The AAA+ ATPase ClpX Is Critical for Growth and Development of Chlamydia trachomatis

Abstract Chlamydia trachomatis (Ctr) is an obligate intracellular bacterium that undergoes a complex developmental cycle in which the bacterium differentiates between two functionally and morphologically distinct forms, each of which expresses its own specialized repertoire of proteins. The transitions between the infectious, non-dividing elementary body (EB) and the non-infectious, replicative reticulate body (RB) are not mediated by division events that re-distribute intracellular proteins. Rather, both primary (EB to RB) and secondary (RB to EB) differentiation require protein turnover. The Clp protease system is well conserved in bacteria and, minimally, relies on a serine protease subunit, ClpP, and a AAA+ ATPase, such as ClpX, that recognizes and unfolds substrates for ClpP degradation. In Chlamydia, clpX is encoded within an operon adjacent to clpP2. We present evidence that the chlamydial ClpX ortholog, and the co-transcribed ClpP2, play a key role in organism viability and development. We demonstrate here that chlamydial ClpX is a functional ATPase and forms the expected homohexamer in vitro. Overexpression of a ClpX mutant lacking ATPase activity had a limited impact on DNA replication or secondary differentiation but, nonetheless, reduced EB viability. Conversely, the overexpression of an inactive ClpP2 mutant significantly impacted later developmental cycle progression by reducing the overall number of organisms. Blocking clpP2X transcription using CRISPR interference led to a decrease in bacterial growth, which did not occur when the non-essential gene incA was targeted. Taken together, our data indicate that ClpX and the associated ClpP2 play a critical role in developmental cycle progression and differentiation. Importance Chlamydia trachomatis is the leading cause of infectious blindness globally and the most reported bacterial sexually transmitted infection both domestically and internationally. Given the economic burden, the lack of an approved vaccine, and the use of broad-spectrum antibiotics for treatment of infections, a further understanding of chlamydial growth and development is critical for the advancement of novel, targeted antibiotics. The Clp proteins comprise an important and conserved protease system in bacteria. Our work highlights the importance of the chlamydial Clp proteins to this clinically important bacterium. Additionally, our study implicates the Clp system playing an integral role in chlamydial developmental cycle progression, which may help establish models of how Chlamydia spp. and other bacteria progress through their respective developmental cycles. Our work also contributes to a growing body of Clp-specific research that underscores the importance and versatility of this system throughout bacterial evolution and further validates Clp-proteins as drug targets.


Results 104
The chlamydial ClpX retains conserved motifs of, and exhibits predicted structural homology 105 to, ClpX orthologs. To initiate our study, we first performed bioinformatic and ab initio 106 structural modeling analyses to determine whether the chlamydial ClpX (ClpX Ctr ) possesses the 107 expected conserved regions and motifs consistent with its proposed function as an AAA+ 108 ATPase. Using multiple sequence alignment, we aligned ClpX Ctr to ClpX orthologs and 109 annotated conserved motifs identified in other studies (Fig. 1a). ClpX Ctr retains the N-terminal 110 metal binding domain (24,25), the Walker A and B motifs for ATP binding and hydrolysis, 111 respectively (21, 23), the sensor motifs for recognition of nucleotide bound state (26), the RKH 112 motif and pore loops for substrate recognition (27-29) and unfolding (30, 31), the arginine finger 113 for inter-subunit sensing of nucleotide state in the ClpX hexamer (22,32), and the IGF Loop for 114 interaction with ClpP (33,34). Interestingly, the predicted secondary structure of ClpX Ctr shows 115 few notable aberrations (see Discussion) from other prototypical bacterial ClpX orthologs and is 116 predicted to form the expected homohexamer by structural modeling (Fig. 1b, two subunits 117 removed for clarity). The spatial conservation of AAA+ and ClpX-specific motifs (colored in 118 Fig. 1b as in the multiple sequence alignment) indicates that the chlamydial ClpX likely 119 functions using a mechanism similar or identical to other ClpX orthologs. Taken together, these 120 in silico studies suggest ClpX Ctr functions as a canonical AAA+ ATPase. 121 122 Chlamydial ClpX forms the expected homohexamer and possesses ATPase activity. To 123 determine the oligomeric state of ClpX Ctr in vitro, we purified recombinant protein and analyzed 124 its migration by native PAGE. At the same time, we also constructed a Walker B ATPase mutant 125 (E187A) ClpX Ctr as a control for biochemical studies. Following the incubation of 10 g of wild-126 not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission.
The copyright holder for this preprint (which was this version posted December 8, 2019. ; https://doi.org/10.1101/868620 doi: bioRxiv preprint type or mutant ClpX Ctr for 20 minutes in a HEPES based buffer, we loaded the entire volume 127 into a 4-20% gradient gel. We observed the ClpX Ctr proteins migrating above the 242 kDa band 128 of the molecular weight ladder, which is close to the expected hexameric size of 283 kDa (Fig. 129 2a). We then sought to assess ATPase activity of recombinant wild-type and ATPase mutant 130 ClpX Ctr using the Biomol Green endpoint assay to measure free phosphate levels following ATP 131 hydrolysis, which served as a proxy for ATPase activity. Indeed, ClpX Ctr hydrolyzed ATP, while 132 the inactive mutant isoform showed a significant defect in ATP hydrolysis (Fig. 2b). These data 133 indicate that ClpX Ctr (i) forms a homohexamer of the predicted size and (ii) possesses ATPase 134 activity that is abrogated by a mutation in the Walker B motif. 135 We next tested whether wild-type and ATPase mutant ClpX Ctr interact with each other 136 using the Bacterial Adenylate Cyclase Two-Hybrid (BACTH) assay. This system is predicated 137 on the reconstitution of adenylate cyclase activity by bringing two complementary fragments of 138 the enzyme (T25 and T18) into close proximity by interacting proteins. Generation of cAMP by 139 the reconstituted adenylate cyclase drives ß-galactosidase production that can be measured 140 qualitatively by the presence of blue colonies and growth on minimal medium (Fig. 2c) or 141 quantitatively by measuring enzyme activity directly (Fig. 2d). We performed a series of 142 pairwise interaction tests between the wild-type and mutant ClpX Ctr . In each instance, we 143 observed a positive interaction that was quantifiable and on par with the positive control (T25-144 Zip vs T18-Zip). We conclude from these data that the mutant isoform can interact with the wild-145 type isoform. 146 147

Overexpression of inactive ClpX or inactive ClpP2 has both overlapping and independent 148
effects. We previously measured the effects of overexpression of both wild-type and catalytically 149 not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission.
The copyright holder for this preprint (which was this version posted December 8, 2019. ; https://doi.org/10.1101/868620 doi: bioRxiv preprint inactive ClpP2 Ctr on chlamydial growth and observed a modest reduction in growth at 24 hours 150 post-infection (hpi) (17). We wanted to more carefully assess growth differences during the 151 chlamydial developmental cycle in the presence of overexpressed wild-type and mutant ClpX Ctr 152 (and ClpX E187A ), ClpP2 Ctr (and ClpP2 S98A ), or both together (ClpP2X Ctr /ClpP2 S98A ClpX E187A ). To 153 do this, we performed growth curves where we induced expression, or not, at 10 hpi and 154 quantified growth at various timepoints after induction. Immunofluorescence analysis (IFA) of 155 replicate treatments and quantification of recoverable inclusion forming units (IFUs; a proxy for 156 EBs) revealed distinct effects upon overexpression of the individual components  as 157 well as with the entire operon (Fig. 3d&e). We noted that overexpression of wild-type ClpP2 Ctr 158 showed no appreciable effect at either 24 or 48 hpi (14 and 38 h pulses of induction, 159 respectively), whereas overexpression of ClpP2 S98A appeared to reduce the number of organisms 160 present within the inclusion at 48 hpi but not 24 hpi (Fig. 3a). These observations correlated with 161 the marked impact on EB production in the later time points of mutant ClpP2 S98A but not wild-162 type ClpP2 Ctr overexpression (Fig. 3b). Conversely, inactive ClpX E187A overexpression resulted 163 in smaller inclusions and a decrease in IFUs that was not observed for overexpression of the 164 wild-type ClpX (Fig. 3a&c). These IFU recovery data suggest that Ctr is more sensitive to 165 ClpX Ctr rather than ClpP2 Ctr disruption earlier in the developmental cycle, as the IFU reduction is 166 exacerbated sooner with ClpX E187A overexpression (note the differences at 24hpi in Fig. 3b&c). 167 As noted for the overexpression of individual wild-type isoforms, there was no significant impact 168 on IFU recovery of overexpressing both wild-type ClpP2 Ctr and ClpX Ctr in tandem. Consistent 169 with the effects of overexpressing individual mutant isoforms, overexpression of the inactive 170 ClpP2 S98A and ClpX E187A isoforms in tandem showed an exacerbated phenotype throughout the 171 developmental cycle as noted by both IFA and IFU assays (Fig. 3d&e). Importantly, the wild-172 not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission.
The copyright holder for this preprint (which was this version posted December 8, 2019. ; https://doi.org/10.1101/868620 doi: bioRxiv preprint type chromosomal copies of ClpP2 Ctr and ClpX Ctr continue to be expressed during these 173 overexpression assays. Therefore, the true impact of overexpression of the mutant isoforms is 174 likely underrepresented. 175 176 Functional disruption of ClpP2 blocks developmental cycle progression while ClpX disruption 177 reduces EB viability. Given that the IFU assay only measures EB viability from a population and 178 not total bacterial numbers or differentiation status, we next wanted to address these nuances of 179 the chlamydial developmental cycle. We first measured genomic DNA as a proxy for total 180 number of bacteria (i.e. both RBs and EBs). From 24 hpi to 48 hpi, we observed a significant 181 drop in gDNA levels when ClpP2 S98A was overexpressed alone or in the mutant operon 182 configuration (Fig. 4a). Conversely, overexpression of any wild-type protein had no significant 183 impact on DNA accumulation. Surprisingly, overexpression of the ClpX E187A also had no 184 significant impact on DNA levels in spite of the reduction in IFUs, suggesting total bacterial 185 numbers are unaffected. To determine differentiation status, we next assessed HctB levels, an 186 EB-specific gene product (35-37), by western blot as an indicator of secondary differentiation. 187 We normalized the integrated density of HctB to the integrated density of MOMP (major outer 188 membrane protein; present in both EBs and RBs) to ensure that we were comparing HctB levels 189 to the total number of bacteria. The relative HctB levels in samples where ClpP2 S98A was 190 overexpressed were reduced substantially, suggesting the generation of fewer EBs and consistent 191 with IFU and genomic DNA data, whereas the other experimental conditions showed no changes 192 in relative HctB levels (Fig. 4b&c). These data suggest that overexpression of ClpX E187A does 193 not impact bacterial replication, as measured by gDNA levels, or RB-to-EB differentiation, as 194 measured by HctB levels. Therefore, we prepared samples for transmission electron microscopy 195 not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission.
The copyright holder for this preprint (which was this version posted December 8, 2019. ; https://doi.org/10.1101/868620 doi: bioRxiv preprint to examine at higher resolution the morphology of EBs and RBs from ClpP2 S98A and ClpX E187A 196 overexpressing strains. Consistent with other measured effects, ClpP2 S98A overexpression 197 resulted in smaller inclusions with fewer organisms (Suppl. Fig. 1a). In contrast, but consistent 198 with its measured effects, ClpX E187A overexpression did not have an obvious effect on RB size or 199 numbers per se; rather, more bacteria with unusual, multi-nucleate staining were observed, as 200 indicated by the arrows (Suppl. Fig. 1b&d compared to uninduced in panel C). These abnormal 201 forms may potentially be EBs with defects in chromosomal packaging or intermediate bodies 202 that have not completed chromosomal condensation. Taken together with the IFU data ( Fig. 3), 203 these results suggest differential effects of overexpression of ClpP2 S98A and ClpX E187A and, by 204 inference, differential effects of these Clp components in the physiology of the organism. 205 206 Knockdown of the clpP2X operon reduces recoverable progeny and results in reduced plasmid 207 retention. Overexpression of mutant isoforms of ClpP2 Ctr and/or ClpX Ctr was sufficient to 208 disrupt chlamydial development in the presence of endogenous ClpP2X Ctr . However, we wanted 209 to directly block the chromosomal copies by employing an improved version of the chlamydial 210 CRISPR interference (CRISPRi) strategy previously described by us ((38) and Ouellette,in 211 prep). CRISPRi relies on the inducible expression of a catalytically inactive Cas9 (dCas9) in 212 combination with a guide RNA (gRNA) to block transcription at specific chromosomal sites 213 (39). We transformed Ctr L2 with vectors encoding the dCas9 and gRNAs targeting either the 214 clpP2X or incA intergenic regions. IncA knockdown served as a control since incA is a non-215 essential gene (40). The CRISPRi transformants were used to infect HEp2 cells. When dCas9 216 expression was induced at 10hpi, we observed a marked and rapid decrease in both clpP2 and 217 clpX transcript levels compared to the uninduced controls at 14hpi (Fig. 5a). Similar results were 218 not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission.
The copyright holder for this preprint (which was this version posted December 8, 2019. ; https://doi.org/10.1101/868620 doi: bioRxiv preprint observed when dCas9 expression was induced at 4hpi (data not shown). Importantly, we did not 219 observe a decrease in transcript levels for clpP1, euo, and omcB (Suppl. Fig. 2; (12,17,41,42)). 220 As previously observed, IncA expression was uniformly blocked after dCas9 induction (Fig. 5b;221 (38)). 222 We next assayed chlamydial growth as measured by IFU recovery after inducible 223 knockdown of the target genes. Expression of dCas9 was induced at 4hpi, and IFUs were 224 harvested at 24 and 48hpi and titred on fresh cell monolayers in the presence of penicillin, the 225 selection agent. When clpP2X expression was blocked at 4hpi, we noted a 5-fold decrease in 226 penicillin-resistant (i.e. transformants containing the CRISPRi plasmid) IFUs at 24hpi but a more 227 than 200-fold decrease at 48hpi (Fig. 5c). In performing these assays in the presence of 228 penicillin, we observed numerous penicillin-sensitive organisms (i.e. aberrant RBs (43)) during 229 the titration step, suggesting that the plasmid conferring resistance and encoding the CRISPRi 230 system was being lost after induction of dCas9 expression. To test this, we quantified plasmid 231 retention in the clpP2X knocked down samples and observed that blocking clpP2X expression 232 resulted in ~75% plasmid loss at 24hpi and greater than 90% loss at 48hpi (Fig. 5d). These 233 effects on IFUs and plasmid retention were not observed for incA knockdown ( Fig. 5c and d). 234 We note that incA knockdown did result in a reproducible, but transient, increase in IFUs at 235 24hpi that returned to "normal" levels at 48hpi (Fig. 5c). The reasons for this are not clear. 236 Nonetheless, we conclude from these data that blocking clpP2X expression is deleterious to 237 Chlamydia, further highlighting its essentiality to this pathogen. 238 239

Chemical disruption of ClpX function is detrimental to Ctr. Recently, ClpX-specific inhibitors 240
were synthesized by the Sieber group and shown to interfere with ClpX ATPase activity. One 241 not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission.
The copyright holder for this preprint (which was this version posted December 8, 2019. ; https://doi.org/10.1101/868620 doi: bioRxiv preprint compound, identified as 334, was shown to have potent inhibitory activity towards ClpX 242 whereas a derivative, 365, was inactive (44). We performed ab initio modelling and molecular 243 dynamics simulations (45) to determine if these compounds could interact with an ADP-bound 244 hexameric ClpX Ctr . For 334, a high scoring model (-9.1 kcal/mol binding affinity, RMSD ~ 0) 245 was predicted with the drug binding near to the ATP binding pocket, suggesting a mechanism of 246 action where 334 likely occludes the ATPase site (Suppl. Fig. 3). Whether the effect stems from 247 the blocking of ATP binding and subsequent destabilization of the complex, attenuation of 248 ATPase function by preventing a conformational change of the complex, or steric hindrance of 249 complex formation remains to be elucidated. Conversely, compound 365 bound outside of the 250 ATP pocket with a much lower score (Suppl. Fig. 4). 251 Given the predicted effects of the ClpX inhibitors on the structure of ClpX Ctr , we next 252 leveraged these compounds to assess the effect of specifically disrupting ClpX Ctr on chlamydial 253 growth. We initiated our studies by treating or not C. trachomatis L2 infected HEp2 cells at 8 hpi 254 with 25 g of drug to target specifically RBs early in development. At 24 hpi, we either 255 harvested and froze IFUs or replaced the medium containing either the drug or the vehicle 256 control with fresh medium lacking these. The latter samples were harvested at 48 hpi and frozen, 257 and then all collected samples were titred in the absence of drug treatment. Initial assessment of 258 immunofluorescent controls showed a marked reduction in inclusion size after 334 treatment for 259 both the 24 and 48 h timepoints (8-24h and 8-48h; Fig. 6a&b). This was accompanied by a 260 severe decrease in recovery of IFUs to near the limit of detection (Fig. 6c). As expected, 365 261 treatment had little effect on IFU recovery at 24 hpi but did reduce IFU numbers by a log 262 following prolonged treatment (8-24h and 8-48h; Fig. 6a-c), supporting our docking simulation 263 that showed lower affinity of 365 to ClpX Ctr . Moreover, 334 had a bacteriostatic effect on C. 264 not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission.
The copyright holder for this preprint (which was this version posted December 8, 2019. ; https://doi.org/10.1101/868620 doi: bioRxiv preprint trachomatis, as removal at 24 hpi allowed for a substantial recovery in IFU counts (8-24h; Fig.  265 6b&c). 266 We then sought to assess the importance of ClpX Ctr function throughout the 267 developmental cycle by treating either early, to target primary differentiation and inclusion 268 establishment, or later, to target pre-formed EBs. Treatment with 334 from 0 to 8 hpi resulted in 269 over a log reduction in recoverable IFUs, demonstrating the importance of ClpX Ctr early during Given the unique roles and protein repertoires of the chlamydial developmental forms 280 (EB/RB), we hypothesize that protein degradation is a critical factor in the differentiation process 281 from one form to the other. The Clp system is highly conserved in both prokaryotic and 282 eukaryotic systems where it has been described to perform important functions in both 283 proteostasis and pathogenesis (46). The Clp system is nominally composed of a proteolytic 284 subunit, ClpP, and a AAA+ ATPase that functions as an unfoldase to recognize substrates and 285 feed them into the ClpP barrel for degradation (23). The work presented here expands our 286 understanding of the chlamydial Clp protease system. Focusing on an initial characterization of 287 not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission.
The copyright holder for this preprint (which was this version posted December 8, 2019. ; https://doi.org/10.1101/868620 doi: bioRxiv preprint ClpX Ctr and the role of the clpP2X operon, we demonstrated the importance of the Clp protease 288 system during chlamydial growth and development. 289 Multiple lines of evidence support that the chlamydial ClpX is a bona fide AAA+ 290 ATPase. Firstly, multiple sequence alignment of ClpX Ctr to orthologs of other bacteria revealed a 291 perfect conservation of the motifs involved in nucleotide binding, ATP hydrolysis, and 292 nucleotide-state sensing ( Fig. 1) (47,48). Secondly, homology-directed and ab initio modelling 293 of ClpX Ctr revealed that the spatial orientation of these domains is conserved as well ( Fig. 1), 294 though we acknowledge that structural studies are critical to drawing conclusions about ClpX Ctr 295 conformational states. Thirdly, ClpX Ctr interacts with itself to form a homohexamer that 296 possesses ATPase activity (Fig. 2). Importantly, this ATPase activity could be disrupted by a 297 targeted mutation in the Walker B motif while having no effect on the oligomerization properties 298 of the protein. Fourthly, a characterized ClpX inhibitor that disrupts its ATPase activity also 299 disrupted the growth of C. trachomatis serovar L2 (Fig. 6). Finally, overexpression of a ClpX Ctr 300 ATPase mutant negatively impacted chlamydial growth and development (Figs. 3&4). 301 While we have characterized the ATPase function of ClpX Ctr and its role in chlamydial 302 growth, further work remains to determine whether this ClpX ortholog functions as an unfoldase. 303 Nevertheless, our bioinformatics analysis supports this as ClpX Ctr retains substrate recognition 304 motifs, including both pore loops and the RKH motif for gripping and translocation of substrates 305 (27-31). Chlamydia spp. also encode the tmRNA/ssrA tagging system for ribosomal rescue (18, 306 49-52), which fits a model where ClpX Ctr may play an integral role in turnover of tagged, 307 partially translated peptides. Whether ClpX Ctr can actually target SsrA-tagged substrates, and 308 whether this tagging is for ribosomal rescue or more specific purposes (53, 54), remains to be 309 not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission.
The copyright holder for this preprint (which was this version posted December 8, 2019. ; https://doi.org/10.1101/868620 doi: bioRxiv preprint determined and is currently under investigation by our research group. A recent article, using an 310 SsrA-tagged GFP, suggests this function of chlamydial ClpX may be conserved (55). 311 One unique feature of ClpX Ctr is the TSSTSSP link between the zinc binding domain 312 (ZBD) and the rest of the protein. To date, the structure of the ZBD has not been crystallized 313 with the rest of the protein due to its apparent disorder; yet, the ZBD of ClpX is important for its 314 function in other bacteria to, for example, recognize specific substrates (24, 56, 57). We 315 hypothesize that the TSSTSSP residues may serve a function in flexibility (58, 59) or extension 316 of the N-terminus, which in turn may modulate its unfoldase/chaperone activity (60). 317 Interestingly, an SP motif has been implicated in initiation of a Type I -hairpin turn (61, 62), 318 which may serve as a mechanism through which the ClpX Ctr N-terminus adopts a unique 319 conformation to recognize uncharacterized adaptors. This linker may be phosphorylated, leading 320 to a conformational switch of the intrinsically disordered N-terminus and enhancing the stability 321 of the otherwise disordered ZBD. We are investigating the potential for a phosphorylation state 322 to activate or attenuate ClpX Ctr function. We hypothesize that at least one of these situations aids 323 in selectivity of ClpX Ctr in vivo activity, but we cannot rule out that any combination may 324 function to yield multiple layers of control. 325 In Chlamydia, clpX is encoded in an operon with clpP2. Our data indicate that, not 326 surprisingly, the ClpP2X Ctr system is highly regulated and essential. We previously demonstrated 327 that unregulated ClpP Ctr activity, through the use of ClpP-activating antibiotics, is detrimental to 328 Chlamydia (17). Here, we performed a systematic analysis of the effects of overexpression of 329 wild-type or inactivated ClpP2X Ctr components. The overexpression of wild-type ClpP2 Ctr and/or 330 ClpX Ctr had no biologically or statistically significant effect on chlamydial growth that we could 331 measure. However, overexpressing inactive ClpP2 Ctr (S98A) and/or ClpX Ctr (E187A) resulted in 332 not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission.
The copyright holder for this preprint (which was this version posted December 8, 2019. ; https://doi.org/10.1101/868620 doi: bioRxiv preprint abrogation of chlamydial growth as measured by recovery of infectious progeny. Three 333 observations should be noted. Firstly, the effect of inducibly-expressed proteins is measured in 334 the presence of the endogenous chromosomally-expressed proteins. Therefore, it is likely that the 335 inactive mutants would have even more dramatic effects on chlamydial growth in the absence of 336 the wild-type chromosomal copy. For ClpX Ctr , this is supported by the effects of the ClpX 337 inhibitor on Chlamydia (Fig. 6), which effectively stopped chlamydial growth. Secondly, we 338 demonstrated that the mutant proteins could interact in vitro with wild-type isoforms (Fig. 2). 339 Therefore, we can infer that overexpression of the mutant proteins leads to their incorporation 340 into the endogenous ClpX machinery to disrupt or impair its function. Thirdly, to our knowledge, 341 ours is the first study to ectopically express two different tagged proteins in Chlamydia, showing 342 both the feasibility of this approach and its potential utility to dissect chlamydial biology. 343 The overexpression of the catalytically inactive mutant Clp proteins in Chlamydia 344 revealed potentially subtle differences in the role of each component in chlamydial growth and 345 development. Surprisingly, we noted a roughly 50% reduction in detectable genomes (Fig. 4A) 346 when ClpX Ctr (E187A) was expressed whereas IFUs were reduced roughly 20-fold (Fig. 3). The 347 production of EBs as measured by HctB levels did not appreciably change (Fig. 4b&c). This 348 suggests that, while development is hindered, the drop in IFUs may be due to defective EB 349 viability, infectivity, or inclusion establishment and not a defect in secondary differentiation per 350 se. Support for this comes from electron microscopy images, which revealed unusual 351 morphologies after overexpression of the mutant ClpX Ctr isoform (Suppl. Fig. 1). Conversely, for 352 ClpP2 Ctr (S98A) overexpression, the substantial IFU decrease coupled with a sharp drop in gDNA 353 levels indicate that ClpP2 Ctr plays a role in developmental cycle progression. HctB levels are also 354 significantly reduced, which is consistent with the lack of EB generation. Taken together, these 355 not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission.
The copyright holder for this preprint (which was this version posted December 8, 2019. ; https://doi.org/10.1101/868620 doi: bioRxiv preprint data may indicate that ClpP2 Ctr is integral to developmental cycle progression or differentiation 356 and that its function is tightly regulated. We cannot, however, exclude that secondary 357 differentiation is directly affected due to the fact that total organism numbers are severely 358 reduced. Rather, our proposed model suggests that ClpP2 Ctr disruption may affect both factors by 359 a mechanism that we are currently working to identify. Conversely, ClpX Ctr may serve a more 360 prominent ClpP2 Ctr -independent function in differentiation of the organism (Suppl. Fig. 5). 361 We successfully generated chlamydial transformants with an inducible knockdown 362 system to repress ClpP2X Ctr expression. To date, this study is the first of its kind in Chlamydia to 363 knock down genes that are essential, highlighting the utility of CRISPRi in studies of chlamydial 364 biology while providing insight into possible ClpP2X Ctr function. Notably, we observed a large 365 decrease in IFU production coupled with an increase in plasmid loss after inhibition of clpP2X 366 expression (Fig. 5). These effects were not observed when targeting a non-essential gene. Of 367 note, penicillin does not kill chlamydiae but blocks cell division (63, 64), which keeps the 368 organism transcriptionally in an RB-like state (65). This suggests that knocking down an 369 essential gene(s) puts selective pressure on the chlamydiae to lose the plasmid encoding the 370 CRISPRi system. This has important ramifications for long-term experiments and functional 371 analyses. Nevertheless, the CRISPRi system represents a significant advance for our ability to 372 study essential systems in this obligate intracellular bacterium. 373 In conclusion, we have demonstrated the importance of the ClpP2X Ctr system to 374 chlamydial development, but many questions remain unanswered. These include why ClpP2 Ctr 375 and ClpX Ctr may serve independent purposes and what substrates this system may be targeting. 376 Additionally, we need to identify any cofactors, chaperones, adaptor proteins, or a lack thereof 377 that may be pertinent to this system. We plan to dissect the structural motifs of ClpP2 Ctr and 378 not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission.
The copyright holder for this preprint (which was this version posted December 8, 2019. ; https://doi.org/10.1101/868620 doi: bioRxiv preprint ClpX Ctr to determine if any of the noted differences from other bacterial Clp proteins may alter 379 activity, which may aid in our goal of further functional assessment. Finally, we need to continue 380 experimentation to address our overarching hypothesis that protein turnover plays a role in 381 chlamydial differentiation, and that the Clp system is a significant aspect of this model. Overall, 382 we conclude that the chlamydial ClpP2X Ctr system is critical to the development of these 383 obligate intracellular bacteria. 384 not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. The PCR products were then incubated with a pDONR TM 221 entry vector (containing attP 417 recombination sites) in the presence of BP Clonase II (Invitrogen) to insert the gene via flanking 418 attP recombination sites and remove the ccdB insert, resulting in an entry vector containing the 419 gene of interest flanked by attL sites. These constructs were transformed into DH5α chemically 420 competent E. coli and plated onto kanamycin-containing LB agar. Plasmid was isolated and used 421 for the LR reaction into one of three destination vectors (pST25-DEST, pSNT25-DEST, or 422 pUT18C-DEST). The same entry vector for any given gene was used for all three LR reactions 423 to insert into the destination vector. Entry vector and destination were incubated in a 1:1 ratio. 424 DH5α E. coli were transformed with 2 μL of the reaction mix. Purified plasmid from an 425 individual colony was sequence verified prior to use in the BACTH assay (see below). 426 Constructs for chlamydial transformation were created using the HiFi Cloning (New 427 England Biolabs) protocol. Primers were designed to add a poly-Histidine (6xHis) tag to the 428 gene of interest with the overlap to insert into the shuttle vector. Primers were generated using 429 the NEBuilder ® assembly tool available from New England BioLabs (http://nebuilder.neb.com). 430 not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission.

Eurofins Genomics. Sequence verified plasmids were transformed into dam-/dcm-E. coli (New 441
England BioLabs) to produce demethylated plasmid, which was verified as described earlier 442 prior to transformation into C. trachomatis (see below). 443 For mutation of ClpX Walker B motif, Q5 mutagenesis (New England BioLabs) was 444 used. Primers were designed encoding the E187A mutation for PCR linearization of the plasmid. 445 ClpX BACTH constructs were used as a template for the PCR amplification, and plasmids were 446 re-circularized by KLD reaction. The resulting reactions were transformed into DH5α E. coli for 447 plasmid production. Plasmids were isolated, and mutations were verified by Sanger sequencing 448 (Eurofins Genomics) prior to use in the BACTH system. These plasmids also served as template 449 for the PCR reactions to produce PCR products for insertion of the mutant clpX gene into the 450 pTLR2 plasmid. 451 Strains created or used in this study are listed in the supplementary material. Transformed 452 E. coli strains were maintained on LB agar plates, with antibiotics as necessary. To extract 453 not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. 200 mM KCl, 20 mM MgCl 2 , and 10% glycerol) was performed using a Millipore Amicon Ultra 472 15 filtration units (3 kDa cut-off). ClpX proteins were quantified using the Bio-Rad Protein 473 assay, assessed for purity on 10% SDS-PAGE gels with Coomassie staining (Suppl. Fig. 6), and 474 identified using anti-His-tag western blot. Blotting was performed using a mouse monoclonal 475 not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission.

Assessment of ClpX ATPase activity in vitro. 486
A 49 µl reaction containing 1.5 µg of recombinant wild-type ClpX or ClpX (E187A) in ATPase 487 assay buffer (see above) was preincubated for 10 minutes at room temperature without ATP. 488 Next, ATP dissolved in ATPase assay buffer was added to 1 mM giving a final volume of 50 µl, 489 and the reaction was incubated at 30⁰ C for 2 hours. After the 2 hours, 200 µl of BIOMOL Green 490 reagent (Enzo Life Sciences) was added and incubated at room temperature for 20 minutes. The 491 absorbance of each reaction was then measured at 620nm using a BioTek Synergy HT plate 492 reader. Reactions were performed in duplicate at least four times with at least two independent 493 protein preparations. 494 495

Determining Protein-Protein Interactions with the BACTH System: The Bacterial Adenylate 496
Cyclase Two-Hybrid (BACTH) assay was utilized to test interactions between wild-type and 497 mutant ClpX (86). The genes of interest are translationally fused to one of either subunit, 498 not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission.
The copyright holder for this preprint (which was this version posted December 8, 2019. ; https://doi.org/10.1101/868620 doi: bioRxiv preprint denoted as T18 and T25, of the B. pertussis adenylate cyclase toxin, which can complement 499 adenylate cyclase deficient (Δcya) DHT1 E. coli. Wild-type and mutant clpX genes cloned into 500 one of the pST25, pSNT25, or pUT18C Gateway ® vectors was tested for both homotypic and 501 heterotypic interactions (9, 82). Plasmids from each background were co-transformed into 502 chemically competent DHT1 E. coli, which were plated on a double antibiotic minimal M63 503 medium selection plate supplemented with 0.5 mM IPTG for induction of the protein, 40 μg/mL 504 Xgal, 0.04% casein hydrolysate, and 0.2% maltose. Leucine zipper motifs were used for controls 505 in pKT25 and pUT18C backgrounds on the appropriate antibiotic selection plates because these 506 have been previously shown to interact (87). Blue colonies, indicative of positive interaction, 507 were screened using the β-galactosidase assay. Random positive colonies were selected and 508 grown in M63 minimal media with the appropriate antibiotics. 0.1% SDS and chloroform were 509 used to permeabilize the bacteria prior to addition of 0.1% o-nitrophenol-β-galactoside (ONPG). 510 1 M NaHCO 3 was used to stop the reaction after precisely 20 minutes of incubation at room 511 temperature. Absorbance at the 405 nm wavelength was recorded and normalized to bacterial 512 growth (OD 600 ), dilution factor, and time (in minutes) of incubation prior to stopping the 513 reaction. Totals were reported in relative units (RU) of β-galactosidase activity. 514 515 Chlamydial Transformation: The protocol followed was a modification of the method 516 developed by Mueller and Fields (88) and as previously described (17). For transformation, 10 6 517 C. trachomatis serovar L2 EBs (25667R) naturally lacking the endogenous plasmid were 518 incubated with 2 μg of unmethylated plasmid in a volume of 50 μL CaCl 2 at room temperature 519 for 30 minutes. Reaction volume was sufficient for one well of a six well plate of McCoy mouse 520 fibroblasts. Transformants were mixed with 1 mL of HBSS and added to 1 mL of HBSS in a six 521 not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission.
The copyright holder for this preprint (which was this version posted December 8, 2019. ; https://doi.org/10.1101/868620 doi: bioRxiv preprint well plate. The plates were centrifuged at room temperature for 15 minutes, 400 xg. The plate 522 was then incubated at 37° C for 15 minutes. After incubation, the HBSS was aspirated and 523 replaced with antibiotic-free DMEM+10% FBS. 8 hours post-infection, the media was replaced 524 with DMEM containing 1 μg/mL cycloheximide and 1 U/mL penicillin. Cells infected with 525 transformants were passaged every 48 hours until a population of penicillin resistant bacteria was 526 established. EBs were harvested and frozen in sucrose/phosphate (2SP; (66)) solution at -80° C. 527 528

Immunofluorescence and Inclusion Forming Unit Analysis: C. trachomatis transformants 530
containing plasmids encoding the 6xHis-tagged protein of interest were used to infect a confluent 531 monolayer of HEp2 cells. Penicillin treatment was maintained throughout the duration of the 532 infection. At 10 hpi, samples were induced or not with 10 nM anhydrotetracycline (aTc). At the 533 given timepoints, three wells of a 24 well plate were scraped in 2SP, vortexed with three 1 mm 534 glass beads, and frozen at -80 C. At the same timepoint, a coverslip was fixed in 3.25% 535 formaldehyde and 0.025% glutaraldehyde for two minutes, followed by permeabilization with 536 cold 90% methanol for one minute. Coverslips were labeled with primary goat anti-major outer 537 membrane protein (MOMP; Meridian, Cincinnati, OH), rabbit anti-6xHis (Abcam, Cambridge, 538 MA), and DAPI. Appropriate donkey secondary antibodies were used (Invitrogen, Carlsbad, 539 CA). Images were acquired on an Axio ImagerZ.2 equipped with Apotome.2 optical sectioning 540 hardware and X-Cite Series 120PC illumination lamp. Frozen IFU samples were titrated onto a 541 fresh monolayer of HEp2s without antibiotics. At 24 hpi, samples were fixed with methanol for 542 10 minutes, stained for MOMP, and enumerated. 543 544 not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission.

Genomic DNA Isolation and qPCR Enumeration of Genomic Equivalents:
At 24 or 48 hpi, 545 one well of a six well plate was scraped into the media overlay and pelleted at 17000 xg, 4 C for 546 15 minutes. Each sample was resuspended in 500 L of cold PBS, frozen three times at -80 C, 547 and processed using the Qiagen DNeasy Blood and Tissue Kit according to the manufacturer's 548 specifications. DNA concentrations were assessed using a spectrophotometer prior to dilution 549 down to 5 ng/L. 5 L of the resulting dilution was used for a 25 L qPCR reaction volume 550 using SYBR ® Green PCR Master Mix (Applied Biosystems). Each reaction was performed in 551 triplicate. A standard curve using Ctr L2 genomic DNA was generated for interpolation of 552 sample Ct values. This experiment was performed three times for three biological replicates. 553 554

Analysis of HctB Levels Upon Clp Overexpression:
At 24 or 48 hpi, one well of a six well plate 555 per test condition was rinsed twice with HBSS. To lyse the cells, 500 L of denaturing lysis 556 buffer (8 M Urea, 10 mM Tris, 2.5% 2-mercaptoethanol, 1% SDS) was added to each well and 557 incubated for 15 minutes at room temperature. 300 units of Universal Nuclease (Pierce) per mL 558 of lysis buffer was added immediately prior to addition to the wells. Following incubation, 559 samples were centrifuged at 17000 xg, 4 C for 15 minutes to remove any insoluble material. 560 Samples were quantitated using the EZQ Protein Quantitation Kit (Pierce). 50 g of each sample 561 was run in a 4-20% gradient SDS-PAGE gel (BioRad) and transferred to a PVDF 0.45 m pore 562 size membrane for 1 h at 300 mA. The membrane was probed using goat anti-MOMP (Meridian) 563 and rabbit anti-HctB (generously provided by Dr. T. Hackstadt, NIH) primary antibodies 564 followed by staining with donkey anti-goat 680 and donkey anti-rabbit 800 (LI-COR) secondary 565 antibodies. The membrane was imaged on an Azure c600 imaging system. The channels were 566 gray-scaled and equally contrast corrected, and the resulting images were used for integrated 567 not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission.
The copyright holder for this preprint (which was this version posted December 8, 2019. ; https://doi.org/10.1101/868620 doi: bioRxiv preprint density measurement with FIJI software (89). To assess relative HctB levels, the HctB integrated 568 density of each sample was normalized to its respective MOMP integrated density to avoid bias 569 due to lower overall organism numbers. The ratios were then used to compare induced versus 570 uninduced relative HctB levels. These experiments were performed three times for a total of 571 three biological replicates. 572 573

Transmission electron microscopy (TEM) assessment of the effect of inactive Clp 574
overexpression. Samples were infected and induced as previously discussed (see above). At 48 575 hpi, samples were fixed using 2% Glutaraldehyde, 2% Formaldehyde in 0.1M Sorensen's 576 phosphate buffer, pH 7.2. Samples were then stained post-fixation in 1% Osmium Tetroxide in 577 water for 1 hour. Samples were dehydrated in an Ethanol series 50%, 70%, 90%, 95%, 100% 3 578 changes of 100%, all steps 15 minutes each and were then soaked in Propylene Oxide 100% 3 579 changes for 15 minutes each. Samples were left overnight in a fume hood in a 1:1 mixture of 580 Propylene Oxide and Embed 812. The following day the samples were placed in molds with 581 fresh Embed 812 and polymerized overnight in an oven set at 65 C. Blocks were thin sectioned 582 90 nanometers thick on a Leica UC6 Ultramicrotome using a Diatome diamond knife. Sections 583 were placed on uncoated 200 mesh copper grids and stained with 2% Uranyl Acetate and 584 Reynold's Lead Citrate. Sections were examined on a FEI Tecnai G2 TEM operated at 80Kv. 585 586 Confirmation of clpP2X and incA knockdown. Briefly, two wells of a six-well plate per 587 condition were infected with pBOMBLCRia-clpP2X transformed CtrL2 at an MOI of 0.8. At 588 either four or ten hpi, samples were or were not induced with 10 nM aTc. At each given 589 timepoint, total RNA was collected using Trizol reagent (Invitrogen) and was extracted with 590 not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission.
The copyright holder for this preprint (which was this version posted December 8, 2019. ; https://doi.org/10.1101/868620 doi: bioRxiv preprint chloroform as described previously (17,85,(90)(91)(92). The aqueous layer was precipitated using 591 isopropanol, as per the manufacturer's instructions. Samples were DNase treated using the 592 TURBO DNA-free kit (Ambion), and 1 µg of the resulting RNA was reverse transcribed using 593 SuperScript III reverse transcriptase (Invitrogen). Equal volumes of cDNA were loaded for each 594 qPCR reaction. To extract genomic DNA, one well per condition was harvested and processed 595 using the DNeasy blood and tissue kit (Qiagen) according to the manufacturer's instructions as 596 noted above. Samples were diluted to 5 ng/µL, and 5 µL of the resulting dilution was used per 597 qPCR reaction. cDNA and gDNA samples were quantified using 25 µL reactions with 2x SYBR 598 PowerUP Green Master Mix (Invitrogen) analyzed on a QuantStudio 3 (Applied Biosystems) 599 thermal cycler using the standard cycling conditions. A standard curve using purified wild-type 600 infected at an MOI of 0.8 with either pBOMBLCRia-clpP2X or pBOMBLCRia-incA 611 transformed into CtrL2. Samples were induced or not at 4 hpi and were harvested, fixed, and 612 titered as previously described. Each titration was fixed using 4% formaldehyde and 0.025% 613 not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission.
The copyright holder for this preprint (which was this version posted December 8, 2019. ; https://doi.org/10.1101/868620 doi: bioRxiv preprint glutaraldehyde to preserve GFP fluorescence. IFU counts of GFP positive inclusions are 614 displayed as a percentage of the uninduced sample at the given timepoint. Plasmid retention for 615 each condition is displayed as the percent of GFP positive to total number of inclusions for each 616 condition. 617 618

Effect of ClpX-targeting compounds on chlamydial growth and host cell viability. Stocks of 619
ClpX-specific inhibitor 334 and its derivative, 365, were synthesized as previously reported (44)  928 not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission.

B.
not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission.

ClpX
ClpX ( T18 not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission.
The copyright holder for this preprint (which was this version posted December 8, 2019. ; https://doi.org/10.1101/868620 doi: bioRxiv preprint not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission.

Fig. 4
not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission.
The copyright holder for this preprint (which was this version posted December 8, 2019. ; https://doi.org/10.1101/868620 doi: bioRxiv preprint not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission.