Molecular mechanism of the one-component regulator RccR on bacterial metabolism and virulence

Abstract

The regulation of carbon metabolism and virulence is critical for the rapid adaptation of pathogenic bacteria to host conditions. In Pseudomonas aeruginosa, RccR is a transcriptional regulator of genes involved in primary carbon metabolism and is associated with bacterial resistance and virulence, although the exact mechanism is unclear. Our study demonstrates that PaRccR is a direct repressor of the transcriptional regulator genes mvaU and algU. Biochemical and structural analyses reveal that PaRccR can switch its DNA recognition mode through conformational changes triggered by KDPG binding or release. Mutagenesis and functional analysis underscore the significance of allosteric communication between the SIS domain and the DBD domain. Our findings suggest that, despite its overall structural similarity to other bacterial RpiR-type regulators, RccR displays a more complex regulatory element binding mode induced by ligands and a unique regulatory mechanism.

Graphical Abstract

Graphical Abstract.

Introduction

Bacteria have evolved numerous strategies for hierarchically adapting to different environmental conditions, with one-component systems (OCSs) being a dominant and primitive mechanism (1). Unlike the well-known two-component system (TCS), which consists of a transmembrane sensor kinase and a cytoplasmic response regulator. OCS is simpler and uses only a single transcription factor (TF) for dynamic and rapid metabolic control (1). OCS TFs consist of a fused signal-sensing domain and functional response domain, but lack His-Asp phosphotransfer domains. In most bacterial growth scenarios, especially regarding efficient switching among different carbon sources, OCS TFs such as CcpE from Staphylococcus aureus (2) and XvgA from Xanthomonas campestris (3) are directly involved in the regulation of genes encoding key enzymes and transporters responsible for metabolite utilization.

HexR, an OCS TF belonging to the RpiR-type sugar catabolism regulator family was first identified in Pseudomonas aeruginosa and characterized as a specific repressor of hex regulon gene expression (4,5). HexR has a N-terminal DNA binding domain (DBD) and a C-terminal sugar isomerase domain (SIS). In the shortage of simple carbohydrates sources, HexR-DBD recognizes target palindromic sequences within the hex regulon, repressing transcription of the downstream genes. 2-keto-3-deoxy-6-phosphogluconate (KDPG), the catabolic intermediate converted from the gluconate-6-phosphate, acts as a specific negative effector of HexR by binding to its SIS. When the fermentable carbohydrates uptaking is adequate, the increase intracellular KDPG level abolishes HexR binding to the operator DNA and derepresses the expression of hex regulon genes (6). KDPG is a key and unique metabolite in the Entner-Doudoroff (ED) pathway, making HexR a critical regulator of central carbon metabolism for many bacteria species (5). Homologs of HexR from Shewanellaceae and Vibrionales also regulate diverse pathways, including glucose transport, mannitol utilization, nucleoside metabolism, glutamate biosynthesis and the glycine cleavage system (4,7).

In Pseudomonas species, besides HexR, another RpiR-like transcriptional regulator RccR has also been identified as a primary carbon metabolism regulator (4). PfRccR from Pseudomonas fluorescens shares 43% sequence identity with PfHexR and uses KDPG as a negative effector as well. Both PfRccR and PfHexR play similar essential roles in bacterial colonization and growth, but they display distinct regulatory properties and target different downstream genes. The novel features of PfRccR include autoregulation and switchable regulation of pyruvate metabolism and glyoxylate shunt/gluconeogenesis to coordinate diverse carbon limitation responses (8,9). Importantly, KDPG binding reduces the high affinity of the PfRccR for a short pseudo-palindromic site (such as within the promoter of pyruvate dehydrogenase gene aceE) but enables it to recognize another longer pseudo-palindromic site (such as within the promoter of glyoxylate shunt gene aceA), suggesting a more complex regulation mechanism. Thus, compared with HexR, RccR exhibits bifunctional regulatory characteristics, allowing it to simultaneously suppress and activate target genes involved in different carbon metabolism pathways. In particular, P. aeruginosa PaRccR (encoded by gene PA5438) variant with 275–277ΔSLR deletion has been closely associated to antibiotic resistance and virulence in CF isolates, but the structural basis for its dynamic DNA recognition mode remains unclear (10).

In this work, we focused on characterizing the molecular mechanisms of P. aeruginosa PaRccR. We confirmed that PaRccR was highly similar to PfRccR, it conservatively binds to the TGTAGT/ACTACA motif and can modify its gene targeting specificity by binding to KDPG. We also demonstrated that PaRccR directly modulates bacterial virulence and resistance by downregulating genes mvaU and algU. To understand the detailed structural mechanism of KDPG-induced regulatory element specificity switching, we determined the crystal structure of PaRccR bound to KDPG at 2.0 Å resolution. The structural data revealed the effector binding mode and molecular dynamic simulation showed that KDPG induces a conformational change in PaRccR, which in turn switches its DNA recognition specificity. Biochemical and genetic experiments further validated the structural findings and confirmed the role of PaRccR in regulating virulence and resistance in P. aeruginosa. We also revealed why ΔSLR deletion on PaRccR in the clinical isolates result in high virulence and resistance phenotypes. Taken together, our study provides new insights into the regulatory mechanism of RccR-type transcriptional regulators and sheds light on the importance of these proteins in controlling bacterial metabolism and virulence. The findings of this study have the potential to inform the development of new strategies for treating infections caused by P. aeruginosa.

Materials and methods

Gene cloning and protein purification

The gene encoding RccR from P. aeruginosa was codon optimized for expression in E. coli, synthesized, and ligated into the pET22b expression vector, which contained a C-terminal His6 tag. All mutants are linker by phosphorylation using a Blunting Kination Ligation (BLK) Kit (TaKaRa, Beijing). The expression vector pET22b-RccRHis6 was transformed into E. coli BL21 (DE3), and cells were cultured in Luria Bertani (LB) medium contained 50ug/ml ampicillin at 37°C. Protein expression was induced by 0.4 mM isopropy-β-d-thiogalactoside (IPTG) when cells were at an optical density at 600 nm (OD₆₀₀) of 0.6–0.8. The cells were further grown for 14 h at 16°C and then harvested by centrifugation at 4000 × g for 15 min. The cell pellet was resuspended in buffer A (20 mM Tris, pH 8.0, 150 mM NaCl) and lysed using an Avestin EmulsiFlex-C5 high-pressure homogenizer (Avestin, Ottawa, Canada). After that, the cell lysate was centrifuged at 18 000 × g for 30 min and then co-incubated with Ni-NTA resin (Qiagen) for 1 h. Then, imidazole was added to 35 mM and 300 mM to wash mixture protein and elute RccR protein, respectively. The protein purity was determined by SDS-PAGE analysis. The protein was further purified with size-exclusion chromatography Superose 6 10/300 GL (GE Healthcare), which was pre-equilibrated with solution buffer A. Peak fractions were determined by SDS-PAGE analysis. All the primers used in this study were detail in Supplementary Table S1.

Electrophoretic mobility shift assays

The reaction systems contained RccR protein with different concentrations varying from 0 to 8 μM and various target DNA motifs. All the double-stranded DNA substrate molecules were cloned from P. aeruginosa genomic DNA or annealed from single-stranded oligonucleotides. The DNA fragments (1 μM) were incubated with different concentrations of purified protein on ice for 30 min in the reaction buffer (25 mM Tris–HCl, pH 8.5, 150 mM NaCl, 5 mM MgCl₂, 1 mM EDTA, 0.8% Tween 20 and 5% glycerol). After incubating, the mixtures were added with 6× orange loading buffer and then subjected to 8% non-denaturating polyacrylamide gels, prepared in 1× TBE buffer (Tris/boric acid/EDTA) at 120 V for 90 min. The gels were stained by ethidium bromide and visualized using a gel imager. To demonstrated the effect of KDPG, the DNA, certain concentration of purified RccR protein and KDPG (5 μM) were incubated on ice for 30 min in the reaction buffer and then verified in 8% non-denaturating polyacrylamide gels with same method. In order to demonstrate the amount of DNA actually bound to RccR, ImageJ was employed for the quantification of free DNA across all EMSA datasets in this article by densitometry.

β-Galactosidase assay

The aceA and aceE probes were cloned ahead of a promoterless lacZ genes in pRG970km to construct lacZ fusion as previous report (11). The rccR gene was cloned into a pET-22b with isopropy-β-d-thiogalactoside (IPTG) inducible promoter. These two plasmids were co-transformed into Escherichia coli BL21 (DE3), and cells were cultured in LB media and M9 medium with various carbon source (2% w/v glucose or acetate) containing kanamycin (50 ug/ml) and ampicillin (50 ug/ml). When the OD₆₀₀ reached to 0.5 at 37°C, we added 0.2 μM IPTG to induce protein expression for 1.5 h. After that, the cells were centrifugated and β-galactosidase activities were measured as previously reported (12). All mutants were measured by the same method. The results show the percentages of β-galactosidase enzyme activity in the experimental group compared with the control group by average value of the measured data of the control group.

RNA extraction and quantitative real-time PCR

Pseudomonas aeruginosa PA14 and PA14-ΔrccR deletion strains were cultured in LB to OD₆₀₀ ∼0.6 and harvested at the exponential phase. The total RNAs were extracted using the AG21027 (Accurate Biotechnology, Hunan), and used as templates for reverse transcription to complementary DNA (cDNA) using a PrimeScript RT reagent kit (TaKaRa, Beijing). The obtained cDNA was stored in –80°C. After check for DNA concentration, 2 × ChamQ™SYBR®qPCR Master Mix (Vazyme, Nanjing) was used to perform qPCR assays. At least three wells were run for each sample. The detailed reaction system was 2 × ChamQ SYBR qPCR Master Mix (Without ROX) 10 μl, Primer 1 (10 μM) 0.4 μl, Primer 2 (10 μM) 0.4 μl, Template cDNA 1 μg and ddH₂O up to 20 μl. The detailed qPCR procedure used in the study was: stage 1 (95°C 30 s, repeat once), stage 2 (95°C 15 s, repeat for 40 times) and stage 3 (95°C 15 s, 60°C 60 s, 95°C 15 s, repeat once). The oprL gene was used as a normalizer. All the primers used in this study were detail in Supplementary Table S1.

Crystallization, data collection and structure determination

Protein samples were mixed with ten-fold molar excess of KDPG prior to crystallization trials. Crystallization of native RccR was carried out using hanging-drop vapor diffusion methods with a crystallization solution containing 0.1 M Tris (pH 8.0), 8% PEG8000. Crystals were flash frozen using 10% glycerol as a cryoprotectant. And single-wavelength diffraction data for native crystals were collected under cryogenic conditions at the beamline. Then, the X-ray intensity data were processed with XDS and HKL3000, and the native data were collected and processed to 2 Å resolution. The crystal assumed the space group I 2 2 2, with the following unit cell dimensions: 74.319, 82.855 and 138.622. We use the structure of NanR (PDB: 4IVN) as model to displace the structure of RccR. Then, we used COOT to iterative rounds of model and utilize Phenix to refine and validate the model. Finally, the native RccR structure was solved to 2 Å resolution. We found differential electron density that interacts with the α4-helix connecting the N- and C-terminal domains. Two polyethylene glycol tails fit well into this electron density and PEG 3350 was included in the crystallization buffer.

Molecular dynamic simulation

Gaussian accelerated molecular dynamic (GaMD) simulation (13) is an enhanced sampling method that smoothes the potential energy surface of system by adding a Gaussian-distributed harmonic boosting potential, which allows the reconstruction of the original free energy landscape by using cumulant expansion to the second order. Here, the whole MD process was performed in the Amber 20 program. First, the topology and coordinate files were generated using LeaP module. The protein and ligand were described by the FF19SB (14) and GAFF force field (15), respectively, and the system was immersed in an explicit TIP3P (16) water box with 15 Å from the edge of the solute, and the counterion (Na⁺) was added to neutralize the system. Subsequently, energy minimization was performed to eliminate unreasonable atomic collisions. The system was heated from 0 K to 300 K under the NVT ensemble in the case of proteins and ligands subject to harmonic constraints. The equilibrium simulation was proceeded under the NPT ensemble with the restrained force on the protein gradually reduced. Then, 5 ns conventional MD simulation was performed to calculate the acceleration parameters (V_max, V_min, V_avg and σ_v) for GaMD. After that, 15 ns GaMD equilibrium simulation was performed to adding and updating the boost potential. Finally, a 500 ns dual boost production simulation was performed. The structure was saved every 5 ps.

Construction of P. areuginosa rccR deletion and mutation strains

The P. aeruginosa ΔrccR strain was constructed according to our previous study (17,18). We cloned the upstream and downstream (500 bp) fragments of rccR from whole P. aeruginosa genome and purified them. Then, the productions were recombined to the linearized DNA of pEX 18 Gm with a ligation-free cloning system (5 × ligation-free cloning master Mix; abm). The recombinant plasmid was transformed into E. coli S17-1 and the mobilized into P. aeruginosa strain PA14. Colonies were screened using antibiotic-resistant and 15% sucrose, and the succeed strains were further verified by PCR and DNA sequencing. For mutation strains, we amplified rccR fragment from genome and recombined it with linearized DNA of pRK415 like the method of deletion plasmid. Then, the plasmid was transformed into ΔrccR strains (19). Finally, the strains were screened using Pseudomonas isolation ager (PIA) plates containing 160 ug/ml tetracycline. All the primers used in this study were detail in Supplementary Table S1.

Biofilm formation assay

Biofilm formation was determined as previous study (20). P. aeruginosa strain PA14 were cultured overnight and diluted 100-fold in fresh LB minimal and M9-medium with various carbon sources. The whole liquid system (1 ml) was transferred into each well of a 24-well polyvinyl chloride (PVC) plate (Sigma) and incubated at 37°C for 48 h. After that, the medium was removed and the wells were washed twice with a sterilized phosphate-buffered saline (PBS). Then the cells which adhered to the wells were stained with 0.1% crystal violet for 30 min and washed twice with PBS, repeatedly. Finally, the cell-bound dye was eluted in 2 ml of 95% ethanol, and the absorbance of the eluted solution was measured using a microplate reader at 570 nm. The experiments were done for three times independently with at least four replicates per experiment.

Measurement of pyocyanin production

Pyocyanin was extracted from culture supernatants and measured using a microplate reader. Briefly, 3 ml chloroform was added to 5 ml culture supernatant. Then, the chloroform layer was transferred to a fresh tube and mixed with 1 ml 0.2 N HCI. At centrifugation, the top layer was removed and its absorbance at 520 nm was measured. The pyocyanin concentrations, expressed as μg/ml/OD₆₀₀, were determined by multiplying the A520 by 17.072.

Antibiotic susceptibility test

For antibiotic susceptibility tests, ciprofloxacin was used. Wildtype PA14, the PA14-ΔrccR and all complementation strains were precultured overnight in LB. Then bacterial preculture was diluted in LB and M9-medium with various carbon sources containing 0.125 μg/ml ciprofloxacin and the starting OD₆₀₀ nm was adjusted to 0.01 for low-speed shaking culture. OD₆₀₀ nm was measured per 1.5 h until 36 h.

G. mellonella killing assay

The P. aeruginosa PA14 WT strain, ΔrccR strain and rccR mutant complementation strains were grown in M9 medium with glucose and acetate as carbon source to an optical density at 600 nm of 0.3 and 0.4. Then collected the cells and washed them with PBS three times. Last instar larvae, 2–3 cm long and 200–300 mg in weight were selected and placed into an empty plate to starve them 12 h before injection at 37°C. Each G. mellonella was injected with 10 μl of P. aeruginosa dilution (5 × 103 CFU/ml, serial dilution and plate counts) by 50 μl Hamilton syringe. The control group of G. mellonella were injected with PBS. All the G. mellonella were grown and monitored every 6 h at 37°C.

Results

RccR displays variable DNA recognition property and effective repressor activity

PaRccR shares 89% sequence identity with PfRccR. Previous studies have shown that PfRccR exhibits conditional repression effects on mutiple carbon metabolism genes, including aceEF, aceA, glcB, pntAA, pckA, gapetc. (9). These gene targets, which have upstream TGTAGT/ACTACA palindromic sites, are also conserved in P. aeruginosa. And this motif was identified recognized by RccR through DNase I footprinting (9). To further assess the recognition capability of PaRccR to the DNA sequences, we conducted Electrophoretic Mobility Shift Assays (EMSAs) using FAM-5ʹ-labeled consensus RccR binding elements from the promoter regions of aceA and aceE. These two sequences contain similar palindromic sequences but different length of the spacer between the motifs. As shown in Figure 1A, purified RccR was able to bind to P_aceA (−460 to −435 bp from the start of aceA) or P_aceE (−127 to −113 bp from the start of aceE). For P_aceA incubated with 2 μM RccR, insignificant labeled probe was shifted (Figure 1A, lane 3), and the amount of shifted probe was increased with an increasing amount of RccR (Figure 1A, lanes 3–5). However, even when incubated with 8 μM RccR, the probe was not completely shifted (Figure 1A, lane 5). By calculating the densitometry of the free DNA, it was determined that only approximately 17% of DNA is bound by the protein, which indicated a rather weak interaction between RccR and P_aceA. For P_aceE, 1 μM RccR shifted more than half of the probe (Figure 1B, lane 2), almost 54% DNA was bound by RccR. When incubated with 2 μM RccR, nearly 65% probe was shifted (Figure 1B, lane 3). Thus, the P_aceA probe seemed to have less affinity for RccR binding than the P_aceE probe.

Figure 1.

RccR is capable of recognizing and repressing targeting pseudo-palindromic sequences. (A, B) EMSA experiments were conducted to confirm the binding effect of RccR to the promoter regions of P_aceA (the probe region of aceA, which sequence is TCGTGTAGTAGTCTTGAAAAAACACTACAAAA) and P_aceE (the probe region of aceE, which sequence is ATCTGTAGTAAAACTACAAG) in the presence or absence of KDPG. As the concentration of RccR increased, a prominent migration band were observed. (A) In the presence of KDPG, RccR shows stronger binding capacity with P_aceA. (B) In the absence of KDPG, RccR exhibits stronger binding capacity with P_aceE. (C, D) The β-galactosidase assay demonstrated the repression of RccR to targeting genes when binding to the P_aceA (C) and P_aceE (D) in LB and M9-medium with various carbon sources. *P < 0.05; **P < 0.01; ***P < 0.001 and ****P < 0.0001 by two-way analysis of variance (ANOVA) statistical test for (C) and (D).

We next addressed the question of how KDPG alters the DNA binding properties of RccR. Upon adding a final concentration of 5 μM of KDPG in presence of RccR and P_aceA, the free DNA is shifted towards a complex band, showing the affinity of RccR for aceA is strongly increased in the presence of KDPG (Figure 1A, lane 7–10). At these conditions, maximum binding is achieved at 8 μM of RccR, almost 93% of DNA has migrated. In contrast, for P_aceE, the addition of KDPG alleviated the formation of complex bands and did not completely shift the DNA band even with 8 μM RccR (Figure 1B, lane 10), only 40% of DNA was bound. These results confirm that the capability to sense KDPG allows RccR to flexibly switch the selectivity towards the spacer length between motifs. To validate the significance of the palindromic sequence and the spacer sequence length for RccR binding, P_aceA and P_aceE probe variants were synthesized and subjected to EMSA experiment (Table 1). Substitutions on the palindromic TGTAGT/ACTACA motif all demonstrated very weak interactions with RccR, and insertion within the spacer region also showed largely reduced RccR binding efficiency (variants 2–7 in Supplementary Figure S1). However, substitution in the spacer region (variant 1 in Supplementary Figure S1) still retained significant RccR binding capacity. It indicates that the RccR-DNA recognition is sensitive to the altered distance between the palindrome but is tolerant to base substitutions within the spacer region.

Table 1.

The sequence of P_aceA and P_aceE and their variants

DNA Probe name	Sequence
PaceE	ATCTGTAGTAAAACTACAAG
	CTTGTAGTTTTACTACAGAT
PaceA	TCGTGTAGTAGTCTTGAAAAAACACTACAAAA
	TTTTGTAGTGTTTTTTCAAGACTACTACACGA
Variant1	ATCTGTAGTgggACTACAAG
	CTTGTAGTcccACTACAGAT
Variant2	ATCTGTAGTAAAACggCAAGT
	ACTTGccGTTTTACTACAGAT
Variant3	ATCTGggGTAAAACTACAAGT
	ACTTGTAGTTTTACccCAGAT
Variant4	ATCTGTAGTaaaaaAAAACTACAAGT
	ACTTGTAGTTTTtttttACTACAGAT
Variant5	TCGTGccGTAGTCTTGAAAAAACACTACAAAA
	TTTTGTAGTGTTTTTTCAAGACTACggCACGA
Variant6	TCGTGTAGTAGTCTTGAAAAAACACccCAAAA
	TTTTGggGTGTTTTTTCAAGACTACTACACGA
Variant7	TCGTGTAGTttttttttttttttACTACAAAA
	TTTTGTAGTaaaaaaaaaaaaaaACTACACGA

Letters in lowercase italics represent the mutant in sequence of P_aceA and P_aceE.

We constructed the P_aceA-lacZ and P_aceE-lacZ reporter plasmids and measured β-galactosidase activity in E. coli (Supplementary Figure S2). Comparing with bacteria that was co-transformed with the empty pET22b, BL21 (DE3) strain and harboring IPTG inducible pET22b-rccR vector exhibited lower β-galactosidase activities in both systems (Figure 1C and D, Supplementary Figure S3), suggesting that RccR negatively regulates the expression of aceA and aceE in vivo. In addition, the P_aceA-lacZ transcription was relatively more inhibited than the P_aceE-lacZ in LB medium, this might be due to the available intracellular KDPG under this condition. To confirm this, we used glucose and acetate as the sole carbon sources to induce two different states of intracellular KDPG synthesis in bacteria, one with high levels and the other with low levels as described previously (5,9). As expected, RccR tightly repressed the P_aceA-lacZ expression in glucose as well as in LB media (Figure 1C), while instead, the expression of P_aceE-lacZ in acetate was remarkably inhibited by RccR than that in glucose-containing media (Figure 1D). This observation is consistent with the EMSA results and revalidated that KDPG-regulated DNA recognition of RccR is an effective mechanism to respond to different carbon sources.

Different impacts of rccR knockout and ΔSLR mutant on P. aeruginosa virulence and resistance

A Galleria mellonella larvae infection model was employed to evaluate the contributions of RccR in P. aeruginosa infection. PA14 wild type (WT) and a rccR knockout (ΔrccR) strains were grown to log phase in glucose and acetate carbon source media, respectively (Figure 2A). An inoculum of ∼10⁵ cfu was sufficient to kill the majority of infected larvae, with melanisation evident at 12 h post infection. And mortality was recorded within 36 h of injection. As Figure 2A shown, the WT infection groups were completely killed after 16 h, while the ΔrccR infection groups all exhibited significant lower morality, with the strains grown in glucose displayed even more attenuated virulence compared to that in acetate medium. Meanwhile, we also measured the susceptibilities of WT and ΔrccR strains to ciprofloxacin under different culture medium (Figure 2B). In LB medium, bacteria exposed to ciprofloxacin of 0.125 μg/ml were still cultured to an OD₆₀₀ above 1.2 after 14 h, and the ΔrccR strain showed slightly but significant reduction of survival. Almost the same level of reduction in ciprofloxacin resistance was observed in M9-glucose cultures. However, when the sole carbon source was replaced with glycerol or acetate, the survival rate of ΔrccR strain were dramatically reduced to 50–70% of the WT.

Figure 2.

RccR contributes to P. aeruginosa virulence phenotype and drug resistance. (A) The survival of G. mellonella larvae was analyzed following the injection with P. aeruginosa cells (WT, ΔrccR, ΔrccR+ pRK415-rccR complementation strain and ΔrccR+ pRK415-rccR- ΔSLR mutant complementation strain) cultured in M9-medium with various carbon sources. Each G. mellonella was injected with 10 μl of P. aeruginosa dilution (5 × 10³ CFU/ml). And the PBS-injected larvae were the negative control. The larvae were monitored for 36 h after the infection. (Mantel−Cox test for statistics, *P < 0.05; **P < 0.01; ***P< 0.001 and ****P < 0.0001). (B) Resistance of WT and ΔrccR strains to ciprofloxacin was measured. These two strains were inoculated in different carbon source culture media containing 0.125 μg/ml ciprofloxacin. And the optical density was measured after 14 h. (C) Pyocyanin production of the WT and ΔrccR with different carbon source. The experiments were repeated three times and error bars were shown. (D) Biofilm formation of the WT and ΔrccR strain with different carbon source. Quantification of biofilm biomass via crystal-violet staining and A570 was measured using a microplate reader. Data are shown as the change relative to WT and represent three independent experiments. (E) Resistance of ΔrccR+ pRK415-rccR complementation strain and ΔrccR+ pRK415-rccR- ΔSLR complementation strain to ciprofloxacin with kinds of carbon sources. *P < 0.05; **P < 0.01; ***P< 0.001 and ****P < 0.0001 by two-way analysis of variance (ANOVA) statistical test for (B)–(E).

Since pyocyanin and biofilm formation are of the important mechanisms used by P. aeruginosa to establish infection and resistance (21), we further compared the pyocyanin and biofilm level between WT and ΔrccR strains. Although there was no obvious difference in pyocyanin synthesis under the LB medium, once fixed to a single carbon source, ΔrccR showed significantly reduced levels (Figure 2C). As for biofilm phenotype, the effect of the mutant strain was more intuitive, showing noticeable reduced biofilm under all culture conditions (Figure 2D). Collectively, these results indicated that the effects of RccR on virulence and antibacterial drug resistance varied significantly with the use of carbon sources, and the obvious differences in these phenotypes under glucose and acetate also implied that KDPG-regulated DNA recognition mechanism also plays a key role.

Furthermore, A C-termial ‘SLR’ deletion in RccR has been reported to be highly correlated with CF P. aeruginosa isolates (10), suggesting that this is likely a gain-of-function mutation that promotes bacterial persistent infection. To verify this mechanism, plasmids pRK415-rccR and pRK415-rccR-ΔSLR were separately reintroduced into the ΔrccR strain, and their resistance and virulence phenotypes were compared. As shown in Figure 2A and B, the virulence against Galleria mellonella larvae can be restored in both complemented strains. However, under the in vitro culture conditions, we can observe that the strain complemented with pRK415-rccR-ΔSLR exhibits higher levels of ciprofloxacin resistance (Figure 2E), pyocyanin (Supplementary Figure S4A) and biofilm production (Supplementary Figure S4B) than that complemented with the wild type. The enhancement of ciprofloxacin resistance by pRK415-rccR-ΔSLR also depends on the type of carbon source, suggesting that its gain of function effect may be related to the regulatory mechanism of KDPG.

RccR directly regulates virulence-associated genes mvaU and algU

The abovementioned data inspired us to further investigate how RccR regulates the expression of virulence and stress-resistance related genes. We used motifs GTAGn(16)Cn(2)CA and GTAGTn(3)ACTAC to search for possible targeting motifs in the intergenic regions of genes in P. aeruginosa genome. Among 100 possible targets obtained (Supplementary Table S2), we noticed two locations that happened to be non-coding regulatory regions flanking the genes mvaU (PA2667) and algU (PA0762), respectively. MvaU is a H-NS type regulators, as well as its paralog MvaT, acts as a global repressor of motility, resistance, and virulence genes (22,23). AlgU is a stress-related sigma factor that activates transcription of alginate biosynthesis and is required for the full resistance of P. aeruginosa (24). We then carried out EMSA experiment to verify whether RccR recognizes the sought motif of mvaU and algU (Figure 3A). DNA migration still occurred, reaching a maximum of 32%-70% in the presence of KDPG at 8 μM of RccR. The pseudo-palindromic sequences found in those regulatory elements are varied in the spacer region (Supplementary Figure S5), but their spacer region length of 16 bp is consistent with P_aceA. Therefore, the presence of KDPG can enhance the binding affinity of both P_mvaU and P_algU to RccR. However, variations in the second palindromic site significantly reduced the RccR binding affinity of the regulating element, and approximately 10-fold higher RccR concentration was required for P_mvaU and P_algU to form protein-DNA complexes compared to P_aceA and P_aceE.

Figure 3.

RccR could regulate virulence-associated gene mavU and algU. (A) EMSA experiments to validate the binding effect of RccR to P_mavU and P_algU in the presence or absence of KDPG. (B) Relative mRNA levels of mavU and algU in ΔrccR compared with WT cultured in different carbon source M9-medium. (C) The β-galactosidase assay demonstrated the repression of RccR to targeting genes when binding to the P_mavU and P_algU in LB and M9-medium with various carbon sources. *P < 0.05; **P < 0.01; ***P < 0.001 and ****P < 0.0001 by two-way analysis of variance (ANOVA) statistical test for (B) and (C).

The subsequent q-PCR data indicate that, although the interaction in vitro is much weaker, the in vivo inhibitory effects of RccR on the expression of mvaU and algU are obvious (Figure 3B). When cultured in M9-glucose medium, the mRNA level of mvaU and algU in the ΔrccR strain were approximately 1.3–1.9-fold higher than in the WT strain. In contrast, there is no obvious difference between the knockout and WT strain in M9-acetate culture. This observation further supports the notion that the regulation of mvaU and algU by RccR primarily occurs under metabolic conditions with sufficient KDPG synthesis. Similar carbon source-dependent and KDPG enhanced regulation effects were also detected in β-galactosidase reporter analysis (Figure 3C). Under the same conditions of M9-glucose medium, comparing the inhibitory effect of RccR on P_aceA (β-galactosidase relative activity decreased to about 25%, Figure 1C) and P_mvaU/P_algU (β-galactosidase relative activity also decreased by half, Figure 3C), it is evident that, despite the nearly tenfold weaker binding ability of RccR to P_mvaU/P_algU in vitro, it is sufficient to significantly suppress their expression in vivo. These results not only demonstrate that RccR can modulate the expression of mvaU and algU to control bacterial infection and drug resistance but are also consistent with the concept that RccR acts as a molecular switch by sensing KDPG and binding to distinct regulatory elements.

Crystal structure of RccR reveals a tetramer with symmetrically arranged six-helical bundle HTH domain

The structure of RccR-KDPG was determined using molecular replacement, with the C-terminal SIS-domain of Vibrio vulnificus NanR (PDB ID: 4ivn) as a search model. The final model was refined to a resolution of 2.0 Å. Most residues were successfully modeled, except for residues 281–293 and the structure was refined to an R factor of 18.20% and R free of 20.65% (Supplementary Table S3). In the RccR-KDPG complex, one molecule was found in the asymmetric unit. However, gel filtration experiments indicated that RccR forms a tetramer in solution (Supplementary Figure S6). Therefore, a dimer-of-dimers tetramer was generated by crystal symmetry (Supplementary Figure S7), where two monomers contribute α8, α10 and α14 to form a tightly associated dimer (dimer interface area: 3221.5Ų). Pairs of dimers further oligomerize to form a homotetramer (total tetramer interface area: 2711 Ų). Dimer formation involves extensive charge-charge interactions and hydrogen bonding, while hydrophobic interactions primarily drive the tetrameric assembly. Additionally, several direct or water molecule-mediated hydrogen bonds contribute to its stability (Supplementary Table S4). This quaternary arrangement bears similarities to the biologically relevant assembly observed in the RpiR/AlsR family regulators MurR (Supplementary Figure S7) (25), indicating that the tetramer represents the functional unit required for the full transcriptional activity of RccR.

Consistent with the known structural information of RpiR/AlsR type proteins (25), the DBD (residue 1–74) of RccR exhibits a multi-helical version of the HTH motif wherein α3, α4 and α5 form the tri-helical HTH core, which is packed by three additional helices (Figure 4A and 4B). This type of HTH fold is relatively rare in bacteria and is suggested to be a derivative of the tetrahelical HTH superclass (26). The C-terminal SIS domain (27) (residue 91–280) is characterized by a five-stranded parallel β-sheet flanked by eight α-helices. It typically functions as a catalytic domain in keto–aldol sugar isomerases or as a ligand binding domain (LBD) in phosphosugar biosynthesis pathway regulators. The linker region (residue 74–90) connecting the HTH and SIS has not been resolved in the previously reported structures of RpiR-like regulators (26,27). In our RccR structure, the linker region clearly forms a short helix-coil hairpin, which is securely held in the interdomain cleft and stabilized by hydrophobic contacts and water-mediated hydrogen bonds (Supplementary Table S5).

Figure 4.

Crystal structure of RccR-KDPG and critical residues for DNA binding. (A) The P. aeruginosa RccR monomer has two domains—an N-terminal DNA-binding domain (orange) and C-terminal effector binding domain (green). The DNA binding domain contains a helix-turn-helix domain wherein α3, 4 and 5 form the core and packed by three additional helices. The C-terminal domain is a SIS domain characterized by five-stranded pa rallel β-sheets and eight α-helices. And these two domains were connected by α7. (B) The topological structure diagram of RccR monomer. (C) Superposing the C-terminal domains of NanR (gray) and RccR. The angle of NanR between two domain is 155.9° while the RccR is 176.4°, which RccR shows an extened conformation. And the detailed interaction of N-terminal neutral residues N2 and Q5-mediated DBD-linker associations. The side chain of N2 form a direct hydrogen bond with the hydroxyl group of I85 and S73 backbone, and the side-chain amine group of Q5 forms 2 direct hydrogen bonds simultaneously with the backbone amide of A78 and main-chain oxygen atom of F83. (D) EMSA experiments to validate the binding effect of RccR variants to P_aceA and P_aceE in the presence or of KDPG. (E and F) The β-galactosidase assay demonstrated the repression of RccR variants when bind to the P_aceA (E) and P_aceE (F) in the presence or absence of KDPG. *P < 0.05; **P < 0.01; ***P< 0.001 and ****P < 0.0001 by one-way ANOVA statistical test for (E) and (F).

Although the overall topology of the RccR monomer closely resembles that of NanR, there is a notable difference in the relative orientation between the DBD and SIS domains of these two homologs. The inter-domain rotation angles are approximately 155.9° in NanR and 176.4° in RccR. Upon superposition of the SIS domains between the two regulators, it was observed that the DBD of RccR adopts a more extended conformation and establishes several hydrogen bonds with the linker region (Figure 4C). In detail, the side chain of N2 forms a direct hydrogen bond with the hydroxyl group of I85 and the S73 backbone. Additionally, the side-chain amine group of Q5 simultaneously forms two hydrogen bonds, one with the backbone amide of A78 and the other with the main-chain oxygen atom of F83 (Figure 4C). It is worth noting that the corresponding linker residues in the NanR structure are disordered, which may contribute to its significantly different inter-domain orientation compared to that of RccR.

Although the high-resolution DNA-bound structure of RpiR/AlsR regulator is currently unavailable, we can still identify key residues involved in DNA recognition based on the conservation of the HTH motif. Positively charged residues, including R15, K16, K20, R53 and R56, which are centrally distributed on ɑ2 and ɑ5, are believed to constitute the basic patch for DNA recognition (28) (Supplementary Figure S8A). Among these residues, R53 and R56 are the most conserved, and their corresponding positions in NanR have been shown to be critical for DNA binding (Supplementary Figure S8B) (29). To investigate the importance of these residues in RccR, we generated and purified RccR mutants R53A and R56A for EMSA assay. As expected, both mutants exhibited significantly reduced DNA migration compared to the wild type (Figure 4D). The maximum amount of DNA binding is only 35%. Subsequent β-galactosidase reporter assays provided further insights. The repression effects of RccR on P_aceA (Figure 4E) and P_aceE (Figure 4F) were almost completely abolished by these single mutations, confirming the crucial role of R53 and R56 in regulatory function.

Structural basis of the interaction between KDPG and RccR

The electron density observed for the ligand KDPG in our RccR structure was well defined (Figure 5A). Similar to other regulators containing SIS domains (27), the ligand resides within a cavity formed by the α13–α14 loop, β1–α10 loop and β3–α12 loop. In the tetramerized RccR, the α10 and C-terminal α15 regions from the opposite dimer subunit, along with α11 from the adjacent tetramer subunit, further enclose this region to create a tightly binding pocket. These structural elements participate in stabilizing KDPG by establishing extensive hydrogen bonds and van der Waals contacts (Figure 5A and B). The central cavity predominantly interacts with the KDPG phosphate group, which directly interacts with the main-chain amide of Q184, S185 and side-chain hydroxyl groups of four serine residues (S139, S183, S185 and S188). Furthermore, the main-chain amide of F136, I182, and the side chain of T231 form additional water-mediated hydrogen bonds with phosphate group. Notably, these residues, particularly the four conserved neutral polar residues, serve as structural constraints to precisely determine the position and orientation of KDPG binding (Supplementary Figure S8B). Conversely, residues involved in sugar unit binding are primarily from neighboring subunits, wherein H148, R152 and K270 form hydrogen bonds with the hydroxyl groups at position C1 of KDPG. A138 interact with C4 hydroxyl. And two other basic residues, H164 and R277, interact with the C5 hydroxyl. It is noteworthy that R277 from α15 is also involved in the interaction with C2 hydroxyl and phosphate moieties. This arginine residue marks the end of two consecutive SLR motifs at the C-terminus of RccR. Hence, the clinical isolate with ΔSLR mutation is likely to lose recognition of KDPG at this critical site, resulting in dysregulation of downstream genes and high drug resistance.

Figure 5.

Structural analysis of the interaction between RccR and KDPG. (A) The overall structure and detailed structural analysis of RccR-KDPG complex and the detail view of the electron density of KDPG simulated annealing omit map (1.0) is shown as pink mesh. Chain A (orange), Chain B (yellow), Chain C (gray), Chain D (indigo). (B) The detailed interactions with RccR-KDPG. Hydrogen bonds, black; water mediated hydrogen bonds, blue. (C) EMSA experiments to validate the key residues impact binding ability of RccR to P_aceE and P_aceA in the presence of KDPG. (D) The β-galactosidase assay demonstrated the repression of RccR variants to P_aceA and P_aceE in the presence KDPG. *P < 0.05; **P< 0.01; ***P < 0.001 and ****P < 0.0001 by one-way ANOVA statistical test for (D).

To validate this inference and examine the impact of impaired KDPG binding on the biological activity of RccR, we generated site-directed mutants H148A, R152A, H164A, T231A, K270A and the quadruple alanine substitutions on conserved serine residues S139, S183, S185 and S188 (SSSS). We purified the RccR mutants, including the ΔSLR version, and subjected them to EMSA assays in the presence of KDPG. All mutations displayed noticeable migration with P_aceE, up to 70% of DNA, consistent with the aforementioned structural analysis, indicating that the mutations greatly reduced or completely abolished the negative regulatory effect of KDPG on RccR’s binding to P_aceE (Figure 5C). This mechanism was further confirmed in the β-galactosidase reporter assay, where all mutations effectively suppressed the transcriptional activity of P_aceE when glucose was supplied in minimal media (Figure 5D). Additionally, the positive regulatory effect of KDPG on RccR-P_aceA recognition also showed varying degrees of reductions ranging from 37% to 100% in these mutants, reflecting the strong or weak contribution of the corresponding site to KDPG recognition. For example, mutations at H148 and H164, which form weak hydrogen bonds with KDPG, and T231, which requires water-mediated interaction, had relatively less impact on P_aceA in the EMSA assay. However, the effect of these mutations on P_aceA transcriptional activity was not as pronounced as that on P_aceE, all mutations except SSSS retained a centain level of repression on P_aceA. Collectively, these results illustrate the crucial role of the intact interaction between RccR and KDPG in the regulatory activity of RccR.

Exploring the conformational rearrangements of RccR through molecular dynamics simulation

It is well-established that conformational flexibility is a prevalent characteristic commonly observed in single-component regulatory systems, where the linkage between the DBD and LBD is responsible for conformational changes induced by effectors. To gain further insights into the dynamic changes and the inter-domain allosteric communications in RccR, we conducted 500 ns gaussian-accelerated molecular dynamics (GaMD) simulations of both the Apo and KDPG-Bound RccR tetramers. During the simulations, we tracked the root-mean-square deviation (RMSD) values of the carbon atoms in the DBD and signal input sensor domain (SIS) relative to the initial structure throughout the entire molecular dynamic trajectory. The RMSD profiles revealed significant conformational changes in the DBD, while the SIS exhibited relatively higher stability (Figure 6A). The analysis of root-mean-square fluctuations (RMSF) demonstrated that large fluctuations occurred in the DBD domain and the C-terminal α15 region (residues 260–280), suggesting their high flexibility (Figure 6B). Moreover, the comparison of RMSF values between the Apo and KDPG-Bound systems revealed showed that KDPG binding enhanced the flexibility of DBD while reduced the flexibility of α15.

Figure 6.

KDPG-mediated confirmational changes of DNA to recognize variable DNA. 500 ns gaussian-accelerated molecular dynamics (GaMD) of APO and KDPG-Bound tetramer were performed respectively. (A) In the Apo and KDPG-Bound systems, the changes in RMSD of DBD domain (residue 1–74) and SIS domain (residue 91–280) over simulation time. These results of MD show the DBS domain exhibited a tendency for significant conformational changes and SIS domain was more stable. (B) The RMSF analysis of DBD domain and α15 (residue 260–280) indicated them perform substantial oscillations or swinging motions during the simulation. (C) Two-dimensional descriptors were generated to analysis the motion of RccR. The angle formed by the centroid of the DBD domain, the Cα of residue V91, and Cα of residue H112 was set as the X axis. And the dihedral formed by the centroid of DBD domain, the centroid of residues P’260-R’274, Cα of residue V91, and Cα of residue H112 was set as the Y-axis. Furthermore, a free-energy was exhibited to assess these conformations. Representative structures were obtained for four Apo systems and three KDPG-Bound systems. Crystal structure (red pot). (D) Superposing the stable domains (91–259) of crystal with four Apo systems and three KDPG-Bound systems. Crystal structure (orange), Apo structure (blue scheme), KDPG-Bound (green scheme). The balls with corresponding colors represent their centroid. (E) Two-dimensional distribution of distances. The distance between R274 (CZ) and D’23 (CG) is set as X axis. And the distance between R271 (CZ) and E’’’200 (CD) is set as Y axis. Apo conformation (orange), KDPG-Bound conformation (green), Crystal structure (red). In Apo system, R271 tens to interact with E’’’200. And in KDPG-Bound system, R274 tends to interact with D’23. (F) EMSA experiments to validate these positions impact binding ability of RccR to P_aceE and P_aceA in the presence of KDPG. (G) The β-galactosidase assay demonstrated the repression of RccR variants to P_aceA and P_aceE in the presence KDPG. *P < 0.05; **P < 0.01; ***P < 0.001 and ****P< 0.0001 by one-way ANOVA statistical test for (G).

The free-energy landscape along the two variables was reweighted to analyze the motions of the DBD relative to the SIS (Figure 6C). The X-axis was defined as the angle formed by the centroid of the DBD domain, the Cα atom of residue V91, and the Cα atom of residue H112. The Y-axis represented the dihedral angle formed by the centroid of the DBD domain, the centroid of residues P’260-R’274 (residues with upper indices are from neighboring subunit), the Cα atom of residue V91, and the Cα atom of residue H112. The analysis revealed several intermediate conformations corresponding to the structures with the lowest energy. The results indicate that upon KDPG binding, RccR tends to adopt or even stabilize in more extended conformations (referred to as KDPG-Bound state II), with the DBD moving further away from the SIS. Conversely, when KDPG is removed, a substantial distribution of compact conformations is observed, with a subset of intermediate conformations driving the DBD closer to the SIS. Therefore, it is suggested that conformational transformation occurs from compact states to intermediate or extended states upon KDPG binding.

We extracted several conformations from the basin of energy landscape to represent the metastable states of Apo and KDPG-Bound systems. These conformations were then superimposed with the original RccR tetramer by aligning the tetramerization core structures (residues 91–259, which were found to be relatively stable in the RMSF analysis) (Figure 6D). Subsequent analysis focused on the inter-domain rearrangements of the mobile DBD and α15 domains. Initially, we observed that the interdomain angle varied from 140.19–178.69 degrees, with asymmetrically arranged DBD in different subunits. To facilitate DBD shifting and contact with the neighboring C-terminal α15 (Supplementary Figure S9A), the associations mediated by N-terminal neutral residues N2 and Q5 in the DBD-linker region needed to be disrupted. Furthermore, we identified new charge-charge interactions in the metastable states, including R11-D230/D232, D23-R’274/R’277, D28-K’267 (Supplementary Figure S9B). Additionally, in different states, the basic residues R271 and R274 along α15 need to form salt bridges with a critical acidic residue E200 from the neighboring subunit at the tetramer interface to stabilize the local conformation (Supplementary Figure S9C). Compared to the Apo RccR, the KDPG-Bound RccR exhibited a relatively narrow distribution of DBD-α15-associated conformations. In the Apo states, separate portions of both the R271-E’’’200 and R274-D’23 bonds formed conformations representing more compacted DBD orientations (Figure 6E). These local conformational differences were attributed to the direct interaction of K270 and R277 from the neighboring subunits' α15 with the ligands (Figure 5B), which also increased the stability of the D232-R187 interaction and enhanced the probability of D232 interacting with R277 (Supplementary Figure S10). Thus, the C-terminus of α15 serves as a key gating mechanism that senses KDPG binding and drives conformational changes.

Based on the above analysis and results obtained from MD, we generated the following mutants: N2A-Q5A, D23A, N27A, D28A, E200A, D230A and K267A (Supplementary Figure S11). Subsequently, we performed EMSA and β-galactosidase reporter analyses on each mutant. Compared to the wild-type, nearly all mutations significantly impaired the regulatory effect of KDPG on RccR, either by decreasing the binding to P_aceE or enhancing the binding to P_aceA (Figure 6F and Supplementary Figure S12). While the E200A mutation also weakened the down-regulation effect of KDPG on P_aceE binding, it displayed a more pronounced enhancement effect on P_aceA recognition compared to the wild-type. Consistent with the EMSA results, in cultures with sufficient KDPG synthesis, all mutations exhibited a stronger negative regulatory effect on the aceE promoter in vivo compared to the wild-type. Conversely, the negative regulatory effects on the aceA promoter were largely diminished for all mutations (except for the E200A mutation, which exhibited a stronger negative regulatory effect than the wild-type) (Figure 6G). Collectively, these mutations influence the regulation of KDPG, while maintaining a certain level of target DNA recognition ability. This suggests that these specific positions play a crucial, yet not irreplaceable, role in the regulation of KDPG-mediated conformational changes in DNA specificity recognition by RccR.

Functional significance of the KDPG-dependent switching mechanism of RccR

Based on the structural analysis, we have identified key sites in RccR that are involved in DNA recognition, KDPG binding, and allosteric regulation. These sites are distributed across different regions and provide the structural basis for effector molecules to regulate the specificity of RccR-DNA recognition. To further investigate their functional implications, we introduced three sets of mutations into the ΔrccR strain using a complement plasmid and compared their phenotypes with the wild-type complemented strain. Similar with the clinically relevant mutant ΔSLR (Figure 2E), most mutations exhibited a significant enhancement in pyocyanin synthesis (2.3- to 4-fold increase), with the degree of enhancement being positively correlated with their effects on KDPG or DNA recognition (T231A served as a negative control) (Figure 7A). These mutations also promoted biofilm synthesis, albeit to a lesser extent (Figure 7B). Furthermore, drug resistance experiments conducted under different media conditions revealed that each mutation significantly enhanced the drug resistance of the strains (1- to 7-fold increase) when KDPG was present in sufficient amounts (Figure 7C). As expected, the E200A mutation did not exhibit a different phenotype from the other mutations, despite its enhanced ability to repress the aceA promoter when mediated by KDPG. This suggests that the regulatory effects of RccR on bacterial virulence and resistance are coordinated with its recognition of both short- and long-pseudo-palindromic sites.

Figure 7.

Effect of RccR key residues on virulence phenotype and drug resistance. (A) Pyocyanin production of complementation transformed with plasmid pRK415-rccR variants strain compared with that of complementation transformed with plasmid pRK415-rccR in PB medium. (B) Biofilm formation of complementation transformed with plasmid pRK415-rccR variants strain compared with that of complementation transformed with plasmid pRK415-rccR. (C) Resistance complementation transformed with plasmid pRK415-rccR and variants strains to ciprofloxacin. These strains were inoculated in different carbon source culture media containing 0.125 μg/ml ciprofloxacin. And the optical density was measured after 24 h. All the experiments were repeated three times and error bars were shown. *P < 0.05; **P < 0.01; ***P < 0.001 and ****P < 0.0001 by one-way ANOVA statistical test for (A)–(C).

In summary, the results from mutagenesis and functional studies confirm that key residues responsible for the biochemical activity of RccR are also crucial for regulating bacterial virulence and resistance phenotypes. This provides significant evidence supporting the dual function of RccR in modulating energy metabolism in response to diverse carbon sources.

Discussion

RpiR family regulators: versatile signal transduction systems in bacteria

OCSs are signal transduction systems comprised of a single protein that contains input and output domains. They are evolutionarily older and more widely distributed among bacteria compared to TCSs. OCSs have the ability to sense and respond to various environmental stimuli, including light, oxygen, temperature, pH, and others (30). Among OCSs, the RpiR family regulator is noteworthy for its unique effector-binding domain, enabling it to detect and respond to sugar derivatives (31). RpiR family regulators can function as either gene expression activators or repressors, depending on the specific ligand or environmental cue they sense.

RpiR homologs have been identified in various bacterial species and are commonly associated with the regulation of carbohydrate metabolic genes. In Escherichia Coli (31), Lactococcus lactis, Sinorhizobium meliloti and Neisseria meningitidis (32), RpiRs such as EcRpiR, GlaR (YugA), IolR, HexR regulate the expression of genes involved in the metabolism of ribose, galactose, inositol, glucose (33,34). Additionally, there are RpiR family regulators involved in the utilization and synthesis of special sugar derivatives or intermediate metabolites in certain bacteria. In Clostridium perfringens (35) and Haemophilus influenzae, RpiR-like regulators NanR and SiaR are associated with sialic acid metabolism and are necessary for maintaining a balance between sialyation and catabolism to support their survive within the host (36). MurR (25) (from E. coli) and OngR (37) (from marine bacteria Pseudoalteromonas strains) negatively control the expression of enzymes responsible for the catabolism of bacterial or fungal cell wall-derived MurNAc and chitooligosaccharides. HpxU (38) from Klebsiella pneumoniae has been shown to act as a transcriptional repressor of allantoate catabolism. On the other hand, RpiRs have been found to coordinate with other pathways to regulate environmental signal sensing and even bacterial virulence. A RpiR-like regulator protein from Bacillus cereus is co-transcribed with casK and casR and plays a role in the optimal unsaturation of fatty acids (FAs) necessary for bacterial cold adaptation (30). Salmonella Typhimurium flagellar synthesis regulator AsiR is critical for acidic pH signal-mediated regulation and Brucella melitensis RpiR1-3 are putatively involved in the metabolism of rhizopinea, a putative signal molecule (39,40). RpiRA-C from Staphylococcus aureus and Salmonella Typhimurium IdoR have been shown to alter the expression of pathogenicity island genes (41,42). Therefore, RpiR family regulators exhibit versatile biological functions beyond central metabolism.

In Pseudomonas species, three members of the RpiR regulators have been identified. Both HexR and RccR are necessary for remodelling central metabolism in response to intracellular KDPG levels (4,9). QapR represses gene expressions that modulates Pseudomonas quinolone signal (43). Notably, RccR not only exhibits a unique regulatory element specificity switching mechanism but also has been found that its gain-of-function mutations are closely associated with drug resistance in clinical strains (10). However, our understanding of the mechanisms and specificities of RpiR family proteins remains limited due to the lack of molecular mechanism studies, especially the structural data. Previous studies have demonstrated that the recognition switch of RccR between short spacer pseudo-palindromic sequence (P_aceE) and long spacer pseudo-palindromic sequence (P_aceA) is achieved through the release and binding of KDPG (9). In this study, EMSA experiments of DNA motif variants revealed that RccR exhibits some tolerance towards the second half of the palindromic sequence when recognizing long spacer regulatory elements (Supplementary Figure S1). This mechanism explains why there are fewer potent RccR target genes with short spacer regulatory elements compared to those with long spacer regulatory elements (Supplementary Table S2). Additionally, RccR has an on/off effect on the regulation of the short spacer regulatory element, which is much stronger than the ‘turn-down’ effect on the long spacer regulatory element (including P_mvaU and P_algU identified in this study) (Figure 1C and D, Figure 3B). Therefore, in the absence of KDPG, RccR rapidly transitions the metabolism by effectively shutting down several key genes in bacteria. When the intracellular KDPG concentration increases, RccR switches to recognize numerous target genes, exhibiting relatively lower affinity for the long spacer regulatory element, thus acting as a finely-tuned repressor. Through this unique mechanism, RccR plays more complex and flexible roles as a global regulator. This also suggests that RpiR family members possess additional more unknown functions and properties that require further research for discover and explain.

Understanding the mechanism of RccR in bacterial survival and persistence within CF patients

The crucial roles of carbohydrate metabolic genes in bacteria have been widely recognized, particularly their contribution to bacterial resistance and pathogenicity (44). Numerous regulators play a pivotal role in connecting virulence, stress resistance and complex carbohydrate utilization. LysR-type carbohydrate-related transcriptional regulators, such as Xanthomonas axonopodis Pv lcrX, L. monocytogenes lysR and Staphylococcus aureus RpvA have been demonstrated to be involved in persister formation and virulence (45). Classical carbon catabolite repression (CCR) regulators from various pathogens have also been found to directly impact the expression of up to 19% of the genome, including toxin genes and their regulators (46–48). An Rgg-like regulator (encoded by 05ZYH33) associated with non-glucose carbohydrate metabolism also contributes to the virulence of Streptococcus suis (49). A pleiotropic regulator in Bacillus cereus exerted influence on carbohydrate metabolism and pathogenesis affecting the survival of this foodborne pathogen in diverse environments (50). Regarding RpiR family regulators, the presence of SIS domains renders RpiRs intrinsically sensitive to the cellular concentration of sugars and derivatives. Increasing reports on the regulation of bacterial infection by RpiR-like regulators suggest that this type of regulator is a key mechanism directly linking carbohydrate utilization and virulence (51–53). However, the specific pathway regulated by RpiR and its effect on the expression of genes related to bacterial infection remain elusive.

The widely-distributed ED pathway, characterized by the cleavage of 2-keto-3-deoxy-6-phosphogluconate into pyruvate and D-glyceraldehyde-3-phosphate, serves as the primary or exclusive route for glucose catabolism in many bacteria (54). The studies on the RipR regulator RccR provide further insight into the general mechanism of the ED pathway as the essential carbohydrate pathway regulating virulence genes and drug resistance in bacteria (9,10). We discovered and confirmed that the downstream genes mvaU and algU, which are directly regulated by RccR, are both involved in P. aeruginosa infection and drug resistance (Figure 3). MvaU belongs to the the H-NS family of proteins and acts as a transcriptional repressor, particularly of Pf4 genes that suppress mammalian immunity, and the small RNA rsmZ, which is crucial for the switch to the virulent phenotype (55). It also interacts with MvaT to coordinate gene expression and control various physiological functions in P. aeruginosa, including QS-mediated virulence factor activation, pyocyanin synthesis and resistance to copper ions (23,56). AlgU is an evolutionarily conserved and extensively studied regulator of bacterial-host interactions, with potential implications for stress tolerance management and promoting virulence in various bacteria, including Pseudomonas species (57). It regulates the synthesis and export of alginate, a virulence factor that contributes to the mucoid phenotype of P. aeruginosa strains during cystic fibrosis lung infections (58). AlgU also controls the Hrp type III secretion system, virulence effectors such as phytotoxins, and transcription regulators, playing significant roles in plant and animal pathogenesis (24). We speculate that the direct regulation of virulence- and resistance-associated genes may be a common mechanism for many RipR regulators in affecting bacterial pathogenesis.

Importantly, structural analysis revealed that R277 of RccR establishes extensive interactions with KDPG during its binding. EMSA assays and in vivo experiments further demonstrated that the ΔSLR deletion reduced the KDPG-induced suppression of long spacer regulatory elements while maintaining negative regulation on short spacer regulatory elements. This mutation also enhanced biofilm, pyocyanin synthesis and drug resistance (Figure 2 and Supplementary Figure S4). Additionally, similar phenotypes were observed for mutations introduced based on structural analysis that disrupted KDPG binding and signaling (Figures 6G and 7). This mechanism enabled RccR-ΔSLR to retain the tight regulation of gluconeogenesis while reducing repression of downstream genes, including mvaU and algU. Considering that mvaU itself acts as a repressor of multiple virulence factors expression, and inactivation of algU has been shown to increase systemic virulence during acute P. aeruginosa infections, this explains how P. aeruginosa can evolve RccR-ΔSLR to survive and persist within the CF patients.

Implications of the structural analysis of RccR

Over 20 OCS families have been identified in prokaryotic species. Many OCS TFs, including proteins from the LuxR, AraC, TetR, GntR and LysR families, have been extensively studied and practically recruited in synthetic biology applications (59–61). The RpiR-like TFs belong to a small yet significant family of OCS regulators, but only a few members have structures available, either with only the SIS or NBD domain resolved or with the linker region missing (e.g. 2O3F, 3IWF, 4IVN, 7EN7, 7EN5, 7EN6, 7EN7 in the PDB database). The structure of RccR presented in this study represents the first complete structure of an RpiR regulator, providing a foundation for understanding the molecular mechanism of this family.

Toward the functional oligomerization state of RpiR regulators, the structure of NanR in complex with N-acetylmannosamine 6-phosphate (ManNAc-6P) suggests that it binds the DNA as a dimer, while a recent study by Zhang et al. reported that MurR functions as a tetramer (25). In the case of RccR, the gel filtration result (Supplementary Figure S6) and the strong interactions between the symmetry mates of RccR suggest that it functions as a homotetramer, similar to MurR. Future investigations are required to clarify whether different RpiR homologs function in different oligomeric forms or if there are unique members that can switch between different oligomerization states under specific regulatory signals.

Despite sequence diversity in the DBD of RipR regulators (Supplementary Figure S8B), there are several conserved basic residues in this family. Structural-based sequence alignment revealed that residues R53 and R56, located on α5, are critical for DNA binding. Additionally, secondary contacts with DNA may be mediated by other negative charged residues in the HTH bundle or even by C-terminal extensions from neighboring subunit, such as R271 and R274. Therefore, further studies on the complex structure of RccR with DNA will help elucidate the molecular mechanism of RpiR binding to DNA. The ligand binding domain of RpiR is homologous to the SIS domain of glucosamine-6-phosphate synthetase. While the catalytic function of the SIS domain in RpiR has been lost during evolution, it has retained the recognition and binding properties of phosphorylated sugar molecules. Consequently, the effectors of NanR, MurR, and RccR all contain phosphate groups. Structural and sequence alignment analyses have shown that four highly conserved hydrophilic residues (S/T) play a key role in recognizing phosphate groups, thereby ensuring stable binding and proper positioning of small molecules.

In the OCS family, members act as regulatory switches, and conformational rearrangements induced by effector binding are a common mechanism, often mediated by interdomain linkers. In the RccR structure, a dense network of hydrogen bonds and hydrophobic or polar contacts was found to stabilize the linker conformation (Supplementary Table S4). Considering that this region has been too flexible to be accurately modeled in previously solved RpiR/AlsR family structures, we can speculate that the linker region undergoes significant conformational changes along with interdomain movements.

The MD simulations reveal that the dynamic behavior of RccR is governed by the C-terminus α15, which tends to move towards the DNA-binding domain (DBD) in KDPG-bound states, resulting in a stabilized distance between adjacent DBD domains within a range of approximately 61.1–69.1 angstroms. The breaking and formation of charge-charge interactions, facilitated by C-terminal basic residues, allow α15 to transmit ligand-binding-induced changes in DBD orientation relative to the SIS domain. In the Apo state, the adjacent DBD domains tend to stretch in the opposite direction, making them suitable for recognizing long spacer regulatory elements. Moreover, the flexible interdomain motions within Apo RccR confer greater inclusiveness in recognizing DNA motifs. Collectively, the structure comparison and MD simulations demonstrate that the binding of KDPG leads to different conformations of the DBD, thereby elucidating the molecular basis of RccR’s ability to sense intracellular signals and regulate carbon metabolism.

Conclusion

Collectively, our findings demonstrate that RccR is a crucial regulator in the carbon metabolism pathway, and its dysfunction profoundly impacts the drug resistance and virulence of P. aeruginosa. The structural analysis of RccR provides valuable insights into its precise binding to the signal molecule KDPG, thereby enhancing the sensitivity and specificity of its transcriptional response. Additionally, the structures obtained through molecular dynamics simulations shed light on the conformational dynamics of RccR during signal response and transcriptional regulation. Furthermore, our work elucidates the biochemical mechanisms by which clinical mutations located in the ligand-binding domain of RccR confer drug resistance. Taken together, our results contribute to a deeper understanding of the function and structural basis of RpiR-type regulators in bacteria, paving the way for future advancements in this field.

Data availability

Atomic coordinates of the refined structures have been deposited in the Protein Data Bank (www.pdb.org) with the PDB code 8JU9.

Supplementary data

Supplementary Data are available at NAR Online.

Funding

Ministry of Science and Technology of China [MoST 2022YFC2303700, 2021YFA1301900]; National Natural Science Foundation of China [81871615, 32222040, 32070049]. Funding for open access charge: Ministry of Science and Technology of China [MoST 2022YFC2303700, 2021YFA1301900]; National Natural Science Foundation of China [81871615, 32222040, 32070049].

Conflict of interest statement. None declared.