A nutrient-dependent division antagonist is regulated post-translationally by the Clp proteases in Bacillus subtilis

Background Changes in nutrient availability have dramatic and well-defined impacts on both transcription and translation in bacterial cells. At the same time, the role of post-translational control in adaptation to nutrient-poor environments is poorly understood. Previous studies demonstrate the ability of the glucosyltransferase UgtP to influence cell size in response to nutrient availability. Under nutrient-rich medium, interactions with its substrate UDP-glucose promote interactions between UgtP and the tubulin-like cell division protein FtsZ in Bacillus subtilis, inhibiting maturation of the cytokinetic ring and increasing cell size. In nutrient-poor medium, reductions in UDP-glucose availability favor UgtP oligomerization, sequestering it from FtsZ and allowing division to occur at a smaller cell mass. Results Intriguingly, in nutrient-poor conditions UgtP levels are reduced ~ 3-fold independent of UDP-glucose. B. subtilis cells cultured under different nutrient conditions indicate that UgtP accumulation is controlled through a nutrient-dependent post-translational mechanism dependent on the Clp proteases. Notably, all three B. subtilis Clp chaperones appeared able to target UgtP for degradation during growth in nutrient-poor conditions. Conclusions Together these findings highlight conditional proteolysis as a mechanism for bacterial adaptation to a rapidly changing nutritional landscape. Electronic supplementary material The online version of this article (10.1186/s12866-018-1155-2) contains supplementary material, which is available to authorized users.


149
UgtP accumulation is subject to nutrient-dependent post-translational regulation 150 In our initial investigation, we observed that the intracellular concentration of a UgtP-151 6XHis fusion protein was several times lower when cells were cultured under nutrient-  Consistent with our previous findings, UgtP-His levels increased linearly with growth 160 rate, as evidenced by a quantitative immunoblot probed with an α-His antibody (Fig. 1). 161 The intracellular concentration of UgtP-His was ~ 3-fold lower in cells cultured in 162 minimal sorbitol, the most nutrient-poor condition, than in those cultured in LB, the most 163 nutrient-rich condition. To control for the possibility that the His-tag was impacting the 164 stability of UgtP-His, we also measured YFP-UgtP levels in a strain expressing the fusion 165 protein from a xylose-inducible promoter (P xyl -yfp-ugtP), cultured in both LB + 0.5% 166 xylose and minimal sorbitol + 0.5% xylose. As we observed with UgtP-His, YFP-UgtP 167 levels were ~3-fold lower in minimal sorbitol compared to LB supporting a model in 168 which UgtP (and not the His or YFP tag) is the primary target for degradation during 169 growth in nutrient-poor medium (Additional File 1: Fig. S1). 170 To determine if UgtP is subject to transcriptional or post-transcriptional modes of 172 regulation, we generated two ugtP-lacZ fusion constructs. In the first construct, a reporter 173 for ugtP transcription, the 700 base pairs immediately upstream of the ugtP start codon 174 were fused to lacZ, leaving the lacZ Shine-Dalgarno sequence intact. In the second was inversely proportional to growth rate. lacZ expression from the transcriptional fusion 183 was 2-fold higher in cells cultured in minimal sorbitol than LB ( Fig. 2A) and expression 184 from the translational fusion was 4-fold higher in minimal sorbitol than LB (Fig. 2B). 185 qRT-PCR data also indicated that ugtP expression is elevated during growth in nutrient-186 poor conditions (Fig. 2C). ugtP levels were 1.3-fold higher in minimal glucose, 1.7-fold 187 higher in minimal glycerol, and 2.5-fold higher in minimal sorbitol compared to LB. In 188 all, these data strongly support a model where nutrient-dependent changes in UgtP  To understand the mechanism responsible for the post-translational regulation of UgtP, 196 we sought to identify the protease responsible for its degradation. As an initial step, we 197 screened B. subtilis strains defective in five well-studied proteases: YluC, CptA, ClpP, 198 YvjB, and Lon for aberrant UgtP regulation. We rationalized that if one of these 199 proteases were responsible for the growth rate-dependent degradation of UgtP, then UgtP 200 would inappropriately accumulate in its absence during growth in nutrient-poor 201 conditions. For this experiment we used the P xyl -ugtP-his construct described above to 202 enhance our ability to detect UgtP by quantitative immunoblot.  (Fig. 3A). UgtP-His levels 207 were approximately 5-fold higher during growth in minimal sorbitol in the ∆clpP strain 208 than in the parental strain or in the four other protease mutants. ClpX. Since each chaperone identifies a unique set of substrates with limited target 216 overlap, we reasoned that either ClpC, ClpE, or ClpX would be responsible for growth 217 rate-dependent degradation of UgtP. To distinguish the Clp chaperone involved, we 218 examined accumulation of UgtP-His from a xylose-inducible promoter in ∆clpC, ∆clpE, 219 and ∆clpX cells cultured in nutrient-poor medium. UgtP-His levels in minimal sorbitol. UgtP-His accumulated to levels similar to that 223 observed in the ∆clpP mutant only in a strain defective for all three chaperones (Fig. 3B). 224 These data suggest that all three Clp chaperones function redundantly to control UgtP To confirm that the Clp proteases are indeed responsible for degradation of UgtP, as our 231 data suggested, we employed an in vivo proteolysis assay similar to that described in (29).

232
Briefly, we cultured P xyl -ugtP-his and the congenic ∆clpP strains in minimal sorbitol with 233 xylose, inhibited new protein synthesis once mid-log was reached, sampled cells every 30 234 minutes for 2 hours, and then measured UgtP-His levels via quantitative immunoblot.    250 In an effort to identify regions of UgtP required for interaction with the Clp chaperones, 251 we screened a series of UgtP mutants for sensitivity to Clp-mediated degradation in vivo.

252
Of particular interest were regions of UgtP that mediate interaction with UDP-glucose, 253 FtsZ, or itself. Three putative UgtP mutants were employed for this experiment: one 254 defective in its putative uracil-binding size (∆URA, F112A V117A), one defective in the 255 putative hexose-binding site (∆HEX, E306A N309A), and a putative oligomerization 256 mutant (∆OLI, I142A E146A). Mutations were generated in conserved residues based on 257 structural data from the UgtP homologue, MDGD synthase (30). Structural data indicate 258 that MDGD has two binding sites for UDP-glucose conserved in UgtP: one to coordinate 259 interactions with the uracil moiety on UDP-glucose and one for the hexose moiety. In 260 addition, structural analysis identified a dimerization site on MDGD. We speculated that 261 conserved residues within the analogous region of UgtP are involved in oligomerization.

262
All three UgtP mutants behaved as predicted based on structural and biochemical data 264 from MDGD synthase (30). Both UgtP∆URA and UgtP∆HEX are defective as sugar 265 transferases, exhibit a punctate localization pattern during growth in nutrient-rich 266 medium, and fail to complement a ugtP null strain for cell size, all of which is consistent 267 with increased self-association and a loss of interaction with FtsZ (24). In support of our prior to immunoblotting (Fig. 4). This difference may reflect conformational changes in  Of the three mutants, only UgtP-His(∆HEX) exhibited ClpP-dependent differential in 283 accumulation compared to wild type UgtP-His, suggesting hexose binding might protect 284 UgtP from proteolysis. In lysate from cells cultured in LB, UgtP-His(∆HEX) 285 accumulated to only ~35% of UgtP-His(WT) levels (Fig. 4A). Even in minimal sorbitol 286 UgtP-His(∆HEX) levels were only ~65% of UgtP-His(WT) (Fig. 4B). Importantly, each 287 of the constructs exhibited congruent mRNA levels measured by qRT-PCR and UgtP-288 His(∆HEX) is at WT levels in a ∆clpP background strain in both LB and minimal 289 sorbitol, indicating that the change in stability is directly due to ClpP-mediated 290 degradation (Additional File 4: Fig. S4). In contrast, UgtP-His(∆URA) and UgtP-291 His(∆OLI) did not exhibit ClpP-dependent differential in accumulation, suggesting that 292 neither interaction with FtsZ nor homo-oligomerization impact Clp recognition.

294
Together these data suggest that interaction with the hexose moiety of UDP-glucose, 295 which we have measured as being more prevalent when cultured in nutrient-rich   Based on our in vivo data supporting a model in which hexose binding may shield UgtP 326 from Clp-mediated proteolysis (Fig. 4), we speculated that adding UDP-glucose or 327 simply glucose to the ClpXP proteolysis assays might hinder UgtP degradation. To test 328 this model we added UDP-glucose or glucose 6-phosphate in excess to the ClpXP in vitro 329 proteolysis reactions described above. However, we saw no difference in UgtP 330 proteolysis in the presence of either sugar (Fig. 5B).

332
Clp-mediated UgtP degradation during growth in nutrient-poor medium does not 333 significantly impact cell size or diglucosyl-diacylglycerol production 334 In an effort to illuminate the potential "rationale" for Clp mediated degradation of UgtP    To test this hypothesis, we took advantage of two strains capable of producing a range of Somewhat surprisingly, we did not observe a significant difference in cell size between  Average cell length of strains BH736 (∆ugtP; P xyl -ugtP-his) and BH12 (P ugtP -ugtP-his; 364 P xyl -ugtP-his) cultured in minimal sorbitol ± 0.5% xylose, determined by measuring the 365 distance between the mid-points of adjacent cell wall septa (n = 600, error = SD). when growing P xyl -ugtP strains). Briefly, we purified membranes from these strains, 373 performed a methanol:chloroform lipid extraction, separated the lipids using thin-layer chromatography, stained for DGD using iodine gas, and quantified DGD using ImageJ.

375
Lipids were also extracted from ∆ugtP cells for use as a negative control.   (Fig. 4). These data suggest ligand binding is largely irrelevant to UgtP stability and the immediately have it proteolyzed, consuming ATP at each step. We were unable to 471 identify a phenotypic explanation for the nutrient-dependent degradation of UgtP. UgtP 472 levels had no effect on cell size during growth in nutrient-poor medium and diglucosyl-473 diacylglycerol production was largely independent of ClpP ( Fig. 6 and Table 1). Given    Bio-Rad Trans-Blot ® Turbo TM instrument as described on pg. 15 of the operation manual.

576
Total protein was stained using 1X Ponceau S solution in 5% acetic acid for 5 minutes.

577
Membranes were then blocked with 5% nonfat milk in PBS for 1 hr.

615
Oligonucleotides used in this study are described in (Additional File 11: French press. Lysates were centrifuged at 12,000 g for 20 minutes to pellet membranes.

652
Pellets were weighed and resuspended to 0.4 g/mL in 100 mM sodium citrate pH 4.7.

653
Methanol and chloroform were added to obtain a ratio of 2:1:0.8

794
The datasets used and/or analyzed during the current study are available from the 795 corresponding author on reasonable request.

Additional File 1: YFP-UgtP is degraded in nutrient-poor growth conditions
Quantitative immunoblot of ectopically expressed YFP-UgtP in nutrient-rich and nutrient-poor growth conditions. PL2423 (Pxyl-yfp-ugtP) was grown in both LB and minimal sorbitol medium. Total protein was used as a loading control via Ponceau staining. Protein levels in LB are set as the reference in the relative expression values shown (n=3, error = SD). Representative immunoblot of UgtP-His from both BH10 (Pxyl-ugtP-his) and BH129 (Pxyl-ugtP-his ; ∆clpP) cultured in S750 sorbitol + 0.5% xylose. Spectinomycin was added to cells to inhibit new protein synthesis, and then cultures were sampled every 30 minutes for 2 hours. Ponceau staining was performed as a loading control.  Additional File 4: ugtP∆HEX is expressed at the same level as ugtP and UgtP∆HEX is stabilized in a ∆clpP background (A) ugtP transcript levels of WT versus the ∆URA, ∆HEX, and ∆OLI mutants cultured in LB + 0.5% xylose measured by qRT-PCR (n = 3, error = SD). The Pxyl-ugtP-his variants in a ∆clpP background (BH767, BH769, BH771, BH773) grown in either (B) LB + 0.5% xylose or (C) minimal sorbitol + 0.5% xylose. FtsZ is used as a loading control. The relative expression (%), using WT UgtP as the reference, is shown below (n = 3, error is SD). A Coomassie stained gel examining ClpXP-mediated proteolysis of a known substrate, Spx (15.5 kDa), and a non-targeted protein, Thioredoxin-His (17.1 kDa). A mixture of 3μM ClpX, 6μM ClpP, 5mM ATP, and either 3μM of Spx or Thio-His was incubated for 60 minutes. Western blots of UgtP-His from BH736 (∆ugtP; Pxyl-ugtP-his) and BH12 (PugtP-ugtP-his ; Pxyl-ugtP-his) cells cultured in minimal sorbitol, with and without xylose, using total protein as a loading control via Ponceau staining. Protein levels of BH12 -xylose are set as the reference in the relative expression values shown, as they represent WT expression (n = 3, error = SD). of protein (no frozen stocks were used). Briefly, 1L of LB medium was 49 inoculated 1:100 with overnight culture from a single colony. Cells were grown at 37°C 50 with the exception of ClpP and ClpX, which were grown at room temperature. When A600 ~ 51 0.6, cells were induced with isopropyl 1-thio-β-D-galactopyranoside to a final 52 concentration of 1mM. Cells were grown for an additional 4-8 h and then cells were 53 harvested by centrifugation, and cell pellets were stored at -80°C. ClpP and ClpX were purified 54