P. aeruginosa by identification of proteins that exhibit reduced abundance and stability in lon-overexpressing cells (Fig 1A). We generated a lon overexpression strain (lonOE) of P. aeruginosa clone C clinical urine isolate 8277 containing plasmid-borne lon under the control of an arabinose-inducible promoter. Immunoblot analysis confirmed an approximately 2-fold upregulation of Lon levels within 90 min of induction (S1A Fig) in comparison to a vector control (VC) strain carrying the empty pJN105 vector.
(A) Strategy of searching for Lon substrate candidates based on changes in protein levels and stabilities upon lon overexpression (WT – wild type, lonOE – lon overexpression). (B) Volcano and dot plots representing the proteomics data in P. aeruginosa clone C isolate 8277. Volcano plots show proteome-wide effects of lonOE in comparison to the vector control (VC) after 90 min of lon induction (top left) as well as proteome-wide changes in protein levels 30 min (top center) and 60 min (top right) after spectinomycin-induced synthesis shut-down in the lonOE strain in comparison to the 0 min time point. The dotted line indicates −log10 p-value of 1.3 (corresponding to a p-value of 0.05). The dot plots below show comparisons of proteome-wide protein stabilities between the lonOE strain and the VC. Log2 fold-changes between 0 min and 30 min (bottom left) as well as 0 min and 60 min (bottom right), respectively, for the VC and the lonOE strains were plotted against each other. Motility-associated proteins and other selected top-ranked proteins as well as Lon protease itself are marked. p-values were calculated using a two-tailed t-test. Three independent biological replicates of 0 and 60 min and two independent biological replicates of 30 min for each of the control and lon overexpression strains were used. (C) Venn diagram showing selection of substrate candidates by applying three different criteria. Criterion I, green circle — proteins showing a lower steady-state level in lonOE compared to the vector control. Criterion II, orange circle — proteins showing a larger reduction between 0 min and 30 or 60 min in lonOE compared to the VC. Criterion III, red circle — proteins that are significantly degraded in lonOE, i.e., showing a significant (p < 0.05) downregulation between 30 or 60 min and 0 min. (D) Pie chart of the functional classification analysis of the 64 proteins satisfying all the criteria in Fig 1C. (E) Table of motility-associated proteins and other selected proteins found in the final list of Lon substrate candidates in Fig 1D along with their primary function in the cell, subcellular location and their functional classification. FliA and FlgE did not satisfy all three criteria, but were included in this list due to their functions in motility and their previous connections with Lon in other bacteria [15,34].
To assess relative protein stabilities between the lonOE and VC strain, we grew these strains in the presence of L-arabinose for 90 min before protein synthesis was shut off by addition of spectinomycin, to then follow the decay of proteins over time by collecting samples at 0, 30 and 60 minutes. We chose a spectinomycin concentration of 1 mg/mL, at which culture growth was strongly reduced, indicating efficient inhibition of protein synthesis (S1B Fig). Time point 0 min was sampled before spectinomycin addition and thus provided information about the steady-state levels of proteins after 90 min of induction of lon expression with L-arabinose. SDS-PAGE analysis showed no striking changes in the total protein fractions across the collected samples (S1C Fig). However, immunoblot analysis with the Lon specific antiserum revealed a mild reduction in Lon levels in both strains over time, indicating that Lon itself is subject to degradation (S1C Fig), a result that is consistent with a previous report [35].
To quantify proteome-wide protein levels and stabilities, we used a quantitative mass spectrometry approach coupled to TMT (Tandem Mass Tag) isobaric labelling and HiRIEF (High Resolution Iso-Electric Focusing) fractionation. A total of 2264 proteins were detected in all samples. Consistent with the immunoblot data (S1A Fig), Lon levels were significantly upregulated in the lonOE strain compared to the VC control strain (Fig 1B).
By comparing the 0 min time points of the lonOE and VC strains, we found that 1123 proteins were downregulated in response to lon overexpression (Fig 1B). To analyze Lon-dependent changes in protein stabilities, we first analyzed the log2 fold-changes between 0 min and 60 min as well as 0 min and 30 min for both the VC and the lonOE strains and compared the log2 fold-changes between both strains. A larger log2 fold change in the lonOE strain at 30 or 60 min compared to the VC indicated lower protein stability in response to lon overexpression (Fig 1B, bottom), and we found 1154 proteins to show this behavior. Finally, we applied a more stringent criterium of selecting only proteins that showed significant degradation in lonOE, i.e., a significant downregulation between 0 min and 30 or 60 min after spectinomycin addition (Fig 1B), which resulted in 182 proteins. It is noteworthy that a fraction of proteins showed an increase in abundance after inhibition of protein synthesis by spectinomycin (Fig 1B). This might be the result of some ongoing protein synthesis even after addition of spectinomycin (S1B Fig).
In order to shortlist candidates for potential Lon substrates, we combined all three criteria and selected proteins that showed reduced steady-state levels in the lonOE strain compared to the VC strain as well as significantly unstable proteins in lonOE (i.e., showing a significant decrease between 30 or 60 min and 0 min after spectinomycin addition) that also show reduced stability compared to the VC strain. This resulted in a final list of 64 proteins which satisfied all of these criteria and were considered to be potential candidates for further investigation (Fig 1C). Among the previously reported substrates, SulA and LasI were not detected, while Hfq showed mildly reduced levels upon lon overexpression, but did not satisfy the other criteria we applied.
Analysis of the annotated functions of these proteins showed that they represented all major functional categories, including a large fraction of metabolic proteins, transcriptional regulators and structural proteins, but also many uncharacterized proteins (Fig 1D). Metabolic proteins such as BioB, RlmN and SpeH as well as the heat shock protein IbpA, which is a known Lon substrate in E. coli [36], were among the top candidates in the putative substrates list sorted by highest steady-state level changes and were chosen for further analysis (Fig 1E). Interestingly, nine of the 64 shortlisted putative substrates were all associated with flagella- or type IV pilus-mediated motility, six of which were classified as structural proteins and three as transcriptional regulators. Additionally, we found the flagellar transcriptional regulator FliA and the flagellar hook protein FlgE, both of which are Lon substrates in other bacteria [15,34], in the list of proteins downregulated by lonOE. The high representation of motility-related proteins among the shortlisted candidates was promising, considering that motility is critical for P. aeruginosa virulence and has been shown to be affected by the Lon protease [26]. Hence, we also selected this group of motility-associated proteins for further investigation (Fig 1E).
In vitro degradation assays confirm Lon proteolysis of nine novel substrates
To test whether the selected substrate candidates are directly degraded by Lon, in vitro degradation assays using purified components were performed. Out of the eleven putative Lon substrates that were annotated to have functions in motility, five are cytoplasmic, including the alternative sigma factors RpoN and FliA, the pleiotropic transcriptional regulator AmrZ and two structural proteins of the flagellum (FliG, FliS), while the flagellar hook protein FlgE is extracellular. BioB, RlmN, SpeH and IbpA are all cytoplasmic proteins. We selected these proteins for their ease of purification, when compared to membrane-bound proteins. Some strains of P. aeruginosa, including clone C 8277, harbor a second gene, hereafter called FliS2 (SG17M genome locus tag: K0E51_RS23370), adjacent and homologous to canonical FliS (SG17M genome locus tag: K0E51_RS23365). Since the mass spectrometry analysis might have failed to discriminate peptides from these two FliS proteins, FliS2 was also included to be tested due to its homology to FliS. Further, we included the previously validated Pseudomonas Lon substrate SulA in our analysis [32].
First, in order to validate the activity of our purified Lon protease from P. aeruginosa, we monitored degradation of the model substrate β-casein [12]. Degradation of β-casein was observed in an ATP-dependent manner indicating that the purified Lon is enzymatically active (S2A Fig). As expected, we also observed fast in vitro degradation of SulA in the presence of ATP and creatine kinase that regenerates ATP, while remaining stable in the absence of ATP and creatine kinase (Fig 2A), confirming that this protein is efficiently degraded by Pseudomonas Lon [32]. Despite being strong candidates according to the proteomics data, BioB and RlmN did not show degradation by Lon in vitro (S2B and S2C Fig). This could mean that these proteins still undergo Lon-dependent degradation in vivo, but require additional factors which are absent in our in vitro system. SpeH (also known as SpeD2), a17 kDa S-adenosyl methionine decarboxylase proenzyme involved in an alternative spermidine biosynthesis pathway showed two bands on SDS-PAGE, as it is known to undergo autocatalysis into two chains which then form a hetero-tetramer [37,38]. Both of these bands showed robust degradation by Lon in vitro (Fig 2A), while IbpA exhibited a subtle but reproducible level of degradation (Fig 2A), which matches the comparably slow Lon-mediated degradation previously observed for E. coli IbpA [36]. Next, we tested the seven motility proteins in our in vitro degradation system. Notably, all of them showed robust degradation by Lon in the presence of ATP (Fig 2B), but no notable degradation in the absence of ATP. Together, these data demonstrate that all seven motility-associated proteins as well as IbpA and SpeH are bona fide substrates of Lon in P. aeruginosa, and also indicate that our proteomics-based search was successful in identifying novel Lon substrates.
(A) In vitro degradation assays of the previously reported Lon substrate SulA as well as the new substrates SpeH and IbpA in the presence (orange) and absence (grey) of ATP and creatine kinase (CK). 0.4 μM of Lon6 hexamer and 4 μM of substrate were used in all assays. Quantification of each of the degradation assays were done using Image Lab 6.1 and plotted using the arithmetic mean ± SD of the quantified values of remaining substrate normalized relative to corresponding Lon levels for each time point. Each graph represents quantifications of at least three independent experiments. One representative SDS-PAGE gel of the assay is shown for each protein. (B) In vitro degradation assays with quantifications as in (A), but with the motility-associated substrates FliG, FliS, FliS2, FlgE, RpoN, FliA and AmrZ. 0.4 μM of Lon6 hexamer and 3 μM of substrates were used, with the exception of FliG that was added at a concentration of 1.4 μM. Asterisks (*) depict unspecific bands.
lon disruption does not affect the steady-state levels of motility-associated substrates, but causes strong accumulation of SulA
According to our proteomics data, the newly validated substrate proteins are affected by Lon in vivo under lon-overexpressing conditions (Fig 1). To assess how much Lon contributes to regulating its substrate proteins in wild-type cells grown under optimal conditions, we wanted to compare the levels and stabilities of these proteins in a lon loss-of-function mutant to the wild type background. Since a lon mutant of P. aeruginosa clone C strain 8277 was not available, and since we did not succeed in generating a new lon deletion mutant in this background, we turned for the remaining experiments of this study to P. aeruginosa PAO1. This strain possesses Lon and most validated substrate proteins identical to clone C strain 8277, with the exceptions of FliS, which shares 76% similarity between the strains and FliS2, which is not present in PAO1. We again used quantitative MS-based proteomics, but this time to specifically analyze the steady-state levels and stabilities of the new Lon substrates as well as SulA between the PAO1 wild type strain and a lon- strain, in which the lon gene has been disrupted by a transposon insertion [39]. SulA was either not detected or present at very low levels in the wild-type replicates, but was present at high levels in all replicates of the lon- strain, indicating that Lon is critical for limiting SulA levels in wild-type cells during optimal conditions through degradation (Fig 3A). In contrast to SulA, none of the other validated substrates, showed a clear change in steady-state levels in the lon- strain (Fig 3A).
(A) Relative steady-state levels of various Lon substrates in PAO1 WT (grey) and lon- (orange) backgrounds as determined by quantitative proteomics. Relative means ± SD of three independent replicates shown, with the exception of SulA in WT that was not detected in two of the replicates. (B) Protein stabilities of the newly identified Lon substrates in WT and lon- as determined by quantitative proteomics. Changes in protein levels over 0, 30 and 60 min after spectinomycin-induced protein synthesis shut-off are shown for the WT (grey) and the lon- (orange) mutant. Plotted points represent the mean values ± SD for three independent replicates relative to time point 0. A simple linear regression analysis shows that the slopes of the fitted lines significantly differ for FliG (-0.003919 for WT and 0.0003186 for lon-, p = 0.0227, *) and FlgE (-0.01380 for WT and -0.002926 for lon-, p = 0.0071, **). For all other proteins, the differences were not significant (p > 0.05).
To also assess the stability of the newly identified substrates in the wild-type and lon- strains, we monitored the decay of these proteins after spectinomycin-induced protein synthesis shut-off. This experiment showed that FliG and FlgE were significantly stabilized in the lon- strain compared to the WT, indicating that Lon contributes to their degradation under optimal conditions (Fig 3B). In the cases of FliS, RpoN, SpeH and IbpA, we observed notable protein degradation in the wild type within 60 min after spectinomycin addition, which was however not significantly affected by absence of Lon in the lon- strain. Finally, FliA and AmrZ showed little to no degradation even in the wild type strain within 60 min after protein synthesis shut-off, indicating that these proteins might be protected from proteolysis under the chosen condition (Fig 3B).
Collectively, these results show that while Lon is critical for maintaining low cellular levels of SulA, absence of Lon proteolysis does not cause major changes in the abundance of the other substrate proteins. This result might be due to compensatory effects, either by other proteases or through transcriptional regulation, as well as mechanisms protecting substrates from degradation under the chosen conditions, as previously reported for FliA in E. coli that is stabilized by its anti-sigma factor FlgM during optimal growth [34].
lon disruption leads to defects in growth, morphology and swimming motility
With the increased knowledge about Lon substrates in P. aeruginosa, we wondered if we can elucidate the molecular basis of the previously reported phenotypes of the lon- mutant. First, we wanted to re-assess some of the reported phenotypes of the lon- mutant in comparison to the wild type, lon overexpression as well as a complemented lon- strain containing a plasmid-borne arabinose-inducible copy of lon. Immunoblot analysis confirmed that Lon was absent in the lon- strain, but present in the WT and the complemented lon- strain (S3 Fig). Consistent with previous work [26], we observed that the lon- strain showed reduced swimming motility in soft agar (Fig 4A) as well as reduced twitching motility (Fig 4B) and impaired biofilm formation when quantifying bacterial attachment in microtiter plates using crystal violet staining (Fig 4C). Additionally, we observed a lesser degree of pigmentation of the lon- strain, indicative of reduced secretion of the secondary metabolite pyocyanin (Fig 4D). Importantly however, while the swimming and pigmentation defects could be complemented by expressing plasmid-borne lon, this was not the case for the biofilm and twitching phenotypes. In these cases, complementation of the lon disruption mutant strain with plasmid-borne lon exhibited even lower levels of twitching and biofilm compared with the vector control. lon overexpression (WT Para lon) had no significant effect on any of the phenotypes (Fig 4A–D).
(A) Swimming motility in soft agar with arabinose after 40–48 hours at 25°C for the strains shown. The means of motility area (cm2) ± SD of three independent biological replicates plotted for each strain. Statistical significance was calculated using one-way ANOVA (Tukey’s multiple comparisons test) for comparison of pairs: WT VC vs WT Para lon, p = 0.8999 (not significant, ns); WT VC vs lon- VC and lon- VC vs lon- Para lon, p < 0.0001, ****. (B) Twitching motility in agar plates with arabinose after 8 days at 25°C of the strains shown. Means of motility zone area ± SD of three independent biological replicates plotted. Statistical significance was calculated using one-way ANOVA (Tukey’s multiple comparisons test) for comparison of pairs: WT VC vs WT Para lon, p = 0.0948 (not significant, ns); WT VC vs lon- VC and lon- VC vs lon- Para lon, p < 0.0001, ****. (C) Absorbance of crystal violet stained biofilms grown for 8 hours and measured at 550 nm normalized against cell growth at OD600. The means ± SD of three independent biological replicates plotted for each strain. Statistical significance was calculated using one-way ANOVA (Tukey’s multiple comparisons test) for comparison of pairs: WT VC vs WT Para lon, p = 0.5083 (not significant, ns); WT VC vs lon- VC, p < 0.0079, **; lon- VC vs lon- Para lon, p < 0.0030, **. (D) Pyocyanin produced by late exponential cultures of the four strains grown at 37°C as visualized by supernatants. Right — Means ± SD of absorbance scans of the supernatants from three biological replicates at 500–900 nm. (E) Growth curves of the four strains grown at 37°C and measured at OD600. Means ± SD of three independent biological replicates plotted. (F) Phase-contrast microscopy images of the four strains when grown at 37°C to logarithmic phase. Right — Violin plot shows quantifications of 2850 cells for each strain from three biological replicates. Statistical significance was calculated using one-way ANOVA (Tukey’s multiple comparisons test) for pairs: WT VC vs WT Para lon, p = 0.9963 (not significant, ns); WT VC vs lon- VC and lon- VC vs lon- Para lon, p < 0.0001, ****.
We reasoned that changes in growth rate and cell morphology might contribute to the observed Lon-dependent swimming defect. Hence, we also recorded growth curves of the different strains and analyzed cell morphology by phase contrast microscopy. These data show that the lon- strain exhibited a mild growth defect (Fig 4E) and also showed clear cell division defects resulting in cell elongation (Fig 4F), consistent with previous data [21]. Both these phenotypes were fully complemented by providing lon in trans.
Together, our phenotypic analyses confirm the previous finding that Lon is required for proper growth, cell division and swimming motility. In contrast, however, the previously reported phenotypes in twitching and biofilm formation could not be clearly attributed to Lon.
Suppressor mutations in sulA restore motility of the lon disruption strain
Based on the phenotypic results, we reasoned that the swimming defect of the lon- mutant is either caused by one or several of the motility proteins that are Lon substrates, or stabilization of another substrate that affects motility more indirectly, for example through defects in cell division and slower growth.
To uncover the genetic cause of the motility defect of the lon- strain, we decided to take an unbiased approach, in which we isolated suppressor mutations restoring motility in the lon- mutant. For this, we inoculated soft agar motility plates with the lon- strain and briefly exposed the plates to UV light to promote mutagenesis. After 48–72 hours of incubation, we observed suppressor mutants forming sectors of increased swimming diameters that emerged from the swimming zone of the original lon- strain (Fig 5A). Independent suppressor mutants from different biological replicates were isolated and swimming motility of the suppressors was confirmed to be significantly higher than that of the lon- strain (Fig 5A). Whole genome sequencing of six of these suppressors (referred to as Sup1–6 hereafter) alongside the WT and lon- strains revealed that each of them contained a mutation in the sulA gene when compared to the original lon- strain (Fig 5B). Sup1, 2, 3 and 4 carried the same point mutation leading to a H127L amino acid substitution in SulA, Sup5 contained another point mutation leading to a D109V amino acid substitution and Sup6 harbored a deletion in sulA (Δ133–148 nt) resulting in a frameshift mutation, which in turn caused a truncation of the corresponding SulA protein. All suppressor mutants restored not only motility, but also normal cell division (Fig 5C), suggesting that all of them are loss-of-function mutations that disrupt SulA’s function as a cell division inhibitor. On the other hand, they did not restore pyocyanin production and rescued growth only partially, indicating that Lon might be affecting these phenotypes through other substrates (S4A and S4B Fig).
(A) Image showing suppressor zones emerging from the motility zone of the lon- strain on a soft agar plate incubated at 25°C for 48–72 hours after UV-irradiation. Right — The means of quantified motility zone area (cm2) ± SD of two (WT, lon-) or three independent biological replicates (Sup1–6) plotted. Statistical significance was determined using one-way ANOVA (Tukey’s multiple comparisons test) for pairs: WT vs lon-, p = 0.0230, *; lon- vs Sup1, 2, 3, 4, 6, p < 0.0001, ****; lon- vs Sup5, p = 0.0003, ***. (B) Illustration of the location and nature of the six suppressor mutations (red) on a simplified structure of SulA depicting alpha helices (green) and beta sheets (blue). (C) Phase-contrast microscopy images of WT, lon- and the six suppressor strains after arabinose induction grown at 37°C to logarithmic phase. Right — Violin plot shows the quantification of 2850 cells for each strain from three independent biological replicates. Statistical significance was calculated using one-way ANOVA (Tukey’s multiple comparisons test) for pairs: WT vs lon- and lon- vs Sup1, 2, 3, 4, 5, 6, p < 0.0001, ****. (D) Phase-contrast microscopy images of the strains shown after arabinose induction, grown at 37°C to logarithmic phase. F – 3xFLAG tag. fs – frameshift. (E) Representative immunoblot of the strains shown after arabinose induction, using an anti-FLAG antibody. ‘*’ on the blot indicates the expected location of F-SulAfs. Right — Quantification of immunoblots based on three independent biological replicates. Statistical significance was calculated using one-way ANOVA (Tukey’s multiple comparisons test) for pairs: F-SulA vs F-SulAH127L, p = 0.0297, *; F-SulA vs F-SulAD109V, p = 0.0144, *. N.D – not detected. (F) Swimming motility in soft agar with arabinose after 40–48 hours at 25°C for the strains shown. (G) Top — Phase-contrast microscopy images of WT and WT treated with 0.5 μg/mL piperacillin for two hours grown at 37°C to logarithmic phase. Middle — Swimming motility in soft agar without and with 0.5 μg/mL piperacillin after 40–48 hours at 25°C. Bottom left — Violin plot showing the quantification of 2250 cells for each condition from three independent biological replicates. Statistical significance was calculated using unpaired t-test for LB vs LB + pip 0.5, p < 0.0001, ****. Bottom right — The means of motility zone area (cm2) ± SD of three independent biological replicates plotted. Statistical significance was calculated using an unpaired t-test LB vs LB + pip 0.5; p < 0.0001, ****. Pip – piperacillin.
To analyze more carefully how the suppressor mutations affect SulA function, we ectopically expressed SulAH127L (Sup1– 4), SulAD109V (Sup5) and the truncated SulAfs (frameshift in Sup6) with a 3xFLAG-tag in the wild-type strain background and compared their effects on cell division to a strain expressing 3xFLAG–tagged wild-type SulA (F-SulA). While the wild-type F-SulA caused very strong cell filamentation, consistent with SulA’s function as cell division inhibitor [40], expression of the mutants led either to no or moderate filamentation (Figs 5D and S5A), showing that they either completely or partially lost their ability to block cell division. Interestingly, the moderate cell filamentation caused by F-SulAD109V was not seen in the corresponding suppressor mutant in the lon- background, which completely restored motility and cell division (Fig 5C). This can likely be explained by higher protein levels when expressed from the pBAD plasmid compared to the native expression level from PsulA. Probing the levels of the different SulA variants with an anti-FLAG antibody showed that F-SulAH127L and F-SulAD109V were clearly detectable, but at reduced levels compared to wild-type SulA (Fig 5E). The truncated SulAfs variant was not detected, likely due to the truncated protein being unstable.
Next, we wondered how expression of the different SulA variants in the wild-type background would affect motility. While expression of F-SulA resulted in almost complete inhibition of motility, expression of F-SulAH127L affected motility only mildly compared to the vector control strain (Figs 5F and S5B). Expression of the F-SulAD109V variant, which caused moderate cell filamentation (Fig 5D), also impaired motility, but not as much as expression of wild-type F-SulA (Figs 5F and S5B). In the absence of the inducer arabinose, leaky expression of F-SulA was still enough to strongly inhibit motility, while no significant swimming defects were observed under this condition for the strains harboring F-SulAH127L or F-SulAD109V (S5C Fig). Together, these results demonstrate that SulA-induced cell filamentation correlates with reduced soft-agar motility, and suggest that the SulA suppressor mutants restore motility of the lon- mutant by reducing cell filamentation.
In order to further establish the connection between cell division inhibition and soft agar motility, we opted to inhibit cell division in a SulA-independent way and tested how this affects motility. Piperacillin is a penicillin-like antibiotic known to inhibit cell division by binding FtsI, preventing peptidoglycan synthesis and inhibiting FtsZ ring formation [41,42]. Consistent with previous data [43], we found that it induces strong cell filamentation of P. aeruginosa at a sub-inhibitory concentration of 0.5 μg/mL (Fig 5G), at which growth is not notably affected (S6 Fig). Importantly, adding this piperacillin concentration to swimming agar plates resulted in a severe motility defect (Fig 5G), confirming an inverse correlation between cell filamentation and soft-agar motility. Based on this result, we suggest that Lon indirectly promotes motility by lowering SulA levels and hence ensuring proper cell division. Hence, accumulation of SulA in the lon- results in cell division defects that cause the formation of filamentous cells that are less motile in soft agar.
A secondary mutation in pilI causes twitching and biofilm defects in the lon- strain
Having sequenced the genome of the lon- strain, we also searched for mutations that may explain the twitching and biofilm phenotypes of this strain that were not possible to complement with plasmid-borne lon expression (Fig 4B and 4C). Indeed, we found a single base pair deletion causing a frameshift mutation in the pilI coding region that was absent in the parental PAO1 strain. This frameshift mutation results in a stop codon immediately downstream of the single base pair deletion and thus in the translation of a short, truncated PilI peptide (Fig 6A). pilI encodes a putative type IV fimbrial biogenesis protein and is part of the chemotaxis-like gene cluster that mediates twitching motility [44] and is also predicted to affect biofilm formation [45]. To test whether the background mutation in pilI is the reason for the defects in these phenotypes observed in the lon- strain, we provided a functional copy of the pilI gene on a plasmid to the lon- strain. Indeed, the defect in biofilm formation (Fig 6B) was fully recovered with a functional pilI gene despite the strain still lacking functional Lon. We also assayed twitching motility and observed that expression of wild-type pilI significantly increased the twitching of the lon- strain, but not to wild type levels (Fig 6C). This could be due to the SulA-dependent cell division defect of the lon- strain partially contributing to the lowered twitching motility. These results indicate that the biofilm and twitching defects of the lon- strain can be attributed to a large extent to the background mutation in pilI. It also shows that the PilI protein of the pilus biogenesis gene cluster, which remains to date poorly characterized, plays an important role in biofilm formation and type IV pilus-mediated twitching of P. aeruginosa.
(A) Illustration of the single base pair deletion mutation on the pilI gene on the pilGHI gene cluster on the genome causing a frameshift leading to a stop codon immediately downstream. (B) Absorbance of crystal violet stained biofilms grown with arabinose for 8 hours and measured at 550 nm normalized against cell growth at OD600. The means ± SD of at least three independent biological replicates plotted. Statistical significance was calculated using one-way ANOVA (Tukey’s multiple comparisons test) for comparison of pairs: WT VC vs WT Para pilI, p = 0.2359 (not significant, ns); WT VC vs lon- VC, p < 0.0222, *; lon- VC vs lon- Para pilI, p < 0.0005, ***. (C) Twitching motility in agar plates with arabinose after 5 days at 25°C of the strains WT VC, WT Para pilI, lon- VC and lon- Para pilI. Means of motility area (cm2) ± SD of at least three independent biological replicates plotted. Statistical significance was calculated using one-way ANOVA (Tukey’s multiple comparisons test) for comparison of pairs: WT VC vs WT Para pilI p = 0.9939 (not significant, ns); WT VC vs lon- VC p < 0.0001, ****; lon- VC vs lon- Para pilI p = 0.0005, ***.
Discussion
The Lon protease is increasingly recognized as a critical regulatory protease that contributes to the coordination of cellular behaviors and stress responses by specifically degrading functional proteins [13,14,20,46,47]. In P. aeruginosa, Lon has been associated with numerous cellular processes that are connected to its pathogenic and highly adaptable life style [21–23,26,33]. However, the mechanistic basis by which Lon affects these processes remained poorly understood, partly due to limited knowledge about native substrate proteins in this organism. In this study, we filled these knowledge gaps by describing novel substrates of Pseudomonas Lon with diverse functions and by linking the Lon-dependent degradation of specific substrate proteins to previously reported phenotypes (Fig 7).
Under optimal conditions Lon is critical for maintaining low SulA levels which allows cell division to take place, which in turn is needed for proper motility. Lon also degrades various motility proteins, including FliG, FliS, FliS2, FlgE, FliA, RpoN and AmrZ. The degradation of these proteins likely contributes to the downregulation of motility under specific growth conditions, for example during the transition from a motile into a sessile life style.
Using a proteomics-based approach, we identified a large group of candidate substrate proteins based on changes in their steady-state levels and stabilities in cells overexpressing lon. We confirmed nine of them as bona fide Lon substrates. Interestingly, despite showing robust Lon-mediated degradation in vitro, the levels of the validated Lon substrates, with the exception of SulA, were not notably affected by lon loss-of-function. In case of the motility proteins, their Lon-dependent degradation was also dispensable for proper motility under optimal growth conditions. This suggests that absence of Lon is either compensated, for example by transcriptional changes or degradation by other proteases, or that Lon proteolysis of these substrates is more critical under alternative growth conditions. Our experiments were conducted in exponentially growing cultures in rich media conditions. Under this condition, motility is known to be favored [48] and thus, the need for degrading motility proteins is expected to be low. However, at the entry to stationary phase and under various conditions during host colonization and invasion, motility is downregulated and subpopulations of cells will instead transition to sessile lifestyles [49]. It is possible that Lon proteolysis of motility proteins is critical during such growth transitions. Alternatively, Lon proteolysis might play a homeostasis-promoting role and remove flagella proteins that are in stoichiometric excess and not bound to their proper binding partners, as previously suggested [50]. Similarly, IbpA degradation is expected to be particularly needed for homeostatic purposes or under conditions during stress recovery, when there is a need for downregulating IbpA. Deciphering the exact conditions under which Lon-mediated degradation of the newly identified substrates takes place is however challenging due to the difficulty in monitoring protein degradation precisely during growth transitions and in individual cells.
In spite of significant differences in the flagellar regulon cascade and amino acid sequences of homologous proteins, Lon was shown to degrade motility-associated proteins in several other bacteria. The flagellar hook protein FlgE and the flagellar sigma factor FliA, which we have identified here as Pseudomonas Lon substrates, have also been described as Lon substrates in E. coli [15,34]. In C. crescentus, the flagellar hook length regulator FliK, the flagellar assembly regulator FliX as well as FlgE were reported as Lon substrates [20] and in B. subtilis Lon degrades the master flagellar activator SwrA [19]. These data suggest an evolutionarily conserved role of Lon in degrading motility-associated proteins independent of protein identity and precise amino acid sequences (as exemplified by FliS and FliS2).
In addition to revealing novel direct substrates of Pseudomonas Lon, our study explains some of the previously reported phenotypes of lon loss-of-function. Our data demonstrate that stabilization and strong accumulation of the cell division inhibitor SulA in lon- cells results in cell division defects and cell filamentation, which in turn impairs motility (Fig 7). The filamentous morphology of cell-division inhibited cells is expected to hamper their penetration through the soft agar. Moreover, given that P. aeruginosa is monotrichous, its single flagellum is expected to operate less efficiently on cells with strongly increased cell mass.
Although SulA was previously shown to be a substrate of Pseudomonas Lon [32], it is noteworthy that it possesses a C-terminus that is clearly dissimilar from the C-terminal degron sequence of E. coli SulA that mediates recognition by Lon [51]. The physiological consequences of Lon-dependent degradation of SulA remained also unclear in P. aeruginosa and have not been linked to motility behaviors and other phenotypes of the lon- strain. Our data show that SulA stabilization in this strain is the major cause of its swimming motility and cell division defects and it is possible that other phenotypes observed in lon- mutants, such as reduced virulence in infection models or increased antibiotic susceptibility can be explained by the presence of excess SulA [21,22]. It is notable that the suppressor mutations D109V and H127L in SulA were both in close proximity to regions (Gln106 to Cys108 and Val120 to Gly125) predicted to be involved in FtsZ binding [52]. Hence, these mutations might disrupt the interaction between SulA and FtsZ, thus resulting in a complete or partial loss of SulA’s inhibitory function on cell division. The importance of Lon in limiting SulA levels in wild type cells is interesting and potentially suggests a role for Lon in adjusting cell division under certain environmental or stress conditions, for example, caused by exposure to antibiotics. In this scenario, reduced availability or activity of Lon in response to certain conditions would trigger cell filamentation. Based on prior studies it was suggested that SulA-induced cell filamentation constitutes a survival mechanism that helps bacteria to protect themselves against phagocytosis by immune cells [53] and to accelerate resistance development through homologous recombination [54].
In conclusion, our study has shed light onto the cellular roles of Lon in the opportunistic pathogen P. aeruginosa by revealing novel substrates through quantitative proteomics and by explaining several of the previously reported phenotypes of lon loss of function. Our findings reveal a role of Lon in degrading motility-associated proteins and its importance in limiting the amounts of the cell division inhibitor SulA, and we also report a background mutation in the lon- strain, thus clarifying previously reported phenotypes in twitching and biofilm formation [22,26]. Furthermore, our study has revealed many other putative Lon substrates that remain to be further studied in the future. Altogether, our work illustrates how cytoplasmic proteases such as Lon contribute to the regulation of functional proteins thereby affecting critical cellular activities and also contributes to a better understanding of the molecular pathways underlying P. aeruginosa pathogenicity.
Materials and methods
Growth conditions
P. aeruginosa strains Clone C 8277 and PAO1 strains were routinely grown in LB medium at 37°C while shaking at 200 rpm unless otherwise indicated. For all experiments involving plasmid induction, 0.2% arabinose was added to both the expression strain as well as the control strain with the empty vector, for at least 90 min before sampling. When needed, the medium was supplemented with either gentamicin (30 µg/ml) or tetracycline (50 µg/ml).
The salt-inducible BL21-SI/pCodonPlus E. coli strain for protein expression was grown at 30°C using LBON/2xYTON no salt medium. When necessary, media was supplemented with antibiotics at following concentrations (concentration in liquid/solid media in µg/ml): gentamicin (15/15), kanamycin (30/50) and chloramphenicol (20/40). E. coli DH5α strain used for cloning was grown at 37°C.
Plasmid construction
All plasmids used in this study are listed in S2 Table.
Expression plasmids.
For constructing plasmid pJN105:lon, the lon gene from Clone C isolate 8277 was cloned into a pJN105 vector with gentamicin resistance cassette for selection, under the control of an L-arabinose-inducible promoter. The gene was inserted between restrictions sites of NheI and PstI on the plasmid backbone through restriction cloning. The same was done for constructing pJN105:pilI. For sulA and its mutants, a 3x FLAG tag was inserted in the pJN105 plasmid between the NheI and NdeI restriction sites and the sulA mutants were cloned into the NdeI site to have the 3xFLAG tag at the N-terminus of the proteins. All primers used are mentioned in S3 Table. The reaction was used to transform E. coli DH5α competent cells which were selected on LB plates containing gentamicin. The plasmid was isolated and sequence-verified. The pJN105 empty vector (VC) and cloned overexpression plasmids were inserted through electroporation into electrocompetent P. aeruginosa cells of either WT or lon- background to create the vector control, overexpression or complementation strains. The successful transformants were selected on LB plates containing gentamicin after overnight growth at 37°C.
Protein expression plasmids.
Plasmids for protein expression were generated by amplifying the corresponding gene from chromosomal DNA of Clone C strain 8277 using primer pairs as listed in S3 Table and subsequent cloning of the insert into a pSUMO-YHRC vector backbone by Gibson assembly. The plasmid was amplified as two fragments from the original using primer pairs oMJF34/oMJF36 and oMJF37/oMJF38 (S3 Table) disrupting the kanamycin resistance cassette. The two fragments and the gene to be cloned were then joined using Gibson assembly and the reaction was used to transform chemically competent DH5α cells which were selected on LB plates containing kanamycin. The plasmid was then extracted and sequence-verified. The primer pairs used for cloning of each gene are as listed in S3 Table.
Sample collection and preparation for proteomics experiments
For the lon overexpression proteomics experiment, samples were collected and prepared as follows: Cultures of P. aeruginosa Clone C 8277 containing either the empty vector pJN105 or pJN105:lon were grown in triplicates of 200 mL LB cultures containing gentamicin for plasmid maintenance, into exponential phase cultures by back diluting 3–4 times. After collecting 1 mL pre-induction samples for immunoblot analysis, 0.2% arabinose was added to both cultures and grown at 37°C for 90 min. 1mg/ml spectinomycin was added to both the cultures and 50 mL of the culture was immediately retrieved into pre-cooled 50 mL tubes, spun down in a pre-cooled centrifuge and the supernatant discarded. The pellet was frozen using liquid nitrogen and stored at –20°C. OD600 values were measured throughout the experiment at hourly intervals. This was repeated at 30 min and 60 min after addition of spectinomycin. Three replicates of 0 and 60 min samples and two replicates of 30 min samples were collected.
For the lon- proteomics experiment, P. aeruginosa strain PAO1 WT and lon transposon mutant lon- were grown in triplicates. After back-diluting 3–4 times, the exponential phase cultures were treated with 1 mg/mL spectinomycin and samples were collected at 0, 30 and 60 min as described before. Three replicates of 0, 30 and 60 min samples were collected.
Immunoblot samples were taken at all time points. The samples were later transferred to –80°C for storage until delivery to the MS facility.
Quantitative proteomics
Proteomics analyses were conducted by the Clinical Proteomics Mass Spectrometry Facility, Karolinska Institute/Karolinska University Hospital/Science for Life Laboratory. The samples were lysed and subjected to protein digestion using trypsin followed by multiplex TMT (Tandem Mass Tag) isobaric labelling and HiRIEF (High Resolution Iso-Electric Focusing) fractionation before loading onto the LC-MS/MS. The multiplexing allowed all samples in each proteomics set to be analyzed simultaneously. The TMT16-plex was used for lonOE proteomics (16 samples) which included 3 replicates each of time points 0 and 60 min and two replicates of time point 30 min for each strain background. TMT18-plex was used for lon- proteomics (18 samples) which included 3 replicates each of time points 0, 30 and 60 min for each strain background. The reference genome used for identifying the proteins from unique peptides was that of P. aeruginosa Clone C SG17M (NCBI RefSeq: JALF01000000) for lonOE proteomics and P. aeruginosa PAO1 (NCBI RefSeq: NC_002516.2) for lon- proteomics. The source data for lonOE proteomics is available as S1 Data and the source data for lon- proteomics is available as S2 Data in supporting information.
Proteomics data analysis
In the lonOE proteomics experiment, a total of 2264 proteins were detected in all samples. The proteins were selected using three criteria: based on reduction in steady-state levels, reduction in stability over time in lonOE compared to VC and significant instability in lonOE. The data was analyzed for steady-state level changes by calculating the log2 fold-change between the 0 min samples of lonOE and VC for each protein and all proteins which show reduced relative levels in lonOE were chosen. The log2 fold-change between 0 min and 60 min as well as 0 min and 30 min of each protein of each strain was analyzed and all proteins which show greater fold change in lonOE over time in 30 or 60 min were selected. Finally, in order to shortlist candidates with the most significant instability in lonOE, proteins with the highest log2 fold-change between 0 min and 30min or 60 min with p < 0.05 were selected. All criteria were independently applied on the original list of 2264 total proteins detected. In order to shortlist as many relevant candidates as possible, the selection was not stringent and all the proteins that satisfied these criteria were included without setting more cut-offs. Venn diagram was created using https://www.deepvenn.com [55]. The functional analysis of shortlisted candidates was performed using the ontology ‘Protein class’ in the PANTHER 19.0 database on www.pantherdb.org [56] and the unclassified proteins were further assigned functions based on The Pseudomonas Genome Database available on https://www.pseudomonas.com [57].
In the lon- proteomics experiment, the proteins of interest as shown in Fig 3 were selected for analysis. The steady-state level changes for each protein in WT and lon- was compared using the protein level at 0 min in the two strains. The relative stability was analysed by dividing the 30 min and 60 min values of protein levels by the corresponding 0 min value for each protein in each strain background.
Protein purification
Purification of proteins was adapted from Holmberg et al. [58]. BL21-SI/pCodonPlus cells were transformed using the plasmids mentioned in the S2 Table by electroporation and selected on LBON agar plates supplemented with kanamycin (Kan) and chloramphenicol (Chlor). Pre-cultures (LBON or 2xYTON + Kan + Chlor) were inoculated with about 20 colonies and cultivated at 30°C overnight. One litre of 2xYTON + Kan + Chlor was innoculated by 1:100 dilution of the pre-culture and cultured until an approximate OD600 of 1.0. The expression was started by addition of 0.5 mM IPTG and 0.3 M NaCl (final concentrations). Incubation continued at 30°C for 4 hours and cells were harvested subsequently by centrifugation (6 800 × g, 4°C, 10 min) and cell pellets stored at −80°C.
For purification, pellets were resuspended in HNG10Im (40 mM HEPES–KOH pH 7.5, 500 mM NaCl, 10% glycerol, 10 mM imidazole), supplemented with 1 mM PMSF, 1 mg/mL Lysozyme and 3 µL Benzonase/10 mL and topped up to 30 mL total volume. Cells were then lysed by 2–3 passes through an EmulsiFlex-C3 high-pressure homogenizer and peak pressure was kept between 25 000 and 30 000 psi. Lysate was cleared by centrifugation at 32 500 × g at 4°C for 1.5 h. Tagged proteins were bound to 3 mL (=1.5 mL bed volume) pre-equilibrated Talon SuperFlow Metal Affinity Resin from TaKaRa per liter culture on ice for 30 min with shaking. After washing 5 times with approximately 50 mL HNG10Im, bound proteins were eluted using HNG + 250 mM imidazole and fractions with protein concentrations ≥0.2 mg/mL were pooled. For 6 × His-SUMO tag removal, 4 µg/mL Ulp1–6 × His was added and imidazole was removed in parallel by dialysis against HNG (40 mM HEPES-KOH pH 7.5, 500 mM NaCl, 10% glycerol). Tag depletion was achieved by binding to 1 g dry Protino Ni-IDA beads and flow through and/or wash containing purified protein was collected. Protein concentration was checked afterwards via SDS-PAGE (Bio-Rad 4–20% Mini-PROTEAN TGX Stain-Free protein gel) and InstantBlue protein stain (Expedeon) or ReadyBlue (Sigma-Aldrich) and quantified using Bio-Rad ImageLab 6.0.1. When necessary, proteins were concentrated using a Pall Advanced Centrifugal Device or an Amicon Ultra Filter with a pore size around one-third of the respective protein. Before storage, 1 mM DTT and 1 mM EDTA was added to the protein. Proteins were snap-frozen in liquid nitrogen and stored at −80°C.
IbpA purification.
The purification protocol was the same as for others except for the buffers since IbpA is an aggregation prone protein and requires different purification buffers, as mentioned in [36]. LG-5 buffer containing 50 mM HEPES–KOH (pH 8.0), 4 M urea, 1 mM BME, 400 mM potassium glutamate, 5 mM imidazole was used for wash and 50 mM HEPES–KOH (pH 8.0), 3 M urea, 1 mM BME, 400 mM potassium glutamate, 500 mM imidazole, 10% glycerol was used for elution. The dialysis was done against buffer with 50 mM HEPES–KOH (pH 8.0), 2 M urea, 10% glycerol, 200 mM potassium glutamate, 1 mM BME and the storage buffer was 50 mM HEPES–KOH (pH 8.0), 800 mM potassium glutamate, 20% sucrose, 1 mM BME.
In vitro degradation assay
In vitro degradation assays were performed as described previously [20]. The reaction was carried out in Lon reaction buffer (25 mM Tris-HCl pH 8.0, 100 mM KCl, 10 mM MgCl2, 1 mM DTT) using 0.4 µM Lon6 hexamer, respective amounts of substrate as mentioned in the figure legends and an ATP regeneration system (4 mM ATP, 15 mM creatine phosphate, 75 µg/mL creatine kinase). The reaction mix and the ATP regeneration system were prepared separately and pre-warmed to 37°C. The reaction was started by adding the ATP regeneration system to the reaction mix. Samples were taken at indicated time points and quenched by 1 volume of 2 × Laemmli SDS loading buffer and snap frozen in liquid nitrogen. Samples were heated at 65°C for 10 min and separated by SDS-PAGE (Bio-Rad 4–20% Mini-PROTEAN TGX Stain-Free protein gel) visualized by InstantBlue protein stain (Expedeon) or ReadyBlue (Sigma-Aldrich) and quantified using Bio-Rad ImageLab 6.0.1. Substrate levels were normalized to the Lon levels of the respective time points.
Immunoblot analysis
For whole cell extract analysis, 1 mL culture samples were collected at the respective time points and cell pellets were obtained by centrifugation. Cell pellets were resuspended in 200 μL of 1 × SDS sample buffer per OD600 1.0, to ensure normalization of the samples. Samples were boiled at 98°C for 10 min and run on an SDS-PAGE using Mini-PROTEAN TGX Stain-Free gels (usually 4–20%, Bio‐Rad). The proteins were transferred to nitrocellulose membranes by a semi-dry blotting procedure as per manufacturer guidelines. The protein gels and membranes were imaged using a Gel Doc Imager before and after the transfer, respectively, to assess equal loading of total protein as well as the quality of the transfer.
Membranes were blocked for 1 h at room temperature in 5% skim milk powder in TBS-Tween (TBST) and protein levels were detected using the anti-Lon (1:10000 dilution; kindly provided by R.T. Sauer) or anti-FLAG M2 (1:5000, Sigma-Aldrich) primary antibodies in 3% skim milk powder in TBST. Secondary antibodies, 1:5000 dilutions of anti‐rabbit or anti-mouse (Thermo Fisher Scientific) and SuperSignal Femto West (Thermo Fisher Scientific) were used to detect primary antibody binding. Immunoblots were scanned using a Chemidoc (Bio‐Rad) system or a LI-COR Odyssey Fc system. Relative signal intensities were quantified using the Image Lab software package (Bio‐Rad).
Motility assays
Swimming motility assays were performed on plates with 0.3% agar added to media containing 1% tryptone and 1% NaCl. Twitching motility was assayed on LB plates containing 1% agar. When needed, 0.2% arabinose and 30 μg/ml of gentamicin was added. The plates were dried for 4h and inoculated with 1 μL of P. aeruginosa cultures that were diluted to OD600 = 0.1. For swimming, the culture was inoculated 3 mm into the vertical centre of the agar. For twitching, the culture was inoculated all the way to the bottom of the agar, between the plate and the agar. The swimming motility plates were incubated at room temperature (approximately 25°C) for 40 – 48 hours for various experiments. The twitching motility plates were incubated at room temperature (approximately 25°C) for five to eight days. The plates were imaged using a ChemiDoc under the setting: Blots – Colorimetric. The motility zone area was quantified using the application Fiji (ImageJ).
Plate reader-based growth curve measurements
Overnight cultures were grown and diluted to an OD600 of 0.05 in LB with the addition of gentamicin and arabinose when applicable. 200 μL of the dilution were added into sterile 96 well transparent plates. LB medium (with or without antibiotic and/ or inducer when applicable) were used as blank in multiple wells. The plate was set up in a Spark microplate reader (Tecan) at 37°C for an OD600 measurement every 10 min for 24 hours with shaking.
Biofilm assays
Overnight cultures were grown and diluted 100× in LB with the addition of gentamycin and arabinose when applicable. 200 μL of the dilution were added into sterile 96 well transparent plates. LB medium (with or without antibiotic and/ or inducer when applicable) were used as blank in multiple wells. The plate was incubated for 8 hours at 37°C and the OD600 was measured at the end of the incubation. The liquid culture was poured out and washed gently by immersing in water twice. All work from this point was carried out in a ventilated hood. After removing residual water, 200 μL of 0.1% crystal violet solution was added to each well and incubated at room temperature for 15 min. The crystal violet was discarded and the wells were washed again with water twice. The crystal violet staining the biofilms was then solubilised using 200 μL of 10% acetic acid and incubated for 15 min. The solution from each well was transferred to a fresh 96-well plate and the absorbance at 550 nm was measured using a SpectraMax i3x multi-mode microplate reader.
Pyocyanin production
Cultures were grown overnight and diluted to a starting OD600 of 0.05 and grown for 4–5 hours until stationary phase OD600 of about 3.0, when WT cultures develop a green colour. The cultures were spun down and the supernatant of each strain was extracted. 100 μL of the supernatant from each replicate of each strain was transferred to a 96-well plate in three or four technical replicates and an absorbance scan from 450 to 900 nm corresponding to pyocyanin secretion, was recorded on a Spark microplate reader (Tecan).
Phase contrast microscopy
The strains were grown to exponential phase and samples were collected. 1% final concentration of formaldehyde was used to fix the cells to be stored at 4°C. The cells were transferred to 1% agarose pads on glass slides and covered with cover slips for imaging. A Ti eclipse inverted research microscope (Nikon) with 100 × /1.45 numerical aperture (NA) objective (Nikon) was used to collect phase-contrast images. The images were processed and quantified using Fiji (ImageJ) and its MicrobeJ plugin.
Generation and analysis of motility suppressors
Swimming motility plates containing the lon- strain after incubation at room temperature for 48 hours was exposed to UV irradiation for 0.4 to 1 second using the ChemiDoc imager under the setting: DNA gels – UV light. The plates were incubated for further one to two days and imaged again. The emerging suppressors were extracted from the outer edge of the suppressor zone using a pipette and plated on LB agar plates. They were grown overnight at 37°C, grown in overnight cultures and tested on motility agar plates to check if they are true suppressors. Six of the suppressors that showed motility equal to or more than the WT strain were selected and their genomic DNA was extracted using the Promega Wizard Genomic DNA Purification kit. The DNA from WT, lon- and Sup 1–6 was sent to SeqCenter, Pennsylvania, USA. The Illumina reads and variant calling revealed all the mutations present compared to the P. aeruginosa PAO1 whole genome NC_002516.2.
Statistical analysis
All quantifications were made from arithmetic mean ± standard deviation of independent replicates. The value of replicate number, sample sizes and the statistical test used along with p-values are mentioned in the respective figure legends. The application Graphpad Prism Version 10.4.1 was used for the analyses and statistical tests.
Supporting information
S1 Fig. Lon levels and stability in the lon overexpression strain.
(A) Representative immunoblot of WT with empty vector (VC) and lon overexpression (WT Para lonOE) using an anti-Lon antibody before and 90 min after induction using 0.2% arabinose. Bottom – Quantification of Lon relative to the VC from the immunoblots based on at least three independent replicates. (B) Growth curves of WT in the presence of 0, 500, 800 and 1000 μg/mL spectinomycin grown at 37°C and measured at OD600. Means ± SD of three independent biological replicates plotted for each concentration. Spec – spectinomycin. (C) Representative immunoblot showing the change in Lon levels of the WT with empty vector (VC) and lon overexpression (WT Para lonOE) over time after addition of 1 mg/ml spectinomycin to shut off protein synthesis and representative stain-free gel showing the total proteome of the proteomics samples.
https://doi.org/10.1371/journal.ppat.1013288.s001
(EPS)
S2 Fig. In vitro degradation assays of β-casein, BioB and RlmN.
In vitro degradation assays testing Lon-dependent degradation of β-casein (A), BioB (B) and RlmN (C) in the presence and absence of ATP. Substrate concentrations used were 200 μg/mL for β-casein, 2.1 μM for BioB and RlmN. 0.4 μM of purified Lon6 hexamer was added in each reaction. Quantifications of each of the degradation assays have been done using Image Lab 6.1 and plotted using the arithmetic mean ± SD of the quantified values of remaining substrate normalized relative to corresponding Lon levels for each time point. Each graph represents quantification of at least three independent experiments. One representative SDS-PAGE gel of the assay is shown for each protein. CK – creatine kinase.
https://doi.org/10.1371/journal.ppat.1013288.s002
(EPS)
S3 Fig. Immunoblot showing Lon levels in wild type and lon- strains.
WT and lon- with empty vector (VC), lon overexpression (WT Para lon) and lon- complemented with lon (lon- Para lon) strains visualized using an anti-Lon antibody 0 and 90 min after induction with 0.2% arabinose.
https://doi.org/10.1371/journal.ppat.1013288.s003
(EPS)
S4 Fig. Growth and pyocyanin levels of motility suppressor strains.
(A) Growth curves of WT, lon-, Sup3, Sup5 and Sup6 grown at 37°C and measured at OD600. Means ± SD of three independent biological replicates plotted for each strain. (B) Pyocyanin produced by late exponential cultures grown at 37°C as visualized by the absorbance scan of supernatants of WT, lon-, Sup3, Sup5 and Sup6 strains from 500–900 nm. Means of three biological replicates plotted.
https://doi.org/10.1371/journal.ppat.1013288.s004
(EPS)
S5 Fig. Quantification of microscopy and motility of strains expressing sulA and its mutants.
(A) Violin plot shows quantification of the phase-contrast microscopy images of the strains shown, grown with arabinose induction at 37°C to logarithmic phase. A total of 900 cells for each strain from a total of at least three biological replicates were used. Statistical significance was calculated using one-way ANOVA (Dunnett’s multiple comparisons test) for pairs: WT VC vs WT Para F-sulA and WT VC vs WT Para F-sulAD109V, p < 0.0001, ****; WT VC vs WT Para F-sulAH127L, p = 0.7164, (not significant, ns); WT + VC vs WT Para F-sulAfs, p = 0.7439, (not significant, ns). (B) Quantification of swimming motility in Fig 5F. Means of motility zone area (cm2) ± SD of three independent biological replicates plotted. Statistical significance was calculated using one-way ANOVA (Dunnett’s multiple comparisons test) for comparison of pairs: WT VC vs WT Para F-sulA and WT VC vs WT Para F-sulAD109V, p < 0.0001, ****; WT VC vs WT Para F-sulAH127L, p = 0.0006, ***. (C) Swimming motility in soft agar without arabinose after 40–48 hours at 25°C for the strains shown. Means of motility zone area (cm2) ± SD of three independent biological replicates are plotted. Statistical significance was calculated using one-way ANOVA (Dunnett’s multiple comparisons test) for comparison of pairs: WT VC vs WT Para F-sulA, p = 0.0004, ***; WT VC vs WT Para F-sulAH127L, p = 0.7508, (not significant, ns); WT VC vs WT Para F-sulAD109V, p = 0.4674, (not significant, ns). Right – Representative image of the swimming agar plate without arabinose.
https://doi.org/10.1371/journal.ppat.1013288.s005
(EPS)
S6 Fig. Growth in piperacillin at different concentrations.
Growth curves of WT in the presence of 0, 0.25, 0.5 and 1 μg/mL piperacillin grown at 37°C and measured at OD600. Means ± SD of three independent biological replicates are plotted for each condition.
https://doi.org/10.1371/journal.ppat.1013288.s006
(EPS)
S1 Data.
Quantitative proteomics (lonOE) source data, related to Fig 1.
https://doi.org/10.1371/journal.ppat.1013288.s010
(XLSX)
S2 Data.
Quantitative proteomics (lon-) source data, related to Fig 3.
https://doi.org/10.1371/journal.ppat.1013288.s011
(XLSX)
Acknowledgments
We thank members of the Jonas group for helpful discussions and feedback on the manuscript, the Hancock lab (University of British Columbia) for providing strains H103 and H1105 and the Clinical Proteomics Mass Spectrometry Core Facility at KI/KS for excellent service, support and advice.
References
- 1. Gur E, Biran D, Ron EZ. Regulated proteolysis in Gram-negative bacteria--how and when?. Nat Rev Microbiol. 2011;9(12):839–48. pmid:22020261
- 2. Sauer RT, Baker TA. AAA+ proteases: ATP-fueled machines of protein destruction. Annu Rev Biochem. 2011;80:587–612. pmid:21469952
- 3. Gur E, Sauer RT. Degrons in protein substrates program the speed and operating efficiency of the AAA+ Lon proteolytic machine. Proc Natl Acad Sci U S A. 2009;106(44):18503–8. pmid:19841274
- 4. Burgos R, Weber M, Martinez S, Lluch-Senar M, Serrano L. Protein quality control and regulated proteolysis in the genome-reduced organism Mycoplasma pneumoniae. Mol Syst Biol. 2020;16(12):e9530. pmid:33320415
- 5. Joshi KK, Chien P. Regulated Proteolysis in Bacteria: Caulobacter. Annu Rev Genet. 2016;50:423–45. pmid:27893963
- 6. Baker TA, Sauer RT. ATP-dependent proteases of bacteria: recognition logic and operating principles. Trends Biochem Sci. 2006;31(12):647–53. pmid:17074491
- 7. Liu J, Francis LI, Jonas K, Laub MT, Chien P. ClpAP is an auxiliary protease for DnaA degradation in Caulobacter crescentus. Mol Microbiol. 2016;102(6):1075–85. pmid:27667502
- 8. Gupta M, Johnson ANT, Cruz ER, Costa EJ, Guest RL, Li SH-J, et al. Global protein turnover quantification in Escherichia coli reveals cytoplasmic recycling under nitrogen limitation. Nat Commun. 2024;15(1):5890. pmid:39003262
- 9. Basta DW, Angeles-Albores D, Spero MA, Ciemniecki JA, Newman DK. Heat-shock proteases promote survival of Pseudomonas aeruginosa during growth arrest. Proc Natl Acad Sci U S A. 2020;117(8):4358–67. pmid:32029587
- 10. Gur E. The Lon AAA+ protease. Subcell Biochem. 2013;66:35–51. pmid:23479436
- 11. Charette MF, Henderson GW, Markovitz A. ATP hydrolysis-dependent protease activity of the lon (capR) protein of Escherichia coli K-12. Proc Natl Acad Sci U S A. 1981;78(8):4728–32. pmid:6458036
- 12. Chung CH, Goldberg AL. The product of the lon (capR) gene in Escherichia coli is the ATP-dependent protease, protease La. Proc Natl Acad Sci U S A. 1981;78(8):4931–5. pmid:6458037
- 13. Van Melderen L, Aertsen A. Regulation and quality control by Lon-dependent proteolysis. Res Microbiol. 2009;160(9):645–51. pmid:19772918
- 14. Akar R, Fink MJ, Omnus DJ, Jonas K. Regulation of the general stress response sigma factor σT by Lon-mediated proteolysis. J Bacteriol. 2023;205(11):e0022823. pmid:37930077
- 15. Arends J, Griego M, Thomanek N, Lindemann C, Kutscher B, Meyer HE, et al. An Integrated Proteomic Approach Uncovers Novel Substrates and Functions of the Lon Protease in Escherichia coli. Proteomics. 2018;18(13):e1800080. pmid:29710379
- 16. Christensen SK, Mikkelsen M, Pedersen K, Gerdes K. RelE, a global inhibitor of translation, is activated during nutritional stress. Proc Natl Acad Sci U S A. 2001;98(25):14328–33. pmid:11717402
- 17. Jonas K, Liu J, Chien P, Laub MT. Proteotoxic stress induces a cell-cycle arrest by stimulating Lon to degrade the replication initiator DnaA. Cell. 2013;154(3):623–36. pmid:23911325
- 18. Mizusawa S, Gottesman S. Protein degradation in Escherichia coli: the lon gene controls the stability of sulA protein. Proc Natl Acad Sci U S A. 1983;80(2):358–62. pmid:6300834
- 19. Mukherjee S, Bree AC, Liu J, Patrick JE, Chien P, Kearns DB. Adaptor-mediated Lon proteolysis restricts Bacillus subtilis hyperflagellation. Proc Natl Acad Sci U S A. 2015;112(1):250–5. pmid:25538299
- 20. Omnus DJ, Fink MJ, Szwedo K, Jonas K. The Lon protease temporally restricts polar cell differentiation events during the Caulobacter cell cycle. Elife. 2021;10:e73875. pmid:34693909
- 21. Brazas MD, Breidenstein EBM, Overhage J, Hancock REW. Role of lon, an ATP-dependent protease homolog, in resistance of Pseudomonas aeruginosa to ciprofloxacin. Antimicrob Agents Chemother. 2007;51(12):4276–83. pmid:17893152
- 22. Breidenstein EBM, Janot L, Strehmel J, Fernandez L, Taylor PK, Kukavica-Ibrulj I, et al. The Lon protease is essential for full virulence in Pseudomonas aeruginosa. PLoS One. 2012;7(11):e49123. pmid:23145092
- 23. Breidenstein EBM, Bains M, Hancock REW. Involvement of the lon protease in the SOS response triggered by ciprofloxacin in Pseudomonas aeruginosa PAO1. Antimicrob Agents Chemother. 2012;56(6):2879–87. pmid:22450976
- 24. Hall BM, Breidenstein EBM, de la Fuente-Núñez C, Reffuveille F, Mawla GD, Hancock REW, et al. Two Isoforms of Clp Peptidase in Pseudomonas aeruginosa Control Distinct Aspects of Cellular Physiology. J Bacteriol. 2017;199(3):e00568-16. pmid:27849175
- 25. Kamal SM, Rybtke ML, Nimtz M, Sperlein S, Giske C, Trček J, et al. Two FtsH Proteases Contribute to Fitness and Adaptation of Pseudomonas aeruginosa Clone C Strains. Front Microbiol. 2019;10:1372. pmid:31338071
- 26. Marr AK, Overhage J, Bains M, Hancock REW. The Lon protease of Pseudomonas aeruginosa is induced by aminoglycosides and is involved in biofilm formation and motility. Microbiology (Reading). 2007;153(Pt 2):474–82. pmid:17259618
- 27. Gellatly SL, Hancock REW. Pseudomonas aeruginosa: new insights into pathogenesis and host defenses. Pathog Dis. 2013;67(3):159–73. pmid:23620179
- 28. Rice LB. Federal funding for the study of antimicrobial resistance in nosocomial pathogens: no ESKAPE. J Infect Dis. 2008;197(8):1079–81. pmid:18419525
- 29.
WHO Bacterial Priority Pathogens List 2024: Bacterial Pathogens of Public Health Importance, to Guide Research, Development, and Strategies to Prevent and Control Antimicrobial Resistance. 1st ed. Geneva: World Health Organization. 2024.
- 30. Laventie B-J, Jenal U. Surface Sensing and Adaptation in Bacteria. Annu Rev Microbiol. 2020;74:735–60. pmid:32905753
- 31. Kazmierczak BI, Schniederberend M, Jain R. Cross-regulation of Pseudomonas motility systems: the intimate relationship between flagella, pili and virulence. Curr Opin Microbiol. 2015;28:78–82. pmid:26476804
- 32. Fernández L, Breidenstein EBM, Taylor PK, Bains M, de la Fuente-Núñez C, Fang Y, et al. Interconnection of post-transcriptional regulation: The RNA-binding protein Hfq is a novel target of the Lon protease in Pseudomonas aeruginosa. Sci Rep. 2016;6:26811. pmid:27229357
- 33. Takaya A, Tabuchi F, Tsuchiya H, Isogai E, Yamamoto T. Negative regulation of quorum-sensing systems in Pseudomonas aeruginosa by ATP-dependent Lon protease. J Bacteriol. 2008;190(12):4181–8. pmid:18408026
- 34. Barembruch C, Hengge R. Cellular levels and activity of the flagellar sigma factor FliA of Escherichia coli are controlled by FlgM-modulated proteolysis. Mol Microbiol. 2007;65(1):76–89. pmid:17537210
- 35. Barros BB, Mahmoud SA, Chien P, Zeinert RD. Degradation of Lon in Caulobacter crescentus. J Bacteriol. 2020;203(1):e00344-20. pmid:33020222
- 36. Bissonnette SA, Rivera-Rivera I, Sauer RT, Baker TA. The IbpA and IbpB small heat-shock proteins are substrates of the AAA+ Lon protease. Mol Microbiol. 2010;75(6):1539–49. pmid:20158612
- 37. Anton DL, Kutny R. Escherichia coli S-adenosylmethionine decarboxylase. Subunit structure, reductive amination, and NH2-terminal sequences. J Biol Chem. 1987;262(6):2817–22. pmid:3546296
- 38. Bolard A, Schniederjans M, Haüssler S, Triponney P, Valot B, Plésiat P, et al. Production of norspermidine contributes to aminoglycoside resistance in pmrAB mutants of Pseudomonas aeruginosa. Antimicrob Agents Chemother. 2019;63(10):e01044-19. pmid:31383668
- 39. Lewenza S, Falsafi RK, Winsor G, Gooderham WJ, McPhee JB, Brinkman FSL, et al. Construction of a mini-Tn5-luxCDABE mutant library in Pseudomonas aeruginosa PAO1: a tool for identifying differentially regulated genes. Genome Res. 2005;15(4):583–9. pmid:15805499
- 40. Gottesman S, Halpern E, Trisler P. Role of sulA and sulB in filamentation by lon mutants of Escherichia coli K-12. J Bacteriol. 1981;148(1):265–73. pmid:7026534
- 41. Iida K, Hirata S, Nakamuta S, Koike M. Inhibition of cell division of Escherichia coli by a new synthetic penicillin, piperacillin. Antimicrob Agents Chemother. 1978;14(2):257–66. pmid:358919
- 42. Pogliano J, Pogliano K, Weiss DS, Losick R, Beckwith J. Inactivation of FtsI inhibits constriction of the FtsZ cytokinetic ring and delays the assembly of FtsZ rings at potential division sites. Proc Natl Acad Sci U S A. 1997;94(2):559–64. pmid:9012823
- 43. Glen KA, Lamont IL. Penicillin-binding protein 3 sequence variations reduce susceptibility of Pseudomonas aeruginosa to β-lactams but inhibit cell division. J Antimicrob Chemother. 2024;79(9):2170–8. pmid:39001778
- 44. Darzins A. Characterization of a Pseudomonas aeruginosa gene cluster involved in pilus biosynthesis and twitching motility: sequence similarity to the chemotaxis proteins of enterics and the gliding bacterium Myxococcus xanthus. Mol Microbiol. 1994;11(1):137–53. pmid:7908398
- 45. Luo Y, Zhao K, Baker AE, Kuchma SL, Coggan KA, Wolfgang MC, et al. A hierarchical cascade of second messengers regulates Pseudomonas aeruginosa surface behaviors. mBio. 2015;6(1):e02456-14. pmid:25626906
- 46. Zhou X, Wang J, Herrmann J, Moerner WE, Shapiro L. Asymmetric division yields progeny cells with distinct modes of regulating cell cycle-dependent chromosome methylation. Proc Natl Acad Sci U S A. 2019;116(31):15661–70. pmid:31315982
- 47. Zeinert RD, Baniasadi H, Tu BP, Chien P. The Lon Protease Links Nucleotide Metabolism with Proteotoxic Stress. Molecular Cell. 2020;79(5):758-767.e6.
- 48. Cremer J, Honda T, Tang Y, Wong-Ng J, Vergassola M, Hwa T. Chemotaxis as a navigation strategy to boost range expansion. Nature. 2019;575(7784):658–63. pmid:31695195
- 49. Guttenplan SB, Kearns DB. Regulation of flagellar motility during biofilm formation. FEMS Microbiol Rev. 2013;37(6):849–71. pmid:23480406
- 50. Gottesman S. Proteases and their targets in Escherichia coli. Annu Rev Genet. 1996;30:465–506. pmid:8982462
- 51. ISHII Y, AMANO F. Regulation of SulA cleavage by Lon protease by the C-terminal amino acid of SulA, histidine. Biochemical Journal. 2001;358(2):473–80.
- 52. Cordell SC, Robinson EJH, Lowe J. Crystal structure of the SOS cell division inhibitor SulA and in complex with FtsZ. Proc Natl Acad Sci U S A. 2003;100(13):7889–94. pmid:12808143
- 53. Justice SS, Hunstad DA, Seed PC, Hultgren SJ. Filamentation by Escherichia coli subverts innate defenses during urinary tract infection. Proc Natl Acad Sci U S A. 2006;103(52):19884–9. pmid:17172451
- 54. Bos J, Zhang Q, Vyawahare S, Rogers E, Rosenberg SM, Austin RH. Emergence of antibiotic resistance from multinucleated bacterial filaments. Proc Natl Acad Sci U S A. 2015;112(1):178–83. pmid:25492931
- 55. Hulsen T. DeepVenn -- a web application for the creation of area-proportional Venn diagrams using the deep learning framework Tensorflow.js. arXiv. 2022.
- 56. Thomas PD, Ebert D, Muruganujan A, Mushayahama T, Albou L-P, Mi H. PANTHER: Making genome-scale phylogenetics accessible to all. Protein Sci. 2022;31(1):8–22. pmid:34717010
- 57. Winsor GL, Griffiths EJ, Lo R, Dhillon BK, Shay JA, Brinkman FSL. Enhanced annotations and features for comparing thousands of Pseudomonas genomes in the Pseudomonas genome database. Nucleic Acids Res. 2016;44(D1):D646-53. pmid:26578582
- 58. Holmberg MA, Gowda NKC, Andréasson C. A versatile bacterial expression vector designed for single-step cloning of multiple DNA fragments using homologous recombination. Protein Expr Purif. 2014;98:38–45. pmid:24631626