Elongasome core proteins and class A PBP1a display zonal, processive movement at the midcell ofStreptococcus pneumoniae

28 Ovoid-shaped bacteria, such as Streptococcus pneumoniae (pneumococcus), have 29 two spatially separated peptidoglycan (PG) synthase nanomachines that locate zonally 30 to the midcell of dividing cells. The septal PG synthase bPBP2x:FtsW closes the septum 31 of dividing pneumococcal cells, whereas the elongasome located on the outer edge of 32 the septal annulus synthesizes peripheral PG outward. We showed previously by sm-33 TIRFm that the septal PG synthase moves circumferentially at midcell, driven by PG 34 synthesis and not by FtsZ treadmilling. The pneumococcal elongasome consists of the 35 PG synthase bPBP2b:RodA, regulators MreC, MreD, and RodZ, but not MreB, and 36 genetically associated proteins Class A aPBP1a and muramidase MpgA. Given its zonal 37 location separate from FtsZ, it was of considerable interest to determine the dynamics of 38 proteins in the pneumococcal elongasome. We found that bPBP2b, RodA, and MreC 39 move circumferentially with the same velocities and durations at midcell, driven by PG 40 synthesis. However, outside of the midcell zone, the majority of these elongasome 41 proteins move diffusively over the entire surface of cells. Depletion of MreC resulted in 42 loss of circumferential movement of bPBP2b, and bPBP2b and RodA require each other 43 for localization and circumferential movement. Notably, a fraction of aPBP1a molecules 44 also moved circumferentially at midcell with velocities similar to those of components of 45 the core elongasome, but for shorter durations. Other aPBP1a molecules were static at 46 midcell or diffusing over cell bodies. Last, MpgA displayed non-processive, subdiffusive 47 motion that was largely confined to the midcell region and less frequently detected over 48 the cell body. 49 50 (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. The copyright holder for this preprintthis version posted January 27, 2024. ; https://doi.org/10.1101/2024.01.10.575112doi: bioRxiv preprint 3 SIGNIFICANCE 51 T hese results demonstrate that unlike in rod- shaped bacteria, the core elongasome 52 of S. pneumoniae exhibits zonal , circumferential motion. This motion is independent of 53 FtsZ treadmilling or the presence of MreB filaments and is separate from the 54 circumferential motion of the septal PG synthase that closes the septal annulus. Also 55 unlike in rod- shaped bacteria, a Class A PBP moves processively at midcell, distinctly 56 from components of the core PG elongasome or septal PG synthase. Thus, processive, 57 circumferential motion in pneumococcal cells follows spatially separate linear tracks that 58 may reflect a common ordered structure in the existing peptidoglycan itself. In contrast, 59 the MpgA muramidase displays a different kind of subdiffusive motion that is largely 60 confined to midcell by an unknown mechanism. 61 62

63 The peptidoglycan (PG) cell wall protects bacteria from osmotic stress, determines 64 cell shape, size, and chaining critical for adaptation and host interactions, and serves as 65 a scaffolding in Gram-positive bacteria for the attachment of wall-teichoic acids, capsules, 66 and extracellular enzymes and virulence factors (1-5). PG synthesis has provided targets 67 for a large number of clinically relevant antibiotics and remains one of the most fertile 68 sources for the discovery of new antibiotic targets in drug- resistant pathogens (6-11). 69 Although a general picture of PG synthesis has emerged (12-16), many fundamental 70 questions remain unanswered about the composition, coordination, chronology, and 71 regulation of the nanomachines that carry out PG synthesis in bacteria. Moreover, major 72 differences have recently emerged in how different bacteria carry out PG synthesis and 73 (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. The copyright holder for this preprintthis version posted January 27, 2024. ; https://doi.org/10.1101/2024.01.10.575112doi: bioRxiv preprint 4 use homologous proteins in different ways (see (17-23)). Streptococcus pneumoniae 74 (pneumococcus; Spn) has emerged as a model for PG synthesis in ovoid-shaped bacteria 75 that is tractable to genetic, biochemical, and cell biological approaches (reviewed in (4, 76 24, 25)). S. pneumoniae is a human nasopharyngeal commensal bacterium that becomes 77 a serious opportunistic pathogen, killing over one million people annually worldwide (26-78 28). Moreover, S. pneumoniae is an antibiotic -resistant “superbug,” for which new 79 antibiotic targets are urgently needed (10, 29-31). 80 Unlike in rod-shaped bacteria, all PG synthesis is zonal and confined to a band at the 81 middle of dividing S. pneumoniae cells (Fig. S1) (24, 32, 33). At the start of cell division, 82 the septal and elongation (peripheral) PG synthesis nanomachines locate to an FtsZ ring 83 at the equators of predivisional daughter cells (Fig. S1A) (34, 35). Unlike Bacillus subtilis 84 and Staphylococcus aureus (36-39), septal ring closure and cell separation occur 85 simultaneously during the S. pneumoniae cell cycle, as do septal and elongation PG 86 synthesis, with elongation PG synthesis likely starting slightly before septal PG synthesis 87 in the initial FtsZ ring (24, 32, 33, 35, 40) . As septation progresses, the septal and 88 elongation PG synthesis nanomachines physically separate to the inside and outside 89 edges, respectively, of the midcell annular disk (Fig. S1B and S1C) (24, 33, 41). Septal 90 PG synthesis in the inner ring, which contains FtsZ and the Class B bPBP2x:FtsW (SEDS) 91 complex, separates the dividing cells (24, 32- 35). Elongation PG synthesis , which is 92 carried out by the pneumococcal core elongasome in the outer ring (24, 32, 33), emanates 93 outward and has been postulated to push MapZ-nascent FtsZ/FtsA/EzrA rings toward the 94 equators of new daughter cells (Fig. S1A) (34, 42, 43). Notably, S. pneumoniae lacks Min 95 and Noc systems (44, 45) and forms FtsZ rings over the nucleoid (22, 46, 47) . 96 (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. The copyright holder for this preprintthis version posted January 27, 2024. ; https://doi.org/10.1101/2024.01.10.575112doi: bioRxiv preprint 5 Consequently, dividing pneumococcal cells characteristically contain three FtsZ rings 97 throughout most of division, one at the closing septum and two in the developing 98 equatorial rings (Fig. S1A) (34, 46, 48). Interestingly, proteins that mediate bundling FtsZ 99 filaments, such as SepF, ZapA, and ZapJ (ZipA is absent), arrive to the equatorial rings 100 very late in cell division (34, 41, 49). 101 Septal ring closure is zonal and confined to a midcell ring in many bacteria (50- 52). 102 Dynamics studies showed that the pneumococcal bPBP2x:FtsW PG septal synthase 103 moves circumferentially at midcell driven by PG synthesis itself, and not by FtsZ 104 treadmilling (34), as was initially concluded for B. subtilis (53). Recent papers 105 demonstrate that zonal septal closure is indeed driven by circumferential septal PG 106 synthesis in Escherichia coli (54), S. aureus (55), and B. subtilis (56), rather than by FtsZ 107 treadmilling. 108 However, elongation PG synthesis in rod-shaped cells is not zonal but is distributed 109 along the curved cylindrical body of growing cells (57-59). This synthesis is carried out by 110 the Rod- complex elongasome consisting of MreB and the PG elongasome complex. 111 Actin-like MreB polymerizes into multiple short, curved filaments perpendicular to the cell 112 long axis along the cell membrane, which is the region of maximal negative Gaussian 113 curvature (16, 60-62). MreB interacts with the cytoplasmic helix -turn-helix domain of the 114 bitopic RodZ protein that also interacts with the elongasome PG synthase complex facing 115 outside the cell membrane (62-65). The core elongasome consists of the polytopic shape, 116 elongation, division and sporulation ( SEDS) family protein RodA (glycosyltransferase; 117 GT), an essential bitopic Class B bPBP (transpeptidase; TP), and bitopic and polytopic 118 positive regulators MreC and MreD, respectively (16, 66- 70). Similar to septal PG 119 (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. The copyright holder for this preprintthis version posted January 27, 2024. ; https://doi.org/10.1101/2024.01.10.575112doi: bioRxiv preprint 6 synthesis, movement of the assembled Rod elongasome is driven by side- wall PG 120 synthesis itself, rather than by ATPase- dependent treadmilling of MreB (57 -59). Thus, 121 MreB filaments serve as curvature-sensing “rudders” for processive synthesis of separate 122 parallel strands of side-wall PG (61). In addition, Class A PBPs, which are bitopic proteins 123 with extracellular GT and TP activities, move largely in non-ordered paths and often stop, 124 possibly filling in or reinforcing gaps left by the Rod system (16, 71-74). 125 In contrast to rod-shaped bacteria, pneumococcal elongasome PG synthesis is zonal 126 and confined to the initial FtsZ ring and later to the outer edge of the septal annular disk 127 (24, 32, 33). The core pneumococcal elongasome contains RodA, Class B bPBP2b, 128 RodZ, MreC, and MreD, but does not contain an MreB homolog (19, 25, 35, 46, 75- 79). 129 RodZ(Spn) serves as an organizer of elongasome assembly and is required for MreC 130 midcell localization followed by localization of RodA and bPBP2b (19, 80) . 131 Immunofluorescence microscopy (IFM) indicated that Class A aPBP1a and the MpgA 132 muramidase co-localize with bPBP2b and MreC, but not with bPBP2x, throughout the cell 133 cycle, including after the outer elongasome PG synthesis ring forms at midcell (Fig. S1) 134 (35, 46, 80). Thus, aPBP1a and MpgA localize with peripheral PG synthesis, although it 135 has not been determined whether they are persistent members of the core elongasome. 136 Paradoxically, the absence of aPBP1a suppresses the requirement for MreCD and 137 RodZ (80, 81), whereas inactivation of pbp1b, which encodes Class A aPBP1b , 138 suppresses the requirement for RodZ, but not for MreCD (19) . In both cases, the 139 bPBP2b:RodA PG synthase is still required for viability, suggesting that the absence of 140 aPBP1a or aPBP1b may activate bypass pathway s that obviate the requirement for 141 certain regulators of the core elongasome (19, 80) . However, amino acid changes that 142 (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. The copyright holder for this preprintthis version posted January 27, 2024. ; https://doi.org/10.1101/2024.01.10.575112doi: bioRxiv preprint 7 reduce the catalytic activity of the MpgA muramidase (80, 82) abolish the requirement for 143 the entire pneumococcal PG elongasome, including the bPBP2b:RodA PG synthase (24, 144 80). The CozE protein was also implicated in pneumococcal elongation PG synthesis 145 (83). However, CozE did not behave equivalently to MreCD in transformation assays of 146 function (84), and the exact function of pneumococcal CozE remains unknown. A 147 pneumococcal TseB homolog renamed as CopD was recently reported to bind to bPBP2b 148 and aPBP1a. A bsence of CopD produces wider cells, consistent with a defect in 149 elongation PG synthesis (85); h owever, the mechanism underlying this morphology 150 change is not known. Last, FtsX is located with the outer ring elongasome proteins later 151 in pneumococcal division (32). FtsX is a subunit of the FtsEX:PcsB PG hydrolase that is 152 thought to play an essential role in PG remodeling between septal and peripheral PG (86-153 89). Notably, the FtsX(Bsu) homolog is also associated with side-wall PG synthesis (90, 154 91). 155 In this paper, we show that the pneumococcal bPBP2b, RodA, and MreC core 156 elongasome proteins move circumferentially with the same velocity and processivity at 157 midcell at different division stages . Unless indicated otherwise, midcell refers to the 158 central rings at the septa of dividing cells and to the equators of predivisional daughter 159 cells about to start division. The processive movement of the core elongasome 160 components is driven by PG synthesis, and is not affected by FtsZ treadmilling, consistent 161 with the absence of FtsZ in the outer ring of elongasome PG synthesis later in division. 162 Strikingly, at WT protein expression levels in exponentially growing cells, a large majority 163 of bPBP2b, RodA, and MreC molecules diffus e in the membrane over the body of 164 pneumococcal cells without synthesizing PG. Unexpectedly, Class A aPBP1a also moves 165 (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. The copyright holder for this preprintthis version posted January 27, 2024. ; https://doi.org/10.1101/2024.01.10.575112doi: bioRxiv preprint 8 circumferentially at midcell with a similar velocity , but in shorter tracks , than the core 166 elongasome components. Processive movement of aPBP1a at midcell depended on PG 167 synthesis, and again, a majority of aPBP1a molecules were found diffusing over the cell 168 body without synthesizing PG. Finally, the MpgA muramidase show ed a distinctive 169 pattern of non-processive, subdiffusive movement, largely confined to midcell regions . 170 These results demonstrate that the pneumococcal PG elongasome moves zonally, which 171 is different from elongasome movement in rod- shaped bacteria, that a Class A PBP 172 moves circumferentially as well as diffusively, and that a PG hydrolase shows unusual 173 fast non-processive movement confined largely to the midcell region. The implications of 174 these modes of movement to the functions of the pneumococcal PG elongasome, Class 175 A PBPs, and PG muramidases and to patterns of PG -glycan strand placement are 176 discussed. 177 178

179 Construction of functional fluorescent-protein fusions. Fusions to pneumococcal 180 elongasome proteins bPBP2b, RodA, MreC, aPBP1a, and MpgA, were constructed at 181 native chromosomal loci to contain an N- terminal expression-enhancing “i-tag”, fused to 182 an intracellular HaloTag (HT) domain, followed by a 15 amino-acid linker, fused to the 183 protein of interest (Tables S1 and S2), as described previously (32, 34, 92-95). Growth of 184 strains expressing iHT-fused proteins in C+Y, pH 6.9 medium was indistinguishable from 185 that of WT (Fig. S2A). Localization of iHT-fused proteins labeled with saturating amounts 186 of HT-TMR ligand matched previously published localization patterns of epitope- tagged 187 or GFP-fused proteins imaged by IFM or two- dimensional epifluorescence microscopy 188 (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. The copyright holder for this preprintthis version posted January 27, 2024. ; https://doi.org/10.1101/2024.01.10.575112doi: bioRxiv preprint 9 (2D-FM), respectively (Fig. S2B) (19, 32, 34, 35, 42, 46, 80, 81, 85, 96). Western blots of 189 these fusions with intracellular iHT domains probed with antibody to native MreC, 190 bPBP2b, aPBP1a, or MpgA showed expression of full -length fusions with minimal 191 cleavage of the iHT domain (Figs. S3A, S4B, S5B and S6). In contrast, western blots of 192 an extracellular C-terminal HT fusion to MreC (MreC-HT) probed with anti-MreC antibody 193 showed extensive cleavage of the HT domain not detectable in blots probed with anti-HT 194 antibody (Fig. S3A, middle and right ). Cleavage of extracellular HT domains was also 195 reported in E. coli (97), and our results emphasize the need to validate protein fusion 196 stability in blots with antibody to native proteins. 197 We reported previously that some tagged proteins, such as iHT-bPBP2b and sfGFP-198 bPBP2b, were expressed from their chromosomal loc us in significantly lower amounts 199 than WT bPBP2b in cells grown in BHI broth, with only minor effects on cell growth and 200 morphology under this growth condition (32). We observed similar results for cells grown 201 in C+Y, pH 6.9 medium, which is required for live-cell fluorescence microscopy (34). The 202 cellular amounts of iHT -bPBP2b and iHT -aPBP1a were 15 ± 5 % and 10 ± 10% (mean 203 ±SD) of the WT bPBP2b and aPBP1a amounts, respectively (Fig. S4B and S5B; Table 204 S3). Underproduced iHT-bPBP2b caused minor increases in cell width and size (Fig. S4 205 C and D). However, the reduced amount of iHT-aPBP1a partly phenocopied a Δ pbp1a 206 mutant (Fig. S5 C and D). The cellular amounts of iHT -MreC and iHT-MpgA expressed 207 from their chromosomal loci were 82 ±18% and 171 ± 11% of the WT MreC and MpgA 208 amounts, respectively (Figs. S3A and S6; Table S3). Cells expressing iHT-MreC showed 209 minimal cell morphology differences from WT (Figs. S2B and S3B). 210 (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. The copyright holder for this preprintthis version posted January 27, 2024. ; https://doi.org/10.1101/2024.01.10.575112doi: bioRxiv preprint 10 To increase cellular amounts of iHT- bPBP2b, iHT-aPBP1a, and iHT-RodA (antibody 211 to native RodA was not available), we constructed merodiploid strains in which each iHT-212 tagged protein was also expressed from an ectopic site under the control of a zinc -213 inducible promoter (98) . Addition of Zn inducer ( [ZnCl2] indicated + 1/10 [MnSO4] to 214 reduce zinc toxicity (41, 80, 99) ) to cultures of these merodiploid strains did not affect 215 growth (Figs. S4A and S5A). Zn addition complemented the low cellular amounts of iHT-216 bPBP2b and iHT- aPBP1a back up to 82 ± 6% and 1 01 ± 41% of the WT bPBP2b and 217 aPBP1a amounts, respectively (Figs. S4B and S5B; Table S3). Complementation also 218 alleviated minor cell shape defects detected in strains with reduced protein amounts 219 (Figs. S4C and D, S5C and D, and S7). iHT-fused proteins localized strongly to midcell 220 of Zn- induced merodiploid strains, but were also more dispersed over bod ies of cells 221 compared to haploid or uninduced merodiploid strains (Figs. S4C and S5C). Finally, 222 merodiploid strains expressing iHT -bPBP2b or iHT-RodA showed minimal amounts of 223 cleavage products when 0.2 mM or 0.25 mM Zn inducer was added to the growth medium 224 (Figs. S4B and S7B). Effects of protein amounts on dynamics are described below. 225 bPBP2b, RodA, and MreC display similar processive circumferential motion at 226 midcell. Strains expressing functional iHT -fused proteins were labeled with limiting 227 concentrations of HT -JF549 ligand (Table S4) , and single-molecule total internal 228 reflection fluorescence microscopy ( sm-TIRFm) was performed for 180 s, as described 229 previously (34) (Materials and Methods). Three criteria were used to determine different 230 motion types. Processive circumferential motion was defined as a molecule moving in 231 one direction for more than six consecutive frames (imaging rate = 1 frame per second 232 (FPS)), with a linear velocity of ≥ 5 nm/s determined by kymograph analysis (Fig. 1A; 233 (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. The copyright holder for this preprintthis version posted January 27, 2024. ; https://doi.org/10.1101/2024.01.10.575112doi: bioRxiv preprint 11

and Methods). Static molecules were defined as being visible for more than six 234 consecutive frames with a velocity < 5 nm/s. Diffusively moving molecules were defined 235 as moving, but not in a consistent direction, for 6 or more non-consecutive frames within 236 a period of 90 s. 237 Single molecules of iHT -bPBP2b, iHT -RodA, and iHT-MreC displayed similar 238 processive circumferential motion exclusively at midcell (Fig. 1A and S8 and Movies S1 239 to S3), that resembled the circumferential motion of i HT-bPBP2x and FtsW-HT reported 240 previously (34) . T he frequency of circumferential, static, and diffusive molecules w as 241 compiled for each strain (Fig. 1B). Examination of the circumferential tracks showed that 242 73% (54/74) of iHT- bPBP2b and 74% (57/77) of iHT- MreC molecules moved 243 unidirectionally at midcell. The remaining 27% of iHT- bPBP2b or 26% of iHT -MreC 244 molecules moved in one direction for at least six consecutive frames and then reversed 245 direction and moved in the opposite direction for at least six frames at approximately the 246 same velocity (angle in kymograph; Fig. 1A, right). Slow or non-moving static molecules 247 were mainly detected at midcell (Fig. S8), whereas diffusive molecules displayed rapid, 248 erratic movement throughout the cell (Fig. 1B and S8 and Movies S1 to S3). Fewer than 249 5% of molecules of these core elongasome proteins were observed to transition between 250 movement states (e.g., static to diffusive or circumferential). 251 A limited number of bPBP2b, RodA , and MreC molecules engage in active PG 252 synthesis. The distribution of single-molecule movement states of iHT-bPBP2b and iHT-253 RodA was highly dependent on cellular protein levels. When expressed at 15% of the WT 254 level, ≈60% of iHT-bPBP2b moved circumferentially and only 32% moved diffusively (Fig. 255 1B and Movie S1). The remaining 7 % of iHT -bPBP2b molecules were static , mainly at 256 (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. The copyright holder for this preprintthis version posted January 27, 2024. ; https://doi.org/10.1101/2024.01.10.575112doi: bioRxiv preprint 12 midcell. iHT -RodA molecules expressed from the native locus showed a similar 257 distribution, suggesting that iHT -RodA was under produced (Fig. 1 B and Movie S2). 258 However, when iHT-PBP2b and iHT -RodA protein amounts were increased by ectopic 259 expression (to ≈82% the WT level for iHT-bPBP2b), only 16% or 9% of single iHT-bPBP2b 260 or iHT -RodA molecules, respectively, moved circumferentially, while ≈70% moved 261 diffusively, with the remaining 12% iHT- bPBP2b or 23% iHT -RodA static, mostly at 262 midcell (Fig. 1B and Movies S4 and S5). Uninduced merodiploid strains expressing iHT-263 bPBP2b or iHT-RodA displayed movement distributions similar to those of haploid strains 264 (Fig. 1B). The amount iHT-MreC expressed from its native locus was nearly (82%) that 265 of WT (Fig. S 3A and Table S3). Approximately 24% of iHT- MreC molecules moved 266 circumferentially, while 52% moved diffusively and 24% were static , similar to the 267 movement distributions of iHT -bPBP2b and iHT- RodA when they were ectopically 268 expressed (Fig. 1B and Movie S3). 269 High-resolution 3D structured- illumination microscopy (3D -SIM) on fixed cells 270 corroborated conclusions based on movement distributions. At protein amounts 271 corresponding to ≈15% of WT, iHT-bPBP2b localized primarily to regions of active PG 272 synthesis at midcell and equators of dividing cells , as indicated by HADA fluorescent D -273 amino-acid labeling (Fig. 2A). When expressed near the WT level, iHT-bPBP2b remained 274 at the HADA-labeled midcell and equators but was mostly dispersed over the body of the 275 entire cell where HADA labeling was not detected (Fig. 2B ). These results support the 276

that iHT-PBP2b molecules that move diffusively in the cell membrane outside 277 of midcell and equators are not actively synthesizing PG (Fig. 2B). Control experiments 278 using a GFP fusion construct to bPBP2b and labeling with a different fluorescent D-amino 279 (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. The copyright holder for this preprintthis version posted January 27, 2024. ; https://doi.org/10.1101/2024.01.10.575112doi: bioRxiv preprint 13 acid (TADA) gave similar results ( Fig. S9). 2D-FM images also showed an increase in 280 dispersed iHT-bPBP2b localization when protein expression was increased to near the 281 WT amount (Fig. S4C). We conclude that at any given time, a limited number of bPBP2b, 282 RodA, and MreC molecules display processive circumferential motion at midcell where 283 PG synthesis occurs. Conversely, bPBP2b, RodA, and MreC are expressed in excess in 284 WT cells, and most molecules are diffusing over the cell surface and not synthesizing PG 285 in cells under these and other growth conditions (Figs. 1B, 2B and S4C; Table S3) (32). 286 bPBP2b, RodA, and MreC form a stable elongasome complex that undergoes 287 circumferential motion. If bPBP2b, RodA, and MreC form an elongasome complex, we 288 would expect that each molecule would move with the same velocity. To test this idea, 289 velocities of circumferentially moving molecules were determined by kymograph analysis 290 (Fig. 3A ) (34). The length of time these molecules were observed moving 291 circumferentially, referred to here as the circumferential duration, was also measured 292 (Fig. 3B), and the particle distance each circumferential molecule moved was calculated 293 by multiplying the velocity by the duration (Fig. S10A). When expressed near WT levels 294 (+Zn in merodiploid and haploid mreC constructs) the mean circumferential velocities, 295 durations, and particle distances of iHT -bPBP2b, iHT -RodA, and iHT -MreC were 296 statistically the same, at ≈11 nm/s, ≈23 s, and ≈230 nm, respectively (Fig. 3A and B; Fig. 297 S10A). Additional strains containing different fusion constructs of bPBP2b or RodA moved 298 at similar circumferential velocities (Fig. S11). These data support the idea that bPBP2b, 299 RodA, and MreC form a stable, circumferentially moving elongasome complex at midcell. 300 Circumferential movement of bPBP2b and RodA is independent of FtsZ 301 treadmilling and driven by PG synthesis. Circumferential movement of the 302 (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. The copyright holder for this preprintthis version posted January 27, 2024. ; https://doi.org/10.1101/2024.01.10.575112doi: bioRxiv preprint 14 pneumococcal septal PG synthase was shown to be independent of FtsZ treadmilling and 303 driven by PG synthesis {Perez, 2019 #136}. A similar conclusion was later reported for 304 the septal PG synthases of S. aureus (55) and B. subtilis (56) and for the population of 305 active septal PG synthases in E. coli (54, 100). In early divisional cells, the pneumococcal 306 PG elongasome is located in an FtsZ-ring that also contains the septal PG synthase (Fig. 307 S1A), whereas at later stages of division, the elongasome separates from the FtsZ- ring 308 and locates to the outer edge of the septal annular disk (Fig. S1 B and C) (24, 32, 33). 309 Notably, the same circumferential velocity of core elongasome protein iHT-MreC, which 310 was expressed near the WT level (Fig. S3A), was detected at equators of early divisional 311 cells, before ostensible constriction, and in later -divisional cells with clear constrictions, 312 when septal (inner) and elongasome (outer) rings are separated at midcell (32) (Figs. 313 S1B and S10B). Since the pneumococcal PG elongasome is initially in an FtsZ ring, we 314 tested whether iHT- bPBP2b velocity is decreased when FtsZ treadmilling is greatly 315 reduced by ectopic expression of GTPase mutant FtsZ(D214A), which reduces FtsZ 316 treadmilling velocity from ≈32 nm/s to ≈4 nm/s (34). To the contrary, overproduction of 317 FtsZ(D214A) caused a statistically minimal increase in iHT-bPBP2b velocity (Fig. 4A and 318 Movie S6), with a marginal decrease in growth rate (Fig. S12). Consistent with a lack of 319 dynamic coupling, the treadmilling velocity of FtsZ-sfGFP filaments was ≈2.4-fold greater 320 than the circumferential velocity of iHT-bPBP2b single molecules expressed in the same 321 cell (Fig. 4B; Movie S 7). We conclude that circumferential movement of the 322 pneumococcal PG elongasome is independent of FtsZ treadmilling. 323 Two lines of evidence indicate that pneumococcal PG elongasome circumferential 324 movement is driven by PG synthesis. First, we determined iHT- bPBP2b or iHT-RodA 325 (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. The copyright holder for this preprintthis version posted January 27, 2024. ; https://doi.org/10.1101/2024.01.10.575112doi: bioRxiv preprint 15 circumferential velocity in a Δ murZ (previously called Δ murA1) mutant, which lacks the 326 predominant MurZ homolog that catalyzes the first committed step of PG precursor 327 synthesis (34, 101). The circumferential velocity of iHT- bPBP2b or iHT-RodA was 328 substantially reduced by ≈1.6-fold in the ΔmurZ mutant compared to WT (Fig. 4 C and D, 329 Movies S8 and S9). Previously, we showed that the velocity of FtsZ treadmilling is 330 unchanged in the ∆murZ mutant (34). Together, these results demonstrate that the rate 331 of midcell elongasome movement is dependent on PG synthesis precursor amounts. 332 Second, catalytic mutants of bPBP2b TP or RodA GT activity did not exhibit 333 processive, circumferential movement. No β-lactam antibiotic is known that preferentially 334 inhibits bPBP2b alone (102). Both bPBP2b and RodA are essential in S. pneumoniae (35, 335 75, 103). Consequently, we constructed merodiploid strains with iht -pbp2b(S391A) (TP 336 catalytic mutation) or iht-rodA(D283A) (GT catalytic mutation) (69) at chromosomal native 337 sites and pbp2b + or rodA+ under the control of the zinc -inducible promoter (P Zn) in an 338 ectopic site (Fig. 5A ). The merodiploid strains were constructed and grown in the 339 presence of Zn inducer, which was removed to deplete the WT proteins (Fig. 5 B). 340 Interestingly, we were able to construct a rodA (D283A)//PZn-rodA+ strain, but we were 341 unable to construct a pbp2b(S391A)//P Zn-pbp2b+ strain in the presence of Zn inducer. 342 This result suggest ed that pbp2b(S391A) was dominant-negative to WT pbp2b+. The 343 reason we were able to construct the iht-pbp2b(S391A)//PZn-pbp2b+ merodiploid was the 344 low (≈15%) intrinsic expression of iHT- bPBP2b compared to WT bPBP2b (Fig. S4B ). 345 Consistent with a dominant -negative effect, depletion of WT bPBP2b in the iht -346 pbp2b(S391A)//PZn-pbp2b+ merodiploid resulted in a more severe drop in OD 620 than in 347 Δpbp2b//PZn-pbp2b+ (Fig. 5B), possibly due to the toxicity of uncrosslinked PG produced 348 (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. The copyright holder for this preprintthis version posted January 27, 2024. ; https://doi.org/10.1101/2024.01.10.575112doi: bioRxiv preprint 16 from GT activity without TP activity (104). Western blot analysis confirmed that both iHT-349 bPBP2b(S391A) and iHT- RodA(D283A) were expressed and stable relative to iHT-350 bPBP2b+ and iHT-RodA+, respectively (Fig. S13A), and that bPBP2b+ expressed from the 351 ectopic locus was fully depleted after 3h (Fig. S13B). 352 No midcell circumferential movement of iHT- bPBP2b(S391A) was detected in the 353 presence or absence of WT bPBP2b (Fig. 5 C; Movies S10 and S11). Instead, iHT-354 bPBP2b(S391A) was static at midcell (Fig. 5D ) or was detected diffusing over the body 355 of cells (Fig. 5C ). Likewise, no midcell circumferential movement of iHT- RodA(D283A) 356 was detected in the absence of RodA + (Fig. 5 C and Movie S12). Like iHT -357 bPBP2b(S391A), iHT-RodA(D283A) was static at midcell (Fig. 5D ) or diffusing over the 358 body of cells (Fig. 5C). Notably, both iHT-bPBP2b(S391A) and iHT-RodA(D283A) located 359 to midcell (Fig. 5E ), indicating that catalytic activity was not required for normal 360 localization. From these combined results, we conclude that processive, circumferential 361 movement of the pneumococcal elongasome PG synthase is driven by PG synthesis, but 362 not by FtsZ treadmilling, even in predivisional cells. 363 bPBP2b localization becomes diffuse and circumferential movement is lost 364 when MreC is depleted. We reported previously that the pneumococcal PG elongasome 365 is assembled sequentially (19) . RodZ directs localization of MreC, which then directs 366 localization of the bPBP2b:RodA synthase. In agreement with this conclusion, iHT-367 bPBP2b localization became diffuse when MreC was depleted (Fig. 6A). In addition, there 368 was a large (≈30- fold) drop in the number of iHT -bPBP2b molecules mov ing 369 circumferentially when MreC was depleted ( Fig. 6B and Movie S 13). Growth curve 370 analysis confirm ed reduced growth yield of cells depleted for MreC ( Fig. S1 4A), and 371 (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. The copyright holder for this preprintthis version posted January 27, 2024. ; https://doi.org/10.1101/2024.01.10.575112doi: bioRxiv preprint 17 western blots confirmed the presence of iHT-bPBP2b during MreC depletion (Fig. S14 B 372 and C). We conclude that MreC is required for proper midcell localization and for 373 elongasome PG synthesis by the bPBP2b:RodA synthase. This result is consistent with 374 the previous conclusion that MreC allosterically activates the GT/TP activity of 375 RodA:PBP2 in the PG elongasome of E. coli (16, 67, 105). 376 b PBP2b and RodA require each other for localization. We further tested whether 377 RodA was required for bPBP2b localization, and vice versa . Indeed, upon RodA 378 depletion, iHT- bPBP2b went from locating primarily to midcell to being distributed at 379 midcell and over the surface of rounded cells (Figs. 7 A and S15C), indicative of loss of 380 peripheral PG synthesis (35, 75, 81). Additionally, a growth yield phenotype was observed 381 at 4 to 5 h into depletion (Fig. S15 A and B). Upon bPBP2b depletion to ≈6% of WT after 382 3 h (Fig. S13 B), iHT-RodA also relocated primarily from midcell to being distributed at 383 midcell and over the surface of rounded cells , with a similar reduction in growth yield 384 (Figs. 7 B and S15). Therefore, there is an interdependence for normal localization of 385 bPBP2b and RodA in the pneumococcal PG elongasome. 386 We next determined the dynamics of iHT -bPBP2b upon depletion of RodA or iHT-387 RodA upon depletion of bPBP2b. In both cases, the motion distribution shifted from 388 primarily circumferential (≈59%) to diffusive (≈65%) movement (Fig. 7C; Movies S14 and 389 S15), where iHT-bPBP2b and likely iHT-RodA were underproduced in these stains (Figs. 390 1B, S7 and S13B ). Remaining circumferential movement at the midcell (≈24% of 391 molecules) likely reflected bPBP2b :RodA complexes that assembled before depletion 392 occurred. Consistent with this interpretation, the velocity of circumferentially moving 393 proteins was unchanged by depletion of its cognate partner (Fig. 7 D and E). These data 394 (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. The copyright holder for this preprintthis version posted January 27, 2024. ; https://doi.org/10.1101/2024.01.10.575112doi: bioRxiv preprint 18 support the idea that when bPBP2b or RodA is depleted, its unpaired cognate partner 395 stops elongation PG synthesis at midcell and is released to diffuse over the cell body. 396 Notably, released, unpaired iHT-b PBP2b or iHT-RodA did not move circumferentially at 397 an increased velocity, suggesting that FtsZ filament end-tracking does not occur for these 398 proteins (100). 399 aPBP1a displays circumferential dynamics at midcell driven by PG synthesis 400 but is not persistently part of the elongasome. M olecules of iHT- aPBP1a 401 unexpectedly displayed circumferential, as well as diffusive and static dynamics (Fig. 1 A 402 and B; Movies S16 to S 18). Like iHT-bPBP2b and likely iHT- RodA, iHT-aPBP1a was 403 underproduced in the haploid strain and required ectopic expression to reach or surpass 404 WT levels (Fig. 1B and Fig. S5B). However, in contrast to core elongasome components, 405 the relative frequency of circumferentially moving iHT- aPBP1a molecules (≈10%) was 406 largely independent of protein expression level (Fig. 1B), and a high percentage (>60%) 407 of iHT-aPBP1a molecules moved diffusively, even at low expression levels (Fig. 1B; Fig. 408 S5B; Table S3). 409 The mean circumferential velocity of iHT-aPBP1a expressed near WT levels was ≈15 410 nm/s, which was similar to, if not somewhat faster than, that of iHT-bPBP2b, iHT-RodA, 411 and iHT-MreC (≈11 nm/s) (Figs. 1A and 3A). Strikingly, the mean circumferential duration 412 of iHT-aPBP1a molecules was ≈10 s, more than two-fold shorter than that of iHT-bPBP2b, 413 iHT-RodA, or iHT-MreC (≈23 s) (Fig. 1A and 3B). Altogether, these results indicate that 414 circumferential motion of aPBP1a is distinct from that of bPBP2b, RodA , and Mre C, 415 suggesting that aPBP1a is not persistently part of the core elongasome complex, 416 although shorter interactions are still possible. 417 (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. The copyright holder for this preprintthis version posted January 27, 2024. ; https://doi.org/10.1101/2024.01.10.575112doi: bioRxiv preprint 19 To confirm that circumferential movement of iHT-aPBP1a molecules depends on PG 418 synthesis, we constructed an iht-pbp1a(S370A)//PZn-iht-pbp1a(S370A) merodiploid strain 419 expressing catalytically inactive (TP null) iHT-aPBP1a(S370A) from both the native site 420 and ectopic loc us (Table S1) . The growth rate of the catalytically inactive iHT-421 aPBP1a(S370A) merodiploid strain was reduced slightly (≈1.3-fold) when Zn inducer was 422 added to the growth medi um (Fig. S16A), but protein expression and localization were 423 comparable to that of iHT -aPBP1a merodiploid strains (Fig. S16 B and C). 424 Circumferential, diffusive , and static mo lecules w ere sorted by cellular location into 425 midcell or locations other than midcell (non- midcell) (Fig. 8). Circumferential movement 426 of iHT -aPBP1a decreased substantially from a relative frequency of ≈21% for iHT-427 aPBP1a to a marginal ≈3% for iHT -aPBP1a(S370A) at midcell, with the remainder of 428 midcell iHT -aPBP1a or iHT -aPBP1a(S370A) molecules static (Fig. 8, Movie S19 ). 429 Essentially no circumferentially moving molecules of iHT- aPBP1a or iHT -430 aPBP1a(S370A) were detected outside of midcell (Fig. 8). A similar conclusion was 431 reached when data from Fig. 1B was sorted into midcell or non- midcell locations (Fig. 432 S17A). Since no PG synthesis was detectable by FDAA labeling away from the midcell 433 (Figs. 2 and S9), and nearly all aPBP1a molecules away from midcell were moving 434 diffusively (Figs. 8 and S17A), this further suggests that diffusively moving aPBP1a were 435 not actively synthesizing PG in exponentially growing cells. Taken together, these data 436 indicate that circumferentially moving iHT -aPBP1a molecules are actively synthesizing 437 PG, while inactive molecules are either static at midcell or moving diffusively throughout 438 the cell. A similar conclusion was reached for elongasome proteins iHT- bPBP2b, iHT-439 RodA, and iHT-MreC (Figs. 1B and 2). 440 (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. The copyright holder for this preprintthis version posted January 27, 2024. ; https://doi.org/10.1101/2024.01.10.575112doi: bioRxiv preprint 20 Other data were consistent with aPBP1a not being a constant component of the PG 441 elongasome. Notably, the velocity of iHT -aPBP1a in a merodiploid iht-pbp1a//PZn-iht-442 pbp1a ∆murZ strain was similar to that of iHT-aPBP1a in a murZ+ merodiploid strain (Fig. 443 S17B, Movie S20). Growth rate, protein amount, localization and relative frequency of 444 circumferential molecules of iHT -aPBP1a were also similar in the ∆ murZ mutant and 445 murZ+ strain (Figs. S16 and S17C). Therefore, unlike iHT-bPBP2b and iHT-RodA (Fig. 4 446 C and D), the circumferential velocity of iHT -aPBP1a was not decreased when PG 447 precursor amounts were reduced. In addition, we tested whether the relative frequency 448 of circumferential iHT-aPBP1a was changed in a mpgA (Y488D) Δpbp2b mutant, which 449 lacks the core PG elongasome (80, 82). Reduced catalytic activity of MpgA(Y488D) 450 bypasses the requirement for the PG elongasome, but mpgA(Y488D) Δpbp2b mutations 451 are synthetically lethal with Δpbp1a (80), possibly indicating that aPBP1a can substitute 452 for the PG elongasome. Consistent with this notion, iHT- aPBP1a dynamics were the 453 same in a mpgA(Y488D) Δpbp2b merodiploid strain as in a WT strain (Fig. S17C, Movie 454 S21). 455 MpgA displays subdiffusive movement restricted to midcell. We investigated the 456 dynamics of iHT-MpgA (previously MltG(Spn)) muramidase, which has been genetically 457 linked to elongasome PG synthesis (80, 82) . At an image acquisition rate of 1 FPS, t he 458 majority of iHT-MpgA molecules (≈87%) appeared to move diffusively, but seemed to be 459 constrained mainly to the midcell region of cells (Fig. 1B and Movie S22), consistent with 460 a previous report that sfGFP-MpgA localizes at midcell (34). The remaining ≈13% of iHT-461 MpgA molecules were static at midcell (Fig. 1B ). There was also a small fraction ( 50 nm/s) movement of very short 463 (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. The copyright holder for this preprintthis version posted January 27, 2024. ; https://doi.org/10.1101/2024.01.10.575112doi: bioRxiv preprint 21 duration at midcell (Figs. 3A, 3B, and 9A). However, this motion was difficult to track and 464 at the limit of the criterion used for continuous motion (six consecutive frames). 465 Consequently, we increased the data acquisition rate to 10 FPS. At this higher rate, 466 iHT-MpgA did not exhibit the processive, unidirectional movement observed for the 467 elongasome components and for aPBP1a (Figs. 1 A and 9B). Instead, iHT-MpgA moved 468 erratically, mainly in the midcell region (see below; Fig. 9B and Movie S23). Measured at 469 10 FPS, the mean velocity of iHT -MpgA moving in short , directional stretches was 470 relatively fast (≈327 nm/s) compared to circumferential movement of other proteins (1,000 nm/s) (Fig. 9C). 472 These distinctive subdiffusive dynamics did not support persistent association of MpgA 473 with the PG elongasome or aPBP1a during PG synthesis. 474 Given these unusual dynamics, we confirmed the functionality of the iHT-MpgA fusion 475 construct. The growth and cell morphology of strains expressing iHT-MpgA were similar 476 to those of WT, and iHT-MpgA showed expected midcell localization (Fig. S2). A mutant 477 expressing partially active MpgA(Y488D) suppress ed the essentiality of the 478 pneumococcal PG elongasome in transformation assays (80) . However, the iht -mpgA 479 construct failed to suppress Δ pbp2b, ΔrodA, ΔmreC, or ΔrodZ in transformation assays 480 (Table S5), indicating iHT-MpgA possessed muramidase activity. Unlike ΔmpgA mutants, 481 which accumulate suppressor mutations that inactivate aPBP1a (80); no pbp1a mutations 482 were detected in the iht -mpgA fusion strain by Sanger DNA sequencing. These results 483 indicated that the iHT-MpgA fusion protein was substantially active and that its dynamics 484 accurately reflected those of WT MpgA. 485 (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. The copyright holder for this preprintthis version posted January 27, 2024. ; https://doi.org/10.1101/2024.01.10.575112doi: bioRxiv preprint 22 C ore PG elongasome components, aPBP1a, and MpgA display different 486 diffusion patterns. We next investigated diffusion dynamics by performing sm-TIRFm at 487 20 FPS (Materials and Methods). Individual molecular trajectories were tracked (Fig. 10A) 488 and used to calculate mean-squared displacements (MSD) and diffusion coefficients (Fig. 489 10B) (106, 107). Diffusion coefficients were between 0.040 and 0.055 µm2/s, comparable 490 to values determined previously for PG synthesis proteins in S. pneumoniae, E. coli and 491 B. subtilis (56, 100, 108, 109) . Consistent with sm- TIRFm imaging at 1 FPS (Fig. 1B ), 492 single molecules displayed rapid diffusive motion in non-midcell regions and much slower, 493 processive motion at midcell, where static molecules were also observed (Fig. 10A and 494 Movies S24 to S30). MSD curves generated for diffusively moving molecules were not 495 linear and plateaued as time increased, with alpha values <1, which is indicative of some 496 form of subdiffusion (Fig. 10B ). Subdiffusion is commonly due to diffusion within a 497 confined space, but can also be caused by membrane crowding or interactions with other 498 molecules (110). For bacterial membrane proteins, this confined space is often the 499 surface area of the cell membrane (110). 500 S ingle-molecule trajectories of diffusively moving PG elongasome core components 501 iHT-bPBP2b, iHT-RodA, and iHT-MreC covered most of the cell area, while non-diffusive 502 processive and static molecules located to midcell, where active PG synthesis occurred 503 (Fig. 10A and Movies S24 to S26), consistent with results acquired at 1 FPS (Fig. S8) . 504 This suggests that the subdiffusive behavior of the core elongasome proteins was due to 505 confinement within the cell membrane. The diffusion coefficients of iHT-bPBP2b and iHT-506 MreC were similar (≈0.046 µm2/s), and the MSD curves of diffusively moving molecules 507 plateaued in the same range ( ≈0.07 µm 2), indicating that iHT -bPBP2b and iHT -MreC 508 (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. The copyright holder for this preprintthis version posted January 27, 2024. ; https://doi.org/10.1101/2024.01.10.575112doi: bioRxiv preprint 23 molecules were confined to a similarly sized area (Fig. 10B) (111). By contrast, iHT-RodA 509 had a slightly higher diffusion coefficient (0.050 µm2/s) with a higher MSD curve plateau 510 (≈0.08 µm 2). These data suggest that bPBP2b and MreC may diffuse as a complex 511 through the membrane separately from RodA. Additional experiments are needed to test 512 this idea. 513 D iffusively moving iHT -aPBP1a molecules had an even greater diffusion coefficient 514 (≈0.055 µm2/s) than those of the core components of the PG elongasome (Fig. 10B). iHT-515 aPBP1a molecules also had a higher plateau in MSD plots (≈0.1 µm 2), indicative of 516 diffusive movement that was observed over the whole cell (Fig. 10A and Movie S27). 517 Again, non-diffusive, processively moving and static iHT-aPBP1a molecules were mostly 518 confined to midcell (Fig. 10A), as was observed by sm-TIRFm at 1 FPS (Fig. 8 and S17A). 519 Interestingly, we also observed some diffusive iHT- aPBP1a molecules that stopped 520 moving, paused for a short time (<1 s), and then resumed diffusive movement (Movie 521 S28). This pausing behavior was not readily detectable for diffusing components of the 522 PG elongasome. We conclude that diffusing aPBP1a molecules move differently from 523 components of the core PG elongasome in pneumococcal cells and that there is an 524 intermediate paused state during diffusion of aPBP1a. 525 Finally, diffusion analysis confirmed the unusual pattern of confined movement of iHT-526 MpgA molecules at midcell. Unlike iHT-bPBP2b, iHT-RodA, iHT-MreC, and iHT-aPBP1a, 527 the majority of iHT -MpgA molecules did not diffuse over the whole cell but were largely 528 confined to midcell (Fig. 10A and Movie S29). This conclusion was further corroborated 529 by manually counting the location of iHT -MpgA tracks from data in Figure 10A. This 530 analysis showed that 74% (148/200) of the tracks of subdiffusive iHT-MpgA tracks were 531 (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. The copyright holder for this preprintthis version posted January 27, 2024. ; https://doi.org/10.1101/2024.01.10.575112doi: bioRxiv preprint 24 located within 125 nm of midcell septa or equatorial rings. This confinement is likely 532 reflected by the low plateau (≈0.05 µm 2) in the MSD plot. In addition, the diffusion 533 coefficient (≈0.04 µm2/s) of iHT -MpgA was considerably lower than those of the other 534 proteins that diffused over the body of cells (Fig. 10B ), consistent with the sm- TIRFm 535 measurements at 10 FPS (Fig. 9). Similar to iHT -aPBP1a, iHT -MpgA subdiffusive 536 molecules were observed transitioning between diffusive motion and paused states 537 (Movie S30). Together, these results indicate that the confined, subdiffusive movement 538 of MpgA molecules is fundamentally different from that of components of the 539 pneumococcal core PG elongasome. 540 541

542 This paper shows that the core PG elongasome moves circumferentially around the 543 midcell of dividing S. pneumoniae cells (Fig. 1) . The velocity of core elongasome 544 members bPBP2b, RodA, and MreC was statistically the same (≈ 11 nm/sec) (Fig. 3), 545 which is slightly slower than the velocity of the bPBP2x and FtsW ( ≈20 nm/sec) 546 components of the septal PG synthase (34) . The processive motion of elongasome 547 members was driven by PG synthesis and was independent of FtsZ treadmilling (Fig. 4 548 and 5), and this motion was not dependent on stage of cell division (Fig. S10B). Strikingly, 549 the majority of bPBP2b or RodA molecules moved circumferentially in cells that severely 550 underproduced bPBP2b ( ≈15% of WT) or likely RodA (Fig. 1B ). In contrast, in cells 551 expressing nearly WT protein levels, a minority of bPBP2b, RodA, and MreC molecules 552 either moved circumferentially and synthesized PG at midcell or exhibited nonmoving 553 static behavior at midcell (Fig. 1B and S8). The majority of core elongasome components 554 (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. The copyright holder for this preprintthis version posted January 27, 2024. ; https://doi.org/10.1101/2024.01.10.575112doi: bioRxiv preprint 25 moved diffusively in the membrane over cell bodies, where PG synthesis was not 555 detected by FDAA labeling of exponentially growing cells (Fig. 2 and S9). 556 These observations indicate that pneumococcal elongation PG synthesis results from 557 processive, circumferential movement of PG elongasomes that are confined to a narrow 558 zone at midcell (Fig. 11). This pattern contrasts with the elongasome movement guided 559 by short MreB filaments and driven by PG synthesis that occurs over the body of many 560 rod-shaped bacteria (16) . MreB -guided PG synthesis results in a dense mesh of 561 concentrically oriented PG glycan strands along the inner , sidewall surface of B. subtilis 562 cells (112). The circumferential movement of the pneumococcal PG elongasome 563 anticipates that peripheral PG glycan strands may likewise be concentrically aligned in S. 564 pneumoniae cells. These observations also indicate that in exponentially growing WT 565 cells, the core elongasome components are present in excess, with only a minority of 566 bPBP2b, RodA, and MreC engaged in active PG synthesis at midcell rings (Fig. 2 and 567 S9). 568 The results presented here further raise the issue of what directs circumferential 569 movement of elongasome components in tracks at midcell. The velocity of elongasome 570 components was not changed when treadmilling was greatly impeded in an ftsZ(GTPase) 571 mutant (Fig. 4A). Independence of FtsZ treadmilling fits the observation that FtsZ was not 572 detected in the outer midcell ring where peripheral PG synthesis occurs later in division 573 (Fig. S1B) (32). The similar velocity of MreC in predivisional cells when elongasomes are 574 still located in the FtsZ- ring as in later -divisional cells when the elongasome is in outer 575 midcell ring (Fig. S10B) is also consistent with independence of FtsZ treadmilling. In 576 addition, when RodA or bPBP2b was depleted, no unassociated bPBP2b or RodA was 577 (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. The copyright holder for this preprintthis version posted January 27, 2024. ; https://doi.org/10.1101/2024.01.10.575112doi: bioRxiv preprint 26 detected being driven at the rate of treadmilling FtsZ (≈34 nm/sec; Fig 7 D and E), as 578 occurs for components of the septal PG synthase of E. coli (100). Moreover, a severe 579 decrease in FtsZ treadmilling speed led to only a marginal (≈17%) decrease in the speed 580 of the septal PG synthase in S. pneumoniae (34) and no decrease in S. aureus ( 55). In 581 B. subtilis, a slightly greater decrease ( ≈40%) in the speed of the septal PG synthase 582 when FtsZ treadmilling was moderately or greatly reduced has recently been attributed 583 to indirect effects (56). Thus, all evidence supports the conclusion that the movement and 584 speed of the septal and elongasome PG synthases in WT S. pneumoniae cells are driven 585 by PG synthesis, independent of FtsZ treadmilling. 586 It remains unknown what determines the track s for the movement of the 587 pneumococcal septal or elongasome PG synthases in the inner and outer midcell rings, 588 respectively (Fig. S1). The muropeptide-crosslinked PG glycan strands to which new PG 589 is added may provide a rigid structure that directs processive PG synthesis in a linear 590 direction. Moreover, there appears to be a limited number of available sites in midcell 591 rings for septal and elongasome PG synthases to load and synthesize PG . A s noted 592 above, only a minority of elongasome components are engaged in PG synthesis in WT 593 cells (Fig. 1B, 2, and 11). However, when the cellular amount of bPBP2b was reduced to 594 ≈15% of the WT level, the majority of bPBP2b molecules moved processively to 595 synthesize PG (Fig. 1B), consistent with a limited number of spots on rings for active PG 596 synthase complexes. Interestingly, ≈26% of circumferentially moving bPBP2b or MreC 597 molecules reversed direction within the timeframe of these experiments. D irection 598 reversal was also reported for processively moving members of the septal PG synthases 599 of S. aureus (55) and B. subtilis (113). This directional transitioning may be indicative of 600 (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. The copyright holder for this preprintthis version posted January 27, 2024. ; https://doi.org/10.1101/2024.01.10.575112doi: bioRxiv preprint 27 PG synthase complexes stopping and reinitiating synthesis of a new glycan strand in the 601 opposite direction (55, 113). 602 We also observed nonmoving , static molecules of bPBP2b, RodA, and MreC 603 expressed at WT concentrations, primarily at midcell and to a lesser ( ≈5-fold) extent 604 elsewhere in cells (Fig. 1B and S8). W hen elongasome PG synthesis was halted by 605 expression of catalytically inactive proteins, circumferential movement also stopped and 606 molecules became static (Fig. 5C and 5D; Fig. 8) (69). To our knowledge, there is no 607 direct evidence that PBPs static for several seconds are actively synthesizing PG, as 608 suggested for aPBPs in E. coli (71, 74). To the contrary, when PG synthases stop 609 synthesis, they stop moving (Fig. 5 C and D; Fig. 8) (34, 55, 56, 114). Therefore, we think 610 it more likely that static PG synthases at midcell in WT cells may be incompletely 611 assembled complexes or complexes waiting to assume an available site for PG synthesis, 612 possibly stabilized by noncovalent binding of PBP TP domains to acceptor peptides in 613 PG (56, 109). Consistent with this idea, we observed transitions of MreC molecules from 614 the processive to static states (<5%), and vice versa (<1%). Outside of midcell, diffusion 615 analysis showed that inactive bPBP2b and MreC may diffuse together as a complex over 616 the cell body, while RodA diffuses separately (Fig. 10B ). The mechanisms that arrange 617 and limit the number of active pneumococcal PG elongasomes remain to be determined. 618 The processive movement of pneumococcal Class A aPBP1a strikingly contrasts with 619 the diffusive motion for Class A PBPs reported previously in E. coli and B. subtilis (71, 620 74). Earlier IFM studies colocalized aPBP1a with elongasome components bPBP2b and 621 MreC (35). Instead of diffusive motion, aPBP1a molecules at midcell moved 622 circumferentially or were static (Fig. 8 and S17), and processive movement was abolished 623 (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. The copyright holder for this preprintthis version posted January 27, 2024. ; https://doi.org/10.1101/2024.01.10.575112doi: bioRxiv preprint 28 in a pbp1a (TP) mutant (Fig. 8) . Processively moving aPBP1a also reversed direction 624 infrequently (<9%). However, several results indicate that aPBP1a is not a persistent 625 member of the core PG elongasome. T he duration of aPBP1a processive motion was 626 considerably shorter (≈10 s) than that of the PG elongasome compo nents (≈23 s) (Fig. 627 3B). In addition, t he relative frequency of circumferential motion of aPBP1a was 628 unchanged in a suppressed Δ pbp2b mpgA(Y488D) mutant that lacks an intact PG 629 elongasome (Fig. S17C). Like iHT-bPBP2b, iHT-aPBP1a was underproduced ( 10% of 630 WT) when expressed from its native chromosomal locus (Figs. 1B and S5B). But, unlike 631 iHT-bPBP2b, circumferentially moving aPBP1a was not the majority species at the low 632 expression level (Fig. 1B ). In addition, unlike bPBP2b and RodA, aPBP1a processive 633 motion was not decreased in a Δ murZ mutant (Fig. S17 B and C), which may reflect a 634 different kinetic dependence for Lipid II substrate. Outside of midcell, aPBP1a moved 635 diffusively in the membrane over the whole length of cells (Figs. 8 and 10A). Together, 636 these results indicate that aPBP1a is not a persistent member of the pneumococcal core 637 PG elongasome (85), although shorter or transient interactions are possible (Fig. 11). 638 aPBPs have been proposed to play roles in normal PG synthesis and in repair of 639 damaged PG (1, 39, 73, 115) . In rod- shaped bacteria, it was postulated that when 640 diffusing aPBPs encounter damaged PG, they cease moving and become static for 641 relatively long times ( ≈5 s) (71, 74). For reasons discussed above, long static PG 642 synthases may not be synthesizing PG. We also detected another aPBP1a motion away 643 from midcell in exponentially growing cells. Diffusing aPBP1a molecules infrequently 644 paused for short (<1 s) intervals, before resuming diffusive motion (Movie S28). It remains 645 (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. The copyright holder for this preprintthis version posted January 27, 2024. ; https://doi.org/10.1101/2024.01.10.575112doi: bioRxiv preprint 29 to be determined whether static or paused aPBP1a molecules are repairing damaged PG 646 in S. pneumoniae cells. 647 In Gram-positive bacteria, the PG in the septal annular disk has a core of concentric 648 ordered PG glycan strands covered by a dense mesh of randomly oriented PG strands 649 that face the cell membrane (112). It has been postulated that aPBPs may synthesize this 650 layer of randomly oriented, inner -facing PG strands (39), although remodeling by PG 651 hydrolases followed by resynthesis could potentially contribute to this disordered pattern. 652 Circumferential, processive movement suggests that aPBP1a synthesis may result in 653 concentric, ordered PG glycan strands, rather than randomly oriented strands. Glycan 654 strands synthesized by the shorter tracks of aPBP1a motion are presumably shorter than 655 those synthesized by the more processive core PG elongasome. These shorter strands 656 may be remodeled and crosslinked into the longer strands to provide additional strength 657 to the peripheral PG layer. The dynamics of the other pneumococcal Class A PBPs, 658 aPBP2a and aPBP1b , remains to be determined. The circumferential movement of an 659 aPBP potentially has another implication to bacteria that elongate only from their poles, 660 such as Rhizobiales species, including Agrobacterium tumefaciens (116, 117) . Polar 661 elongasome complexes in these bacteria lack bPBP:SEDS PG synthases, and PG 662 elongation is carried out only by a single essential aPBP (116, 117), which may move 663 circumferentially during PG elongation. 664 Finally, the subdiffusional, confined motion of the MpgA muramidase at midcell was 665 unexpected. IFM showed that MpgA (formerly MltG(Spn)), like aPBP1a, colocalized with 666 MreC to the outer ring in later-divisional cells (80). MpgA is an essential muramidase that 667 cleaves newly synthesized PG glycan chains 7 disaccharides from the undecaprenol 668 (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. The copyright holder for this preprintthis version posted January 27, 2024. ; https://doi.org/10.1101/2024.01.10.575112doi: bioRxiv preprint 30 membrane anchor (82). A LysM domain in MpgA correctly places the cleavage point in 669 the glycan chain. Therefore, MpgA is thought to act as a PG release factor that frees 670 newly synthesized PG for crosslinking into existing PG (82) . Instead of moving 671 circumferentially, most MpgA molecules moved at subdiffusion speeds ( ≈300 nm/sec) 672 that were slower than that of freely diffusing MpgA (>1,000 nm/sec). However, instead of 673 moving over the whole body of cells, MpgA molecules were largely confined to regions of 674 PG synthesis at midcell (Fig s. 1B and 10A). Therefore, MpgA functions separately from 675 the core PG elongasome and from aPBP1a (Fig. 11). Only a fraction ( ≈13%) of MpgA 676 molecules were static, and we detected subdiffusing MpgA molecules that paused (<1 s) 677 and then resumed movement (Movie S30), similar to aPBP1a. Whether static or paused 678 MpgA are carrying out PG strand cleavage requires further study. Likewise, proteins that 679 interact with and regulate MpgA activity (118) and the mechanism that confines MpgA 680 movement to midcell await discovery. 681 682

683 Bacterial strains used were unencapsulated ( Δcps) derivatives of Streptococcus 684 pneumoniae (Spn) serotype 2 strain D39W and are listed in Table S1. IU1945 and IU1824 685 were used as parent strains (Table S1). Detailed methods are described in SI Appendix, 686

, including: bacterial strain construction and growth conditions; 687 ectopic expression and depletion conditions; cell labeling with HaloTag ligand; 2D -688 epifluorescence microscopy and analysis; sm- TIRFm sample preparation; sm- TIRFm 689 imaging; sm-TIRFm image analysis; single-molecule tracking and diffusion analysis; 3D-690 (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. The copyright holder for this preprintthis version posted January 27, 2024. ; https://doi.org/10.1101/2024.01.10.575112doi: bioRxiv preprint 31 structured illumi nation microscopy and analysis; quantitative western blotting; 691 transformation assay. 692 693

694 We thank laboratory members and Jie Xiao (Johns Hopkins) for discussions about this 695 work; John D. Richardson and Ziyun A. Ye for technical assistance with some 696 experiments, Jim Powers (Indiana University Bloomington) for advice about light 697 microscopy; Mike VanNieuwenhze (Indiana University Bloomington) for FDAA reagents; 698 Luke Lavis (Janelia Lab) for Fluor JF549; Reinhold Brückner, and Regine Hakenbeck 699 (Kaiserlautern University) for anti -bPBP2x antibody; and Suzanne Walker and David Z. 700 Rudner (Harvard Medical School) for antibodies against pneumococcal PG synthesis 701 proteins. This work was supported by NIH g rant R35GM131767 (to MEW) , NSF grant 702 MCB1027504 (to SLS); NIH grant RO1AI148752 (to Suzanne Walker); NIH grant T32 703 GM109825 (to AJP); NIH grant F31AI138430 (to MML); NIH grants T32 GM007753 and 704 F30 AI156972 (to JEP), and NIH equipment grant S10OD024988 to the Indiana University 705 Bloomington (IUB) Light Microscopy Imaging Center. 706 707 CONFLICT OF INTEREST 708 The authors declare that they have no conflicts of interests. 709 710 AUTHOR CONTRIBUTIONS 711 AJP, MML, KEB, SLS, and MEW contributed to the conception or design of this study. 712 AJP, MML, KEB, MAT, JEP SLS, HCTT, and MEW contributed to the acquisition, 713 (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. The copyright holder for this preprintthis version posted January 27, 2024. ; https://doi.org/10.1101/2024.01.10.575112doi: bioRxiv preprint 32 analysis, and interpretation of the data. KEB, HCTT, and MEW contributed to the writing 714 of the manuscript with input from the other authors. 715 716 DATA AVAILABILITY 717 The data that support the findings of this study are presented in the paper, including the 718 Supplemental Information and Appendix A. 719 720 ORCID 721 Amilcar J. Perez https://orcid.org/0000-0001-6729-7564 722 Melissa M. Lamanna https://orcid.org/0000-0002-6535-7903 723 Kevin E. Bruce https://orcid.org/0009-0005-3300-8863 724 Julia E. Page https://orcid.org/0000-0002-9884-2382 725 Sidney L. Shaw https://orcid.org/0000-0001-9195-6128 726 Ho-Ching T. Tsui https://orcid.org/0000-0003-0849-874X 727 Malcolm E. Winkler https://orcid.org/0000-0002-1482-2588 728 729

730 1. E. A. Mueller, P. A. Levin, Bacterial Cell Wall Quality Control during Environmental 731 Stress. mBio 11 (2020). 732 2. M. Rajagopal, S. Walker, Envelope Structures of Gram-Positive Bacteria. Curr Top 733 Microbiol Immunol 404, 1-44 (2017). 734 3. E. R. Rojas, The Mechanical Properties of Bacteria and Why they Matter. Adv Exp 735 Med Biol 1267, 1-14 (2020). 736 4. W. Vollmer, O. Massidda, A. Tomasz, The Cell Wall of Streptococcus pneumoniae. 737 Microbiol Spectr 7 (2019). 738 5. K. D. Young, Bacterial shape: two- dimensional questions and possibilities. Annu 739 Rev Microbiol 64, 223-240 (2010). 740 6. K. Bush, P. A. Bradford, Interplay between beta- lactamases and new beta-741 lactamase inhibitors. Nat Rev Microbiol 17, 295-306 (2019). 742 (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. The copyright holder for this preprintthis version posted January 27, 2024. ; https://doi.org/10.1101/2024.01.10.575112doi: bioRxiv preprint 33 7. T. den Blaauwen, J. M. Andreu, O. Monasterio, Bacterial cell division proteins as 743 antibiotic targets. Bioorg Chem 55, 27-38 (2014). 744 8. K. Lewis, The Science of Antibiotic Discovery. Cell 181, 29-45 (2020). 745 9. L. L. Ling et al., A new antibiotic kills pathogens without detectable resistance. 746 Nature 517, 455-459 (2015). 747 10. L. T. Sham, H. C. Tsui, A. D. Land, S. M. Barendt, M. E. Winkler, Recent advances 748 in pneumococcal peptidoglycan biosynthesis suggest new vaccine and 749 antimicrobial targets. Curr Opin Microbiol 15, 194-203 (2012). 750 11. J. Zhou et al., Breaking down the cell wall: Still an attractive antibacterial strategy. 751 Front Microbiol 13, 952633 (2022). 752 12. S. Booth, R. J. Lewis, Structural basis for the coordination of cell division with the 753 synthesis of the bacterial cell envelope. Protein Sci 28, 2042-2054 (2019). 754 13. A. J. F. Egan, J. Errington, W. Vollmer, Regulation of peptidoglycan synthesis and 755 remodelling. Nat Rev Microbiol 18, 446-460 (2020). 756 14. S. Garde, P. K. Chodisetti, M. Reddy, Peptidoglycan: Structure, Synthesis, and 757 Regulation. EcoSal Plus 9 (2021). 758 15. S. Kumar, A. Mollo, D. Kahne, N. Ruiz, The Bacterial Cell Wall: From Lipid II 759 Flipping to Polymerization. Chem Rev 122, 8884-8910 (2022). 760 16. P. D. A. Rohs, T. G. Bernhardt, Growth and Division of the Peptidoglycan Matrix. 761 Annu Rev Microbiol 75, 315-336 (2021). 762 17. C. Billaudeau et al., Contrasting mechanisms of growth in two model rod- shaped 763 bacteria. Nat Commun 8, 15370 (2017). 764 18. P. J. Eswara, K. S. Ramamurthi, Bacterial Cell Division: Nonmodels Poised to 765 Take the Spotlight. Annu Rev Microbiol 71, 393-411 (2017). 766 19. M. M. Lamanna et al., Roles of RodZ and class A PBP1b in the assembly and 767 regulation of the peripheral peptidoglycan elongasome in ovoid- shaped cells of 768 Streptococcus pneumoniae D39. Mol Microbiol 118, 336-368 (2022). 769 20. S. P. Ouellette, J. Lee, J. V. Cox, Division without Binary Fission: Cell Division in 770 the FtsZ-Less Chlamydia. J Bacteriol 202 (2020). 771 21. A. R. Pereira et al., FtsZ-Dependent Elongation of a Coccoid Bacterium. mBio 7 772 (2016). 773 22. M. G. Pinho, M. Kjos, J. W. Veening, How to get (a)round: mechanisms controlling 774 growth and division of coccoid bacteria. Nat Rev Microbiol 11, 601-614 (2013). 775 23. A. M. Randich, Y. V. Brun, Molecular mechanisms for the evolution of bacterial 776 morphologies and growth modes. Front Microbiol 6, 580 (2015). 777 24. N. S. Briggs, K. E. Bruce, S. Naskar, M. E. Winkler, D. I. Roper, The Pneumococcal 778 Divisome: Dynamic Control of Streptococcus pneumoniae Cell Division. Front 779 Microbiol 12, 737396 (2021). 780 25. O. Massidda, L. Novakova, W. Vollmer, From models to pathogens: how much 781 have we learned about Streptococcus pneumoniae cell division? Environ Microbiol 782 15, 3133-3157 (2013). 783 26. M. J. Cox, N. Loman, D. Bogaert, J. O'Grady, Co-infections: potentially lethal and 784 unexplored in COVID-19. Lancet Microbe 1, e11 (2020). 785 27. V. Sender, K. Hentrich, B. Henriques -Normark, Virus -Induced Changes of the 786 Respiratory Tract Environment Promote Secondary Infections With Streptococcus 787 pneumoniae. Front Cell Infect Microbiol 11, 643326 (2021). 788 (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. The copyright holder for this preprintthis version posted January 27, 2024. ; https://doi.org/10.1101/2024.01.10.575112doi: bioRxiv preprint 34 28. J. N. Weiser, D. M. Ferreira, J. C. Paton, Streptococcus pneumoniae: 789 transmission, colonization and invasion. Nat Rev Microbiol 16, 355-367 (2018). 790 29. CDC (2019) Antibiotic resistance threats in the United States, 2019. in AR Threats 791 Rep., U.S. Dep. Health Hum. Serv., Cent. Dis. Control Prev. (AR Threats Report, 792 U.S.A. Dept Health and Human Services, Center for Disease Control and 793 Prevention, Atlanta, GA). 794 30. P. S. Gibson, J. W. Veening, Gaps in the wall: understanding cell wall biology to 795 tackle amoxicillin resistance in Streptococcus pneumoniae. Curr Opin Microbiol 796 72, 102261 (2023). 797 31. WHO (2017) List of bacteria for which new antibiotics are urgently needed. (World 798 Health Organization, Geneva, Switzerland). 799 32. A. J. Perez et al., Organization of peptidoglycan synthesis in nodes and separate 800 rings at different stages of cell division of Streptococcus pneumoniae. Mol 801 Microbiol 115, 1152-1169 (2021). 802 33. J. Trouve et al., Nanoscale dynamics of peptidoglycan assembly during the cell 803 cycle of Streptococcus pneumoniae. Curr Biol 31, 2844-2856 e2846 (2021). 804 34. A. J. Perez et al., Movement dynamics of divisome proteins and PBP2x:FtsW in 805 cells of Streptococcus pneumoniae. Proc Natl Acad Sci U S A 116, 3211-3220 806 (2019). 807 35. H. T. Tsui et al., Pbp2x localizes separately from Pbp2b and other peptidoglycan 808 synthesis proteins during later stages of cell division of Streptococcus pneumoniae 809 D39. Mol Microbiol 94, 21-40 (2014). 810 36. J. Errington, L. J. Wu, Cell Cycle Machinery in Bacillus subtilis. Subcell Biochem 811 84, 67-101 (2017). 812 37. V. A. Lund et al., Molecular coordination of Staphylococcus aureus cell division. 813 Elife 7 (2018). 814 38. B. M. Saraiva et al., Reassessment of the distinctive geometry of Staphylococcus 815 aureus cell division. Nat Commun 11, 4097 (2020). 816 39. D. Straume, K. W. Piechowiak, M. Kjos, L. S. Havarstein, Class A PBPs: It is time 817 to rethink traditional paradigms. Mol Microbiol 116, 41-52 (2021). 818 40. R. Wheeler, S. Mesnage, I. G. Boneca, J. K. Hobbs, S. J. Foster, Super-resolution 819 microscopy reveals cell wall dynamics and peptidoglycan architecture in ovococcal 820 bacteria. Mol Microbiol 82, 1096-1109 (2011). 821 41. A. J. Perez et al. , FtsZ -Ring Regulation and Cell Division Are Mediated by 822 Essential EzrA and Accessory Proteins ZapA and ZapJ in Streptococcus 823 pneumoniae. Front Microbiol 12, 780864 (2021). 824 42. A. Fleurie et al. , MapZ marks the division sites and positions FtsZ rings in 825 Streptococcus pneumoniae. Nature 516, 259-262 (2014). 826 43. N. Holeckova et al., LocZ is a new cell division protein involved in proper septum 827 placement in Streptococcus pneumoniae. mBio 6, e01700-01714 (2014). 828 44. V. W. Rowlett, W. Margolin, The Min system and other nucleoid- independent 829 regulators of Z ring positioning. Front Microbiol 6, 478 (2015). 830 45. L. J. Wu, J. Errington, Nucleoid occlusion and bacterial cell division. Nat Rev 831 Microbiol 10, 8-12 (2011). 832 46. A. D. Land et al., Requirement of essential Pbp2x and GpsB for septal ring closure 833 in Streptococcus pneumoniae D39. Mol Microbiol 90, 939-955 (2013). 834 (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. The copyright holder for this preprintthis version posted January 27, 2024. ; https://doi.org/10.1101/2024.01.10.575112doi: bioRxiv preprint 35 47. R. van Raaphorst, M. Kjos, J. W. Veening, Chromosome segregation drives 835 division site selection in Streptococcus pneumoniae. Proc Natl Acad Sci U S A 836 114, E5959-E5968 (2017). 837 48. M. Jacq et al., Remodeling of the Z-Ring Nanostructure during the Streptococcus 838 pneumoniae Cell Cycle Revealed by Photoactivated Localization Microscopy. 839 mBio 6 (2015). 840 49. A. Mura et al., Roles of the Essential Protein FtsA in Cell Growth and Division in 841 Streptococcus pneumoniae. J Bacteriol 199 (2017). 842 50. P. A. Levin, A. Janakiraman, Localization, Assembly, and Activation of the 843 Escherichia coli Cell Division Machinery. EcoSal Plus 9, eESP00222021 (2021). 844 51. C. R. Mahone, E. D. Goley, Bacterial cell division at a glance. J Cell Sci 133 (2020). 845 52. L. M. Y. Mertens, T. den Blaauwen, Optimising expression of the large dynamic 846 range FRET pair mNeonGreen and superfolder mTurquoise2(ox) for use in the 847 Escherichia coli cytoplasm. Sci Rep 12, 17977 (2022). 848 53. A. W. Bisson -Filho et al. , Treadmilling by FtsZ filaments drives peptidoglycan 849 synthesis and bacterial cell division. Science 355, 739-743 (2017). 850 54. X. Yang et al., A two-track model for the spatiotemporal coordination of bacterial 851 septal cell wall synthesis revealed by single- molecule imaging of FtsW. Nat 852 Microbiol 6, 584-593 (2021). 853 55. S. Schäper et al., Processive movement of Staphylococcus aureus 854 essential septal peptidoglycan synthases is independent of FtsZ treadmilling and 855 drives cell constriction. bioRxiv 10.1101/2023.06.29.547026 (2023). 856 56. K. D. Whitley, J. Grimshaw, E. Karinou, P. J. Stansfeld, S. Holden, A one -track 857 model for spatiotemporal coordination of Bacillus subtilis septal cell 858 wall synthesis. bioRxiv 10.1101/2023.06.29.547024, 2023.2006.2029.547024 859 (2023). 860 57. J. Dominguez-Escobar et al., Processive movement of MreB -associated cell wall 861 biosynthetic complexes in bacteria. Science 333, 225-228 (2011). 862 58. E. C. Garner et al. , Coupled, circumferential motions of the cell wall synthesis 863 machinery and MreB filaments in B. subtilis. Science 333, 222-225 (2011). 864 59. S. van Teeffelen et al., The bacterial actin MreB rotates, and rotation depends on 865 cell-wall assembly. Proc Natl Acad Sci U S A 108, 15822-15827 (2011). 866 60. B. P. Bratton, J. W. Shaevitz, Z. Gitai, R. M. Morgenstein, MreB polymers and 867 curvature localization are enhanced by RodZ and predict E. coli's cylindrical 868 uniformity. Nat Commun 9, 2797 (2018). 869 61. S. Hussain et al. , MreB filaments align along greatest principal membrane 870 curvature to orient cell wall synthesis. Elife 7, e32471 (2018). 871 62. R. M. Morgenstein et al., RodZ links MreB to cell wall synthesis to mediate MreB 872 rotation and robust morphogenesis. Proc Natl Acad Sci U S A 112, 12510-12515 873 (2015). 874 63. R. Ago, D. Shiomi, RodZ: a key -player in cell elongation and cell division in 875 Escherichia coli. AIMS Microbiol 5, 358-367 (2019). 876 64. F. O. Bendezu, C. A. Hale, T. G. Bernhardt, P. A. de Boer, RodZ (YfgA) is required 877 for proper assembly of the MreB actin cytoskeleton and cell shape in E. coli. EMBO 878 J 28, 193-204 (2009). 879 (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. The copyright holder for this preprintthis version posted January 27, 2024. ; https://doi.org/10.1101/2024.01.10.575112doi: bioRxiv preprint 36 65. F. van den Ent, C. M. Johnson, L. Persons, P. de Boer, J. Lowe, Bacterial actin 880 MreB assembles in complex with cell shape protein RodZ. EMBO J 29, 1081-1090 881 (2010). 882 66. A. J. Meeske et al., SEDS proteins are a widespread family of bacterial cell wall 883 polymerases. Nature 537, 634-638 (2016). 884 67. P. D. A. Rohs et al., A central role for PBP2 in the activation of peptidoglycan 885 polymerization by the bacterial cell elongation machinery. PLoS Genet 14, 886 e1007726 (2018). 887 68. P. D. A. Rohs et al., Identification of potential regulatory domains within the MreC 888 and MreD components of the cell elongation machinery. J Bacteriol 889 10.1128/JB.00493-20 (2021). 890 69. M. Sjodt et al. , Structure of the peptidoglycan polymerase RodA resolved by 891 evolutionary coupling analysis. Nature 556, 118-121 (2018). 892 70. M. Sjodt et al. , Structural coordination of polymerization and crosslinking by a 893 SEDS-bPBP peptidoglycan synthase complex. Nat Microbiol 5, 813-820 (2020). 894 71. H. Cho et al. , Bacterial cell wall biogenesis is mediated by SEDS and PBP 895 polymerase families functioning semi -autonomously. Nat Microbiol 1, 16172 896 (2016). 897 72. M. F. Dion et al. , Bacillus subtilis cell diameter is determined by the opposing 898 actions of two distinct cell wall synthetic systems. Nat Microbiol 4, 1294- 1305 899 (2019). 900 73. M. Pazos, W. Vollmer, Regulation and function of class A Penicillin- binding 901 proteins. Curr Opin Microbiol 60, 80-87 (2021). 902 74. A. Vigouroux et al. , Class -A penicillin binding proteins do not contribute to cell 903 shape but repair cell-wall defects. Elife 9, e51998 (2020). 904 75. K. H. Berg, G. A. Stamsas, D. Straume, L. S. Havarstein, Effects of low PBP2b 905 levels on cell morphology and peptidoglycan composition in Streptococcus 906 pneumoniae R6. J Bacteriol 195, 4342-4354 (2013). 907 76. J. Philippe, T. Vernet, A. Zapun, The elongation of ovococci. Microb Drug Resist 908 20, 215-221 (2014). 909 77. G. A. Stamsas et al. , Identification of EloR (Spr1851) as a regulator of cell 910 elongation in Streptococcus pneumoniae. Mol Microbiol 105, 954-967 (2017). 911 78. D. Straume, G. A. Stamsas, K. H. Berg, Z. Salehian, L. S. Havarstein, Identification 912 of pneumococcal proteins that are functionally linked to penicillin- binding protein 913 2b (PBP2b). Mol Microbiol 103, 99-116 (2017). 914 79. J. J. Zheng, A. J. Perez, H. T. Tsui, O. Massidda, M. E. Winkler, Absence of the 915 KhpA and KhpB (JAG/EloR) RNA-binding proteins suppresses the requirement for 916 PBP2b by overproduction of FtsA in Streptococcus pneumoniae D39. Mol 917 Microbiol 106, 793-814 (2017). 918 80. H. C. Tsui et al., Suppression of a deletion mutation in the gene encoding essential 919 PBP2b reveals a new lytic transglycosylase involved in peripheral peptidoglycan 920 synthesis in Streptococcus pneumoniae D39. Mol Microbiol 100, 1039- 1065 921 (2016). 922 81. A. D. Land, M. E. Winkler, The requirement for pneumococcal MreC and MreD is 923 relieved by inactivation of the gene encoding PBP1a. J Bacteriol 193, 4166-4179 924 (2011). 925 (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. The copyright holder for this preprintthis version posted January 27, 2024. ; https://doi.org/10.1101/2024.01.10.575112doi: bioRxiv preprint 37 82. A. Taguchi, J. E. Page, H. T. Tsui, M. E. Winkler, S. Walker, Biochemical 926 reconstitution defines new functions for membrane- bound glycosidases in 927 assembly of the bacterial cell wall. Proc Natl Acad Sci U S A 118 (2021). 928 83. A. K. Fenton, L. El Mortaji, D. T. Lau, D. Z. Rudner, T. G. Bernhardt, CozE is a 929 member of the MreCD complex that directs cell elongation in Streptococcus 930 pneumoniae. Nat Microbiol 2, 16237 (2016). 931 84. M. M. Lamanna, A. T. Maurelli, What Is Motion? Recent Advances in the Study of 932 Molecular Movement Patterns of the Peptidoglycan Synthesis Machines. J 933 Bacteriol 204, e0059821 (2022). 934 85. C. Lenoir et al. , The morphogenic protein CopD controls the spatio- temporal 935 dynamics of PBP1a and PBP2b in Streptococcus pneumoniae. mBio 14, 936 e0141123 (2023). 937 86. S. G. Bartual et al. , Structural basis of PcsB -mediated cell separation in 938 Streptococcus pneumoniae. Nature communications 5, 3842 (2014). 939 87. B. E. Rued et al. , Structure of the Large Extracellular Loop of FtsX and Its 940 Interaction with the Essential Peptidoglycan Hydrolase PcsB in Streptococcus 941 pneumoniae. mBio 10 (2019). 942 88. L. T. Sham, S. M. Barendt, K. E. Kopecky, M. E. Winkler, Essential PcsB putative 943 peptidoglycan hydrolase interacts with the essential FtsXSpn cell division protein 944 in Streptococcus pneumoniae D39. Proc Natl Acad Sci U S A 108, E1061-1069 945 (2011). 946 89. L. T. Sham, K. R. Jensen, K. E. Bruce, M. E. Winkler, Involvement of FtsE ATPase 947 and FtsX extracellular loops 1 and 2 in FtsEX -PcsB complex function in cell 948 division of Streptococcus pneumoniae D39. mBio 4 (2013). 949 90. P. Domínguez-Cuevas, I. Porcelli, R. A. Daniel, J. Errington, Differentiated roles 950 for MreB-actin isologues and autolytic enzymes in Bacillus subtilis morphogenesis. 951 Mol Microbiol 89, 1084-1098 (2013). 952 91. J. Meisner et al. , FtsEX is required for CwlO peptidoglycan hydrolase activity 953 during cell wall elongation in Bacillus subtilis. Mol Microbiol 89, 1069-1083 (2013). 954 92. M. J. Catalão, J. Figueiredo, M. X. Henriques, J. P. Gomes, S. R. Filipe, 955 Optimization of fluorescent tools for cell biology studies in Gram-positive bacteria. 956 PLoS One 9, e113796 (2014). 957 93. J. B. Grimm et al. , A general method to improve fluorophores for live- cell and 958 single-molecule microscopy. Nat Methods 12, 244- 250, 243 p following 250 959 (2015). 960 94. M. X. Henriques, M. J. Catalão, J. Figueiredo, J. P. Gomes, S. R. Filipe, 961 Construction of improved tools for protein localization studies in Streptococcus 962 pneumoniae. PLoS One 8, e55049 (2013). 963 95. G. V. Los et al., HaloTag: a novel protein labeling technology for cell imaging and 964 protein analysis. ACS Chem Biol 3, 373-382 (2008). 965 96. K. Peters et al., Streptococcus pneumoniae PBP2x mid- cell localization requires 966 the C-terminal PASTA domains and is essential for cell shape maintenance. Mol 967 Microbiol 92, 733-755 (2014). 968 97. N. Ke, D. Landgraf, J. Paulsson, M. Berkmen, Visualization of Periplasmic and 969 Cytoplasmic Proteins with a Self-Labeling Protein Tag. J Bacteriol 198, 1035-1043 970 (2016). 971 (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. The copyright holder for this preprintthis version posted January 27, 2024. ; https://doi.org/10.1101/2024.01.10.575112doi: bioRxiv preprint 38 98. A. Eberhardt, L. J. Wu, J. Errington, W. Vollmer, J. W. Veening, Cellular localization 972 of choline -utilization proteins in Streptococcus pneumoniae using novel 973 fluorescent reporter systems. Mol Microbiol 74, 395-408 (2009). 974 99. F. E. Jacobsen, K. M. Kazmierczak, J. P. Lisher, M. E. Winkler, D. P. Giedroc, 975 Interplay between manganese and zinc homeostasis in the human pathogen 976 Streptococcus pneumoniae. Metallomics : integrated biometal science 3, 38 -41 977 (2011). 978 100. J. W. McCausland et al. , Treadmilling FtsZ polymers drive the directional 979 movement of sPG -synthesis enzymes via a Brownian ratchet mechanism. Nat 980 Commun 12, 609 (2021). 981 101. H. T. Tsui et al., Negative regulation of MurZ and MurA underlies the essentiality 982 of GpsB - and StkP -mediated protein phosphorylation in Streptococcus 983 pneumoniae D39. Mol Microbiol 120, 351-383 (2023). 984 102. O. Kocaoglu, H. C. Tsui, M. E. Winkler, E. E. Carlson, Profiling of β -lactam 985 selectivity for penicillin -binding proteins in Streptococcus pneumoniae D39. 986 Antimicrob Agents Chemother 59, 3548-3555 (2015). 987 103. K. H. Berg, D. Straume, L. S. Havarstein, The function of the transmembrane and 988 cytoplasmic domains of pneumococcal penicillin- binding proteins 2x and 2b 989 extends beyond that of simple anchoring devices. Microbiology (Reading) 160, 990 1585-1598 (2014). 991 104. H. Cho, T. Uehara, T. G. Bernhardt, Beta- lactam antibiotics induce a lethal 992 malfunctioning of the bacterial cell wall synthesis machinery. Cell 159, 1300-1311 993 (2014). 994 105. I. Shlosman et al. , Allosteric activation of cell wall synthesis during bacterial 995 growth. Nat Commun 14, 3439 (2023). 996 106. N. Tarantino et al., TNF and IL-1 exhibit distinct ubiquitin requirements for inducing 997 NEMO-IKK supramolecular structures. J Cell Biol 204, 231-245 (2014). 998 107. J. Y. Tinevez et al., TrackMate: An open and extensible platform for single-particle 999 tracking. Methods 115, 80-90 (2017). 1000 108. T. K. Lee et al., A dynamically assembled cell wall synthesis machinery buffers cell 1001 growth. Proc Natl Acad Sci U S A 111, 4554-4559 (2014). 1002 109. G. Ozbaykal et al. , The transpeptidase PBP2 governs initial localization and 1003 activity of the major cell-wall synthesis machinery in E. coli. Elife 9 (2020). 1004 110. C. H. Bohrer, J. Xiao, Complex Diffusion in Bacteria. Adv Exp Med Biol 1267, 15-1005 43 (2020). 1006 111. A. Kusumi, Y. Sako, M. Yamamoto, Confined lateral diffusion of membrane 1007 receptors as studied by single particle tracking (nanovid microscopy). Effects of 1008 calcium-induced differentiation in cultured epithelial cells. Biophys J 65, 2021-2040 1009 (1993). 1010 112. L. Pasquina-Lemonche et al., The architecture of the Gram -positive bacterial cell 1011 wall. Nature 582, 294-297 (2020). 1012 113. S. Middlemiss et al., Molecular motor tug- of-war regulates elongasome cell wall 1013 synthesis dynamics in Bacillus subtilis. bioRxiv 1014 10.1101/2023.05.10.540107, 2023.2005.2010.540107 (2023). 1015 (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. The copyright holder for this preprintthis version posted January 27, 2024. ; https://doi.org/10.1101/2024.01.10.575112doi: bioRxiv preprint 39 114. Z. Lyu et al. , FtsN maintains active septal cell wall synthesis by forming a 1016 processive complex with the septum -specific peptidoglycan synthases in E. coli. 1017 Nat Commun 13, 5751 (2022). 1018 115. D. Straume et al., Class A PBPs have a distinct and unique role in the construction 1019 of the pneumococcal cell wall. Proc Natl Acad Sci U S A 117, 6129-6138 (2020). 1020 116. W. M. Figueroa- Cuilan, P. J. B. Brown, Cell Wall Biogenesis During Elongation 1021 and Division in the Plant Pathogen Agrobacterium tumefaciens. Curr Top Microbiol 1022 Immunol 418, 87-110 (2018). 1023 117. M. A. Williams et al., Unipolar Peptidoglycan Synthesis in the Rhizobiales Requires 1024 an Essential Class A Penicillin-Binding Protein. mBio 12, e0234621 (2021). 1025 118. A. R. Winther, M. Kjos, M. L. Herigstad, L. S. Havarstein, D. Straume, EloR 1026 interacts with the lytic transglycosylase MltG at midcell in Streptococcus 1027 pneumoniae R6. J Bacteriol 10.1128/JB.00691-20 (2021). 1028 1029 FIGURE LEGENDS 1030 Fig. 1. Elongation PG synthesis proteins display processive circumferential motion, 1031 and a limited number of molecules engage in PG synthesis simultaneously. Sm-TIRFm 1032 was performed at 1 FPS as described in Materials and Methods on strains iht-pbp2b 1033 (IU15928), iht-pbp2b // PZn-iht-pbp2b (IU16553), iht-rodA (IU15970), iht-rodA // PZn-iht-1034 rodA (IU16496), iht-mreC (IU16344), iht-pbp1a (IU16320), iht-pbp1a // P Zn-iht-pbp1a 1035 (IU16497) and iht -mpgA (IU15997). (A) Representative time- lapse images and 1036 kymographs of molecules displaying processive circumferential movement in strains 1037 IU15928, IU15970, IU16344 and IU16497. Cell outlines are green (DICm) and HT-labeled 1038 molecules are magenta. Location of lines used to make kymographs are shown in the 1039 time lapse images (yellow dashed lines) . F or each protein, the left kymograph was 1040 generated from the cell depicted in the time -lapse images. Numbers (1-3) denote when 1041 each image in the time lapse w as taken. The other two kymographs are from cells that 1042 are not shown. Calculation of circumferential velocities from kymographs are described 1043 in Materials and Methods . Scale bars are 1 µm. (B) Movement patterns of HT- labeled 1044 molecules. Bars represent the mean relative frequency of each type of motion. For a given 1045 (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. The copyright holder for this preprintthis version posted January 27, 2024. ; https://doi.org/10.1101/2024.01.10.575112doi: bioRxiv preprint 40 strain or condition, the r elative frequencies of circumferential, diffusive, and static 1046 molecules were determined for each biological replicate. Frequencies were then 1047 averaged over all replicates to determine the mean ( error bars represent ± SD). Mean 1048 values are annotated above each bar, and the total number of molecules analyzed (n) 1049 from the number of biological replicates (R) for each strain is below each strain or 1050 condition. Circumferential (red), diffusive (blue), and static (grey) movement patterns are 1051 shown. Circumferentially moving molecules were defined as molecules moving in one 1052 direction for 6 or more frames (at 1 FPS imaging rate) with a linear velocity ≥ 5 nm/s. 1053 Static molecules were defined as molecules not moving or moving for 6 or more frames 1054 with a velocity < 5 nm/s. Diffusive molecules were defined as molecules that moved, but 1055 not in a consistent direction, for 6 or more non- consecutive frames within a period of 90 1056 s. Full criteria for determining movement patterns are described in Materials and Methods. 1057 The expression level of iHT-fusion proteins relative to the untagged WT level is shown as 1058 a percentage under the strain names. iHT-fusion proteins are expressed solely from the 1059 native chromosomal locus or with an additional copy of the gene encoding the iHT-fusion 1060 protein at an ectopic site under control of a zinc-inducible promoter, with 0 or 0.25 mM Zn 1061 inducer added. Unpaired t -tests were performed to compare relative frequencies of 1062 motion types. ns (nonsignificant); *P < 0.05; **P < 0.01; ***P < 0.001; ****P < 0.0001. 1063 Fig. 2. Diffusively moving components of the core PG elongasome do not actively 1064 synthesize PG (blue) in non-midcell regions of exponentially growing pneumococcal cells. 1065 3D-SIM images, where iHT-bPBP2b is red, regions of PBP transpeptidase activity is blue, 1066 and midcell regions include septa of dividing cells and equators of predivisional daughter 1067 cells. Cells were grown in BHI ± 0.3 mM Zn inducer, labeled with a saturating amount of 1068 (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. The copyright holder for this preprintthis version posted January 27, 2024. ; https://doi.org/10.1101/2024.01.10.575112doi: bioRxiv preprint 41 HT-ligand for 15 min, washed, labeled with 400 µM of the fluorescent D-amino acid HADA 1069 (blue) for 2.5 min, fixed, and imaged as described in Materials and Methods. Images are 1070 summed from 15 Z plane sections ( Z-stack) representing a total depth of 1.875 µm. 1071 Percentages indicate the amount of iHT-bPBP2b expressed relative to the untagged WT 1072 bPBP2b+ level (32). (A) iht-pbp2b (IU15928) and (B) iht-pbp2b // PZn-iht-pbp2b (IU16553). 1073 Scale bar is 1 µm. 1074 Fig. 3. bPBP2b, RodA, and MreC form a stable complex during active PG synthesis. 1075 Dot plots of (A) circumferential velocities and (B) circumferential durations of elongation 1076 PG synthesis proteins determined by sm-TIRFm at 1 FPS. Strains are the same as in Fig. 1077 1. +Zn condition indicates 0.25 mM Zn inducer was added. Black and red lines are median 1078 ± interquartile, and mean ± SD are indicated. n = total molecules analyzed from 2- 5 1079 biological replicates. Dotted grey lines indicate the minimum thresholds for (A) velocity (5 1080 nm/s) and (B) duration (6 s). A Kruskal-Wallis with Dunn’s multiple comparisons test was 1081 used to compare circumferential velocities or circumferential durations in different strains. 1082 ns (nonsignificant); **P < 0.01; ***P < 0.001; ****P < 0.0001. 1083 Fig. 4. Circumferential movement of bPBP2b and RodA is independent of FtsZ 1084 treadmilling and reflective of PG synthesis. Sm-TIRFm was performed at 1 FPS on strains 1085 expressing iHT-bPBP2b or iHT-RodA. Dot plots of circumferential velocities are shown. 1086 Black and red lines are median ± interquartile, and mean ± SD are indicated. (A) Both iht-1087 pbp2b (black, IU15928) and iht-pbp2b P Zn-ftsZ(D214A) (red, IU16091) strains were 1088 supplemented with 0.25 mM Zn inducer. (B) IU16056 expresses both iHT- bPBP2b and 1089 FtsZ-sfGFP. The treadmilling velocity of FtsZ- sfGFP filaments (green) and the 1090 circumferential velocity of iHT -bPBP2b single molecules (red) were determined. (C) iht-1091 (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. The copyright holder for this preprintthis version posted January 27, 2024. ; https://doi.org/10.1101/2024.01.10.575112doi: bioRxiv preprint 42 pbp2b (black, IU15928) and iht-pbp2b ∆murZ (red, IU16110) strains. (D) iht-rodA (black, 1092 IU15970) and iht-rodA ∆murZ (blue, IU16112) strains. n = total molecules analyzed from 1093 2 biological replicates. Circumferential velocities in different strains were compared using 1094 a Mann-Whitney test. *P < 0.05; ****P < 0.0001. 1095 Fig. 5. The catalytic activities of bPBP2b and RodA are required for cell viability and 1096 PG synthesis, but not localization. (A) Diagram of merodiploid strains containing catalytic 1097 mutant alleles fused to the HaloTag at the native locus and the Zn-inducible WT allele at 1098 the ectopic site. ( B) Representative growth curves of WT ( black square, IU1824), iht-1099 pbp2b (red square, IU15928), iht -rodA (blue square, IU15970), ∆ pbp2b // P Zn-pbp2b 1100 (green, IU11258), iht-pbp2b(S391A) // PZn-pbp2b (red, IU16232), and iht-rodA(D283A) // 1101 PZn-rodA (blue, IU16239) strains grown in C+Y medium with or without 0.2 mM Zn inducer 1102 as described in the section on ectopic expression and depletion conditions in Materials 1103 and Methods. The experiment was repeated twice with similar results . (C and D) sm-1104 TIRFm was performed at 1 FPS on strains IU16232 and IU16239. (C) Movement patterns 1105 of iHT-bPBP2b(S391A) (IU16232) and iHT-RodA(D283A) (IU16239). The layout is the 1106 same as Fig. 1B (see legend for details ). Unpaired t-tests were performed to compare 1107 relative frequencies of motion types. ns (nonsignificant); **P < 0.01; ***P < 0.001; ****P < 1108 0.0001. (D) Representative montages and kymographs of static molecules from strains 1109 IU16232 and IU16239 (without Zn inducer for 3 h) are displayed as described for Fig. 1A, 1110 except the scale bar is 0.5 µm. ( E) 2D -FM of catalytically inactive proteins. Strains 1111 IU16232 and IU16239 were grown without Zn (see Fig. 5B) and labeled with saturating 1112 amounts of HT-ligand. HT-fusion proteins are shown in magenta. Scale bar is 1 µm. The 1113 experiment was repeated twice with similar results. 1114 (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. The copyright holder for this preprintthis version posted January 27, 2024. ; https://doi.org/10.1101/2024.01.10.575112doi: bioRxiv preprint 43 Fig. 6. bPBP2b localization becomes diffuse and circumferential movement is lost 1115 upon MreC depletion. (A) 2D-FM of strains iht-pbp2b (left, IU15928) and iht-pbp2b ∆mreC 1116 // PZn-mreC (right, IU16281). Cells were grown in C+Y media with or without (depletion) 1117 0.2 mM Zn inducer, labeled with saturating amounts of HT-ligand, and imaged after 3 h. 1118 See the section on ectopic expression and depletion conditions in Materials and Methods 1119 for details. The experiment was repeated twice with similar results. (B) sm-TIRFm was 1120 performed at 1 FPS on strains iht-pbp2b (IU15928) and iht-pbp2b ∆mreC // PZn-mreC 1121 (IU16281) after 3 h of growth in C+Y media without Zn inducer (depletion). The layout is 1122 the same as Fig. 1B (see legend for details). Unpaired t-tests were performed to compare 1123 relative frequencies of motion types. ns (nonsignificant); *P < 0.05; **P < 0.01. 1124 Fig. 7. bPBP2b and RodA require each other for localization and motion. 2D-FM and 1125 sm-TIRFm at 1 FPS were performed on strains iht-pbp2b ∆rodA // PZn-rodA (IU16204) 1126 and iht-rodA ∆pbp2b // PZn-pbp2b (IU16202). (A and B) 2D-FM showing localization in 1127 cells of strains (A) iHT-bPBP2b (IU16204) and (B) iHT-RodA (IU16202) grown with or 1128 without (depletion; depl.) 0.2 mM Zn inducer for 3 h and labeled with saturating amounts 1129 of HT-ligand. See the section on ectopic expression and depletion conditions in Materials 1130 and Methods for additional details. The experiment was repeated twice with similar 1131 results. (C to E) sm-TIRFm was performed at 1 FPS on IU16204 and IU16202 without Zn 1132 inducer (depletion) for 3 hours. (C) Movement patterns of HT-labeled molecules. Results 1133 for the iht-pbp2b ∆rodA // PZn-rodA (IU16204) and iht-rodA ∆pbp2b // PZn-pbp2b (IU16202) 1134 strains were determined and are graphed with data for iht-pbp2b (IU15928) and iht-rodA 1135 (IU15970) replotted from Fig. 1B for comparison. The layout is the same as Fig. 1B (see 1136 legend for details ). Unpaired t-tests were performed to compare relative frequencies of 1137 (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. The copyright holder for this preprintthis version posted January 27, 2024. ; https://doi.org/10.1101/2024.01.10.575112doi: bioRxiv preprint 44 motion types. ns (nonsignificant), *P < 0.05; **P < 0.01; ***P < 0.001. (D and E) Dot plots 1138 of circumferential velocities with mean values ± SD indicated for ( D) strains IU15928 1139 (black) and IU16204 (red), and (E) strains IU15970 (black) and IU16202 (red). Data from 1140 stains IU15928 and IU15970 are replotted from Fig. 3A for comparison. Black and red 1141 lines are median ± interquartile, and mean ± SD are indicated. n = total molecules 1142 analyzed from 2 biological replicates. A Mann-Whitney test was used to c ompare 1143 velocities upon protein depletion (depl.). ns (nonsignificant). 1144 Fig. 8. Circumferential movement of aPBP1a reflects active PG synthesis. Sm-TIRFm 1145 was performed at 1 FPS as described in Materials and Methods on strains iht-pbp1a // 1146 PZn-iht-pbp1a (IU16497) and iht-pbp1a(S370A) // P Zn-iht-pbp1a(S370A) (IU19168) with 1147 0.25 mM Zn added. Movement patterns of HT-labeled molecules at sites of PG synthesis 1148 (midcell) or elsewhere in the cell (non-midcell) were analyzed. The layout is the same as 1149 Fig. 1B (see legend for details). A two-way ANOVA with Tukey’s multiple comparison test 1150 was used to compare relative frequencies of motion types among strains. *P < 0.05. 1151 Fig. 9. MpgA displays confined subdiffusive movement at midcell. Sm-TIRFm was 1152 performed on strains iht-mpgA (IU15997) and iht-mreC (IU16344). Representative 1153 kymographs of (A) iHT-MpgA molecules moving at sites of PG synthesis imaged at 1 FPS 1154 and (B) iHT-MpgA and iHT-MreC molecules imaged at 10 FPS. (C) Dot plots of velocities 1155 determined at 10 FPS over short runs of iHT -MpgA molecules in midcell regions (black) 1156 and at non-midcell regions (red). Black and red lines are median ± interquartile, and mean 1157 ± SD are indicated. n = total molecules analyzed from 2 biological replicates. Velocities 1158 in different regions of cells were compared using a Mann-Whitney test. ****P < 0.0001. 1159 (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. The copyright holder for this preprintthis version posted January 27, 2024. ; https://doi.org/10.1101/2024.01.10.575112doi: bioRxiv preprint 45 Fig. 10. Components of the pneumococcal core PG elongasome and proteins 1160 implicated in elongation PG synthesis exhibit different patterns of confined diffusion. Sm-1161 TIRFm was performed at 20 FPS on strains iht-pbp2b // PZn-iht-pbp2b (IU16553), iht-rodA 1162 // PZn-iht-rodA (IU16496), iht-mreC (IU16344), iht-pbp1a // PZn-iht-pbp1a (IU16497) and 1163 iht-mpgA (IU15997). 0.25 mM Zn inducer was added to strains IU16553, IU16496 and 1164 IU16497. ( A) Representative fields of cells containing multiple single -molecule 1165 trajectories (colored lines) classified as diffusive (top row) and non-diffusive (bottom row). 1166 Trajectories with a displacement > 0.13 µm and a velocity standard deviation > 0.63 µm/s 1167 were defined as diffusive, and the remaining trajectories were defined as non- diffusive. 1168 Full details of trajectory constructions and classification criteria are described in Materials 1169 and Methods. The color of each trajectory line segment represents the displacement (in 1170 nm) of the molecule from one frame to the next. Scale bar is 1 µm. ( B) Mean square 1171 displacements (MSD) were calculated for diffusive and non-diffusive trajectories. Circles 1172 show MSD values, and error bars represent SEM. Lines represent MSD curves fit to the 1173 data as described in Materials and Methods . To the right of the graph are the mean 1174 diffusion coefficients (± SEM) of diffusive trajectories, alpha values of diffusive trajectories 1175 (α), the percent of total trajectories that were classified as diffusive (% diffusive), and the 1176 total number of trajectories (n) analyzed from 2 biological replicates. α values < 1 indicate 1177 that molecules are exhibiting confined diffusion (or subdiffusion), discussed in the text. 1178 The diffusion coefficients were compared between strains using a Brown- Forsythe and 1179 Welch’s ANOVA with a Games-Howell multiple comparisons test. **P < 0.01. 1180 Fig. 11. Summary model of ( A) the composition and dynamics of core elongasome 1181 components (bPBP2b, RodA, and MreC) and proteins linked to PG elongation synthesis 1182 (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. The copyright holder for this preprintthis version posted January 27, 2024. ; https://doi.org/10.1101/2024.01.10.575112doi: bioRxiv preprint 46 (aPBP1a and MpgA) i n growing S. pneumoniae cells; and ( B) the distribution of these 1183 proteins when underexpressed (left) or expressed at WT levels (right). ( A) The 1184 circumferential, processive movement of bPBP 2b, RodA, and MreC reported here likely 1185 apply to the other core elongasome components RodZ and MreD. In early, predivisional 1186 cells, the core PG elongasome is located in the FtsZ- ring at the equator midcell of a 1187 daughter cell. In later divisional cells, the core PG elongasome is in the outer ring of the 1188 midcell septal annular disk. The motion of MreC is the same in early or late divisional 1189 cells. Class A aPBP1a also displays circumferential, processive motion at midcell that 1190 depends on PG synthesis, but its dynamics indicate that aPBP1a is not a persistent 1191 member of the core PG elongasome, although short interactions are possible. In contrast, 1192 MpgA moves in a distinctive type of subdiffusion that is largely confined to the midcell 1193 region. Components of the core elongasome and aPBP1a move diffusively outside of 1194 midcell rings. There are also nonmoving, static molecules (> 6 s) of elongasome proteins 1195 and aPBP1a, mainly in midcell regions. It is not clear whether static molecules are 1196 synthesizing PG or are in a non- synthesizing transition state. In addition, diffusing 1197 molecules of aPBP1a and MpgA were infrequently detected pausing (< 1s) before 1198 resuming motion. See text for additional information. ( B) When underexpressed, 1199 components of the core PG elongasome are largely confined to PG synthesis at midcell 1200 septa and the equators of predivisional daughter cells starting to divide. In contrast, when 1201 expressed at WT levels, most components of the PG elongasome (and aPBP1a ) are 1202 diffusing over the cell surface and are not synthesizing PG. Thus, elongasome 1203 components and aPBP1a are in excess in growing cells, and only a limited number 1204 engage in midcell PG synthesis. In contrast, subdiffusion of MpgA molecules expressed 1205 (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. The copyright holder for this preprintthis version posted January 27, 2024. ; https://doi.org/10.1101/2024.01.10.575112doi: bioRxiv preprint 47 at the WT level is largely confined to the midcell region by an unknown mechanism. See 1206 text for additional details. 1207 (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. The copyright holder for this preprintthis version posted January 27, 2024. ; https://doi.org/10.1101/2024.01.10.575112doi: bioRxiv preprint Fig. 1. 100 s 100 s 100 s B A iHT-RodA Time Lapse Kymographs 100 s iHT-bPBP2b Time Lapse Kymographs 1 2 3 iHT-MreC Time Lapse Kymographs iHT-aPBP1a Time Lapse Kymographs 1 2 3 1 2 3 1 2 3 1 2 3 1 2 3 1 2 3 1 2 3 C D S C D S C D S0.0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1.0 0.09 0.67 0.23 0.40 0.38 0.22 0.57 0.26 0.17 Relative frequency iht-rodA // PZn-iht-rodA -Zn -Zn +Zn iht-rodA (n=118, R=4) (n=74, R=2) (n=75, R=2) ✱✱ ns ✱✱ ns ns ns ✱✱✱ ✱✱✱ ✱✱✱ iHT-RodA C D S C D S C D S0.0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1.0 0.16 0.71 0.12 0.46 0.36 0.19 0.60 0.32 0.07 Relative frequency iht-pbp2b // PZn-iht-pbp2b -Zn -Zn +Zn iht-pbp2b 15% 13% 82% (n=88, R=4) (n=68, R=2) (n=170, R=3) ✱ ✱✱ ✱✱✱ ns ns ✱ ✱✱✱ ✱✱✱✱ ns iHT-bPBP2b C D S0.0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1.0 0.24 0.52 0.24 Relative frequency -Zn iht-mreC 82% (n=96, R=3) ✱✱✱ ✱✱ ns iHT-MreC C D S C D S C D S0.0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1.0 0.06 0.80 0.140.10 0.61 0.29 0.12 0.62 0.26 Relative frequency iht-pbp1a // PZn-iht-pbp1a -Zn -Zn +Zn iht-pbp1a 9% 101% (n=122, R=2) (n=82, R=2) (n=125, R=5) 10% ✱ ns ns ✱ ns ns ✱✱✱✱ ✱✱✱✱ ✱ iHT-aPBP1a C D S0.0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1.0 1.1 0 0.87 0.13 Relative frequency -Zn iht-mpgA 171% (n=148, R=3) ✱✱✱✱ ✱✱✱ ✱ iHT-MpgA Motion Type C = Circumferential D = Diffusive S = Static % Protein amount relative to untagged WT (n= measurements, R= biol. replicates) (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. The copyright holder for this preprintthis version posted January 27, 2024. ; https://doi.org/10.1101/2024.01.10.575112doi: bioRxiv preprint Fig. 1. Elongation PG synthesis proteins display processive circumferential motion, and a limited number of molecules engage in PG synthesis simultaneously. Sm-TIRFm was performed at 1 FPS as described in Materials and Methods on strains iht-pbp2b (IU15928), iht-pbp2b // PZn-iht-pbp2b (IU16553), iht-rodA (IU15970), iht-rodA // P Zn-iht-rodA (IU16496), iht-mreC (IU16344), iht-pbp1a (IU16320), iht-pbp1a // PZn-iht-pbp1a (IU16497) and iht-mpgA (IU15997). (A) Representative time- lapse images and kymographs of molecules displaying processive circumferential movement in strains IU15928, IU15970, IU16344 and IU16497. Cell outlines are green (DICm) and HT-labeled molecules are magenta. Location of lines used to make kymographs are shown in the time lapse images (yellow dashed lines). For each protein, the left kymograph was generated from the cell depicted in the time-lapse images. Numbers (1-3) denote when each image in the time lapse was taken. The other two kymographs are from cells that are not shown. Calculation of circumferential velocities from kymographs are described in Materials and Methods. Scale ba rs are 1 µm. ( B) Movement patterns of HT-labeled molecules. Bars represent the mean relative frequency of each type of motion. For a given strain or condition, the relative frequencies of circumferential, diffusive, and static molecules were determined for each biological replicate. Frequencies were then averaged over all replicates to determine the mean (error bars represent ± SD). Mean values are annotated above each bar, and the total number of molecules analyzed (n) from the number of biological replicates (R) for each strain is below each strain or condition. Circumferential (red), diffusive (blue), and static (grey) movement patterns are shown. Circumferentially moving molecules were defined as molecules moving in one direction for 6 or more frames (at 1 FPS imaging rate) with a linear velocity ≥ 5 nm/s. Static molecules were defined as molecules not moving or moving for 6 or more frames with a velocity < 5 nm/s. Diffusive molecules were defined as molecules that moved, but not in a consistent direction, for 6 or more non-consecutive frames within a period of 90 s. Full criteria for determining movement patterns are described in Materials and Methods. The expression level of iH T -fusion proteins relative to the untagged WT level is shown as a percentage under the strain names. iHT-fusion proteins are expressed solely from the native chromosomal locus or with an additional copy of the gene encoding the iHT-fusion protein at an ectopic site under control of a zinc-inducible promoter, with 0 or 0.25 mM Zn inducer added. Unpaired t- tests were performed to compare relative frequencies of motion types. ns (nonsignificant); *P < 0.05; **P < 0.01; ***P < 0.001; ****P < 0.0001. (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. The copyright holder for this preprintthis version posted January 27, 2024. ; https://doi.org/10.1101/2024.01.10.575112doi: bioRxiv preprint ≈15% iHT-bPBP2b HADA iht-pbp2b (IU15928) B Z-stack summations HADA ≈100% iHT-bPBP2b BHI +0 mM Zn inducer iht-pbp2b // PZn-iht-pbp2b (IU16553) A BHI +0.3 mM Zn inducer Fig. 2. Diffusively moving components of the core PG elongasome do not actively synthesize PG (blue) in non-midcell regions of exponentially growing pneumococcal cells. 3D-SIM images, where iHT-bPBP2b is red, regions of PBP transpeptidase activity is blue, and midcell regions include septa of dividing cells and equators of predivisional daughter cells. Cells were grown in BHI ± 0.3 mM Zn inducer, labeled with a saturating amount of HT-ligand for 15 min, washed, labeled with 400 µM of the fluorescent D-amino acid HADA (blue) for 2.5 min, fixed, and imaged as described in

and Methods. Images are summed from 15 Z plane sections ( Z-stack) representing a total depth of 1.875 µm. Percentages indicate the amount of iHT-bPBP2b expressed relative to the untagged WT bPBP2b+ level ( 32). ( A) iht-pbp2b (IU15928) and ( B) iht-pbp2b // P Zn-iht-pbp2b (IU16553). Scale bar is 1 µm.. (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. The copyright holder for this preprintthis version posted January 27, 2024. ; https://doi.org/10.1101/2024.01.10.575112doi: bioRxiv preprint 0 10 20 30 40 60 80 100Circumferential Velocity (nm/s) Native: Ectopic: Zn: iht-pbp2b PZn-iht-pbp2b- - - + iht-rodA PZn-iht-rodA- - +- iht-mreC - - iht-mpgA - - iht-pbp1a PZn-iht-pbp1a- +- - 13.1 ±5.4 (85) 10.4 ±4.4 (28) Mean ±SD (n) 11.6 ±4.5 (33) 12.3 ±4.4 (75) 9.1 ±2.6 (12) 10.7 ±4.7 (34) 12.5 ±7.5 (69) 15.9 ±15.3 (14) 14.6 ±11.5 (43) 19.7 ±13.2 (8) 54.3 ±24.8 (27) ns ns ✱✱✱✱ ns ns ns ✱✱✱✱ ns ✱✱✱✱ ✱✱ 0 10 20 30 40 50 60 70 90Circumferential Duration (s) 26.6 ±11.8 (85) 22.4 ±9.8 (28) Mean ±SD (n) 30.4 ±14.9 (75) 23.8 ±11.7 (12) 23.4 ±13.0 (69) 10.6 ±4.4 (14) 9.8 ±3.9 (43) 7.3 ±1.7 (27) Native: Ectopic: Zn: iht-pbp2b PZn-iht-pbp2b- - +- iht-rodA PZn-iht-rodA- - +- iht-mreC - - iht-mpgA - - iht-pbp1a PZn-iht-pbp1a- +- - 25.2 ±13.5 (33) 21.2 ±9.2 (34) 10.0 ±3.8 (8) ns ns ✱✱✱✱ ✱✱✱✱ ns ✱✱✱ ✱✱✱✱ ✱✱✱✱ ✱✱✱✱ ns Fig. 3. B A (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. The copyright holder for this preprintthis version posted January 27, 2024. ; https://doi.org/10.1101/2024.01.10.575112doi: bioRxiv preprint Fig. 3. bPBP2b, RodA, and MreC form a stable complex during active PG synthesis. Dot plots of (A) circumferential velocities and (B) circumferential durations of elongation PG synthesis proteins determined by sm-TIRFm at 1 FPS. Strains are the same as in Fig. 1. +Zn condition indicates 0.25 mM Zn inducer was added. Black and red lines are median ± interquartile, and mean ± SD are indicated. n = total molecules analyzed from 2-5 biological replicates. Dotted grey lines indicate the minimum thresholds for (A) velocity (5 nm/s) and (B) duration (6 s). A Kruskal-Wallis with Dunn’s multiple comparisons test was used to compare circumferential velocities or circumferential durations in different strains. ns (nonsignificant); **P < 0.01; ***P < 0.001; ****P < 0.0001. (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. The copyright holder for this preprintthis version posted January 27, 2024. ; https://doi.org/10.1101/2024.01.10.575112doi: bioRxiv preprint 0 5 10 15 20 25 30 35Circumferential Velocity (nm/s)WT 14.7 ±5.4 nm/s (53) ∆murZ 10.0 ±4.0 nm/s (40) iHT-bPBP2b ✱✱✱✱ mean ±SD (n) 0 5 10 15 20 25 30 35Circumferential Velocity (nm/s)WT 14.2 ±4.8 nm/s (50) ∆murZ 9.1 ±3.0 nm/s (32) iHT-RodA ✱✱✱✱ mean ±SD (n) 0 5 10 15 20 25 30Circumferential Velocity (nm/s)WT 12.0 ±5.0 nm/s (70) FtsZ(D214A) ectopic expression 14.0 ±5.3 nm/s (62) iHT-bPBP2b ✱ mean ±SD (n) 0 10 20 30 40 50 60 70 80 iHT-bPBP2b (single molecule) 15.2 ±5.2 nm/s (38) FtsZ-sfGFP (treadmilling) 35.2 ±15.2 nm/s (48) Circumferential Velocity (nm/s) ✱✱✱✱ mean ±SD (n) A B C D +Zn +Zn Fig. 4. Circumferential movement of bPBP2b and RodA is independent of FtsZ treadmilling and reflective of PG synthesis. Sm-TIRFm was performed at 1 FPS on strains expressing iHT-bPBP2b or iHT-RodA. Dot plots of circumferential velocities are shown. Black and red lines are median ± interquartile, and mean ± SD are indicated. (A) Both iht-pbp2b (black, IU15928) and iht-pbp2b PZn- ftsZ(D214A) (red, IU16091) strains were supplemented with 0. 25 mM Zn inducer. ( B) IU16056 expresses both iHT-bPBP2b and FtsZ-sfGFP. The treadmilling velocity of FtsZ-sfGFP filaments (green) and the circumferential velocity of iHT-bPBP2b single molecules (red) were determined. (C) iht-pbp2b (black, IU15928) and iht-pbp2b ∆murZ (red, IU16110) strains. ( D) iht-rodA (black, IU15970) and iht-rodA ∆murZ (blue, IU16112) strains. n = total molecules analyzed from 2 biological replicates. Circumferential velocities in different strains were compared using a Mann- Whitney test. *P < 0.05; ****P < 0.0001. (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. The copyright holder for this preprintthis version posted January 27, 2024. ; https://doi.org/10.1101/2024.01.10.575112doi: bioRxiv preprint Fig. 5. A 0 1 2 3 4 5 6 7 8 9 0.001 0.01 0.1 1 +Zn -Zn +Zn -Zn +Zn -Zn WT ∆pbp2b // PZn-pbp2b iht-pbp2b iht-pbp2b(S391A) // PZn-pbp2b iht-rodA iht-rodA(D283A) // PZn-rodA Time (hours) OD620 -Zn -Zn -Zn B C Native locus Ectopic locus pbp2b(S391A)iht PZn * rodA(D283A)iht PZn IU16232 IU16239 * Time Lapse Kymographs 180 s iHT-bPBP2b(S391A) [bPBP2b+ depl.] Scale bar = 0.5 µm D iHT-RodA(D283A) [RodA+ depl.] 180 s Time Lapse Kymographs -Zn 3 h Microscopy E pbp2b+ rodA+ -Zn 3.5-4 h iHT-bPBP2b(S391A) [bPBP2b+ depl.] HTHT + Phase iHT-RodA(D283A) [RodA+ depl.] iht-rodA(D283A) // PZn-iht-rodA+ C D S C D S C D S0.0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1.0 0 0.48 0.52 0 0.58 0.42 0 0.69 0.31 Relative frequency iht-pbp2b(S391A) // PZn-iht-pbp2b+ +Zn -Zn (n=68, R=2) (n=52, R=2) -Zn (n=60, R=2) ✱✱ ✱✱ ✱✱ ✱✱ ns ✱✱ ✱✱✱✱ ✱✱✱ ✱✱✱✱ C = Circumferential D = Diffusive S = Static (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. The copyright holder for this preprintthis version posted January 27, 2024. ; https://doi.org/10.1101/2024.01.10.575112doi: bioRxiv preprint Fig. 5. The catalytic activities of bPBP2b and RodA are required for cell viability and PG synthesis, but not localization. (A) Diagram of merodiploid strains containing catalytic mutant alleles fused to the HaloTag at the native locus and the Zn-inducible WT allele at the ectopic site. ( B) Representative growth curves of WT (black square, IU1824), iht-pbp2b (red square, IU15928), iht- rodA (blue square, IU15970), ∆pbp2b // P Zn-pbp2b (green, IU11258), iht-pbp2b(S391A) // P Zn- pbp2b (red, IU16232), and iht-rodA(D283A) // P Zn-rodA (blue, IU16239) strains grown in C+Y medium with or without 0.2 mM Zn inducer as described in the section on ectopic expression and depletion conditions in Materials and Methods. The experiment was repeated twice with similar results. ( C and D) sm-TIRFm was performed at 1 FPS on strains IU16232 and IU16239. ( C) Movement patterns of iHT-bPBP2b(S391A) (IU16232) and iHT-RodA(D283A) (IU16239). The layout is the same as Fig. 1B (see legend for details). Unpaired t-tests were performed to compare r elative frequencies of motion types. ns (nonsignificant); **P < 0.01; ***P < 0.001; ****P < 0.0001. (D) Representative montages and kymographs of static molecules from strains IU16232 and IU16239 (without Zn inducer for 3 h) are displayed as described for Fig. 1A, except the scale bar is 0.5 µm. ( E) 2D- FM of catalytically inactive proteins. Strains IU16232 and IU16239 were grown without Zn (see Fig. 5B) and labeled with saturating amounts of HT-ligand. HT-fusion proteins are shown in magenta. Scale bar is 1 µm. The experiment was repeated twice with similar results. (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. The copyright holder for this preprintthis version posted January 27, 2024. ; https://doi.org/10.1101/2024.01.10.575112doi: bioRxiv preprint A HTPhase iht-pbp2b ΔmreC // PZn-mreC (IU16281) Merge iht-pbp2b mreC+ (IU15928) HTPhase Merge 1 µm +Zn -Zn +Zn -Zn MreC+ MreC+ MreC+ MreC depl. (3 h) B C D S C D S0.0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1.0 0.02 0.54 0.45 0.65 0.19 0.16 Relative frequency MreC+ -Zn (n=120, R=2, IU15928) MreC depletion (3 h) -Zn (n=101, R=2 IU16281) iHT-bPBP2b ✱ ✱ ns ✱ ns ✱✱ C = Circumferential D = Diffusive S = Static Fig. 6. bPBP2b localization becomes diffuse and circumferential movement is lost upon MreC depletion. (A) 2D-FM of strains iht-pbp2b (left, IU15928) and iht-pbp2b ∆mreC // PZn-mreC (right, IU16281). Cells were grown in C+Y media with or without (depletion) 0.2 mM Zn inducer, labeled with saturating amounts of HT-ligand, and imaged after 3 h. See the section on ectopic expression and depletion conditions in Materials and Methods for details. The experiment was repeated twice with similar results. (B) sm-TIRFm was performed at 1 FPS on strains iht-pbp2b (IU15928) and iht- pbp2b ∆mreC // P Zn-mreC (IU16281) after 3 h of growth in C+Y media without Zn inducer (depletion). The layout is the same as Fig. 1 B (see legend for details). Unpaired t- tests were performed to compare relative frequencies of motion types. ns (nonsignificant); *P < 0. 05; **P < 0.01. (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. The copyright holder for this preprintthis version posted January 27, 2024. ; https://doi.org/10.1101/2024.01.10.575112doi: bioRxiv preprint 0 5 10 15 20 25 30 35 40 Circumferential Velocity (nm/s) bPBP2b+ 12.3 ±4.4 nm/s (75) bPBP2b depl. 14.1 ±6.3 nm/s (25) iHT-RodA ns mean ±SD (n) -Zn Fig. 7. C B D E A Phase + HT HT iht-pbp2b ∆rodA // PZn-rodA+ (IU16204) +Zn -Zn RodA+ RodA depl. (3 h) iht-rodA ∆pbp2b // PZn-pbp2b+ (IU16202) Phase + HT HT +Zn -Zn bPBP2b+ bPBP2b depl. (3 h) iHT-bPBP2b iHT-RodA 0 5 10 15 20 25 30 35 40 Circumferential Velocity (nm/s) RodA+ 13.1 ±5.4 nm/s (85) RodA depl. 11.9 ±5.1 nm/s (27) iHT-bPBP2b ns mean ±SD (n) -Zn C D S C D S C D S C D S0.0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1.0 0.28 0.62 0.10 0.57 0.26 0.170.21 0.67 0.11 0.60 0.32 0.07 Relative frequency (n=88, R=4 IU15928) iHT-bPBP2b RodA+ RodA depl. (n=96, R=2 IU16204) (n=118, R=4 IU15970) iHT-RodA bPBP2b+ bPBP2b depl. (n=77, R=2 IU16202) -Zn -Zn -Zn -Zn ✱ ✱✱✱ ✱✱ ✱ ✱ ns ✱✱ ns ✱✱ ✱✱ ✱✱ ✱ C = Circumferential D = Diffusive S = Static (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. The copyright holder for this preprintthis version posted January 27, 2024. ; https://doi.org/10.1101/2024.01.10.575112doi: bioRxiv preprint Fig. 7. bPBP2b and RodA require each other for localization and motion. 2D-FM and sm-TIRFm at 1 FPS were performed on strains iht-pbp2b ∆rodA // PZn-rodA (IU16204) and iht-rodA ∆pbp2b // PZn-pbp2b (IU16202). ( A and B) 2D- FM showing localization in cells of strains ( A) iHT-bPBP2b (IU16204) and (B) iHT-RodA (IU16202) grown with or without (depletion; depl.) 0.2 mM Zn inducer for 3 h and labeled with saturating amounts of HT-ligand. See the section on ectopic expression and depletion conditions in Materials and Methods for additional details. The experiment was repeated twice with similar results. (C to E) sm-TIRFm was performed at 1 FPS on IU16204 and IU16202 without Zn inducer (depletion) for 3 hours. ( C) Movement patterns of HT-labeled molecules. Results for the iht-pbp2b ∆rodA // PZn-rodA (IU16204) and iht-rodA ∆pbp2b // PZn-pbp2b (IU16202) strains were determined and are graphed with data for iht-pbp2b (IU15928) and iht-rodA (IU15970) r eplotted from Fig. 1B for comparison. The layout is the same as Fig. 1B (see legend for details). Unpaired t- tests were performed to compare relative frequencies of motion types. ns (nonsignificant), *P < 0.05; **P < 0.01; ***P < 0.001. (D and E) Dot plots of circumferential velocities with mean values ± SD indicated for (D) strains IU15928 (black) and IU16204 (red), and (E) strains IU15970 (black) and IU16202 (red). Data from stains IU15928 and IU15970 are replotted from Fig. 3A for comparison. Black and red lines are median ± interquartile, and mean ± SD are indicated. n = total molecules analyzed from 2 biological replicates. A Mann-Whitney test was used to compare velocities upon protein depletion (depl.). ns (nonsignificant). (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. The copyright holder for this preprintthis version posted January 27, 2024. ; https://doi.org/10.1101/2024.01.10.575112doi: bioRxiv preprint Non-Midcell (n=79, R=2)(n=136, R=2) Non-Midcell 0.0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1.0 1.1 0.97 0.040 0.96 0.03 0 0.79 0.010 0.99 0.21 0 Relative frequency (n=30, R=2) iHT-aPBP1a+ (IU16497, +Zn) Midcell (n=71, R=2) Midcell iHT-aPBP1a(S370A) (IU19168, +Zn) ✱ ✱ C D S C D S C D S C D S C = Circumferential D = Diffusive S = Static Fig. 8. Circumferential movement of aPBP1a reflects active PG synthesis. Sm-TIRFm was performed at 1 FPS as described in Materials and Methods on strains iht-pbp1a // PZn-iht-pbp1a (IU16497) and iht-pbp1a(S370A) // P Zn-iht-pbp1a(S370A) (IU19168) with 0. 25 mM Zn added. Movement patterns of HT-labeled molecules at sites of PG synthesis (midcell) or elsewhere in the cell (non-midcell) were analyzed. The layout is the same as Fig. 1B (see legend for details). A two- way ANOVA with Tukey’s multiple comparison test was used to compare relative frequencies of motion types among strains. *P < 0.05. (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. The copyright holder for this preprintthis version posted January 27, 2024. ; https://doi.org/10.1101/2024.01.10.575112doi: bioRxiv preprint 0 500 1000 1500 2000Velocity (nm/s) Midcell 327 ±267 nm/s (42) Non-midcell 1,026 ±410 nm/s (12) iHT-MpgA (10 FPS) ✱✱✱✱ mean ±SD (n) Kymographs (1 FPS)A B Kymographs (10 FPS) 70 s 10 s iHT-MpgA iHT-MpgA iHT-MreC (for comparison) 10 s circumf. diffusive 10 s C Kymographs (10 FPS) Fig. 9. MpgA displays confined subdiffusive movement at midcell. Sm-TIRFm was performed on strains iht-mpgA (IU15997) and iht-mreC (IU16344). Representative kymographs of (A) iHT-MpgA molecules moving at sites of PG synthesis imaged at 1 FPS and ( B) iHT-MpgA and iHT-MreC molecules imaged at 10 FPS. (C) Dot plots of velocities determined at 10 FPS over short runs of iHT-MpgA molecules in midcell regions (black) and at non-midcell regions (red). Black and red lines are median ± interquartile, and mean ± SD are indicated. n = total molecules analyzed from 2 biological replicates. Velocities in different regions of cells were compared using a Mann-Whitney test. ****P < 0.0001. (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. The copyright holder for this preprintthis version posted January 27, 2024. ; https://doi.org/10.1101/2024.01.10.575112doi: bioRxiv preprint A (μm2/s) ±SEM α % diffusive iht-pbp1a // PZn-iht-pbp1a +Zn 0.055 ±0.0005 0.28 77 914 iht-rodA // PZn-iht-rodA +Zn 0.050 ±0.0002 0.23 90 2,116 iht-mreC 0.047 ±0.0003 0.22 68 1,351 iht-pbp2b // PZn-iht-pbp2b +Zn 0.046 ±0.0003 0.20 74 1,143 iht-mpgA 0.040 ±0.0003 0.18 60 1,291 n Diffusion coefficient 0.0 0.1 0.2 0.3 0.4 0.5 0.6 0.00 0.02 0.04 0.06 0.08 0.10 Diffusive Time (s) MSD (µm2) Non-diffusive (processive and static) ** B Fig. 10. DiffusiveNon-Diffusive iHT-bPBP2b iHT-RodA iHT-MreC iHT-MpgAiHT-aPBP1a Displacement (nm per step) >150<25 (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. The copyright holder for this preprintthis version posted January 27, 2024. ; https://doi.org/10.1101/2024.01.10.575112doi: bioRxiv preprint Fig. 10. Components of the pneumococcal core PG elongasome and proteins implicated in elongation PG synthesis exhibit different patterns of confined diffusion. Sm-TIRFm was performed at 20 FPS on strains iht-pbp2b // PZn-iht-pbp2b (IU16553), iht-rodA // PZn-iht-rodA (IU16496), iht- mreC (IU16344), iht-pbp1a // P Zn-iht-pbp1a (IU16497) and iht-mpgA (IU15997). 0. 25 mM Zn inducer was added to strains IU16553, IU16496 and IU16497. ( A) Representative fields of cells containing multiple single-molecule trajectories (colored lines) classified as diffusive (top row) and non-diffusive (bottom row). Trajectories with a displacement > 0. 13 µm and a velocity standard deviation > 0. 63 µm/s were defined as diffusive, and the remaining trajectories were defined as non-diffusive. Full details of trajectory constructions and classification criteria are described in

and Methods. The color of each trajectory line segment represents the displacement (in nm) of the molecule from one frame to the next. Scale bar is 1 µm. (B) Mean square displacements (MSD) were calculated for diffusive and non-diffusive trajectories. Circles show MSD values, and error bars represent SEM. Lines represent MSD curves fit to the data as described in Materials and Methods. To the right of the graph are the mean diffusion coefficients ( ± SEM) of diffusive trajectories, alpha values of diffusive trajectories (α), the percent of total trajectories that were classified as diffusive (% diffusive), and the total number of trajectories (n) analyzed from 2 biological replicates. α values < 1 indicate that molecules are exhibiting confined diffusion (or subdiffusion), discussed in the text. The diffusion coefficients were compared between strains using a Brown-Forsythe and Welch’s ANOVA with a Games-Howell multiple comparisons test. **P < 0.01. (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. The copyright holder for this preprintthis version posted January 27, 2024. ; https://doi.org/10.1101/2024.01.10.575112doi: bioRxiv preprint A Active: circumferential, processive movement at ≈11 nm/s at midcell; dependent on PG synthesis and independent of FtsZ treadmilling Inactive: diffusive in non-midcell regions; static at midcell (?) Subdiffusive movement, largely confined to midcell region; some static and paused at midcell; no circumferential movement; some diffusive in non- midcell regions Active: circumferential, processive movement at ≈15 nm/s at midcell; dependent on PG synthesis; paused diffusive molecules (?) Inactive: diffusive in non-midcell regions; static at midcell (?) MpgA Muramidase (PG glycan-chain release in elongation) Core Elongasome (elongation PG synthesis) aPBP1a PG synthase (possibly reinforce peripheral PG and/or repair damaged PG) B bPBP2b or aPBP1a expression WT levels≈10% of WT bPBP2b MpgA aPBP1a Circumferential Diffusive Static FtsZ ring Key PG release PG release out in MreC (dimer) bPBP2b RodA RodZ MreD Short or transient interactions possible Fig. 11. (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. The copyright holder for this preprintthis version posted January 27, 2024. ; https://doi.org/10.1101/2024.01.10.575112doi: bioRxiv preprint Fig. 11. Summary model of (A) the composition and dynamics of core elongasome components (bPBP2b, RodA, and MreC) and proteins linked to PG elongation synthesis (aPBP1a and MpgA) in growing S. pneumoniae cells; and (B) the distribution of these proteins when underexpressed (left) or expressed at WT levels (right). (A) The circumferential, processive movement of bPBP2b, RodA, and MreC reported here likely apply to the other core elongasome components RodZ and MreD. In early, predivisional cells, the core PG elongasome is located in the FtsZ-ring at the equator midcell of a daughter cell. In later divisional cells, the core PG elongasome is in the outer ring of the midcell septal annular disk. The motion of MreC is the same in early or late divisional cells. Class A aPBP1a also displays circumferential, processive motion at midcell that depends on PG synthesis, but its dynamics indicate that aPBP1a is not a persistent member of the core PG elongasome, although short interactions are possible. In contrast, MpgA moves in a distinctive type of subdiffusion that is largely confined to the midcell region. Components of the core elongasome and aPBP1a mo ve diffusively outside of midcell rings. There are also nonmoving, static molecules (> 6 s) of elongasome proteins and aPBP1a, mainly in midcell regions. It is not clear whether static molecules are synthesizing PG or are in a non-synthesizing transition state. In addition, diffusing molecules of aPBP1a and MpgA were infrequently detected pausing (< 1s) before resuming motion. See text for additional information. (B) When underexpressed, components of the core PG elongasome are largely confined to PG synthesis at midcell septa and the equators of predivisional daughter cells starting to divide. In contrast, when expressed at WT levels, most components of the PG elongasome (and aPBP1a ) are diffusing over the cell surface and are not synthesizing PG. Thus, elongasome components and aPBP1a are in excess in growing cells, and only a limited number engage in midcell PG synthesis. In contrast, subdiffusion of MpgA molecules expressed at the WT level is largely confined to the midcell region by an unknown mechanism. See text for additional details. (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. The copyright holder for this preprintthis version posted January 27, 2024. ; https://doi.org/10.1101/2024.01.10.575112doi: bioRxiv preprint