Identification of essential alphaproteobacterial genes reveals operational variability in conserved developmental and cell cycle systems

Abstract

The cell cycle of Caulobacter crescentus is controlled by a complex signalling network that co-ordinates events. Genome sequencing has revealed many C. crescentus cell cycle genes are conserved in other Alphaproteobacteria, but it is not clear to what extent their function is conserved. As many cell cycle regulatory genes are essential in C. crescentus, the essential genes of two Alphaproteobacteria, Agrobacterium tumefaciens (Rhizobiales) and Brevundimonas subvibrioides (Caulobacterales), were elucidated to identify changes in cell cycle protein function over different phylogenetic distances as demonstrated by changes in essentiality. The results show the majority of conserved essential genes are involved in critical cell cycle processes. Changes in component essentiality reflect major changes in lifestyle, such as divisome components in A. tumefaciens resulting from that organism's different growth pattern. Larger variability of essentiality was observed in cell cycle regulators, suggesting regulatory mechanisms are more customizable than the processes they regulate. Examples include variability in the essentiality of divJ and divK spatial cell cycle regulators, and non-essentiality of the highly conserved and usually essential DNA methyltransferase CcrM. These results show that while essential cell functions are conserved across varying genetic distance, much of a given organism's essential gene pool is specific to that organism.

Figure 1

Phylogenetic tree of select Alphaproteobacteria. Brackets indicate orders within the Alphaproteobacteria. Organisms of interest to this study are indicated with arrows. To generate the phylogenetic tree, gyrA sequences from representative species were aligned using MUSCLE (Edgar, 2004). RAxML reconstructed the maximum likelihood phylogeny using the JTT amino acid substitution matrix and a four-category discrete gamma distribution of sequence rate variation among sites (Stamatakis, 2006).

Figure 2

Distribution of internal transposon densities. The transposon density of the internal 60% of each ORF was plotted on a histogram. The distribution shows that each organism has a population of densities close to the Y-axis (essential genes) and a larger population with higher density (non-essential genes). Genes are present in the trough area between the peaks that likely contains essential genes as well as genes that have a significant growth defect when disrupted. Numbers in parenthesis represent the height of bars exceeding the Y-axis (97 for above 0.4 of A. tumefaciens and 295 for the 0-0.0025 range of B. subvibrioides). To the right of each distribution is a portion of the Ward’s cluster analysis performed using the internal transposon densities. The portions displayed cover the essential gene peak, trough, and early non-essential gene peak. Dashed lines indicate where major cluster divisions correspond to portions of the distribution. Genes from the Y-axis to line 1 were classified as essential, from line 1 to line 2 as unresolved, and from line 2 onward as non-essential.

Figure 3

Non-essentiality of septum site selection mechanisms and characterization of B. subvibrioides ΔmipZ. A) Transposon map of the min locus in A. tumefaciens, demonstrating it is likely non-essential. B) Transposon map of the parA, parB and mipZ genes in B. subvibrioides, demonstrating it is likely non-essential. C) Phase contrast image of wild-type and ΔmipZ. White arrows indicate incorrectly placed septum sites in ΔmipZ, either placed asymmetrically or multiple sites in the same cell. D) Histogram of wild-type and ΔmipZ cell sizes demonstrating deletion of mipZ causes increased variability of cell size (both smaller and larger than wild-type). E) Lectin staining of holdfast material of wild-type and ΔmipZ. F) Short-term adhesion assay of ΔmipZ. Equal numbers of cells for wild-type and ΔmipZ were inoculated into 24-well petri dishes and incubated for 2 h (in triplicate). After incubation, each well was washed three times with PYE and stained with crystal violet. After washing away excess crystal violet, bound stain was eluted and measured spectrophotometrically. Values were normalized to wild-type. Each assay was performed three times and average and standard deviation were calculated.

Figure 4

Swim expansion of developmental and cell cycle B. subvibrioides mutants. Equal numbers of cells for each strain were inoculated into low concentration agar PYE medium and incubated in a humid chamber for 5 days. Swim expansion is demonstrated by movement of the colony outwards from the inoculation point.

Figure 5

Characterization of the B. subvibrioides ΔgcrA strain. A) Transposon map demonstrating that gcrA is non-essential in this organism. Each bar is a unique insertion site and height represents frequency of detection for that insertion site. B) Lectin staining of holdfast material of wild-type and ΔgcrA. C) Growth of ΔgcrA compared to wild-type in PYE medium. Each strain was analyzed in triplicate. D) Short-term adhesion assay of ΔgcrA. Equal numbers of cells for wild-type and ΔgcrA were inoculated into 24-well petri dishes and incubated for 2 h (in triplicate). After incubation, each well was washed three times with PYE and stained with crystal violet. After washing away excess crystal violet, bound stain was eluted and measured spectrophotometrically. Values were normalized to wild-type. Each assay was performed three times and average and standard deviation were calculated.

Figure 6

Characterization of the B. subvibrioides ΔccrM strain. A) Transposon map of ccrM demonstrating that it is likely essential in A. tumefaciens and non-essential in B. subvibrioides. Each bar is a unique insertion site and height represents frequency of detection for that insertion site. B) Growth of ΔccrM compared to wild-type in PYE medium. Each strain was analyzed in triplicate. C) Lectin staining of holdfast material of wild-type and ΔccrM. C) Growth of ΔgcrA compared to wild-type in PYE medium. Each strain was analyzed in triplicate. D) Comparison of ftsZ promoters from C. crescentus (C.c.) and B. subvibrioides (B.s.). Annotation of features was based up on Kelly et al. (Kelly et al., 1998). E) Short-term adhesion assay of ΔccrM. Equal numbers of cells for wild-type and ΔccrM were inoculated into 24-well petri dishes and incubated for 2 h (in triplicate). After incubation, each well was washed three times with PYE and stained with crystal violet. After washing away excess crystal violet, bound stain was eluted and measured spectrophotometrically. Values were normalized to wild-type. Each assay was performed three times and average and standard deviation were calculated.

Figure 7

Characterization of the B. subvibrioides ΔdivJ, ΔdivK and ΔdivJΔdivK strains. A) A histidine kinase, DivJ, is localized to the stalked pole in the growing predivisional cell (Wheeler & Shapiro, 1999). DivJ phosphorylates the single domain response regulator DivK (Wheeler & Shapiro, 1999, Matroule et al., 2004). At the pole opposite the stalk (flagellar pole), another histidine kinase, PleC, de-phosphorylates DivK (Deich et al., 2004, Wheeler & Shapiro, 1999). DivJ and PleC have antagonistic activities on DivK until the time of cell division, at which point the activities are physically separated, leading to DivK becoming highly phosphorylated in the stalked cell and de-phosphorylated in the swarmer cell. DivK~P prevents a functional interaction between a degenerate histidine kinase DivL and a hybrid histidine kinase CckA (Tsokos et al., 2011). CckA, along with the histidine phosphotransfer protein ChpT (not pictured), phosphorylates and activates CtrA when CckA functionally interacts DivL (Jacobs et al., 2003, Biondi et al., 2006b). CckA and ChpT also phosphorylate another response regulator CpdR (not pictured) which controls the stability of CtrA (Iniesta et al., 2006). In the nascent stalked cell, DivK becomes phosphorylated, disrupting functional DivL-CckA interaction, reversing phosphate flow, and deactivating CtrA, which permits chromosome replication initiation (Tsokos et al., 2011, Chen et al., 2009). In the nascent swarmer cell, DivK is de-phosphorylated, permitting functional DivL-CckA interaction, activating the phosphorelay that activates CtrA, resulting in repression of chromosome replication. The ability of CtrA to transcriptionally activate genes in the swarmer cell is inhibited by a small regulatory protein SciP (Tan et al., 2010, Gora et al., 2013, Gora et al., 2010). B) Transposon map demonstrating the essentiality or non-essentiality of divJ and divK in A. tumefaciens and B. subvibrioides. Each bar is a unique insertion site and height represents frequency of detection for that insertion site. C) Lectin staining of holdfast material of wild-type and ΔdivJ, ΔdivK and ΔdivJΔdivK strains. D) Short-term adhesion assay of ΔdivJ, ΔdivK and ΔdivJΔdivK strains. Equal numbers of cells were inoculated into 24-well petri dishes and incubated for 2 h (in triplicate). After incubation, each well was washed three times with PYE and stained with crystal violet. After washing away excess crystal violet, bound stain was eluted and measured spectrophotometrically. Values were normalized to wild-type. Each assay was performed three times and average and standard deviation were calculated. E) Transmission electron micrographs demonstrating that the ΔdivJ and ΔdivK strains still produce flagella. F) Histogram of wild-type and ΔdivK cell sizes demonstrating deletion of divK causes a slight decrease in cell length.