A KDPG sensor RccR governs Pseudomonas aeruginosa carbon metabolism and aminoglycoside antibiotic tolerance

Abstract Pseudomonas aeruginosa harbors sophisticated transcription factor (TF) networks to coordinately regulate cellular metabolic states for rapidly adapting to changing environments. The extraordinary capacity in fine-tuning the metabolic states enables its success in tolerance to antibiotics and evading host immune defenses. However, the linkage among transcriptional regulation, metabolic states and antibiotic tolerance in P. aeruginosa remains largely unclear. By screening the P. aeruginosa TF mutant library constructed by CRISPR/Cas12k-guided transposase, we identify that rccR (PA5438) is a major genetic determinant in aminoglycoside antibiotic tolerance, the deletion of which substantially enhances bacterial tolerance. We further reveal the inhibitory roles of RccR in pyruvate metabolism (aceE/F) and glyoxylate shunt pathway (aceA and glcB), and overexpression of aceA or glcB enhances bacterial tolerance. Moreover, we identify that 2-keto-3-deoxy-6-phosphogluconate (KDPG) is a signal molecule that directly binds to RccR. Structural analysis of the RccR/KDPG complex reveals the detailed interactions. Substitution of the key residue R152, K270 or R277 with alanine abolishes KDPG sensing by RccR and impairs bacterial growth with glycerol or glucose as the sole carbon source. Collectively, our study unveils the connection between aminoglycoside antibiotic tolerance and RccR-mediated central carbon metabolism regulation in P. aeruginosa, and elucidates the KDPG-sensing mechanism by RccR.


Introduction
Pseudomonas aeruginosa is a Gram-negative opportunistic pathogen that causes severe infections in humans, in particular in cystic fibrosis patients ( 1 ).Aminoglycoside antimicrobials, such as tobramycin, gentamicin, neomycin and amikacin, are vital for treating P. aeruginosa infections by targeting the 30S ribosome and inhibiting protein synthesis to achieve bactericidal effect ( 2 ,3 ).Pseudomonas aeruginosa has become multidrug-resistant to widely used antibiotics by developing various antibiotic resistance mechanisms, including biofilm formation, persister formation, target modification, enzymatic drug modification, multidrug efflux systems and low membrane permeability ( 2 ,4 ).
Pseudomonas aeruginosa can rapidly adjust its metabolism for acclimating to the changes in the host environment, owing to the large number of regulatory proteins encoded by its genome ( 5 ,6 ).Transcription factors (TFs) sense the environmental stimulants and alter the expression of genes whose products are involved in drug resistance, metabolic pathways, pathogenicity of bacteria and so on.For example, RpiR family members usually control the expression of enzymes in different metabolic pathways (7)(8)(9)(10)(11)(12), and MarR family numbers regulate the antibiotic resistance of bacteria (13)(14)(15).
Previous studies showed that certain metabolic stimulants, such as glucose, mannitol, fructose and pyruvate, can increase the susceptibility of Esc heric hia coli persisters to aminoglycoside antibiotics, and fructose also can sensitize Staphylococcus aureus persisters to aminoglycoside antibiotic killing ( 16 ).For P. aeruginosa , fumarate can boost the antimicrobial activity of tobramycin, and in contrast, glyoxylate can promote bacterial tolerance to tobramycin ( 17 ).Moreover, inactivation of the glycolytic enzyme, triosephosphate isomerase ( tpiA ), significantly reduced the bacterial resistance to aminoglycoside antibiotics ( 18 ).These previous findings well demonstrated that bacterial carbon flux is crucial for its tolerance to aminoglycoside antibiotics.However, the specific regulatory elements in P. aeruginosa that respond to these central metabolites and switch the antimicrobial susceptible states (tolerant or susceptible) remain elusive.
In this study, we screened a P. aeruginosa TF inactivation library generated by site-specific transposon-assisted genome engineering (STAGE) technology ( 19 ), and identified that a mutant carrying the inactivation of the PA5438 ( rccR ) gene, a member of the RpiR TF family, was significantly enriched under tobramycin selection.Transcriptomic analyses indicated that RccR can alter the bacterial carbon flux via regulating the expression of metabolic enzymes in pyruvate synthesis ( aceE and aceF ) and glyoxylate shunt pathway ( aceA and glcB ), and consequently, affecting the tolerance to aminoglycoside antibiotics.Moreover, through biochemical approaches, we confirmed that RccR responds to a metabolic intermediate of the Entner-Doudoroff pathway, 2-keto-3-deoxy-6-phosphogluconate (KDPG), for its regulatory role.To further understand the molecular mechanism of KDPG responding by RccR, we solved the structure of the RccR / KDPG complex and explored the functions of key residues that interact with KDPG.Collectively, these findings shed light on the molecular mechanism of RccR-mediated metabolic regulation and provided a plausible model for the development of aminoglycoside tolerance in P. aeruginosa .

Bacterial strains, plasmids, primers and growth conditions
The strains and plasmids used in this study are listed in Supplementary Table S2 .The primers used in this study are listed in Supplementary Table S3 .The E. coli strains and P. aeruginosa strains were grown at 37 • C in Luria-Bertani broth (LB).Antibiotics were added at the following concentrations: 15 μg / ml tetracycline, 50 μg / ml carbenicillin and 50 μg / ml kanamycin for E. coli strains; 100 μg / ml tetracycline and 150 μg / ml carbenicillin for P. aeruginosa strains.

Construction of bacterial strains and plasmids
The pACRISPR / pCasPA system was used to construct the rccR -deletion strain in PAO1 as previously described ( 20 ).The pA CRISPR -NN1-rccR plasmid was constructed to contain a spacer targeting the rccR gene and a repair template consisting of ∼500-bp flanking sequences at the upstream and the downstream of the rccR gene.The pA CRISPR -NN1-rccR was then electroporated into the PAO1 wild-type electrocompetent cells containing pCasPA.Deletion of rccR was validated by polymerase chain reaction (PCR) and Sanger sequencing.Plasmids were then cured in successfully mutated cells.The aceE -deletion strain was similarly constructed.
To complement the rccR mutation, the pA CRISPR -NN2-rccR plasmid was constructed to contain a spacer targeting the deletion junction and a repair template consisting of the complete rccR gene with a synonymous mutation, and ∼500bp flanking DNA sequences at the upstream and the downstream of the rccR gene.The pA CRISPR -NN2-rccR plasmid was electroporated into the rccR -deletion electrocompetent cells containing pCasPA.Complementation of rccR was validated by PCR and Sanger sequencing.Plasmids were then cured in successfully complemented cells.To introduce certain residue substitution in RccR, mutations were created on the pA CRISPR -NN2-rccR plasmid by circular polymerase extension cloning (CPEC) ( 21 ) using 2 × Taq Master Mix (Gen-Script).Complementation strains of certain RccR mutants were similarly created.
For expression and purification of RccR, the coding sequence of rccR gene from P. aeruginosa PAO1 was inserted into a pET28a backbone with a C-terminal His 6 tag using Hieff Clone ® Universal One Step Cloning Kit (Yeasen).Mutations of rccR gene were then introduced into pET28a-RccRhis via CPEC.The constructed plasmids were extracted by using TIANprep Mini Plasmid Kit (TIANGEN).
For overexpressing the aceA gene or glcB gene in PAO1 wild-type strain, the coding sequences of aceA gene and glcB gene were inserted into a pAK1900 backbone with rpsL promoter of PAO1 strain, respectively.Then, the pAK1900rpsl-aceA and pAK1900rpsl -glcB plasmids were electroporated into the PAO1 electrocompetent cells.

Protein expression and purification
The pET28a-RccR-his plasmid was transformed into E. coli BL21(DE3), and the cells were grown in LB with 50 μg / ml kanamycin at 37 • C until the optical density at 600 nm (OD 600 ) reached 0.6.The isopropylβ-dthiogalactopyranoside was supplemented at a final concentration of 0.25 mM to induce the protein expression overnight at 16 • C. Cells were harvested by centrifugation and the pellet was resuspended in buffer A (10 mM Tris-HCl, pH 7.5, 1 M NaCl and 1 mM DTT), and then the suspension was disrupted by sonication and clarified by centrifugation.The supernatant was filtered and then loaded onto a 5-ml HisTrap Ni-NTA column (Cytiva).The column was washed with buffer A containing 62.5 mM imidazole for unbound proteins, and the Histagged RccR protein was eluted with buffer A containing 500 mM imidazole.The eluted protein was concentrated to 2 ml and loaded onto a HiLoad 16 / 600 Superdex 200pg column (Cytiva) for further elution with buffer A. The purified protein was concentrated for subsequent experiments.

Protein crystallization, data collection and structure determination
The RccR / KDPG complex was crystallized at 16 • C by using the sitting-drop vapor-diffusion method.The RccR protein (6 mg / ml in 400 mM NaCl, 10 mM Tris-HCl, pH 7.5) was incubated with KDPG (the molar ratio of RccR and KDPG was ∼1:10) on ice for 30 min.One microliter of RccR / KDPG complex solution was mixed with an equal volume of reservoir solution containing 0.15 M dl -malic acid (pH 7.0), 0.1 M imidazole (pH 7.0) and 22% polyethylene glycol monomethyl ether 550 (v / v).The crystals were protected in liquid nitrogen before data collection.
The data were collected at BL18U1 beamline of the Shanghai Synchrotron Radiation Facility and processed by HKL3000 ( 22 ).The phase of the RccR / KDPG complex structure was determined by Phaser ( 23 ) from CCP4i ( 24 ) using the structure of Vibrio vulnificus NanR (PDB code: 4IVN) ( 25 ) as the search model.The model was built by Autobuild from PHENIX ( 26 ).The model of RccR / KDPG was refined using Refmac5 from CCP4i and further improved manually by Coot ( 27 ).The final structure figures were prepared by Pymol ( http://www.pymol.org).

Antibiotic screening assay of P. aeruginosa mutant library
The antibiotic screening assay of P. aeruginosa TF mutant library was performed as previously described ( 19 ).The TF mutation inactivation library of P. aeruginosa was constructed using the CRISPR / Cas12k transposition system ( 19 ).A glycerol stock of P. aeruginosa TF mutant library was thawed and mixed on ice.Then, 100 μl of cells were diluted into 100 ml of LB medium with or without 3 μg / ml tobramycin.The culture was shaken at 37 • C overnight and harvested by centrifugation.The genomic DNA was extracted by the Ezup column bacteria genomic DNA purification kit (Sangon, Shanghai) and then subjected to NGS (HaploX Genomics Center, Jiangxi).

Isothermal titration calorimetry
The association constants were determined by the MicroCal ITC200 system (Malvern).The RccR wild-type protein, RccR mutant proteins and the KDPG were prepared in the same buffer containing 10 mM Tris-HCl (pH 7.5) and 350 mM NaCl.The proteins and the KDPG were diluted to a final concentration of 60 and 600 μM, respectively.The KDPG solutions in the syringe were slowly titrated into the reaction cell containing the protein solutions.The whole process of the assay was carried out at 25 • C with a stirring speed of 750 rpm.The ligand solutions were injected 20 times with 120 s intervals between two injections.The raw data were analyzed with the Origin7 software package (Malvern).

Electrophoretic mobility shift assay
The DNA substrates for electrophoretic mobility shift assay (EMSA) were prepared by PCR amplification using FAMlabeled primers from the P. aeruginosa PAO1 genome and then gel purification by a SPARKeasy Gel / PCR Purification Kit (Shandong Sparkjade Biotechnology Co., Ltd).The labeled DNA fragments (10 nM) were incubated with different concentrations of purified protein on ice for 30 min in the reaction buffer (20 mM Tris-HCl, pH 8.5, 150 mM KCl, 5 mM MgCl 2 , 1 mM ED TA, 1 mM D TT, 0.8% Tween 20 and 2.5% glycerol).The mixtures were added with 5 × bromophenol blue loading buffer and then subjected to 6% native polyacrylamide gel electrophoresis in 0.5 × TBE buffer at 120 V for 30 min.Images were acquired by the GelDoc System (Bio-Rad).
The labeled DNA fragments (10 nM), a certain concentration of purified RccR protein and different concentrations of KDPG were incubated on ice for 30 min in the reaction buffer and then separated on 6% native polyacrylamide gels using the same protocol.

Growth curve assays
Two microliters of overnight cultures of different P. aeruginosa PAO1 strains ( ∼OD 600 = 1.4) were diluted into 200 μl of fresh LB medium or MOPS minimal medium and transferred to a BioScreen micro-well plate.The plate was shaken continuously at 37 • C in the automated microbe growth curve analysis system BioScreen C (OY Growth Curves Ltd, Finland) and the OD 600 was measured every 1 h.The MOPS minimal medium was prepared as previously described ( 28 ), and supplemented with 20 mM glucose, 20 mM glycerol and 20 mM acetate as the sole carbon sources, respectively.All experiments were performed in triplicate.

Spotting assay
Overnight cultures of different P. aeruginosa PAO1 strains were subjected to eight consecutive 10-fold dilutions in LB medium.The diluted bacteria were spotted on the LB agar plates supplemented with or without antibiotics.The plates were incubated overnight at 37 • C.

RNA extraction and transcriptome sequencing
Pseudomonas aeruginosa PAO1 wild-type and rccR -deletion mutant strains were cultured in LB medium at 37 • C and harvested at the exponential phase and stationary phase, respectively.The total RNAs were extracted using the MiniBEST Universal RNA Extraction Kit (Takara) and used as templates for reverse transcription to complementary DNA (cDNA) using the TransScript ® One-Step gDNA Removal and cDNA Synthesis SuperMix (TransGen Biotech Co.) according to the manufacturer's instructions.The cDNA products were subjected to transcriptome sequencing by Majorbio (Shanghai).

Quantitative real-time PCR
Quantitative real-time PCR (qRT-PCR) was carried out in the CFX96™ Touch Real-Time PCR Detection System (Bio-Rad) by using ChamQ Universal SYBR qPCR Master Mix (Vazyme).Gene expression level was normalized to the internal reference, the gyrB gene of P. aeruginosa .Three biological replicates with three technical replicates were performed for all experiments.

rccR is associated with aminoglycoside antibiotic tolerance
To explore the TFs associated with aminoglycoside antibiotic tolerance in P .aeruginosa , we treated the TF mutant library that was constructed by CRISPR / Cas12kguided STAGE technology with tobramycin (3 μg / ml) as previously described ( 19 ).The TF library was recovered from frozen cultures and cultured in fresh LB medium or LB containing tobramycin ( Supplementary Figure S1 A).The drugtolerant mutants survived and were enriched during this process, whereas the antibiotic-sensitive mutants were gradually eliminated.We performed next-generation sequencing (NGS) analysis on the samples with and without tobramycin treatment and compared the guide reads of each gene in the tobramycin-treated sample with those in the untreated control sample.We found that disruption of rccR substantially increased P .aeruginosa tolerance to tobramycin (Figure 1 A and Supplementary Figure S1 B), and the guides targeting PA5438 ( rccR ) were enriched (Figure 1 B), indicating that the disruption of rccR indeed contributes to bacterial growth in the presence of tobramycin.
To further validate the rccR -mediated aminoglycoside antibiotic tolerance, we constructed the rccR -deletion mutant and rccR -complementation strains using the CRISPR / Cas9 genome editing system in P .aeruginosa ( 20 ).As shown in Figure 1 C and D, deletion of rccR significantly increased bacterial tolerance to tobramycin, and this phenotype can be restored by introducing a wild-type copy of rccR into the rccR -deletion mutant strain.Moreover, consistent results were observed with the treatment of two additional aminoglycoside antibiotics, gentamicin and amikacin ( Supplementary Figure S1 C  and D).However, we did not observe the growth differences between the wild-type strain and rccR -deletion mutant in the treatment with other classes of antibiotics ( Supplementary Figure S2 A-F), suggesting that the rccR -deletion mutant exclusively increases bacterial tolerance to aminoglycoside antibiotic in P .aeruginosa .

RccR regulates the pyruvate catabolism and glyoxylate shunt pathway
To investigate the regulatory roles of RccR, we performed the transcriptome analysis of the PAO1 wild-type and rccRdeletion mutant strains.As shown in Figure 2 A and B, Supplementary Figure S3 and Supplementary Table S4 , the expression of most of the identified genes was significantly downregulated when rccR was deleted, whereas the expression of aceE , aceF , aceA and glcB was substantially increased (Figure 2 A and B), suggesting that RccR was a transcriptional repressor of these genes.We further measured the mRNA transcription levels of aceE , aceF , aceA and glcB by qRT-PCR, showing a result consistent with the RNA sequencing (RNAseq) result (Figure 2 C), thereby confirming the regulatory role of RccR in pyruvate catabolism ( aceE and aceF ) and glyoxylate shunt pathway ( aceA and glcB ) (Figure 2 D).
A previous study confirmed that the tricarboxylic acid (TCA) cycle and cellular respiratory activity are both essential for aminoglycoside antibiotics lethality ( 17 ).Therefore, we hypothesized that the tolerance of rccR -deletion strain to aminoglycoside antibiotic may be related to the activation of the glyoxylate shunt pathway (Figure 2 D), which disrupted downstream TCA cycle activity.To verify this hypothesis, we overexpressed aceA gene and glcB gene into the PAO1 wild-type strain by a pAK1900 plasmid ( rpsl promoter), respectively, and measured the growth curve in the presence or absence of aminoglycoside antibiotics.As shown in Figure 2 E and Supplementary Figure S4 A and B, overexpression of aceA gene or glcB gene in wild-type strain enhanced the tolerance to aminoglycoside antibiotics compared to wild-type and complementation strains, suggesting that increasing the expression of enzymes in the glyoxylate shunt pathway may lead to aminoglycoside antibiotic tolerance.

RccR is a KDPG-responsive regulator
A previous study demonstrated that KDPG is the signal molecule for the RccR homolog in P. fluorescens ( 8 ).To explore whether KDPG is also an effector of RccR in P. aeruginosa , we performed the isothermal titration calorimetry (ITC) to determine the binding affinity between RccR and different carbon metabolites.The ITC data showed that except for KDPG, no response was observed for citrate, succinate or malonate when they were used for the titration (Figure 3 A and Supplementary Figure S5 A-C).RccR bound to KDPG with a dissociation constant ( K d ) value of 3.31 μM and nearly 1:1 binding stoichiometry (Figure 3 A), indicating that one RccR molecule binds one molecule of KDPG.
To investigate the impact of KDPG binding on RccRmediated regulation, we performed the EMSA between RccR and its operator DNA in the presence or absence of KDPG.EMSA showed that RccR can effectively bind to the promoter DNA, and the binding affinity between RccR and the promoter DNA of aceE was stronger than that of aceA or glcB (Figure 3 B and Supplementary Figure S6 A).Intriguingly, the addition of KDPG significantly relieved the retardation of aceE promoter DNA migration, suggesting that KDPG interfered the binding of RccR to the aceE promoter DNA.In contrast, KDPG addition strengthened the binding affinity of RccR to both promoters of aceA and glcB (Figure 3 C and Supplementary Figure S6 A).These results showed that RccR is a KDPG-sensing regulator with two distinct regulatory effects upon KDPG binding: one is to relieve the repression of pyruvate metabolism and the other is to increase the repression of the glyoxylate shunt pathway ( Supplementary Figure S6 B).

Structural characterizations of the RccR / KDPG complex
To elucidate the detailed KDPG-sensing and regulatory mechanism of RccR, we attempted to determine the crystal structures of apo-RccR and the RccR / KDPG complex.We screened hundreds of crystallization conditions and obtained the crystals of the RccR / KDPG complex.The complex was crystallized with the space group I222 , and the crystal structure was refined to 1.9 Å resolution ( Supplementary Table S1 ).Structural analysis revealed that the RccR / KDPG complex contains four monomers (A-D) with each monomer consisting of two domains (Figure 4 A and B): an N-terminal DNAbinding domain (DBD) and a C-terminal sugar isomerase domain (SIS).The DBD comprises seven α-helices and the SIS forms an α/ β structure, which consists of four-stranded parallel β-sheets and eight α-helices (Figure 4 B and C).The last αhelix is absent in the SIS due to the poor electron density.Each monomer in this complex structure harbors a well-defined signal molecule KDPG (Figure 4 A), which resides in the central pocket surrounded by three individual SISs.
We failed to obtain the crystal of apo-RccR, likely because of the high flexibility of RccR in the absence of KDPG.We thereby predicted the structure of apo-RccR using Colab-Fold (29)(30)(31).As shown in Supplementary Figure S7 A, compared to the structure of the RccR / KDPG complex, apo-RccR displays an open conformation with the two domains relatively scattered.We overlapped SISs in one monomer of apo-RccR with that of the RccR / KDPG complex and observed that all the DBDs were separate and significant conformational changes occurred in SISs of the other three monomers ( Supplementary Figure S7 B).Taken together, these results indicated that KDPG is a specific signal molecule that stabilizes the conformation of RccR protein, binding of which may trigger a conformational change to RccR.

KDPG recognition mechanism by RccR
Extensive interactions between RccR and KDPG were revealed by a close inspection of the ligand-binding pocket of the RccR / KDPG complex structure.In total, nine residues

KDPG binding by RccR is critical for RccR-mediated regulation
To verify the roles of these residues in KDPG binding, we mutated these residues (S139, H148, R152, H164, S183, S185, S188, K270 and R277) to alanine (A), respectively.We purified the mutant proteins and performed EMSA to assess the promoter DNA binding activity of mutant proteins.Compared with wild-type RccR, R152A, S183A, S185A, S188A, K270A and R277A mutant proteins exhibited similar binding activities to the target DNAs (Figure 5 A), whereas mutation of S139, H148 or H164 to alanine abolished the binding activity of RccR to the same target DNAs ( Supplementary Figure S8 ).Furthermore, in contrast to the wild-type RccR, the R152A, S183A, S185A, S188A, K270A and R277A mutant proteins were almost unresponsive to KDPG treatment (Figure 5 A), in-dicating that the mutation of these residues impaired the interaction between RccR and KDPG.
Next, we employed ITC to quantitatively examine the binding affinities between the mutant proteins (R152A, S183A, S185A, S188A, K270A and R277A) and KDPG.As shown in Figure 5 B and Supplementary Figure S9 , compared with wild-type RccR, the S183A, S185A and S188A mutant proteins showed significantly weakened binding affinities to KDPG with K d of 28.66, 20.97 and 15.69 μM, respectively ( Supplementary Figure S9 A-C); the R152A, K270A and R277A mutant proteins exhibited undetectable binding activities to KDPG ( Supplementary Figure S9 D-F), suggesting that these three residues are essential for KDPG recognition.
Given the vital role of RccR in regulating pyruvate catabolism and glyoxylate shunt pathway, we assessed the   growth of these mutant strains in MOPS minimal medium supplemented with different sole carbon sources.The R152A, S183A, S185A, S188A, K270A and R277A mutations were individually introduced into P. aeruginosa by using CRISPR / Cas9-based genome editing strategy ( 20 ), and then the growth curves of these mutant strains were measured in different culture conditions.Interestingly, three mutant strains (R152A, K270A and R277A) showed impaired growth in MOPS medium supplemented with glucose or glycerol, but not acetate, as the sole carbon sources (Figure 5 C), whereas all the other strains displayed similar growth irrespective of the carbon source (Figure 5 C).
Mutation of R152, K270 or R277 to alanine fully abolished KDPG binding by RccR (Figure 5 B and Supplementary Figure S9 D-F), which may permanently suppress the pyruvate catabolism pathway by constantly inhibiting the expression of aceE / F .Permanent suppression of pyruvate catabolism can result in attenuated acetyl-CoA production and downstream TCA cycle activity, and finally lead to growth impairment when glucose or glycerol is used as the carbon source (Figure 5 C and D).However, acetate can be directly converted to acetyl-CoA and then fluxed to TCA cycle by bypassing the pyruvate catabolism pathway; thus, no growth inhibition was observed in the R152A, K270A and R277A mutants, in which pyruvate catabolism is constantly suppressed (Figure 5 C and D).The role of the pyruvate catabolism pathway and aceE / F for bacterial growth when glucose or glycerol is used as the sole carbon source was further demonstrated as the PAO1 aceE -deletion strain grows slower in LB medium ( Supplementary Figure S10 A) and barely grows in MOPS medium supplemented with glucose compared with the PAO1 wild-type strain ( Supplementary Figure S10 B).Moreover, a similar growth phenotype was observed in E .coli aceE -deletion strain ( 32 ).

Discussion
Pseudomonas aeruginosa is a Gram-negative opportunistic pathogen with metabolic versatility and the capacity to cause severe infections in humans ( 1 ,5 ).The superior metabolic plasticity of P. aeruginosa can be largely attributed to the sophisticated transcriptional regulatory networks ( 5 ,6 ), enabling its success in infection and antibiotic tolerance.Aminoglycoside is a class of commonly used antibiotics for treating P. aeruginosa infections ( 2 ,3 ).Metabolic perturbation by some central metabolites, such as fumarate and glyoxylate, was found to alter P. aeruginosa aminoglycoside antibiotic tolerance by adjusting TCA cycle activity ( 17 ).However, the linkage among transcriptional regulation, metabolic status and antibiotic tolerance in P. aeruginosa remains largely unclear.
Through genetic screening with the TF mutant library ( 19 ), we identified three TFs associated with aminoglycoside antibiotic tolerance, including a RoxS / RoxR two-component system regulator PA4493 ( 33 ), a RpiR family regulator PA5438 and an uncharacterized protein PA3895 (Figure 1 A and Supplementary Figure S1 B).The RpiR family regulators commonly act as transcriptional activators or repressors of sugar catabolism, such as maltose, glucose, lactose and galactose ( 7-10 ,12 ).Additionally, the homolog of PA5438 has been demonstrated to be a regulator of central carbon metabolism in P .fluorescens ( 8 ).We thereby chose PA5438 as a candidate regulator for central carbon metabolism and antibiotic tolerance for subsequent studies.
Transcriptome analysis combined with biochemical assays demonstrates the direct regulatory roles of RccR in pyruvate metabolism and glyoxylate shunt pathway.The enhancement of the glyoxylate shunt pathway and the resulting altered TCA cycle activity is likely a mechanism for the increased aminoglycoside antibiotic tolerance in the rccR mutant.In addition to the upregulation of aceE , aceF , aceA and glcB genes, many genes involving transport, metabolism, bacterial chemotaxis, quorum sensing and biofilm formation are found to be downregulated in the rccR -deletion mutant ( Supplementary Figure S3 and Supplementary Table S4 ).Therefore, other mechanisms may also be associated with aminoglycoside antibiotic tolerance, which is further evidenced by the fact that overexpression of aceA or glcB cannot completely render the strain aminoglycoside antibiotic tolerance to the same level as that of the rccR -deletion mutant (Figure 2 E and Supplementary Figure S4 ).
Taken together, we identified RccR as a TF that is associated with aminoglycoside antibiotic tolerance and regulates pyruvate metabolism and glyoxylate shunt pathway in P .aeruginosa , revealing the linkage among carbon metabolism, transcriptional regulation and drug tolerance ( Supplementary Figure S11 ).Moreover, we determined the structure of the RccR / KDPG complex, elucidating the molecular mechanisms of KDPG sensing and RccR-mediated transcriptional regulation.This study provides insights into new antimicrobial strategy development against P .aeruginosa infections.

Figure 1 .
Figure 1.The rccR -deletion mutant strain is tolerant to tobramycin.( A ) Enrich scores of each gene are ranked and plotted.PA5438 ( rccR ) is enriched after tobram y cin treatment and mark ed in blue color.( B ) Scatterplot sho ws the insertion counts of all target sites in tobram y cin-treated and control samples.The sites belonging to PA5438 are marked in red color.( C ) Spotting assa y f or the rccR -deletion mutant.The strains are grown in LB agar plates in the presence or absence of tobram y cin.( D ) Growth curve assay for the rccR -deletion mutant.The strains are grown in LB medium in the presence or absence of tobram y cin.Data are represented as mean ± standard deviation (SD) ( n = 3).
(A.S139, B.H148, B.R152, C.H164, A.S183, A.S185, A.S188, B.K270 and B.R277), involving three different monomers (A-C) of the tetrameric protein, engage in direct hydrogen bonds with KDPG (Figure 4 D).Specifically, KDPG interacts with RccR via the sites formed by α10 helix (S139), β3 sheet (S183) and L14 loop (S185 and S188) of monomer A, α10 helix (H148), L11 loop (R152), α15 helix (K270) and L19 loop (R277) of monomer B, and α11 helix (H164) of monomer C (Figure 4 D).The side-chain hydroxyl groups of four serine residues (A.S139, A.S183, A.S185, A.S188) play a main role in recognizing the phosphate group and forming hydrogen bonds with the phosphate oxygen atoms of KDPG (Figure 4 D and E).The side-chain amines of B.H148, B.R152 and B.K270 provide hydrogen-bond interactions with the carboxyl oxygen atom at position C1 of KDPG, and the carbonyl group at position C2 also forms hydrogen bonds with the side-chain amines of B.K270 and B.R277.Moreover, the hydroxyl group at position C5 of KDPG forms hydrogen bonds with the side-chain imidazole nitrogen atom of C.H164 and the side-chain amine of B.R277 (Figure 4 D and E).

Figure 2 .
Figure 2. RccR is a regulator of p yru v ate metabolism and gly o xylate shunt pathw a y .V olcano plots for significance difference analysis of gene expression betw een PA O1 wild-t ype and rccR -deletion mut ants in the logarithmic phase ( A ) and st ationary phase ( B ). R ed circles indicate upregulated e xpression, blue circles indicate downregulated expression and gray circles indicate no significant difference.( C ) The relative mRNA transcription levels of aceE , aceF , aceA and glcB of the wild-type PAO1 and the rccR -deletion mutant in the logarithmic phase and stationary phase.The gyrB gene of PAO1 is used as the reference gene.Data are represented as mean ± SD ( n = 3).( D ) The genes in pyruvate metabolism and glyoxylate shunt pathway are regulated by RccR.( E ) Growth curve assay of the PAO1 strains overexpressing aceA or glcB gene.The strains are grown in LB medium in the presence or absence of tobram y cin.Data are represented as mean ± SD ( n = 3).

Figure 3 .
Figure 3. Assessment of the effect of KDPG on RccR.( A ) ITC assay for the binding between RccR and KDPG.K d , dissociation constant; N , number of binding sites per RccR.EMSA analysis of the interaction between RccR and its operator DNA in the absence ( B ) or presence ( C ) of 0.0625, 0.125, 0.25, 0.5, 1, 2, 4 and 8 mM KDPG.

Figure 4 .
Figure 4. Str uct ural characterizations of the RccR / KDPG comple x. ( A ) T he o v erall str uct ure of RccR / KDPG comple x in tetrameric f orm and the detailed view of the electron density of KDPG.The monomers (A-D) are marked with wheat, light blue, aquamarine and gray, respectively.KDPG is colored in green.( B ) Cartoon of RccR in its monomeric str uct ure.α-Helices, β-sheets and loops are colored in cyan, magenta and pink, respectively.( C ) Schematic representation of the topology of RccR based on the tertiary str uct ure of RccR / KDPG.( D , E ) The detailed direct h y drogen-bond interactions between RccR and KDPG.Dashed lines represent direct h y drogen bonds.Residues located in the same monomer are marked with the same color.

Figure 5 .
Figure 5. KDPG recognition by RccR is crucial for the regulation activity.( A ) EMSA analysis of the binding abilities of RccR single mutant proteins to different DNAs in the absence or presence of increasing amounts of KDPG.The gel electrophoresis images of RccR WT protein are from Figure 3 B and C. WT, wild type.( B ) ITC assays for the binding affinities between KDPG and different RccR single mutant proteins.K a , association constant; N / A, no detectable binding by ITC. ( C ) Growth curves of PAO1 WT and rccR mutant strains in MOPS minimal medium supplemented with glucose, glycerol or acetate as the sole carbon source.WT, wild type.Data are represented as mean ± SD ( n = 3).( D ) Carbon metabolism in P. aeruginosa with glucose, glycerol or acetate as the sole carbon source.