Strains and plasmids used in this study are listed in the Supplementary Tables 1, 2. All E. coli strains used in the reported experiments are derivatives of MG1655 (Guyer et al., 1981). Bacterial cells were grown in lysogeny broth (LB) [1% tryptone, 0.5% yeast extract, 0.5% NaCl] or minimal M9 medium supplemented with 0.2% casamino acids and 0.2% glucose. Antibiotics were used at 25 μg/ml (chloramphenicol; Cm), 37.5 μg/ml (kanamycin; Kan) or 5 μg/ml (tetracycline; tet).

Gene deletion mutants were constructed to resemble those in the Keio knockout collection (Baba et al., 2006). The Kan^R cassette was amplified by using pKD13 as a template with sequences homologous to the 5′ and 3′ ends of the target genes in the chromosome. The sequences of the primer pairs used for gene deletion are listed in Supplementary Table 3. The amplified DNA was electroporated into MG1655 harboring either pKD46 or pBO51 that expresses the lambda Red genes from an arabinose-inducible promoter (pKD46) or high temperature-inducible promoter (pBO51), and recombinants were selected on LB agar supplemented with 37.5 μg/ml kanamycin (Johnson et al., 2004).

Strains with multiple deletion mutations were made by sequential introduction of each deletion via P1 transduction followed by removal of the Kan^R cassette using FLP recombinase expressed from pCP20, leaving a frt scar sequence at each deletion locus (Datsenko and Wanner, 2000). The correct orientation of the DNA flanking frt sequences in multiple deletion mutants was confirmed for all deletions in each mutant.

MG1655 was mutagenized with the Ez-Tn5 < Kan-2 > transposome (Epicenter) as previously described (Bernhardt and Boer, 2004). Mutants were selected on agar for kanamycin resistance at 30°C, yielding libraries consisting of approximately 4 × 10⁵ independent transposon insertions. The mutant library was harvested by scraping colonies from the agar surface and suspending them in LB broth. The suspension of the mutant library was mixed with glycerol to a final concentration of 15%, aliquoted, and kept frozen at −80°C.

A transposon mutant library of MG1655 was thawed and incubated in LB for roughly two doublings. The resuscitated mutant library culture was diluted to an OD₆₀₀ of 0.04 in 25 mL LB lacking or supplemented with 0.1 μg/ml aztreonam or 2.5 μg/ml mecillinam and incubated for 2 h at 30°C with agitation (250 rpm). After incubation, the cells of each culture were pelleted, washed with 1X PBS, and resuspended in 0.5 mL DNase I solution [0.25 mg DNase I (Roche) in 10 mM Tris–HCl, pH 7.5, 2.5 mM MgCl₂, 0.1 mM CaCl₂]. The cell suspensions in DNase I solution were incubated for 10 min at 37°C to remove extracellular DNA released from lysed cells. DNase I-treated cells were pelleted, resuspended in 1X PBS supplemented with 10 mM EDTA, and incubated for 10 min at 80°C to inactivate DNase I. The cells were pelleted and kept frozen at −80°C until preparation of genomic DNA from the unlysed cells.

Genomic DNA from each cell pellet was extracted, fragmented, and poly-C tailed as previously described (Lai et al., 2017). The transposon-chromosome junctions in the resulting DNA were amplified by using Easy-A Hi-Fi Cloning System (Agilent Technologies). The primers used were the poly-C tail-specific primer 5′- GTGACTGGAGTTCAGACGTGTGCTCTTCCGAT CTGGGGGGGGGGGGGGGG-3′ and the transposon-specific primer 5′-ACCTGCAGGCATGCAAGCTTCAGGG-3′. The transposon-chromosome junctions were further amplified in a second nested PCR with primers that add sequencing barcodes to each mutant library, NEBNext Multiplex Oligos for Illumina (NEB) and the transposon-specific primer 5′- AATGATACGGCGACCACCGAGATCTACACTCTTTTCAGG GTTGAGATGTGTATAAGAGA-3′. The final PCR products were quantified and equal amounts of each barcoded library were mixed. The pooled sequencing library was run on a 2% agarose gel, and DNA fragments ranging from 200 and 500 bp were excised and purified using QIAquick Gel Extraction Kit (Qiagen). The resulting library was sequenced using a MiSeq reagent kit V3 (150-cycle) (Illumina) with the custom primer 5′-ACACTCTTTTCAGGGTTGAGATGTGTATAAGAGACAG-3′. Sequencing reads were trimmed using trimmomatic (Bolger et al., 2014) to remove adaptor sequences, and mapped to the E. coli MG1655 genome (NC_000913) using bowtie 1.0.0 (Langmead et al., 2009). Differences in the total number of reads at any given transposon insertion site between untreated and beta lactam-treated samples were determined using a Mann–Whitney U test. Transposon insertion profiles were visualized using the Sanger Artemis Genome Browser and Annotation tool.

Cultures of MG1655 and SP10 (Δtol-pal) strains were grown to an exponential phase (OD₆₀₀ between 0.4 and 0.5) in LB at 30°C. Aztreonam or mecillinam was added to 15 mL cultures at the final concentrations of 0, 0.1, 1, or 10 μg/ml and incubated for 5 min at 30°C. Cells were then collected by centrifugation at 4°C, washed with ice-cold 1X PBS twice, resuspended in 500 μl 1X PBS supplemented with 10 mM EDTA and 15 μM Bocillin (Invitrogen), and incubated at room temperature for 15 min. After incubation, the cell suspensions were washed with 1X PBS once, resuspended in 500 μl 1X PBS, and disrupted by sonication. After a brief spin for 1 min at 4 Krcf to remove undisrupted cells, membrane fractions were pelleted by ultracentrifugation at 200 Krcf for 20 min at 4°C. The membrane fractions were then washed with 1X PBS and resuspended in 50 μl 1X PBS. Resuspended samples were mixed with 50 μl 2X Laemmli sample buffer and boiled for 10 min at 95°C. After measuring the total protein concentrations of each sample with the NI-protein assay (G-Biosciences), 25 μg of total protein for each sample was then separated on a 4–20% gradient SDS–PAGE gel and the labeled proteins were visualized using a Typhoon 9400 fluorescence imager (GE Healthcare) with excitation at 488 nm and emission at 520 ñm.

Growth conditions prior to microscopy are described in the figure legends. Prior to imaging, cells were immobilized on 2% agarose pads containing 1X M9 salts and covered with #1.5 coverslips. Micrographs were obtained using a Leica DM2500 LED microscope equipped with a Leica DFC7000 GT camera, Fluo Illuminator LRF 4/22, HC PL APO 100x/1.40 Oil Ph3 objective lens, and Leica Las X acquisition software. Images in the green channel were obtained using an L5 filter cube.

In a transposon screen to identify factors important for maintaining cell integrity upon beta lactam treatment, we used mecillinam and aztreonam that bind specifically to PBP2 of the Rod system and PBP3 of the divisome, respectively (Curtis et al., 1979; Kocaoglu and Carlson, 2015). These drugs were chosen because phenotypic analyses would be easier with beta-lactams that mainly target a single PBP. We also reasoned that it would be advantageous for identification of mutations that accelerate lysis because bPBP-specific beta-lactams result in slower lysis than drugs that bind to multiple PBPs (Satta et al., 1995). For the actual screen, wild-type E. coli MG1655 cells were mutagenized with the EZTn5-Kan2 transposome to generate a mutant library consisting of approximately 4 × 10⁵ independent insertions. The mutant library was grown in LB to the exponential phase (OD₆₀₀ = 0.4), diluted in LB lacking or containing beta-lactams, and incubated for 2 hrs. After drug treatment, genomic DNA was isolated from the unlysed cells of each sample and the transposon insertion sites were mapped by high-throughput sequencing (Figure 1A). Comparison of the transposon insertion profiles between each sample revealed that transposon insertion in several genes significantly decreased in the sample collected after drug treatment relative to the sample prepared without drug treatment (Figure 1B). The decrease in transposon insertion densities indicated that the mutants bearing the transposon insertion lysed much faster than other mutants in the library upon beta-lactam treatment.

The genes that showed a clear reduction in transposon insertion density after beta-lactam treatment were ponB, lpoB, tolQ-pal, mepS, and slt (Figure 1B). Among these genes, ponB, lpoB, and tolQ-pal showed reduction regardless of the drug used. On the other hand, mepS and slt showed beta lactam-specific reduction: mepS upon mecillinam treatment and slt upon aztreonam treatment. To validate the hits identified from the transposon insertion site sequencing (Tn-seq) experiment, we made null mutants of the identified genes and compared the lysis kinetics and shape change of the mutant strains with the wild-type strain after drug treatment (Figures 1C–H, 2A-B and Supplementary Figure 1). Each mutant showed lysis phenotypes consistent with the Tn-seq results. The ΔponB,ΔlpoB, ΔtolQ, ΔtolR, ΔtolA, ΔtolB, and Δpal strains lysed faster than the wild-type strain upon either mecillinam or aztreonam treatment. The ΔmepS strain lysed faster upon mecillinam treatment, but not much upon aztreonam treatment. In contrast, the Δslt strain showed accelerated lysis upon aztreonam treatment, but not upon mecillinam treatment.

While transposon insertions in ponB and lpoB genes encoding PBP1b and LpoB decreased dramatically upon treatment of mecillinam or aztreonam, transposon insertions in ponA and lpoA genes encoding another aPBP-Lpo factor pair did not decrease noticeably upon treatment of either beta lactam drug (Supplementary Figure 2A). Although E. coli cells need only one of these aPBP-Lpo systems for survival under normal laboratory growth conditions, it has been observed that inactivation of the PBP1b-LpoB system causes a severe defect in maintaining cell integrity under various cell wall-perturbing conditions (del Portillo and de Pedro, 1990, 1991; Paradis-Bleau et al., 2010; Typas et al., 2010; Vigouroux et al., 2020). In addition, an elevated lysis phenotype of the ponB and lpoB mutants under normal growth conditions was also identified in a genetic screen with CPRG (Paradis-Bleau et al., 2014). In contrast, the ponA and lpoA mutants displayed only minor phenotypes except for the synthetic lethality with ponB-lpoB mutations (Yousif et al., 1985; Mueller et al., 2019; Vigouroux et al., 2020). Consistent with the previous reports and our Tn-seq results, the ΔponA and ΔlpoA strains did not show accelerated lysis upon treatment of mecillinam or aztreonam when we examined the lysis kinetics of these strains upon treatment of either beta lactam (Supplementary Figures 2B,C). These results indicated that PBP1b-LpoB pair plays a major role in maintenance of cell wall integrity upon inhibition of TP activity of the Rod system or the divisome, but PBP1A-LpoA function is dispensable under the same circumstances.

Accelerated lysis of the tol-pal mutant upon treatment of aztreonam or mecillinam suggested that the Tol-Pal system plays an important role in maintaining the integrity of the cell wall. However, because the tol-pal mutant has a permeabilized OM (Cascales et al., 2000; Lloubès et al., 2001), it also seemed possible that the OM permeability defect might be responsible for the accelerated lysis of the tol-pal mutant upon beta-lactam treatment. Beta-lactams might accumulate to a higher concentration in the periplasm of the tol-pal mutant than the wild-type strain due to the OM permeability defect, which might in turn cause non-specific inhibition of other PBPs such as aPBPs that are important for maintaining cell integrity. Thus, we examined if the OM permeability defect of the Δtol-pal strain affected the PBP-binding profile of aztreonam and mecillinam by using a Bocillin assay. A noticeable difference was not observed between the wild type and Δtol-pal strains across a broad range of concentrations (Figures 3A,B), indicating that the permeability defect of the Δtol-pal strain does not significantly affect the PBP-binding specificity of mecillinam and aztreonam. Furthermore, the Δtol-pal strain lysed faster than the wild-type strain even when treated with 100-fold less amounts of aztreonam or mecillinam than the wild-type strain (Figures 3C,D). These results suggested that the permeability defect is not likely to be responsible for the accelerated lysis of the Δtol-pal strain upon treatment of aztreonam or mecillinam, and the Tol-Pal system likely plays a more direct role in maintaining cell integrity.

The accelerated lysis of the Δtol-pal strain upon mecillinam treatment was especially interesting because the Tol-Pal system localizes at the division sites and most of the reported phenotypes of the tol-pal mutant are related to the division process (Gerding et al., 2007; Egan, 2018; Szczepaniak et al., 2020b). Upon mecillinam treatment, wild-type E. coli forms spherical cells that become larger and eventually lyse (Spratt, 1975). However, mecillinam-treated tol-pal mutant cells lysed early before forming spherical cells, indicating that the tol-pal mutant has a severe defect in maintaining cell integrity upon inhibition of the Rod system (Figure 2B). Because most known functions of the Tol-Pal system are related to cell division, we wondered if the tol-pal mutant somehow lysed at the division sites upon mecillinam treatment even though mecillinam targets the Rod system responsible for cell elongation.

To examine this possibility, we tested if inhibition of divisome assembly suppresses the accelerated lysis of the tol-pal mutant upon treatment of beta lactams. Assembly of the divisome was inhibited by overexpressing the FtsZ antagonist SulA, and beta lactams were added after confirming the inhibition of divisome assembly by visualizing the Z-ring with the zapA-gfp fusion expressed from the native zapA genomic locus. The effect of division inhibition was tested in M9 minimal medium because we found that sulA overexpression somehow enhanced the lysis of mecillinam-treated wild type cells when the cultures were grown in LB. As expected, accelerated lysis of the tol-pal mutant upon aztreonam treatment was almost completely suppressed by inhibition of division, consistent with the idea that lysis upon aztreonam treatment occurs at the division sites (Figure 4). However, the accelerated lysis of the tol-pal mutant upon mecillinam treatment was not suppressed by sulA overexpression, suggesting that the cell integrity defect of the tol-pal mutant is not restricted to the division sites (Figure 4).

Ineffectiveness of divisome inhibition in suppressing the accelerated lysis of the tol-pal mutant upon mecillinam treatment suggested that the Tol-Pal system plays a role in maintaining cell integrity not only in the dividing cells but also in the elongating cells. Thus, we examined another lysis phenotype related to the tol-pal mutation to gauge if the Tol-Pal system is involved in maintaining cell integrity at locations other than the division sites. Mutation of the tol-pal genes has been shown to cause a synthetic lysis phenotype with inactivation of the PBP1b-LpoB system (Typas et al., 2010; Markovski et al., 2016). It was assumed that this synthetic lysis would occur mostly at the division sites because the Tol-Pal system was known to function mainly at the division sites and the PBP1b-LpoB system activity was also thought to be important for maintaining cell integrity at the division sites (Typas et al., 2010). We reasoned that this synthetic lysis would also occur during cell elongation if the Tol-Pal system plays an important role in maintaining cell integrity during elongation. If so, the synthetic lysis caused by inactivation of both the Tol-Pal system and the PBP1b-LpoB system would not be efficiently suppressed by division inhibition.

To examine the effect of division inhibition on the synthetic lysis phenotype, we constructed the strain SP200 [P_ara:ponB, Δtol-pal, zapA-gfp] in which the tol-pal locus is deleted and PBP1b can be depleted by growth in medium lacking arabinose. SP200 lysed upon depletion of PBP1b when it was grown in no salt LB. This synthetic lysis would be efficiently suppressed by division inhibition if the Tol-Pal system is mainly involved in maintaining cell integrity at the division sites. On the other hand, the synthetic lysis would only be partially suppressed by division inhibition if Tol-Pal system activity is also required for cell elongation. The synthetic lysis was not suppressed efficiently by sulA overexpression, indicating that the synthetic lysis occurs not only at the division sites but also during elongation (Figure 5).

The synthetic phenotype between inactivation of the Tol-Pal and PBP1b-LpoB systems prompted us to test if inactivation of the PBP1a-LpoA system also causes a synthetic phenotype with the tol-pal mutation. However, unlike inactivation of the PBP1b-LpoB system, mutation of the PBP1a-LpoA system in the tol-pal mutant did not cause an obvious defect in maintaining cell integrity in normal growth conditions (Figure 6A). This result was not surprising because mutation of the PBP1a-LpoA system does not usually cause an obvious defect in E. coli except for the synthetic lethality with the inactivation of the PBP1b-LpoB system (Supplementary Figure 2; Mueller et al., 2019; Vigouroux et al., 2020). Next, we tested if inactivation of the PBP1a-LpoA system affects the lysis phenotype of the Δtol-pal strain upon beta lactam treatment. Interestingly, lysis of the tol-pal mutant upon treatment of aztreonam or mecillinam was further accelerated by mutation of the PBP1a-LpoA system, indicating that the synthetic phenotype with the tol-pal mutation is not specific to inactivation of the PBP1b-LpoB system, but can also be observed with inactivation of the PBP1a-LpoA system albeit at a small extent (Figures 6B,C).

The above results suggested that the Tol-Pal system is required for maintaining cell integrity during elongation. Because all other genes identified as required for maintaining cell integrity in our Tn-seq screen are either PG synthetic or remodeling enzymes, we suspected that the Tol-Pal system would also play a role in PG assembly during cell elongation. The similarity of the lysis phenotype of the tol-pal and ponB/lpoB mutants upon beta-lactam treatment suggested that the functions of the two systems might be related. We hypothesized that the Tol-Pal system is required for the optimal activity of aPBP-Lpo systems during elongation. If so, the accelerated lysis of the tol-pal mutant upon beta-lactam treatment might be caused by inefficient activity of the aPBP-Lpo systems, especially the PBP1b-LpoB system. In addition, the synthetic phenotypes observed upon simultaneous inactivation of an aPBP-Lpo system and the Tol-Pal system might occur because the remaining aPBP-Lpo system does not function optimally in the absence of the Tol-Pal system. To examine our hypothesis, we tested if the accelerated lysis of the tol-pal mutant upon beta-lactam treatment is suppressed by enhancing aPBP-Lpo system activity. At first, we tested the effect of increasing expression of aPBPs and their activators. However, overproduction of aPBPs along with their cognate Lpo factors did not suppress the accelerated lysis of the tol-pal mutant (Supplementary Figure 3). We reasoned that simple overproduction of aPBPs and Lpo factors might not suppress the accelerated lysis of the tol-pal mutant if the Tol-Pal system is required for the efficient activation of aPBPs by Lpo factors. If so, expression of aPBPs that do not require activation by Lpo factors might suppress the Tol-Pal defect. Thus, we examined the effects of enhancing aPBP activity by expressing Lpo factor-bypass aPBPs that do not require Lpo factors for their PG synthetic activity (Markovski et al., 2016; Sardis et al., 2021). To our amazement, expression of ponB[E313D] encoding an LpoB-bypass PBP1b protein significantly suppressed the accelerated lysis of the tol-pal mutant as well as that of the lpoB mutant upon mecillinam treatment, while expression of wild-type ponB did not suppress the Tol-Pal defect (Figure 7 and Supplementary Figure 4). Importantly, suppression of the Tol-Pal defect by ponB[E313D] expression did not appear to be due to PBP1b[E313D] being more active than wild-type PBP1b. The ponB[E313D] allele was less efficient than wild-type ponB in suppressing the accelerated lysis of the ΔponB strain upon beta lactam treatment (Supplementary Figure 5). These results suggested that the defect of the tol-pal mutant in maintaining cell integrity during elongation might be at least partially due to inefficient aPBP activity, and the Tol-Pal system might be involved in facilitating the activation of PBP1b by LpoB in the elongating cells.