Peptidoglycan maturation controls outer membrane protein assembly

Gideon Mamou, Federico Corona, Ruth Cohen-Khait, Nicholas G. Housden, Vivian Yeung, Dawei Sun, Pooja Sridhar, Manuel Pazos, Timothy J. Knowles, Colin Kleanthous & Waldemar Vollmer | Nature volume 606, pages 953–959 (2022)

This wiki article is a review of Mamou et al.’s 2022 paper, “Peptidoglycan maturation controls outer membrane protein assembly”, which can be reached by this link.

Gram-negative bacteria are characterized as bacteria containing two cell membranes enclosing a thin layer of peptidoglycan (PG), the primary component of the bacterial cell wall (Mamou et al., 2022). These barriers help Gram-negative bacteria survive harsh environments, but present a problem for growth. By having an outer membrane (OM), Gram-negative bacteria require a mechanism to coordinate OM growth with the underlying cell wall. For a long time, this coordination was poorly understood (Deghelt & Collet, 2022). However, recent evidence has pointed to a protein called BamA as the link.

BamA is a periplasmic protein that helps the OM function by inserting outer membrane proteins (OMPs) into it (Mamou et al., 2022). Previous research showed that BamA distributed uniformly around the periplasm, but data conflicted on where it preferentially inserted OMPs (Gunasinghe et al., 2018). However, some findings showed that OMP biogenesis may be concentrated at the mid-cell region (Rassam et al., 2015). The importance of this is that PG structure differs based on location: in the mid-cell region, PG is nascent and composed of pentapeptides, while at the cell poles, PG is mature and composed of tetrapeptides (Mamou et al., 2022). Thus, it was possible that differences in PG maturation controlled OMP biogenesis via BamA activity.

Based on this knowledge, it was hypothesized that BamA represents the coordinating link between the cell wall and the outer membrane (Mamou et al., 2022). The researchers conducted their experiments using biochemical and genetic techniques, as well as in vivo and in vitro models.The researchers confirmed that BamA is uniformly distributed throughout the periplasm of live E. coli cells during different stages of their development. However, BamA is mainly active and functional at cell division and growth sites, such as the mid cell region. As a result, OMP biogenesis is also concentrated at cell division and growth sites, as it relies on BamA function. Furthermore, OMP biogenesis is preceded by PG biogenesis. BamA is only active around nascent, pentapeptide-rich PG, while it is inhibited by mature, tetrapeptide-rich PG. These phenomena were demonstrated in a few other Gram-negative species, but further experimentation is required to assume universality. These findings help explain how Gram-negative bacteria rapidly respond to environmental changes by altering OMP variety, and give researchers novel targets for antibiotic development (Mamou et al., 2022).

Figure 1. Summary of Mamou et al.’s findings, including the dynamics between PG maturity, BamA localization, and OMP biogenesis, as represented by BAM activity (Mamou et al., 2022).

In the experiment, the researchers used the techniques called epifluorescence microscopy and 3D structured illumination microscopy (SIM) following labelling with a specific, high-affinity monoclonal Fab antibody13 (MAB2) that binds extracellular loop 6 in BamA (Mamou et al., 2022). This was done so that BamA could be visualized and identifies the location of BamA (Mamou et al., 2022).

Epifluorescence microscopy, also known as widefield microscopy, is a type of fluorescence microscopy that is used to image fluorescently labeled samples. The fluorophores within the sample are excited to produce light of a different colour when exposed to a certain colour of intense incident light (Webb and Brown, 2014). The microscope's objective lens then collects the produced light and routes it through a number of filters to separate the fluorescent signal from the background. Lastly, a detector is used to photograph and record this signal (Webb and Brown, 2014).

3D structured illumination microscopy (SIM) is another form of fluorescence microscopy that uses a patterned illumination to obtain high-resolution, 3D images of a sample (Kraus et al., 2017). This is by illuminating the sample with a series of structured patterns which are created by interfering two or three light beams at different angles. This process allows the microscope to capture information about the sample that would otherwise be lost in traditional epifluorescence microscopy. The resulting image is reconstructed using advanced computational algorithms that combine the patterned images to create a high-resolution, 3D image of the sample. SIM can achieve a better resolution than conventional fluorescence microscopy, making it a good tool for visualizing subcellular structures and processes bridging the gap between light microscopy and structural biology (Kraus et al., 2017).

The researchers studied two TonB-dependent transporters (TBDTs), FepA and BtuB, which were labelled with high-affinity fluorescently labelled colicins that use these OMPs as receptors. To locate FepA, ColB was fused with either mCherry or GFP, while to locate BtuB, ColE9 was labeled using Alexa Fluor 488 (AF488). These OMPs were expressed separately in E. coli from plasmids and induced with arabinose.

Pull-down assay is a technique used to examine in vitro protein-protein interactions. In the experimental set up, one protein tested would be set as the immobile bait, which would be captured by an immobile tag. On the other hand, the other protein would be set as mobile prey, which would pass through the column containing the immobile bait. The column would then be wash and the proteins would be eluted in elution fraction. Bands of proteins interacting would be observed when the elution solution is analyzed using SDS-page. In the paper, Mamou and his colleagues used pull-down assays to investigate the binding of Bam proteins (all components) to peptidoglycan, POTRA domains of BamA to peptidoglycan, and Bam proteins to both nascent and mature peptidoglycan (Mamou et al., 2022).

Microscale thermophoresis (MST) is a commonly used biophysical technique to measure molecule interactions (Huang and Zhang, 2020). The basis of the technique is based on the thermophoresis of molecules, which is a phenomenon describing the movement of molecules in response to temperature gradients (Huang and Zhang, 2020). In an MST experiment, samples consist of a fluorescent molecule and a non-fluorescent binding ligand (tested) in serial diluted concentrations. (Huang and Zhang, 2020). Samples are put into capillaries of the MST machine, and an infrared laser will pass through the sample solution to generate a temperature gradient (Biocompare, 2013). The thermophoresis of the fluorescent molecules would be detected and recorded into thermographs (Biocompare, 2013). If the two molecules interact, the mobility of the fluorescent ligand would be affected (Huang and Zhang 2020). The dissociation constant (Kd) could be calculated from the fitting curve plotted in normalized fluorescence against ligand concentration (Biocompare, 2013). The lower the Kd value, the higher the affinity.

According to the paper, Manou et al. used MST to quantify the binding of Bam proteins and soluble uncross-linked disaccharide tetrapeptide chains (Tetran) of E coli. peptidoglycan. The researchers labeled the purified Bam proteins with Red-NHS and mixed them with serial diluted Tetran or mock digestions. The mixed solutions were put into capillaries for LED excitation, and the steady-state region of thermographs were analyzed according to a 1:1 binding model using the MO-Affinity Analysis software form NanoTemper Technologies (Mamou et al., 2022).

The goal of the in vitro BAM activity assay is to measure BAM activity. BAM activity directly correlates with OmpT activity. OmpT is an outer membrane protease (Yun, 2009). In this paper, Mamou et al. added a fluorogenic peptide to measure OmpT activity in vitro. They prepared three replicates for each of their three experiments that contained pentapeptide-rich PG, tetrapeptide-rich PG, or no PG. In each replicate, OmpT cleaved the fluorogenic peptide and produced a fluorescence emission. They measured this fluorescence emission over 1 hour and 20 minutes, and so they inferred BAM activity from these results.

The goal of this experiment was to determine which BAM proteins bind to PG. In this paper, Mamou et al. adapted a technique called “In Vivo and In Vitro Protein-Peptidoglycan Interactions” outlined in Bacterial Protein Secretion Systems.

The first step is to grow the bacterial strains in conditions that would produce the protein of interest (Li, 2017). Therefore, Mamou et al. grew their E. coli MC1061 in 50 ml of LB at 37 degrees Celsius. Then, they pelleted these cells by centrifugation at 4500 x g for 10 minutes at room temperature (Mamou et al., 2022). The pellets were washed with PBS three times (Mamou et al., 2022). They then added DTSSP to the cells until it reached a final concentration of 0.5 mM (Mamou et al., 2022). After incubating in room temperature for 10 minutes, the cross-linking reaction was stopped by adding Tris-HCl (Mamou et al., 2022).

Then, they took aliquots of cross-linked samples for later SDS-PAGE analysis (Li, 2017). The samples were boiled in 8% SDS, vigorously stirred for 30 minutes, and then pelleted by ultracentrifugation at 130,000 x g for 1 hour at room temperature (Mamou et al., 2022). The samples were then washed in 2% SDS, resuspended, and mixed with SDS-PAGE buffer with our without 10% B-mercaptoethanol at 100 degrees celsius for 10 minutes (Mamou et al., 2022). After briefly centrifuging, they loaded the samples onto 15% polyacrylamide gels for Western blot analysis (Mamou et al., 2022).

BamA is uniformly distributed throughout the cell surface and forms small clusters, whereas newly synthesised OMPs insert along the cell's long axis and at division sites. Researchers found no evidence of BamA enrichment at cell-division sites. In addition, they demonstrated that the process is not restricted to plasmid-based production of OMP genes and BamA's OMP insertion activity is cell cycle-dependent. These discovery was made using epifluorescence microscopy and 3D structured illumination microscopy (SIM) followed labelling with a high-affinity monoclonal Fab antibody13 (MAB2) that binds BAM extracellular loop 6.

Figure 1. BamA and FepA distribution on cell surface (Mamou et al., 2022).

Previous research has found that the pattern of OMP biogenesis is similar as PG biogenesis in E. coli. Mamou and collageus investigated the correlation of the patterns through co-labeling experiments. OMP biogenesis was imaged by labelling induced FepA with ColB–GFP, and PG biogenesis was imaged by fluorescent D-amino acid 7-hydroxycoumarincarbonylamino-D-alanine (HADA). Similar patterns of fluorescence label in PG and OMP biogenesis was indeed observed, and the florescence was especially strong at division sites. Co-localization analysis revealed a strong correlation between PG and OMP fluorescence patterns, which is even greater than the correlation between OMP and BamA patterns. PG biogenesis was also observed to occur consistently earlier than OMP biogenesis. With a closer look at the division site, cells were separated into two groups. Group 1 consists of cells with only PG labelling, while group 2 consists of cells with both PG and OMP labelling. No cells with only OMP labelling was found. Comparison between the two groups of cells illustrated a narrower septal width for group 2 cells than group 1 cells. These findings suggest that cell wall biogenesis precedes the biogenesis of OMPs at division sites. The same result was observed in Pseudomonas aeruginosa , suggesting a conserved mechanism for all gram negative bacteria.

Figure 2. (A) Fluorescence intensity of PG and OMP biogenesis in dividing cells. (B) Fluorescence intensity of group 1 cells and group 2 cells. (C) Septal width comparison of group 1 cells and group 2 cells (Mamou et al., 2022).

In order to determine the relationship between outer membrane proteins and peptidoglycan biogenesis, the researchers first determined whether purified BAM proteins physically interacted with peptidoglycan from wild-type E. coli MC1061 in pull-down experiments. The results depicted the specific binding of all BAM proteins other than BamA P1,2, BamCD and BamCDE complexes to peptidoglycan (Fig 3(C)).

Next, they tested the interaction of BAM proteins with soluble uncross-linked disaccharide-tetrapeptide chains, or Tetra, by microscale thermophoresis. Tetra's structure is depicted in Fig 3(D). The interaction of Tetra with certain domains of BamA as well as BamB, BamC, BamE, and BamCD showed that they bind specifically to Tetra chains of peptidoglycan, as illustrated in Fig 3(E).

Then, they treated E. coli MC1061 cultures with a chemical cross-linker (3,3'-dithiobis), followed by the isolation of peptidoglycan, reversal of cross-linking, and the detection of BAM proteins by SDS-PAGE with specific antibodies. While BamA and BamC were cross-linked to peptidoglycan isolated from detergent-treated cells, BamB and BamE were not cross-linked to peptidoglycan. The results displayed that BamA and BamC interact with peptidoglycan in E. coli cells (Fig 3(F)).

As such, BAM proteins are likely to be close to the cell wall due to the dimensions of the BAM complex inferred from structural studies but relative positions vary due to the dynamic nature of the OMP folding cycle which may explain why peptidoglycan is detected to bind to some BAM components in vitro but is not in vivo.

Figure 3. (C) Interaction of BAM proteins with purified PG by SDS-PAGE and Coomassie blue staining. (D) Structural scheme of poly-disaccharide-tetrapeptide chains (Tetra). (E) Interaction of BamA POTRA domain constructs with Tetra shown by MST. (F) Apparent Kd values for BAM protein/sub-complex interactions with Tetra, measured by MST where ND = no interaction detected (Mamou et al., 2022).

The researchers previously knew that sites of PG biosynthesis are transiently rich in pentapeptides. Coincidentally, pentapeptides are present in PG precursors, but not mature PG, which have tetrapeptides. These tetrapeptides form after DD-carboxypeptidase cleaves the pentapeptides.

They wanted to know if the difference in PG composition affects OMP biogenesis. Specifically, they wanted to know if stem peptide position of PG affects BAM-PG interactions. In the following experiments, tetrapeptide (mature) PG came from MC1061 E. coli. Meanwhile, pentapeptide (nascent) PG came from CS703-1 E. coli, which have non-functional DD-carboxypeptidase due to a mutation.

First, they completed pull-down assays with tetrapeptide-rich and pentapeptide-rich PG. They found that the BAM complex interacts with mature PG, but almost no interaction with nascent PG.

Fig. 4 shows the results of the pull-down assays. Each figure has three steps, the supernatant, the wash fraction, and the resulting pellet. On the left, the tetrapeptide-rich pellet contains Bam components, which indicates a successful pull-down. That is, the tetrapeptide-rich PG bind to the Bam subunits, but the pentapeptide-rich PG does not. A negative control on the right shows that Bam subunits are not pulled down at all in the absence of PG.

Figure 4. Adapted from Fig. 3b in Mamou et al. From left-to-right, these show the results of the pull-down assays of BAM in the presence of tetra-rich PG, penta-rich PG, and no PG as a negative control.

Now that they have discovered that mature PG more strongly interacts with nascent PG, they wanted to know whether the composition of peptides in PG affects BAM activity in vitro. They used an in vitro BAM activity assay with OmpT to test this. OmpT is an outer membrane protease. Greater BAM activity means that more OmpT will be translocated across the outer membrane. OmpT activity is tracked by the fluorescence emitted following the cleavage of a fluorogenic peptide.

The pertinent results of this assay is the half-maximal effective concentration. The half-maximal effective concentration is the amount of substrate needed to reduce the activity of a protein by half. A lower half-maximal effective concentration means that it is more effective in reducing BAM activity. They found that mature PG has a much lower half-maximal effective concentration. Therefore, mature PG decreases BAM activity more effectively than nascent PG. This means that BAM is more active in pentapeptide-rich regions.

In fig. 5, we have time in the x-axis and fluorescence by fluorogenic peptide cleavage by OmpT in the y-axis. On the left, we see that increasing concentrations of nascent, pentapeptide-rich PG leads to a small decrease in fluorescence, but it is very little. On the contrary, the graph on the right shows that increasing concentrations of mature, tetrapeptide-rich PG leads to much larger decreases in fluorescence.

Figure 5. Adapted from Fig. 5c and 5d in Mamou et al. Increasing concentrations of tetra-rich PG decreases fluorescence which indicates a decrease in BAM activity. In contrast, increasing concentrations of penta-rich PG has very little effect on fluorescence.

In fig. 6, we see the PG-concentration on the x-axis and the relative BAM activity on the y-axis. We see that the half-maximal effective concentration is much lower for tetra-rich PG than penta-rich PG. This means that mature PG inhibits BAM activity. Therefore, regions that have less mature PG have greater BAM activity.

Figure 6. Adapted from Fig. 3e in Mamou et al. Tetra-rich PG has a much lower half-maximal effective concentration of 0.86 mg/ml while penta-rich PG's is more than 5 mg/ml.

With their acquired knowledge, the researchers hypothesized that cell wall composition was the driver of OMP patterning in the outer membrane of E. coli in vivo. To test this hypothesis, they used two approaches, both of which monitor PG incorporation with HADA labeling and OMP biogenesis using FepA labeling.

First, they tried antibiotics that disrupt the peptidoglycan layer. While Mecillinam had a negligible effect on the coordination between PG and OMP biogenesis as division sites were still formed, Aztreonam was able to inhibit the formation of division septa/sites and decreased OMP biogenesis activity was observed when compared to the septa of dividing cells. When it was removed, division sites were reformed and activity was enhanced.

The second approach involved inducing OMP (FepA) expression in a pentapeptide-rich mutant strain (CS703-1) and its tetrapeptide-rich parent (CS109). The distribution of FepA in CS109 was similar to the wild-type, BW25113. However, in the pentapeptide-rich strain, there was a defective biogenesis pattern with reduced mid-cell bias. To determine the effectiveness of mid-cell OMP insertion in highly disrupted cell envelopes, the researchers developed a live-cell assay. This assay saw results that displayed how FepA distribution was defective at the mid-cell in the pentapeptide-rich cell wall mutant but could be restored through PBP5 expression in the background. This indicated that outer membrane instability associated with PG mutants can be rescued by correcting the cell wall.

Together, the data showed that the local density of pentapeptides underpins OMP biogenesis patterns and outer membrane stability in E. coli. Some key takeaways are: nascent pentapeptides do not lower BAM OMP insertion activity like matured ones, pentapeptides are absent in old poles but abundant at division sites in PG, and the opposite is true for inhibitory tetrapeptides such that little or no OMP biogenesis is seen in old polar regions of E. coli.

The synchronization of OMP insertion with PG growth is essential for stabilizing cell membrane formation. Thus, the study of this phenomenon is of great importance to scientists who are currently looking for BAM-targeting compounds. Understanding the enzymes involved in peptidoglycan synthesis and remodeling, and the mechanisms of peptidoglycan incorporation into the bacterial envelope are essential for implications of peptidoglycan metabolism for antibiotic resistance (Deghelt & Collet, 2022). Altering or blocking this synchronization may contribute to the global effort in fighting against antibiotic resistance.

By adding OMP from the division site and binary partitioning, bacteria can rapidly turnover OMPs without activating protein degradation. This process enhances the survival of bacteria in a rapidly changing environment and create new challenges for combating bacterial infections. Finding a way to inhibit turnover and restrict adaptation may effectively slow down the process of antimicrobial resistance development.

The exact mechanism of BAM inhibition by tetrapeptides is still unknown. There are many possibilities, including direct blockage of the access of unfolded OMP substrates to the BamA or indirect restriction of the conformational transitions of the BAM machinery that enable OMP folding and release.

It remains unclear why BAM subunits bind PG differently in vitro and vivo. The paper suggests that the highly dynamic nature of the OMP folding cycle in bacteria's vivo condition may explain this difference. The Bam complex undergoes conformational changes during the folding process, so its structure and its affinity towards PG compartment might not be strictly the same as in vitro condition.

The maturation of PG may not be the only element that controls BAM complex activity. The Sec translocon in the inner membrane, which plays a crucial role in unfolded OMP substrate transportation, is another possible element that contributes to PG-mediated BAM control. The Sec and BAM translocons can associate with each other in the periplasmic space, the region between the inner and outer membranes of bacteria through the interaction between BamA and SecY compartments. The finding provides new insights into the complex process of OMP biogenesis and suggest potential novel explaination of PG-mediated OMP insertion.

Colicin

Colicin is a bacteriocin labeling that can be toxic to some strains of E.coli. It may impact cell function or integrity when apllied as a experimental labelling substance.

MAB2

Without reliable citation backup, the statement 'MAB2 has no effects on growth or function’ may not be entirely reliable.

The pull-down assays may not reflect the true interactions in vivo as the real dynamics can be affected by other molecules, proteins, ions, etc. in native environments.

Figure

The use of red-green color schemes may be problematic for individuals with red-green color blindness, and blurry images can impact interpretation and analysis.

Data

The paper concluded that tetrapeptide had little effect on BAM activity based on a sample size of three (n=3), such small sample size may not be reliable and may not be representative of the true population proportion.

Biocompare and NanoTemper technologies. (2013). Bench Tip Video: Measuring Interactions Using Microscale Thermophoresis. YouTube. Retrieved March 1, 2023, from https://www.youtube.com/watch?v=lod8xeaaAZc.

Deghelt, M., & Collet, J.-F.. (2022). Bacterial envelope built to a peptidoglycan tune. Nature, 606(7916), 866–867. https://doi.org/10.1038/d41586-022-01739-3

Gunasinghe, S. D., Shiota, T., Stubenrauch, C. J., Schulze, K. E., Webb, C. T., Fulcher, A. J., Dunstan, R. A., Hay, I. D., Naderer, T., Whelan, D. R., Bell, T. D. M., Elgass, K. D., Strugnell, R. A., & Lithgow, T.. (2018). The WD40 Protein BamB Mediates Coupling of BAM Complexes into Assembly Precincts in the Bacterial Outer Membrane. Cell Reports, 23(9), 2782–2794. https://doi.org/10.1016/j.celrep.2018.04.093

Huang, L., & Zhang, C. (2020). Microscale thermophoresis (MST) to detect the interaction between purified protein and small molecule. Methods in Molecular Biology, 187–193. https://doi.org/10.1007/978-1-0716-0954-5_17

Kleanthous, C., Rassam, P., & Baumann, C. G. (2015). Protein-protein interactions and the spatiotemporal dynamics of bacterial outer membrane proteins. Current Opinion in Structural Biology, 35, 109–115. https://doi.org/10.1016/j.sbi.2015.10.007

Li, G., & Peter Howard, S. (2017). In Vivo and In Vitro Protein–Peptidoglycan Interactions. Bacterial Protein Secretion Systems: Methods and Protocols, 143-149. Mamou, G., Corona, F., Cohen-Khait, R., Housden, N. G., Yeung, V., Sun, D., Sridhar, P., Pazos, M., Knowles, T. J., Kleanthous, C., & Vollmer, W.. (2022). Peptidoglycan maturation controls outer membrane protein assembly. Nature, 606(7916), 953–959. https://doi.org/10.1038/s41586-022-04834-7

Quantitative 3D structured illumination microscopy of nuclear structures. Nat Protoc 12, 1011–1028 (2017). https://doi.org/10.1038/nprot.2017.020

Rassam, P., Copeland, N. A., Birkholz, O., Tóth, C., Chavent, M., Duncan, A. L., Cross, S. J., Housden, N. G., Kaminska, R., Seger, U., Quinn, D. M., Garrod, T. J., Sansom, M. S. P., Piehler, J., Baumann, C. G., & Kleanthous, C.. (2015). Supramolecular assemblies underpin turnover of outer membrane proteins in bacteria. Nature, 523(7560), 333–336. https://doi.org/10.1038/nature14461

Webb, D. J., & Brown, C. M. (2013). Epi-fluorescence microscopy. Methods in molecular biology (Clifton, N.J.), 931, 29–59. https://doi.org/10.1007/978-1-62703-056-4_2

Yun, T. H., & Morrissey, J. H. (2009). Polyphosphate and omptins: novel bacterial procoagulant agents. Journal of cellular and molecular medicine, 13(10), 4146–4153. https://doi.org/10.1111/j.1582-4934.2009.00884.x