Spatiotemporal control of PopZ localization through cell cycle-coupled multimerization

Abstract

Bacterial cell poles constitute defined subcellular domains where numerous proteins localize, often at specific times, to affect various physiological processes. How pole recognition occurs and what governs the timing of protein localization are often unknown. In this paper, we investigate the mechanisms governing the localization of PopZ, a chromosome-anchoring protein whose unipolar to bipolar localization pattern is critical for cell cycle progression in Caulobacter crescentus. We provide evidence that polar localization of PopZ relied on its self-assembly into a higher-order structure (matrix) and that the unipolar to bipolar transition was coupled to the asymmetric distribution of ParA during the translocation of the origin-proximal ParB-parS partition complex. Collectively, our data suggest a model in which a local increase of ParA concentration promotes the assembly of a PopZ matrix precisely when and where this matrix is needed. Such coupling of protein assembly with a cell cycle-associated molecular asymmetry may represent a principle of cellular organization for controlling protein localization in both time and space.

Figure 1.

Schematics of PopZ localization pattern during C. crescentus cell cycle. See Introduction for details.

Figure 2.

The C-terminal H3H4 domain of PopZ is necessary and sufficient for polar localization in wild-type C. crescentus cells. (A) Schematic overview of a multiple sequence alignment of 100 PopZ orthologues highlighting regions of conservation, displayed with Jalview using the Clustal X color scheme (Waterhouse et al., 2009). Proline (Pro) residues are all depicted in light green. The three main domains and the approximate positions of the α helices (H1–4), predicted with Jpred 3 (Cole et al., 2008), are indicated. term, terminal. (B) Schematic representation of the PopZ truncation variants. Domains, predicted α helices, and their position are indicated as well as the size of each variant (in amino acids). All regions are drawn to scale. (C) Localization of PopZ-YFP variants in C. crescentus cells. Synthesis of the PopZ-YFP variants was induced for 5 h in wild-type cells before imaging (phase contrast and YFP fluorescence). The YFP signal has been scaled for display. Bar, 2 µm. Strains are as follows: CJW3693 for full-length; CJW3695 for ΔH3H4; CJW3696 for ΔC; CJW3801 for N; CJW3694 for ΔN; CJW3697 for C; CJW3802 for H3H4; and CJW3816 for ΔH2. (D) YFP spots were detected from images in C. The percentage of cells with at least one focus of PopZ-YFP variant is shown. All foci were polarly localized. Number of cells counted (n): n = 765 for full-length; n = 701 for ΔH3H4; n = 803 for ΔC; n = 656 for N; n = 568 for ΔN; n = 726 for C; n = 721 for H3H4; and n = 941 for ΔH2. The data shown are from a representative experiment out of three repeats.

Figure 3.

The H3H4 domain mediates PopZ oligomerization. (A) Synthesis of the PopZ-YFP variants was induced for 5 h before imaging. The YFP signal has been scaled for display. Strains are as follows: CJW3707 for full-length; CJW3709 for ΔH3H4; CJW3710 for ΔC; CJW3804 for N; CJW3708 for ΔN; CJW3711 for C; CJW3805 for H3H4; and CJW3818 for ΔH2. Arrows point at stalks. (B) YFP spots were detected from images in A. The percentage of cells with at least one focus of PopZ-YFP variant is shown. All foci were localized at the pole. Number of cells counted: 1,067 for full-length; 699 for ΔH3H4; 592 for ΔC; 448 for N; 786 for ΔN; 576 for C; 755 for H3H4; and 790 for ΔH2. The data shown are from a representative experiment out of three repeats. (C) ΔN-TC synthesis was induced in CJW4432 cells for 4.5 h before FlAsH staining and imaging. An overlay of phase-contrast and FlAsH signal (red) is shown. (D) Distribution of cell lengths for strains shown in A. (E) Western blot detection of PopZ-YFP variants after migration of whole-cell extracts in native gels. Expression of the variants was induced in ΔpopZ (wild-type [wt] PopZ⁻) or wild-type (wild-type PopZ⁺) cells for 3 h. Samples were kept in a native state (SDS⁻) or denatured (SDS⁺) before native PAGE. Reference ticks from the NativeMark Unstained Protein Standard ladder (used in all gels) are shown. Black lines indicate that intervening lanes have been spliced out. Strains (wild-type/ΔpopZ background) are as follows: CJW3693/CJW3707 for full-length; CJW3695/CJW3709 for ΔH3H4; CJW3696/CJW3710 for ΔC; CJW3694/CJW3708 for ΔN; CJW3697/CJW3711 for C; and CJW3802/CJW3805 for H3H4. Bars, 2 µm.

Figure 4.

PopZ assembles into a selective matrix at the pole in E. coli. (A) Synthesis of YFP-tagged PopZ variants was induced for 2 h in E. coli cells grown in LB medium. Overlays of phase-contrast (green background) and YFP fluorescence signal (red) are shown. Strains are as follows: CJW3997 for full-length; CJW4684 for ΔC; CJW4685 for N; CJW4659 for ΔN; CJW4002 for H3H4; and CJW4001 for ΔH2. Note that all YFP fusions were at the C terminus except for ΔN, which was tagged at the N terminus because the C-terminal fusion was unstable. (B) Synthesis of PopZ-TC was induced for 3 h in CJW3991 cells. Arrows point at poles with a FlAsH-stained PopZ-TC focus. (C) Cells and conditions are the same as in B. Brackets delimit the polar LRI area visualized by DIC microscopy and labeled with FlAsH-stained PopZ-TC. The inset shows a zoomed example of a polar LRI region. (D) PopZ-TC synthesis was induced for 2 h in L1-GFP–producing CJW4673 cells before DAPI staining and imaging. Brackets delimit polar LRI areas visualized by DIC microscopy. (E) PopZ-TC production was induced for 2 h in strain CJW4744. PopZ-TC was labeled with FlAsH. Brackets delimit a polar LRI area visualized by DIC microscopy. (F) CJW3997 cells were treated with cephalexin for 2 h before induction of PopZ-YFP and CFP-ParB synthesis for 1 h. Cells were stained with DAPI before imaging. Overlays of DAPI and PopZ-YFP (top) or CFP-ParB (bottom) with the MicrobeTracker cell outline are shown. Arrowheads indicate polar and nonpolar foci. (G) PopZ-YFP synthesis was induced in CJW3997 cells for 1 h before addition of cephalexin for an additional 3 h. CFP-ParB synthesis was induced during the last hour of treatment before DAPI staining and imaging. Overlays are displayed as in F. Arrows delimit the accumulation of PopZ-YFP and CFP-ParB at one pole. Bars: (A–C [main image], D, and E) 2 µm; (C, insets) 1 µm; (F and G) 5 µm.

Figure 5.

Bipolar localization of PopZ can happen in the absence of DNA replication. (A) CJW2214 cells were grown in liquid medium with xylose to induce the expression of YFP-PopZ. Swarmer cells were treated or not treated with 25 µg/ml novobiocin and imaged every 30 min. The mean fraction of cells showing bipolar YFP-PopZ localization (from three independent experiments per condition) is shown for each time point. Error bars show SEM. (B) Representative cells from one experiment described in A are shown for each time point, under untreated and novobiocin-treated conditions. Arrows point at bipolar YFP-PopZ foci. (C) CFP-ParB was imaged at each time point in the same cells as in A. The mean fraction of cells having two CFP-ParB foci (from three independent experiments per condition) is shown for each time point. Error bars show SEM. (D) Swarmer CJW4721 cells were grown on an M2G agarose pad containing 5 µg/ml novobiocin. The fraction of cells with two PopZ-YFP foci or with at least one DnaN-mCherry focus (indicative of DNA replication) is shown for each time point. Outlines and fluorescent signals of representative cells at 150 min after synchrony are shown. Arrowheads point at bipolar PopZ-YFP. (E) Swarmer CJW2237 cells were grown on an M2G agarose pad containing 5 µg/ml novobiocin. Shown are kymographs of the PopZ-YFP and CFP-ParB signals along the cell length over time from two representative cells. Relative positions of 0 and 1 represent the old pole and the new pole, respectively. Note that C. crescentus grows slower on agarose pads (E) than in liquid cultures (as in A). Bars, 2 µm.

Figure 6.

Bipolar localization of PopZ correlates with the completion of parS segregation. (A) CJW3544 cells were grown with (TipN⁺) or without (TipN⁻) xylose before synchrony. Swarmer cells were grown on M2G agarose pads with (TipN⁺) or without xylose (TipN⁻), and PopZ-YFP and MipZ-CFP were imaged every 5 min. (left) The fraction of cells with two PopZ-YFP foci for each time point. (right) All cells from all time points were sorted by cell length, and the fraction of cells with two PopZ-YFP foci is shown for each bin of 0.05 µm. (B) Kymographs of the MipZ-CFP (proxy for parS) and PopZ-YFP signal intensity profiles along the cell length over time, averaged for the population of TipN⁺ cells and TipN⁻ cells imaged in A. Total cell numbers are indicated (n). Relative position of 0 and 1 represent the old pole and the new pole, respectively. Arrowheads indicate the appearance of PopZ-YFP at the new pole. (C) Distribution of the time after synchrony (left) or the cell length (right) at which MipZ-CFP (parS) reaches the new pole in each cell population imaged in A. (D) Distribution of the time after synchrony (left) or of cell length (right) when a second PopZ-YFP focus is first detected. (E) For each TipN⁺ and TipN⁻ cell imaged in A, the cell length at which PopZ-YFP (x axis) or MipZ-CFP (y axis) reached the new pole was recorded. Each dot represents one cell. Correlation coefficient (r), p-value, and cell counts (n) are indicated. (F) PopZ-CFP and ParA-YFP were imaged in synchronized cells (CJW4626; n: cell counts) grown on an M2G agarose pad. The mean fluorescence intensity profile along the cell length is shown for each fusion protein during cell cycle progression. PopZ and ParA as seen by epifluorescence microscopy are represented on schematics for three representative steps during the cell cycle.

Figure 7.

ParA mediates spatiotemporal control of PopZ localization. (A) The synthesis of PopZ-YFP variants and ParA_R195E-CFP was induced for 5 and 2 h, respectively. The bar graph shows the fraction of cells with ParA_R195E-CFP spots in C. crescentus cells producing PopZ-YFP (CJW4769) or ΔN-YFP (CJW4770). Cell counts (n) are as follows: 456 out of 478 CJW4769 cells and 137 out of 817 CJW4770 cells had at least one PopZ-YFP focus. The data shown are from a representative experiment out of two repeats. (B) Overproduction of PopZ-TC or ΔN-TC was induced for 5 h in CJW4845 or CJW4846 C. crescentus cells, respectively, and ParA_R195E-CFP synthesis was induced for 2 h, before FlAsH staining. Arrows indicate the location of the PopZ-TC or ΔN-TC–rich area. (C) PopZ-TC (CJW4746) or ΔN-TC (CJW4745) was imaged with FlAsH after overproduction for 5 h in C. crescentus cells. Arrows are displayed as in B. The double arrow points at one CFP-ParB focus. (D) PopZ-YFP variants and ParA_R195E-CFP were imaged in E. coli after induction, for 2 and 1 h, respectively. Strains are as follows: CJW4835 for full-length and CJW4836 for ΔN. Insets indicate that the cell was imaged on a different field of view. (E) PopZ-YFP variants and CFP-ParB were imaged after 2 h of induction in E. coli cells grown in LB. Strains are as follows: CJW3997 for full-length and CJW4659 for ΔN. (F) Kymograph of CFP-ParB and PopZ-YFP in a representative ParA_K20R-producing C. crescentus cell. Swarmer cells (CJW4441) were imaged every 2 min. (G) Swarmer C. crescentus cells (CJW4613) producing ParA_K20R were imaged every 7 min. The kymograph of the ParA-YFP signal is shown for a representative cell, along with the relative position of the PopZ-CFP foci (white circles). (H) Mean fraction of the ParA-YFP signal located in the vicinity of the new pole before and after the formation of a second PopZ-CFP focus. The time when two PopZ-CFP spots were first detected was defined as time 0 for each cell from the experiment described in G. The fraction of ParA-YFP signal in the new pole-proximal quarter of the cell was averaged for all cells at each time point using time 0 as a reference (red line). (I) E. coli cells (CJW4917) producing DivIVA-ParA_R195E-CFP were grown in M9 + glucose at 30°C and treated with cephalexin for 1 h before induction of PopZ-TC synthesis by washing the cells for 30 min in M9 + glycerol containing arabinose. Cells were first stained with FlAsH for 30 min and then with DAPI before imaging. Arrows point to LRI regions. (J) E. coli cells (CJW4918) producing DivIVA-ParAR195E-CFP were grown in M9 + glucose, washed for 15 min in M9 + glycerol containing arabinose to briefly induce the synthesis of PopZ-YFP, and washed in M9 + glucose for 15 min before growth and time-lapse imaging on an M9 + glucose pad (repression of popZ-yfp expression) containing cephalexin. Representative cells are shown for selected time points. Arrows point to new foci. Bars, 2 µm.

Figure 8.

Model for uncontrolled and cell cycle–controlled pole localization. (A) Accumulation of a self-assembling protein (here PopZ) through multimerization in favorable regions, such as poles as a result of their low DNA content, is a stochastic process that leads to the expansion of the protein structure (PopZ matrix) at one pole. (B) Spatial and temporal regulation of the spontaneous multimerization process in A can be achieved through coupling with a cell cycle event (ParA-dependent ParB–parS segregation) that involves the local concentration of an interacting protein (ParA). Accumulation of the protein partner (ParA) results in a local increase in the concentration of diffusing self-assembling proteins (PopZ oligomers) to a level that promotes and sustains assembly into a higher-order structure (PopZ matrix) where and when the cell cycle event takes place. In the case of PopZ, a coupling with the ParA-dependent segregation of ParB–parS allows for the controlled assembly of a PopZ matrix at the new pole in time to capture the partitioning ParB–parS complex. N-term, N-terminal.