Glutamate 83 and arginine 85 of helix H3 bend are key residues for FtsZ polymerization, GTPase activity and cellular viability of Escherichia coli: lateral mutations affect FtsZ polymerization and E. coli viability

Abstract

Background: FtsZ is an essential cell division protein, which localizes at the middle of the bacterial cell to mediate cytokinesis. In vitro, FtsZ polymerizes and induces GTPase activity through longitudinal interactions to form the protofilaments, whilst lateral interactions result within formation of bundles. The interactions that participate in the protofilaments are similar to its eukaryotic homologue tubulin and are well characterized; however, lateral interactions between the inter protofilaments are less defined. FtsZ forms double protofilaments in vitro, though the key elements on the interface of the inter-protofilaments remain unclear as well as the structures involved in the lateral interactions in vivo and in vitro. In this study, we demonstrate that the highly conserved negative charge of glutamate 83 and the positive charge of arginine 85 located in the helix H3 bend of FtsZ are required for in vitro FtsZ lateral and longitudinal interactions, respectively and for in vivo cell division.

Results: The effect of mutation on the widely conserved glutamate-83 and arginine-85 residues located in the helix H3 (present in most of the tubulin family) was evaluated by in vitro and in situ experiments. The morphology of the cells expressing Escherichia coli FtsZ (E83Q) mutant at 42°C formed filamented cells while those expressing FtsZ(R85Q) formed shorter filamented cells. In situ immunofluorescence experiments showed that the FtsZ(E83Q) mutant formed rings within the filamented cells whereas those formed by the FtsZ(R85Q) mutant were less defined. The expression of the mutant proteins diminished cell viability as follows: wild type > E83Q > R85Q. In vitro, both, R85Q and E83Q reduced the rate of FtsZ polymerization (WT > E83Q >> R85Q) and GTPase activity (WT > E83Q >> R85Q). R85Q protein polymerized into shorter filaments compared to WT and E83Q, with a GTPase lag period that was inversely proportional to the protein concentration. In the presence of ZipA, R85Q GTPase activity increased two fold, but no bundles were formed suggesting that lateral interactions were affected.

Conclusions: We found that glutamate 83 and arginine 85 located in the bend of helix H3 at the lateral face are required for the protofilament lateral interaction and also affects the inter-protofilament lateral interactions that ultimately play a role in the functional localization of the FtsZ ring at the cell division site.

Figure 1

(A) Front view of the E. coli FtsZ structural model. N-and C-terminal domains are represented in brown and gray, respectively. Arrows indicate longitudinal and lateral polymerization axes and the face of FtsZ is described as top, bottom, right and left according to the orientation of GTP (yellow). Residues E83 and R85 locate in the bend within the H3 helix. (B) Primary sequence alignment of the H3 region of FtsZ. The most conserved positions are colored in dark blue. The secondary structure of FtsZ is shown in the last line: alpha helix, orange rectangle; beta strand, blue arrow; coil, grey. The arrows indicate residues E83 and R85 of the E. coli FtsZ sequence.

Figure 2

Viability of E. coli VIP2 cells expressing FtsZ mutants. Exponentially-grown cells of E. coli VIP5 without (−) or with plasmids that express wild-type FtsZ (WT), FtsZ(E83Q) or FtsZ(R85Q) were plated by duplicate on two LB agar plates and incubated at 42°C or 30°C. Cell viability was determined the following day. The average and the standard deviation (SD) of 3 independent experiments were obtained using the program EXCEL. The H3 bend mutations cause significant decrease in cell viability.

Figure 3

Cell morphology and localization of FtsZ wild type and mutants E83Q and R85Q in cells with (30°C) and without (42°C) co-expression of FtsZ(WT) from pLAR9. Immunofluorescence of FtsZ (green) overlaid with the bright field image (gray) shows the localization of FtsZ in VIP5 cells (for strain information see Methods). Scale Bar, 5μm. Immunofluorescence was performed as described in Methods. Briefly, cells were grown until OD600 reached 0.5, cultures were split in half and incubated at 30°C and 42°C respectively for 4 h. Cells were harvested, fixed and immunofluorescent labeled with anti-FtsZ as first antibody and the second antibody conjugated with Alexa 488.

Figure 4

GTPase activity of the FtsZ mutants in the presence and absence of ZipA. (A) Progress curves for the GTPase activity in a solution containing 50 mM Mes pH 6.5, 50 mM KCl, 10 mM MgCl₂ with 12.5 μM of wild type FtsZ (O), E83Q (∇ ) or R85Q (□) at 30°C. The reaction was started by the addition of GTP at 1 mM final concentration, and GDP was determined by HPLC, as described in Material and Methods. (B) Progress curves for the GTPase activity of R85Q. The hydrolysis of GTP was measured with the following R85Q concentrations: (▼ ) 6.2; (● ) 12.4; (∇ ) 18.6; (O) 24.8; (∎) 31.0 and 49.6 μM (). The lag time (τ) was determined as the abscissa value where the line of the steepest slope of the progress curves intersects the x-axis. The inset shows the dependence of τ on the inverse of protein concentration. The values of picomoles indicated in the ordinate axe correspond to the amount of GDP in 100 μL of the reaction mixture (as indicated in Methods). (C) GTPase activity of 12.5 μM wild type FtsZ, E83Q and R85Q in the presence of 12 μM ZipA; the control with ZipA alone showed no GTPase activity (not shown). Inorganic phosphate was quantified using malachite green as described in Methods and the values of the ordinate axe are multiplied by 20.

Figure 5

Polymerization of E83Q and R85Q in presence of ZipA. (A) The polymerization of the FtsZ mutants and wild type was followed by light scattering; the buffer and protein concentrations were the same as described in Figure  4. After a stable base line of the buffer containing 1mM GTP, with or without ZipA, FtsZ was added to initiate the polymerization reaction. The change in light scattering at 350 nm was recorded during the polymerization reaction and the difference between the base line and the maximum value reached after FtsZ addition is shown in the bar graph. (B) Electron microscopy of negatively stained polymers of wild type FtsZ, E83Q and R85Q polymerized as in (A). The samples for FtsZ WT, E83Q and R85Q were taken at 23 min, 53 min and 2h 50 min, respectively (see Additional file 1: Figure S4 for polymerization curves). Electron micrographs were taken at a magnification of 28,500 for wild type FtsZ and E83Q, and of 52,000 for R85Q. Scale bar, 50 nm. (C) Histogram of FtsZ polymers width distribution shows that the mutations E83Q and R85Q decrease the number of protofilaments per polymer. At least 31 polymers were measured for each case using ImageJ.