Uncovering the mechanism for polar sequestration of the major bacterial sugar regulator by high-throughput screens and 3D interaction modeling

Summary

The poles of rod-shaped bacteria emerge as regulatory hubs. We have shown that enzyme I (EI), the major bacterial sugar metabolism regulator, is sequestered when not needed in TmaR phase-separated condensates in Escherichia coli cell poles. Here, we combined genetic and automated microscopy screens to identify residues in EI and TmaR that are important for their interaction and colocalization. Mutating these residues affects EI-TmaR interaction in bacteria and impairs co-phase separation in yeast. The results were used to generate an EI-TmaR interaction model, which agrees with coevolution data and is supported by conservation of the interacting residues and EI-TmaR colocalization in other species. Mutating residues predicted to interact electrostatically further supports our model. The model explains how TmaR controls EI activity and its interaction with the phosphoprotein HPr and, hence, sugar uptake. Our study highlights the importance of sugar metabolism spatial regulation during evolution and presents a way to unravel protein-protein interactions.

Graphical abstract

Highlights

• •EI-TmaR complex model agrees with coevolution, conservation, and localization data
• •EI-TmaR interaction model explains how TmaR controls EI activity
• •Mutating the interacting residues impairs EI-TmaR co-phase separation in yeast
• •We suggest a way to unravel protein interactions by combining screens with modeling

The bacterial pole-localizer TmaR sequesters the sugar regulator EI in polar condensates. Albocher-Kedem et al. identify residues in EI and TmaR that impair their interaction and colocalization by genetic and microscopy screens. They then use these residues to generate a 3D interaction model that explains how TmaR controls EI activity.

Introduction

The complexity of bacterial cells has been underestimated for decades. The scarcity of membrane-bounded organelles led to the assumption that macromolecules are not specifically localized in bacterial cells. The first reports of proteins localizing to specific domains in Escherichia coli, e.g., to the septa¹ and to the cell poles,² started to appear three decades ago. In the last two decades, it became clear that macromolecules have distinct spatial distribution also in bacteria.^3,4 The structural and biochemical features of the poles of rod-shaped bacterial cells make them special regions to which various proteins, as well as RNAs localize.^3,4,5 The list of proteins that localize to the E. coli cell poles is constantly growing, and they are emerging as hubs for regulation of central cellular functions, including motility, cell cycle, metabolism, differentiation, pathogenesis, and secretion.^3,6,7 We have previously proposed that colocalization of key sensory and regulatory macromolecules at the pole facilitates integration of extracellular information to generate an optimal response, suggesting that the response to many environmental stimuli is regulated at the poles.⁸

Proteins might be targeted to the poles of symmetrically dividing bacteria for various reasons. One example is the assembly of supramolecular assemblies that exert their effect from the pole, as in the case of the E. coli chemotaxis complex.^9,10 Another example is the sequestration of proteins that are kept inactive until they are needed, as in case of MurG, FtsZ, and enzyme I (EI), which are involved in peptidoglycan biosynthesis, cell division, and regulation of sugar consumption, respectively.^11,12,13,14 The cues that recruit proteins to the poles and maintain them there include recognition of phospholipids that are enriched in this domain, recognition of the specific geometry of the membrane at the poles (concave), binding to pole-tethered proteins, and nucleoid occlusion.³

Here, we focus on EI, the major regulator of hierarchical sugar uptake and utilization by the phosphotransferase system (PTS) in almost all bacterial species.¹⁵ The PTS system has been intensively studied, due to its importance in linking glycolysis to sugar uptake and the involvement of the PTS in all aspects of cellular physiology.¹⁶ By phosphorylating itself, with the donor being phosphoenolpyruvate (PEP), the final product of glycolysis, EI initiates a phosphorylation cascade that includes the phosphoprotein HPr and one of the various sugar-specific permeases, called enzyme II (EII), which culminates in simultaneous transport and phosphorylation of energetically preferred sugars, termed PTS sugar. The phosphorylated sugars are then hydrolyzed to phosphorylated glucose or fructose that enters glycolysis.¹⁵ The PTS also regulates non-PTS sugar utilization by controlling the activity of other proteins (e.g., adenylate cyclase).¹⁵ In addition to globally modulating carbon metabolism, the PTS carries out many regulatory functions related to nitrogen and phosphate metabolism, chemotaxis, potassium transport, and virulence.¹⁷ The involvement of the PTS in carbohydrate-mediated chemotaxis is mediated by EI, which exerts an inhibitory effect on the activity of the sensor kinase CheA.¹⁷ We have previously shown that EI forms clusters at the poles of E. coli cells¹⁸ and that it localizes to regions of negative membrane curvature, present in the poles and in the septa of dividing cells (future poles).¹⁹ Of note, the chemotaxis system also localizes to the cell poles,²⁰ facilitating its crosstalk with the PTS, which is required for the production of optimal responses.⁸ The polar clusters of EI form stochastically by the assembly of pre-existing dispersed molecules during growth, with the polar fraction of EI acting as a reservoir of ready-to-act EI molecules.^13,14

EI subcellular localization and activity were shown by us to be controlled by TmaR (previously called YeeX), the first pole-localizer to be discovered in E. coli, which sequesters EI at the poles, thus preventing it from functioning and releasing it upon need.¹⁴ TmaR was recently discovered to form a membraneless organelle at the poles and to include non-active EI in the polar condensates, thus providing a mechanism for EI regulation by polar sequestration.²¹ Since phase separation (PS) emerges as a way to reduce variability in protein concentration outside the condensates,²² this finding is in line with our report that TmaR limits the degree of heterogeneity in EI activity by sequestering it at the poles.¹⁴ Although TmaR was shown to phase separate by heterotypic interactions with RNA and possibly with other cellular factors, TmaR is a crucial component in the formation of the polar condensates.²¹ The finding that EI is not recruited to the polar condensate in the absence of TmaR¹⁴ implies that TmaR serves as a scaffold for EI condensation. More recently, we found that TmaR controls bacterial motility by protecting flagellar RNAs in the polar condensates, thus regulating flagella production.²¹ The control of metabolism and motility in TmaR-EI condensates provides an explanation for the linkage between metabolism and motility. The recent finding that TmaR also controls sRNA-mediated regulation by recruiting their chaperon, Hfq, to the polar condensates upon stress²³ implies that TmaR-EI polar condensates are hubs that link major regulatory and survival functions and, hence, their interaction warrants studying.

The structure of EI in multiple states has been extensively studied by various experimental approaches. Structure determination of phosphorylated EI by X-ray diffraction at 2.7 Å resolution identified two discrete domains: the N-terminal domain (EIN), also called the activation domain, which contains the phosphorylation site and interacts with HPr, and the C-terminal domain (EIC), also called the dimerization domain, which mediates EI dimerization and binds to PEP.²⁴ The two domains are connected by the linker, which determines the relative positioning of the two domains. The structures of EI and of its complex with HPr, determined by X-ray scattering and NMR, shed light on EI dynamic equilibrium between closed, partially closed, and open states,^24,25,26,27 and highlight the role of the linker during the cycle of EI phosphorylation and phosphate transfer. Notably, EIC mediates EI targeting to the pole.¹⁸ The structure of TmaR has not been determined yet, but it is predicted to form a coiled coil.¹⁴

In this study, we set out to explore the mode of EI-TmaR interaction and its implications on EI activity. To this end, we developed a novel methodology for screening mutants in EI and TmaR, which fail to interact, by combining random mutagenesis and high-throughput automated microscopy. We isolated EI mutants that maintain activity but do not localize to TmaR polar clusters, and TmaR mutants that localize to the poles but fail to recruit EI. The mutant characteristics were validated by assaying EI-TmaR interaction, EI activity, and the ability of EI and TmaR to phase separate together in a heterologous yeast system. The screens uncovered predicted interaction interfaces and these could be applied to predict EI-TmaR 3D molecular interaction. Our predicted interaction model, which is based on our experimental data, is supported by coevolution data and provides an explanation for how TmaR controls EI activity by controlling its localization.

Results

Isolation of EI functional mutants that do not cluster with TmaR at the cell poles

To identify the residues in EI that are required for its interaction with TmaR, which controls EI activity by polar sequestration and release,¹⁴ we designed a genetic screen for the isolation of EI mutants that fail to localize to the poles but are still active in regulating sugar consumption (Figure 1A). To this end, we constructed a library of EI mutants by random mutagenesis. The library was obtained by error-prone PCR of the ptsI gene, which codes for EI, fused to mCherry, present on a low-copy plasmid in the context of the entire pts operon with its native promoter. To ensure that aberrant expression levels are not affecting our screen, we measured the level of EI-mCherry expressed from this plasmid and could verify that it is only modestly elevated (2.9 ± 0.4 times higher than EI-mCherry expressed from the chromosomal pts operon) (Figures S1A and S1B). Indeed, we found a slightly higher fraction of it is detected in polar clusters compared to chromosome-encoded EI-mCherry (Figures S1C and S1D). The library of EI mutants was then introduced into cells deleted for the pts operon (Δpts). To isolate mutants that retained EI ability to regulate sugar consumption, we streaked 3,552 mutants on MacConkey-glucose indicator plates and picked the 2,304 colonies that generated a red color, implying that they consumed the glucose. These colonies were grown in 96-well plates and subjected to a high-throughput automated fluorescence microscopy screen. Six EI mutants no longer localized to polar clusters and were detected as homogenously spread within the cytoplasm despite their ability to consume sugar.

Figure 1.

Isolation of EI mutants that are functional but do not cluster

(A) The screen scheme. A library of EI mutants was constructed by error-prone PCR of the ptsI gene, fused to mCherry, present on a low-copy plasmid in the context of the entire pts operon with its native promoter. The library was introduced into MG1655 cells deleted for the pts operon. Out of 3,552 colonies that were streaked on MacConkey-glucose plates, 2,304 generated red color, indicating that they express functional EI, which enables the cells to consume the sugar. These mutants were screened by automated fluorescence microscopy to identify those that express EI protein that does not cluster at the poles. The six mutants expressing functional and diffuse EI were sequenced.

(B) The screen results. The position of each amino acid that has been replaced in the six functional and mislocalized EI variants is shown. Each row shows the mutated residue(s)in one of the mutants, represented by dots, along the x axis. An orange background marks EIN, purple background marks its linker region, and green background marks EIC.

(C) The 3D structure of the EI dimer (PDB: 2HWG). The two models on the left are cartoons showing the EI 3D structure from different angles, and the one on the right shows the protein surface. EIN is colored in orange, the linker region in purple, and EIC in green. Amino acids whose substitution yielded functional but mislocalized EI mutants are marked in red. G266 in both subunits is presented as a red sphere.

(D) Fraction of cells with clusters of mCherry-tagged WT EI or EI mutants with single point mutations, all expressed from the pts native promoter and locus in the chromosome. The bars show the standard deviations (see representative images for all EI variants in Figure S1G; n is between 731 and 1,120 cells). Statistical analysis for the differences between WT and the mutants was conducted using the unpaired t test based on 12 fields. ^∗p < 0.01, ^∗∗p < 0.001, and ^∗∗∗p < 10⁻⁹.

(E) Close-up on the β turn in EI, right after the linker, in the 3D structure model (see C). EIN is colored in orange, the linker between the domains in purple, and EIC in green. G266 is depicted as a red sphere.

(F) Images of cells expressing TmaR-YFP (yellow) and mCherry-tagged WT EI or G266C (red), all expressed from their respective native promoter and locus in the chromosome. Scale bars, 2 μM (see the respective results with EI variants G266S and G226D in Figure S3A).

(G) Fraction of cells in the images shown in (F) and Figure S3A with mCherry-tagged EI variants (red bars) and TmaR-YFP (yellow bars) calculated based on fluorescence microscopy. The bars show the SDs (n is between 467 and 1,175 cells). Statistical analysis for the differences between WT and the mutants was conducted for each protein using the t test based on three fields. ^∗p < 0.001.

(H) Cells expressing mCherry-tagged EI, WT, or G266C, from the native pts promoter and locus in the chromosome, and harboring a plasmid expressing TmaR from an isopropyl ß-D-1-thiogalactopyranoside (IPTG)-inducible promoter, grown with increasing concentrations of IPTG (0, 0.1, or 1 mM) were imaged. The CVIs of the respective EI variants were calculated based on fluorescence intensity throughout each cell. Violin plots depict the CVI distribution for mCherry-tagged EI variants, with the blue line showing the median value. Between 360 and 850 cells were analyzed per variant. Representative cells above each violin illustrate the quantitative data. Scale bars, 2 μm. The same analysis for EI variants G266S and G226D in provided in Figure S3B.

(I) CoIP of the EI variants with TmaR. Cells deleted for tmaR and expressing mCherry-tagged WT EI or one of the three EI-G266 mutant proteins from the native pts promoter and locus in the chromosome, and harboring a plasmid expressing TmaR-His or just the His tag, were crosslinked by formaldehyde. TmaR or the His tag were immunoprecipitated using nickel beads. Samples were fractionated by SDS-PAGE, blotted onto a nitrocellulose membrane, and probed with anti-mCherry antibody to detect EI-mCherry, followed by stripping of the membrane and probing it with anti-His antibody to detect TmaR-His. The lower panel shows TmaR-His and the mCherry-tagged EI variants that were co-precipitated with it, detected by anti-His and anti-mCherry antibodies, respectively, on a representative membrane. The upper panel shows the ratio between the band intensities of mCherry-tagged EI mutants expressed in the presence of TmaR-His minus the intensities in the absence of TmaR-His and the intensity of TmaR-His band. The signal obtained when WT TmaR and WT EI are co-expressed is defined as 1. The bars show the SDs between three biological replicates. Statistical analysis for the differences between WT and the mutants was conducted using the two-sided Mann-Whitney test. ^∗p < 0.05.

DNA sequence analysis of the six EI mutants that are active but fail to localize to the poles revealed that three had a single amino acid substitution, one had two substitutions, and two had either four or five substitutions (Table S1; Figure 1B). Looking at the mutated residues in these six mutants, we noticed that (1) there is a high-mutation region between residues 261 and 325, which is at the beginning of EIC, next to the linker that connects the two domains of EI (Figures 1B and S1E); (2) most mutations are in residues that are on the surface of EI (Figures 1C and S1F; Video S1); and (3) glycine 266 (G266) was replaced by three different residues (cysteine, serine, and aspartate; mutants 1, 3, and 4). When introducing 11 different mutations from the six mutants as single base substitutions to the chromosomal ptsI gene fused to mCherry, only the three G266 mutants were detected as diffused throughout the cell, whereas the others were detected in polar clusters (Figures 1D and S1G). To ensure that the observed distribution of the three G266 mutants is not a result of low expression, we verified that their level is comparable to that of wild-type (WT) EI by western blot analysis (Figure S1H). Finally, we verified that their activity is comparable to that of WT EI by comparing the red color that the different EI proteins produce on MacConkey-glucose plates (Figure S1I). Hence, the three EI variants with substitutions of G266 have the phenotype that we set out to isolate, that is, failure to localize to the polar TmaR condensates and, yet, preservation of the ability to regulate sugar consumption.

Video S1. The screen identified residues in EI in the context of its 3D structure

A video of the 3D structure of EI dimer (PDB: 2HWG) showing first a cartoon model and then the protein surface. The N-terminal domain (EIN), the linker region, and the C-terminal domain (EIC) are colored in orange, magenta, and green, respectively. Residues whose substitution yielded functional, but mislocalized, EI mutants are marked in red. G266 is presented as a red sphere in both monomers.

Based on the 3D structure of EI, determined by X-ray crystallography²⁴ and NMR,^25,26,27 G266 (marked as red spheres in Figures 1C, 1E and S2), which is located in a β turn right after the linker that connects the two domains of EI, seems important for generating flexibility of EI N′-terminal domain (EIN) relative to EI C′-terminal domain (EIC) (see structurer in Figures S2B–S2D, and Normal mode analysis below), because the small glycine residue bestows conformational flexibility to proteins’ backbones.^28,29 We assume that the mutation in G266 leads to a stiffer but nonetheless open conformation of EI, as it allows binding to and activation of HPr, although future investigation is required (see Discussion). Taken together, our screen for EI mutants that retained their functionality but lost subcellular distribution highlights EI G266, which serves as a pivotal joint for opening and closing the EI dimer to enable binding to other proteins (Figure S2), as essential for EI localization to the cell poles.

Mutations replacing G266 in EI impair EI-TmaR interaction in vivo

Since TmaR is responsible for EI polar localization, and in its absence EI does not cluster at the poles,¹⁴ we hypothesized that G266, whose substitution by three different residues prevents EI localization to the poles, is important for EI interaction with TmaR. To test our hypothesis, we examined the interplay between the EI G266 mutants and TmaR, compared to WT EI-TmaR interplay. We first co-imaged TmaR-YFP together with mCherry-tagged WT EI or with one of the EIG266 mutants, all expressed from their respective native promoters and loci in the chromosome, by fluorescence microscopy. As expected, only WT EI clustered with TmaR at the poles (Figures 1F, 1G, and S3A). We next asked if the inability of the EI G266 mutants to be recruited to the TmaR polar clusters can be overcome by increasing TmaR level in the cell. To address this question, we expressed TmaR from a plasmid with an inducible promoter in cells expressing the mCherry-tagged EI proteins from the chromosome and monitored the subcellular distribution of mCherry by fluorescence microscopy upon a gradual increase in TmaR level (Figures 1H and S3B). The distribution pattern of the EI proteins within the cells was inferred from the coefficient of variation fluorescence intensity (CVI), calculated by dividing the standard deviation by the mean intensity of the mCherry fluorescent signal (Figures 1H and S3B), since lower CVI implied more even distribution of the fluorescence signal in the cells. The results show that, although increasing TmaR level increased the polar fraction of WT EI, as inferred from the increase in the fraction of cells with EI polar clusters and the increase in clusters size, it did not affect the distribution of the EI G266 mutants, and they remained diffuse independently of TmaR levels.²¹

To directly examine if the substitutions of G266 in the isolated mutants impair the association with TmaR, we used co-immunoprecipitation (coIP). For this analysis, mCherry-tagged WT EI or one of the three G266 mutants were expressed from the native pts promoter and locus in the chromosome together with TmaR-His or just the His tag, which were expressed from a plasmid, in ΔtmaR cells. Following crosslinking by formaldehyde, the His-tagged proteins and their interactors were isolated using nickel beads. Because the EI variants stuck to the nickel beads nonspecifically, we calculated the TmaR-specific binding of EI by comparing the difference (delta) between the band intensity of mCherry-tagged EI variants expressed in the presence and absence of TmaR-His, and then the ratio between the calculated difference (delta) and the TmaR-His band (see STAR Methods). We present the ratio of each EI variant relative to WT EI. The results show that WT EI associates with TmaR in vivo, as indicated by their co-precipitation, whereas substitution of G266 by other residues significantly reduces the ability of EI to associate with TmaR (Figure 1I).

To see if EI alone is sufficient for recognizing TmaR and interacting with it and if the replacement of G266 impairs this interaction, we used the far-western assay, which is widely accepted as an indication for direct interaction between two proteins.^30,31 Lysates of ΔtmaR cells, either overexpressing TmaR from a plasmid or not, were fractionated by SDS-PAGE and blotted onto a nitrocellulose membrane. Four membranes with equal amounts of the two lysates (see Figure S3C) were probed with either purified His-tagged WT EI or with each of the EI G266 mutants, followed by incubation with antibodies against the His tag. While it is clear that the two proteins can directly interact, the effect of substituting G266 seems mild and not significant compared to the effect observed by coIP (Figures 1I and S3C). Of note, the EI variants used for the far-western assay were overexpressed, while, for the coIP assay, they were not overexpressed. Based on the two assays, we conclude that the EI mutants, in which G266 was replaced by other residues, have reduced potential to interact with TmaR compared to WT EI. This reduced affinity can be compensated by overexpression of the EI mutant proteins.

To further validate the influence of the mutations in EI G266 on the interaction with TmaR in vivo, we conducted a bacterial-two-hybrid assay.³² To conduct this assay, we expressed WT TmaR from one plasmid together with WT EI or one of its three variants with different replacements of G266 from a second plasmid. The results in Figure S3D show that, although TmaR interacts with WT EI, as demonstrated by the growth of blue colonies on X-Gal plates, due to expression of the β-galactosidase reporter gene, TmaR interaction with the three EI mutants in G266 is significantly reduced. These results are in accord with the results obtained in the coIP assay and support our conclusion that replacing G266 in EI mutants with other residues significantly impairs the interaction with TmaR.

Together, our results show that EI and TmaR associate in vivo and are capable of directly interacting, and that G266 in EI is important for this association, although additional factors might affect interaction of the two proteins.

Isolation of TmaR mutants that prevent EI recruitment to the poles

In light of our successful trial to identify residues in EI that are important for its recruitment to TmaR polar clusters, we decided to perform the complementary screen to identify residues in TmaR that contribute to its ability to recruit EI, that is, whose substitution would impair recruitment of EI to the poles (Figure 2A). We thus constructed a library of TmaR mutants by random mutagenesis of the tmaR gene fused to YFP and expressed it from a low-copy plasmid regulated by the tmaR native promoter. The library was introduced into cells deleted for tmaR (ΔtmaR) and expressing EI-mCherry from the native promoter and locus in the chromosome. The level of the plasmid-expressed TmaR-YFP is 5.6 ± 0.3 times higher than TmaR-YFP expressed from the chromosome, as indicated by western blot analysis (Figures S4A and S4B). Accordingly, the fraction of cells with TmaR and EI clusters in these cells is higher than in cells expressing EI-YFP from the chromosome and the clusters are bigger (Figures S4C and S4D). Important for our screen, the increase in polar localization of TmaR, when expressed from a plasmid, was accompanied by a comparable increase in EI polar localization (Figures S4C and S4D). Therefore, we used this setup to isolate TmaR mutants that prevent EI recruitment by high-throughput automated fluorescence microscopy. Out of 1,152 TmaR-YFP variants, each originating from a single colony, we detected 17 variants that did not lose their polar localization but no longer recruited EI-mCherry to poles; that is, EI-mCherry was detected as homogenously spread within the cytoplasm.

Figure 2.

Isolation of TmaR mutants that impair EI-TmaR colocalization

(A) The screen scheme. A library of TmaR mutants was constructed by error-prone PCR of the tmaR gene present on a low-copy plasmid with its native promoter and expressing TmaR fused to GFP. The library was introduced into MG1655 cells deleted for the tmaR gene. Out of 1,152 colonies for which localization of EI and TmaR was monitored by automated fluorescence microscopy, 17 were found to express a TmaR mutant that clusters at the poles but did not recruit EI, which was distributed in the cytoplasm. These TmaR mutants were sequenced.

(B) The screen results. The position of each amino acid that has been replaced in the 17 TmaR variants that cluster at the poles but causes EI to mislocalize is shown. Each row shows the mutated residue(s) in one of the mutants, along the x axis. Positions whose replacement were isolated as a single substitution are depicted by blue spheres. The blue circles represent the two positions whose replacement was accompanied by an additional mutation, which was isolated also as a single mutation.

(C) A structural model of TmaR generated by trRosetta.^33,34,35 Amino acids substituted in patch I and patch II are marked in orange and magenta, respectively.

(D) Fraction of cells with clusters of EI-mCherry (red bars) and YFP-tagged TmaR variants (yellow bars), all expressed from their native promoter and locus in the chromosome. The bars show the SDs (n is between 404 and 812 cells). Statistical analysis of the differences between the WT and the mutants was conducted for each mutant using unpaired t test, based on 10 fields. ^∗p < 0.001, ^∗∗p < 0.0001, ^∗∗∗p < 10⁻⁵; no asterisk means non-significant (NS) > 0.01.

DNA sequence analysis of the 17 TmaR mutant proteins, which fail to localize EI to the poles, revealed that 14 of them had a single base substitution, and three had two substitutions. Since, in two of the latter variants, one substitution was also isolated as a single point mutation, we assumed that the repeating mutation is the one responsible for the observed phenotype. In the third variant, both substitutions were isolated also as a single point mutation (Table S2; Figure 2B). Altogether, there were eight different substitutions, five of them isolated more than once, which mapped to two regions in TmaR (Table S2; Figure 2B). Four mutations, F8L, V11I, L12M, and R16P, localized to a region near TmaR N terminus and the other four, D78H, A84G, E85D, and K95E, localized to a region close to the C terminus (marked as “patch I” and “patch II,” respectively, in Figure 2B). To get an idea of the physical context of the mutated residues, we generated a model of the structure of TmaR using trRosetta (yanglab.nankai.edu.cn/trRosetta).^33,34,35 The model consists of two α helices that fold into a coiled-coil structure, as we previously suggested,¹⁴ and the mutated residues are located in these two helices with their side chains facing out (Figure 2C; Video S2).

Video S2. Predicted 3D structure of TmaR

A video of the 3D structure model of TmaR generated by trRosetta. Amino acids, which were substituted in the sticky and stabilizing patches in the mutants obtained in our screen, are colored in orange and magenta, respectively, with their protruding side chains presented as sticks.

When we introduced the above eight mutations to the chromosomal YFP-tagged tmaR gene in cells that also express EI-mCherry from the chromosome, they all exhibited the same phenotype, which we observed with the TmaR-YFP mutants expressed from a plasmid. That is, although the TmaR mutants were detected in polar clusters, the EI-mCherry protein co-expressed with them was spread out, hardly detected at the poles (Figures 2D and S5). The reduction in the fraction of cells with a TmaR cluster in the case of F8L and R16P (see yellow columns in Figure 2D) can be attributed to a lower level of these mutant proteins, as demonstrated by western blot analysis (Figure S6A) and their lower mean intensity (MI) (Figure S6B). The reduction in the fraction of cells with a TmaR cluster in the case of R16P and D78H is apparently due to a higher fraction of diffused TmaR (Figure S5), inferred also from the low CVI of the YFP fluorescent signal in these cells (Figure S6C). Of note, the YFP tag decreases the stability the D78H mutant protein, whereas the His tag does not (Figure S6A). Still, despite the quantitative reduction in the fraction of cells with clusters of these three TmaR mutants, the absolute phenotype is the same; that is, in cells in which they are detected in polar clusters, EI-mCherry is spread out. Hence, all eight mutants were followed up by further analysis.

As a secondary screen, we tested the effect of the fluorescent tags on EI-TmaR colocalization since the fraction of cells with a polar EI-mCherry cluster in populations expressing chimeric TmaR-YFP has been shown to be lower than in populations expressing untagged TmaR.¹⁴ Hence, we compared the fraction of cells with EI-mCherry polar clusters that co-express YFP-tagged or untagged TmaR variants (Figure S6D), as well as of cells expressing each YFP-tagged TmaR variants with mCherry-tagged or untagged EI (Figure S6E). The results suggest that tagging EI with mCherry had no effect on the localization of TmaR-YFP variants (Figure S6E). However, tagging TmaR variants with mutations in patch I (F8L, V11I, and L12M) with YFP had an effect on the localization of EI-mCherry, although the effect in the case of F8L, V11I was minor (Figure S6D). We therefore did not continue to analyze the TmaR L12M mutant further.

Overall, in the screen for TmaR residues that are important for recruiting EI to the poles, we identified seven residues, residing in two patches on TmaR helices, whose substitution either does not affect TmaR polar clustering or affects it to some extent, but completely or almost completely impairs TmaR ability to localize EI to the poles.

The screen-isolated mutations in TmaR and EI impair their interaction

To examine whether the TmaR mutants, isolated in the screen described above, fail to recruit EI because they lost the ability to interact with EI, we tested the effect of the mutations on the association with EI in vivo by co-immunoprecipitation (coIP) and the bacterial two-hybrid assay, as well as their ability to directly interact with EI in vitro by the far-western assay. To examine the ability of TmaR mutants to associate with EI in vivo by coIP, we asked if EI is co-precipitated with them from cell lysates. To answer that, we expressed each of the TmaR mutants, as well as WT TmaR, all fused to a His tag as well as just the His tag, from a plasmid in ΔtmaR cells expressing WT EI-mCherry from the native pts promoter and locus in the chromosome. Of note, the effect of the plasmid-expressed His-tagged TmaR mutants on EI-mCherry localization (see Figure S7) is the same as that of the chromosome-expressed TmaR mutants shown in the previous section. Hence, independent of their expression level, TmaR mutants are unable to recruit EI to the poles. Following crosslinking, the His-tagged proteins were precipitated with nickel beads and fractionated by SDS-PAGE, and the amount of EI-mCherry that co-precipitated with them was monitored by western blot analysis. The fraction of EI-mCherry that co-purified with all TmaR mutants constituted only 0.15–0.4 of the EI-mCherry amount that co-purified with WT TmaR (Figure 3A). Hence, substituting each of the seven residues in TmaR, which were identified as important for EI recruitment to the poles, impairs TmaR association with EI in vivo.

Figure 3.

The screen-isolated mutations in TmaR and EI impair their interaction and the effect of isolated TmaR mutants on EI function

(A) CoIP of the TmaR variants with EI. Cells deleted for the tmaR gene and expressing EI-mCherry from the pts native promoter and locus in the chromosome, as well as one of the His-tagged TmaR variants or just the His tag from a plasmid, were crosslinked by formaldehyde. The His-tagged proteins were precipitated with nickel beads and fractionated by SDS-PAGE, and the amount of EI-mCherry that co-precipitated with them was monitored by western blot analysis using anti-mCherry antibodies. The membrane was then stripped and probed with anti-His antibody to detect the TmaR-His variants. The lower panel shows the levels of EI-mCherry and the His-tagged TmaR variants that were co-precipitated with it on a representative membrane. The upper panel shows the ratio between the band intensity of EI-mCherry expressed in the presence of or TmaR-His minus the band intensity of EI-mCherry expressed in the absence of TmaR-His and the intensities of His-tagged TmaR mutant bands. The ratio obtained when WT TmaR and WT EI are co-expressed is defined as 1. The bars show the SDs between three biological replicates. Statistical analysis for the differences between WT and the mutants was conducted using the two-sided Mann-Whitney test. ^∗p < 0.05.

(B) Far-western analysis of the interaction between EI and TmaR variants. Equal amounts of lysates of ΔtmaR cells overexpressing the His tag only or His-tagged WT TmaR or mutant proteins were fractionated on three SDS-polyacrylamide gels and blotted onto nitrocellulose membranes, which were probed with either anti-His antibody (western) or first with EI-mCherry or just mCherry (I) followed by anti-mCherry antibody (II). The observed bands match His-TmaR size. Of note, none of the TmaR variants interacted with the mCherry protein, which served as a negative control. The upper panel shows the ratio between the band intensity of EI-mCherry to His-tagged TmaR variant bands. The ratio obtained when WT TmaR and WT EI are co-expressed is defined as 1. The bars show the SDs between three biological replicates. Statistical analysis for the differences between WT and the mutants was conducted using the two-sided Mann-Whitney test. ^∗p < 0.05.

(C) The upper panel shows images of yeast cells co-expressing variants of both mCherry-tagged EI and Venus-tagged TmaR. From left to right: EI (WT) with TmaR (WT); EI (G266C) with TmaR (WT); EI (WT) with TmaR (F8L, V11I, R16P), termed TmaR (patch-I mut); and EI(WT) with TmaR (D87H, A84G, E85D), termed TmaR (patch-II mut). Scale bars, 5 μm. The lower panel shows the phase diagrams of yeast cells expressing the indicated E. coli EI and TmaR variant proteins as functions of median fluorescence intensities of individual cells in the RFP and GFP channels. Red and black dots indicate cells with or without a TmaR condensate, respectively.

(D) Examining the effect of the mutations in TmaR on EI function. A representative MacConkey plate supplemented with 0.4% fructose and containing 1 mM IPTG streaked with ΔtmaR cells harboring a plasmid that expresses WT TmaR, one of TmaR variants that cause EI mislocalization or the His tag only (EV). Red color indicates sugar consumption and, hence, that EI is functional. Cells deleted for the ptsI gene, which encodes EI (ΔptsI), grow as white colonies, providing a negative control.

(E) The rate of glucose consumption by cells deleted for the pts operon and for the tmaR gene and expressing each of the indicated His-tagged TmaR mutants, or by cells expressing only a His tag. The bars show the CVs between four biological repeats. Statistical analysis of the rate of glucose consumption between the mutant and WT TmaR proteins was conducted using the t test. ^∗∗∗p < 10⁻⁶, ^∗∗p < 0.01, and ^∗p < 0.05.

As another approach for testing the in vivo interaction of TmaR variants with WT EI, we used the bacterial-two-hybrid assay.³² For this assay, WT EI was expressed from one plasmid and either WT TmaR or one of its mutants were expressed from a second plasmid. The results, presented in Figure S8, demonstrate once again that WT TmaR interacts with WT EI, as indicated by the growth of blue colonies, with a comparable color intensity to that of the positive control (Zip-Zip). All TmaR mutants in patch I failed to interact with EI, whereas the patch-II mutants showed reduction in interaction with TmaR to various degrees. Although E85D was completely impaired for interaction with TmaR, the other three mutants showed reduced level of interaction compared to WT TmaR, with K95E showing the lowest affinity to EI. Together, the mutations in TmaR, which did not recruit EI, negatively affect the interaction with EI, with the effect of the patch-I mutants being the strongest.

To examine the effect of the mutations in TmaR on its ability to interact directly with EI we applied the far-western assay. For this purpose, we blotted equal amounts of lysates of ΔtmaR cells expressing the different His-tagged TmaR variants or only the His tag from a plasmid onto three nitrocellulose membranes. We then probed the membranes with either anti-His antibody to compare the expression level of the TmaR variants (Figure 3B, top membrane), or purified EI-mCherry followed by anti-mCherry antibodies to monitor interaction with EI (Figure 3B, middle membrane), or purified mCherry followed by anti-mCherry antibodies (Figure 3B, lower membrane), which served as a negative control). The level of the different TmaR-His variants varies, either due to different expression levels from a plasmid or to a difference in stability (Figure 3B upper membrane; Figure S6A). After quantification and statistical analysis, the far-western results show that the three mutations in patch I significantly impaired the ability TmaR to interact directly with EI, whereas the effect of the four mutations in patch II was weak in three cases, but not significant, indicating once the importance of patch I for direct interaction with EI, compared to the marginal effect of patch II.

The intriguing in vitro results with the mutants of TmaR patch II and the recent evidence about TmaR PS²¹ triggered us to examine the ability of EI and TmaR to interact in a phase-separated condensate in a heterologous non-bacterial cellular environment. Therefore, we turned to the new pipeline that maps high-resolution interaction-dependent phase diagrams of two proteins that co-phase separate in live yeast cells.³⁶ To this end, EI-mCherry and Venus-TmaR were cloned in separate low-copy centromeric plasmids to achieve stochastic expression of each protein; that is, different concentrations and ratios of the proteins are expressed in each yeast cell. The results in Figures S9A and S9B show that, when each protein was expressed alone, neither TmaR nor EI self-assembled to form phase-separated condensates in yeast, with a minor deviation from this rule by TmaR patch-II mutants. These results do not contradict the results obtained in E. coli cells, in which TmaR forms condensates in the absence of EI, since TmaR was shown to phase separate in E. coli by heterotypic interactions with RNA and possibly other factors,²¹ which are probably not present in yeast cells.

For the phase diagrams of the two proteins expressed together, we plotted the cells' median RFP and GFP intensities, with each dot representing a cell. The red dots are cells in which the EI and TmaR proteins are together in a condensate, whereas the black dots are cells without a condensate. Because PS prohibits a certain range of concentrations of the two components, it leads to a lack of data points at the center of the plot. This range of prohibited concentrations is delineated by the phase boundary, beyond which there is low density of red dots, which represent cells with condensates. In contrast, non-phase separating protein pairs are able to cover the entire range of concentrations along both axes and the phase boundaries are no longer clear.³⁶ The phase diagram obtained with the WT versions of EI and TmaR supports a two-component PS of the two proteins (Figures 3C, S9B, and S9C), providing an additional and different type of evidence for direct interaction between EI and TmaR. Notably, when the mutations in patch I (F8L, V11I, and R16P) or in patch II (D87H, A84G, and E85D) of TmaR were co-expressed with WT EI, or when the mutation in G266 of EI (G266C) was co-expressed with TmaR, the two proteins no longer co-phase separated to form a condensate (Figures 3C, S9B, and S9C). Not only do the phase diagrams obtained with EI and TmaR in yeast cells provide strong support for EI-TmaR interaction and for the stability of the complex but they also highlight the importance of patch II for the interaction in a cellular environment. The results further suggest that EI-TmaR interaction is likely to be direct, as EI and TmaR are not conserved in yeast, so the likelihood that a third partner, except for RNA, brings them together is very low. Moreover, disruption of the interaction in yeast by mutating the residues that came up in our screens supports the specific nature of the interaction and the accuracy of the predictions. Finally, the nature of the phase diagrams demonstrates that the interaction is concentration and partner dependent.

The effect of TmaR mutants that fail to interact with EI on the activity of EI

Bearing in mind that only the diffuse fraction of EI is active in sugar uptake and that cells lacking TmaR consume more sugar than cells expressing TmaR,^13,14 we examined the effect of the mutations in TmaR, which prevent EI polar sequestration, on sugar consumption. For this purpose, we streaked ΔtmaR cells expressing WT TmaR or one of TmaR mutants from a plasmid, as well as ΔtmaR cells harboring the vector only (EV) on MacConkey-fructose plates, with ΔptsI cells serving as a negative control. Cells expressing the three TmaR mutants with substitutions in patch I generated an intense red color, comparable to the color generated by cells lacking TmaR (EV), which is visibly more intense than the color generated by cells expressing WT TmaR (Figure 3D, upper plate). This observation suggests that the patch-I mutants are incapable of regulating EI activity due to their inability to interact with EI and sequester it at the poles. The four TmaR mutants with substitutions in patch II generated a red color that was either comparable to the color generated by cells expressing WT TmaR (D78H) or more intense (A84G, E85D, and K95E) (Figure 3D, bottom plate), suggesting that mutants that partially retained ability to interact with EI are also defective in regulating EI to some extent. The results on the indicator plates are in agreement with the quantification of glucose consumption rates (Figure 3E), which showed that EI activity is higher when it is expressed with two of the patch-I mutants of TmaR, F8L and TmaR V11I, than when expressed with WT TmaR. In the case of the patch-II mutants, only TmaR A84G led to a significantly higher sugar consumption by EI compared to WT TmaR.

The results in this section suggest that the three substitutions in patch I of TmaR lead to a complete loss in the ability to interact with EI and to sequester it at the poles and, hence, to an increase in EI-mediated sugar consumption, comparable to cells lacking TmaR. The patch-II mutants, which are only partially defective in recognizing and associating with EI, as indicated by their ability to interact with purified EI directly, also only partially affect EI activity. Hence, failure of TmaR mutants to interact with EI correlates with impairment in regulating its activity. Overall, our screen identified one region in TmaR that plays a critical role in the interaction with EI and a second region that is involved in this interaction. Because mutants in patch II impaired the ability of TmaR to co-phase separate in yeast cells, the interaction mediated by this patch might be strengthened by additional factors and/or by biochemical constituents in the intracellular environment.

Applying the results of the screens to predict EI-TmaR 3D molecular interaction

Our screens identified G266 in EI and seven residues located in two patches on TmaR as important for EI-TmaR interaction in vivo. To reveal the structural details of this interaction, we used ClusPro (https://cluspro.bu.edu)^37,38,39,40 to dock TmaR to EI. For this, we used the solved EI dimer structure (PDB: 2HWG^24,41) and the TmaR structural model (predicted by trRosetta; see Figure 2C) without considering the results of our screens. The four models shown in Figure S10, which represent the top six ClusPro models (the two models that are not shown are symmetric copies of two models that are shown) suggest that patch I in TmaR may stick to different regions in EI and, hence, we termed this patch, which includes two adjacent hydrophobic residues (F8 and V11), the “sticky patch.” However, except for model #03, the other ClusPro models predict interaction with EIN, in contrast to our previously published experimental results that showed that EIC and not EIN has the capacity to localize EI to the cell poles.¹⁸ Moreover, the four models in Figure S10, including #03, fail to explain the involvement of patch II of TmaR in the interaction. Of note, three out of four mutations in patch II of TmaR are replacement of charged amino acids (D78, E85, and K95), with two of them replaced by a residue with an opposite charge. We therefore focused on the model generated by ClusPro that gave the best score for electrostatic-favored interactions (Figure 4A). This model is similar to model #03 in Figure S10 but predicts electrostatic attraction of patch II of TmaR to EI, since the EI surface electrostatic potential in this model features a negative patch near TmaR K95 in patch II of EI (colored magenta in Figure 4B). We therefore termed patch II the “stabilizing patch.” Notably, the ClusPro predictions for EI-TmaR molecular interaction, which suggest that patch I is crucial for the interaction, whereas patch II is secondary, agree with our results, which show that patch II is crucial for EI-TmaR interaction in vivo (Figure 3A) but does not seem to mediate the interaction (Figure 3B). Taken together, the experimental data and the ClusPro predictions suggest that the model shown in Figures 4A, 4B, and 4C is likely to exist in vivo.

Figure 4.

Computational modeling of EI-TmaR interaction and experimental validation by mutating EI residues predicted to be involved in the interaction with TmaR

(A–C) EI-TmaR interaction model obtained by ClusPro protein-protein docking,^37,38,39,40 using the solved structure of the EI dimer (see Figure 1E) and the structural model of TmaR (see Figure 2C). Shown is the highest-rated electrostatic-favored interaction model that agrees best with our experimental data (see text). TmaR is in cyan, with the residues in the sticky and stabilizing patches substituted in the mutants obtained in our screen colored in orange and magenta, respectively, and their protruding side chains presented as sticks. (A and C) EI dimer is in shades of gray, with one subunit in dark gray and the other in light gray. The residues in EI and TmaR predicted by RaptorX ComplexContact to coevolve with the highest scores (>0.3) are presented as spheres: magenta in TmaR stabilizing patch, orange in TmaR sticky patch, and blue and green in EI two subunits, respectively. EI G266 is marked as red sphere in (A). (B) The electrostatic potential of the surface of the EI dimer structure. Positively and negatively charged regions are highlighted in shades of red and blue, respectively, according to the heatmap shown below. Arrows point at mutations in the stabilizing patch. (C) Magnified view of the EI-TmaR interaction region.

(D) Major motions of EI, as suggested by normal mode analysis (generated using the Anisotropic Normal Mode [ANM] server, anm.csb.pitt.edu). The EI dimer solved structure was provided as input (see Figure 1E). Shown is mode n = 3, which shows how EI opens and closes the binding site for TmaR, involving a hinge movement centered near G266. Major movements are in red. See also Video S3.

(E) Images of mCherry-tagged WT EI or one of its variants. Scale bars, 5 μm.

(F) Boxplot showing mean intensity (MI) of the mCherry signals calculated for cells expressing mCherry-tagged WT EI or one of its variants, all expressed from the native ptsI promoter and locus in the chromosome. Green triangles indicate mean values.

(G) Boxplot showing the CVI of the mCherry signals calculated for cells expressing mCherry-tagged WT EI or one of its variants, all expressed from the native ptsI promoter and locus in the chromosome. Green triangles indicate the mean values.

(H) Bar plot showing the fraction of cells with clusters of mCherry-tagged WT EI or EI mutants with single point mutations, all expressed from the pts native promoter and locus in the chromosome. The bars show the standard deviations. Statistical analyses for the differences between WT and the mutants were conducted using the unpaired t test based on nine fields. ^∗∗∗p <10⁻⁹.

(F–H) n is between 1,043 and 1,574 cells.

We attempted to model EI-TmaR interaction also by AlphaFold, an artificial intelligence (AI) system that predicts 3D structures of proteins (https://alphafold.ebi.ac.uk).⁴² The EI structure predicted by AlphaFold is similar to the solved closed structure determined by NMR (PDB: 2N5T) shown in Figure S2B. Despite some differences in TmaR predictions, the same interaction interfaces are predicted in all top four models generated by AlphaFold (Figure S11A). These are also the same interaction interfaces predicted as a top rank by ClusPro (Figure S10A). Of note, in the three top-ranked models generated by AlphaFold, TmaR is predicted to fold into a coiled coil, as in the model predicted for TmaR by trRosetta (Figure S11B). However, the AlphaFold models are less likely to exist than the ClusPro model, on which we chose to focus (Figure 4A), for the following reasons: (1) they predict interaction with EIN rather than with EIC; (2)they fail to explain the involvement of patch II of TmaR in the interaction; (3)residues that coevolved in both proteins are spatially far from the predicted interaction interfaces (see below). Hence, despite the high accuracy achieved by AlphaFold for the prediction of proteins’ 3D structure, it seems that more data on protein-protein interactions are required to achieve a similar level of accuracy in predicting 3D interactions between proteins by an AI system.

Support for the suggested model is provided by the residue identified in EI to be involved in EI-TmaR binding, namely G266. Since G266 in EI did not come up in EI-TmaR coevolution analysis (see below), nor is it in close proximity to TmaR in our proposed EI-TmaR interaction model, and, because it is located in a region that connects the two domains of EI (Figure 4A), we asked if G266 is of importance for EI flexibility, which in turn would be required for binding to TmaR. To this end, we applied normal mode analysis of the EI dimer structure (using the Anisotropic Normal Mode [ANM] server at anm.csb.pitt.edu⁴³), which identifies the major motions that are possible in a protein while disturbing the protein structure minimally. Model #3 suggests how EI could open and close the binding site for TmaR, with G266 serving as a hinge joint that provides the flexibility required for this motion (Figure 4D; Video S3), even though it is located far away from the binding site. Replacement of this small glycine residue by the bulkier residues that came up in our screen (cysteine, serine, or aspartate) is expected to destabilize the turn where it is located and dramatically influence binding to TmaR, explaining why it came up in our screen for residues that are required for the interaction with TmaR time and again.

Video S3. Major motions of EI

A video showing the major motions of EI, as predicted by normal mode analysis, using the ANM server (anm.csb.pitt.edu), based on 2HWG (see Figure 1E). Shown is mode #3, which demonstrates how EI opens and closes the binding site for TmaR.

To further validate our model for EI-TmaR molecular interaction, we chose four pairs of residues in the two proteins that are predicted by our preferred model to interact electrostatically. The putative interactions between the selected residues in EI, E109, EE116/117, R366, and D486, and their predicted interacting partners in TmaR, R98, R95, N22, and R19, are shown in Figure S12. Of note, two of the four residues in EI are predicted to interact with residues in patch I of TmaR, and the other two with residues in patch II. Additionally, R95 in TmaR may interact electrostatically with either E116 and/or E117 in EI. We then constructed mCherry-tagged EI mutants with an opposite electrical charge to the native charge in the positions that we chose, generating E109K, E116K, E117K, R366D, and D468R. Next, we monitored the localization of the new EI variants compared to WT EI (Figure 4E). Importantly, the expression levels of the mCherry-tagged EI mutants were all similar to that of EI-mCherry, as judged by the comparable mCherry mean intensities (Figure 4F). Although all the EI mutants could create clusters (Figure 4E), they all exhibited a bigger fraction that is diffused throughout the cell, as indicated by the lower coefficient of variance (CV) intensity of the fluorescence signal in the EI mutants compared to that of the WT EI (Figures 4E and 4G). Moreover, the EI mutants had significantly fewer cells with EI clusters than WT EI (Figure 4H). These results suggest that the EI residues that we chose, based on our preferred 3D model, play a role in EI-TmaR interaction, although the interaction does not rely solely on either one of them; therefore, these mutants could not be isolated in our screen.

Evolution conservation of EI-TmaR interaction and colocalization

Additional support for our preferred model for EI-TmaR molecular interaction was obtained by analyzing the coevolution of the TmaR and EI proteins in bacteria, using the RaptorX ComplexContact server^33,34,35 (Figure 5A). Eight out of the nine residues in TmaR, which coevolved with highest confidence (scores > 0.3) with residues in EI (Table 1), map to the two patches of TmaR identified in the screen as important for the interaction with EI (Table S2; see orange and magenta spheres in Figure 4A). Moreover, three TmaR residues that strongly coevolved with EI, F8, V11, and A84 were identified in our screen as important for recruiting EI to the poles. In our model, coevolving pairs, involving residues V15 and R19 in patch I of TmaR, are located proximal to EI residue P353 and nearby E351 and L355 of both subunits of the EI dimer (see green and blue spheres in Figures 4A and 4C). Of note, these EI residues might also play a role in EI dimerization, which is essential for its autophosphorylation,⁴⁴ explaining why we could not isolate mutations in these residues, since only cells with functional EI were chosen for this screen. Still, it is reassuring that the residues in EI that coevolved with TmaR are in close proximity to the TmaR sticky patch in our model (Figure 4C). Hence, the coevolution analysis of TmaR and EI further supports our proposed model for the EI-TmaR interaction.

Figure 5.

Conservation of EI-TmaR interaction in bacteria

(A) The results of coevolution of TmaR and EI proteins in bacteria produced by RaptorX ComplexContact. Shown is a heatmap of the probabilities of two residues being in contact (i.e., their minimum backbone non-hydrogen atom distance falling in the range [0.6 Å]). Higher probabilities are represented by darker color.

(B) Images of ΔtmaR ΔptsI E. coli cells expressing TmaR-YFP (yellow) and EI-mCherry (red) from plasmids, both originating from E. coli or from the S. typhimurium as well as each originating from a different species. Scale bars, 2 μm.

(C) Images of S. typhimurium cells expressing TmaR (YFP tagged, green) and EI (mCherry tagged, red) both from E. coli (EC). Scale bars, 2 μm.

(D) Images of ΔtmaR and ΔptsI E. coli cells expressing EI-mCherry (red) originating from E. coli (upper images) or from S. typhimurium (lower images) from plasmids. Scale bars, 2 μm.

Table 1.

The amino acid pairs that coevolve in EI and TmaR, obtained by coevolution analysis

Coevolved amino acid positions		Strength
EI	TmaR
353	15	0.37
389	51	0.36
141	8∗	0.36
353	84∗	0.33
353	19	0.33
141	11∗	0.33
351	19	0.32
353	80	0.32
355	15	0.30

Coevolution of EI and TmaR was analyzed by RaptorX ComplexContact.^45,46,47,48 The position of residues predicted to coevolve in EI and TmaR with the most significant scores (>0.3) are presented. Asterisks indicate residues identified in our screen (Table S2). In italics, residue pairs located in proximity in our interaction model (Figures 4A–4C).

For a broader perspective on the conservation of the interaction between EI and TmaR, we asked whether it occurs also in other species that express the two proteins. Hence, we asked if TamR from Salmonella typhimurium can recruit E. coli EI to the polar cluster and vice versa. To address this question, we expressed mCherry-tagged EI and YFP-tagged TmaR of either E. coli or S. typhimurium from plasmids in E. coli deleted for the tmaR and ptsI genes. We observed that the EI and TmaR from S. typhimurium co-localized to the E. coli cell poles (Figure 5B, right panel). Likewise, EI and TmaR from E. coli co-localized to the S. typhimurium cell poles (Figure 5C). More importantly, the S. typhimurium EI was recruited to the polar clusters of E. coli TmaR and vice versa (Figure 5B). Notably, in cells deleted for tmaR, EI of both species was detected as homogeneously spread through the cytoplasm (Figure 5D). Comparison of the TmaR and EI protein sequences from E. coli and S. typhimurium (Figure S13) shows no difference between the species in the positions that were identified in our screens as important for EI-TmaR interaction.

Taken together, the strong coevolution and conservation of the residues that are important for EI-TmaR interaction and colocalization, and the preservation of their ability to interact and colocalize in other bacterial species, highlight the importance of EI-TmaR interaction during evolution.

Additional evidence supporting the predicted model for EI-TmaR 3D molecular interaction

Our working model predicts that TmaR interacts with the closed conformation of EI dimer (Figure 4A). The model further predicts that TmaR stabilizes EI, because the TmaR sticky patch interacts with EIC and then the complex is stabilized by an electrostatic interaction between K95 in the TmaR stabilizing patch and E117 in EIN (Figure S14). These interactions are not predicted by ClusPro to co-occur with the open dimer. To check these predictions, we applied chemical crosslinking to ΔtmaR cells that express EI-mCherry from the chromosome and the different His-tagged TmaR variants, either the WT or the mutants that fail to recruit EI, or only the His tag, from a plasmid. We then blotted equal amounts of lysates onto a nitrocellulose membrane and probed it with anti-mCherry antibodies (Figure 6A). Only cells expressing WT TmaR, but not the mutant TmaR proteins, presented a distinct band that corresponds to the size of EI dimer. Importantly, the same phenomenon of losing the EI dimer band was also observed with the three mutants of EI G266 replacement (Figure 6B). Together, these results support the hypothesis that the interaction between EI and TmaR stabilizes the EI dimeric form.

Figure 6.

A suggested mechanism for the control of EI by TmaR and evidence supporting it

(A and B) Western blots of chemically crosslinked ΔtmaR cells expressing EI-mCherry from the chromosome and His-tagged TmaR variants or only the His tag from a plasmid (A) or mCherry-tagged EI variants from the chromosome and the His-tagged TmaR or only the His tag from a plasmid (B). The estimated molecular weight (MW) of monomeric EI-mCherry is 91 kDa. The monomer detected by anti-mCherry antibodies is marked by a blue arrow and the putative dimer by an orange arrow. The bar plots represent the calculated ratios between the dimeric and the monomeric forms in each lane, based on three biological replicates. The error bars represent the SD from three biological replicates. Statistical analysis for the differences between cells expressing WT EI and WT TmaR to the mutants were calculated by t test ^∗p < 0.05, ^∗∗p < 0.01.

(C) Schematic presentations of (I) EI dimer closed conformation (EIC is in blue and EIN is in gray). (II) EI dimer open conformation, which allows EI binding to HPr (olive green). (III) An interaction model of TmaR (cyan) with the EI dimer: TmaR sticky patch (orange) interacts with EIC of one EI monomer, whereas K95 in TmaR (purple X) interacts with EIN of the second EI monomer. (IV) The interaction with TmaR does not allow the second monomer EI to open and bind to HPr.

To summarize, our working model (Figure 4A) explains a significant part of the experimental findings, both in terms of the effects of individual mutations as well as interacting residue pairs. The fact that none of the docking models could account for the entire array of the experimental results, in particular for the mutations in patch II, suggests that some of the interactions might involve a different conformation of EI or require a third protein.

Discussion

Studies published in the last couple of decades have shown that, despite their small size and the scarcity of membrane-bound organelles, bacterial cells are highly organized (see Introduction). Still, the spatiotemporal distribution pattern of only a small fraction of proteins in leading bacterial models, such as E. coli, B. subtilis, and Caulobacter crescentus, has been characterized in detail. In recent years, tools to enable high-throughput imaging of protein localization in bacteria, such as special slides and agar pads, have been developed and applied to depict bacterial localizomes.^49,50 Efforts have been made to obtain a genome-scale quantitative characterization of the localizome dynamics.^51,52 However, unlike in eukaryotes, screens to correlate genetic perturbations with changes in protein localization in bacteria are considered challenging due to the difficulty of obtaining high-resolution images near the diffraction limit. Here we show that automated large-scale mutant screens can be successfully pursued to distinguish pole-localized from cytoplasmic delocalized signals. Moreover, we could carry out reciprocal screens by randomly mutagenizing two interacting proteins, one at a time, to determine specific residues that are required for the interaction in each of them. Notably, mutating these residues impairs EI-TmaR interaction and colocalization not only in bacteria but also in yeast cells. The acquired information was used, together with the solved structure of one partner and the predicted structure of the other to predict their mode of interaction. The validity of our interaction model was reinforced by coevolution analysis, by the conservation of residues that are predicted to interact in the two proteins, and by colocalization of EI and TmaR in other species. Together, the results highlight the importance of controlling the subcellular localization of the major bacterial sugar metabolism factor during evolution. Notably, our approach can be applied to predict molecular interactions between other proteins in bacteria.

The role of G266 in EI, which came up time and again in our screen, although it is not in direct interaction with TmaR, is very much in agreement with the role predicted for the helical linker (residues 231–260), which precedes G266, in the rotation of EIN versus EIC, based on EI crystal structure and its solution analysis by NMR.^24,25,26,27 Notably, glycines confer conformational flexibility to peptide backbones^28,29 to the extent that even addition of a single methyl group to their side chain (e.g., substituting it with an alanine) has been shown to lead to substantial reduction in flexibility of the polypeptide backbone.⁵³ The suggested swiveling of the two domains relative to each other, which is needed for orienting the phosphorylated histidine in EI and the HPr-binding domains away from the PEP-binding domain, as well as for inserting TmaR between two subunits of the EI dimer, as predicted by our interaction model, require a wide range of rotation angles between the domains (see Figure S2). It has been suggested that loose packing of the linker against the rest of the structure confers the required flexibility.²⁴ Our finding that G266 acts as a hinge joint for the swivel, which is backed by the algorithm that identifies EI’s major possible motions that minimally disturb its structure (Video S3), sheds new light on the structural elements that provides EI with the dynamics required to promote large-scale rearrangement of the domains.

Based on the many studies of EI structure cited above, the catalytic cycle of EI is accompanied by conformational changes that enable its transition between a closed conformation that allows autophosphorylation by PEP, an intermediate partially closed conformation, and an open conformation that allows HPr phosphorylation (see Figure S2). Our results imply that TmaR interacts with the closed conformation of EI for three reasons. First, as we stated above, TmaR sticky patch interacts with EIC, and the complex is then stabilized by an electrostatic interaction between K95 in TmaR stabilizing patch and E117 in EIN (Figure S14). These interactions are not predicted to co-occur with the open dimer. Second, a closer look at the predicted interaction model reveals that each of the patches in TmaR that are predicted to interact with EI, the sticky and the stabilizing, interacts with both monomers of the EI dimer, and this interaction stabilizes EI closed conformation or is enabled by it (Figures 6C and S14). Third, an apparent accumulation of WT EI dimer is observed in cells expressing WT EI and WT TmaR, but not in cell expressing TmaR mutants, which were isolated in our screen as not recruiting EI (Figure 6A), or in cells expressing EI mutants isolated in our screen as not recruited to TmaR (Figure 6B). These results imply that binding of EI to TmaR is mutually exclusive with binding of EI to HPr, since HPr interacts with the open conformation of EI (Figures 6C, S2, and S14). Interestingly, AlphaFold-generated models also predict the same mutual exclusiveness. This is in line with our previous findings that TmaR-associated EI is inactive in sugar uptake,^13,14 and that EI and HPr localize in the cell independently,¹⁸ with EI but not HPr localization depending on TmaR.²¹ Since HPr needs to be phosphorylated by EI to be active in sugar uptake, which is compatible with the finding that HPr is released from the pole in an EI-dependent manner,¹⁸ including EI in the membraneless organelle formed by TmaR guarantees its inactivity not only by secluding it but also by keeping it in a state that cannot bind to HPr and phosphorylate it. Our results cannot determine whether EI must be in a phosphorylated state when in a condensate with TmaR, since we screened only for active EI, thus preventing the isolation of EI mutants that cannot get phosphorylated. However, regardless of EI phosphorylation state in the polar condensates, the mode of its association with TmaR impedes the phosphorylation cascade, composed of EI, HPr, and the sugar permeases, which permits sugar uptake. An important issue that needs to be fully resolved in future studies is the effect of G266 substitutions in EI on its affinity and kinetics of interaction with HPr and TmaR. Solving the structure of EI-TmaR complex will corroborate the effect of the different mutations in EI and TmaR on their interaction.

PS-driven condensates typically contain scaffold and non-scaffold molecules. In the case of TmaR and EI, TmaR serves as the scaffold that recruits EI and induces its phase transition.²¹ Although the expression levels of non-scaffolds may affect phase behavior of the scaffold protein,⁵⁴ EI overexpression does not affect TmaR PS, so in this case the effect is unidirectional from TmaR to EI.¹⁴ Since we screened for mutations in TmaR that do not affect its ability to phase separate, but only its ability to recruit EI to its condensates, the nature of the mutations can only shed light on how TmaR interacts with EI rather than on the interaction between TmaR molecules. In this regard, the involvement of F8 and R16 in TmaR, which came up in our screen, suggests that the interaction between TmaR and EI complies with the stickers-and-spacers theory. This theory states that interaction between proteins that undergo PS is mediated by reversible noncovalent interactions between short multivalent motifs, termed stickers, which are presented as patches on the surface of folded proteins and whose valence impacts the driving forces for phase transition.^55,56,57 Non-sticker regions on the surface of proteins, termed spacers, may only weakly affect PS.⁵⁶ Based on studies of proteins that undergo PS, the aromatic phenylalanine and the positively charged arginine are effective stickers that often drive PS by forming electrostatic interactions.^57,58 Besides the F8 and R16, whose substitution impaired recruitment of EI, there are two additional phenylalanines and three arginines in the first 20 residues of TmaR. Since the number of residues that may act as stickers, that is, the summed valence, governs saturation concentration and, thus, PS, we hypothesize that this region in TmaR forms reversible noncovalent interactions with EI, thus affecting EI phase behavior. Of note, the first residues at TmaR N terminus are predicted to form a disordered region (IDR) and, as such, are considered a sticker domain that promotes PS.⁵⁶ It might very well be that some molecules of TmaR interact with additional constituents of the condensate, for which TmaR serves as a scaffold that fosters their phase transition, via this domain. However, the strong evidence for EI-TmaR co-PS in a heterologous eukaryotic cell system implies that these putative constituents of the TmaR condensates are conserved through the different kingdoms of living cells.

TmaR was previously reported by us to be phosphorylated on tyrosine 72, a modification that is essential for TmaR clustering at the poles.¹⁴ In a recent study, tyrosine phosphorylation was shown to be crucial for TmaR PS,²¹ consistent with the involvement of phosphorylation in PS of eukaryotic proteins,⁵⁹ and for its ability to control bacterial motility.²¹ Relevant for this study, mutants in which Y72 or all three tyrosines in TmaR were replaced by phenylalanine could no longer regulate EI activity, although they were still able to interact with EI in vitro.¹⁴ Hence, such mutants could not be isolated by our screen, because they are diffused throughout the cytoplasm. Of note, the phosphorylation event, although not part of results used for predicting the interaction model, are compatible with it. Anyway, it seems that TmaR tyrosine phosphorylation is more relevant for TmaR and its role in condensate formation than for its mode of interaction with EI.

The transient nature of the interactions within the condensates, in general, and between TmaR and EI, in particular, provides a possible explanation for the difference between the in vitro and in vivo results, that is, the reduced ability of TmaR mutants in patch II (the stabilizing patch) and of EI G266 mutants to interact with the WT EI and WT TmaR, respectively, in vivo, although they interact in vitro. It might very well be that these weak electrostatic interactions take place in the context of the certain concentration and the degree of liquidity that occur in the membraneless organelle but not in vitro. Although, by and large, the results obtained by the different interaction assays convey the same message, the differences in the interaction of EI with TmaR that bears mutations in patch II may be explained by the relative levels in which the two proteins are expressed. In the coIP assay, only TmaR was overexpressed from a plasmid; in the two-hybrid and the far-western assays, both EI and TmaR were overexpressed; and, in the yeast co-PS assay, EI and TmaR were expressed stochastically at different concentrations and ratios. Interestingly, an alternative explanation is that EI-TmaR interaction in vivo involves a third partner, which either brings them together and/or stabilizes the supramolecular complex. However, this explanation is less likely, since EI and TmaR are not conserved in eukaryotes, so expression of this third partner in yeast is very dubious. Rather, the finding that each of the proteins expressed alone did not undergo PS in yeast, whereas their co-expression generated an unblemished phase diagram, implies that the presence of TmaR and EI is both necessary and sufficient to undergo PS. In case there is a third protein, its expression noise should be quite low, as the phase boundary in the yeast phase diagram is quite sharp. In either case, our results support the notion that it is not the localization per se that influences EI activity but its interaction with TmaR and its inclusion in the membraneless organelle.

More broadly, our study uncovers the importance of spatial regulation of signal transduction pathways in bacteria, particularly that of sugar metabolism. On the practical side, it demonstrates how high-throughput screening combined with 3D molecular modeling and coevolution analysis can shed light on modes of interaction between proteins and suggests a new approach to uncover interaction-mediated regulation of proteins.

Limitations of the study

Two limitations of our screens are that they were not fully saturated and that the mutant selection was performed manually, so only mutants in which EI was fully diffused were selected for further analysis, thus ignoring intermediate interactions. Additionally, because we screened for EI mutants that fail to interact with TmaR but are still active, we cannot conclude whether EI may interact and phase separate with TmaR when in its unphosphorylated and inactive state. Although our results indicate that EI and TmaR interact directly, we cannot rule out the involvement of a third partner, which is not absolutely required for the interaction but stabilizes it. Additionally, because either one or both proteins were expressed from a plasmid in the in vivo interaction assays, we cannot estimate TmaR:EI stoichiometry in vivo. Finally, our study did not establish the affinity and kinetics of EI-TmaR interaction, nor is our EI-TmaR interaction model verified by structural analysis of the complex yet.

Resource availability

Lead contact

Further information and requests should be directed to the lead contact, Orna Amster-Choder (ornaam@ekmd.huji.ac.il).

Materials availability

Bacterial strains and plasmids will be made available upon request.

Data and code availability

• •This paper does not report new or original code.
• •The raw data for the manuscript is available upon request from the lead contact Prof. Orna Amster-Choder (ornaam@ekmd.huji.ac.il).
• •Any additional information required to reanalyze the data reported in this paper is available from the lead contact upon request.

Acknowledgments

We thank Dan Tawfik, Lianet Noda-Garcia, and members of the Amster-Choder lab for fruitful discussions and suggestions. Research in the O.A.-C. lab was supported by the Israel Science Foundation (ISF) founded by the Israel Academy of Sciences and Humanities (grant no. 1274/19). O.A.-C. is an incumbent of the Dr. Jacob Grunbaum Chair in Medical Sciences. Research in the M.S. lab was supported by a Volkswagen (VW) Foundation LIFE grant 93092. M.S. is an incumbent of the Dr. Gilbert Omenn and Martha Darling Professorial Chair in Molecular Genetics. Research in the O.S.-F. lab was supported by ISF grant no. 301/2021. E.D.L. acknowledges support from the European Research Council (ERC) under the European Union’s Horizon 2020 research and innovation program (grant agreement no. 819318), by the ISF (grant no. 1452/18), and by the Abisch-Frenkel Foundation.

Author contributions

O.A.-C. and N.A.-K. designed the study. A.N.-S. and N.A.-K. generated the strains and plasmids. N.A.-K., M.H., A.F., E.S., and O.G. conducted the experiments. N.A.-K., M.H., E.S., and O.G. analyzed the data. N.A.-K., A.N.-S., E.D.L., O.S.-F., M.S., and O.A.-C. contributed to data interpretation. N.A.-K. and O.A.-C. wrote the manuscript. O.A.-C., E.D.L., O.S.-F., and M.S. supervised the study, provided advice, and contributed to the experimental design.

Declaration of interests

The authors declare no competing interests.

Declaration of generative AI and AI-Assisted technologies

During the preparation of this work, the author(s) used no AI technologies.

STAR★Methods

Key resources table

REAGENT or RESOURCE	SOURCE	IDENTIFIER
Antibodies
Anti mCherry	Abcam	Cat#ab167453 RRID:AB_2571870
Anti GroEL	Abcam	Cat#ab90522 RRID:AB_2049715
His-tag	Genscript	Cat#A00612 RRID:AB_915573
Bacterial strains
MG1655 (WT)	Lopian et al.60	N/A
Δpts-operon MG1655, Δpts::kan	Lopian et al.18	N/A
EI-mCherry MG1655 ptsI-mCherryϕkanR	This work	N/A
NA-EI(Q87H)-mCherry MG1655 ptsI(Q87H)-mCherryϕkanR	This work	N/A
NA-EI(R186H)-mCherry MG1655 ptsI(R186H)-mCherryϕkanR	This work	N/A
NA-EI(A261F)-mCherry MG1655 ptsI(A261F)-mCherryϕkanR	This work	N/A
NA-EI(G266C)-mCherry MG1655 ptsI(G266C)-mCherryϕkanR	This work	N/A
NA-EI(G266S)-mCherry MG1655 ptsI(G266S)-mCherryϕkanR	This work	N/A
NA-EI(G266D)-mCherry MG1655 ptsI(G266D)-mCherryϕkanR	This work	N/A
NA-EI(T277A)-mCherry MG1655 ptsI(T277A)-mCherryϕkanR	This work	N/A
NA-EI(P308S)-mCherry MG1655 ptsI(P208A)-mCherryϕkanR	This work	N/A
NA-EI(A323V)-mCherry MG1655 ptsI(A323V)-mCherryϕkanR	This work	N/A
NA-EI(A538H)-mCherry MG1655 ptsI(A538H)-mCherryϕkanR	This work	N/A
NA-TmaR-YFP_EI-mCherry MG1655, tmaR-YFPϕcat, ptsI-mCherryϕkanR	This work	N/A
NA-TmaR-YFP_EI(G266C)-mCherry MG1655, tmaR-YFPϕcat, ptsI(G266C)-mCherryϕkanR	This work	N/A
NA-TmaR-YFP_EI(G266S)-mCherry MG1655, tmaR-YFPϕcat, ptsI(G266S)-mCherryϕkanR	This work	N/A
NA-TmaR-YFP_EI(G266D)-mCherry MG1655, tmaR-YFPϕcat, ptsI(G266D)-mCherryϕkanR	This work	N/A
NA-EI-mCherry_TmaR(F8L)-YFP MG1655 ptsI-mCherryϕkanR, tmaR(F8L)-YFPϕcat	This work	N/A
NA-EI-mCherry_TmaR(V11I)-YFP MG1655 ptsI-mCherryϕkanR, tmaR(V11I)-YFPϕcat	This work	N/A
NA- EI-mCherry_TmaR(L12M)-YFP MG1655 ptsI-mCherryϕkanR, tmaR(L12M)-YFPϕcat	This work	N/A
NA-EI-mCherry_TmaR(R16P)-YFP MG1655 ptsI-mCherryϕkanR, tmaR(R16P)-YFPϕcat	This work	N/A
NA-EI-mCherry_TmaR(D78H)-YFP MG1655 ptsI-mCherryϕkanR, tmaR(D78H)-YFPϕcat	This work	N/A
NA-EI-mCherry_TmaR(A84G)-YFP MG1655 ptsI-mCherryϕkanR, tmaR(A84G)-YFPϕcat	This work	N/A
NA-EI-mCherry_TmaR(E85D)-YFP MG1655 ptsI-mCherryϕkanR, tmaR(E85D)-YFPϕcat	This work	N/A
NA-EI-mCherry_TmaR(K95E)-YFP MG1655 ptsI-mCherryϕkanR, tmaR(K95E)-YFPϕcat	This work	N/A
ΔtmaR MG1655, ΔtmaR:kanR	This work	N/A
ΔtmaR EI-mCherry MG1655, ptsI-mCherry, ΔtmaR:kanR	This work	N/A
OG543 MG1655, tmaR- HisϕkanR,	This work	N/A
NA-TmaR(F8L)- His MG1655, tmaR(F8L)- HisϕkanR,	This work	N/A
NA-TmaR(V11I)- His MG1655, tmaR(V11I)- HisϕkanR,	This work	N/A
NA- TmaR(L12M)- His MG1655, tmaR(L12M)- HisϕkanR,	This work	N/A
NA-TmaR(R16P)- His MG1655, tmaR(R16P)- HisϕkanR,	This work	N/A
NA-TmaR(D78H)- His MG1655, tmaR(D78H)- HisϕkanR,	This work	N/A
NA-TmaR(A84G)- His MG1655, tmaR(A84G)- HisϕkanR,	This work	N/A
NA-TmaR(E85D)- His MG1655, tmaR(E85D)- HisϕkanR,	This work	N/A
NA-TmaR(K95E)- His MG1655, tmaR(K95E)- HisϕkanR,	Szoke et al.21	N/A
NA-TmaR-YFP MG1655, tmaR-YFPϕcat	This work	N/A
NA-TmaR(F8L)-YFP MG1655, tmaR(F8L)-YFPϕcat	This work	N/A
NA-TmaR(V11I)-YFP MG1655, tmaR(V11I)-YFPϕcat	This work	N/A
NA- TmaR(L12M)-YFP MG1655, tmaR(L12M)-YFPϕcat	This work	N/A
NA-TmaR(R16P)-YFP MG1655, tmaR(R16P)-YFPϕcat	This work	N/A
NA-TmaR(D78H)-YFP MG1655, tmaR(D78H)-YFPϕcat	This work	N/A
NA-TmaR(A84G)-YFP MG1655, tmaR(A84G)-YFPϕcat	This work	N/A
NA-TmaR(E85D)-YFP MG1655, tmaR(E85D)-YFPϕcat	This work	N/A
NA-TmaR(K95E)-YFP MG1655, tmaR(K95E)-YFPϕcat	This work	N/A
NA-EI-mCherry_TmaR(F8L) MG1655 ptsI-mCherryϕkanR, tmaR(F8L)	This work	N/A
NA-EI-mCherry_TmaR(V11I) MG1655 ptsI-mCherryϕkanR, tmaR(V11I)	This work	N/A
NA-EI-mCherry_TmaR(L12M) MG1655¬ ptsI-mCherryϕkanR, tmaR(V11I)	This work	N/A
NA-EI-mCherry_TmaR(R16P) MG1655 ptsI-mCherryϕkanR, tmaR(L12M)	This work	N/A
NA-EI-mCherry_TmaR(D78H) MG1655 ptsI-mCherryϕkanR, tmaR(D78H)	This work	N/A
NA-EI-mCherry_TmaR(A84G) MG1655 ptsI-mCherryϕkanR, tmaR(A84G)	This work	N/A
NA-EI-mCherry_TmaR(E85D) MG1655 ptsI-mCherryϕkanR, tmaR(E85D)	This work	N/A
NA-EI-mCherry_TmaR(K95E) MG1655 ptsI-mCherryϕkanR, tmaR(K95E)	This work	N/A
Salmonella typhimurium SL1344	OAC lab collection	N/A
Chemicals, peptides, and recombinant proteins
Bacto™ Yeast Extract	BD	212750
Bacto™ Agar	BD	214010
Sodium chloride (NaCl)	Bio-lab	001903059100
Bacto™ Tryptone	BD	211705
Bacto Difco Casamino Acids	BD	223050
BD Difco™ M9 Minimal Salts	BD Biosciences	248510
Difco MacConkey Agar base	BD Biosciences	212123
MgSO4	Sigma-Aldrich	M5921
Thiamin (vitamin B1)	Sigma-Aldrich	T4625
Glucose	Sigma-Aldrich	16325
Fructose	Sigma-Aldrich	F2543
Sucrose	Sigma-Aldrich	Cat#S0389
Ampicillin	AppliChem Panreac	A0389
Chloramphenicol	Sigma-Aldrich	C0378
Kanamycin sulfate	Biological Industries	25389-94-0
X-Gal	Ornat	1758–0300
Tris	Bio-lab	002009239100
Glycine	Sigma-Aldrich	G8898
Sodium lauroyl sarcosinate (Sarkosyl)	Sigma-Aldrich	L9150
GeneMorph II Random Mutagenesis Kit	Agilent	200550
InstantBlue® Protein Stain	Expedeon	Cat#ISB1L
Ponceau S	Sigma-Aldrich	Cat#P7170
DpnI	NEB	R0176S
AatII	NEB	R0117S
AvrII	NEB	R0174S
AccI	NEB	R0161S
XhoI	NEB	R0146S
SalI	NEB	R3138S
BamHI	NEB	R3136S
KpnI	NEB	R3142S
HindIII	NEB	R3104S
NdeI	NEB	R0111S
SacI	NEB	R3156S
T4 DNA ligase	NEB	M0202L
Nunc™ Lab-Tek™ Chambered Coverglass	Thermo Fisher	155361
GeneMorph II Random Mutagenesis Kit	Agilent	200550
96-well cell culture microplate μclear®, black	Greiner	655090
Nunc™ 96-well Flat-Bottom Microplates	Thermo Fisher	167008
Critical commercial assays
HisPur™ Ni-NTA Resin	Thermo Fisher	88223
Glucose Consumption Assay	Sigma-Aldrich	G3293
Experimental models: Organisms/strains
Saccharomyces cerevisiae BY4741	B. Brachmann et al.61	N/A
Oligonucleotides
F-EI-500 GCGCGAAATTAATCGTTAC	This work	N/A
R-EI+500 GGTTTCACCCACGGTTACG	This work	N/A
G261T-Q87H(F) GAGGAGCTGGAGCATGAAATCATAGCC	This work	N/A
G261T-Q87H(R) GGCTATGATTTCATGCTCCAGCTCCTC	This work	N/A
G557A-R186H(F) GACGCGGGTGGCCATACTTCCCACACC	This work	N/A
G557A-R186H(R) GGTGTGGGAAGTATGGCCACCCGCGTC	This work	N/A
C782T-A261F(F) GAAAGATCTGCCAGTTATTACGCTGGAC	This work	N/A
C782T-A261F(R) GTCCAGCGTAATAACTGGCAGATCTTTC	This work	N/A
G796A-G266C(F) CTATTACGCTGGACTGTCACCAGGTAG	This work	N/A
G796A-G266C(R) CTACCTGGTGACAGTCCAGCGTAATAG	This work	N/A
G796A- G266S (F) CTATTACGCTGGACAGTCACCAGGTAG	This work	N/A
G796A- G266S (R) CTACCTGGTGACTGTCCAGCGTAATAG	This work	N/A
G797A-G266D(F) CTATTACGCTGGACGATCACCAGGTAG	This work	N/A
G797A-G266D(R) CTACCTGGTGATCGTCCAGCGTAATAG	This work	N/A
C830T-T277A(F) CTAACATTGGTATGGTTCGTGACGTTG	This work	N/A
C830T-T277A(R) CAACGTCACGAACCATACCAATGTTAG	This work	N/A
C922T-P308S(F) CCGCGACGCACTGTCCACTGAAGAAG	This work	N/A
C922T-P308S(R) CTTCTTCAGTGGACAGTGCGTCGCGG	This work	N/A
C968T-A323V(F) CAGTGGCTGAAGTGTGTGGCTCGCAAG	This work	N/A
C968T-A323V(R) CTTGCGAGCCACACACTTCAGCCACTG	This work	N/A
G1613A-R538H(F) CATTAAGAAGATTATCCATAACACGAACTTC	This work	N/A
G1613A-R538H(R) GAAGTTCGTGTTATGGATAATCTTCTTAATG	This work	N/A
F-end of yeeX-his-TAA-kan TCTCCAAAAAGCTGAAAGCTATGGGCGAAATGAAAAACGGCGA AGCGAAGCATCACCATCACCATCACTAAcacgtcttgagcgattgtg	This work	N/A
R-kan-intergenic after yeeX ATCAATGAGCAATGAGGGTTGCCGGGCAACCC TCATTGAATAAAACGGGAAatatcctccttagttcctattcc	This work	N/A
F-TmaR-500 CCAGCAAAGAAACCCGTATTCC	This work	N/A
R-TmaR+500 AGTGGAGGTGCGTTATGGCC	This work	N/A
F_F8L(C22A) CTACCAAGCCTTCATTACAGGACGTAC	This work	N/A
R_F8L(C22A) GTACGTCCTGTAATGAAGGCTTGGTAG	This work	N/A
F_V11I(G31A) CATTCCAGGACATACTGGAATTTGTTCG	This work	N/A
R_V11I(G31A) CGAACAAATTCCAGTATGTCCTGGAATG	This work	N/A
F_L12M(C34A) CATTCCAGGACGTAATGGAATTTGTTCG	This work	N/A
R_L12M(C34A) CGAACAAATTCCATTACGTCCTGGAATG	This work	N/A
F_R16P(G47C) CTGGAATTTGTTCCTCTGTTCCGTCGTAAG	This work	N/A
R_R16P(G47C) CTTACGACGGAACAGAGGAACAAATTCCAG	This work	N/A
F_D78H(G223C) GTTGACCATTACATCATCAAAAATGCCGAGC	This work	N/A
R_D78H(G223C) GCTCGGCATTTTTGATGATGTAATGGTCAAC	This work	N/A
F_A84G(C251G) GTTGACGATTACATCATCAAAAATGGCGAGC	This work	N/A
R_A84G(C251G) GCTCGCCATTTTTGATGATGTAATCGTCAAC	This work	N/A
F_E85D(G255T) GCCGATCTCTCCAAAGAACGCCGCGATATC	This work	N/A
R_E85D(G255T) GATATCGCGGCGTTCTTTGGAGAGATCGGC	This work	N/A
F_K95E(A283G) GAACGCCGCGATATCTCCGAAAAGCTGAAAG	This work	N/A
R_K95E(A283G)	This work	N/A
F-AatII -231TSS PTS CTTTCAGCTTTTCGGAGATATCGCGGCGTTC	This work	N/A
R-AvrII 21TT PTS CCGACGTCTGCAACAGTAATGCCAGC	This work	N/A
F-EI_486-Ins GACCTAGGGCCGCCGCTGGCGGAAGC	This work	N/A
R-mCherry-Ins CATTCTGGTTGCCGCTGAC	This work	N/A
F-Vec_pZS∗1PTS GTAAATTGGGCCGCATCTCGTGGAttacttgtacagctcgtccatg	This work	N/A
R-Vec_pZS∗1PTS_EI_486 TCCACGAGATGCGGCCCAATTTAC	This work	N/A
F-EI-Vac GTCAGCGGCAACCAGAATG	This work	N/A
R-EI-Vac GGAACTGCGCGACGAAGGTAAAGCGTTTGACG	This work	N/A
F-EI-Ins CCAGAGAACGCGCCATGATAGAGGTGTGGGAAG	This work	N/A
R-EI-Ins CTTCCCACACCTCTATCATGGCGCGTTCTCTGG	This work	N/A
F-pCAN CGTCAAACGCTTTACCTTCGTCGCGCAGTTCC	This work	N/A
R-pCAN CTATGGGCGAAATGAAAAACGGCGAAGCGAAG	This work	N/A
F-TmaR GAAGGCTTGGTAGTTTCCATTTTTATACCCCTG	This work	N/A
R-TmaR CTTCGCTTCGCCGTTTTTCATTTCGCCCATAG	This work	N/A
AccIII_F-TmaR CAATCCGGAATGGAAACTACCAAGCC	This work	N/A
XhoI_R-TmaR CAACTCGAGTTACTTCGCTTCGCCG	This work	N/A
XmaI_F-GB1 CAACCCGGGATGCAGTACAAGCTTATCC	This work	N/A
SalI_R-TmaR CAAGTCGAGTAACTTCGCTTCGCCG	This work	N/A
R- mut-pathcI GAACAGAGGAACAAATTCCAGTATGTCCTGTAATGAAG	This work	N/A
F-Phospho_pathcI /5Phos/CGTCGTAAGAACAAACTGCAACG	This work	N/A
R- mut-pathcII GAGAGATCGCCATTTTTGATGATGTAATGGTCAAC	This work	N/A
F-Phospho-pathcII /5Phos/CAAAGAACGCCGCGATATCTCC	This work	N/A
F-BamHI-Kozak-EI TGGATCCCTGCAGGACAAAATGATTTCAGGCATTTTAGC	This work	N/A
R-SalI-mCherry ggacgagctgtacaagTAAGTCGACC	This work	N/A
F-KpnI-TmaR(coli) GAAAGGTACCGCATGGAAACTACCAAGCC	This work	N/A
F-HindIII-Venus AATTAAGCTTAGCTAGCTTACTTGTACAGCTCGTC	This work	N/A
F-YeeX(Sal) CATTAAAGAGGAGAAAGGTACCGCATGGAAACAACCAAGCCTTC	This work	N/A
R-YeeX(Sal) CTTCTCTTTTCCATGGATTCGGCTTTTACATCGGCG	This work	N/A
F-pZA32-venus CGCCGATGTAAAAGCCGAATCCATGGAAAAGAGAAG	This work	N/A
R-pZA32-venus GAAGGCTTGGTTGTTTCCATGCGGTACCTTTCTCCTCTTTAATG	This work	N/A
F-NdeI_EI(sal) GGCAGCCATATGATGATTTCAGGCATTTTAGC	This work	N/A
R-SacI_EI(sal) GTGGAGAGCTCGCAGATTGTTTTTTCTTC	This work	N/A
Recombinant DNA
pZS∗1-PTSoperon_EI-mCherry	This work	N/A
pZE12-TmaR	Szoke et al.21	N/A
pQELL-EI	OAC lab collection	N/A
pQELL-E(G266C)	This work	N/A
pQELL-EI(G266S)	This work	N/A
pQELL-EI(G266D)	This work	N/A
pBADLLEI-mCherry	Lopian et al.18	N/A
pBAD18-mCherry	Szoke et al.14	N/A
pTrc-eYFP-TmaR	Szoke et al.21	N/A
pZS∗1-TmaRr-YFP	Szoke et al.21	N/A
pCA24N	Kitagawa et al.62	N/A
JW1989-YeeX (pCA24N-TmaR)	Kitagawa et al.62	N/A
pCA24N-TmaR(F8L)	This work	N/A
pCA24N-TmaR(V11I)	This work	N/A
pCA24N-TmaR(R16P)	This work	N/A
pCA24N-TmaR(D78H)	This work	N/A
pCA24N-TmaR(A84G)	This work	N/A
pCA24N-TmaR(E85D)	This work	N/A
pCA24N-TmaR(K95E)	This work	N/A
O-03b-p416-GPD-NES13-Im2-4ltb-mFusionRed-URA3	EDL lab collection	N/A
P-05-p413-GPD-Venus-E9-p53-HIS3	EDL lab collection	N/A
GB1 vector	EDL lab collection	N/A
pYeast-YFP-TmaR	This work	N/A
pYeast-YFP-TmaR(pathcII)	This work	N/A
pYeast-YFP-TmaR(pathcI)	This work	N/A
pYeast-EI-mCherry	This work	N/A
pYeast-EI(G266C)-mCherry	This work	N/A
pZA32-TmaR(EC)-YFP	This work	N/A
pZA32-TmaR(Sal)-YFP	This work	N/A
BAD18EI(Sal)-mCherry	This work	N/A
Software and algorithms
NIS Elements Advanced Research (AR) version 4.5	Nikon	N/A
Image Lab v6.0.1	Bio-Rad	http://www.bio-rad.com/en-il/product/image-lab-software
PyMOL Molecular Graphics. 2010	System Schrodinger, LLC	https://pymol.org/2/

Experimental model and subject details

Bacterial strains and growth conditions

Bacterial strains and plasmids used in this study are listed in Key resources table, respectively. Unless otherwise indicated, overnight cultures were grown at 30°C in minimal M9-glucose media supplemented with 0.0002% vitamin B1, 1 mM MgSO₄, 0.2% CAA and 0.4% glucose and with the appropriate antibiotics at the following concentrations: kanamycin (30 μg/mL), chloramphenicol (25 μg/mL) and ampicillin (200 μg/mL). The overnight cultures were diluted 1:100 into fresh M9-glucose media, and growth continued until mid-logarithmic phase. Proteins were expressed from their native promoters and locus in the chromosome or overexpressed from the indicated plasmids induced by 0.1 mM IPTG or 0.1% arabinose, depending on the promoter, from the start of growth.

Method details

Construction of strains and plasmids

All primers used in this study are listed in Key resources table.

Strains construction

Mutations in the chromosome-expressed EI-mCherry were constructed by two overlapping PCR products amplified from the ptsI-mCherry fused genes, using the primers F-EI-500 and R-EI+500 together with the primers containing the respective mutations (as described below), which were ligated using Gibson assembly procedure.⁶³ The assembled sequences were amplified by PCR and introduced into MG1655 strain using positive and negative selection.⁶⁴ The primers used to introduce the indicated mutations were:

G261T-Q87H(F) and G261T-Q87H(R) were used to construct NA-EI(Q87H)-mCherry.

G557A-R186H(F) and G557A-R186H(R) were used to construct NA-EI(R186H)-mCherry. C782T-A261F(F) and C782T-A261F(R) were used to construct NA-EI(A261F)-mCherry.

G796A-G266C(F) and G796A-G266C(R) were used to construct NA-EI(G266C)-mCherry. G796A- G266S (F) and G796A- G266S (R) were used to construct NA-EI(G266S)-mCherry. G797A-G266D (F) and G797A-G266D(R) were used to construct NA-EI(G266D)-mCherry. C830T-T277A(F) and C830T-T277A(R) were used to construct NA-EI(T277A)-mCherry.

C922T-P308S(F) and C922T-P308S(R) were used to construct NA-EI(P308S)-mCherry.

C968T-A323V(F) and C968T-A323V(R) were used to construct NA-EI(A323V)-mCherry.

G1613A-R538H(F) and G1613A-R538H(R) were used to construct NA-EI(A538H)-mCherry.

NA-TmaR-YFP_EI-mCherry, NA-TmaR-YFP_EI(G266C)-mCherry, NA-TmaR-YFP_EI(G266S)-mCherry and NA-TmaR-YFP_EI(G266D)-mCherry strains were constructed by P1⁶⁵ transduction of tmaR-YFPϕcat cassette to NA-EI-mCherry, NA-EI(G266C)-mCherry, NA-EI(G266S)-mCherry and NA- EI(G266D)-mCherry, respectively.

NA-TmaR-YFP strain was generated by P1 transduction⁶⁵ of tmaR-YFPϕcat from SX1989⁶⁶ to MG1655.

ΔtmaR and ΔtmaR EI-mCherry strains was generated by P1 transduction⁶⁵ of ΔtmaR::kan from JW1989-1⁶⁷ to MG1655. The Kan^R cassette was removed from ΔtmaR using pCP20⁶⁸ followed by P1 transduction⁶⁵ of ptsI-mCherryϕkanR.

Strain OG543, which carries tmaR-hisϕkan, was constructed as follows: the kan cassette was amplified from strain OG167,²³ using the primers “F-end-of-yeeX-his-TAA-kan” that contained homology to the end of the tmaR gene, as well as the His-tag sequence, and the primer “R-kan-intergenic-after-yeeX”. The PCR fragment was introduced into the BW25113 strain containing pKD46 plasmid. The cells were grown at 42⁰C and tested for the loss of pKD46. The tmaR-hisϕkan fragment was transferred to strain MG1655 by P1 transduction.⁶⁵

TmaR mutations in the chromosome were introduced by two overlapping PCR products of tmaR, tmaR-YFP or His-tmaR amplified from strains MG1655, NA-TmaR-YFP and OG543, respectively, using the primers F-TmaR-500 and R-TmaR+500 together with primers containing the respective mutations (as described below), which were ligated using Gibson assembly.⁶³ The assembled sequences were amplified by PCR and introduced into MG1655 using positive and negative selection.⁶⁴ The primers and templates used to introduce the mutations were:

F_F8L(T22C) and R_F8L(T22C) primers were used to construct NA-TmaR(F8L)-YFP, TmaR(F8L)-His and NA-TmaR(F8L). F_V11I(G31A) and R_V11I(G31A) primers were used to construct NA-TmaR(V11I)-YFP, TmaR(V11I)-His and NA-TmaR(V11I). F_L12M(C34A) and R_L12M(C34A) primers were to construct NA-TmaR(L12M)-YFP, TmaR(L12M)-His and NA-TmaR(L12M). F_R16P(G47C) and R_R16P(G47C) primers were to construct NA-TmaR(R16P)-YFP, TmaR(R16P)-His and NA-TmaR(R16P). F_D78H(G223C) and R_D78H(G223C) primers were to construct NA-TmaR(D78H)-YFP, TmaR(D78H)-His and NA-TmaR(D78H). F_A84G(C251G) and R_A84G(C251G) primers were to construct NA-TmaR(A84G)-YFP, TmaR(A84G)-His and NA-TmaR(A84G). F_E85D(G255T) and R_E85D(G255T) primers were to construct NA-TmaR(E85D)-YFP, TmaR(E85D)-His and NA-TmaR(E85D). F_K95E(A283G) and R_K95E(A283G) primers were to construct NA-TmaR(K95E)-YFP, TmaR(K95E)-His and NA-TmaR(K95E).

The above TmaR and TmaR-YFP variants were transduced with P1⁶⁵ containing ptsI-mCherryϕkanR to generate NA-EI-mCherry_TmaR(F8L)-YFP, NA-EI-mCherry_TmaR(F8L), NA-EI-mCherry_TmaR(V11I)-YFP, NA-EI-mCherry_TmaR(V11I), NA-EI-mCherry_TmaR(R16P)-YFP, NA-EI-mCherry_TmaR(R16P), NA-EI-mCherry_TmaR(D78H)-YFP, NA-EI-mCherry_TmaR(D78H), NA-EI-mCherry_TmaR(A84G)-YFP, NA-EI-mCherry_TmaR(A84G), NA-EI-mCherry_TmaR(E85D)-YFP, NA-EI-mCherry_TmaR(E85D), NA-EI-mCherry_TmaR(K95E)-YFP and NA-EI-mCherry_TmaR(K95E) strains.

Plasmids construction

pZS^∗1PTSoperon_EI-mCherry was constructed in two steps: First, the PTS operon was amplified from MG1655 using the primers: F-AatII −231TSS PTS and R-AvrII 21TT PTS. The PCR products and the vector pZS^∗13luc⁶⁹ were cleaved with AatII and AvrII and ligated using T4 DNA ligase to construct pZS^∗1PTSoperon. Second, mCherry was fused to the 3′ end of the ptsI gene by Gibson assembly.⁶³ The vector PCR was amplified from pZS^∗1PTSoperon using the primers F-Vec_pZS^∗1PTS and R-Vec_pZS^∗1PTS_EI_486. The ptsI-mCherry fusion was amplified from pBAD18EI-mCherry using the primers F-EI_486-Ins and R-mCherry-Ins to construct pZS^∗1PTSoperon EI-mCherry.

To construct pQELL-EI(G266C), pQELL-EI(G266S) and pQELL-EI(G266D), two overlapping PCR products of the respective amplified vector and inserts described below, were ligated by Gibson assembly.⁶³ The vector PCR, which was the same for the three plasmids, was amplified from pQELL-EI with the primers F-EI-Vac and R-EI-Vac. The inserts were amplified by PCR from strains NA-EI(G266C)-mCherry, NA-EI(G266S)-mCherry and NA-EI(G266D)-mCherry, respectively, using the primers F-EI-Ins and R-EI-Ins.

To construct pCA24N-TmaR(F8L), pCA24N-TmaR(V11I), pCA24N-TmaR(R16P), pCA24N-TmaR(D78H), pCA24N-TmaR(A84G), pCA24N-TmaR(E85D) and pCA24N-TmaR(K95E), two overlapping PCR products were ligated by Gibson assembly.⁶³The vector pCA24N-TmaR was the same in all seven plasmids with the N and the C termini of the tmaR gene amplified from pCA24N-TmaR with the following primers F-pCAN and R-pCAN. The insert PCR fragments were amplified from cells carrying the respective tmaR mutation, using the primers F-TmaR and R-TmaR to construct the strains NS-TmaR(F8L)-YFP, NS-TmaR(V11I)-YFP, NS-TmaR(R16P)-YFP, NS-TmaR(D78H)-YFP, NS-TmaR(A84G)-YFP, NS-TmaR(E85D)-YFP and NS-TmaR(K95E)-YFP.

The plasmids pYeast-YFP-TmaR, pYeast-YFP-TmaR(pathcII), pYeast-YFP-TmaR(pathcI), pYeast-EI-mCherry, pYeast-EI(G266C)-mCherry were generated by modifying plasmids p413 and p416,⁷⁰ which express the tetrameric and dimeric proteins reported in,³⁶ as following:

pYeast-YFP-TmaR was constructed in two steps. First, tmaR sequence was amplified by PCR from an MG1655 colony, using primers AccII_F-TmaR and XhoI_R-TmaR, then cleaved with AccII and XhoI and ligated with T4 ligase to GB1, which has been cleaved by the same restriction enzymes, thus generating GB1-eYFP-tmaR. Second, eYFP-tmaR sequence was amplified by PCR from GB1-eYFP-tmaR, using primers XmaI_F-GB1 and SalI_R-TmaR, then cleaved with SalI and XmaI and ligating to P-05-p413-GPD-Venus-E9-p53-HIS3 (p413 backbone), cleaved by the same restriction enzymes.

pYeast-YFP-TmaR(pathcII) was constructed by inverse PCR reaction on pYeast-YFP-TmaR plasmid with the primers R-mut-pathcII and F-Phospho-pathcII. The PCR products were treated with DpnI and ligated using T4 DNA ligase.

pYeast-YFP-TmaR(pathcI), was constructed by inverse PCR reaction on pYeast-YFP-TmaR plasmid with the primers R-mut-pathcIand F-Phospho-pathcI. The PCR products were treated with DpnI and ligated using T4 DNA ligase.

pYeast-EI-mCherry was constructed by first amplifying ptsI-mCherry from pBADLLEI-mCherry with primers F-BamHI-Kozak-EI and R-SalI-mcherry. Next the PCR product and pO-03b-p416-GPD-NES13-Im2-4ltb-mFusionRed-URA3 (p416 backbone) were cleaved with SalI and BamHI and ligated using T4 ligase.

To construct pYeast-EI(G266C)-mCherry, pYeast-EI-mCherry was amplified with G796A-G266C(F) and G796A-G266C(R) primers to create overlapping PCR products that were ligated by Gibson assembly.⁶³

To construct pZA32-TmaR(EC)-YFP, the TmaR-YFP fragment was amplified from the pZS^∗1TmaR-venus using the primers F-KpnI-TmaR and F-HindIII-Venus. The PCR products and the vector pZA32-luc were cleaved by KpnI and HindIII and ligated using T4 DNA ligase.

To construct pZA32-TmaR(Sal)-YFP, two overlapping PCR products were ligated by Gibson assembly¹: 1. tmaR(Sal) was amplified using the primers F-YeeX(Sal) an R-YeeX(Sal) from SL1344 chromosome. 2. tmaR(EC)-YFP was amplified from pZA32-TmaR(EC)-YFP using the primers F-pZA32-venus and R-pZA32-venus.

To construct BAD18EI(Sal)-mCherry, S. typhimurium ptsI gene was amplified from an SL1344 colony using the primers F-NdeI_EI(sal) and R-SacI_EI(sal). The PCR products and the vector pBADLL-EImCherry were cleaved by NdeI and SacI and ligated using T4 DNA ligase.

Fluorescence microscopy

Fluorescence microscopy was carried out as described previously.¹³ For snap-shot imaging, 10–20 μL (according to the cell density) were spotted on 1% agarose PBS pads. Unless otherwise indicated, the scale bar size was 5 μm. Cells are shown in phase (gray), in the appropriate fluorescent channel (colored) or as merged.

Image analysis

Image analysis was performed with NIS-Elements Advanced Research (AR) version 4.4 software (Nikon). The detection of cells and clusters were carried out automatically, using the MicroAnalyzer tool⁷¹. Unless otherwise indicated, for data processing, including plots, we used custom code on matplotlib and pandas libraries using python. The bar plots were generated using Microsoft Excel.

Mutant library construction

EI library

For the random creation of single point mutations, the ptsI gene from pZS^∗1PTSoperon ptsI-mCherry (2,052 ng target DNA) was amplified with primers corresponding to its 5′ and the 3' by error prone PCR using the GeneMorph II Random Mutagenesis Kit (Agilent #200550) following manufacture recommendations. The PCR library was cloned into the low copy plasmid pZS^∗1-PTSoperon_EI-mCherry, which expresses the mutants from the pts operon native promoter. The final library was expressed in Δpts-operon cells.

TmaR library

For the random creation of single point mutations, the tmaR gene, present on pTrc-eYFP-TmaR (308.5 ng target DNA), was amplified with primers corresponding to its 5′ and the 3' termini by error prone PCR using the GeneMorph II Random Mutagenesis Kit (Agilent #200550) following manufacture recommendations. The PCR library was cloned into the low copy plasmid pZS^∗1-TmaRr-YFP, which expresses the mutants from tmaR native promoter. The final library was expressed in ΔtmaR_EI-mCherry cells.

Automated fluorescence microscopy

Microscopic screening was performed using an automated microscopy setup as described previously.⁷² In short, cells were picked from 96 wells plates and inoculated into 384-well polystyrene growth plates containing M9-glucose minimal medium with casamino acids (CAA) using a RoToR arrayer (Singer). Liquid cultures were grown overnight in in a shaking incubator (LiCONiC Instruments) at 30°C. A JANUS liquid handler (PerkinElmer), which is connected to the incubator, was used to back-dilute the strains 1:50 into plates containing M9 glucose minimal medium without CAA. Plates were then transferred back to the incubator and were allowed to grow for 3 h at 30°C to reach logarithmic growth phase, as validated in a preliminary calibration. The liquid handler was then used to transfer strains into glass-bottom 384-well microscopy plates (Matrical Bioscience) coated with Poly-L-lysine (Sigma-Aldrich) for 20 min to allow cell adhesion. Wells were washed twice in medium to remove floating cells and to obtain a cell monolayer. The wells were then filled with M9-glucose minimal medium without CAA. Plates were then transferred into an automated inverted fluorescent microscopic ScanR system (Olympus) using a swap robotic arm (Hamilton). Imaging of plates was performed in 384-well format using an X60 air lens (NA 0.9) at 24°C with a cooled charge-coupled device camera (ORCA-ER; Hamamatsu). Images were acquired at GFP (excitation at 490/20 nm, emission at 535/50 nm) and mCherry (excitation at 572/35 nm, emission at 632/60 nm) channels.

Western blot analysis

Equal amounts of cells grown to the indicated OD₆₀₀ were collected from each strain, washed with PBS and resuspended in Laemmli buffer. Each sample was then heated to 95°C for 10 min, and the lysates were separated on 12% SDS–polyacrylamide gels (unless otherwise indicated). Gels were subjected to Western blot analysis as described previously.⁷³ The proteins were probed with the indicated antibodies. For quantification of the Western blot bands’ intensity, we used Image Lab.

Co-immunoprecipitation (CoIP) analysis

Cells were diluted 1:100 in 25 mL of fresh LB supplemented with 0.1 mM IPTG and grown until mid-logarithmic phase. Subsequently, cells were pelleted, washed in 1X PBS and resuspended in 1mL 1X PBS. Cells were crosslinked by adding 0.1% formaldehyde for 20 min at room temperature, then quenched with 250 mM glycine and incubating for 10 min at room temperature. The crosslinked cells were lysed by Mixer Mill MM400 using glass beads in lysis buffer supplemented with 2% sarkosyl, purified using HisPur Ni-NTA Resin (Thermo Scientific) as suggested by the manufacturer. The eluted proteins were resuspended in Laemmli sample buffer, boiled at 95°C for 10 min, and fractionated on a 12% SDS-polyacrylamide gels. The proteins were blotted onto a nitrocellulose membrane. The membrane was incubated overnight with anti-mCherry antibody (ab167453, Abcam) at 4°C, washed, incubated with anti-rabbit HRP antibody for 1 h and imaged. The membrane was stripped by ReBlot Plus Strong Antibody Stripping Solution (Merck), incubated overnight at 4°C with anti-His antibody (Genscript 6G2A9) and then probed with anti-mouse HRP antibody and imaged. Image LAB was used for bands intensity quantification. For Figure 1I, cells expressing EI-mCherry, NA-EI(G266C)-mCherry, NA-EI(G266S)-mCherry or NA-EI(G266D)-mCherry, which also harbor pCA24N or pCA24N-TmaR plasmid were used. The normalized ratio between EI to TmaR was calculated by introducing the band intensities into the flowing equation:

$$\frac{\left({EI\left(mut\right)}_{TmaR\left(WT\right)}{-EI\left(mut\right)}_{noTmaR}\right)/TmaR}{\left({EI\left(WT\right)}_{TmaR\left(WT\right)}{-EI\left(WT\right)}_{noTmaR}\right)/TmaR}$$

For Figure 3A, cells expressing EI-mCherry and harboring pCA24N, pCA24N-TmaR WT, pCA24N-TmaR mutants (F8L, V11I, R16P, D78H, A84G, E85D, K95E) or pCA24N-His-tag were used. The normalized ratio between EI to TmaR was calculated by introducing the band intensities into the flowing equation:

$$\frac{\left({EI}_{TmaR\left(mut\right)}-E{I}_{noTmaR}\right)/TmaR\left(mut\right)}{\left({EI}_{TmaR\left(WT\right)}-E{I}_{noTmaR}\right)/TmaR\left(WT\right)}$$

Far-western analysis

Far Western analyses were carried out essentially as described previously.⁷³ For purification of EI-His, EI(G266C)-His, EI(G266S)-His, EI(G266D)-His, EI-mCherry or mCherry, overnight cultures of MG1655 cells harboring pQELL-EI, pQELL-EI(G266C), pQELL-EI(G266S), pQELL-EI(G266D), pBADLLEI-mCherry or pBAD18-mCherry, respectively, were diluted 1:100 in 40 mL of fresh LB supplemented with either 1mM IPTG for pQELL vectors or 0.1% arabinose for pBAD vectors. Cells were grown until mid-logarithmic phase, lysed by Mixer Mill MM400 using glass beads and purified using HisPur Ni-NTA Resin (Thermo Scientific), as suggested by the manufacturer.

For Figure S3C, overnight cultures of ΔtmaR cells with or without pZE12-TmaR were diluted 1:100 in 5mL of fresh LB supplemented with 0.1 mM IPTG and grown until the mid-logarithmic phase. Subsequently, cells were pelleted, washed three times with 1X PBS, resuspended in Laemmli sample buffer, boiled at 95°C for 10 min, and run on four 12% SDS-polyacrylamide gels. The proteins were blotted onto four nitrocellulose membranes that were stained with Ponceau S (Sigma-Aldrich). The membranes were incubated with purified EI-His, EI(G266C)-His, EI(G266S)-His or EI(G266D)-His proteins in 7% Difco skim milk (BD biosciences), dissolved in 10 mL phosphate buffered saline with 20% Tween 20 (PBST) for overnight at 4°C, washed, incubated overnight at 4°C with anti-His antibody (Genscript 6G2A9), washed and incubated with anti-mouse HRP antibody for 1 h. The membranes were washed three times with PBST and imaged. The normalized ratio between EI bound to TmaR to the background was calculated by introducing the band intensities into the flowing equation:

$$\frac{{EI}_{mut}/backgrundband}{{EI}_{WT}/backgrundband}$$

For Figure 3B, overnight cultures of ΔtmaR cells, overexpressing TmaR WT or its variants (F8L, V11I, R16P, D78H, A84G, E85D, K95E) or His-tag from the respective pCA24N plasmid derivatives, were diluted 1:100 in 5 mL of fresh LB supplemented with 0.1 mM IPTG and grown until mid-logarithmic phase. Subsequently, cells were pelleted, washed with 1X PBS, resuspended in Laemmli sample buffer, boiled at 95°C for 10 min, and fractionated on three 12% SDS-polyacrylamide gel. The proteins were blotted onto three nitrocellulose membranes. The first membrane (for Western analysis) was incubated overnight at 4°C with anti-His antibody (Genscript 6G2A9) and then with anti-mouse HRP antibody. The second and the third membranes (for Far-Western analysis) were incubated overnight with purified EI-mCherry or mCherry respectively in 7% Difco skim milk (BD biosciences) dissolved in 10mL PBST at 4°C, washed, incubated with anti-mCherry antibody (Abcam ab167453) overnight at 4°C, washed and incubated with anti-rabbit HRP antibody for 1 h. The membranes were washed three times with PBST and imaged. The normalized ratio between EI bound to TmaR to the background was calculated by introducing the band intensities into the flowing equation:

$$\frac{EI/{TmaR}_{mut}}{EI/{TmaR}_{WT}}$$

Bacterial two-hybrid assay

Bacterial two-hybrid assays were carried out essentially as described previously.⁷⁴ Briefly, the BTH101 strain was co-transformed with a combination of pKT25 and pUT18C expressing either zipper domain, TmaR variants or EI variants, as indicated in the corresponding figure. Cells were grown overnight in biological replicates in LB containing kanamycin, ampicillin and 0.5 mM IPTG. Interaction between the two hybrid proteins was monitored on LB X-GAL plates [LB agar supplemented with ampicillin (100 μg/mL), kanamycin (50 μg/mL), X-Gal (40 μg/mL), and IPTG (0.5 mM)].

Phase separation analysis in yeast

BY4741 yeast cells⁶¹ were transformed via heat shock with plasmids encoding the indicated variants of mCherry⁷⁵-tagged EI (p4X6 backbone⁷⁰) and Venus⁷⁶-tagged tmaR (p4X3 backbone⁷⁰). Three biological replicates of each transformant were grown in synthetic defined media lacking uracil and/or histidine for three days to generate a saturated culture. Subsequently, 1 μL culture was transferred to 35 μL media in an optical glass bottom plate (Greiner), and grown for 6 h. The cells were imaged, cells and puncta were automatically identified, RFP and GFP intensities were recorded, and median RFP and GFP intensities were plotted against each other, as previously described.³⁶ For phase diagrams, cells without puncta were defined as cells that do not show puncta in either channel.

MacConkey plates

MacConkey plates were made by mixing 40 mg/mL of Difco MacConkey Agar base (Becton Dickinson and company) with 0.4% fructose or glucose as indicated. Bacteria were streaked from single fresh colonies, and the plates were incubated overnight at room temperature unless otherwise indicated. Cells deleted for the pts operon (PTS-KO) served as a negative control for consumption of PTS sugars on the MacConkey plates.

Glucose consumption assay

The glucose consumption assay was done as described before¹⁴ with cells deleted for the pts operon or MG1655 ΔtmaR cells harboring one of the following plasmids: pCA24N, pCA24N-TmaR, pCA24N-TmaR(F8L), pCA24N-TmaR(V11I), pCA24N-TmaR(A84G) or pCA24N-TmaR(K95E). Samples were taken in 5 min intervals and up to 20 min after the addition of glucose. The glucose consumption rate was calculated from the slope of the calibration curve.

EI monomer to dimer ratio analysis

Cells were diluted 1:100 in 25 mL of fresh LB supplemented with 0.1 mM IPTG and grown until mid-logarithmic phase. Subsequently, cells were pelleted, washed in 1X PBS and resuspended in 1 mL 1X PBS. Cells were crosslinked by adding 0.1% formaldehyde for 20 min at room temperature, then quenched with 250 mM glycine and incubating for 10 min at room temperature. The crosslinked cells were lysed by Mixer Mill MM400 using glass beads in lysis buffer supplemented with 2% sarkosyl, purified using HisPur Ni-NTA Resin (Thermo Scientific), as suggested by the manufacturer. The eluted proteins were resuspended in Laemmli sample buffer, and kept on ice to prevent reverse cross-linking. The samples were fractionated on 4–20% Mini-PROTEAN TGX precast protein gels (cat #4561096) gels. The proteins were blotted onto a nitrocellulose membrane. The membrane was incubated overnight with anti-mCherry antibodies (ab167453, Abcam) at 4°C, washed, incubated with anti-rabbit HRP antibody for 1 h and imaged. For quantification of the Western blot bands Image Lab software was used.

For Figure 6A, cells expressing EI-mCherry from the chromosome and harboring pCA24N, pCA24N-TmaR WT or its mutants (F8L, V11I, R16P, D78H, A84G, E85D, K95E), or only a His-tag from the respective pCA24N plasmid derivatives were used.

For Figure 6B, cells expressing EI-mCherry, NA-EI(G266C)-mCherry, NA-EI(G266S)-mCherry and NA-EI(G266D)-mCherry from the chromosome and harboring pCA24N or pCA24N-TmaR plasmid were used.

Modeling the structure of TmaR

The 3D structure of TmaR was modeled using the TrRosetta web-server (https://yanglab.nankai.edu.cn/trRosetta)21–23.

Modeling the structure of the EI-TmaR interaction

TmaR was docked to EI using the ClusPro protein-protein docking server (https://cluspro.bu.edu).^37,38,39,40 For EI structure, PDB ID 2HWG was used, and for TmaR structure, the model generated as described above was used. The electrostatically favored energy function was used.

The input for the generation of the predicted 3D structures of EI and TmaR proteins and their complex by AlphaFold2, a neural network-based deep learning system developed by Google DeepMind,⁷⁷ using Google Colab notebooks by Sergey Ovchinnikov,⁷⁸ was the protein sequence of two EI (two subunits) and one TmaR. the models were predicted using the notebook accessible at:

https://colab.research.google.com/github/sokrypton/ColabFold/blob/main/AlphaFold2.ipynb.

EI electrostatic surface analysis

The electrostatic potential (distribution of positively and negatively charged regions) on the EI surface was calculated using the Adaptive Poisson–Boltzmann Solver⁷⁹ as implemented in PyMOL (Schrodinger LLC. The PyMOL Molecular Graphics System. 2010).

Co-evolution analysis

To identify potentially interacting residues between EI and TmaR, we applied co-evolution analysis using the RaptorX Contact Prediction Server (http://raptorx.uchicago.edu).^45,46,47,48

EI normal mode analysis

To inspect main internal motions of EI, we provided the coordinates of 2HWG to the Anisotropic Network Model Web Server 2.1 (http://anm.csb.pitt.edu)⁴³ We selected mode n = 3.

Figures of structures were generated using PyMOL (Schrodinger LLC. The PyMOL Molecular Graphics System. 2010)

Quantification and statistical analysis

The statistical analyses were done as described in the figure legends, using python custom code by the scipy.stats package.

Published: March 17, 2025