IF proteins in eukaryotes share a rod-like domain between two variable domains. The most common IF proteins are keratin and lamin. They self-assemble in a nucleotide-independent manner (Lowe & Amos, 2009). The most studied IF-like bacterial protein is crescentin (CreS), from Caulobacter crescentus (Figure 7A). It was the first IF-like protein to be discovered, in 2003 (Ausmees, Kuhn, & Jacobs-Wagner, 2003). It has mild sequence similarity with eukaryotic IF, but does not depend on nucleotides to form slightly curved filaments in vitro, as eukaryotic IFs. Lack of CreS causes Caulobacter to lose its typically curved form and to become a straight rod (Ingerson-Mahar & Gitai, 2012).
The main similarity between IF and CreS is that they are coiled coil-rich proteins. There are a number of coiled coil-rich proteins in bacteria that self-assemble into filaments in a nucleotide-independent manner. Some act as scaffolds or localization factors of other proteins, but their exact functions are still unknown (Lin & Thanbichler, 2013). FilP is an example of a coiled coil-rich protein from Streptomyces coelicolor that plays a role in the formation of hyphae (Figure 7B). Deletion of filP induces abnormal hyphal morphology and decreased stability (Cabeen & Jacobs-Wagner, 2010). Other examples are RsmP from Corynebacterium glutamicum and Ccrp from Bdellovibrio bacteriovorus (Ingerson-Mahar & Gitai, 2012).
Endosomal Sorting Complexes Required for Transport proteins (ESCRTs)
ESCRT proteins are present in eukaryotic cells where they are involved in the late stages of mammalian cytokinesis. They are also present in archaea and some bacterial species like Chlamydia that do not have FtsZ (Ingerson-Mahar & Gitai, 2012). ESCRTs may be linked with cell division.
Prokaryotic-specific cytoskeleton proteins
It was believed that cytoskeletal proteins fitted into the previous three groups: actin, tubulin and intermediate filaments. Nevertheless, other cytoskeletal proteins have been discovered that could have a role in cellular organization and cell form, only present in bacterial cells.
Walker A Cytoskeletal ATPases (WACAs)
Walker A Cytoskeletal ATPases (WACAs) are a subfamily of the P-loop NTPase family; which includes GTPases, signal recognition particle proteins and eukaryotic cytoskeleton-associated proteins. WACAs are found in most bacteria and some archaea. The two main examples of the WACA proteins are ParA and MinD (Ingerson-Mahar & Gitai, 2012).
ParA is a plasmid and chromosome segregation protein that determines the position of other proteins in the cell (Pogliano, 2008). In its free state, it is a monomer that dimerizes and then polymerizes when bound to ATP. It also associates with additional factors like DNA and ParB, that stimulates ATP hydrolysis and the dissociation of the filament (Ingerson-Mahar & Gitai, 2012). It seems that in Caulobacter and Vibrio cholerae, ParA binds the parS origin-proximal region and its polymerization-depolymerization pulls the chromosome through the cell during its segregation. There are other ParA homologs implicated in plasmid segregation (Ingerson-Mahar & Gitai, 2012). MinD polymerizes at the inner leaflet of the cytoplasmic membrane, attached to it by an amphipatic helix, in an ATP-dependent manner (Figure 8) (Lowe & Amos, 2009). It associates with MinC and, together, they inhibit FtsZ polymerization, acting as spatial regulators of the site of cell division (Lutkenhaus, 2007). In E. coli, MinE binds to MinD and promotes ATP hydrolysis, creating a gradient of MinD bound to the membrane and allowing the polymerization of FtsZ at midcell. B. subtilis lacks MinE, but it still recruits MinD to the cell poles and recent division sites.
C. crescentus does not have the Min system, but has a different WACA protein, namely MipZ, that disables the mislocalization of FtsZ (Ingerson-Mahar & Gitai, 2012). At the beginning of the cell cycle, MipZ binds the origin of replication (oriC) of the chromosome via ParB. After DNA replication, one of the daughter origins migrates to the opposite cell pole with MipZ, activating the release of FtsZ from the area. This mechanism creates a gradient of proteins from both cell poles, directing FtsZ assembly at mid-cell (Pogliano, 2008).
Biochemical properties of MreB
MreB biochemistry has proven to be a though area of investigation. It is a difficult protein to purify in its soluble, active form. This explains why for a long time MreB’s biochemistry was entirely studied on the thermophilic organism T. maritima (Bean & Amann, 2008; Esue, Cordero, Wirtz, & Tseng, 2005; Popp et al., 2010). With years, more MreBs have been purified in E. coli, Leptospira interrogans, Magnetospirillium gryphiswaldense, Candidatus Magnetobacterium casensis (Mcas), Chlamydophila pneumoniae (Barko et al., 2016; Bean & Amann, 2008; Deng et al., 2016; Gaballah, Kloeckner, Otten, Sahl, & Henrichfreise, 2011; Nurse & Marians, 2013; Sonkaria et al., 2012). Still, only a single -and controversial- publication has been released in a Gram-positive organism, B. subtilis (Mayer & Amann, 2009). Table 3 summarizes the biochemical properties of the MreBs studied and mentioned above.
Table 3 Summary of biochemical properties of MreB in different organisms. (1) Dependence on ATP for polymerization. (2) Dependence on Mg for polymerization. (3) Dependence on monovalent ion; max., maximal polymerization; Inh., polymerization inhibited. (4) pH at which polymerization was maximal (max.) or inhibited (inh.). (5) Method used to differentiate between aggregation and polymerization; EM, electron microscopy; FL, fluorescent labelling. (6) (Nurse & Marians, 2013). (7) (Bean & Amann, 2008). (8) (Barko et al., 2016). (9) (Gaballah et al., 2011). (10) (Sonkaria et al., 2012). (11) (Deng et al., 2016). (12) (Mayer & Amann, 2009).
The only studies that didn’t differentiate polymerization from aggregation by further analyzing light scattering and sedimentation results where those performed on B. subtilis MreB (Bs-MreB) and C. pneumoniae MreB (Cp-MreB) (Gaballah et al., 2011; Mayer & Amann, 2009). What is most striking is that these two studies are the only ones to defend that MreB’s polymerization/aggregation is non-dependent on ATP. It is very interesting to note that A22, an inhibitor of MreB polymerization, does not affect neither Bs-MreB (Noguchi et al., 2008) nor Cp-MreB (Gaballah et al., 2011) while it 35 inhibits MreB polymerization in many other organisms (Noguchi et al., 2008). A22 binds to the nucleotide binding (NB) pocket of MreB preventing ATP binding and inhibiting MreB polymerization (Bean et al., 2009). The NB pocket is highly conserved in all MreBs, however, the protein conformation could vary in each MreB in a way that it prevents A22 binding to this NB pocket. C. pneumoniae is a coccoid, intracellular bacteria and no peptidoglycan has been reliably detected in its CW. It is thought that MreB is needed in these bacteria to maintain the proper functionality of the divisome.
Localization and dynamics of MreB
MreB’s localization is one of its most polemic aspects. The subcellular localization of MreB was first performed in 2001 by the Errington group (Jones et al., 2001). This work, together with results from another group on MreB, Mbl and MreBH of B. subtilis (Defeu Soufo & Graumann, 2004), described the MreB proteins to form filaments in the cytoplasm, very close to the membrane. These filaments followed a helical pattern along the cell long axis and were dynamic. Further confirmation came from studies on MreB from E. coli and C. crescentus (Figge, Divakaruni, & Gober, 2004; Kruse, Moller-Jensen, Lobner-Olesen, & Gerdes, 2003). However, a decade after the first localization study of MreB was published, three independent groups reported results contradicting the helical pattern of MreB and the formation of micrometer long filaments (Dominguez-Escobar et al., 2011; Garner et al., 2011; van Teeffelen et al., 2011). They observed MreB homologs forming discrete patches that moved along the cell circumference, perpendicularly to the cell long axis, in B. subtilis and in E. coli. These results were supported by works from the Jensen group who, moreover, demonstrated that the strain used to study MreB in E. coli, MC1000/pLE7, caused artifacts due to the YFP fusion of the protein (Swulius et al., 2011; Swulius & Jensen, 2012). Since then, new publications have argued against this model, showing high resolution pictures of long structures (Olshausen et al., 2013). Based on these, a new model reviving « the helices » was proposed (Errington, 2015). There is still an ongoing conflict between helices vs. perpendicular tracks and filaments vs. diffraction-limited clusters. It must be noted that, because of the resolution power of microscopes, a « diffraction limited » object only means that it is the smallest form distinguishable by a microscope. As a consequence, everything smaller than this limit (~ 300 nm), no matter its shape or level of organization, will look like a globular 300 nm patch and smaller filaments will also look as round blobs. Therefore, the debate between filaments and patches is not about the existence of filamentous structures but rather about their size.
Role of MreB in cell shape determination and cell wall synthesis
The most obvious phenotype and probably the primary function of MreB, seems to be the control of rod shape during elongation, although its precise role and way of action are not known. Curiously, most non rod-shaped bacteria lack mreB, but have mreC and mreD. We could hypothesize that MreB is linked to lateral CW elongation in rods and that is why it is absent in non-rods, but there are some exceptions of coccoids with MreB (Ouellette, Karimova, Subtil, & Ladant, 2012) and this explanation wouldn’t fit those cases. Anyway, from pioneering studies on MreB, the shape defect of mreB was reported (Doi et al., 1988; Wachi et al., 1987). These mutants have a characteristic phenotype, they bulge and curve, even forming telephone cord-like cell chains, and, finally, they lyse (Figure 10). A study from the Shaevitz group (Ouzounov et al., 2016) uses sub-lethal concentrations of A22 (a drug that inhibits MreB polymerization) and a series of MreB mutants to study the polymerization properties of MreB and its effect in cell shape in E. coli. They obtained a number of MreB point mutations that caused a higher resistance to A22. This allowed them to find two exciting correlations, the strongest being between cell diameter and MreB polymer angle and the second between cell diameter and polymer number. Interestingly, their data shows that a reduction of the helical angle of MreB entails an increase of cell width.
Cell wall integrity is tightly linked to cell morphology. An impressive study from (Ursell et al., 2014) measures simultaneously cell shape dynamics, CW insertion and cytoskeletal localization in E. coli.
Their conclusions are that MreB’s localization is biased to negatively curved areas along the cell cylinder, directing PG insertion to those areas, causing the straightening of the cell. There are multiple evidences of MreB from different organisms (including B. subtilis) interacting with Mur and DAP proteins as well as with proteins from the CW machinery (Divakaruni, Baida, White, & Gober, 2007; Favini-Stabile, Contreras-Martel, Thielens, & Dessen, 2013; Gaballah et al., 2011; Rueff et al., 2014; White, Kitich, & Gober, 2010). Kawai and co-workers showed that the localization and correct function of PBP1 is MreB-dependant (Kawai, Daniel, et al., 2009). Also, in pulldown assays, both MreB and Mbl associate with PBP1, PBP2a, PBP4 and possibly PBP5, independently (Kawai, Asai, et al., 2009). Furthermore, PbpH, PBP2a, MreC, MreD and RodA move similarly and colocalize with the MreB/Mbl/MreBH complex (Dominguez-Escobar et al., 2011; Garner et al., 2011). Further validation of these results comes from single-molecule experiments that revealed how PBP2 and MreB colocalize transiently to coordinate cell wall synthesis (Lee et al., 2014). Their results show how, by PBP2 having a diffusive motion and MreB following directed paths, their transitory interactions are beneficial to buffer growth through variable enzyme abundances and changing environmental conditions. It is possible that MreB acts as the link between PG precursor synthesis and its insertion to the CW.
Table of contents :
1.1. Bacterial envelope
1.1.1 Cytoplasmic membrane
1.1.2 Cell wall
1.1.3 Outer membrane
1.2.1 Peptidoglycan structural models
1.2.2 Peptidoglycan synthesis
1.3. Bacterial cytoskeleton
1.3.1 Bacterial cytoskeletal proteins
1.3.2 Bacterial actin homologs
1.4.1 MreB isoforms
1.4.2 Biochemical properties of MreB
1.4.3 Localization and dynamics of MreB
1.4.4 Role of MreB in cell shape determination and cell wall synthesis
1.4.5 Other roles of MreB
1.5. Aims of the thesis
2. Materials and Methods
2.1. Media Composition:
2.2. Media supplements:
2.3. Strains and plasmids
2.4. Experimental procedures
2.4.1 Cloning procedures:
2.4.2 Manipulation in B. subtilis
2.4.3 Protein procedures:
2.4.4 RNA procedures:
2.4.5 Microscopy methods:
3.1. Functional analysis of ydcF, ydcG and ydcH.
3.1.1 The ydcFGH operon is composed of three genes of unknown functions
3.1.2 Construction of knock-out mutants of ydcF, ydcG and ydcH
3.1.3 Phenotypic characterization of ydc genes exposes an inappropriate strain frame
3.1.4 YdcF, YdcG and YdcH are not involved in stress resistance
3.2. Transcriptional study of ydcFGH
3.2.1 ydcH is under the control of two promoters
3.2.2 YdcH, but not YdcF nor YdcG, is involved in the control of Pydc1 expression
3.2.3 The absence of MreB is not responsible for Pydc1 induction
3.3. YdcH, a new regulator for carbon metabolism?
3.4. MreB mutagenesis
3.4.1. Setting up a genetic screen for MreB loss-of-function mutants
3.4.2. Random mutagenesis of mreB
3.4.3. Site directed mutagenesis of mreB
3.4.4. Phenotypic characterization reveals different categories of MreB*s
3.4.5. Growth defect of WeB and ΔmreB mutants can be suppressed by addition of fructose
4.1 YdcH: a repressor/activator MarR transcription regulator?
4.2 YdcH: a new transition state regulator
4.3 A library of MreB mutants with impaired functionality
4.4 MreB may play a role in CW synthesis, cell morphology and cell metabolism
4.5 Some MreB*s have atypical colony morphologies
4.6 Possible connection between the mreB deletion and the ydcH frame-shift
Appendix 1: Phenotypic analysis of ydcFGH
A1.1 ydcF, -G and –H deletion mutants are not impaired for cell morphology
A1.2. Defects during stationary phase
Appendix 2: The absence of MreB is not responsible for Pydc1 induction
A2.1. Absence of mreB complementation is not due to chromosomal positioning of the gene
A2.2. ydcFGH induction is not due to decreased expression of minC
A2.3. ydcFGH induction is not caused by the expression of a remnant peptide of MreB
A2.4. ydcFGH induction is not caused by abnormal levels of MreCD
A2.5. Absence of the MreB protein is not the cause of ydcFGH induction
A2.6. ydcFGH induction is unlinked to the mreB locus
Appendix 2: Differentially expressed genes in the ΔydcH strain
Appendix 4: MreB*s TIRFM acquisitions