Identification of Potential Regulatory Domains within the MreC and MreD Components of the Cell Elongation Machinery

Cell shape in bacteria is largely determined by the cell wall structure that surrounds them. The multiprotein machine called the Rod system (elongasome) has long been implicated in rod shape determination in bacilli.

ABSTRACT

The bacterial peptidoglycan (PG) cell wall maintains cell shape and prevents osmotic lysis. During growth of rod-shaped cells, PG is incorporated along the cell cylinder by the RodA-PBP2 synthase of the multiprotein Rod system (elongasome). Filaments of the actin-like MreB protein orient synthesis of the new PG material. They are connected to the RodA-PBP2 synthase in part through the RodZ component. MreC and MreD are other conserved components of the system, but their function is not well understood. Amino acid changes in RodA-PBP2 that bypass a requirement for MreC and MreD function were recently identified, suggesting that the Mre proteins act as activators of the synthase. To further investigate their function, we developed a genetic strategy to identify dominant negative alleles of mreC and mreD in Escherichia coli. Residues essential for Rod system function were identified at the junction of two subdomains within MreC and in a predicted ligand-binding pocket of MreD. Additionally, we found that although the proline-rich C-terminal domain of MreC is nonessential, substitutions within this region disrupt its function. Based on these results, we propose that the C terminus of MreC and the putative ligand-binding domain of MreD play regulatory roles in controlling Rod system activity.

IMPORTANCE Cell shape in bacteria is largely determined by the cell wall structure that surrounds them. The multiprotein machine called the Rod system (elongasome) has long been implicated in rod shape determination in bacilli. However, the functions of many of its conserved components remain unclear. Here, we describe a new genetic system to dissect the function of these proteins and how we used it to identify potential regulatory domains within them that may modulate the function of the shape-determining machinery.

INTRODUCTION

In order to grow and divide, bacterial cells must expand the peptidoglycan (PG) cell wall matrix that surrounds their cytoplasmic membrane and protects it from osmotic rupture (1). The wall structure consists of glycan strands of alternating N-acetylglucosamine (GlcNAc) and N-acetylmuramic acid (MurNAc) units, with interstrand cross-links formed by short peptides connected to the MurNAc sugars (1). Two types of enzymes are capable of synthesizing the PG network: the class A penicillin-binding proteins (aPBPs) (2) and complexes between a SEDS (shape, elongation, division, and sporulation) protein and a class B PBP (bPBP) (3–5). The aPBPs possess two enzymatic domains in a single polypeptide, a PG glycosyltransferase (PGTase) domain for glycan polymerization and a transpeptidase (TPase) domain for cross-linking (2). These activities are split in the SEDS-bPBP complexes, with the SEDS proteins forming a distinct family of PGTase enzymes (6) and the bPBPs employing a TPase domain similar to that of the aPBPs (2).

Multiprotein complexes involving the PG synthases are thought to direct the major morphogenic processes of cell elongation and division in most bacteria (7, 8). In rod-shaped cells, the Rod system (elongasome) expands the lateral wall during cell elongation, whereas the cytokinetic ring (divisome) constructs the new cell poles during division (7, 8). Each of these complexes relies on a cytoskeletal filament for their activity, with the Rod system utilizing polymers of the actin-like protein MreB to promote rod shape and the divisome using dynamic polymers of the tubulin-like FtsZ protein to organize the divisome at midcell (9–12).

The essential PG synthase of the Rod system is a SEDS-bPBP complex formed between the polymerase RodA and TPase PBP2 (3, 5). Localization studies have found that the synthase and other Rod system components localize as dynamic foci along the cell cylinder and that these foci rotate circumferentially around the cellular long axis (13–15). PG synthesis is thought to provide the propulsive force for these dynamics (13–15), with MreB filaments associated with the RodA-PBP2 synthases serving as rudders to orient the motion (16). Overall, the current model is that directional insertion of new PG glycans perpendicular to the long axis is what promotes rod shape. The role of the aPBPs in this process is unclear, but recent results suggest that they may work outside of the Rod complex to fortify and expand the oriented material laid down by the Rod system (17, 18), with the relative ratio of Rod system versus aPBP activity serving as a mechanism to modulate cell width (19).

MreB and RodA-PBP2 work with the membrane proteins RodZ, MreC, and MreD to promote cell elongation (7, 8). The roles of these additional components in Rod system function are not well understood. Among them, RodZ (20–22) has been the best characterized. A cocrystal structure of RodZ with MreB, mutant analysis, and molecular dynamics simulations suggest that the cytoplasmic domain of RodZ anchors MreB to the Rod system and alters MreB filament bending (23–26). MreC and MreD are also highly conserved and essential for Rod system function, but it has remained unclear whether they act simply as scaffold proteins for complex assembly or have additional functional roles in cell elongation (27–31). We recently presented evidence that MreC induces a conformational change in PBP2 that allosterically activates the RodA-PBP2 synthase (3). This work raised the possibility that MreC, possibly in conjunction with MreD, functions to modulate Rod system activity in response to as-yet-unknown cellular or environmental stimuli.

Here, we report a general genetic strategy for identifying dominant negative variants of Rod system proteins in Escherichia coli and its use to further investigate the functions of MreC and MreD. Residues critical for Rod system function were identified in MreC at the junction of two of its core subdomains. We also found that although the proline-rich C-terminal domain of MreC is nonessential, substitutions within this region can disrupt its function. Additionally, amino acid substitutions in a predicted ligand-binding pocket of MreD were also found to disrupt Rod system function. Based on these results, we propose that the C terminus of MreC and the putative ligand-binding pocket of MreD may play regulatory roles in controlling the activity of the cell elongation machinery.

RESULTS

Strategy to identify functionally important residues in MreC and MreD.

To select for mutants defective for Rod system activity, we took advantage of the antibiotic mecillinam, a beta-lactam that specifically targets the transpeptidase active site of PBP2 (32), and the conditional essentiality of the Rod system (29) (Fig. 1). We previously showed that mecillinam induces a lethal malfunctioning of the Rod system involving the futile synthesis of un-cross-linked glycans that are destined for rapid degradation by the lytic transglycosylase Slt (33). Thus, loss-of-function mutations in Rod system genes can be isolated by plating for mecillinam resistance under conditions that allow cells to survive without the normally essential Rod machinery (33, 34) (Fig. 1).

FIG 1.

Summary of growth phenotypes resulting from a Rod system defect. Shown is a schematic diagram of the Rod system highlighting its components and a chart summarizing the growth phenotypes of strains with or without a defective Rod system. MreC and MreD are abbreviated as C and D, respectively. Elevated FtsZ levels are generated using pTB63, a low-copy-number plasmid carrying the ftsQ-ftsA-ftsZ operon (ftsQAZ) under the control of native promoters. The rich medium used was LB, and minimal medium was M9. Mecillinam resistance (R) or sensitivity (S) was assessed on LB. NA, not assayed because cells do not grow even without drug.

Our goal was to identify amino acid residues required for the function of MreC and MreD. However, if we were to simply select for nonfunctional variants of these factors, the majority of the mutants isolated would likely encode proteins that are either unstable or poorly expressed, which would not be especially informative for functional analysis. We therefore chose to select for dominant negative mutants of mreC and mreD because such mutants should, by definition, produce stable proteins that form at least partial complexes with their partners to disrupt Rod system activity. To this end, a plasmid library encoding a PCR-mutagenized copy of the mreCD operon under the control of the IPTG (isopropyl-β-d-thiogalactopyranoside)-inducible tac promoter (P_tac) was generated. It was then introduced into a wild-type strain (MG1655) harboring the low-copy-number plasmid pTB63, which carries the ftsQAZ operon and renders the Rod system nonessential due to the overproduction of the division protein FtsZ (29) (Fig. 1). Transformants were then plated on LB agar containing IPTG to induce expression of the mreCD genes and mecillinam to select for defects in the Rod system. Plasmids were isolated from mecillinam-resistant colonies and transformed back into the parental background to confirm that resistance was plasmid linked. The mreCD operon from such plasmids was then sequenced to identify the mutation conferring mecillinam resistance. In cases where the plasmid contained multiple mutations, the causative mutation was identified using site-directed mutagenesis to generate alleles with single changes, followed by assessment of dominant negative activity.

Alterations in two regions of MreC result in dominant negative activity.

A total of 13 dominant negative mreC alleles were identified based on their ability to confer mecillinam resistance to cells with the pTB63 (ftsQAZ) plasmid. Ten of these alleles encoded MreC proteins with a single amino acid substitution, and three encoded proteins with C-terminal truncations (Fig. 2). Based on a model of the E. coli MreC structure derived from the solved structure of the Listeria monocytogenes protein (35, 36), eight of the substitutions conferring dominant negative activity mapped to the interface between the alpha and beta domains (Fig. 2). The remainder of the substitutions mapped to the C-terminal domain, which was absent in the structure (Fig. 2). We refer to this region of MreC as the gamma domain.

FIG 2.

Dominant negative alleles of mreC that confer mecillinam resistance. (A) Cultures of strain MG1655/pTB63 (WT; ftsQAZ) harboring vectors expressing the indicated alleles of mreC were serially diluted and spotted on LB agar containing no drug, mecillinam (2.5 μg/ml), or mecillinam (2.5 μg/ml) and IPTG (50 μM), as indicated. Plates were incubated at 30°C for 16 h (no drug) or for 40 h (with mecillinam) prior to being photographed. The mreC alleles were expressed in the context of the mreCD operon under the control of the tac promoter (P_tac) from plasmids pHC800 (empty vector), pPR11 [mreC(WT)], pPR17 [mreC(L107P)], pPR44 [mreC(S110F)], pPR47 [mreC(R113P)], pPR46 [mreC(E116V)], pPR50 [mreC(G156D)], pPR55 [mreC(P230Q)], pPR20 [mreC(L267M)], pPR45 [mreC(V285M)], pPR49 [mreC(R292H)], pPR21 [mreC(V120-stop)], pPR22 [mreC(Y131-stop)], and pPR13 [mreC(W270-stop)]. “stop” refers to a premature stop codon. Note that some of the alleles confer mecillinam resistance without IPTG induction, which is presumably due to leaky basal expression from the P_tac promoter. (B) E. coli MreC, modeled based on a crystal structure from L. monocytogenes using Phyre2 (35, 36). Residues included in the indicated domains are as follows: transmembrane, 13 to 35; alpha domain, 36 to 110; beta domain, 118 to 270; and gamma domain, 271 to 367. Amino acid substitutions resulting in dominant negative activity are highlighted in red. See the text for details.

The defects caused by the amino acid substitutions in MreC were further characterized by testing their ability to induce a growth or shape defect in cells producing a normal level of FtsZ (Fig. 3). To test their ability to inhibit cell growth, the MreC variants were produced from the same P_tac-containing vector used to assess their ability to promote mecillinam resistance. A wide range of growth phenotypes was observed, with the overproduction of MreC(L107P), MreC(S110F), MreC(R113P), MreC(G156D), MreC(L267M), MreC(V285M), and MreC(R292H) inducing a particularly severe defect (Fig. 3A). Unexpectedly, overexpression of the remaining variants identified in the mecillinam selection resulted in a mild or no effect on the growth of cells with normal FtsZ levels (Fig. 3A). Because the overproduction of MreC(WT) (where WT means wild type) from the strong P_tac promoter resulted in shape defects, we switched to an expression vector with the weaker lactose promoter (P_lac) to assess the effects of the MreC variants on cell shape. At this lower level of overproduction, most of the variants were found to alter cell shape to some extent, with the MreC(S110F), MreC(R113P), and MreC(R292H) variants inducing the strongest cell shape defects, as indicated by their ability to induce cell widening and an appreciable reduction of the cellular aspect ratio (length/width) (Fig. 3B and C). The finding that some of the variants identified in the mecillinam selection resulted in little to no effect on growth or morphology when overproduced suggests that mecillinam toxicity is a more sensitive assay for Rod system function than monitoring cell shape or growth under conditions where the machinery is essential. In other words, only a minor defect in Rod system activity is required to promote survival during mecillinam challenge, whereas more severe defects are required to observe slowed growth and shape changes.

FIG 3.

Dominant negative growth and shape defects induced by MreC variants. (A) Overexpression of mreC dominant negative alleles is toxic to cells with normal levels of FtsZ. Cultures of MG1655 (WT) harboring vectors expressing the indicated alleles of mreC were serially diluted and spotted on LB agar with or without IPTG (100 μM) as indicated. Plates were incubated overnight at 37°C and photographed. The mreC alleles were expressed in the context of the mreCD operon under the control of the tac promoter (P_tac) from the same plasmids listed in the legend for Fig. 2. (B) Overexpression of dominant negative mreC alleles confers a shape change. Overnight cultures of wild-type cells (TB28) harboring vectors expressing the indicated alleles of mreC were diluted 1/200 in 5 ml of LB supplemented with 25 μg/ml chloramphenicol and 1 mM IPTG. When the OD₆₀₀ reached ∼0.4 to 0.5, the cells were fixed and then imaged by phase-contrast microscopy. Scale bar, 5 μm. Note that the mreC alleles were expressed in the context of the mreCD operon under the control of the lac promoter (P_lac) and therefore were expressed at lower levels than for the experiments in panel A that used the stronger tac promoter. Plasmids used were pMS5 [mreC(WT)], pMS7 [mreC(S110F)], pMS21 [mreC(R113P)], pMS20 [mreC(E116V)], pMS19 [mreC(K155E)], pMS10 [mreC(G156D)], pMS22 [mreC(P230Q)], pMS3 [mreC(L267M)], pMS8 [mreC(V285M)], pMS9 [mreC(R292H)], pMS6 [mreC(V120-stop)], and pMS1 [mreC(W270-stop)]. (C) The aspect ratio (length/width) of cells grown as described in the legend for panel B was quantified using MicrobeJ (49). The ratio calculated for each cell is plotted as a single dot, with the mean ratio and standard deviation of the analyzed population (>200 cells for each strain) indicated by the bars.

The set of MreC variants were also tested for their ability to complement the shape defect of strains deleted for mreC (Fig. 4). Generally, those variants with single amino acid substitutions that displayed the strongest dominant negative activity, MreC(L107P), MreC(G156D), and MreC(R292H), were the most defective in terms of their ability to complement the growth and shape defect caused by MreC inactivation. They failed to promote the growth of the mutant or even to partially ameliorate its shape defect at any of the inducer concentrations tested (Fig. 4). The truncated variants were similarly unable to restore growth or morphology to the mutant (Fig. 4), whereas the remaining missense alleles retained partial function. They restored growth of the mreC mutant on LB agar, with some of the alleles comparable to mreC(WT) in promoting growth at the lowest concentration of inducer tested [mreC(E116V), mreC(P230Q), mreC(L267M), mreC(V285M)], whereas others required a higher induction level to complement growth [mreC(S110F), mreC(R113P), mreC(K155E)] (Fig. 4). These variants also partially corrected the shape defect of the mreC mutant, converting the spherical/amorphous cells to short rods with a wider than normal diameter (Fig. 4). There was no apparent correlation between the location of the amino acid substitution on the protein and the severity of the phenotype, with substitutions like G156D at the alpha-beta interface and the R292H substitution in the gamma domain both among the changes causing a strong defect. We conclude that the interface between the alpha and beta domains of MreC as well as the gamma domain are important for its ability to promote Rod system function.

FIG 4.

Functionality of mreC alleles. (A) Cells of MT4 (ΔmreC ftsQAZ) expressing the indicated mreC alleles in the context of the mreCD operon under the control of the lac promoter (P_lac) were grown overnight in M9 glucose medium with chloramphenicol. The resulting overnight cultures were normalized for OD₆₀₀, spotted, and grown on the indicated media as described in the legend for Fig. 3. (B) Overnight cultures of the strains from panel A were grown in M9 glucose medium with 25 μM IPTG, diluted to an OD₆₀₀ of 0.05 in the same medium, and grown at 30°C to an OD₆₀₀ of 0.4 to 0.5. Cells were then fixed, immobilized, and imaged by phase-contrast microscopy. Bar, 5 μm. (C) Graph showing the aspect ratio quantified for a population of cells of each strain from panel B (n  > 200). The analysis was performed as described in the legend for Fig. 3. The same set of plasmids used in Fig. 3B was used for the complementation experiments shown here.

Potential regulatory activity of the MreC gamma domain.

The gamma domain of MreC is proline rich, and it had been previously noted that several Gram-negative species encode an MreC with similar C-terminal extensions (37). To better understand the phylogenetic distribution of this domain, we compared 20,301 protein sequences containing the MreC Pfam domain. We found that 12% of these sequences contained an identifiable gamma domain, which we defined as a sequence C-terminal to the MreC beta domain that is at least 45 residues long and has greater than 15% proline content (Fig. 5A). The domain is most prevalent among the Proteobacteria with an apparent enrichment among the Gammaproteobacteria (Fig. 5A). Outside of this clade, the gamma domain is largely absent from MreC homologues, with the exception of a subset of Actinobacteria that encode a recognizable MreC protein (Fig. 5A).

FIG 5.

Intragenic suppressors rescue the growth defect of an mreC(R292H) mutant. (A) Phylogenetic tree showing the distribution of species encoding MreC and the subset of these species in which MreC contains a gamma domain. (B) Strains containing the indicated mreC allele at the native genomic locus (HC555, PR5, PR72, PR73, PR74, and PR75) were serially diluted and spotted on either M9 agar supplemented with 0.2% glucose and Casamino Acids (Rod system nonessential) or LB agar (Rod system essential). Plates were incubated at 30°C for either 40 h (M9) or 16 h (LB) prior to being photographed. (C) Overnight cultures of the above-described strains were grown in M9 with 0.2% glucose and Casamino Acids, diluted to an OD₆₀₀ of 0.05 in the same medium, and grown to an OD₆₀₀ of 0.2. Cells were gently pelleted, resuspended, and diluted in LB (OD₆₀₀ = 0.025) and then grown until the OD₆₀₀ reached 0.2. At this time, cells were fixed, immobilized, and imaged using phase-contrast microscopy. All growth was at 30°C. Scale bar, 5 μm. (D) Cultures of MT4 (ΔmreC) harboring the mreD expression plasmid pFB128 plus vectors expressing the indicated mreC alleles under P_lac control were serially diluted and spotted on M9 agar supplemented with 0.2% glucose and Casamino Acids or LB agar supplemented with 50 μM IPTG as indicated. All plates also contained 25 μg/ml chloramphenicol and 100 μM spectinomycin to maintain the plasmids. M9 plates were incubated at 30°C for 40 h and LB plates were incubated at 37°C for 16 h before being photographed. Plasmids used were pPR66, pPR70, pMS1, and pPR108.

The enrichment of MreC proteins with gamma domains within the Proteobacteria suggests that this C-terminal region might play an important role in MreC function in these organisms. Accordingly, one of the most defective mreC alleles isolated in the selection for dominant negative mutants encodes MreC(R292H), which has a substitution in this region (Fig. 2B). We previously used allelic replacement to engineer the mreC(R292H) allele into the native mre locus (3). As expected for a mutant with a poorly functional Rod system, this strain lost its rod shape and failed to grow on rich medium (3) (Fig. 1). We took advantage of this growth phenotype to select for suppressors that restored growth of the mreC(R292H) mutant on rich medium (3). A subset of these suppressors were described previously and encoded variants of PBP2 that resulted in the hyperactivation of the RodA polymerase in vivo and in vitro (3). These suppressors suggested a role for MreC in the activation of PG synthesis by the Rod system through its interaction with PBP2 (3). Here, we report the characterization of additional suppressors of mreC(R292H), which have provided insight into the potential function of the MreC gamma domain.

The newly characterized set of mreC(R292H) suppressors were intragenic. Three of them encoded MreC variants with an additional amino acid substitution to overcome the defect caused by the R292H change: S110F, H286P, or L293R. Although the intragenic suppressors varied in their effectiveness, they all at least partially restored growth on rich medium and rod shape to the parental mreC(R292H) strain (Fig. 5B and C). Notably, the S110F substitution is located at the alpha/beta domain junction and was also identified as a change that alone resulted in an MreC variant with dominant negative activity (Fig. 2 and 5B). The other missense changes either restored an Arg residue (L293R) at a position neighboring the original R292H substitution or altered a residue within the gamma domain (H286P) in close proximity to the change at position 292. Although the mechanism by which these substitutions partially restore function to the MreC(R292H) variant is not clear, they suggest that the physiochemical properties of the gamma domain may be more important for MreC function than its exact sequence and that there may be a functional interplay between the gamma domain and the alpha/beta domain junction of the protein (see Discussion).

Surprisingly, an in-frame deletion allele encoding an MreC variant lacking residues 281 to 295, which removes the offending R292H substitution, was also isolated as a suppressor of mreC(R292H). The deletion variant was as good as, if not better than, the missense alleles at restoring growth and rod shape to the parental mreC(R292H) strain (Fig. 5B and C). This observation indicates that residues 281 to 295 are not critical for MreC function and prompted us to investigate whether the gamma domain is generally dispensable for Rod system activity. To test the gamma domain requirement, we made expression vectors that produce different C-terminal truncations of the 367-amino-acid MreC protein. Production of MreC(1–278) lacking the vast majority of the gamma domain was found to be sufficient to complement the growth defect of a ΔmreC strain, whereas a slightly larger truncation, MreC(1–270), was only partially functional (Fig. 5D). Because the R292H substitution within the gamma domain abolishes MreC function, but the domain as a whole is dispensable, we conclude that the R292H substitution causes the gamma domain to adopt a conformation that inhibits Rod system activity.

Characterization of MreD variants with dominant negative activity.

Following PCR mutagenesis of the entire mreCD operon in the expression plasmid, the vast majority of the dominant negative mutants identified had changes in mreC. To enrich for mreD mutations, we repeated the mecillinam resistance selection using a plasmid library carrying wild-type mreC and PCR-mutagenized mreD. Most of the dominant negative alleles of mreD encoded truncations, including frameshift or nonsense mutations at codons 27, 34, 55, 60, 101, 102, 109, 129, 133, 142, 145, and 150. Additionally, a dominant negative frameshift allele was even isolated in the third to last codon (codon 160) of mreD, resulting in an 11-amino-acid C-terminal extension following the small deletion. The elevated mecillinam resistance phenotypes resulting from the overproduction of three representative C-terminal truncations are shown in Fig. 6. Although we had hoped that the selection design would help us avoid the isolation of mutants encoding significantly truncated proteins, the observation that truncations of MreD that produce proteins including only the first 27 or 34 amino acids have dominant negative activity suggests that its N-terminal region is sufficient to interface with at least one other Rod system component to block binding with native MreD. These residues are predicted to encompass just the first transmembrane domain of MreD, suggesting that this binding partner is one of the transmembrane components of the elongation machinery. Similarly, the dominant negative activity of the frameshift mutation at codon 160 indicates that the extreme C terminus of MreD may also be required for normal Rod complex formation or activity.

FIG 6.

Dominant negative alleles of mreD that confer mecillinam resistance. Cultures of strain MG1655/pTB63 (WT; ftsQAZ) harboring vectors expressing the indicated alleles of mreD were serially diluted and spotted on LB agar containing no drug, mecillinam (2.5 μg/ml), or mecillinam (2.5 μg/ml) and IPTG (50 μM), as indicated. Plates were incubated at 30°C for 16 h (no drug) or 40 h (with mecillinam) prior to being photographed. The mreD alleles were expressed in the context of the mreCD operon under the control of the tac promoter (P_tac) from plasmids pPR11 [mreCD(WT)], pMS28 [mreCD(I69F)], pMJQ12 [mreCD(T73M)], pJQ11 [mreCD(K91T)], pMS39 [mreCD(F34-fs)], pMS45 [mreCD(W101-stop)], and pMS46 [mreCD(E133-stop)]. fs and stop, frameshift and premature stop codons, respectively.

In addition to the truncations, three dominant negative missense alleles of mreD were isolated that resulted in single amino acid substitutions in the encoded protein (Fig. 6). In a modeled structure of MreD, two of the changes, I69F and T73M, are predicted to lie in an extracytoplasmic loop, whereas the third change (K91T) is predicted to be located near a cytoplasmic loop (Fig. 7A). MreD is predicted to have remote structural similarity to the ligand-binding subunit (S component) of energy-coupling factor ABC transporters (Fig. 7A) (38–40). Notably, the I69F and T73M substitutions are predicted to lie near the putative ligand-binding pocket (Fig. 7A). A derivative with both substitutions, MreD(I69F, T73M), induced a dominant negative mecillinam resistance phenotype (Fig. 7B). In complementation experiments, the individual mutants appeared to retain partial function (Fig. 7C). However, when they were combined, expression of mreD(I69F, T73M) was incapable of correcting the growth or shape defect of a strain deleted for mreD (Fig. 7C and D). Because the double mutant retained dominant negative activity, we infer that it produces a stable protein and that the substitutions I69F and T73M result in a functional defect.

FIG 7.

Location of and phenotypes induced by MreD missense alleles. (A) Left, predicted structure of E. coli MreD, using the HHPred template selection tool and iTASSER comparative protein structure modeling (38, 50). Residues critical for MreD function are highlighted in red. Right, predicted structure of E. coli MreD (gray) aligned with the RibU S-component (PDB ID 3P5N) from Staphylococcus aureus (51) (blue) with the riboflavin ligand in yellow. (B) Cultures of strain MG1655/pTB63 (WT; ftsQAZ) harboring vectors expressing the indicated alleles of mreD were diluted and plated as described in the legend for Fig. 6. The mreD alleles were expressed in the context of the mreCD operon under the control of the tac promoter (P_tac) from plasmid pPR11 or pJQ17. (C) Overnight cultures of strain JQ23 (ΔmreD) harboring plasmids producing the indicated MreD variant from an mreCD operon under the control of the lac promoter (P_lac) were normalized for OD₆₀₀, serially diluted, and plated on LB agar plates with or without IPTG (5 μM) as indicated and grown at 30°C. Plasmids used were pPR66 (empty vector), pMS5 [mreC(WT)], pJQ16 [mreD(I69F)], pJQ14 [mreD(T73M)], and pJQ19 [mreD(I69F, T73M)]. (D) Overnight cultures of a subset of the strains from panel C were grown in M9 arabinose medium, diluted to an OD₆₀₀ of 0.05 in the same medium supplemented with 50 μM IPTG, and grown at 30°C to an OD₆₀₀ of 0.5 to 0.6. Cells were then fixed, immobilized, and imaged by phase-contrast microscopy. Bar, 5 μm. The graph on the right shows the aspect ratio quantified for a population of cells of each strain (n > 200). The analysis was performed as described in the legend for Fig. 3.

To further investigate the functional importance of the putative ligand-binding pocket in MreD, site-directed mutagenesis was used to generate variants of green fluorescent protein (GFP)-MreD with additional alterations in this domain. Five GFP-MreD derivatives were constructed with either a dramatic change in the size of an amino acid side chain or a charge substitution/change, and they were tested for their ability to complement the growth defect of a ΔmreD strain (Fig. 8A and B). Unlike GFP-MreD(WT), three of the variants were unable to complement the deletion (Fig. 8A and B). Immunoblotting for GFP demonstrated that two of these derivatives, GFP-MreD(D67K) and GFP-MreD(M81K), produced amounts of protein comparable to that of GFP-MreD(WT), indicating that they are likely to be functionally defective. We therefore conclude that residues in the putative ligand-binding pocket of MreD are required for its function.

FIG 8.

Mutagenesis of the predicted ligand-binding pocket of MreD. (A) Overnight cultures of strain JQ23 (ΔmreD) harboring plasmids producing the indicated GFP-MreD variant were normalized for OD₆₀₀, serially diluted, and plated on M9-glucose or LB agar plates with or without IPTG (500 μM) as indicated and grown at 30°C. Plasmids used were pTU242 [GFP-MreD(WT)], pJQ31 [GFP-MreD(W120A)], pJQ37 [GFP-MreD(D115K)], pJQ33 [GFP-MreD(D67K)], pJQ36 [GFP-MreD(V39K)], and pJQ40 [GFP-MreD(M81K)]. (B) Full view and close-up images of the predicted structure of MreD (gray) with the mutated residues highlighted in green. The position of the riboflavin ligand from RibU (PDB ID 3P5N) is depicted in yellow. (C) Anti-GFP immunoblot of GFP-MreD and its derivatives. Each lane contains protein from exponential-phase (OD₆₀₀ = 0.3) whole-cell extracts of JQ23 (ΔmreD) harboring integrated plasmids expressing GFP-MreD(WT) and its derivatives grown with or without 500 μM IPTG as indicated. Plasmids used were pTU242 [GFP-MreD(WT)], pJQ31 [GFP-MreD(W120A)], pJQ37 [GFP-MreD(D115K)], pJQ33 [GFP-MreD(D67K)], pJQ36 [GFP-MreD(V39K)], and pJQ40 [GFP-MreD(M81K). The band corresponding to the molecular weight of GFP-MreD (47 kDa) is indicated by an arrowhead.

DISCUSSION

The roles played by MreC and MreD within the Rod system have remained unclear, even though the structure of a large portion of the MreC periplasmic domain was solved many years ago (36, 37). One of the MreC beta-sheet domains has been observed to share structural similarity with alpha-lytic protease. However, mutagenesis studies found that residues predicted to be important for protease activity were not required for MreC function, suggesting that it is not an active protease (36). Instead, due to the lack of an identified enzymatic function, its relative abundance in cells, detected two-hybrid interactions with several different PBPs (36, 41), and structural models indicating a potential for filament formation (36, 37), many models for MreC function propose that it plays primarily a structural role in the Rod system linking PBPs to the rest of the machinery (41, 42). Similarly, MreD has been proposed to act as a scaffold, bridging interactions between the membrane-bound Rod system proteins and cytoplasmic proteins like MreB and enzymes that synthesize PG precursors (43). Our recent genetic analysis of Rod system regulation combined with a prior structural study of an MreC-PBP2 complex has suggested an alternative function for MreC, one in which it plays a role in activating PG polymerization by RodA through its interaction with PBP2 (3). By analogy, we propose that MreD may also be functioning as a regulatory component within the elongation system.

One barrier to defining the functions of MreC and MreD has been the lack of an unbiased genetic system to identify amino acid residues critical for their activity. In this report, we describe such a system and use it to help define functionally important regions of these conserved cell elongation factors. For MreC, a number of dominant negative mutants had substitutions that mapped to either the interface of the alpha-helical and beta-sheet-rich regions of its large periplasmic domain or its proline-rich C terminus. Because they induce a dominant negative phenotype, the mutants must stably accumulate and retain some ability to interact with other components of the Rod system. Accordingly, in our initial characterization of MreC(G156D) and MreC(R292H) for use in investigating Rod system activation, we found that both proteins accumulated to levels similar to the wild-type protein when they were produced from the native mreC locus (3). Thus, we think it is reasonable to assume that many of the MreC variants identified are functionally impaired rather than being compromised in their ability to accumulate in vivo.

Notably, none of residues implicated in MreC function are located in the MreC-MreC homodimer or MreC-PBP2 heterodimer interfaces identified in structural studies (36, 37, 42). Because both of these interfaces are quite extensive, it may be difficult to disrupt them with a single amino acid substitution, especially one caused by a single nucleic acid change. However, it is possible that one or more of the substitutions identified alters the conformation of MreC to disrupt the MreC-MreC or MreC-PBP2 interaction interface while leaving interactions with other Rod system components intact, thereby inactivating the complex. Additional biochemical and structural analyses are needed to explore these possibilities.

Further support for MreC functioning in a regulatory capacity was revealed by the isolation of intragenic suppressors of the mreC(R292H) mutant. Surprisingly, the R292H substitution was found to induce a Rod system defect, whereas removal of essentially all of the gamma domain where it resides was tolerated. This observation suggests that the gamma domain may control MreC activity by toggling between an inhibitory and a noninhibitory state and that the R292H change locks it in the inhibitory form. Additionally, the finding that the R292H defect is suppressed by an S110F substitution, which is located at the junction of the alpha and beta domains and itself causes a Rod system defect, indicates that the potential regulatory activity of the gamma domain may be mediated in part by a direct or indirect connection with the alpha-beta domain interface. What factors or signals might control the activity of the gamma domain remain unclear. One attractive possibility is that through interactions with other proteins or the PG itself, this domain may help identify areas of the envelope where the insertion of new material is needed. Alternatively, because the domain is primarily conserved among the Proteobacteria, it may facilitate the coordination of Rod system activity with other aspects of envelope biogenesis like outer membrane biogenesis.

Functionally important regions of MreD were also revealed by the genetic analysis. Most of the dominant negative mreD alleles contained frameshifts or premature stop codons. These mutations were found throughout the coding sequence of mreD, with N-terminal fragments as short as 26 amino acids conferring a dominant negative phenotype. This result suggests that the first transmembrane helix is sufficient to interact with at least one other Rod system subunit. Dominant negative missense alleles of mreD were also identified, two of which encoded substitutions in a periplasmic loop of MreD located near a predicted ligand-binding pocket in the protein (39). Site-directed mutagenesis was used to generate additional substitutions in this region of the protein. Two of the resulting variants, MreD(D67K) and MreD(M81K), failed to complement the growth defect of an mreD deletion but produced stable protein, indicating that they are functionally defective and that the putative ligand-binding pocket is required for MreD activity. Further work is required to investigate whether a ligand indeed binds in this pocket and, if so, to determine its identity. Nevertheless, the results suggest the attractive possibility that MreD function is modulated by a small molecule to control cell elongation by the Rod system.

In conclusion, we have identified functionally important residues in both MreC and MreD. The mutants identified have already been useful in uncovering a role for MreC in the activation of the RodA-PBP2 PG synthase (3) and in identifying a potential regulatory role for its C-terminal gamma domain. We anticipate that the future use of these and other variants of the MreC and MreD proteins identified here will continue to yield new insights into the function of these widely distributed yet enigmatic components of the cell elongation machinery.

MATERIALS AND METHODS

Media, bacterial strains, plasmids, and growth conditions.

All of the reported experiments were carried out in derivatives of E. coli strain MG1655 (44). Cells were grown at 37°C in lysogeny broth (1% [wt/vol] tryptone, 0.5% [wt/vol] yeast extract, 0.5% [wt/vol] NaCl) (LB) or minimal M9 medium (45) supplemented with 0.2% Casamino Acids (CAA) and 0.2% sugar, as indicated. Plasmids were maintained in DH5α(λpir). Whenever necessary, antibiotics were used at 50 (ampicillin [Amp]), 25 (chloramphenicol [Cm]), 50 (kanamycin [Kan]), or 7.5 (tetracycline [Tet]) μg/ml. Strains and plasmids used in this study are listed in Tables S1 and S2 in the supplemental material, respectively. Plasmid construction procedures are detailed in the supplemental material. Culture conditions for each experiment are described in the relevant figure legends.

Selection for dominant negative Rod system mutants.

To make a plasmid library that expresses mutagenized mreCD from the P_tac promoter, the mutagenic polymerase Pfu(D473G) and primers 5′-GTCATCTAGACTGCCTGGTCTGATACGAGAATACGCATAACTTATG-3′ and 5′-GTCAAAGCTTTTATTGCACTGCAAACTGCTGACGG-3′ were used to amplify XbaI-nativeRBS-mreCD-HindIII from a plasmid template, as described by Biles and Connolly in 2004 (46). This insert was cloned into the pHC800 vector using restriction enzymes XbaI and HindIII. Ligated plasmids were transformed into DH5α(λpir), yielding ∼800,000 colonies, about half of which contained empty vectors and half of which contained vectors with the mutagenized mreCD insert. Plasmids were isolated from scraped and pooled transformant colonies, and the plasmid library was transformed into MG1655/pTB63.

A second library in which only mreD was mutagenized was generated by amplifying mreC-mreD-HindIII with Pfu(D473G) and primers 5′-GTCATCTAGACTGCCTGGTCTGATACGAGAATACGCATAACTTATG-3′ and 5′-CAGCGAGTGGCAACAGATCATGT-3′. This insert was cloned into the pPR11 (P_tac::mreCD) vector using restriction enzymes BglI and HindIII. There is a naturally occurring BglI cut site at the 3′ end of mreC. The library was passaged through DH5α(λpir) and then transformed into MG1655/pTB63 as for the original library.

A third library was generated in which mutagenized mreD was cloned into the pJQ5 (P_tac::mreCD-phoA) vector. The in-frame PhoA fusion facilitated the identification of plasmids encoding full-length MreC-MreD through the identification of survivors forming blue colonies on medium with the PhoA substrate 5-bromo-4-chloro-3-indolyl phosphate (BCIP). The insert mreC-mreD-XhoI was amplified with Pfu(D473G) and primers 5′-CGCTCTCCACAAAGGGCTAC-3′ and 5′-TGATAAGCTTGCGCCACCCTCGAGTTGCACTGCAAACTGCTGACG-3′ and cloned into pJQ5 using restriction enzymes BglI and XhoI. The library was passaged through DH5α(λpir) and then transformed into CB373/pTB63 (MG1655 ΔlacIZYA ΔphoA/ftsQAZ).

To select for cells expressing a dominant negative allele of mreC or mreD, the libraries were plated on LB supplemented with Cm, Tet, mecillinam (2.5 μg/ml), and IPTG (50 μM) at 30°C. In the case of the third library, the medium also contained 40 μg/ml of BCIP. The libraries were also plated on medium lacking mecillinam to determine the efficiency of plating. Mecillinam-resistant colonies were replica streaked on LB, LB-mecillinam, and LB-mecillinam-IPTG. Plasmids were isolated from transformants exhibiting IPTG-dependent mecillinam resistance and retransformed into MG1655/pTB63 to ensure that the mecillinam resistance phenotype was linked to the plasmid. Dominant negative alleles of mreCD were sequenced. In cases where the insert contained multiple mutations, site-directed mutagenesis was used to identify the causative mutation.

Phylogenetic tree construction.

A phylogenetic tree of all bacterial species encoding MreC was retrieved from http://annotree.uwaterloo.ca/ by searching for the Pfam ID PF04085, using a P value of 0.00001 (47). The full-length MreC sequences were downloaded from AnnoTree and then uploaded to the Galaxy web platform (48). We used the public server at https://usegalaxy.org to make a rough alignment using the MAFFT multiple alignment program with default parameters. The MreC Pfam domain spans from residues 120 to 269. An MreC sequence was considered to have a gamma domain if >45 amino acids with >15% proline content aligned to residues 270 to 367 of E. coli MreC. The TaxIDs of species with an MreC gamma domain were entered into AnnoTree for phylogenetic visualization, and the resulting graphic was manually superimposed on the tree of all species encoding the MreC Pfam domain.

Selection for suppressors of mreC(R292H).

The mreC(R292H) intragenic suppressor mutants were isolated using a previously described selection strategy (3). Briefly, PR5 [mreC(R292H)] cells were grown under permissive conditions (M9-CAA-glucose, 30°C) and then plated under nonpermissive conditions (LB or LB with 1% SDS). Cells that could grow under these nonpermissive conditions were screened for restoration of Rod shape and then subjected to whole-genome sequencing.

Image acquisition and analysis.

Growth conditions prior to microscopy are described in the relevant figure legends. Unless otherwise indicated, cells were fixed in 2.6% formaldehyde with 0.04% glutaraldehyde at room temperature for 1 h, followed by storage at 4°C for up to 3 days. Prior to imaging, cells were immobilized on 2% agarose pads containing the appropriate growth medium and covered with no. 1.5 coverslips. Phase-contrast microscopy was performed on a Nikon TE2000 microscope equipped with a 100× plan apo 1.4 numerical aperture (NA) objective, 0.90 NA condenser lens, and a CoolSNAP HQ2 monochrome camera (Photometrics). Images were acquired using NIS Elements AR 3.2 software.

Construction of GFP-MreD variants and immunoblotting.

Site-directed mutagenesis of plasmid pTU242 (P_lac::gfp-mreD) to generate constructs for the production of GFP-MreD variants with substitutions in the putative ligand-binding pocket were generated by GenScript. For immunodetection of the fusion proteins, cultured cells at an optical density at 600 nm (OD₆₀₀) of 0.3 were pelleted and resuspended in water and 2× Laemmli sample buffer (100 mM Tris-HCl [pH 6.8], 2% SDS, 0.1% bromophenol blue, 20% glycerol) at a 1:1 ratio. Samples were boiled for 10 min at 100°C and sonicated (Qsonica tip sonicator; amplitude, 25%; time, 1 min) two times. Protein concentration was measured using the NI (noninterfering) protein assay (with bovine serum albumin [BSA] protein standard) (G Biosciences catalog no. 786-005). Sample concentrations were then normalized using 1× sample buffer. Samples were run on a 4 to 20% Mini-Protean TGX precast protein gel and transferred to a polyvinylidene difluoride (PVDF) membrane. The membrane was then rinsed in phosphate-buffered saline plus 0.1% Tween (PBS-T) and blocked in 5% milk in PBS-T for 1.5 h. The membrane was then incubated in primary antibody solution of 1% milk in PBS-T plus goat anti-GFP antibody at a 1:5,000 dilution for approximately 16 h at 4°C. The membrane was then washed four times in PBS-T (one time quickly and three times for 10 min per wash). The membrane was then incubated in secondary antibody solution (anti-goat IgG–horseradish peroxidase [HRP], 1:5,000 dilution; Rockland catalog no. 18-8816-33) in 0.2% milk in PBS-T for 2 h. Following five washes with PBS-T, the membrane was developed using SuperSignal West Pico Plus chemiluminescent substrate (Thermo Fisher Scientific catalog no. 34577) and imaged using the c600 Azure Biosystems platform.