Pseudomonas aeruginosa Polynucleotide Phosphorylase Controls Tolerance to Aminoglycoside Antibiotics by Regulating the MexXY Multidrug Efflux Pump

Pseudomonas aeruginosa is an opportunistic pathogen that shows high intrinsic resistance to a variety of antibiotics. The MexX-MexY-OprM efflux pump plays an important role in bacterial resistance to aminoglycoside antibiotics.

ABSTRACT

Pseudomonas aeruginosa is an opportunistic pathogen that shows high intrinsic resistance to a variety of antibiotics. The MexX-MexY-OprM efflux pump plays an important role in bacterial resistance to aminoglycoside antibiotics. Polynucleotide phosphorylase (PNPase) is a highly conserved exonuclease that plays important roles in RNA processing and the bacterial response to environmental stresses. Previously, we demonstrated that PNPase controls the tolerance to fluoroquinolone antibiotics by influencing the production of pyocin in P. aeruginosa. In this study, we found that mutation of the PNPase-encoding gene (pnp) in P. aeruginosa increases bacterial tolerance to aminoglycoside antibiotics. We further demonstrate that the upregulation of the mexXY genes is responsible for the increased tolerance of the pnp mutant. Furthermore, our experimental results revealed that PNPase controls the translation of the armZ mRNA through its 5′ untranslated region (UTR). ArmZ had previously been shown to positively regulate the expression of mexXY. Therefore, our results revealed a novel role of PNPase in the regulation of armZ and subsequently the MexXY efflux pump.

INTRODUCTION

Pseudomonas aeruginosa is an opportunistic pathogen that infects immunocompromised individuals, cystic fibrosis patients, and burn victims (1). The high intrinsic antibiotic resistance of P. aeruginosa greatly impedes clinical treatment. The bacterium possesses multiple resistance mechanisms, including β-lactamase (2), low membrane permeability (3), pyocyanin (4), and multidrug efflux systems (5). There are at least 12 resistance-nodulation-division (RND) multidrug efflux systems in P. aeruginosa. MexA-MexB-OprM is a major multidrug efflux system that contributes to bacterial intrinsic resistance to β-lactams, fluoroquinolones, macrolides, aminoglycosides, tetracycline, and chloramphenicol, etc. (6–9). MexX-MexY-OprM is an inducible system that plays an important role in resistance to aminoglycosides and macrolides (10, 11). The expression of the mexX-mexY operon is mainly regulated by a pathway that is composed of ArmZ and MexZ in response to ribosome-targeting antibiotics (12, 13). In the absence of antibiotics, the transcription of mexXY is repressed by MexZ. Antibiotics that target the ribosome, such as aminoglycosides and macrolides, cause ribosome stalling at the mRNA of the leader peptide (PA5471.1) of armZ, which alters the RNA secondary structure, resulting in the transcription of armZ mRNA. ArmZ then binds to MexZ, derepressing the expression of the mexX-mexY operon (14–16).

Polynucleotide phosphorylase (PNPase) is a highly conserved exonuclease that exists in bacteria and eukaryotes (17). The crystal structures of PNPase in Escherichia coli (18, 19), Streptomyces antibioticus (20), and Homo sapiens (21) have been solved. PNPase is mainly composed of two PH domains at the N terminus and KH and S1 domains at the C terminus (22–24). PNPase plays multiple roles in RNA processing, including 3′-to-5′ exoribonuclease activity in the presence of Mg²⁺ and inorganic phosphate (P_i), 3′-to-5′ exodeoxyribonuclease activity, polymerizing deoxyribonucleoside diphosphates (dNDPs) into single-stranded DNA (ssDNA) and polymerizing ribonucleoside diphosphate (rNDP) into RNA without a template (25–30). PNPase has been shown to be involved in rRNA degradation (31, 32), tRNA processing (33–35), and small RNA (sRNA)-mediated gene regulation (27, 36). In addition, PNPase, RNase E, RNA helicase, and enolase form an RNA degradosome that plays an important role in mRNA decay (37).

Previously, we demonstrated that PNPase is essential for P. aeruginosa viability. However, a strain with a deletion of the KH and S1 domains was viable, increased the stabilities of the small RNAs RsmY and RsmZ, and promoted the expression of the type VI secretion system (H1-T6SS) (38). Meanwhile, PNPase influences the translation of the rhlI mRNA through a small RNA, P27 (39). In addition, PNPase is involved in bacterial resistance to fluoroquinolone antibiotics by influencing the 5′ untranslated region (UTR) of prtR (40). Here, in this study, we found that mutation of the pnp gene increases bacterial tolerance to aminoglycoside antibiotics due to the upregulation of the MexXY multidrug efflux pump by influencing the translation of armZ. We further demonstrate that the 5′ UTR of the armZ mRNA is involved in PNPase-mediated translational repression.

RESULTS

PNPase influences bacterial tolerance to aminoglycosides.

We previously demonstrated a role of PNPase in resistance to ciprofloxacin in P. aeruginosa (40). In this study, we examined whether PNPase controls bacterial resistance to aminoglycoside antibiotics. Wild-type PAK and an isogenic mutant with a deletion of the KH-S1 domains of PNPase (ΔKH-S1) displayed similar levels of resistance (MICs) to amikacin, gentamicin, neomycin, and tobramycin (Tables 1 and 2). However, following treatment with amikacin, gentamicin, neomycin, and tobramycin, the survival rates of the ΔKH-S1 mutant were approximately 10,000-, 500-, 1,000-, and 100-fold higher than those of wild-type PAK, respectively (Fig. 1A to D), indicating a role of PNPase in bacterial tolerance to aminoglycoside antibiotics.

TABLE 1.

Bacterial susceptibilities to ciprofloxacin, ofloxacin, carbenicillin, meropenem, and erythromycin with a standard inoculum of 1 × 10⁵ CFU/well

Strain	MIC (μg/ml)
	Ciprofloxacin	Ofloxacin	Carbenicillin	Meropenem	Erythromycin
PAK	0.25	1.5	150	0.15	125
ΔKH-S1	1	3	150	0.15	125
ΔKH-S1/Tn7T-pnp	0.25	1.5	150	0.15	125

TABLE 2.

Bacterial susceptibilities to tetracycline, tobramycin, amikacin, neomycin, and gentamicin with a standard inoculum of 1 × 10⁵ CFU/well

Strain	MIC (μg/ml)a
	Tetracycline	Tobramycin	Amikacin	Neomycin	Gentamicin
PAK	15	1.25	2.5	25	2.5
ΔKH-S1	7.5	1.25	2.5	25	2.5
ΔKH-S1/Tn7T-pnp	15	1.25	2.5	25	—

^a“—” indicates that the complemented strain is resistant to gentamicin due to the mini-Tn7T insertion.

FIG 1.

PNPase influences bacterial tolerance to aminoglycosides. PAK, the ΔKH-S1 mutant, and the ΔKH-S1/Tn7T-pnp complemented strain were grown to an OD₆₀₀ of 0.8 to 1.0 at 37°C and treatedwith 10 μg/ml amikacin (A), 5 μg/ml tobramycin (B), 100 μg/ml neomycin (C), and 10 μg/ml gentamicin (D). At the indicated time points (2 h, 4 h, 6 h, and 8 h), the bacterial survival rates were determined by serial dilution and plating assays. Complementation was achieved by the insertion of a pnp gene into the chromosome by mini-Tn7, which carries a gentamicin resistance gene. Thus, the survival rate of the complemented strain under gentamicin treatment was not tested.

Upregulation of mexXY contributes to increased tolerance to aminoglycosides in the ΔKH-S1 mutant.

To understand the mechanism of PNPase-mediated regulation in antibiotic tolerance, we determined the expression levels of the mexA, mexB, mexX, and mexY genes, which play major roles in bacterial tolerance against aminoglycosides. Real-time PCR (RT-PCR) results revealed a ≥5-fold upregulation of mexX and mexY in the ΔKH-S1 mutant, whereas the expression levels of mexA and mexB were not altered. Complementation with a pnp gene restored the mRNA levels in the ΔKH-S1 mutant (Fig. 2A). To further confirm the expression of mexX, we utilized C-terminally 6×His-tagged mexX (mexX-His) driven by its native promoter (16, 41). The MexX-His level was higher in the ΔKH-S1 mutant (Fig. 2B). Deletion of the mexXY operon in wild-type PAK and the ΔKH-S1 mutant resulted in similar survival rates upon treatment with amikacin and gentamicin (Fig. 2C), indicating a role of MexXY in the increased aminoglycosides tolerance of the ΔKH-S1 mutant.

FIG 2.

Upregulation of mexXY contributes to the increased tolerance to aminoglycosides in the ΔKH-S1 mutant. (A) PAK, ΔKH-S1, and ΔKH-S1/Tn7T-pnp strains were grown in LB broth to an OD₆₀₀ of 0.8 to 1.0 at 37°C, followed by RNA extraction. The mRNA levels of mexA, mexB, mexX, and mexYwere determined by real-time PCR with rpsL as the internal control. ***, P < 0.001 by Student’s t test; ns, not significant. (B) Protein levels of MexX-His in PAK and the ΔKH-S1 mutant. Bacterial cells were grown in LB broth to an OD₆₀₀ of 0.8 to 1.0, and the MexX-His levels were determined by Western blotting. RpoA was used as the loading control. (C) Wild-type PAK, ΔKH-S1, ΔmexXY, and ΔKH-S1 ΔmexXY strains were grown in LB broth to an OD₆₀₀ of 0.8 to 1.0 at 37°C and treated with 10 μg/ml amikacin and 10 μg/ml gentamicin. At the indicated time points (2 h, 4 h, 6 h, and 8 h), the bacterial survival rates were determined by serial dilution and plating assays. (D) Wild-type PAK and the ΔKH-S1 mutant were grown in LB broth to an OD₆₀₀ of 0.8 to 1.0 at 37°C, followed by treatment with 1.25 μg/ml amikacin or 300 μg/ml chloramphenicol for 1 h. The mRNA levels of mexX were determined by real-time PCR with rpsL as the internal control. ***, P < 0.001 by Student’s t test. (E) Wild-type PAK and the ΔKH-S1 mutant were grown in LB broth to an OD₆₀₀ of 0.8 to 1.0 at 37°C, followed by treatment with 300 μg/ml chloramphenicol for 1 h. The bacteria were washed once with fresh LB broth and then treated with 10 μg/ml amikacin. At the indicated time points (2 h, 4 h, 6 h, and 8 h), the bacterial survival rates were determined by serial dilution and plating assays.

If the upregulation of mexXY plays a major role in the increased tolerance, preinduction of the pump might reduce the difference in the bacterial survival rates between the wild-type strain and the ΔKH-S1 mutant. Treatment with amikacin at 1/2 MIC induced the expression of mexX by 7- and 2-fold in the wild-type strain and the ΔKH-S1 mutant, respectively, resulting in a 2-fold difference between the two strains (Fig. 2D). Treatment with a bacteriostatic antibiotic, chloramphenicol, drastically induced the expression of mexX and resulted in similar mRNA levels in the two strains (Fig. 2D). We then treated wild-type PAK and the ΔKH-S1 mutant with chloramphenicol for 1 h. After the removal of chloramphenicol, we treated the bacteria with amikacin. Indeed, pretreatment with chloramphenicol increased the survival rate of wild-type PAK by approximately 50-fold. However, pretreatment with chloramphenicol did not further increase the survival rate of the ΔKH-S1 mutant, which might be due to the already high tolerance (∼50% survival rate) without pretreatment (Fig. 2E). Therefore, preinduction of the MexXY pump reduced the difference in survival rates between wild-type PAK and the ΔKH-S1 mutant under treatment with amikacin.

We then suspected that in the MIC assay, the surviving wild-type cells with the MexXY pump induced might be able to grow after 24 h, resulting in an MIC reading similar to that of the ΔKH-S1 mutant. To test our hypothesis, we reduced the initial inoculum from 10⁵ to 10³ CFU per well, which resulted in 2-fold-lower MICs of the tested aminoglycosides against wild-type PAK than against the ΔKH-S1 mutant (Table 3). In combination, these results indicated that the upregulation of mexXY contributes to the increased tolerance to aminoglycosides in the ΔKH-S1 mutant.

TABLE 3.

Bacterial susceptibilities to antibiotics at a low inoculum of 1 × 10³ CFU/well

Strain	MIC (μg/ml)a
	Amikacin	Gentamicin	Tobramycin	Neomycin
PAK	1.25	1.25	0.625	6.25
ΔKH-S1	2.5	2.5	1.25	25
ΔKH-S1/Tn7T-pnp	1.25	—	0.625	12.5
PAK ΔmexZ	2.5	2.5	ND	ND
ΔKH-S1 ΔmexZ	2.5	2.5	ND	ND
PAK ΔmexXY	0.3125	0.3125	ND	ND
ΔKH-S1 ΔmexXY	0.3125	0.3125	ND	ND
PAK ΔarmZ	0.3125	0.3125	ND	ND
ΔKH-S1 ΔarmZ	0.3125	0.3125	ND	ND

^a“—” indicates that the complemented strain is resistant to gentamicin due to the mini-Tn7T insertion. ND, not determined.

Upregulation of ArmZ contributes to the increased antibiotic tolerance in the ΔKH-S1 mutant.

Transcription of the mexXY operon is directly repressed by MexZ. The interaction between ArmZ and MexZ releases the repression (12). We then determined the expression levels of MexZ and ArmZ by utilizing 6×His-tagged mexZ (mexZ-His) and armZ (armZ-His) driven by their respective native promoters (Fig. 3A). The protein levels of both MexZ-His and ArmZ-His were increased in the ΔKH-S1 mutant. On the P. aeruginosa chromosome, the mexZ gene and the mexXY operon are located next to each other and transcribed in the opposite directions (Fig. 3B). We suspected that the binding of MexZ to the intergenic region between mexZ and mexXY might repress the transcription of mexZ itself (autorepression), and thus, the upregulation of ArmZ might increase the expression of mexZ. Indeed, by utilizing a transcriptional fusion between the mexZ promoter and a lacZ reporter gene (P_mexZ-lacZ), we found that deletion of mexZ increased its promoter activity, whereas overexpression of mexZ repressed the promoter activity (Fig. 3C), thus proving the autorepressive regulation of mexZ. In addition, deletion of mexZ in wild-type PAK resulted in higher MICs of gentamicin and amikacin (Table 3) and increased the bacterial survival rate to the same level as that of the ΔKH-S1 mutant under treatment with gentamicin and amikacin (Fig. 3D). However, deletion of mexZ in the ΔKH-S1 mutant did not further increase the MICs or bacterial survival rates (Fig. 3D). These results indicated that the inactivation of MexZ might play a major role in the increased resistance of the ΔKH-S1 mutant.

FIG 3.

Upregulation of armZ contributes to increased antibiotic tolerance in the ΔKH-S1 mutant. (A) C-terminally 6×His-tagged mexZ and armZ are driven by their own promoters. The bacteria were grown in LB broth to an OD₆₀₀ of 0.8 to 1.0, and the MexZ-His and ArmZ-His levels were determined by Western blotting. RpoA was used as the loading control. (B) Location and transcription directions of the mexZ gene and the mexXY operon on the chromosome of P. aeruginosa. MexZ binds to the mexZ-mexX intergenic region. (C) Expression of P_mexZ-lacZ in PAK, the ΔmexZ mutant, and the ΔmexZmutant containing an empty vector or the mexZ overexpression plasmid. The bacteria were grown in LB broth to an OD₆₀₀ of 0.8 to 1.0, and the bacterial cells were subjected to a β-galactosidase assay. ***, P < 0.001 by Student’s t test. (D) Wild-type PAK, ΔKH-S1, ΔmexZ, and ΔKH-S1 ΔmexZ strains were grown in LB broth to an OD₆₀₀ of 0.8 to 1.0 at 37°C and treated with 10 μg/ml amikacin or 10 μg/ml gentamicin. At the indicated time points (2 h, 4 h, 6 h, and 8 h), the bacterial survival rates were determined by serial dilution and plating assays.

We next examined the expression level of mexZ in the ΔKH-S1 mutant. The mexZ promoter activity and its mRNA level were increased in the ΔKH-S1 mutant, which were decreased by the deletion of armZ (Fig. 4A and B). In addition, the deletion of armZ in the ΔKH-S1 mutant (ΔarmZ ΔKH-S1) reduced the expression of the mexXY operon (Fig. 4C), and bacterial tolerance to amikacin and gentamicin returned to levels similar to those of the armZ deletion mutant (ΔarmZ) (Fig. 4D). In combination, these results demonstrated that the upregulation of ArmZ in the ΔKH-S1 mutant contributes to the upregulation of mexXY, which subsequently increases bacterial tolerance to aminoglycosides.

FIG 4.

Deletion of armZ eliminates KH-S1-mediated bacterial tolerance to aminoglycosides. (A) PAK, ΔKH-S1, ΔKH-S1/Tn7T-pnp, ΔarmZ, and ΔKH-S1 ΔarmZ strains were grown in LB broth to an OD₆₀₀ of 0.8 to 1.0 at 37°C, followed by RNA extraction. The mRNA level of mexZ was determined by real-time PCR with rpsL as the internal control. ***, P < 0.001 by Student’s t test. (B) Expression of P_mexZ-lacZ in PAK, ΔKH-S1, ΔKH-S1/Tn7T-pnp, ΔarmZ, and ΔKH-S1 ΔarmZ strains. The bacteria were grown in LB broth to an OD₆₀₀ of 0.8 to 1.0. Bacterial cells were collected by centrifugation and then subjected to a β-galactosidase assay. ***, P < 0.001 by Student’s t test. (C) PAK, ΔKH-S1, ΔarmZ, and ΔKH-S1 ΔarmZ strains were grown in LB broth to an OD₆₀₀ of 0.8 to 1.0 at 37°C, followed by RNA extraction. The mRNA levels of mexX and mexY were determined by real-time PCR with rpsL as the internal control. ***, P < 0.001 by Student’s t test. (D) The indicated strains were grown to an OD₆₀₀ of 0.8 to 1.0 at 37°C and treated with 10 μg/ml amikacin and 10 μg/ml gentamicin. At the indicated time points (2 h, 4 h, 6 h, and 8 h), the bacterial survival rates were determined by serial dilution and plating assays.

PNPase controls the translation of ArmZ through the 5′ UTR of its mRNA.

To understand the mechanism of the increased ArmZ protein level, we examined the mRNA level of the armZ gene by real-time PCR. The armZ mRNA levels were similar between the ΔKH-S1 mutant and PAK (Fig. 5A), indicating posttranscriptional regulation. Indeed, when the ArmZ native promoter of armZ-His was replaced with an exogenous inducible P_BAD promoter (P_BAD-PA5471.1-armZ-His), the ArmZ-His level remained high in the ΔKH-S1 mutant (Fig. 5C). However, when we replaced the native 5′ UTR with an exogenous ribosome binding site (RBS) from the vector pET28a (P_BAD-SD-armZ-His), the ArmZ-His levels became similar in the ΔKH-S1 mutant and wild-type PAK (Fig. 5C). These results indicated that the 5′ UTR of armZ is involved in posttranscriptional regulation.

FIG 5.

armZ is regulated by PNPase at the posttranscriptional level through the 5′ UTR of its mRNA. (A) PAK, the ΔKH-S1 mutant strain, and the ΔKH-S1/Tn7T-pnp complemented strain were grown in LB broth to an OD₆₀₀ of 0.8 to 1.0 at 37°C, followed by RNA extraction. The mRNA level ofarmZ was determined by real-time PCR with rpsL as the internal control. (B) Structures of the 6×His-tagged armZ and gfp fusions. C-terminally 6×His-tagged ArmZ was driven by an inducible P_BAD promoter with 275 bp of the 5′-UTR sequence of the armZ gene or with an exogenous ribosome binding site from the plasmid pET28a on the plasmid pUCP20 (without the tac promoter). P_BAD-armZ-His, P_BAD-86-armZ-His, P_BAD-36-armZ-His, and P_BAD-15-armZ-His represent 6×His-tagged armZ driven by the P_BAD promoter with 212, 86, 36, and 15 bp of the 5′-UTR sequence of the armZ gene, respectively. The armZ open reading frame was replaced by a gfp gene, resulting in P_BAD-36-armZ-GFP and P_BAD-15-armZ-GFP. (C) Strains containing the ArmZ-His or gfp expression plasmid were grown in LB broth to an OD₆₀₀ of 0.8 to 1.0 with 0.05% arabinose. Protein levels of ArmZ and GFP were determined by Western blotting. RpoA was used as the loading control. Lanes: 1, PAK; 2, ΔKH-S1; 3, ΔKH-S1/Tn7T-pnp.

To determine the region involved in the posttranscriptional regulation of armZ, we constructed a series of P_BAD-driven armZ-His fusions with truncated 5′ UTRs of the armZ mRNA (Fig. 5B). When the 5′ UTRs were ≥36 bp, the levels of ArmZ-His were higher in the ΔKH-S1 mutant. However, when the 5′ UTR was reduced to 15 bp, similar levels of ArmZ-His were observed in the ΔKH-S1 mutant and wild-type PAK (Fig. 5C).

To examine whether the armZ coding region is involved in posttranscriptional regulation, we replaced the armZ coding sequence with a green fluorescent protein (GFP) gene (gfp), obtaining P_BAD-(36)-gfp and P_BAD-(15)-gfp, respectively (Fig. 5B). Fusion with the 36-bp 5′ UTR of armZ resulted in higher GFP levels in the ΔKH-S1 mutant, whereas fusion with the 15-bp 5′ UTR resulted in similar GFP levels in the ΔKH-S1 mutant and wild-type PAK (Fig. 5C). These results indicated that the 36-bp 5′ UTR of armZ plays an important role in the PNPase-mediated posttranscriptional regulation of armZ.

Analysis with Mfold (42) revealed a potential hairpin structure in the 5′ UTR of armZ mRNA, which might block the ribosome binding site (Fig. 6A). To examine whether the secondary structure affects the translation of ArmZ, we changed the original TTCAATC sequence of the 5′ UTR to AAATTAA (Fig. 6A), which might abolish the potential hairpin structure. In wild-type PAK, the mutation increased ArmZ-His to a level similar to that in the ΔKH-S1 mutant (Fig. 6B). These results demonstrated a critical role of the sequence (TTCAATC) in the PNPase-mediated posttranscriptional regulation of ArmZ, which might be due to its role in affecting the secondary structure of the 5′ UTR of armZ mRNA.

FIG 6.

The sequence of the 5′ UTR of armZ mRNA affects the translation of armZ. (A) Hairpin structure predicted by Mfold in the 5′ UTR of armZ mRNA. The potential ribosome binding sequence (RBS) is indicated by a dashed line. The original TTCAATC sequence is replaced by AAATTAA. (B) C-terminally 6×His-tagged armZ was driven by the P_BAD promoter with the original or mutated 5′-UTR sequence. Strains containing the armZ-His expression plasmid were grown in LB broth to an OD₆₀₀ of 0.8 to 1.0 with 0.05% arabinose. Protein levels of ArmZ were determined by Western blotting. RpoA was used as the loading control. WT, wild type.

DISCUSSION

In this study, we demonstrated that PNPase controls the expression of mexXY and, consequently, bacterial tolerance to aminoglycoside antibiotics. PNPase is a conserved exoribonuclease that plays pleiotropic roles in the bacterial response to environmental stresses. In Yersinia, PNPase is involved in bacterial survival under oxidative and cold stresses (43). In E. coli and Bacillus subtilis, PNPase contributes to bacterial survival under oxidative stresses and UV radiation by promoting DNA repair (25, 44–46). In Yersinia, salmonellae, and Helicobacter pylori, PNPase has further been reported to be involved in the regulation of bacterial virulence (38, 47–49). PNPase may execute its regulatory function by influencing sRNA stability or as a component of the RNA degradosome.

In this study, we explored the mechanism of PNPase-mediated regulation of mexXY. Deletion of the KH-S1 domains results in the upregulation of ArmZ, which releases the MexZ-mediated repression of mexXY. Previous studies demonstrated that armZ is regulated through a transcription attenuation mechanism whereby the bacterium can quickly respond to antibiotic-caused ribosome stalling (16, 41). Our results revealed that PNPase controls the translation of armZ through the 5′ UTR of the armZ mRNA. Alteration of the 5′-UTR sequence in wild-type PAK abolished the PNPase-mediated repression of the translation of armZ.

Previous studies in E. coli demonstrated the autoregulation of PNPase at the posttranscriptional level (50, 51). The 5′ UTR of the PNPase mRNA forms a long stem-loop, which is cleaved by RNase III, resulting in a free 3′ end. As PNPase degrades RNAs from the 3′ end, the stability of the PNPase mRNA is reduced (52). However, we found that deletion of the RNase III-encoding gene rnc did not affect the expression of ArmZ in wild-type PAK (data not shown), indicating that PNPase does not control the expression of ArmZ through a similar mechanism.

An analysis of the 5′ UTR of the armZ mRNA by Mfold (42) revealed a potential hairpin structure, which might block the ribosome binding site (RBS) (Fig. 6A). We speculated that PNPase might play a role in altering the secondary structure. One of the possible mechanisms could be that an sRNA binds to the upstream half of the duplex, which exposes the RBS and facilitates translation. It has been demonstrated that PNPase controls sRNA stabilities in various bacteria (38, 53, 54). We have previously demonstrated that PNPase directly degrades the sRNAs RsmY, RsmZ, and P27 in P. aeruginosa. Therefore, it is likely that mutation of pnp might result in the accumulation of an sRNA that disrupts the hairpin structure. The presence of ribosome-targeting antibiotics might induce the expression of a protein or sRNA that inhibits the function of PNPase, thus derepressing the translation of the armZ mRNA. A previous study identified that an sRNA, PA0805.1, enhances the expression of mexXY (55). Real-time PCR revealed the upregulation of PA0805.1 in the ΔKH-S1 mutant (data not shown). However, deletion of PA0805.1 in the ΔKH-S1 mutant did not reduce the expression of armZ or mexXY (data not shown). Another possibility might be that PNPase directly binds to the 5′ UTR of the armZ mRNA and inhibits its translation. We have tried to examine the direct interaction between the PNPase KH-S1 domain and an RNA spanning the 36-nucleotide (nt) 5′ UTR and the 142-nt coding sequence of the armZ mRNA by an electrophoretic mobility shift assay (EMSA). However, no obvious interaction was observed in our experiment. We suspected that the expression of PNPase might be affected by ribosome-targeting antibiotics at either the transcriptional or translational level. Nevertheless, PNPase-mediated regulation of the translation of armZ adds a novel layer of regulation of the expression of the MexXY efflux system. Further studies are needed to explore whether PNPase controls the expression of armZ through an sRNA or other mechanisms.

MexZ directly represses the transcription of the mexXY operon by binding to its promoter. In this study, we demonstrated the autorepression of MexZ. It is likely that the binding of MexZ to the intergenic region between the mexZ gene and the mexXY operon also represses the transcription of mexZ. The autorepression of MexZ might enable the bacterium to fine-tune the expression of mexXY in response to antibiotics. The presence of ribosome-targeting antibiotics induces the expression of ArmZ, which antagonizes the function of MexZ, resulting in the upregulation of mexXY and mexZ. Reduction of the antibiotic concentration reduces the transcription of armZ. The MexZ protein then represses the transcription of mexXY and its own gene, thus keeping the expression level of mexXY in accordance with the antibiotic level. Therefore, mutation of mexZ in clinical isolates (56, 57) disrupts the regulatory circuity, resulting in the constitutive expression of mexXY. The function of PNPase and the regulatory mechanisms of PA5470, armZ, mexZ, and mexXY are summarized in Fig. 7.

FIG 7.

Regulatory mechanisms and functions of PNPase, armZ, PA5470, and mexX-mexY. The armZ-PA5470 operon is regulated through a transcription attenuation mechanism. The transcription of armZ and its leader peptide (PA5471.1) is driven by a constitutive promoter. Under normal growth conditions, the complete translation of PA5471.1 leads to the formation of a transcription terminator (the hairpin structure formed by the pairing between segments 3 and 4), which blocks the transcription of downstream armZ and PA5470 (14–16). Meanwhile, MexZ binds to the intergenic region between its coding region and the mexXY operon, repressing the transcription of the genes. Ribosome-targeting antibiotics such as aminoglycosides and macrolides cause ribosome stalling at the PA5471.1 mRNA, which alters the RNA secondary structure and abolishes the formation of the transcription terminator structure, leading to the transcription of armZ and PA5470. PNPase represses the translation of the armZ mRNA through its 5′ UTR. A possible mechanism is that an sRNA binds to the 5′ UTR, which eliminates a potential hairpin structure and exposes the ribosome binding site (RBS) of the armZ mRNA. Thus, the degradation of the sRNA by PNPase represses translation. The ArmZ protein binds to MexZ to release its repression of the transcription of its own gene and mexXY (12, 13). The upregulation of the MexXY multidrug efflux system enhances bacterial antibiotic resistance. The autoregulation of mexZ might contribute to the homeostatic expression of mexXY in response to antibiotics. Meanwhile, PA5470 might function as a peptide chain release factor that rescues the stalled ribosomes, thus contributing to antibiotic resistance (63). OM, outer membrane; IM, inner membrane.

In summary, we found that PNPase controls bacterial tolerance to aminoglycoside antibiotics and identified the region of the mRNA that is involved in the PNPase-mediated regulation of armZ. Our results reveal a novel layer of regulation that fine-tunes the expression of mexXY.

MATERIALS AND METHODS

Bacterial strains, growth conditions, plasmids, and primers.

The bacterial strains, plasmids, and primers used in this study are listed in Table 4 (38, 58, 59). E. coli and P. aeruginosa strains were grown in Luria-Bertani (LB) broth (5 g/liter sodium chloride, 5 g/liter yeast extract, and 10 g/liter tryptone [pH 7.4]) at 37°C with agitation at 180 to 200 rpm. The construction of mutant strains was performed as described previously (58).

TABLE 4.

Bacterial strains, plasmids, and primers used in this study

Strain, plasmid, or primer	Description or sequence (5′→3′)	Source, reference,purpose, or function
P. aeruginosa strains
PAK	Wild-type strain of Pseudomonas aeruginosa	David Bradley
ΔKH-S1	PAK with pnp (KH and S1) deletion	38
ΔKH-S1/Tn7T-pnp	PAK ΔKH-S1 with pnp inserted on the chromosome with a mini-Tn7T insertion	38
PAK ΔmexXY	PAK deleted of mexXY	This study
PAK ΔarmZ	PAK deleted of armZ	This study
ΔKH-S1 ΔmexXY	PAK ΔKH-S1 deleted of mexXY	This study
ΔKH-S1 ΔarmZ	PAK ΔKH-S1 deleted of armZ	This study
PAK ΔmexZ	PAK deleted of mexZ	This study
Plasmids
pEX18Tc	Gene replacement vector; Tcr oriT+ sacB+	58
pMMB67EH	Expression vector with a tac promoter; Apr	59
pMMB67EH-mexZ	MexZ on pMMB67EH	This study
pUCP20(no promoter)	Escherichia-Pseudomonas shuttle vector; no promoter; Ampr	This study
pUCP20(no promoter)-PmexZ-LacZ	mexZ promoter and lacZ gene on pUCP20(no promoter)	This study
pUCP20(no promoter)-pRkaraRed-PA5471.1-armZ-His	6×His-tagged armZ driven by the PBAD promoter with 275 bp of the 5′-UTR sequence on pUCP20(no promoter)	This study
pUCP20(no promoter)-pRkaraRed(pET28a)-armZ-His	6×His-tagged armZ driven by the PBAD promoter with the ribosome binding site from pET28a on pUCP20(no promoter)	This study
pUCP20(no promoter)-pRkaraRed-armZ-His	6×His-tagged armZ driven by the PBAD promoter with 212 bp of the 5′-UTR sequence on pUCP20(no promoter)	This study
pUCP20(no promoter)-pRkaraRed(−86)-armZ-His	6×His-tagged armZ driven by the PBAD promoter with 86 bp of the 5′-UTR sequence on pUCP20(no promoter)	This study
pUCP20(no promoter)-pRkaraRed(−36)-armZ-His	6×His-tagged armZ driven by the PBAD promoter with 36 bp of the 5′-UTR sequence on pUCP20(no promoter)	This study
pUCP20(no promoter)-pRkaraRed(−15)-armZ-His	6×His-tagged armZ driven by the PBAD promoter with 15 bp of the 5′-UTR sequence on pUCP20(no promoter)	This study
pUCP20(no promoter)-pRkaraRed(−36)-GFP	GFP driven by the PBAD promoter with 36 bp of the 5′-UTR sequence on pUCP20(no promoter)	This study
pUCP20(no promoter)-pRkaraRed(−15)-GFP	GFP driven by the PBAD promoter with 15 bp of the 5′-UTR sequence on pUCP20(no promoter)	This study
Primers
mexA-RT-S	AGGTGCGTCCCCAGGTG	RT-PCR
mexA-RT-AS	CCTCGTAGGTGGCGGG	RT-PCR
mexB-RT-S	AACGTGCAGATTTCCTCCG	RT-PCR
mexB-RT-AS	GATGTTCTCGAATTGCTCCG	RT-PCR
mexX-RT-S	TGCGAAGAAGCAGCGGA	RT-PCR
mexX-RT-AS	CAGGCGACGGGTGACG	RT-PCR
mexY-RT-S	CTACAACATCCCCTATGACACCT	RT-PCR
mexY-RT-AS	CATCACGGCGAACACCAG	RT-PCR
mexZ-RT-S	ACTACAAGAACAAGATCGAGG	RT-PCR
mexZ-RT-AS	TGGCGTTTTCGTCGGGTA	RT-PCR
armZ-RT-S	GCAAACCACCGCCGAA	RT-PCR
armZ-RT-AS	GCCATGCCACAGCCGA	RT-PCR
SacI-PA5471.1-armZ-S	CGGGAGCTCTGCGCCGCTCATTTCCATCGCG	Translational fusion
BamHI-PA5471.1-armZ-A	CGCGGATCCTCAGTGGTGGTGGTGGTGGTGTCGGCAGCACTCCCCACGGGTC	Translational fusion
XhoI-PBAD-(275)armZ-S	CAACTCTCTACTGTTGGTGGCTCGAGACCGCAAAGGAGGTTACATCGAT	Translational fusion
XhoI-PBAD-(212)armZ-S	CAACTCTCTACTGTTGGTGGCTCGAGCGGCCACCGAAACCTACCGCTC	Translational fusion
XhoI-PBAD-(86)armZ-S	CAACTCTCTACTGTTGGTGGCTCGAGGTTAGACGCCCTTCGCCCTTCCG	Translational fusion
XhoI-PBAD-(36)armZ-S	CAACTCTCTACTGTTGGTGGCTCGAGCCCCGGATCTACCGTTTCAATCA	Translational fusion
XhoI-PBAD-(15)armZ-S	CAACTCTCTACTGTTGGTGGCTCGAGCACATGGATTGGATTATGGGCAA	Translational fusion
XhoI-PBAD-(pET28)armZ-S	CAACTCTCTACTGTTGGTGGCTCGAGGAAGGAGATATACCATGGGCAACTACATCAAGCC	Translational fusion
XhoI-PBAD-(36mut)armZ-S	CAACTCTCTACTGTTGGTGGCTCGAGCCCCGGATCTACCGTAAATTAAACATGGATTGGATTATGGGCAAC	Translational fusion
XhoI-PBAD-(36)GFP-S	CAACTCTCTACTGTTGGTGGCTCGAGCCCCGGATCTACCGTTTCAATCACATGGATTGGATTATGAGTAAAGGAGAAGAACTTTTCAC	Translational fusion
XhoI-PBAD-(15)GFP-S	CAACTCTCTACTGTTGGTGGCTCGAGCACATGGATTGGATTATGAGTAAAGGAGAAGAACTTTTCAC	Translational fusion
HindIII-ArmZ/GFP-AS	GTAAAACGACGGCCAGTGCCAAGCTT	Translational fusion
EcoRI-PmexZ-S	CCGGAATTCAGCGCTTGAGCTTGTCGG	Transcriptional fusion
BamHI-PmexZ-AS	CGCGGATCCGGCGTTTTCGTCGGGTAC	Transcriptional fusion

MIC and survival assays.

The MICs were determined by the 2-fold serial dilution method as previously described (40). The bacteria were grown in LB broth at 37°C until the optical density at 600 nm (OD₆₀₀) reached 0.8 to 1.0. Next, the bacterial concentration was adjusted to 5 × 10⁵ CFU/ml. Two hundred microliters of the bacterial suspension was added to each well of a 96-well plate (Corning). The plate was incubated at 37°C for 24 h without agitation. The lowest antibiotic concentration that inhibited visible bacterial growth was recorded as the MIC. The bacterial survival rate was determined as previously described (40). Briefly, bacteria were cultured in LB broth at 37°C until the OD₆₀₀ reached 1.0. Next, the bacteria were treated with the indicated antibiotics at 37°C with agitation (200 rpm) for up to 8 h. The numbers of live bacterial cells before and after antibiotic treatment were determined by serial dilution and plating.

RNA extraction, reverse transcription, and quantitative RT-PCR.

Total RNA was isolated with a bacterial total RNA kit (Zomanbio, Beijing, China), and cDNA was synthesized using PrimeScript RT master mix (TaKaRa, Dalian, China). In the quantitative RT-PCR experiment, the cDNA was mixed with specific forward and reverse primers (Table 4) and chamQ universal SYBR quantitative PCR (qPCR) master mix (Vazyme, Nanjing, China). Quantitative RT-PCR was performed with the CFX Connect real-time system (Bio-Rad, USA). rpsL, which encodes the 30S ribosomal protein S12, was used as an internal control (60–62).

Western blotting.

Proteins isolated from the same number of bacteria were separated with a 10% or 12% sodium dodecyl sulfate-polyacrylamide gel electrophoresis (SDS-PAGE) gel. Next, the proteins were transferred onto a polyvinylidene difluoride (PVDF) membrane. The membrane was probed with an antibody against GFP (1:1,500; GeneTex), a mouse monoclonal antibody against the 6×His tag (1: 2,000; EMD Millipore Corp., Germany), or an antibody against RpoA (1: 2,000; BioLegend) overnight at 4°C. The PVDF membrane was then washed with 1× phosphate-buffered saline (PBS) containing 0.1% Tween 20 (Solarbio, Beijing, China) 4 times. Next, the PVDF membrane was incubated with anti-mouse IgG (1:2,000; Promega, USA) at room temperature for 45 min. The signals were detected with an ECL Plus kit (Millipore) using a Bio-Rad molecular imager (ChemiDocXRS). The RNA polymerase α subunit RpoA was used as a loading control.

Promoter activity assay.

The promoter region of the mexZ gene was amplified by PCR with the primers P_mexZ-S and P_mexZ-AS (Table 4). The PCR fragment was then cloned into the EcoRI and BamHI sites of pDN19lacZΩ (40), resulting in P_mexZ-lacZ. Bacteria were grown to an OD₆₀₀ of 0.8 to 1.0. A total of 1.0 ml of bacteria was collected by centrifugation (12,000 × g for 2 min) and resuspended in 1.5 ml of Z buffer (50 mM β-mercaptoethanol, 60 mM NaH₂PO₄, 1 mM MgSO₄, 60 mM Na₂HPO₄, and 10 mM KCl). One milliliter of the suspension was used to measure the OD₆₀₀. The remaining 0.5 ml of the suspension was vigorously mixed with 10 μl of chloroform (BBI Life Sciences, Shanghai, China) and 10 μl of 0.1% SDS (BBI Life Sciences, Shanghai, China). One hundred microliters of o-nitrophenyl-β-d-galactopyranoside (ONPG) (40 mg/ml; Sigma, USA) was added to the mixture, followed by incubation at 37°C. Once the color turned light yellow, the reaction was ended by the addition of 0.5 ml 1 M Na₂CO₃. The reaction time was recorded, followed by OD₄₂₀ measurement. The β-galactosidase activity (Miller units) was calculated as (1,000 × OD₄₂₀)/(T × V × OD₆₀₀), where T is the reaction time (in minutes) and V is the bacterial volume (in milliliters).