General protein diffusion barriers create compartments within bacterial cells

Abstract

In eukaryotes, the differentiation of cellular extensions such as cilia or neuronal axons depends on the partitioning of proteins to distinct plasma membrane domains by specialized diffusion barriers. However, examples of this compartmentalization strategy are still missing for prokaryotes, although complex cellular architectures are also widespread among this group of organisms. This study reveals the existence of a protein-mediated membrane diffusion barrier in the stalked bacterium Caulobacter crescentus. We show that the Caulobacter cell envelope is compartmentalized by macromolecular complexes that prevent the exchange of both membrane and soluble proteins between the polar stalk extension and the cell body. The barrier structures span the cross-sectional area of the stalk and comprise at least four proteins that assemble in a cell-cycle-dependent manner. Their presence is critical for cellular fitness because they minimize the effective cell volume, allowing faster adaptation to environmental changes that require de novo synthesis of envelope proteins.

Figure 1. A diffusion barrier compartmentalizes the Caulobacter periplasm

(A–C) Diffusion of a xylose-inducible PstS-mCherry fusion assayed using FLIP (A and C) and FRAP (B). Cells (EK363) were bleached with seven ~4 nsec pulses in the region indicated by yellow circles. Bleaching was performed either multiple times in succession (A) or once followed by a 105 sec recovery (B–C). Insets show schematic representations of the results, and graphs show the quantification of fluorescence in multiple cells (panel A: n = 6, p < 0.0002; panel B: n = 7, p < 2×10⁻⁷; panel C: n = 3, p < 0.002; error bars = SD). Fluorescence intensities were measured in the stalk (blue) and cell body (black) of the bleached cell or the stalk (red) and cell body (green) of a nearby control cell. For intra-stalk bleaching, the bleached (blue) and unbleached (red) portions of the stalk as well as the cell body (black) fluorescence were quantified. The color-maps of the fluorescent images were scaled for easier visualization. However, all quantifications were performed using raw image data. The fluorescence intensity of each region of interest was normalized to its pre-bleach intensity. Abbreviations: PB, pre-bleach; B, bleach. Scale bars: 2 μm. See also Figure S1.

Figure 2. Identification and subcellular localization of novel stalk proteins

(A) Stalk localization of StpA-mCherry (SW33) and StpB-mCherry (SW30) produced from the xylose-inducible P_xyl promoter after 24 h of growth in phosphate-rich (M2G, high PO₄³⁻) or phosphate-poor medium (M2G^−P, low PO₄³⁻) containing 0.3% xylose. (B) Co-immunoprecipitation analysis of stpB-His (SS233) and wild-type cells reveals an interaction between StpA and StpB. Whole-cell lysates (L) and eluates from co-immunoprecipitation experiments (Co-IP) were subjected to immunoblot analysis using anti-His and anti-StpA antibodies. (C) StpA,B,C,D localization in stalks is reminiscent of the distribution of crossbands. Cells of strain SS243 (stpD::stpD-gfp P_xyl::P_xyl-stpA-mcherry), SS88 (stpB::stpB-mcherry stpD::stpD-gfp) and SS389 (stpC::stpC-mcherry stpD::stpD-gfp) were grown in M2G^−P for 24 h. Synthesis of StpA-mCherrry was induced with 0.3% xylose for 24 h. (D) Schematic depicting the domain organization of the stalk proteins with the predicted transmembrane domains (orange), the signal peptide (purple) and the Sel1 motifs (green). (E) Cell fractionation analysis reveals that StpA, StpC, and StpD are membrane-bound proteins, whereas StpB is soluble. Whole-cell lysates (L) and the corresponding membrane (M) and soluble (S) fractions of cells producing His-tagged Stp proteins (SS233, SS220, SS244 and SS247) were subjected to Western blot analysis using an anti-His antibody. Fractionation efficiency was verified by probing the same fractions with anti-CtrA and anti-SpmX antibodies. Note, the absence of stpA and stpAB does not reduce the levels of StpB and StpC, respectively (Figure S2C). (F) The Stp proteins are targeted to the periplasm. The TEM-1 β-lactamase gene (bla) was fused to the 3’ end of stpA, stpB, stpC and stpD, respectively. The gene fusions were placed under the P_xyl promoter in a β-lactam-sensitive reporter strain. Cells (SS165, SS172, SS273, SS274) were patched on PYE agar containing ampicillin and either 0.2% glucose or 0.3% xylose. Scale bars: 3 μm. See also Figure S2.

Figure 3. Crossbands are static multi-protein complexes

(A) StpAB-deficient cells consistently lack crossbands. Cells with and without stpAB (SW51, n = 8) were grown in PYE and imaged by ECT. The images show a longitudinal section of the stalk. Asterisks denote crossbands. Arrowheads point at unidentified structures spanning the stalk core. Scale bars: 100 nm. (B and C) The distribution of StpB-mCherry foci reflects the distribution of crossbands in stalks. Cells of strains CB15N (WT) and SS160 (stpB-mcherry) were grown in M2G^−P and imaged either by electron (EM) or fluorescence (FM) microscopy, respectively. Electron micrographs were acquired of negatively stained wild-type cells. From the respective images, the number of crossbands (n = 68 cells) and StpB-mCherry foci per μm stalk (n = 316 cells) was quantified (*p > 0.2, t-test; error bars = SEM). Asterisks denote crossbands. Scale bars: 500 nm (EM) and 3 μm (FM). (D) StpB spatially overlaps with crossbands. Strain SW30 (P_xyl::P_xyl-stpB-mcherry) was grown in M2G^−P with 0.3% xylose. Cells were fixed on EM grids and imaged first by low-magnification phase contrast/fluorescence microscopy (inset; arrowheads indicate StpB-mCherry foci) and then by ECT. Shown is an ECT slice of a stalk with arrows pointing to crossband structures (left panel) and the respective correlated image showing the ECT slice overlayed with a fluorescence micrograph of the same region (right panel). Scale bar: 100 nm. (E) FRAP analysis reveals that crossbands are static protein complexes. Cells of strain SS160 (stpB-mcherry) were cultured in M2G^−P and imaged by fluorescence microscopy to identify StpB-mCherry localization. A laser pulse was applied to selected regions (yellow circles), and StpB-mCherry signals were bleached. Cells were imaged immediately and 10 min after the laser pulse. Scale bar: 3 μm. See also Figure S3.

Figure 4. Crossband synthesis is cell cycle-dependent and relies on hierarchal self-assembly of the Stp proteins

(A) Western blot analysis of Stp protein levels during the cell cycle. Swarmer cells of SS233 (stpB::stpB-His), SS247 (stpC::stpC-His) and SS244 (stpD::stpD-His) were grown in M2G. Samples were taken from the culture in 20 min intervals and probed with anti-CtrA, anti-StpA and anti-His antiserum. The schematic illustrates the different morphological stages of the cell cycle. (B) Timecourse microscopy of StpB-mCherry localization, starting with isolated swarmer cells of SS160 (stpB-mcherry). Cells were grown in M2G. Scale bar: 3 μm. (C) Localization hierarchy of the Stp proteins. Xylose-inducible fluorescent protein fusions to StpA, StpB, StpC, and StpD were analyzed in the indicated strain backgrounds (SS141, SS142, SS234, SS236, SS240, SS263, SS264, SS265). Cells were grown in M2G^−P and induced for 24 h with 0.3% xylose. Scale bar: 3 μm. (D) Schematic illustrating the order of StpABCD complex assembly. (E) Stalk ultrastructure of cells carrying an inducible copy of stpAB on a multi-copy plasmid (SS214). Cells were cultivated in M2G^−P in the absence inducer, negatively contrasted with uranyl acetate and imaged by transmission electron microscopy. The dashed rectangle in (i) indicates the region magnified in (ii). Asterisks indicate crossbands. Scale bars: 500 nm. A strain carrying the empty plasmid (SS258) showed the wild-type frequency of crossbands (data not shown). (F) Visualization and 3D-reconstruction of helical StpAB assemblies. Cells carrying a plasmid-encoded copy of stpAB under the control of P_xyl (SS214) were pre-cultured in PYE, grown in M2G^−P containing 0.3% xylose, and analyzed by ECT. Shown is a section through a representative tomogram of a stalk (left panel) and a 3D reconstruction of the helical assemblies induced by StpAB overproduction (right panel). Scale bars: 50 nm. (G) Constitutive production of StpA increases the frequency of crossbands. Cells of strain SW33 (P_xyl ::P_xyl-stpA-mcherry) and SW30 (P_xyl::P_xyl-stpB-mcherry) were grown in M2G^−P with 0.3% xylose for 24 h and imaged by fluorescence microscopy. The number of fluorescent foci per μm stalk was determined for cells of SW30 (n = 120) and SW33 (n = 194) (*p < 0.05, t-test; error bars = SEM). See also Figure S4.

Figure 5. Crossbands serve as protein diffusion barriers

(A) Analysis of the compartmentalization of soluble and periplasmic red fluorescent protein (tdimer2) using FLIP. Cells of strain SS269 (stpD::stpD-gfp pP_xyl-TAT-tdimer2) and SS216 (ΔstpAB pP_xyl-TAT-tdimer2) were cultured in M2G^−P containing 0.3% xylose for 24 h. Cells were mounted on an agarose pad and exposed to a laser pulse in the regions indicated by a yellow circle. Scale bar: 3 μm. (B) Crossbands compartmentalize periplasmic, inner- and outer membrane proteins. Cells of strains SS277 (stpD::stpD-gfp P_xyl-gspG-mcherry), SS272 (ΔstpAB P_xyl::P_xyl-gspG-mcherry), SS299 (stpD::stpD-gfp P_xyl::P_xyl-pstS-mcherry), SS302 (ΔstpAB P_xyl::P_xyl-pstS-mcherry), SS283 (stpD::stpD-gfp P_xyl::P_xyl-elpS-mcherry), and SS284 (ΔstpAB P_xyl::P_xyl-elspS-mcherry) were first grown in M2G^−P for 36 h and then induced with 0.3% xylose for 11–13 h. Scale bar: 3 μm. (C) StpX-GFP mobility requires compartmentalization of the stalk from the cell body by the newest crossband. Cells of strain YB5058 (stpX::stpX-gfp P_xyl::P_xyl-stpB-mcherry) were grown in HIGG-30 μM phosphate containing 0.3% xylose and mounted on an agarose pad. First, StpB-mCherry fluorescence was imaged to identify regions of interest (yellow circles). Then, these regions were simultaneously bleached for 50 sec, followed by the acquisition of a post-bleach image. White lines outline the cell bodies. Scale bar: 3 μm. (D) Crossbands affect the processing of the stalk-specific protein StpX. Wild-type, ΔstpAB (SW51) and ΔstpX (YB5231) cells were grown to stationary phase in HIGG-30 μM phosphate and subjected to immunoblot analysis using an antibody raised against the N-terminal domain of StpX (anti-StpX-NTD). Arrowheads denote the full-length version of StpX, asterisks the dominant short fragment. See also Figure S5.

Figure 6. Diffusion barriers are crucial for bacterial fitness

(A) Rate of periplasmic accumulation of an inducible PstS-mCherry protein fusion in the wild-type (EK363) and ΔstpAB (EK389) background. Cells were grown in HIGG-30 μM phosphate with 0.3% glucose to induce long stalks while repressing the synthesis of PstS-mCherry. The cells were seeded on pads with 0.3% xylose, and PstS-mCherry accumulation was monitored by timelapse microscopy. Mean fluorescence/cell body area was measured for ~ 300 cells per strain. Error bars = SD. Fitting the data to an exponential function and solving the equations in the exponential region (t = 165 to 255 min) for equal fluorescence intensities yielded a time difference in the accumulation of PstS-mCherry of 22.4 (± 0.8) min. (B) Competitive growth of wild-type and diffusion barrier-deficient cells. To analyze the effect of diffusion barriers on the rate of recovery from nutrient starvation, wild-type, ΔstpAB and ΔstpCD cells were differentially labeled with the fluorescent proteins YFP (EK392, EK417, EK486) or mCherry (EK416, EK393, EK487) and starved for either phosphate or nitrogen. Mutant and wild-type cells were combined at equal ratios, transferred to nutrient-replete medium and grown to late-exponential phase. More than 1000 cells were analyzed by fluorescence microscopy before and after recovery to determine shifts in the relative composition of the cultures. The differentials were then used to calculate the lag in the onset of cell division (see Figure S6B). Values represent the average of four experiments, including two in which the fluorescent marker was switched (error bars = SD; * p < 0.02; ** p < 0.002). (C) Wild-type cells outcompete a diffusion barrier-deficient mutant. Wild-type and ΔstpAB cells (SW51) were grown in PYE and mixed at equal optical densities. Mixed cultures (n = 5) were repeatedly diluted into fresh PYE and cultured to stationary phase. At the indicated timepoints, cells were withdrawn and spread on PYE agar. The ratio of wild-type and barrier-deficient cells was determined by colony PCR. Error bars show SD (n ~ 450). See also Figure S6.

Figure 7. Model for diffusion barrier formation and function

(A) Diffusion barrier assembly can be envisioned as a nucleation-like process in which StpA (orange) and StpB (green) form the basic scaffold. StpC (blue) and StpD (purple) are accessory components that are required to seal the diffusion barrier. Potential unidentified components of the complex that may establish a connection to the outer membrane or assemble at the cytoplasmic face of the inner membrane are depicted in yellow. (B) The synthesis of diffusion barriers minimizes the effective volume of the periplasmic space and reduces the physiologically active membrane surface area. In the absence of diffusion barriers, newly synthesized proteins that are targeted to the cell envelope are constantly diluted due to diffusion into the stalk extension.