Organization of Peptidoglycan Synthesis in Nodes and Separate Rings at Different Stages of Cell Division of Streptococcus pneumoniae

Abstract

Bacterial peptidoglycan (PG) synthesis requires strict spatiotemporal organization to reproduce specific cell shapes. In ovoid-shaped Streptococcus pneumoniae (Spn), septal and peripheral (elongation) PG synthesis occur simultaneously at midcell. To uncover the organization of proteins and activities that carry out these two modes of PG synthesis, we examined Spn cells vertically oriented onto their poles to image the division plane at the high lateral resolution of 3D-SIM (structured-illumination microscopy). Labelling with fluorescent D-amino acids (FDAA) showed that areas of new transpeptidase (TP) activity catalyzed by penicillin-binding proteins (PBPs) separate into a pair of concentric rings early in division, representing peripheral PG (pPG) synthesis (outer ring) and the leading-edge (inner ring) of septal PG (sPG) synthesis. Fluorescently tagged PBP2x or FtsZ locate primarily to the inner FDAA-marked ring, whereas PBP2b and FtsX remain in the outer ring, suggesting roles in sPG or pPG synthesis, respectively. Pulses of FDAA labeling revealed an arrangement of separate regularly spaced “nodes” of TP activity around the division site of predivisional cells. Tagged PBP2x, PBP2b, and FtsX proteins also exhibited nodal patterns with spacing comparable to that of FDAA labeling. Together, these results reveal new aspects of spatially ordered PG synthesis in ovococcal bacteria during cell division.

1 |. INTRODUCTION

Streptococcus pneumoniae (pneumococcus; Spn) is a Gram-positive, commensal bacterium of humans and a major opportunistic pathogen that causes life-threatening illnesses, including pneumonia, bacteremia, and meningitis (Loughran et al., 2019, Weiser et al., 2018). Spn is a prolate ellipsoid-shaped bacterium, referred to as an ovococcus (Zapun et al., 2008). Its distinct ovoid shape is maintained by a thick peptidoglycan (PG) wall that surrounds the entire cell (Vollmer et al., 2019). PG consists of glycan sugar chains of alternating units of β-(1,4) linked N-acetylglucosamine (GlcNAc) and N-acetylmuramic acid (MurNAc). Glycan chains are crosslinked together by PG peptides attached to MurNAc, thereby forming a mesh-like network (Egan et al., 2020, Egan et al., 2015). As in most eubacteria, Spn PG synthesis is essential for normal growth and cell division and determines cell shape and chaining, which impact colonization and virulence (Vollmer et al., 2019, Siegel & Weiser, 2015, Rodriguez et al., 2012). Besides protecting cells from turgor stress, Spn PG serves as a scaffold for surface-attached virulence factors, including capsule, sortase-attached proteins, and wall teichoic acids (Vollmer et al., 2019).

In Spn, PG synthesis is organized initially by an FtsZ ring at the midcell equator perpendicular to the long axis of newly divided cells (Fig. S1). PG synthesis is carried out by distinct septal PG (sPG) and peripheral PG (pPG) modes (Berg et al., 2014, Rued et al., 2017, Tsui et al., 2014). sPG synthesis produces the cross wall that separates daughter cells (Land et al., 2013, Perez et al., 2019, Straume et al., 2020), whereas concurrent pPG synthesis elongates daughter cells from midcell to form ovoid-shaped cells (Fig. S1) (Land et al., 2013, Perez et al., 2019, Straume et al., 2020). pPG synthesis in Spn functionally resembles preseptal PG elongation that occurs briefly at the beginning of division of rod-shaped bacteria (Aaron et al., 2007, Potluri et al., 2012). However, pPG elongation in Spn cells remains confined to the midcell region, in contrast to lateral PG elongation organized by MreB-containing Rod complexes moving over the body of rod-shaped cells perpendicular to the long axis (Egan et al., 2020).

The location of these two modes of PG synthesis within the same midcell region requires significant coordination and organization. Previously, we used dual-protein 2D-immunofluorescence microscopy (IFM) and high-resolution 3D-SIM to determine the relative localization patterns of pneumococcal division and PG synthesis proteins at different stages of division (Land et al., 2013, Rued et al., 2017, Tsui et al., 2014). Stages of cell division were assigned retrospectively based on static “snap-shot” images of cells from non-synchronized exponential cultures. These studies showed that in newly divided, predivisional (stage 1) daughter cells, division and PG synthesis proteins locate to FtsZ-organized midcell rings (Fig. S1). As nearly simultaneous sPG and pPG synthesis begins (stage 2) (Wheeler et al., 2011), the equatorial ring begins to constrict and becomes the new septal ring. Some FtsZ and associated proteins EzrA and FtsA move outward continuously with MapZ from the septal ring toward the positions of the future equatorial rings in daughter cells (Perez et al., 2019). As constriction and midcell elongation continue (stage 3), FtsZ amount decreases at the midcell of middle- to late-divisional cells and begins to accumulate at the equators of daughter cells. Meanwhile, PG synthesis proteins remain at the midcell septal ring.

At the resolution of standard 2D-epifluorescence microscopy (2D-EFM), essential class B bPBP2x transpeptidase (TP), which interacts with the FtsW SEDS glycosyltransferase (GT) to carry out sPG synthesis (Perez et al., 2019, Taguchi et al., 2019), locates in an apparent inner ring within an outer ring containing the class B bPBP2b TP (Tsui et al., 2014), which interacts with the RodA SEDS GT to carry out pPG synthesis (Fig. S1) (Meeske et al., 2016, Sjodt et al., 2020). Several other proteins implicated in pPG synthesis locate to this apparent outer peripheral ring, including MreC that organizes and may activate bPBP2b (Tsui et al., 2016), class A aPBP1a that has both GT and TP activities (Land et al., 2013), MltG endo-lytic transglycosylase that cleaves glycan chains (Tsui et al., 2016), and StkP serine/threonine kinase that mediates cell division (Land et al., 2013, Tsui et al., 2014). Near the end of cell division (stage 4), remaining PG synthesis proteins that have not moved to developing equatorial rings converge to a small spot between the daughter cells (Fig. S1) (Tsui et al., 2014), before moving to the equatorial rings of the daughter cells (Perez et al., 2019).

Additional evidence for physical separation of the sPG and pPG synthesis machines midway through cell division was obtained by staining with fluorescent vancomycin (FL-Vanco), which labels PG pentapeptides in areas of PG synthesis (Daniel & Errington, 2003), with fluorescent D-amino acids (FDAAs), which label PG in regions of active PBP TP activity (Hsu et al., 2017, Kuru et al., 2012), and with an activity-based β-lactone fluorescent probe (7FL) specific for bPBP2x (Sharifzadeh et al., 2017, Sharifzadeh et al., 2020). FL-Vanco labeling combined with IFM of mid-divisional cells indicated localization of bPBP2x inside of regions of nascent PG synthesis that were surrounded by bPBP2b and other pPG synthesis proteins (Tsui et al., 2014). Long pulses of FDAAs for 5 min followed by 3D-SIM showed labeling of a midcell ring surrounding a prominent central dot of FDAA labeling, resembling a “Saturn-like” pattern, in mid-divisional cells (Tsui et al., 2014). Specific inhibition with the β-lactam antibiotic methicillin and protein depletion experiments indicated that the central dot of septal TP activity is attributable to bPBP2x (Tsui et al., 2014). Likewise, these central dots of FDAA labeling disappeared in rings of non-constricted septa in cells depleted for the GpsB regulatory protein (Rued et al., 2017). Last, spatial separation of bPBP2x was confirmed with a bPBP2x-specific, activity-based β-lactone fluorescent probe (7FL, Lac(L)-Phe-FL), whose labeling recapitulated a “Saturn-like” labeling pattern, indicating that although most active bPBP2x moves toward the center of the septum, some still remains in the outer peripheral ring (Sharifzadeh et al., 2017, Sharifzadeh et al., 2020).

Limitations of these previous fluorescent-protein and -probe studies are the low resolution (≈250 nm) of conventional 2D-EFM and of rotated 3D-SIM images (Holden, 2018). Due to their ovoid shape, Spn cells predominantly lie sideways along their long axis during imaging on a flat surface. Imaging of septal rings and associated structures in horizontal cells by 3D-SIM is limited by the Z-axial resolution of 250–300 nm of the system, which is much lower than the 100 nm resolution in the XY lateral plane (Holden, 2018). This axial resolution limitation remains when reconstructed image volumes of horizontally oriented cells are rotated by 90° to visualize septal rings as circles. To overcome this obstacle, we devised a simple method to reliably orient a small, but sufficient, number of Spn cells vertically onto their poles (i.e., “on their heads”) for 3D-SIM imaging, thereby enabling high lateral resolution of division rings. Using this approach, we were able to detect FDAA-labeled intermediate concentric rings of sPG synthesis TP activity in strains expressing fluorescent PG synthesis proteins. This approach located the proteins to the outer peripheral or inner septal rings and showed that these proteins are organized as nodal structures distributed regularly around these rings in early divisional cells. We further found that short pulses of FDAA labeling also revealed nodal distributions of TP activity that were disrupted in division mutants. Together, these results reveal a highly uniform organization of coordinated PG synthesis and distinct separation of the sPG and pPG synthesis machines at the midcell of dividing ovococcal bacteria.

2 |. RESULTS

2.1 |. “Snap-shots” of vertically oriented Spn cells labeled with FDAAs reveal constriction of concentric septal and peripheral rings

To circumvent the lower resolution during rotation of 3D-SIM images, we devised a simple method to capture a sufficient number of vertically oriented, fixed Spn cells to image division planes at high ≈100 nm × ≈100 nm XY lateral resolution. Unencapsulated D39W (Δcps) was used in this study (Table S1), because encapsulated D39 forms long chains of cells that make microscopy difficult, and capsule tends to mask morphological defects of some PG synthesis mutants (Barendt et al., 2009). D39 Δcps wild-type (WT) cells were first labeled with a blue FDAA (HADA) for several generations and then with a red FDAA (TADA) for 2.5 min to label sites of new PG synthesis (Fig. 1A and 1B). Cells were fixed with formaldehyde and resuspended in a small amount of hardening anti-fade solution pipetted onto small, round glass coverslips onto which slides were placed. The anti-fade suspension medium minimized photobleaching, and its viscosity and the large number of cells in samples trapped some cells in a vertical orientation. This procedure led to a reliable, but small (≤1%), percentage of cells suspended on their poles, which could be spotted as circles (Fig. 1A and 1C), among many horizontally oriented cells (Fig. 1B) in each field.

Figure 1.

3D-SIM of vertically oriented Spn cells labeled with FDAAs reveals spatially distinct, concentric midcell ring intermediates of TP activity at different stages of division. (A) Schematic of labeling procedure (top). WT cells (IU1945) were labeled with 125 μM HADA for several generations (old cell wall; cyan), washed, and labeled with 125 μM TADA for 2.5 min (new cell wall from ≈7% of a generation; red). Cells were fixed and prepared for vertical cell imaging by 3D-SIM (see Experimental Procedures). Representative images are shown of rings from pre-, early, and late-division stages estimated by the diameters of outer TADA rings. Left images, HADA and TADA channels overlaid; right images TADA channel only. (B) Representative horizontally oriented cells from the same field as (A) locating the concentric rings at constricting midcell regions. (C) Montage of manually sorted Spn cells in different stages of cell division. Only the TADA is shown, and images are representative of >50 cells from more than three independent biological replicates. (D) Diagrammatic summary based on data in panel (A), above and on other data in Results of the organization of sPG and pPG synthesis in Spn, including concentric rings of new PG synthesis, membrane invagination, and proteins located in this study. The rings of preexisting and new PG synthesis correspond to the blue and red labeling patterns in panel (A), respectively, and the curved arrows indicate the image rotations.

Vertically oriented Spn cells were captured at different stages of the division cycle (Fig. 1A and 1C). In predivisional cells, red TADA labeling appears as a single ring around the division site (diagrammed in Fig. 1D). Remarkably, as division precedes, TADA labeling is detected in a pair of intermediate concentric rings, which are designated as the outer (peripheral) and inner (septal) rings. The inner ring represents the TP activity at the leading-edge of the closing septum annulus that eventually converges down to a dot, surrounded by the outer ring to form a “Saturn-like” pattern reported before, where the dot results from bPBP2x TP activity (Tsui et al., 2014). In very late divisional cells, all remaining TADA labeling at midcell converges to a single dot at the division site, while new TADA rings appear at the equators of daughter cells. The regions of TP activity, including their relationship to membrane invagination, are summarized in Figure 1D. We conclude that PG synthesis occurs at two distinct, separate locations at the midcell of dividing Spn cells.

2.2 |. bPBP2x and FtsZ primarily locate to the constricting inner septal ring, whereas bPBP2b and FtsX locate to the outer peripheral ring

We reasoned that the spatial separation observed in vertically oriented Spn cells could be used to assign PG synthesis proteins to the sPG and/or pPG synthesis machine(s). We labeled strains expressing isfGFP-bPBP2x or sfGFP-bPBP2b with TADA (125 μM) for a moderate time (2.5 min), and then observed TADA and GFP fluorescence in vertically oriented cells (Fig. 2 and S2). isfGFP-bPBP2x or sfGFP-bPBP2b was expressed from its native chromosomal locus, and fixed cells expressing these constructs exhibited minimal discernible growth or cell morphology defects (Fig. S3A and S3B) (Perez et al., 2019). Western blotting with native antibody showed a single species of each fusion protein (Fig. S3C). Nevertheless, isfGFP-bPBP2x cellular amount is moderately increased (169%) compared to unlabeled WT bPBP2x, whereas sfGFP-bPBP2b is underproduced to only 12–15% of the WT amount, without causing ostensible defects in cell growth and shape in BHI broth (Fig. S3). We characterized two other strains that express fusion constructs without causing defects in growth and morphology. In one strain, iHT-bPBP2x is expressed at ≈164% or ≈88% of WT bPBP2x level when grown in BHI broth or C+Y medium, respectively (Fig. S2C, S2E, and S3 A-C). In the second iHT-bPBP2b//P_Zn-iHT-bPBP2b merodiploid strain grown in BHI broth containing 0.3 mM Zn²⁺/0.03 mM Mn²⁺, iHT-bPBP2b is expressed at ≈WT bPBP2b level (≈107%; Fig. S2D and S3C).

Figure 2.

bPBP2x or bPBP2b locate to the leading edge of the septal ring annulus or to the outer peripheral ring, respectively, at midcell in vertically oriented Spn cells. Cells expressing isfGFP-bPBP2x (IU11157) or sfGFP-bPBP2b (IU9965) were labeled with 125 μM TADA for 2.5 min, fixed, and prepared for vertical cell imaging by 3D-SIM (see Experimental Procedures). (A) Representative images of cells (6 to 29) at different stages of division. isfGFP-bPBP2x and sfGFP-bPBP2b, pseudo-colored cyan; TADA labeling, pseudo-colored red. Brightness and contrast were manually adjusted to show signals associated with division rings and to reduce background lacking TADA labeling. Progression of cell division was indicated by the separation of the inner (septal) and outer (peripheral) rings. Cellular amounts of fluorescent-proteins produced relative to untagged WT protein in a control strain (see Fig. S3C) are in parentheses. (B) Quantitation of nodal distributions of isfGFP-bPBP2x and sfGFP-bPBP2b (but not TADA) in predivisional Spn cells with single, overlapping rings of TADA and GFP labeling. Distributions of ring diameters (left), nodes per ring (middle), and arc distances (right) were determined as described in Experimental Procedures. Graphs show median (interquartile range; whiskers, 5^th-95^th percentile; +, mean). The apparent bimodality in the arc distance distribution is due to a resolution limit on nodes closer together than ≈0.1 μm. Means ± SD from two independent biological replicates are shown below the graphs and compiled in Table S3. Differences between means were compared by a two-tailed t test (GraphPad Prism). *, P<0.05 and ***, P<0.001.

In predivisional cells grown in BHI broth, TADA treatment for 2.5 min labels equatorial rings fairly uniformly, whereas sfGFP-bPBP2x and sfGFP-bPBP2b appear as nodes distributed regularly around the circumference of cells (Fig. 2A and, S2A, row 1). Independent labeling of WT bPBP2x with the 7FL β-lactone probe specific for bPBP2x corroborated a nodal organization in predivisional equatorial rings of daughter cells (Fig. S2F). As division progresses, isfGFP-bPBP2x moves inward resulting in formation of the inner septal ring (Fig. 2A and S2A, row 2), with some bPBP2x remaining in the outer ring, as was also observed for labeling with the 7FL (Fig. S2F) (Sharifzadeh et al., 2017, Sharifzadeh et al., 2020). Displacement between the single TADA ring and the inner isfGFP-bPBP2x ring (Fig. 2A and S2A, row 2) likely reflects the continued movement inward of isfGFP-bPBP2x during the time required to wash away unincorporated TADA and fix cells. As the TADA-labeled inner septal ring becomes more prominent, isfGFP-bPBP2x condenses into a dense, concentric ring largely lacking detectable nodes at this resolution, with some isfGFP-bPBP2x remaining in the outer ring (Fig. 2A and S2A, row 3).

A similar pattern of localization was observed for slightly overexpressed or underexpressed iHT-bPBP2x in cells grown in BHI broth or C+Y medium, respectively (Fig. S2C and S2E), including residual iHT-bPBP2x remaining in the outer ring. In iHT-protein fusion experiments, the blue FDAA HADA replaced TADA due to spectral overlap of the red HT-JF549 ligand with TADA. Nodes of iHT-bPBP2x (Fig. S2C and S2E) were often less distinct and regular than those of intrinsically fluorescent isfGFP-bPBP2x (Fig. 2A and S2A), likely because of non-optimal labeling with the extrinsic HT substrate.

By contrast, as isfGFP-bPBP2x progresses to the inner septal ring during division (Fig. 2A and S2A, row 3), sfGFP-bPBP2b expressed at a low level compared to WT bPBP2b remains confined to the outer peripheral ring, with little, if any sfGFP-bPBP2b detected in the inner ring marked by TADA (Fig. 2A and S2B, row 3). Likewise, iHT-bPBP2b expressed near the WT bPBP2b level remains as nodes in the outer peripheral ring of pre- and early divisional cells (Fig. S2D, row 1), but is not detected in the constricting inner septal ring (Fig. S2D, row 2). Again, extrinsically labeled iHT-bPBP2b nodes vary more than those of intrinsically fluorescent sfGFP-bPBP2b. Together, these results identify intermediate states of spatial separation of the bPBP2x-containing sPG and bPBP2b-containing pPG synthesis machines during Spn division (see Fig. 1D). The nodal localization of bPBP2x and bPBP2b in the midcell rings of predivisional cells is taken up below in the context of FDAA labeling patterns.

We extended this approach to a protein that had not been localized before. FtsX is a polytopic membrane protein that forms a complex with the cytoplasmic FtsE ATPase and the extracellular PcsB PG hydrolase in Spn (Alcorlo et al., 2020, Rued et al., 2019, Sham et al., 2011, Sham et al., 2013). The FtsEX:PcsB complex is essential for Spn growth due to its role in hydrolytic remodeling during PG synthesis (Alcorlo et al., 2020, Bartual et al., 2014, Ng et al., 2004). Depletion of FtsEX:PcsB results in the formation of chains of spheroid cells (Ng et al., 2004, Sham et al., 2013), characteristic of defective pPG synthesis (Alcorlo et al., 2020, Tsui et al., 2014). To test this idea, we localized an FtsX’-isfGFP-FtsX’ sandwich fusion in vertically oriented Spn cells (Fig. S2G). This FtsX sandwich fusion is expressed from the normal chromosomal locus and does not cause cell defects, even though about 60% seemed to be proteolytically cleaved once in the extracellular ECL1 domain near the fusion point (Fig. S3C). Throughout the cell cycle, FtsX-isfGFP-FtsX’ appears as nearly evenly spaced nodes around the outer peripheral ring demarked by TADA labeling and was not detected in the inner septal ring later in division (Fig. S2G). Finally, we determined that FtsZ-sfGFP tracks with TP activity at the leading-edge of the closing septum annulus (Fig. S2H). We conclude that PG remodeling by the FtsEX:PcsB complex is likely confined to pPG synthesis in Spn, and that FtsEX and pPG synthesis is not associated with an FtsZ ring during much of the Spn cell cycle (see Fig. 1D). In addition, this example illustrates the utility of using FDAA labeling as a fiducial marker to distinguish proteins involved in sPG and/or pPG synthesis.

2.3 |. New PG synthesis is organized as a series of regularly spaced nodes around the midcell of predivisional Spn cells

At a labeling time of 2.5 min with 125 μM TADA, we observed apparent nodal variation of TADA intensity in the outer and inner rings of dividing Spn cells (Fig. 1 and 2). As noted above, we also observed nodal positioning of bPBP2x, bPBP2b, and FtsX in these rings (Fig. 2 and S2). To study these patterns further, we labeled WT Spn cells with a lower concentration of TADA (45 μM) for a very short pulse (17 s) before imaging vertically oriented cells (Fig. 3). To maximize the spacing between nodes, we confined our analysis of TADA labeling patterns to predivisional cells that have the largest diameters by inspection and that were relatively plentiful in fields of non-synchronized cultures. Very short pulses resolved TADA labeling into a series of regularly spaced nodes distributed around the equators of vertically oriented predivisional Spn cells (Fig. 3A). A similar nodal pattern was also observed in cells from the same fields that were tilted, but not vertically oriented (data not shown).

Figure 3.

PBP TP activity is organized into regular nodes at the midcell of predivisional Spn cells. WT cells (IU1945) were incubated for 17 s with a low concentration of TADA (45.5 μM), fixed, and prepared for vertical cell imaging by 3D-SIM (see Experimental Procedures). (A) Representative images of 12 predivisional Spn cells pulsed-labeled with TADA. Mean ring diameter ± SD of >60 cells from >3 independent biological replicates was determined as described in Experimental Procedures. (B) Distributions of arc distances (top) and number of nodes per ring (bottom) determined for the data set of cells in (A) and compiled in Table S3.

To quantify the pattern of these nodes, we developed a custom vertical image analysis graphical user interface (VIMA-GUI) using MATLAB (see Experimental Procedures). After manually designating individual nodes in each division ring, the program accurately determines the location of the ring, the diameter of the cell, the number of FDAA nodes, and the arc distance between adjacent nodes. TADA pulse-labeled and fixed WT Δcps D39 cells have an average diameter of 0.84 μm ± 0.05 (SD), indicating that the selected cells are at the same predivisional stage (Fig. 3A; Table S3). WT Δcps D39 cells contain an average of 9.8 ± 1.0 (SD) individual TADA-labeled nodes, with an average arc distance between nodes of 0.27 μm ± 0.07 (SD) (Fig. 3B; Table S3). The consistency of these measurements strongly supports the notion of an ordered placement of TP activity sites early in Spn division.

2.4 |. The pattern of FDAA-labeled nodes is not caused by image processing

To confirm that the regular nodal TADA-labeling pattern was not caused by the processing steps required for rendering structured illumination images, we examined the raw data and data-process parameters for TADA-pulse-labeled WT Δcps D39 cells (Fig. 1 and 3). Nodal labeling was still observed at the equators of pulse-labeled predivisional Spn cells using conventional, low-resolution (≈250 nm) widefield 2D-EFM without image deconvolution (arrows, Fig. 4A). Deconvolution improved the clarity of the nodes in these low-resolution widefield images, but the images remained blurry compared to 3D-SIM images (Fig. 3A). To examine the influence of the major user-defined data smoothing function in SIM reconstruction, we processed 3D-SIM data by adjusting the Wiener filter setting stepwise between 0.001–0.02 (Komis et al., 2015). Higher filter settings (above 0.01) resulted in clear “honeycomb” patterning effects on the sample and in background regions. We found no discernable patterning or grouping (e.g., over-separation of intensity into nodes) of signal with settings below 0.006 (Fig. 4B, row 2) and set the Wiener filter at 0.001 for the data presented in this work (Fig. 4B, row 1). Finally, we asked if the level of fluorescence signal in the pulse-labeled cells was potentially contributing to an artificial separation of signal into nodes. To this end, we determined the maximal pixel peak intensities in lines drawn through nodes in cells labeled for 17s (Fig. 3) and through the nearly contiguous labeling rings of cells labeled for 2.5 min (Fig. 1). Background offsets were determined from regions lacking cells for the two labeling conditions, and these backgrounds were subtracted from the mean maximal pixel intensities. The mean maximal pixel intensity of nodes (425 A. U. ± 105) was similar to that of nearly contiguous rings (461 A. U. ± 169), indicating that the SIM processing can produce nearly contiguous rings and separate nodes at about the same intensity level. This equivalence argues that the observed nodes are not simply being created by grouping together regions of low fluorescence signal. We conclude that the 3D-SIM processing parameters used in these analyses did not create the TADA nodal patterns observed in vertical Spn cells.

Figure 4.

TADA-labeled nodes at the midcell of vertically oriented predivisional Spn cells are observed by widefield microscopy and at different 3D-SIM filter settings. WT cells (IU1945) were labeled with 45.5 μM TADA for 17 s, fixed, and prepared for vertical cell imaging as described in Experimental Procedures. (A) Representative widefield-microscopy images of four separate cells, before and after deconvolution. Arrows point to nodes of TADA labeling. (B) 3D-SIM images processed with different Wiener filter settings of the midcell of a single TADA-labeled vertical cell (top row) or a background region lacking a cell (bottom row). Red box, 0.001 Wiener filter setting used throughout this paper for 3D-SIM imaging. Scale bar = 1 μm.

In support of this conclusion, the nodal pattern of TADA labeling becomes irregular in mutants defective in FtsZ ring assembly or in the regulation of PG synthesis (next section). This irregularity is not consistent with an image reconstruction artifact. In one experiment, FtsZ was expressed from a Zn²⁺-inducible promoter in a ΔftsZ//P_Zn-ftsZ⁺ merodiploid strain grown in BHI broth containing sufficient Zn²⁺ to allow growth comparable to the WT strain. The merodiploid strain was shifted to BHI broth lacking Zn⁺² for 2.5 h, which resulted in enlarged spheroid cells with increased (≈1.4-fold) diameters (Fig. 5; Table S3). The FtsZ-depleted and WT FtsZ⁺ strain were pulse-labeled (17 s) with TADA, and vertically oriented cells were imaged by 3D-SIM. WT cells show a regular nodal pattern in ≈94% of vertical cells, whereas the FtsZ-depleted cells with larger diameters show a regular nodal pattern in only ≈40% of cells (Fig. 5A). The remaining ≈60% of FtsZ-depleted cells show irregular nodal arrangements, often with large gaps between nodes (arrow, Fig. 5A). “Regular” and “irregular” nodal patterns were initially distinguished by visual heuristic criteria, where “regular” refers to nodal patterns that appear by eye to be relatively evenly distributed, whereas “irregular” refers nodal distributions that contain one or more large gaps with spacing estimated to be at least twice that in the regular nodal pattern. Subsequent measurements of arc lengths generally matched the initial visual criteria. FtsZ-depleted cells with regular patterns contained more TADA-labeled nodes than WT or FtsZ-depleted cells with irregular patterns (Fig. 5B, middle; Table S3). The mean distance between nodes was greater for FtsZ-depleted than WT cells, with the variability of gap distances in FtsZ-depleted cells with irregular patterns reflected by a high standard deviation (Fig. 5B, right; Table S3). Taken together, this evidence strongly argues that the FDAA nodes are not an artifact of 3D-SIM imaging and that FtsZ is required to maintain the WT arrangement of FDAA nodes in predivisional Spn cells.

Figure 5.

FtsZ-depleted cells have enlarged midcell diameters, and a majority of cells have irregularly spaced nodes of TP activity. IU8124 (ΔftsZ//P_Zn-ftsZ⁺) cells were grown in BHI broth containing 0.2 mM ZnCl₂ to ectopically express FtsZ for 12 h prior to depletion. IU8124 cells were diluted into BHI broth without added ZnCl₂ and incubated to deplete FtsZ (see Experimental Procedures). After 2.5 h, WT (IU1945) and FtsZ-depleted cells were labeled with 45.5 μM TADA for 17 s, fixed, and prepared for vertical cell imaging by 3D-SIM (Experimental Procedures). (A) Representative 3D-SIM images of nodes of TP activity in midcell rings of predivisional cells. Labeling was classified as regular (Reg) or irregular (Irreg) with gaps of varying sizes between nodes (white arrow) as described in Results. Percentages refer to analysis of 76 WT and 47 FtsZ-depleted cells. Scale bar equals 1 μm. (B) Distributions of midcell ring diameters (left), nodes per ring (middle), and arc distances (right) of WT and FtsZ-depleted cells were determined as described in Experimental Procedures. Ring diameters and nodes per ring were determined for 76 WT cells and for 19 or 28 Reg or Irreg FtsZ-depleted cells, respectively. Arc distance were determined for 746 nodes in WT cells and for 230 or 293 nodes in Reg or Irreg FtsZ-depleted cells, respectively. Graphs show median (interquartile range; whiskers, 5^th-95^th percentile; +, mean). The apparent bimodality in the arc distance distribution is due to a resolution limit on nodes closer together than ≈0.1 μm. Means ± SD are compiled in Table S3. Differences in means relative to WT were determined by one-way ANOVA with Bonferonni’s multiple comparison posttest (GraphPad Prism). ***, P<0.001.

2.5 |. The number of regularly spaced FDAA-labeled nodes is correlated with the diameter of predivisional Spn cells

The increase in the number of nodes in FtsZ-depleted cells with regular spacing (Fig. 5B, middle; Table S3) was suggestive of a regulatory mechanism for node placement. To explore this idea, we determined labeling patterns in five additional division and PG synthesis mutants that were pulsed with TADA for 17 s and imaged vertically (Fig. 6). Amino acid changes in FtsZ(G107S) (divisome ring organizer protein) (Perez et al., 2019), EzrA(T506I) (FtsZ-ring modulator in Gram-positive bacteria) (Land et al., 2014), and GpsB(K96N) (regulator of the balance between sPG and pPG synthesis) (Cleverley et al., 2019, Rued et al., 2017) cause temperature sensitivity (TS) at 42°C. TS mutants expressing EzrA(T506I) or GpsB(K96N) were isolated for this study (Table S1). At the semi-permissive temperature of 37°C in BHI broth, each mutant grows slower and forms cells with enlarged diameters (Fig. S4A; Table S3). The TADA-labeling pattern of each mutant was similar to that of the FtsZ-depleted strain. The majority (53% to 66%) of mutant cells show irregular nodal patterns with gaps and greater arc distances, whereas cells with regular labeling patterns contain an increased number of nodes spaced similarly to WT (Fig. 6A, 6B, and S4B; Table S3).

Figure 6.

Arc distance between regularly spaced midcell TP nodes is constant in predivisional Spn mutants with decreased or increased diameters compared to WT. Mutant strains of S. pneumoniae with altered midcell diameters were labeled with 45.5 μM TADA for 17 s, fixed, and prepared for vertical cell imaging by 3D-SIM as described in Experimental Procedures. Strains used were: Δpbp1a (K164), ΔkhpA (E751), WT (IU1945), ftsZ(G107S) (IU10612), ezrA(T506I) (IU11034), and gpsB(K96N) (IU11956) (Table S1). (A) Representative 3D-SIM images of TP activity in midcell rings of predivisional cells. Percentages indicate the frequency of regularly spaced nodes (Reg) or irregularly spaced nodes (Irreg) with gaps (see white arrows) for the number of cells analyzed of each strain. Irregularly spaced nodes were observed in the majority of ftsZ(G107S), ezrA(T506I) and gpsB(K96N) cells, whose diameters were greater than that of WT (see Fig. S4 and Table S3). (B) Distributions of arc distances between TP nodes of WT and mutant strains were determined as described in Experimental Procedures. Graphs show median (interquartile range; whiskers, 5^th-95^th percentile; +, mean). The apparent bimodality in the arc distance distribution is due to a resolution limit on nodes closer together than ≈0.1 μm. Means ± SD are compiled in Table S3. Cells with irregular midcell nodal patterns exhibit large SDs. Differences in means relative to WT were determined by one-way ANOVA with Bonferonni’s multiple comparison posttest (GraphPad Prism). ***, P<0.001. (C) Linear relationship between ring diameter (X-axis) versus nodes per ring (Y-axis) for WT and mutant strains with regular nodal patterns. Colored crosses represent measurements from single rings determined as described in Experimental Proceduresand shown in Fig. S4. Line of best fit and r² values for the combined data set were determined using GraphPad Prism.

We extended this analysis to Δpbp1a and ΔkhpA mutants that have significantly smaller diameters than WT cells in BHI broth (Fig. S4A, Table S3). Δpbp1a mutants lack aPBP1a and produce skinny, slightly elongated cells (Land & Winkler, 2011, Tsui et al., 2016), while ΔkhpA mutants lack a regulatory RNA-binding protein and form smaller cells with the same aspect ratio as larger WT cells (Zheng et al., 2017). The nodal pattern in the Δpbp1a or ΔkhpA mutant is regular in >90% of cells with an average arc distance similar to that of WT (Fig 6A and 6B; Table S3); consequently, the mutant cells have fewer nodes per cell (Fig. S4B). When plotted, these data show that the number of regularly spaced nodes, when present, increases linearly with ring diameter (Fig. 6C), consistent with a mechanism that maintains WT spacing between sites of PG synthesis in predivisional Spn cells.

2.6 |. Spacing of bPBP2x and bPBP2b nodes is similar to that of FDAA nodes produced by short pulse labeling

The number and arc distance of nodes of isfGFP-bPBP2x and sfGFP-bPBP2b in the midcell rings of predivisional cells (Fig. 2A, S2A, S2B, and 7A) were measured and found to be similar to those of the FDAA nodes in WT cells (Fig. 2B and 7C; Table S3). Notably, the spacing of the TADA (red) and sfGFP-bPBP2b (green) nodes determined in the same cell is similar, but displaced in most cells (Fig. 7A), providing further support that the nodal patterns are not a microscopy artifact (i.e., the nodal patterns are not imaging channel or probe specific and are observed at the slightly different native resolution of the two fluorescence emission colors). To quantitate the relative distributions of these patterns, we calculated correlation coefficients for the number and spatial distribution of TADA and sfGFP-bPBP2b nodes in 48 cells (Fig. S5A and S5B). As expected by the number of nodes (Table S3), there is a positive correlation between the number of sfGFP-bPBP2b and TADA nodes (Fig. S5A). In contrast, spatial correlation coefficients were distributed around 0, indicating no strong interdependence of the positions of the TADA and sfGFP-bPBP2b nodes (Fig. S5B).

Figure 7.

sfGFP-bPBP2b and Bocillin-FL (Boc-FL) labeling are organized in nodal patterns at the midcell of predivisional Spn cells. Cells expressing sfGFP-bPBP2b (IU9965) were labeled with 45.5 μM of TADA for 17 s, fixed, and prepared for vertical cell imaging by 3D-SIM as described in Experimental Procedures. WT (IU1945) cells were pulse-labeled with TADA for 17 s, followed by labeling with 2 μg/mL of Boc-FL for 21 s, fixed, and prepared for imaging (see Experimental Procedures). (A) Representative images of IU9965 localizing sfGFP-bPBP2b and TADA as nodes. Each row is a separate cell. (B) Representative images of WT cells localizing Boc-FL and TADA as nodes. Each row is a separate cell. (C) Distributions of ring diameters (left), nodes per ring (middle), and arc distances (right) for WT cells labeled with TADA alone, IU9965 (sfGFP-bPBP2b) labeled with TADA, IU9965 (sfGFP-bPBP2b) not labeled with TADA, WT labeled with TADA followed by Boc-FL, and WT labeled with Boc-FL alone. Measurements were performed as described in Experimental Procedures. Graphs show median (interquartile range; whiskers, 5^th-95^th percentile; +, mean). The apparent bimodality in the arc distance distribution is due to a resolution limit on nodes closer together than ≈0.1 μm. Means ± SD are compiled in Table S3. Differences in means relative to WT labeled with TADA were determined by one-way ANOVA with Bonferonni’s multiple comparison posttest (GraphPad Prism). *, P<0.05; ***, P<0.001. Correlation coefficient analyses are presented in Fig. S5 and Results.

We performed two additional labeling protocols to gain information about these patterns. First, we performed tandem short-pulse labeling of WT cells with TADA (17 s), which labels regions of active TP activity, followed by Boc-FL (21 s), which labels all active PBPs (Fig. 7B; Table S3). Displaced two-color patterns of TADA and Boc-FL nodes were produced, similar to those of TADA-labeled sfGFP-bPBP2b cells (Fig. 7A). Again, there was a positive correlation between the number of TADA and Boc-FL nodes (Fig. S5C), but no spatial correlation of the positions of the two kinds of nodes in 18 cells (Fig. S5D). The regular nodal pattern after Boc-FL pulse-labeling pattern did not match what would be expected from an earlier study done at long labeling times (Kocaoglu et al., 2012). We reprised this earlier study and obtained different results (Supplemental Data; Fig. S6), consistent with those reported here (Fig. 7B). Last, we tandemly pulse labeled WT cells with three different colors of FDAAs (green, then red, then blue) (Fig. S7). We again observed regular, displaced nodal patterns for each color of FDAA probe (arrows, Fig. S7), supporting the conclusion that there is a distributive, regular nodal pattern of PBP localization and TP activity at the midcell ring of predivisional Spn cells.

2.7 |. FDAA labeling is organized in nodes in other ovoid-shaped bacterial species

To determine if this regular, organized pattern of TP activity by PBPs is present in other ovococcal species, we labeled Streptococcus mitis and Enterococcus faecalis with short (17 s) pulses of TADA and imaged vertically oriented cells by 3D-SIM (Fig. 8). S. mitis is a close evolutionary relative that exchanges DNA with Spn by natural competence, while E. faecalis is more evolutionarily distant in the same Lactobacillale order (Gao et al., 2014, Ludwig et al., 1985, Massidda et al., 2013). Both species demonstrated a pattern of TADA-labeled nodes at midcell with an average arc distance indistinguishable from that of WT Spn (Fig. 8B; Table S3). Thus, the regular organized pattern of PBP TP activity in predivisional cells is widely distributed in ovococcal species.

Figure 8.

The nodal pattern of regularly spaced TP activity at the midcell of predivisional cells is conserved in other ovoid-shaped bacterial species. S. pneumoniae (Spn, IU1945), S. mitis (Smi, ATCC 49456) and E. faecalis (Efa, ATCC 51299) cells were labeled with 45.5 μM TADA for 17 s, fixed, and prepared for vertical cell imaging by 3D-SIM (see Experimental Procedures). (A) Representative images of two predivisional cells pulse-labeled with TADA are shown for Spn (left), S. mitis (middle), and E. faecalis (right) from the total number of cells analyzed (bottom). (B) Distributions of ring diameters (left), nodes per ring (middle), and arc distances (right) of ovoid-shaped cells. Measurements were performed as described in Experimental Procedures. Graphs show median (interquartile range; whiskers, 5^th-95^th percentile; +, mean). The apparent bimodality in the arc distance distribution is due to a resolution limit on nodes closer together than ≈0.1 μm. Means ± SD are compiled in Table S3. Differences in means relative to Spn were determined by one-way ANOVA with Bonferonni’s multiple comparison posttest (GraphPad Prism).**, P<0.01 and ***, P<0.001.

3 |. DISCUSSION

Whether the sPG and pPG synthesis machines of Spn are distributed contiguously in a single midcell ring (“plum pudding” model) or separate from each other during septal constriction has been a point of contention (Fleurie et al., 2014, Tsui et al., 2014, Straume et al., 2016, David et al., 2018). The demonstration of concentric intermediate rings of TP activity reported here provides strong support for the separation hypothesis (Fig. 1D and 9A). Localization of bPBP2x to the inner ring, which corresponds to the constricting leading edge of the septal annulus, further supports closure driven by PG synthesis by the bPBP2x:FtsW complex (Perez et al., 2019, Taguchi et al., 2019). Some bPBP2x remains in the outer ring of mid-to-late divisional cells when iHT-bPBP2x is expressed at near WT level in C+Y medium (Fig. S2E) or when WT bPBP2x is itself labeled with the 7FL probe (Fig. S2F) (Sharifzadeh et al., 2017). Thus, active bPBP2x is both at the constricting leading edge of the septal annulus and at its outer edge, suggesting expansion of the annulus in both places. The constricting FtsZ ring also tracks with sPG synthesis at the leading edge of the septal annulus ring, implying that FtsZ is not detectable in the outer pPG synthesis ring (Fig. S2H).

Figure 9.

Summary diagram of the localization of PG synthesis in Spn based on results in this paper. (A) Separation of sPG and pPG (elongasome) machines at the midcell of dividing Spn cells. Cell division begins at the midcell equator of newly divided predivisional cells in FtsZ-organized divisome rings containing the components of the sPG and pPG synthesis machines. The sPG machine contains bPBP2x TP, its partner FtsW GT, and other components, while the pPG synthesis machine is made up of bPBP2b TP and its partner RodA GT, other “Rod” complex proteins, and the FtsEX:PcsB remodeling PG hydrolase. Some sPG and pPG synthesis proteins are arranged as regularly spaced nodes in predivisional cells (see B, below), but the relative spatial arrangements of nodes of different proteins have not yet been determined. Early in cell division, the septal annulus forms and begins to close. The leading edge of the closing septal annulus separates the sPG machine from the pPG synthesis machine that remains at the outer edge of the annulus and elongates the PG outward from the midcell. The constricting FtsZ ring tracks with the leading edge of the septal annulus, such that the outer peripheral ring is not organized by FtsZ beyond the predivisional stage. Later in division, FtsZ remaining at the septum migrates to the developing equatorial rings in daughter cells. The inner sPG synthesis machine constricts into a dot surrounded by the closing outer pPG synthesis ring, which eventually constricts and merges with the sPG dot as the new cell pole is completed and the PG synthesis enzymes migrate to the new equatorial rings in daughter cells (see Fig. 1D also). (B) In predivisional cells, TP activity, sPG synthesis complexes, and pPG complexes are organized into a pattern of regularly spaced nodes (red circles). The placement of PBPs and TP activity is not fixed at single positions on equators of early divisional cells, but rather, is dynamic and distributive, likely driven by PG synthesis itself (Perez et al., 2019). Aggregate movements of these heterogenous nodal complexes with time would account for the distributive, non-correlated nodal patterns observed in two-color labeling experiments. A constant distance (≈0.27 μm) is maintained between nodes of PBP TP activity and between the PBPs themselves. Midcell rings of predivisional cells with smaller or larger diameters contain fewer or more nodes, respectively, than WT (≈ 10 nodes).

In contrast, bPBP2b expressed at 12% or at WT levels primarily localizes to the outer pPG synthesis ring (Fig. 2A, S2B, S2D, and S3C). These results indicate that only about 12% of bPBP2b cellular amount is sufficient for normal growth and Spn cell morphology (Fig. S3A and S3B), implying that bPBP2b is in excess in cells in this culture condition. An implication of confinement of bPBP2b to the outer midcell ring is that the presumed circumferential movement of the bPBP2b:RodA complex during pPG synthesis is guided by a structure that lacks FtsZ filaments/bundles. Spn does not encode the actin-like MreB protein that mediates PG elongation of rod-shaped bacteria (4), and determination of Spn elongasome composition and organization is an area of active research.

FtsX is also confined to the outer pPG synthesis ring (Fig. S2G). FtsX is an essential, polytopic membrane protein that forms a complex with the cytoplasmic FtsE ATPase and the extracellular PcsB PG hydrolase in Spn (Alcorlo et al., 2020, Bartual et al., 2014, Rued et al., 2019, Sham et al., 2011, Sham et al., 2013). Depletion of essential FtsX, FtsE, or PcsB results in chains of spherical cells, similar to those caused by depletion of bPBP2b (Alcorlo et al., 2020, Berg et al., 2013, Ng et al., 2004, Sham et al., 2011, Sham et al., 2013, Tsui et al., 2014). An FtsX’-isfGFP-FtsX’ fusion was detected in the outer PG synthesis ring, but not in the inner septal ring (Fig. S2G), consistent with a role for FtsEX:PcsB in pPG synthesis, analogous to that played by FtsEX:CwlO in sidewall PG synthesis in Bacillus subtilis (Bsu) (Brunet et al., 2019, Meisner et al., 2013). Thus, Spn FtsX is in proximity to FtsZ and FtsA in the nascent divisome in early predivisional cells (Fig. S2G); however, in contrast to Escherichia coli (Eco) FtsX (Pichoff et al., 2019), as division proceeds, Spn FtsX physically separates from FtsZ and FtsA, which are at the leading edge of the septal annulus (Fig. S2H). This result raises the possibility that a PG remodeling hydrolase other than FtsEX:PcsB mediates sPG synthesis. Based on these examples, dual labeling of vertical cells can be applied to assign other PG synthesis, divisome, and regulatory proteins to the sPG or pPG synthesis machines of Spn. This approach may also explicate PG stress responses, such as possible roles of Class A PBPs in imparting resistance to an exogenously added PG hydrolase following exposure of Spn laboratory strain R6 to a β-lactam antibiotic that inhibits bPBP2x (and DacA (PBP3)) (Straume et al., 2020).

We noted that bPBP2x and bPBP2b are distributed in regular nodal patterns at the equators of predivisional cells (Fig. 2, 7A, and S2 A-F). As the inner ring constricts, this nodal pattern becomes more compact and difficult to resolve (Fig. 2A); hence, we confined this study to the large equators of predivisional cells (Fig. 3A). By shortening the FDAA pulse time down from 2.5 min to ≈17 s, we observed that TP activity also is distributed in a regular nodal pattern with ≈10 nodes separated by an average arc length of 0.27 ± 0.07 μm in WT cells (Fig. 3B). A comparable nodal pattern of FDAA pulse labeling was detected in S. mitis and E. faecalis (Fig. 8), indicating a common organization of TP activity in predivisional ovococcal cells.

Several controls support the conclusion that the nodal pattern was not caused by processing steps required to generate structured illumination images (see Results). Most importantly, the regular nodal pattern of FDAA labeling was disrupted by gaps in a majority of cells depleted for FtsZ (Fig. 5) or in temperature-sensitive ftsZ(G107S), ezrA(T506I), and gpsB(K96N) mutants grown at the semi-permissive temperature of 37°C (Fig. 6). Each of these growing TS mutants formed enlarged cells with increased diameters compared to WT (Fig. S4). Between 34%−47% of these larger mutant cells still showed regular nodal patterns of FDAA labeling with more nodes per ring spaced at the WT arc distance (Fig. 6). Conversely, Δpbp1a or ΔkhpA mutants have smaller diameters than WT with fewer FDAA nodes spaced at the WT arc distance (Fig. 6). Together, these data are consistent with a mechanism that maintains a regular spacing of PG synthesis in predivisional cells, irrespective of equator diameter (Fig. 6C and 9B).

The average arc distance was similar between nodes of FDAA labeling, isfGFP-bPBP2x, isfGFP-bPBP2b, and Boc-FL (Fig. 2, 3, and 7). However, the positioning of nodes in two-color FDAA/sfGFP-bPBP2b or FDAA/Boc-FL experiments showed low correlation (Fig. S5), as did nodes sequentially pulse-labeled with three different colors of FDAAs (Fig. S7). The displacement of FDAA nodes relative to bPBP2b or Boc-FL, may reflect the time (≈1 min) that it takes to wash away unincorporated FDAA. It is also possible that PBPs other than bPBP2b are active in predivisional cells. Altogether, these results indicate that the placement of PBPs and TP activity is not fixed at single positions on equators of predivisional cells, but rather, is dynamic and distributive.

The localization of PG synthesis and PBPs in early divisional cells varies among bacterial species that use different modes of septum formation and division. Similar to the patterns reported here for Spn and other ovococcal species (Fig. 2, 8, and S2), labeling Eco for short pulses, but not for long times, with an FDAA results in nodal (also called punctate) patterns, whose patterns were not determined or quantitated at high resolution (Yang et al., 2017). Consistent with nodal pattern formation, high-resolution microscopy of vertically oriented Eco cells revealed a pattern of “discrete densities” of mCit-FtsI (bPBP3) distributed around the entire septal ring of dividing Eco cells (Söderström et al., 2019).

Different FDAA pulse-labeling patterns were reported for S. aureus (Sau) and Bsu compared to those in Spn and Eco. Discrete foci of PG synthesis were not observed in Sau cells labeled with short FDAA pulses and viewed by 3D-SIM, in an experiment similar to Figure 3, or labeled with a pulse of other D-amino acid probes and viewed by localization microscopy (Lund et al., 2018). However, while not commented upon, SIM images of sfGFP-PBP1 in some vertically oriented Sau cells do appear nodal around closing septal rings (Monteiro et al., 2018), and heterogenous localization of Sau GFP-PBP2 has also been reported (Strauss et al., 2012). Labeling Bsu with short, sequential pulses of two colors of FDAAs or an FDAA followed by Boc-FL (Bisson-Filho et al., 2017) gave a different pattern from the regular nodal pattern in Spn (Fig. 7 and S7). At the lower resolution of rotated 3D-SIM images, sequential labeling with two colors of FDAAs gives a pattern suggestive of a limited number of PG synthesis complexes moving in both directions around the septum. Labeling Bsu with an FDAA followed by Boc-FL again suggests a surprisingly limited number of active PBP complexes labeled by Boc-FL adjacent to newly synthesized PG marked by the FDAA (Bisson-Filho et al., 2017). We conclude that nodal PG synthesis at septa occurs in different patterns in some bacteria and is apparently absent in others, possibly reflecting different mechanisms of sPG synthesis and cell separation. In this regard, Spn and Eco simultaneously close division septa while separating daughter cells (Perez et al., 2019, Yang et al., 2017), whereas septa formation by PG synthesis occurs before cell separation in Sau and Bsu (Bisson-Filho et al., 2017, Lund et al., 2018, Monteiro et al., 2018).

The mechanism that causes the regular nodal distribution of PBPs and their TP activity in predivisional cells of Spn remains to be determined (Fig. 9B). Recent notable studies have demonstrated heterogeneous supramolecular localization of membrane proteins generally and at septa of Sau cells, including interacting enzymes that catalyze phospholipid biosynthesis and the MreD regulator of PG synthesis (Garcia-Lara et al., 2015, Weihs et al., 2018). This heterogenous localization results in punctate patterns of these proteins at division septa, resembling the localization of Spn PBPs reported here (Fig. 2 and S2). Polytopic MreD plays a role in supramolecular localization of the phospholipid biosynthesis enzymes, and heterogeneous punctate patterns are lost upon protein overexpression (Weihs et al., 2018). A favored explanation for punctate supramolecular organization is that membrane protein complexes distort local curvature on membranes and thereby perturb diffusion, resulting in distribution patterns for the majority of membrane proteins (Weihs et al., 2018). Whether membrane protein distribution, the size and composition of PG synthesis complexes, an unknown scaffolding complex, or a combination of these mechanisms causes the nodal distribution of PBPs and TP activity in predivisional Spn cells remains to be determined.

Finally, single molecules of bPBP2x and its interacting partner FtsW move circumferentially around the septa of dividing pneumococcal cells in either direction at ≈21 ± 8 (SD) nm/s (Perez et al., 2019). Based on data in that paper, runs extend for ≈28 ± 15 (SD) s (n = 106 cells). Multiplication of these two numbers gives a distance of 588 ± 316 nm, which because of the large propagated error could accommodate the ≈270 nm arc lengths reported here (Fig. 3B). In contrast, bPBP2x and FtsW not bound to septa move diffusively throughout the cell membrane (McCausland et al., 2020, Perez et al., 2019). In Spn, movement of the bPBP2x:FtsW complex at septa is driven by PG synthesis itself and not by treadmilling movement of FtsZ filaments/bundles (Perez et al., 2019). In the static images here, we do not precisely know how many molecules form the bPBP and FDAA-labeled nodes; but, a reasonable assumption is that each node consists of multiple PBPs and regions of PG labeling. In addition, lower levels of PBPs could be present between the more concentrated nodes. With this in mind, circumferentially moving bPBP2x:FtsW complexes likely move linearly in both directions across nodal regions driven by the PG synthesis echoed by FDAA labeling. Aggregate movements of these heterogenous nodal complexes with time would then account for the distributive, non-correlated labeling patterns in two-color pulse labeling experiments (Fig. 7B and S7). This and other hypotheses about the composition, organization, and dynamic movement of these nodal PG synthesis complexes in Spn and other ovococcal bacteria await future testing.

4 |. EXPERIMENTAL PROCEDURES

4.1 |. Bacterial strains and growth

S. pneumoniae (Spn) strains used in this study were constructed in strains IU1824 and IU1945, which are unencapsulated (Δcps) derivatives of serotype 2 strain D39W (Table S1). Strains were constructed by transforming linear DNA amplicons into competent cells and culturing transformants on trypticase soy agar II plates (BD BBL, 221261) containing 5% (v/v) defibrinated sheep blood as described previously (Tsui et al., 2014), except when noted (see Table S1 footnotes). Primers used for the synthesis of amplicons are listed in Table S2.

Strains were inoculated from frozen glycerol stocks into 5 mL of brain heart infusion (BHI; BD Bacto, 237500) broth containing added ZnCl₂ and MnSO₄ where indicated. Serial dilutions were prepared and incubated for 12–16 h as overnight cultures at 37°C in an atmosphere of 5% CO₂. Extent of growth of liquid cultures was determined by OD₆₂₀ as described previously (Land et al., 2013). Overnight cultures still in exponential phase (OD₆₂₀ ≈0.05–0.4) were diluted in parallel to OD₆₂₀ ≈0.003–0.005 in 5 mL of fresh BHI broth, containing added ZnCl₂ and MnSO₄ where indicated. Cultures were incubated at 37°C in 5% CO₂, and OD₆₂₀ readings were taken every 45–60 min until cultures reached stationary phase.

4.2 |. 2D-epifluorescence microscopy (2D-EFM) and phase-contrast microscopy

Spn cells were prepared for microscopy as described below in sections describing labeling. Cells were viewed on a Nikon Eclipse E-400 epifluorescence phase-contrast microscope using a Nikon Intensilight C-HGFI epifluorescence illuminator. Images were captured using a CoolSNAP HQ2 charge-coupled device (CCD) camera (Photometrics) and processed with Nikon NIS-Elements BR imaging software. sfGFP and isfGFP images were collected using the FITC-HYQ filter (Ex 460–500 nm; Em 560 nm) with a 1 s exposure time. Halo-Tag (HT) images were collected using the Texas Red-HYQ filter (Ex 532–587 nm; Em 650 nm) with a 1 s exposure time. Additional image processing was completed using FIJI (Schindelin et al., 2012).

4.3 |. Fluorescent D-amino acids (FDAAs) used

FDAAs 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 broth were diluted from 500 mM stocks in DMSO, which were stored at −20°C and shielded from light.

4.4 |. Localization of HADA (≈100 min labeling), then TADA (2.5 min labeling) in vertically oriented Spn cells

Cells from overnight cultures were diluted to OD₆₂₀ ≈0.02 in 2 mL of fresh BHI broth containing HADA (125 μM final). At OD₆₂₀ ≈0.22, 575 μL of cultures was centrifuged at 16,100 × g for 5 min at room temperature, and pellets were resuspended in 250 μL of BHI broth containing TADA (125 μM final). Cells were incubated at 37°C for 2.5 min, chilled on dry ice for 20 s, and centrifuged for 2.5 min at 16,100 × g at 4°C. Cell pellets were centrifuged and washed twice with 500 μL ice-cold PBS, then centrifuged a third time and resuspended in 500 μL of 4% (v/v) paraformaldehyde and incubated in the dark for 15 min at room temperature, followed by 45 min on ice in the dark.

Fixed cells were centrifuged, and pellets were resuspended in 100 μL of ice-cold GTE buffer (50 mM glucose, 20 mM Tris-HCl, pH 7.5, 1 mM EDTA). For imaging, cells were centrifuged for 5 min at 16,100 × g at 4°C to remove the GTE buffer and centrifuged once more to remove residual GTE buffer with a P20 pipette. Excess liquid was allowed to evaporate for approximately 1 min before pellets were resuspended in 3.0 μL of Vectashield Hardset Antifade (Vector Laboratories, H-1400) with vortexing. 1.2 μL of resuspended cells was pipetted onto a 12 mm/1.5 round coverslip (EMS, 72230–01), and a microscope slide was carefully placed on top. The slide was incubated coverslip side down at room temperature in the dark for 15 min, and then viewed by 3-Dimensional Structured Illumination Microscopy (3D-SIM). Exposure time and %T settings to acquire images were 5 ms and 50% for HADA and for TADA.

4.5 |. Quantitative western blotting of relative cellular amounts of fusion proteins

The amounts of sfGFP- and HT-tagged fusion proteins present relative to WT untagged protein were determined by quantitative western blotting with antibodies to native proteins. Cells from overnight cultures were diluted to OD₆₂₀ ≈0.005 in 5 mL of fresh BHI broth. At OD₆₂₀ ≈ 0.15–0.2, 2 mL of culture was collected and cells were centrifuged for 5 min at 16,100 × g at 4°C, washed in 2 mL of ice-cold PBS, centrifuged, and the supernatant removed. Cell pellets were placed on dry ice for 15 min, and thawed at room temperature for 5 min. Pellets were resuspended in 80 μL SEDS lysis buffer (0.1% (v/v) deoxycholate, 150 mM NaCl, 0.2% (v/v) SDS, 15 mM EDTA, pH 8.0), vortexed vigorously, and incubated at 37°C with shaking at 300 rpm for 15 min, with brief vortex mixing every 5 min. Protein concentrations of cell lysates were determined using the DC protein assay kit (Bio-Rad, 5000116) with a standard curve of 0.1–1.3 mg/mL of BSA as described by the manufacturer’s instructions. Samples were diluted 1:1 with 2x Laemmli sample buffer (Bio-Rad, 161–0737) containing 5% (v/v) β-mercaptoethanol and incubated at 95°C for 10 min. Cell lysates were loaded onto a 10% SDS-PAGE gel (Bio-Rad, 456–1035), run for 1 h at 150 V and transferred onto a nitrocellulose membrane for 1.5 h at 350 mA. Blots were blocked in 6 mL PBS-Tw (PBS with 0.1% (w/v) Tween 20 (Sigma, P1379–500ML)) containing 50 mg/mL skim milk powder (BD Difco, 232100) for 20 min, then rinsed briefly twice in 5 mL PBS-Tw. Blots were incubated for 1 h in 10 mL PBS-Tw with either rabbit anti-bPBP2b or rabbit anti-bPBP2x antibodies (1:10,000 dilution). Blots were rinsed briefly twice in 5 mL PBS-Tw, then washed in 5 mL PBS-Tw for 15 min. Blots were incubated for 1 h in 10 mL PBS-Tw with ECL anti-rabbit IgG horseradish peroxidase-linked whole antibody (1:10,000 dilution; GE Healthcare, NA93AV), then successively washed for 5, 5, 15, 5, and 5 min in fresh PBS-Tw. Blots were incubated for 1 min with Amersham ECL western blotting detection reagent (GE Healthcare, RPN2106), and imaged with an IVIS imaging system (Xenogen) using a 1 min exposure time, as described before (Wayne et al., 2010).

For experiments to determine FtsX′-isfGFP-FtsX′ levels, the following changes were made, as described in (Rued et al., 2019). Briefly, cells from overnight cultures were diluted to OD₆₂₀ ≈0.002 in 25 mL of fresh BHI broth. At OD₆₂₀ ≈ 0.15–0.2, cultures were centrifuged for 10 min at 8,000 × g at 4°C, and pellets were resuspended in 1.5 mL ice-cold buffer containing 20 mM potassium phosphate, pH 7.5 and 140 mM NaCl. Samples were transferred to 2 mL Lysing Matrix B Fast Prep tubes (MP Biomedicals 116911050) and lysed at 4°C in a FastPrep-24 (MP Biomedicals) using 3 runs of 40 s at 6.0 m/s in a 24×2 adaptor. Samples were placed on ice, then centrifuged for 1 min at 14,000 × g at 4°C. 1.2 mL supernatant and 1 mL buffer were placed in a 3.2 mL polypropylene ultracentrifuge tube (Beckman Coulter, 362333), and tubes were balanced with buffer and centrifuged for 45 min at 100,000 × g at 4°C in a Beckman Coulter Optima Max-XP ultracentrifuge, using a TLA-100 fixed-angle rotor. Pellets were resuspended in 250 μL of 20 mM potassium phosphate, pH 7.5, 140 mM NaCl, and 0.04% (w/v) n-dodecyl β-D-maltoside (Pierce, 89903). Protein assays, gel electrophoresis, transfer to nitrocellulose membranes, and blocking procedures were performed as described above. Blots were incubated for 1 h in 10 mL PBS-Tw containing 100 mg BSA with polyclonal rabbit anti-FtsX (1:500 dilution) (Rued et al., 2019). Secondary antibody labeling, chemiluminescent detection, and imaging were performed as described above.

To construct a standard curve, western blots of three or four different amounts of the same extract sample (ranging from 1.25 to 12.0 μg of total protein) were performed as described above, and the chemiluminescent signal of protein bands was plotted against the amount of protein loaded. Linear regression analysis (GraphPad Prism) was used to determine the linear signal range for each antibody used. For samples determined to be in the linear range, relative protein amounts were determined by dividing each signal by the signal of the untagged WT protein (≡1).

4.6 |. Localization of TADA (2.5 min label) and sfGFP-tagged proteins in vertically oriented Spn cells

For strains IU1824 (WT), IU11157 (isfgfp-pbp2x), IU9965 (sfgfp-pbp2b), and IU16121 (ftsX′-isfgfp-ftsX′), cells from overnight cultures were diluted to OD₆₂₀ ≈0.003 in 5 mL of fresh BHI broth. At OD₆₂₀ ≈0.23, 600 μL of culture was transferred to a 0.22 μm microcentrifuge tube filter (Corning Costar Spin-X, CLS8160) and the media was removed by applying vacuum to the filter with a vacuum pump. The filter was incubated with 250 μL of BHI broth at 37°C for 3 s, which was filtered away quickly with vacuum. 500 μL PBS at 37°C was added to the filter and immediately filtered away. The filter was removed from the vacuum pump and 250 μL of BHI broth containing TADA (125 μM final) was pipetted up and down approximately 10 times to resuspend the cells, which were incubated at 37°C in the dark for 2.5 m. After incubation, the filter was reattached to the vacuum pump and the media was filtered away. 500 μL of PBS at room temperature was added to the filter and immediately filtered away. This wash step was repeated, then the filter was removed from the vacuum pump. 600 μL of 4% (v/v) paraformaldehyde was added to the filter and mixed by pipetting up and down approximately 10 times to resuspend the cells, which were transferred to a 2 mL tube (Eppendorf 022431102) and incubated in the dark for 15 min at room temperature, then 45 min on ice in the dark. After fixing, cells were centrifuged for 5 min at 16,100 × g at 4°C and washed twice with 600 μL ice-cold PBS. Cells were then washed with 200 μL of ice-cold GTE buffer, resuspended in Hardset Antifade, mounted on slides, and viewed by 3D-SIM as described in Section 4.4. Exposure times and %T settings were 5 ms and 50% for TADA and for sfGFP.

4.7 |. Localization of HADA (2.5 min label) and HT-tagged proteins in vertically oriented Spn cells

For strains IU1824 (WT), IU14927 (iht-pbp2x), IU15928 (iht-pbp2b) and IU16553 (iht-pbp2b//P_Zn-iht-pbp2b), cells from overnight cultures were diluted to OD₆₂₀ ≈0.003 in 5 mL of fresh BHI broth (for IU16553, BHI broth was supplemented with 0.3 mM ZnCl₂ and 0.03 mM MnSO₄ throughout). At OD₆₂₀ ≈0.07, JF549 Halo-Tag ligand (400 μM stock in DMSO) was added to 600 μL culture to a final concentration of 500 nM. Cultures were incubated at 37°C in 5% CO₂ in the dark for 30 m, with brief vortex mixing every 10 min. Cells were centrifuged for 2.5 min at 16,100 × g at room temperature and washed twice with 1 mL of 37°C BHI broth, then resuspended in 600 μL BHI broth and incubated at 37°C in 5% CO₂ in the dark for 20 min. After incubation, cells were labeled with HADA (400 μM final), fixed, and prepared for vertical cell imaging as described in Section 4.6. Exposure times and %T settings were 5 ms and 40% for HADA and 5 ms and 50–100% for HT.

4.8 |. Localization of bPBP2x with 2R,3S-β-lactone-L-Phe-fluorescein (7FL) in WT and Δpbp1b mutant Spn cells

WT (IU1945) and Δpbp1b (E193) strains were labeled with 5 μg/mL of the bPBP2x-specific activity probe 7FL for 20 min in the dark as described in (Sharifzadeh et al., 2017, Sharifzadeh et al., 2020). Tilted, horizontally oriented cells were detected by 3D-SIM as described in Section 4.12.

4.9 |. Localization of FDAA nodes (TADA labeling for 17 s) in vertically oriented Spn cells

For experiments using short TADA pulse labeling to resolve FDAA nodes, cells from overnight cultures were diluted to OD₆₂₀ ≈0.02 in 2 mL of fresh BHI broth. At OD₆₂₀ ≈0.23, 600 μL cultures were labeled with TADA as described in Section 4.6, with the following changes: 250 μL of BHI broth containing TADA (45.5 μM final) was added to the filter and incubated for 3 s and filtered away, resulting in a total labeling time of ≈17 s. Cells were washed twice with 500 μL aliquots of room temperature PBS with filtering between washes. Cells were fixed and prepared for vertical cell imaging as described in Section 4.6. Exposure times and %T settings were 15 ms and 100% for TADA and 5 ms and 50% for sfGFP.

For strain IU8124 (ΔftsZ//P_Zn-ftsZ⁺), the following changes were made to the growth conditions: 0.2 mM ZnCl₂ (no MnSO₄) was added to overnight cultures. Following 12 h of incubation, serially diluted cultures of IU1945 and IU8124 still in exponential phase (OD₆₂₀ ≈0.05–0.4) were centrifuged at 16,100 × g for 5 m at room temperature, resuspended in 1 mL of 4°C PBS, and centrifuged again at room temperature. Pellets were resuspended in 2 mL of BHI broth, diluted to OD₆₂₀ ≈0.005 in 2 mL of BHI broth in a second tube (no ZnCl₂), and incubated at 37°C in 5% CO₂. After a 2.5 h depletion of FtsZ, strains were labeled, fixed, and prepared for imaging as described above.

4.10 |. Localization of TADA (17 s) followed by Boc-FL (21 s) in vertically oriented Spn cells

For short pulses with TADA then Boc-FL, the following changes were made to the labeling procedure in Section 4.9: Following the step in which the BHI broth containing TADA is filtered away, cells were washed with 500 μL of room temperature PBS and filtered. 250 μL of PBS containing 2 μg mL⁻¹ Boc-FL at 37°C was added to the filter, incubated for 7 s, and filtered away (total labeling time ≈21 s). Cells were washed with 500 μL of room temperature PBS, filtered, and resuspended in 600 μL of 4% (v/v) paraformaldehyde. Cells were fixed and prepared for vertical cell imaging as described in Section 4.6. Exposure time and %T setting was 15 ms and 100% for TADA and 6 ms and 50% for Boc-FL.

4.11 |. Localization after sequential labeling with FDAAs BADA (green; 40 s), followed by TADA (red; 40 s), followed by HADA (blue; 40 s) in vertically oriented Spn cells

For experiments with three tandem pulses of different colored FDAAs, cells from overnight cultures were diluted to OD₆₂₀ ≈0.02 in 2 mL of fresh BHI broth. At OD₆₂₀ ≈0.23, 600 μL of culture was centrifuged for 5 min at 16,100 × g at room temperature, and pellets were resuspended in 250 μL of BHI broth containing BADA (62.5 μM final). Cells were incubated for 40 s at 37°C, chilled for 20 s on dry ice, and centrifuged for 2.5 min at 16,100 × g at 4°C. After discarding the supernatant, the labeling, incubation, cooling, and centrifugation steps were repeated twice more with the same cells, first with TADA (62.5 μM final), and last with HADA (62.5 μM final). After discarding the supernatant from the HADA labeling, pellets were washed with 500 μL of ice-cold PBS, centrifuged, and resuspended in 600 μL of 4% (v/v) paraformaldehyde. Cells were fixed and prepared for vertical cell imaging as described in Section 4.6. Exposure time and %T settings were 5 ms and 50% for HADA, for BADA, and for TADA.

4.12 |. Image acquisition by 3D-SIM

A DeltaVision OMX Super Resolution system in the Indiana University Bloomington Light Microscopy Imaging Center (http://www.indiana.edu/~lmic/microscopes/index.html#OMX) was used for 3D-SIM. An Olympus PlanApo N 60X/1.42 oil objective was used, and images were acquired using a PCO.edge 4.2 (CMOS) camera system (Kelheim Germany). Laser lines were 405 nm with emission filters of 419–465 nm (for HADA), 488 nm with emission filters of 500–550 nm (for Boc-FL, sfGFP, and BADA), and 561 nm with emission filters of 609–654 nm (for Ceph-CT, TADA, and HT). For each sample tested, 3D-SIM of horizontally oriented cells was performed to ensure that the refractive index of the immersion oil used matched the refractive index of the sample prepared (1.518 oil for samples prepared in hardset antifade and 1.514 oil for samples prepared with Slowfade gold antifade reagent). To initially focus cells, the 561 nm laser line was used with the lowest laser power and exposure times needed to detect TADA-labeled rings in regions of interest. To acquire images for 3D-SIM, between 9–15 z sections (each 0.125 μm thick) were initially imaged using the laser powers indicated above. Image processing was performed with DeltaVision SoftWoRx Software (GE Healthcare), consisting of OMX Reconstruction (Wiener filter value set at 0.001) and OMX Align. Vertically oriented Spn cells were caught at different stages of the cell cycle. To eliminate TADA and other labeling outside of midcell, all final reconstructed images in this work are cross sections of 2–4 z sections (0.25–0.50 μm total thickness) containing only division planes, unless noted otherwise.

4.13 |. Analysis of FDAA node distribution using a custom Matlab GUI

Reconstructed 3D-SIM image slices 2–4 z sections thick were analyzed using a point-and-click vertical image analysis graphical user interface (VIMA-GUI) organized in MATLAB (The Mathworks). Briefly, several arbitrary points were selected around the division site of a single image, and these points were used to create a marker circle around the division site to determine pixel intensities. Nodes in each image were then manually selected, and these designations were used to reformat the circle into a size and position that overlaid the division site as closely as possible for each individual image. Based on the location of the marked nodes, the program placed a pair of concentric circles with a 5-pixel gap between them, such that the fluorescent division site ring was aligned between the concentric rings. The program then measured the intensity and distribution of fluorescence within the space between the concentric rings. Nodes were selected as individual foci, which could be oblong or round shaped. Oblong foci that demonstrated no partial segmentation were considered as one focus. Oblong patterns that contained distinct multiple foci, even if slightly connected, were counted as separate foci. A subroutine in the VIMA-GUI program was used to calculate the correlation coefficients of the number of nodes and their spatial distribution in dual-labeled cells.

4.14 |. Widefield fluorescence microscopy

A DeltaVision personalDV microscope (Applied Precision) equipped with a CoolSNAP HQ2/HQ2-ICX285 camera was used to perform widefield fluorescence microscopy. An 8 s exposure was taken with 100% T for imaging TADA labeling. Samples were prepared as described in Section 4.9 for 17 s pulse labeling with TADA and visualization of vertically oriented cells. Where indicated, image deconvolution processing was performed with DeltaVision SoftWoRx Software, including adjustments to Wiener filter settings.

4.15 |. Expression and purification of Spn bPBP2x and bPBP2b soluble fragments for polyclonal antibody production

Plasmid construction.

A DNA fragment of pbp2b (spd_1486) encoding amino acids Q40 to N685 of the extracellular domain of bPBP2b was amplified from D39W Δcps genomic DNA using primers: oAT90 (GTACATATGCAGGTTTTGAACAAGGATTTTTACG) and oAT91 (ACTAAGCTTATTCATTGGATGGTATTTTTGATACAGA), where restriction sites are underlined. After digestion with NdeI and HindIII, the PCR amplicon product was ligated into vector pET22b, resulting in plasmid pATPL005 that imparts resistance to ampicillin (Amp^R) and expresses Spn PBP2b^Q40-N685-His₆. Similarly, a DNA fragment of pbp2x (spd_0306) encoding amino acids G49 to D750 of the extracellular domain of bPBP2x was amplified from D39W Δcps genomic DNA using primers: oAT92 (GTACATATGGGGACAGGCACTCGCTTTG) and oAT93 (ACTAAGCTTGTCTCCTAAAGTTAATGTAATTTTTTTAATG), where restriction sites are underlined. After digestion with NdeI and HindIII, the PCR amplicon product was ligated into vector pET22b, resulting in plasmid pATPL025 that imparts Amp^R and expresses Spn PBP2x^G49-D750-His₆.

Expression, protein purification, and antibody production.

Each plasmid was transformed into E. coli BL21(DE3), and the resulting strains were grown in 1 L of LB broth supplemented with 50 mg/L carbenicillin at 37ºC with shaking to OD₆₀₀ ≈0.4. Culture were cooled to 16ºC before inducing protein expression with 500 μM IPTG. Cells were harvested 18 h post-induction by centrifugation (4,200 x g for 15 min at 4ºC). Pelleted cells were resuspended in 30 mL lysis buffer (50 mM HEPES pH 7.5, 500 mM NaCl) supplemented with 1 mM PMSF, and resuspended cells were lysed by three passages through a cell disruptor (Emulsiflex C5, Avestin) at 15,000 psi. Insoluble material was removed by ultracentrifugation (100,000 x g for 30 min at 4ºC). 1 mL pre-equilibrated Ni-NTA resin (Qiagen) was added to each supernatant, and mixtures were stirred for 1 h at 4ºC. Sample were loaded onto a gravity column and washed with 20 mL wash buffer (500 mM HEPES pH 7.5, 500 mM NaCl, 0.1% (vol/vol) Triton X-100) containing 20 mM imidazole, and then 20 mL wash buffer containing 50 mM imidazole. Proteins were eluted in 5 mL wash buffer containing 300 mM imidazole. The eluate was further purified by size exclusion chromatography with a Superdex 200 10/300 GL column (GE Healthcare) equilibrated in running buffer (50 mM HEPES pH 7.5, 150 mM NaCl, 0.1% (vol/vol) Triton X-100-Reduced). Fractions containing the target proteins were concentrated by centrifugal filtration and protein concentrations were determined with Pierce BCA Protein Assay Kit (Thermo Fisher Scientific). The purity of soluble Spn bPBP2b and bPBP2x fragments was >95% based on Coomassie Blue staining of PAGE gels. Purified PBP fragments were used to generate rabbit polyclonal antibodies (Covance). For this study, anti-bPBP2x and anti-bPBP2b were used without further purification. Western blotting showed that both antibody preparations showed little non-specific background, and comparisons of blots of native PBPs with PBP-fusion constructs ruled out the presence of overlapping non-specific bands at the molecular mass of the PBP or the PBP-fusion constructs (Fig. S3C).