Pbp2x Localizes Separately from Pbp2b and Other Peptidoglycan Synthesis Proteins during Later Stages of Cell Division of Streptococcus pneumoniae D39

Abstract

The relative localization patterns of class B penicillin-binding proteins Pbp2x and Pbp2b were used as positional indicators of septal and peripheral (side-wall-like) peptidoglycan (PG) synthesis, respectively, in the midcell regions of Streptococcus pneumoniae cells at different stages of division. We confirm that Pbp2x and Pbp2b are essential in the strain D39 genetic background, which differs from that of laboratory strains. We show that Pbp2b, like Pbp2x and class A Pbp1a, follows a different localization pattern than FtsZ and remains at division septa after FtsZ reappears at the equators of daughter cells. Pulse-experiments with fluorescent D-amino acids (FDAAs) were performed in wild-type cells and in cells in which Pbp2x activity was preferentially inhibited by methicillin or Pbp2x amount was depleted. These experiments show that Pbp2x activity separates from that of other PBPs to the centers of constricting septa in mid-to-late divisional cells resolved by high-resolution 3D-SIM microscopy. Dual-protein and protein-fluorescent vancomycin 2D and 3D-SIM immunofluorescence microscopy (IFM) of cells at different division stages corroborate that Pbp2x separates to the centers of septa surrounded by an adjacent constricting ring containing Pbp2b, Pbp1a, and regulators, StkP and MreC. The separate localization of Pbp2x suggests distinctive roles in completing septal PG synthesis and remodeling.

INTRODUCTION

Streptococcus pneumoniae (pneumococcus) is a human commensal bacterium that colonizes the nasopharynx and causes a number of serious respiratory and invasive diseases (Donkor, 2013, Henriques-Normark & Tuomanen, 2013, Vernatter & Pirofski, 2013). Drug-resistant S. pneumoniae was recently listed by the CDC as a serious threat to public health in the U.S.A. (CDC, 2013). S. pneumoniae cells are shaped liked prolate-ellipsoids referred to as ovococci that divide perpendicularly to their long axis and often form chains of cells, especially when capsule is present (Fig. 1A) (Barendt et al., 2009, Massidda et al., 2013, Pinho et al., 2013, Zapun et al., 2008b). Pneumococcal cell shape and chaining play important roles during colonization and infection (Dalia & Weiser, 2011, Rodriguez et al., 2012). The ellipsoid shape of S. pneumoniae and other ovococcus bacteria is largely determined by the peptidoglycan (PG) cell wall that surrounds these Gram (+) bacteria (Massidda et al., 2013, Pinho et al., 2013, Sham et al., 2012). PG is composed of glycan chains that are cross-linked by PG peptides (Egan & Vollmer, 2013, Lovering et al., 2012, Typas et al., 2012). PG provides the major resilient scaffolding for the covalent attachment of surface structures, including wall-teichoic acid (WTA), capsule, and surface proteins linked by sortases, many of which are virulence factors (Denapaite et al., 2012, Eberhardt et al., 2012, Schneewind & Missiakas, 2013, Tomasz & Fischer, 2006). Because of its surface exposure and prokaryotic-specific mechanism of biosynthesis, PG is the target for many clinically important antibiotics used to treat S. pneumoniae and other bacterial infections, including β-lactam (e.g., penicillins and cephalosporins) and glycopeptide (e.g., vancomycin) antibiotics (den Blaauwen et al., 2014, Hakenbeck et al., 2012, Nikolaidis et al., 2014, Sham et al., 2012, Zapun et al., 2008a).

Fig. 1.

Introductory model of PG biosynthesis in ovococcus bacteria, such as S. pneumoniae, and summary of phenotypes caused by disruption of septal and peripheral PG synthesis. A. Top. Ovococci divide perpendicularly to their long axis (Zapun et al., 2008b). Unencapsulated derivatives of wild-type S. pneumoniae form mostly diplococci and chains of two cells, whereas capsulated strains form short chains of 8-10 cells (Barendt et al., 2009). Bottom. Formation of prolate-ellipsoid-shaped bacteria requires two modes of PG synthesis, peripheral (side-wall-like) and septal PG synthesis, that occur in the midcell regions of dividing S. pneumoniae cells (Massidda et al., 2013, Pinho et al., 2013, Sham et al., 2012, Zapun et al., 2008b). At the start of a division cycle, components of both peripheral PG synthesis complexes (orange ovals) and septal synthesis complexes (green rectangles) locate to the equators of cells (bottom). Peripheral PG synthesis (light blue; top) occurs between the future equator and septum of dividing cells and may commence before septal synthesis (Massidda et al., 2013, Wheeler et al., 2011, Zapun et al., 2008b). At some point, septal PG synthesis (medium blue) commences to divide the cell in two. Data presented in this paper show that the complexes that carry out peripheral and septal PG synthesis locate to a large constricting ring throughout the division cycle, with the exception of Pbp2x, which moves to the centers of septa in mid-to-late divisional cells. The grey Pac-Man corresponds to PG hydrolases that remodel the PG and allow septal separation. B. Comparison of morphological changes of cells of an unencapsulated derivative of S. pneumoniae D39 when septal (top) or peripheral PG synthesis (bottom) is blocked. Septal PG synthesis and ring closure is blocked by selective inhibition of Pbp2x by methicillin (Land et al., 2013), depletion of Pbp2x (Results) (Berg et al., 2013, Peters et al., 2014), or GpsB depletion (Land et al., 2013), causing cell elongation. Peripheral PG synthesis is blocked by depletion of Pbp2b (Results) (Berg et al., 2013), or MreC (Land & Winkler, 2011), causing formation of similar-looking chains of nearly spherical cells. See text for additional details. All micrographs are at the same magnification, with scale bar = 1 μm.

Formation of prolate-ellipsoid-shaped cells requires two modes of PG synthesis by penicillin-binding proteins (PBPs) accompanied by PG remodeling by PG hydrolases (Fig. 1A) (Higgins & Shockman, 1970, Massidda et al., 2013, Pinho et al., 2013, Sham et al., 2012, Zapun et al., 2008b). In ovococcus cells, equators of daughter cells become septa during division (Fig. 1A). As in other eubacteria, FtsZ-mediated septal PG synthesis occurs at the midcell and leads to cell division (Fig. 1A) (Egan & Vollmer, 2013, Massidda et al., 2013, Pinho et al., 2013, Typas et al., 2012). Unlike rod-shaped cells, ovococcus cells carry out a form of peripheral PG elongation at midcells instead of at side-walls (light blue, Fig. 1A). Peripheral PG synthesis accounts for the elongation of ovococci that distinguishes them from spherical-shaped cocci (Pinho et al., 2013, Zapun et al., 2008b). S. pneumoniae and other ovococci lack MreB homologues that move in circumferential patterns around the bodies of rod-shaped bacteria and mediate lateral side-wall PG synthesis (Garner et al., 2011, Lee et al., 2014). Therefore, pneumococcal peripheral PG synthesis resembles preseptal PG side-wall synthesis that occurs at midcells early in the division cycle of E. coli and other rod-shaped bacteria [see (Potluri et al., 2012, van der Ploeg et al., 2013)].

Septal and peripheral PG synthesis can separately be inhibited in S. pneumoniae cells (Fig. 1B). Pbp2x and Pbp2b are class B (transpeptidase only) PBPs that mediate septal and peripheral PG synthesis in S. pneumoniae (Berg et al., 2013, Fleurie et al., 2014, Massidda et al., 2013, Peters et al., 2014) and correspond to Pbp3 and Pbp2 in E. coli, respectively [see (Egan & Vollmer, 2013, van der Ploeg et al., 2013)]. Specific inhibition of Pbp2x with antibiotics results in cell elongation, consistent with inhibition of septal PG synthesis (Fig. 1B) (Land et al., 2013, Perez-Nunez et al., 2011), whereas depletion of Pbp2x results in a mixture of round cells and elongated cells, often with pointed ends (Fig. 1B; Results) (Berg et al., 2013, Peters et al., 2014). By contrast, depletion of Pbp2b leads to the formation of rounded cells in chains, consistent with inhibition of peripheral PG synthesis (Fig. 1B; Results) (Berg et al., 2013). These depletion experiments suggest that Pbp2x and Pbp2b are essential in S. pneumoniae (Results) (Berg et al., 2013, Berg et al., 2014, Fleurie et al., 2014, Peters et al., 2014).

Regulatory proteins can also be assigned to septal or peripheral PG synthesis based on cell-shape defects caused by depletion and on the functions of homologues in rod-shaped bacteria (Massidda et al., 2013, Zapun et al., 2008b). Depletion of the essential MreC or MreD protein in the D39 genetic background of S. pneumoniae results in formation of chains of spherical cells remarkably similar to those formed by blocking peripheral PG synthesis by Pbp2b depletion (Fig. 1B) (Berg et al., 2013, Land & Winkler, 2011). In the D39 genetic background, GpsB is essential, and depletion of GpsB causes formation of enlarged, elongated cells unable to close division septa (Fig. 1B) (Land et al., 2013). However, bypass suppressor mutations have become a major complication in determining primary phenotypes of regulatory proteins in pneumococcal cell division. Strain D39 is the encapsulated virulent serotype 2 progenitor strain of most unencapsulated laboratory strains, which have accumulated over 80 mutations that change metabolism and PG composition (Carvalho et al., 2013, Lanie et al., 2007). Neither MreCD nor GpsB is essential in laboratory strains (Fleurie et al., 2014, Land & Winkler, 2011), which contain suppressors that may alter pathways of cell division and PG biosynthesis. In other instances, bypass suppressors likely have accumulated during mutant construction. For example, the StkP serine/threonine kinase certainly regulates some stage of cell division of S. pneumoniae (Beilharz et al., 2012, Fleurie et al., 2012), but ΔstkP mutants exhibit a variety of morphology defects, due in part to suppressor accumulation [see (Massidda et al., 2013)].

Despite complications from suppressor accumulation, Pbp2x, Pbp2b, and some regulatory proteins can be assigned separable functions in septal or peripheral PG synthesis. Therefore, it is likely that separate complexes of some components are present at the midcells of dividing pneumococcal cells (Fig. 1A). However, it is unclear whether some components, such as class A (dual transglycosylase and transpeptidase) Pbp1a, which has been implicated in peripheral PG synthesis (Land & Winkler, 2011), also participate in septal PG synthesis (Massidda et al., 2013, Peters et al., 2014) and whether septal and peripheral PG synthesis complexes are intermixed, adjacent, or physically separate during different stages of division (Fleurie et al., 2014, Land et al., 2013, Massidda et al., 2013, Zapun et al., 2008b). Moreover, measurements of dimensions of dividing pneumococcal cells indicated that peripheral and septal PG synthesis occur simultaneously over much of the pneumococcal cell cycle, although there may be a short period of preseptal PG peripheral synthesis before the start of septal PG synthesis (Wheeler et al., 2011). However, the extent of switching and coordination between peripheral and septal PG synthesis remain largely unknown.

Most localization studies of Pbp2x, Pbp2b, and Pbp1a have been performed using 2D immunofluorescence microscopy (IFM) or epifluorescence microscopy of live cells containing fluorescent-protein fusions (resolution ≈250 nm) (Fleurie et al., 2014, Morlot et al., 2003, Peters et al., 2014, Zapun et al., 2008b). In these studies, numbers of cells showing localization patterns were counted. By these criteria, Pbp2x, Pbp2b, Pbp1a, and other PG synthesis proteins appear to colocalize at all stages of cell division. In a previous study, we used an image analysis graphical user interface (IMA-GUI) to quantitate average fluorescence intensities of immunolabeled proteins along the long axis of images of cells at different division stages obtained by dual-protein 2D IFM (Land et al., 2013). This analysis revealed that FtsZ and GpsB have overlapping, but not identical, patterns of localization. Notably, FtsZ and GpsB migrated from midcell septa to equators of daughter cells at an earlier stage of division than Pbp2x and Pbp1a, consistent with other reports (Morlot et al., 2003, Zapun et al., 2008b). IMA-GUI analysis also revealed that the average diameter of Pbp2x rings appeared to be smaller than that of Pbp1a rings in mid-divisional cells (Land et al., 2013). Results from 2D analyses of FtsZ, GpsB, Pbp2x, and Pbp1a localization were confirmed and extended by high-resolution, dual-protein 3D-SIM (structured illumination microscopy) IFM (resolution XY≈100 nm; Z≈250 nm) (Land et al., 2013). 3D-SIM images showed that Pbp2x and Pbp1a colocalize to an equatorial ring in early-to-mid divisional cells, but Pbp2x separates from the ring containing Pbp1a and locates as a disk at the center of the septum in mid-to-late divisional cells (Land et al., 2013).

In this paper, we report the use of fluorescent D-amino acid (FDAA) probes (Kuru et al., 2012) to independently demonstrate that Pbp2x locates separately from other PBPs in mid-to-late stages of pneumococcal cell division. An additional tool was added to quantitate dual-protein 2D IFM images to show that Pbp2x localizes separately from Pbp1a and other PG synthesis proteins, including Pbp2b, MreC, and StkP, at a later stage of division. 3D-SIM IFM corroborated that PG synthesis proteins remain together in a ring structure until a later stage of cell division, after which Pbp2x locates to the center of division septa, where PG-pentapeptides labeled by fluorescent-vancomycin (FL-V) concentrate. These results show that pneumococcal PG synthesis proteins colocalize contiguously in a constricting ring through the early and middle stages of pneumococcal cell division; but, separate protein complexes in these constricting rings could not be resolved at the resolution of 3D-SIM. The exceptional localization of Pbp2x to the centers of septa in mid-to-late divisional cells suggests that the PG transpeptidase cross-linking activity of Pbp2x functions not only in septal PG synthesis, but may have other roles in PG remodeling.

RESULTS

pbp2x and pbp2b are essential and mediate septal and peripheral PG synthesis, respectively, in S. pneumoniae D39

Tracking the localization patterns of the primary class B PBPs has been an effective strategy to study preseptal (PBP2) and septal (PBP3) PG synthesis in E. coli (van der Ploeg et al., 2013). Based on this precedent, we performed studies of Pbp2x (septal) and Pbp2b (peripheral) localization in S. pneumoniae. Recent studies have shown that Pbp2x and Pbp2b are essential in laboratory strains of S. pneumoniae and their depletion causes distinctive cell morphologies (Berg et al., 2013, Peters et al., 2014). We performed these studies in the strain D39 genetic background that lacks suppressors found in laboratory strains (Introduction) (Land & Winkler, 2011). We first needed to confirm the essentiality of Pbp2x and Pbp2b and their depletion phenotypes in the D39 background. The capsule of strain D39 causes the formation of short chains of cells, whereas unencapsulated (Δcps) derivatives of D39 form single and diplococcal cells that are highly amenable to binning retrospectively into cell division stages (Fig. 1A) (Land et al., 2013, Tsui et al., 2011). Hence, this study was performed using an isogenic Δcps derivative of strain D39 (Table S1). A merodiploid D39 Δcps (Δpbp2x // P_fcsK-pbp2x⁺) strain was constructed to control Pbp2x amount from an ectopic fucose-inducible promoter (Chan et al., 2003, Land et al., 2013) (Table S1). Depletion of Pbp2x upon removal of fucose led to cessation of growth (Fig. S1A) and formation of cells with a variety of abnormal shapes, including elongated cells with pointed ends (3h, Fig. S1B; below). As depletion continued, Pbp2x-depleted cells enlarged, rounded up, and lysed. This pointed-cell phenotype in the D39 strain is similar to that reported for pneumococcal laboratory strains (Berg et al., 2013, Peters et al., 2014), and contrasts with the formation of elongated rod-shaped cells when Pbp2x activity is inhibited preferentially by methicillin (below), as noted before (Land et al., 2013).

A D39 Δcps (Δpbp2b // P_fcsK-pbp2b⁺) merodiploid strain was also constructed for depletion of Pbp2b (Table S1). Depletion of Pbp2b upon removal of fucose caused cessation of growth of the D39 strain (Fig. S2A) and formation of chains of rounded cells (see below) (Fig. S2B), similar to those reported for pneumococcal laboratory strains (Berg et al., 2013). Whisker-plot analysis showed that the cell length of Pbp2b cells remains the same, but the cell widths increase until the cells become nearly spherical (Fig. S2C). Following prolonged incubation, D39 cells depleted of Pbp2b assume irregular shapes and lyse (data not shown), similar to the laboratory strains. We conclude that Pbp2x and Pbp2b are essential in the D39 genetic background, with depletion causing similar cell defects in septal and peripheral PG synthesis, respectively, and autolysis, as was reported in laboratory strains (Berg et al., 2013, Peters et al., 2014). Depletion of Pbp2x or Pb2b is used in the pulse-labeling experiments with FDAAs described below.

Pbp2b and MreC, like Pbp2x and Pbp1a, localize differently from FtsZ at later division stages

An early conclusion that Pbp2b localized exclusively to pneumococcal cell equators (Morlot et al., 2003) was later modified based on additional 2D IFM data to colocalization of Pbp2b with Pbp2x and other PBPs at division equators and septa (Zapun et al., 2008b). A recent study concludes that FtsZ-RFP colocalizes at midcells with GFP-Pbp2x and GFP-Pbp2b (Fleurie et al., 2014). Our previous work (Land et al., 2013) extended earlier work (Morlot et al., 2003) to show that Pbp2x and Pbp1a follow a different localization pattern than FtsZ and GpsB at later stages of division. To distinguish among these patterns of localization of Pbp2b, we performed dual-protein 2D IFM and 3D-SIM IFM to track Pbp2b and FtsZ in the same cells at different stages of pneumococcal cell division (Land et al., 2013, Tsui et al., 2011, Wayne et al., 2010). In parallel, we performed similar studies on MreC, which like Pbp2b, mediates pneumococcal peripheral PG synthesis (Land & Winkler, 2011, Massidda et al., 2013). Proteins were tagged with short epitope peptides (FLAG, Myc, or HA), expressed from single-copy genes in native chromosomal loci, and strains were checked for absence of tag cleavage, lack of cell morphology and growth defects, and normal PBP activity by staining with fluorescent bocillin (Boc-FL) (Fig. S3) (Land et al., 2013). Localization patterns were confirmed using different epitope tags (Experimental procedures). Because S. pneumoniae cells divide perpendicularly to their long axis and remain initially attached as diplococci (Fig. 1A), IFM images of cells can be sorted retrospectively into division stages (Fig. 2A and S4) (Land et al., 2013, Tsui et al., 2011). Protein intensities were averaged in 2D IFM images of cells at different division stages and compared using the IMA-GUI reported previously (Fig. 2A) (Land et al., 2013). Similar to Pbp2x and Pbp1a (Fig. S4) (Land et al., 2013), Pbp2b and MreC localize differently from FtsZ in mid-to-late divisional cells. In pre- and early-divisional cells (stages 1 and 2, Fig. 2A, and S5A), Pbp2b, Pbp2x, Pbp1a, and MreC colocalize as a ring with FtsZ. In mid-to-late divisional cells (stage 3), a ring of Pbp2b, Pbp2x, Pbp1a, MreC, and some FtsZ constricts at division septa, while most FtsZ appears to migrate to the equators of daughter cells. In late-divisional cells (stage 4), Pbp2b, Pbp2x, Pbp1a, and MreC remain at highly constricted septa, but the majority of these proteins locate with FtsZ at equators. 3D-SIM IFM confirmed that Pbp2b and MreC localize differently than FtsZ in mid-to-late divisional cells (Fig. 2C and S5C).

Fig. 2.

Different localization patterns of Pbp2b and FtsZ during later stages of pneumococcal cell division. A. Averaged images and fluorescence intensity traces of strain IU7217 (FtsZ-Myc Pbp2b-HA) grown to mid-exponential phase in BHI broth and processed for dual-protein 2D IFM and DAPI labeling (see Experimental procedures). Cells were binned into division stages 1-4, and images from the indicated number of cells (n) from two independent biological replicates were averaged using the IMA-GUI program described in Experimental procedures. Row 1, cell shapes from phase-contrast images; row 2, nucleoid locations from DAPI labeling; row 3, Pbp2b locations from IFM; row 4, FtsZ locations from IFM; row 5, normalized mean fluorescence intensity distributions along the horizontal cell axis for each channel (black, phase image; blue, DNA; green, Pbp2b; red, FtsZ). B. Scatter plot of the paired widths from the same cells of FtsZ and Pbp2b fluorescent immunolabeled regions at the midcell equators or septa of strain IU7217 at division stages 1-3. Width measurements and plotting were done using the IMA-GUI program (see Experimental procedures). The dotted reference line intercepts the origin with slope = 1. Differences between the paired widths were calculated for each cell in each divisional stage, and the null hypothesis that the mean difference was 0 was tested by a 1-sample Student t test, where **, p<0.01; ***, p<0.001. Septal widths of stage 4 cells were not analyzed, because FtsZ had re-located from septa to equators of daughter cells. Similar results for FtsZ and Pbp2b localization were obtained for strain IU7944 (FtsZ-Myc Pbp2b-HA⁴) (data not shown). C. Representative 3D-SIM IFM and DAPI images of strain IU7944 (FtsZ-Myc Pbp2b-HA⁴) at different division stages. DNA (DAPI stained image) is false-colored white or blue in columns 1 or 5, respectively. FtsZ and Pbp2b are pseudo-colored as green and red respectively, and overlapping signal is colored yellow. The first row of each panel represents images captured in the XY plane, while second row images were obtained by rotating a section of the mid-cell region around the X or Y axis. In stage 3 cells, FtsZ has begun to re-locate to equators, while Pbp2b remains largely at the septum (arrows). Images are representative of >20 examined cells in different division stages from one experiment. Similar 3D-SIM results were obtained for strain IU7217 (FtsZ-Myc Pbp2b-HA). (data not shown). Scale bar = 1 μm.

The widths of Pbp2b, Pbp2x, Pbp1a, and MreC septal rings appeared to be larger than that of FtsZ in averaged 2D IFM images of cells at different division stages (Fig. 2A, S4, and S5A) (Land et al., 2013). We added a quantitation function to the IMA-GUI program to measure the widths of septal rings in 2D IFM images of cells (see Experimental procedures). Septal widths of different proteins ranged from ≈1.2 μm in stage 1 cells to ≈0.66 μm in stage 4 cells. Paired septal widths of two different proteins detected in the same cell were plotted against each other (Fig. 2B, S4, and S5B; Experimental procedures). The points on scatter plots of FtsZ versus Pbp2b or MreC septal ring widths mostly fall below the diagonal line, indicating that diameters of FtsZ septal rings are smaller than those of Pbp2b or MreC septal rings at division stages 1-3. To statistically analyze these data, we took the average differences of the FtsZ and Pbp2b or MreC septal ring widths and performed a Student t-test for the null hypothesis that the difference is zero. For cells at division stages 1-3, the Pbp2b and MreC septal rings are significantly larger than the FtsZ ring (Fig. 2B). This graphical and statistical method is used throughout this paper. A difference in ring diameters was confirmed by 3D-SIM IFM (Fig. 2C and S5C), where the apparent separation of FtsZ and these other proteins across the membrane in rotated mid-cell images is likely enhanced by the width of primary and fluorescent secondary antibodies (≈10 nm each) bound to each protein. Taken together, the 2D and 3D-SIM IFM results are consistent with the cytoplasmic location of FtsZ compared to the extracellular location of the epitope tags at the C-terminus of the PBPs and MreC (Land et al., 2013, Land & Winkler, 2011).

Pulse-labeling with FDAAs demonstrates that Pbp2x activity separates from that of other PBPs in mid-to-late divisional cells

FDAA probes were recently developed and likely label regions where there is active transpeptidase activity catalyzed by PBPs (Kuru et al., 2012). We reasoned that if Pbp2x and other PBPs separate in mid-to-late stages of pneumococcal cell division, then we should observe multiple regions of FDAA incorporation into septal regions. Wild-type D39 Δcps cells were labeled during steady-state growth with one color of FDAA (HADA, pseudo-colored blue, Fig. 3), washed, and then pulse labeled for 5 min with a second color FDAA (TADA, pseudo-colored red, Fig. 3). Cells were fixed and examined by 3D-SIM to determine labeling patterns of cells at different stages of divisions. A single ring of red labeling was present at equators and constricting septa of early-divisional cells (panels 1 and 2, Fig. 3, where rotated views are from sections of mid-cell regions). However, in mid-to-late divisional cells, the red labeling of the septal regions was at two distinct places: a ring of red labeling surrounded a central solid region of labeling (arrow, panel 3, Fig. 3; WT, Fig. 4). In late-divisional cells, a small dot of red labeling remained between cells with most red labeling occurring at the equators of daughter cells (panel 4, Fig. 3). The separation of the FDAA labeling into two distinct regions at the septa of mid-to-late divisional cells could not be resolved by standard 2D fluorescence microscopy without image rotation (data not shown), as was tried previously (Fleurie et al., 2014). We conclude that the PBP transpeptidase activities separate to two locations in the septa of mid-to-late divisional pneumococcal cells.

Fig. 3.

Pulse labeling with FDAAs reveals separate locations of PBP transpeptidase activity in mid-to-late divisional pneumococcal cells. Strain IU1945 (D39 Δcps) growing exponentially in BHI broth at 37°C was pre-labeled with FDAA HADA (pseudo-colored blue), washed, and pulsed for 5 min with FDAA TADA (pseudo-colored red) as described in Experimental procedures. Cells were visualized by 3D-SIM. The first row of each panel represents images captured in the XY plane, while second row images were obtained by rotating a section of the midcell region around the X or Y axis. Numbers indicate stages of cell division. In stage 3 mid-to-late divisional cells, FDAA labeling occurs at two distinct regions of the septum (arrows). Images are representative of >100 3D-reconstructed cells in different division stages from >3 experiments. Scale bar = 1 μm.

Fig 4.

Lack of central septal FDAA labeling in cells treated with a concentration of methicillin that preferentially inhibits the transpeptidase activity of pneumococcal Pbp2x. Cultures of exponentially growing wild-type strain IU1945 (D39 Δcps) were split, duplicate cultures were treated with methicillin (0.1 μg mL⁻¹) for 40 min, and treated and untreated cultures were pulse labeled for 5 min with FDAA TADA (pseudo-colored red) (see Experimental procedures). Two views of 3D-SIM images of untreated (left) and methicillin-treated (right) cells are shown. The first column of each condition is in the XY plane, and the second column is a rotation of the entire cell around X or Y axis as indicated by the line inside the circular arrows. Methicillin-treated cells are elongated and rings lack the central septal disks of FDAA labeling seen in untreated cells (arrows). Images are representative of >60 3D-reconstructed stage-3 cells for each condition from two biological replicates. Scale bar = 1 μm.

Two experiments demonstrate that the central septal labeling with FDAA is due to Pbp2x transpeptidase activity. Previously we showed that addition of 0.1 μg mL⁻¹ methicillin preferentially inhibited >80% of Pbp2x transpeptidase activity (with <20% inhibition of Pbp3 (DacA) transpeptidase activity) and caused cell elongation (Fig. 1B) (Land et al., 2013). Therefore, we treated wild-type cells with methicillin before pulse labeling with TADA FDAA (Fig. 4). Untreated mid-to-late divisional cells showed central septal labeling with TADA surrounded by a ring of labeling in >80% of ≈60 examined cells (Fig. 4, left, where the rotated views are from ends of cells and also show labeled equatorial rings). In contrast, methicillin treatment allowed labeling of rings, but abolished central septal labelling with TADA in >98% of ≈60 examined cells (Fig. 4, right). Since Pbp2x is preferentially inhibited in these cells (Land et al., 2013) and Pbp3 (DacA) does not localize to septa (Barendt et al., 2011, Morlot et al., 2004), these results support the conclusion that central septal labeling is catalyzed by the transpeptidase activity of class B Pbp2x.

In a second experiment, we depleted Pbp2x before pulse labeling with the TADA FDAA (Fig. 5). pbp2x merodiploid cells (Δpbp2x // P_fcsK-pbp2x⁺) grown in 1% (wt/vol) fucose-containing medium were somewhat smaller than wild-type cells, and more irregular cells were observed (Fig. 5A and S1). PBPs were labeled to saturation with Boc-FL (Experimental procedures), which revealed that Pbp2x was overexpressed by ≈2-fold in this merodiploid strain (data not shown). In contrast to the two-position labeling in most (88%) wild-type cells (Fig. 3 and 4), overexpression of Pbp2x resulted in TADA labeling (pseudo-colored red) as contiguous disks (78%) or nearly contiguous thickened rings and dots (22%) at the septa of mid-to-late divisional merodiploid cells (Fig. 5A, panel IU7506 + fucose). These results were confirmed by overexpression of Pbp2x in a pbp2x⁺ merodiploid strain (pbp2x⁺ // P_fcsK-pbp2x⁺) grown in 1% (wt/vol) fucose-containing medium (data not shown). Removal of fucose from the pbp2x merodiploid strain (Δpbp2x // P_fcsK-pbp2x⁺) for ≈3 h depleted the relative Pbp2x amount to <20% based on Boc-FL labeling (data not shown). Pbp2x depletion led to the formation of enlarged, pointed cells containing septal bands without significant central septal TADA labeling in >90% of 27 examined cells (Fig. 5A, right panels). As a control experiment, TADA pulse labeling was performed in pbp2b merodiploid cells (Δpbp2b // P_fcsK-pbp2b⁺) depleted of Pbp2b (Fig. 5B). In this case, pbp2b merodiploid cells grown in fucose resembled wild-type cells. Spherical cells depleted of Pbp2b still showed central septal labeling, consistent with active Pbp2x activity (Fig. 5B, right). Unexpectedly, in ≈30% of ≈100 examined spherical Pbp2b-depleted cells, the relative orientation of the division planes in sister cells was rotated instead of being parallel (Fig. 5B, third panel). Together, these results indicate that inhibition of Pbp2x activity by methicillin or depletion of Pbp2x abrogates the FDAA labeling at the centers of septa, supporting the conclusion that this labeling is due to the transpeptidase activity of Pbp2x, which locates to septal centers in mid-to-late divisional cells (see below) (Land et al., 2013).

Fig 5.

Lack of central septal FDAA labeling in pneumococcal cells depleted of Pbp2x, but not cells depleted of Pbp2b. A. Wild-type strain IU1945 and pbp2x merodiploid strain IU7506 (Δpbp2x//CEP::P_fcsK-pbp2x⁺) were grown exponentially, and IU7506 was depleted of Pbp2x by shifting to BHI broth lacking fucose as described in Experimental procedures. Pre-labeling with FDAA HADA (pseudo-colored blue), pulse-labeling with FDAA TADA (pseudo-colored red) for 5 min, fixation, and 3D-SIM were performed as described in Experimental procedures. Each panel shows four views of the same dividing cell at division stage 3, with the top two images in the XY plane and the bottom two images rotated as indicated. Wild-type and merodiploid cells expressing Pbp2x have central septal disks of FDAA labeling (long arrows), whereas cells depleted of Pbp2x lack central FDAA labeling in septa (short arrows). Pbp2x was overexpressed ≈2X in strain IU7506 + fucose, accounting for the slightly smaller cells and contiguous disks of FDAA labeling (see text). B. pbp2b merodiploid strain IU7397 (Δpbp2b//ΔbgaA::P_fcsK-pbp2b⁺) was grown, depleted for Pbp2b, and labeled with FDAAs as described above for A and in Experimental procedures. Rings and central FDAA labeling of septa of division stage 3 cells were observed in all cells (long arrows). In about 30% of the spherical Pbp2b-depleted cells, the division plane was unhinged (large arrow; see text). For A or B, images are representative of 27 3D-reconstructed stage 3 cells for each condition from two biological replicates.

Pbp2x protein separates from other PG synthesis proteins besides Pbp1a in mid-to-late division cells

Because of their distinct roles in septal and peripheral PG synthesis, we localized Pbp2x and Pbp2b in the same cells. We also tracked the localization of Pbp2x relative to the StkP Ser/Thr kinase, which binds to Pbp2x at some stage of division (Morlot et al., 2013). Analysis of dual-protein 2D IFM images showed that Pbp2x and Pbp2b colocalize in equatorial rings of the same diameter in pre- and early-divisional cells (stage 1, Fig. 6A). However, in mid- and mid-to-late divisional cells, the average diameter of Pbp2b rings was greater than that of Pbp2x rings (stage 2-4, Fig. 6A). Likewise, Pbp2x and StkP colocalized in pre- and early-divisional cells (stage 1, Fig. 6B), but the StkP rings remained larger than the Pbp2x rings in later division stages (stage 2-4, Fig. 6B). Pbp2x and StkP appeared to migrate to equators of daughter cells at similar stages of division (stages 3 and 4, Fig. 6B), consistent with a conclusion in Peters et al. (2014).

Fig. 6.

2D-IFM analysis showing separate positioning of (A) Pbp2b and Pbp2x and (B) StkP and Pbp2x in septa of mid-to-late divisional pneumococcal cells. Top panels are averaged images and normalized mean fluorescence intensity traces, and bottom graphs are scatter plots of labeled widths obtained from (A) strain IU7145 (Pbp2x-FLAG³ Pbp2b-HA) and (B) strain IU7510 (StkP-FLAG² Pbp2x-HA). Dual-protein 2D IFM and DAPI labeling were performed as described in Experimental procedures. Image averaging and generation of scatter plots of paired labeled widths at midcell equators and septa were done as described for Fig. 2 and in Experimental procedures and (Land et al., 2013). NS, not significant (i.e. difference of means of paired widths = 0); **, p<0.01; and ***, p<0.001. The data shown were obtained from two independent biological replicate experiments.

Dual-protein 3D-SIM IFM images (Fig. 7) supported the localization patterns derived from 2D IFM (Fig. 6). Pbp2b or StkP colocalized with Pbp2x in nonuniform rings with similar diameters at midcells of all early-divisional cells examined (n > 10 cells for each protein pair) (stage 1, Fig. 7). As cell division proceeded, Pbp2x moved inside the rings of Pbp2b and StkP in all the cells examined (n ≥ 7) (stages 2 and 3, Fig. 7). In mid-to-late divisional cells, Pbp2x was detected as a disk between layers of Pbp2b in cells where division was nearly complete (n ≥ 5 cells) (stages 3-4, Fig. 7A). This was the same distinct separation pattern that we reported previously for Pbp2x and Pbp1a (Land et al., 2013). We confirmed Pbp1a, Pbp2b, StkP, and MreC located similarly throughout division (Fig. 8, 9, and S6). Septal widths from 2D IFM images of pairs of these proteins fell on diagonal lines in scatter plots and were statistically indistinguishable (Fig. 8 and S6B). Localization patterns determined by 2D and 3D-SIM IFM of pairs of these proteins were remarkably similar (Fig. 8, 9, and S6). Taken together, we conclude that Pbp2x colocalizes with Pbp2b, Pbp1a, StkP, and MreC in equatorial rings during early stages of division. At some point in mid-to-late division, Pbp2x moves to a central septal location separate and adjacent to closing rings of these other proteins. By the end of ring constriction, all of these proteins appear concentrated at the connected poles and in the equators of daughter cells.

Fig. 7.

Representative 3D-SIM IFM images confirming separate positioning of (A) Pbp2b and Pbp2x and (B) StkP and Pbp2x in septa of mid-to-late divisional pneumococcal cells. Representative 3D-SIM IFM images of strain IU7145 (Pbp2x-FLAG³ Pbp2b-HA) and IU7510 (Pbp2x-HA StkP-FLAG²) at different division stages were obtained as described in Experimental procedures. Arrangement of panels and pseudo-coloring is similar to Fig. 2, except red and green correspond to Pbp2x and Pbp2b in (A), and Pbp2x and StkP in (B), respectively. The first row of each panel represents images captured in the XY plane, while second row images were obtained by rotating a section of the midcell region around the X or Y axis. In stage 1 cells, Pbp2x colocalizes with Pbp2b and StkP; however by stages 2 and 3, Pbp2x separates from Pbp2b and StkP and eventually forms a disk in the septum surrounded by rings of Pbp2b and StkP (arrows). Between stage 3 and 4, Pbp2x appears as one layer between two layers of Pbp2b on the sides of the new daughter cells (also see Fig. 9B) (arrows). Images are representative of >30 examined cells in different division stages from two biological replicates. Scale bar = 1 μm.

Fig. 8.

2D-IFM analysis showing that Pbp2b, Pbp1a, MreC, and StkP colocalize in midcell rings of similar diameters at all stages of pneumococcal cell division. Top panels are averaged images and normalized mean fluorescence intensity traces, and bottom graphs are scatter plots of labeled widths obtained from (A) IU7363 (Pbp2b-HA Pbp1a-L-FLAG³), (B) IU7752 (Pbp2b-HA⁴ MreC-L-FLAG³), and (C) IU7512 (StkP-FLAG² Pbp2b-HA). Dual-protein 2D IFM and DAPI labeling were performed as described in Experimental procedures. Image averaging and generation of scatter plots of paired labeled widths at midcell equators and septa were done as described for Figure 2 and in Experimental procedures and (Land et al., 2013). NS, not significant (i.e. difference of means of paired widths = 0); **, p<0.01; and ***, p<0.001. Data shown were obtained from two independent biological replicates.

Fig. 9.

Representative 3D-SIM images confirming that Pbp2b, Pbp1a, MreC, and StkP colocalize in a constricting midcell ring during pneumococcal cell division. Representative 3D-SIM IFM images were obtained from (A) IU7363 (Pbp1a-L-FLAG³ Pbp2b-HA), (B) IU7752 (Pbp2b-HA⁴ MreC-L-FLAG³), and (C) IU7512 (Pbp2b-HA StkP-FLAG²) at different division stages. Arrangements of panels are similar to Fig. 2C and 7. Images shown are representative of >30 examined cells in different division stages from two biological replicates. Scale bar = 1 μm.

Pbp2x locates inside regions of nascent PG synthesis demarcated by FL-V staining

In a final series of experiments, we determined the localization of Pbp2x, Pbp2b, Pbp1a, StkP, and MreC relative to regions of the cell surface labeled with fluorescent vancomycin (FL-V), indicative of the presence of PG pentapeptide substrate and active PG synthesis (Daniel & Errington, 2003, Ng et al., 2004, Pinho & Errington, 2005, Wheeler et al., 2011). Dual-labeled 2D IFM and 3D-SIM IFM again indicated a distinctive localization pattern for Pbp2x compared to the other four proteins (Fig. 10, 11, S7, and S8). In early-divisional cells, all five proteins and FL-V staining colocalized in equatorial rings. However, as the septal rings constricted, rings of Pbp2x became smaller than regions of FL-V labeling (Fig. 10A and 11A), whereas rings of the other Pbp2b, Pbp1a, StkP, and MreC circled regions of FL-Van labeling (Fig. 10, 11, S7, and S8). At mid-to-late stages of division, regions of FL-V staining filled septa and extended along the boundaries of the joined cells. In these cells, Ppb2x located as a closing disk within the region of FL-V labeling (Fig. 11A), whereas Pbp2b, Pbp1a, StkP, and MreC formed rings threaded around the boundaries marked by FL-V labeling (Fig 11 and S8). In fully divided cells, FL-V and proteins that had not migrated to equators of the daughter cells collapsed into single puncta at the poles of the connected cells. We conclude that relative to the cell surface revealed by FL-V labeling, Pbp2x separates from and precedes other PG synthesis proteins during septal closure in pneumococcal cells.

Fig. 10.

2D-IFM analysis showing smaller septal width of Pbp2x, but larger septal widths of other PG synthesis proteins compared to FL-V staining at different stages of pneumococcal cell division. Dual FL-V staining and protein IFM was performed as described in Experimental procedures. Top panels are averaged images and normalized mean fluorescence intensity traces, and bottom graphs are scatter plots of labeled widths for FL-V staining and immunolabeled proteins for (A) IU6929 (Pbp2x-HA), (B) IU7426 (Pbp2b-HA⁴), and (C) IU5544 (Pbp1a-L-FLAG³). Image averaging and generation of scatter plots of paired labeled widths at midcell equators and septa were done as described for Fig. 2 and in Experimental procedures and (Land et al., 2013). NS, not significant (i.e. difference of means of paired widths = 0); **, p<0.01; and ***, p<0.001. Data shown were obtained from two independent biological replicates.

Fig. 11.

Representative 3D-SIM images showing different distributions of the PBPs with respect to FL-V staining of (A) IU6929 (Pbp2x-HA), (B) IU7426 (Pbp2b-HA⁴), and (C) IU5544 (Pbp1a-L-FLAG³). Dual FL-V staining and protein IFM was performed as described in Experimental procedures. FL-V stain is pseudo-colored green, and Pbp2x (A), Pbp2b (B) and Pbp1a (C) are pseudo-colored as red, and overlapping FL-V and PBP signal is represented by yellow. Numbers to the left of each panel indicate division stages. The first row of each panel shows images captured in the XY plane, while second row images were obtained by rotating a section of the mid-cell region around the X or Y axis. In mid-to-late stage 3 divisional cells, FL-V staining is splitting between the daughter cells, Pbp2x (A) is present in septal centers, and Pbp2b (B) and Pbp1a (C) are present in the outer rims of FL-V labeling (arrows). Images shown are representatives of >20 examined cells in different division stages from two biological replicates. Scale bar = 1 μm.

DISCUSSION

This paper provides two independent pieces of evidence that the PG synthesis apparatus separates into two adjacent complexes in mid-to-late divisional cells in S. pneumoniae D39 cells. First, FDAA probes pulse-labeled two distinct regions of active PBP transpeptidase activity in the septa of mid-to-late divisional cells; a ring around a central contiguous region of FDAA staining (Fig. 3 and 4). The central transpeptidase activity can be attributed to Pbp2x, because the central labeling disappeared when Pbp2x transpeptidase activity was inhibited by the β-lactam antibiotic methicillin (Fig. 4) and when Pbp2x was depleted (Fig. 5). Second, dual-protein and protein-fluorescent vancomycin 2D and 3D-SIM IFM showed that Pbp2x separated to septal centers ahead of adjacent closing rings containing Pbp2b, Pbp1a, StkP, and MreC (Fig. 12). The Pbp2x localization pattern is consistent with the FDAA-labeling results (Fig. 3-5) and supports the conclusion that Pbp2x locates separately from other PG synthesis proteins at later stages of pneumococcal cell division. Eventually, the septal ring, which lacks FtsZ at this stage, closes down to a concentrated polar dot, by which time most of PG synthesis proteins have re-appeared at the equators of daughter cells (Fig. 2, S4, S5, 12) (Land et al., 2013).

Fig. 12.

Summary of localization patterns of S. pneumoniae PBPs, PG synthesis regulators StkP and MreC, and regions of fluorescent-vancomycin (FL-V) staining at different stages of cell division. Sections of midcells looking down the long axes of cells are illustrated. In pre-divisional cells, all proteins and FtsZ are located in an equatorial ring that will become the septum. The FtsZ and PG synthesis protein rings likely coincide, although FtsZ rings appear to be smaller than PG synthesis protein rings, because the FtsZ epitope tag is cytoplasmic, whereas the epitope tags on PG synthesis proteins are extracellular (see text). In mid-divisional cells, Pbp2x has started to form a ring inside of the constricting septal ring that contains the other proteins and surrounds regions of FL-V staining; FtsZ has started to disappear from the septum and reorganize at the equators of daughter cells. In mid-to-late division, FtsZ has left septa and is assembling into new equatorial rings, Pbp2x has moved to septal centers overlapping regions of FL-V staining, while the other proteins remain in a constricting septal ring surrounding regions of FL-V staining. In the last stages of division, cells have or are about to fully separate and the remaining proteins and FL-V staining condense into a polar dot; most of these proteins have started to reassemble at the equatorial rings or daughter cells. See text for additional details.

The results in this paper support and extend our previous conclusion that Pbp2x exhibits distinctive, separate localization from Pbp1a and other PG synthesis proteins in mid-to-late divisional cells (Land et al., 2013). In considering this localization pattern (Fig. 12), it is important to note that Pbp2x colocalizes with Pbp2b, Pbp1a, StkP, and MreC in early stages of division and that Pbp2x remains adjacent to the closing septal ring that contains the other proteins. Therefore, there is opportunity in the division cycle for Pbp2x to interact with Pbp1a, as suggested by suppressor studies (Peters et al., 2014, Zerfass et al., 2009), and with StkP, as suggested by biochemical assays (Morlot et al., 2013). Our results indicate that with the exception of Pbp2x in mid-to-late divisional cells, the protein complexes that mediate septal and peripheral synthesis functions are not separated in the midcell region of pneumococcal cells beyond the ≈100 nm resolution of 3D-SIM. Thus, the idea of widely separated PG synthesis machines at different midcell locations is not supported, as discussed recently in (Fleurie et al., 2014). On the other hand, the resolution of 3D-SIM (≈100 nm) cannot distinguish whether PG synthesis functional complexes are further separated or organized topologically within the larger midcell ring structures detected here. Determination of whether there are adjacent separate complexes that carry out septal and peripheral PG synthesis will require higher resolution approaches than 3D-SIM.

Our results add an important qualification to the generalization from recent studies that pneumococcal Pbp2x localizes similarly to Pbp2b, Pbp1a, and StkP throughout the cell cycle (Beilharz et al., 2012, Fleurie et al., 2014, Morlot, et al., 2013, Peters et al., 2014). In one recent study (Fleurie et al., 2014), FDAA labeling failed to reveal separate regions of transpeptidase activity in the septa of mid-to-late divisional cells, as this labeling pattern would not be observable in transversely displayed cells at conventional resolution. Here, the separation of FDAA labeling was readily detected by high-resolution 3D-SIM (Fig. 3 and 4). Moreover, separation of Pbp2x from other PG synthesis proteins occurs in mid-to-late, but not in early stages, of division (Fig. 12). Reliable detection of Pbp2x separation required averaging of 2D images of populations of cells at different stages of division (Fig. 2, 6, 8, and 10), combined with analysis of 3D-SIM images (Fig. 3, 7, 9, and 11). These combined approaches were not used in previous studies. Finally, conclusions about PG synthesis protein colocalization in one study partly relied on patterns observed in elongated ΔgpsB mutant cells of a laboratory strain that likely contains a suppressor (Fleurie et al., 2014) compared to the strain D39 background used here, in which gpsB is essential (Land et al., 2013).

The mechanisms that target Pbp2x to the centers of septa in later stages of pneumococcal division are currently unknown. Recent results establish that the extracellular C-terminal PASTA domains, but not the transpeptidase domain or its activity, are required for normal localization of pneumococcal Pbp2x (Peters et al., 2014). This conclusion was further supported in a new paper showing that the cytoplasmic and transmembrane domains of pneumococcal Pbp2x are essential for function, but not for normal localization (Berg et al., 2014). In another recent study, the extracellular PASTA domains of the StkP Ser/Thr kinase were implicated in normal localization of Pbp2x, possibly through a direct interaction (Morlot et al., 2013). However, the relationship between StkP and Pbp2x localization seems to have broken down in ΔgpsB mutants, which showed mislocalization of StkP, but not Pbp2x (Fleurie et al., 2014). The roles of PASTA and other domains of Pbp2x and StkP in early-stage equatorial and mid-to-late-stage septal localization remain to be determined.

Its exceptional localization in later divisional cells suggests that Pbp2x may play additional roles in septal completion. We tested whether Pbp2x and Pbp1a localization patterns changed in a mutant that was deleted for the other class A PBPs, Pbp2a and Pbp1b. In a Δpbp2a Δpbp1b mutant (IU8447, Table S1), Pbp1a is the only enzyme that catalyzes PG transglycosylase activity that makes glycan chains (see (Massidda et al., 2013)). We found that Pbp1a and Pbp2x localization was the same in the Δpbp2a Δpbp1b mutant as in the wild-type parent (data not shown). Thus, in this mutant, the Pbp1a PG transglycosylase activity does not localize with Pbp2x to the centers of septa in mid-to-late divisional cells. Pbp1a and Pbp2x colocalize again only in the last stages of division, when the septal ring condenses to a dot, and in the next division cycle at the equators of predivisional cells (Fig. 12). Together these results suggest the following model. In early stages of division, Pbp2x is likely required for cross-link formation in septal PG synthesis. In later stages of division after Pbp2x separates to septal centers, this role in septal PG synthesis may continue by using glycan strands synthesized by adjacent Pbp1a and other class A PBPs. In addition, separated Pbp2x may assume a role in PG remodeling by attaching new glycan strands to old glycan strands, whose cross-links have been cleaved by remodeling PG endopeptidases (Singh et al., 2012). Current studies are aimed at comparing the localization patterns of septal remodeling PG hydrolases, such as the PcsB:FtsEX complex (Bartual et al., 2014, Sham et al., 2011, Sham et al., 2013, Sham et al., 2012), to those of Pbp2x and other PG synthesis proteins in septa of mid-to-late divisional cells (Fig. 12).

Finally, the results presented here show that FDAA labeling patterns are strongly correlated with the D,D-transpeptidase activities of pneumococcal PBPs (Fig. 3-5). In wild-type pneumococcal cells, FDAA labeling appears as a single ring at equators of early divisional cells and as a ring surrounded by a central disk at the septa of mid-to-late divisional cells (Fig. 3 and 4). This pattern matches the localization of the Pbp2x protein at different division stages (Fig. 12). As a cautionary note, even modest overexpression (≈2X) of Pbp2x obscures the two-position FDAA labeling pattern at septa, and FDAA labeling now occurs as thickened disks over entire septa (Fig. 5A). This modest overexpression of Pbp2x is accompanied by moderate size and shape changes of cells (Fig. S1B). Inhibition of Pbp2x transpeptidase activity or depletion of Pbp2x amount results in FDAA labeling of equatorial rings, but not central septal disks (Fig. 4 and 5A). HPLC analysis of pneumococcal PG peptides showed that FDAAs predominantly label PG pentapeptide monomers and PG peptide dimers containing a pentapeptide (data not shown). S. pneumoniae lacks homologues of L,D-transpeptidase that are found in E. coli and other bacteria [see (Sanders & Pavelka, 2013)]. Moreover, FDAA labeling of equators and septa is abrogated by prior treatment of cells with penicillin or vancomycin that generally inhibit PBP transpeptidation; after antibiotic addition, FDAA distributes throughout pneumococcal cells (data not shown). Together, these results indicate that FDAA labeling reports regions on cells where PBP transpeptidases are active at different stages of the pneumococcal cell cycle.

EXPERIMENTAL PROCEDURES

Bacterial strains and growth conditions

Bacterial strains used in this study were derived from strain IU1945, an unencapsulated derivative of serotype 2 S. pneumoniae strain D39 (Lanie et al., 2007) (Table S1). Strains containing antibiotic markers were constructed by transforming linear DNA amplicons synthesized by overlapping fusion PCR into competent pneumococcal cells as described previously (Ramos-Montanez et al., 2008, Tsui et al., 2011, Tsui et al., 2010). Primers used for the generation of amplicons are listed in Table S2. All constructs were confirmed by DNA sequencing of chromosomal regions corresponding to the amplicon region used for transformation. Bacteria were grown on plates containing trypticase soy agar II (modified; Becton-Dickinson) and 5% (vol/vol) defibrinated sheep blood (TSAII-BA). Plates were incubated at 37°C in an atmosphere of 5% CO₂. For antibiotic selections, TSAII BA plates contained 250 μg kanamycin mL⁻¹, 150 μg spectinomycin mL⁻¹, 0.3 μg erythromycin mL^-−1, 250 μg streptomycin mL⁻¹, or 2.5 μg chloramphenicol mL⁻¹. Strains were cultured statically in Becton-Dickinson brain heart infusion (BHI) broth at 37°C in an atmosphere of 5% CO₂, and growth was monitored by OD₆₂₀ as described before (Land et al., 2013, Tsui et al., 2011, Tsui et al., 2010). Bacteria were inoculated into BHI broth from frozen cultures or colonies, serially diluted into the same medium, and propagated overnight. For growth experiments, overnight cultures that were still in exponential phase (OD₆₂₀ = 0.1 to 0.4) were diluted back to OD₆₂₀ ≈ 0.003 to start final cultures, which lacked antibiotics.

Transformations of Δpbp2x<>aad9 amplicon into merodiploid strain IU7470 (pbp2x⁺//CEP::P_fcsK-pbp2x⁺) to obtain strain IU7506 or Δpbp2b<>aad9 amplicon into merodiploid strain IU7337 (pbp2b⁺//ΔbgaA::P_fcsK-pbp2b⁺) to obtain strain IU7397, were carried out in the presence of 1% (wt/vol) L-fucose (Sigma, F2252) in the transformation mix and in all subsequent steps. To deplete cells of Pbp2x or Pbp2b, strains IU7506 (Δpbp2x//CEP::P_fcsK-pbp2x⁺) or IU7397 (Δpbp2b//ΔbgaA::P_fcsK-pbp2b⁺) were grown overnight in BHI broth containing 1.0% (wt/vol) L-fucose. Cells from overnight cultures were collected by centrifugation (10 min at 16,000 × g at 25°C) and resuspended to OD₆₂₀ ≈ 0.005 in BHI broth containing 1.0% (wt/vol) L-fucose or no fucose.

Cell length and width measurements

Cell lengths and widths of parent strain IU1945 and strain IU7397 (Δpbp2b//ΔbgaA::P_fcsK-pbp2b⁺) grown in BHI broth with or without fucose were measured from phase-contrast images by using Nikon NIS-Element AR software as described before (Barendt et al., 2009). For IU1945, separated stage 1 cells or daughter cells of stage 4 cells were measured. For strain IU7397 grown with or without fucose, measurements included cells in chains whose widths at constriction sites were < 50% of cell widths.

Boc-FL labeling of PBPs in pbp2x merodiploid and HA-tagged pbp2b strains

Quantitative labeling with fluorescent bocillin (Boc-FL) (Molecular Probes) to determine PBP activity and/or amount was performed for strains expressing Pbp2b-HA (IU6933), or Pbp2b-HA⁴ (IU7426), and for a pbp2x merodiploid strain (IU7506) as described previously (Kocaoglu et al., 2012, Land et al., 2013, Zhao et al., 1999).

Western blotting

Strains were grown exponentially in BHI broth to OD₆₂₀ ≈ 0.15. Lysates were prepared as described previously (Wayne et al., 2010), and polypeptides were separated using 4-15% mini-protean TGX pre-cast gradient gels (Bio-Rad, 456-1083). A Benchmark pre-stained protein standard (Life Technologies, 10748-010) calibrated against a Novex sharp unstained protein standard (Life Technologies, LC5801) was used to determine polypeptide molecular masses. FLAG-, HA-, and Myc-tagged proteins were detected by Western blotting using a 1:1000 dilution of primary anti-FLAG polyclonal antibody (Sigma, F7425), anti-HA rabbit polyclonal antibody (Invitrogen, 71-5500) at 1 μg mL⁻¹, or a 1:1000 dilution of primary anti-Myc polyclonal rabbit antibody (Sigma, C3956), followed by a 1:10000 dilution of HRP-conjugated donkey anti-rabbit antibody (GE Healthcare, NA934), and ECL detection reagent. Chemiluminescent signal from polypeptide bands was detected by using an IVIS imaging system as described in (Wayne et al., 2010).

2D immunofluorescence microscopy (IFM)

Localization of FLAG-, Myc- and HA-tagged proteins by IFM was performed for exponentially growing cells as described before (Land et al., 2013, Tsui et al., 2011, Wayne et al., 2010). Primary antibodies used were anti-FLAG rabbit polyclonal antibody (Sigma, F7425, 1:100 dilution), anti-FLAG mouse monoclonal M2 antibody (Sigma, F1804, 1:100), anti-c-Myc rabbit polyclonal antibody (Sigma, C3956, 1:100 dilution), anti-c-Myc mouse monoclonal antibody (Life Technologies, R950-25, 1:100 dilution), or anti-HA rabbit polyclonal antibody (Invitrogen, 71-5500, 1:50 dilution). Secondary antibodies used were Alexa Fluor 488 or Alexa Fluor 568 highly cross-adsorbed goat anti-rabbit IgG (Life Technologies; A11034 or A11036, respectively, 1:100 dilution) and Alexa Fluor 488 or Alexa Fluor 568 (Life Technologies; A11029 or A11031, respectively, 1:100 dilution) highly cross-adsorbed goat anti-mouse IgG. For-single-protein labeling experiments, polyclonal rabbit antibodies were used followed by Alexa Fluor 488 or Alexa Fluor 568 goat anti-rabbit IgG. For dual labeling of FLAG- and Myc-tagged proteins [(IU6978 (pbp2x-FLAG³ ftsZ-Myc); IU7074 (mreC-L-FLAG³ ftsZ-Myc)], primary antibodies were anti-FLAG rabbit polyclonal antibody and mouse monoclonal antibody against c-Myc, and secondary antibodies were Alexa Fluor 488 goat anti-rabbit and Alexa Fluor 568 goat anti-mouse. For dual labeling of HA- and Myc-tagged proteins [IU7217 (pbp2b-HA ftsZ-Myc)], primary antibodies were anti-HA rabbit polyclonal antibody and anti-Myc mouse monoclonal antibody, and secondary antibodies were Alexa Fluor 488 goat anti-rabbit and Alexa Fluor 568 goat anti-mouse. For dual labeling of HA- and FLAG-tagged proteins [(IU7145 (pbp2b-HA pbp2x-FLAG³); IU7590 (stkP-HA pbp1a-L-FLAG³); IU7752 (pbp2b-HA⁴ mreC-L-FLAG³); IU7363 (pbp2b-HA pbp1a-L-FLAG³); IU7365 (pbp2x-HA pbp1a-L-FLAG³); IU7510 (pbp2x-HA stkP-FLAG²); IU7512 (pbp2b-HA stkP-FLAG²)], primary antibodies were anti-HA rabbit polyclonal antibody and anti-FLAG mouse monoclonal antibody. Secondary antibodies were Alexa Fluor 488 goat anti-rabbit and Alexa Fluor 568 goat anti-mouse for strains IU7145, IU7590 and IU7752, and Alexa Fluor 568 goat anti-rabbit and Alexa Fluor 488 goat anti-mouse for strains IU7363, IU7365, IU7510, and IU7512. Primary antibody incubations were for 1h at 24°C for single tagged strains, and FLAG-Myc and Myc-HA double tagged strains and for 2h at 37°C for FLAG-HA double tagged strains. Secondary antibody incubation was for 1h at 24 °C for all strains. DNA in nucleoids was stained by using mounting media SlowFade gold antifade reagent with DAPI (Life Technologies, S36936).

Control experiments showed similar localization patterns for each protein when tracked in single or dual-protein labeling experiments (data not shown). Similar localization patterns were observed for each protein fused to different tags [Pbp2x fused to HA (IU6929), FLAG³(IU6819), or HA³ (IU7722); Pbp2b fused to HA (IU6933) or HA⁴ (IU7426); StkP fused to FLAG² (IU7434) or HA (IU7438)]. Similar localization patterns were observed when Alexa Fluor 488 or Alexa Fluor 568 combinations were swapped in dual-protein experiments [IU7217 (pbp2b-HA ftsZ-Myc); IU7590 (pbp1a-L-FLAG³ stkP-HA)] (data not shown). Control experiments showed no detectable labeling in cells not expressing tagged proteins (IU1945) with the combinations of antibodies used (data not shown). Labeling of strains containing single-tagged proteins tested with the double labeling procedure that contained both sets of primary and secondary antibodies produced signal only in the expected fluorescence channel (data not shown).

Image analysis of 2D IFM images

2D IFM images of dividing pneumococcal cells were aligned and manually binned into four division stages (pre-, early-, mid-, and late-divisional) using a point-and-click image-analysis graphical user interface (IMA-GUI) organized in MATLAB (The Mathworks) as reported before (Land et al., 2013). Mean normalized cell outlines from phase-contrast images and mean normalized fluorescent intensities of DNA and proteins from all aligned cells at a given stage were calculated to graphically represent the relative distribution of molecules along the longitudinal cell axis during the division process (Land et al., 2013). Numbers of cells averaged for each stage are indicated (n) and were taken from at least two fields from each of two independent biological replicate experiments, unless specified otherwise.

To quantitatively evaluate the relative positions of individual proteins through the process of septation, we extended the IMA-GUI to include methods for independently marking the septal edge positions of different proteins in the same double-labeled cell. The distance between each edge represents the extent to which the protein is distributed within the septal ring (i.e. its septal width). Widths were determined for protein septal rings with sharp edges in images of cells averaged at different division stages. Using the cell morphology as an indicator of cell cycle phase, we reasoned that if two proteins were discretely colocalized, the mean difference in septal width for all cells in a particular phase would be centered at zero with a normal distribution across the cell population. For statistical analyses, paired widths from individual cells were output to a data sheet in GraphPad Prism 5. Some proteins were largely absent from septa of stage 4 cells (e.g., FtsZ, Fig. 2A) or did not form septal protein rings with distinct edges and were not included in width comparisons. For pairs of different proteins, septal widths were compared for 63-100%, 78-99%, 67-100%, and 20-100% of binned cells at stages 1, 2, 3, and 4, respectively. Differences between paired widths were calculated for cells in each division stage. One sample Student-t tests were performed to determine whether mean differences in widths were significantly different from the null hypothesis value of zero (NS, not significant; **, p<0.01; and ***, p<0.001 in graphs). To visualize the extent to which each protein pair changes relative width through different cell divisional stages, paired data from each cell were plotted as a scatterplot with a superimposed diagonal line to indicate the mean position for the assumption that the two proteins are strictly colocalized through the cell cycle.

3D-SIM (structured illumination microscopy) IFM

Samples were prepared as described previously (Land et al., 2013), and 3D-SIM was performed using the OMX 3D-SIM super-resolution system located in the Indiana University Bloomington Light Microscopy Imaging Center (http://www.indiana.edu/~lmic/microscopes/OMX.html). The system used is equipped with four Photometrics Cascade II EMCCD cameras that allow simultaneous four color imaging, and is controlled by DV-OMX software, with image processing by Applied Precision softWoRx 6.0.0 software. Exposure times and %T settings for DAPI, Alexa 488, and Alexa 568 images were 50 to 100 ms and 50%, 50 ms and 1 to 10%, and 50 ms and 10 to 50%, respectively.

3D-SIM of FDAA pre-labeled and pulse-labeled cells

FDAAs HADA (7-hydroxycoumarin-3-carboxylic acid 3-amino-D-alanine) and TADA (tetramethylrhodamine 3-amino-D-alanine) were synthesized as reported in Kuru et al. (2012), with the following change: TADA was synthesized as reported for TDL, except that Boc-D-DAP-OH (N-alpha-t-Butyloxycarbonyl-D-2,3-diaminopropionic acid) was used in place of Boc-D-Lys-OH (N-alpha-t-Butyloxycarbonyl-D-lysine). Working solutions of HADA and TADA in BHI were diluted from 500 mM stocks in DMSO, which were stored at −70°C. For HADA pre-labeling, bacterial cells from overnight cultures were diluted to OD₆₂₀ ≈ 0.01 in 2 mL of fresh BHI broth (37°C) and incubated at 37°C in an atmosphere of 5% CO₂. At OD₆₂₀ ≈ 0.02, 1 μL of 500 mM HADA in DMSO was added to a final concentration of 250 μM. At OD₆₂₀ ≈ 0.2-0.25 (≈ 3.5 doublings), cultures were transferred to a 2-mL microfuge tube, placed in an ice bath for 1 min, and then centrifuged for 5 min at 16,000 × g at 4°C. Supernates were discarded, and pellets were resuspended in 1 mL of cold (4°C) phosphate-buffered saline (PBS; Ambion; AM9625). Cells were centrifuged for 2.5 min at 16,000 × g at 4°C, resuspended in cold PBS, and centrifuged again. After the second wash and centrifugation, cells were pulse-labeled with TADA. Pellets were resuspended in 1 mL of warm (37°C) BHI broth containing 500 μM TADA. Cells were incubated at 37°C in an atmosphere of 5% CO₂ for 5 min (14% of doubling time) and then chilled on ice for 1 min. Cells were centrifuged for 2.5 min at 16,000 × g at 4°C, and pellets were resuspended in 1 mL of cold (4°C) PBS). After the second wash and centrifugation, pellets were resuspended in 1 mL of 4% paraformaldehyde (EMS; 157-4) and incubated for 15 min at room temperature, followed by 45 min on ice. Fixed cells were centrifuged, and pellets were washed three times with PBS at 4°C as described above. After the third wash and centrifugation, cells were resuspended in 0.3 ml of cold (4°C) GTE buffer (50 mM glucose, 1 mM EDTA, 20 mM Tris-HCl, pH 7.5), and stored in the dark at 4°C for up to 20 h. Fifty μL of cell suspension was deposited on a methanol-treated coverslip and incubated for 5 min at room temperature. Unattached cells were aspirated, and attached cells were treated with 50 μL of 0.2% Triton X-100 (Mallinckrodt; H282) in PBS (PBS-T) for 10 s. PBS-T was aspirated from the wells, and the coverslip was air dried completely before immersion in cold (−20°C) methanol and kept at −20°C for 10 min. The coverslip was then removed from methanol and air dried, after which the coverslip was incubated with 50 μL of PBS-T for 5 min, followed by two washes with 50 μL PBS and aspirated immediately, and once with 50 uL PBS for 5 min at room temperature before aspiration. After the coverslip was air dried, 7 μL of Slowfade Gold antifade reagent (Invitrogen; S36936) was applied to the coverslip, which was then inverted onto a cleaned and methanol-treated slide, and sealed. 3D-SIM was performed using the OMX 3D-SIM super-resolution system described above. Laser lines used for examination of samples were 405 with emission filters of 419-465 for HADA, and 561 with emission filters of 609-705 for TADA. Exposure times and %T settings for both HADA and TADA were 50 ms and 30%. Images in Fig. 3 are representative of >100 3D-reconstructed cells in different division stages from >3 independent experiments.

For TADA labeling after methicillin treatment of strain IU1945 (D39 Δcps, WT), duplicate 5 mL cultures of IU1945 were grown to exponential phase (OD₆₂₀ ≈ 0.1) at 37°C in an atmosphere of 5% CO₂. Methicillin was added to a final concentration of 0.1 μg mL⁻¹ to one culture (Land et al., 2013). After 40 min, 250 μL samples of the treated and untreated cultures were added to a microfuge tubes. TADA was added to a final concentration of 500 μM, and incubation was continued for 5 min at 37°C. Samples were prepared for 3D-SIM as described above. Images in Figure 4 are representative of >60 3D reconstructed stage-3 cells from two independent experiments.

FDAA labeling and 3D-SIM microscopy of merodiploid strains IU7506 (Δpbp2x//CEP::P_fcsK-pbp2x⁺) and IU7397 (Δpbp2b//ΔbgaA::P_fcsK-pbp2b⁺) were performed as described above, with the following changes. Overnight cultures of strains IU7506 and IU7397 in BHI broth containing 1.0% (wt/vol) L-fucose were centrifuged, and pellets were resuspended into 5 mL of warm (37°C) BHI broth containing or lacking 1.0% (wt/vol) L-fucose and diluted to a starting OD ≈ 0.01. To conserve FDAA usage, HADA was added to a final concentration of 250 μM to a 500 μL aliquot of each starting culture in a microfuge tube, which was covered by perforated parafilm and incubated at 37°C in an atmosphere of 5% CO₂. Parallel OD₆₂₀ readings were monitored for the remaining 4.5 mL of cultures. At OD₆₂₀ ≈ 0.1 (≈ 2.8 h after the start of the cultures), samples were washed and resuspended in 250 μL BHI broth containing TADA at a final concentration of 500 μM. The cultures were incubated at 37°C for 5 min, washed, and prepared for 3D-SIM as described above. Images shown in Fig. 5 are representative of >24 3D-reconstructed stage-3 cells for each condition from two independent biological replicates.

Dual FL-V staining and protein IFM

100 μg of Bodipy-FL-conjugated vancomycin (FL-V) (Molecular Probes, V-34850) was dissolved in 10 μL of DMSO, and 90 μL of sterile water was added to make a 1 μg μL⁻¹ solution. This solution was mixed with equal volume of unlabeled vancomycin at the same concentration. 3 μL of this mixture was added to 1.5 mL of exponentially growing cells (OD₆₂₀ ≈ 0.15) in BHI broth to give a final concentration of 2 μg mL⁻¹ total vancomycin. Samples were incubated at room temperature for 5 min, placed on ice for 2 min, collected by centrifugation, and washed twice with cold PBS, followed by fixation with 4% paraformaldehyde and subsequent steps used for IFM (Land et al., 2013, Wayne et al., 2010). FLAG-tag strains [IU5544 (pbp1a-L-FLAG³), IU7434 (stkP-FLAG²); IU4970 (mreC-L-FLAG³)] and HA-tagged strains [IU6929 (pbp2x-HA); IU7426 (pbp2b-HA⁴)] were labeled with primary anti-FLAG rabbit polyclonal antibody or anti-HA rabbit polyclonal antibody, respectively, followed by Alexa Fluor 568 highly cross-absorbed goat anti-rabbit IgG. DNA in nucleoids was stained by using mounting media SlowFade gold antifade reagent with DAPI. Similar cell morphology and protein localization were observed in samples labeled or not labeled with FL-V, indicating that the FL-V labeling procedure did not alter localization of Pbp1a, Pbp2b, Pbp2x, MreC, and StkP (data not shown).