ABSTRACT
Despite the identification of many genes and pathways involved in the persistence phenomenon of bacteria, the mechanisms of persistence are not well understood. Here, using Escherichia coli as a model, we identified polynucleotide phosphorylase (PNPase) as a key regulator in persister formation. We successfully constructed pnp knockout mutant strain and its complemented strain, and exposed the pnp knockout mutant and complemented strain to antibiotics and stress conditions. The results showed that, compared with the wild-type W3110, the pnp knockout strain had defect in persistence to antibiotics and stress conditions, and the persistence to antibiotics and stresses was restored upon complementation. RNA-Seq was performed to identify the transcriptome profile in the pnp knockout strain compared with wild-type strain W3110, and the data revealed that 242 (166 up-regulated, and 76 down-regulated) genes were differentially expressed in the pnp knockout mutant strain. KEGG pathway analysis of the up-regulated genes showed that they were mostly mapped to metabolism and virulence pathways, most of which are positively regulated by the global regulator cyclic AMP receptor protein (CRP). Similarly, the transcription level of the crp gene in the pnp-deletion strain increased 3.22-fold in the early stationary phase. We further explored the indicators of cellular metabolism of the pnp-deletion strain, the persistence phenotype of the pnp and crp double-deletion mutant, and the transcriptional activity of crp gene. Our results indicate that PNPase controls cellular metabolism by negatively regulating the crp operon at the post-transcriptional level by targeting the 5’- Untranslated Region (UTR) of the crp transcript. This study offers new insight about the persister mechanisms and provides new targets for development of new drugs against persisters for more effective treatment of persistent bacterial infections.
INTRODUCTION
Persisters are a small fraction of dormant or non-growing bacteria that are tolerant to lethal antibiotics or stresses (1). Persisters are distinct from antibiotic-resistant cells in that they are genetically identical and remain susceptible to antibiotics when they are growing again (2, 3). Persisters pose significant challenges for the treatment of many chronic and persistent bacterial infections, such as tuberculosis, urinary tract infections and biofilm infections (4–6). Therefore, it is of great importance to understand the mechanisms of persistence and develop new strategies to more effectively cure such persistent infections. Although the phenomenon of bacterial persistence was discovered over 70 years ago (7), our understanding of the genetic basis of persister formation remains incomplete.
Because the mechanisms of persistence are highly redundant, new mechanisms of persister formation and survival are continually discovered. In bacteria, RNA decay, which is necessary for recycling of nucleotides and for rapid changes in the gene expression program is mainly carried out by RNA degradosome (8). PNPase encoded by pnp is a major component of the RNA degradosome, which is composed of a complex structure with RNase E, helicase RhlB, and enolase together with PNPase. In bacteria, mRNA degradation is of great significance, as it not only achieves the nucleotide recycling, but also can control gene expression in different growth conditions (9, 10). PNPase has been found to be important in many aspects of RNA metabolism (8, 11–13). In addition, it also plays important roles in post-transcriptional regulation of gene expression (14).
In this study, to address the role of RNA degradation in bacterial persistence, we constructed pnp knockout strain and its complementation strain, and then assessed the survival of the pnp gene knockout mutant and the complementation strain upon exposure to antibiotics and stress conditions. In addition, RNA-Seq was performed to evaluate the transcriptome of the pnp knockout strain compared with the parent strain W3110, and the differential expression was analyzed to shed light on the molecular basis of the pnp mediated persistence. We demonstrate that PNPase controls cellular metabolism by negatively regulating the global regulator cyclic AMP receptor protein(CRP) operon at the post-transcriptional level through targeting the 5’-Untranslated Region (UTR) of the crp transcript to regulate persister formation.
MATERIALS AND METHODS
Bacterial strains and growth media
The strains used in this work were derived from wild type E. coli K12 W3110 (F–mcrAmcrBIN(rrnD-rrnE)1 lambda–). Luria-Bertani (LB) broth (0.5% NaCl) and agar were used for bacterial cultivation.
Construction of E. coli W3110 knockout mutants
The λ Red recombination system was used for construction of pnp knockout in the E. coli chromosome (15). The candidate gene was replaced by the chloramphenicol resistance gene, which can be removed by pCP20. Primers used for knockout and additional external primers used to verify the correct integration of the PCR fragments by homologous recombination are shown in Table 1.
Construction of pBAD202-pnp recombinant plasmid
The plasmid pBAD202 was used for construction of a recombinant containing a functional wild type pnp gene (16). The primers designed based on the pnp gene were F (5’-CATGCCATGGACCCACATAGAGCTGGGTTA-3’) and R (5’-CCCAAGCTTGCAAATGGCAACCTTACT-3’). The PCR products were digested with the restriction enzymes NcoI and HindIII (Thermo Fisher, USA) and ligated to the plasmid digested with the same enzymes. The recombinant constructs containing the pnp gene and the vector control were used to transform the Δpnp deletion strain and wild type W3110 strain by electroporation.
Persister assay for various antibiotics
Persistence was measured by determining bacterial survival in the form of colony forming units (CFUs) upon exposure to three antibiotics, namely, ampicillin at 200 μg/ml, norfloxacin at 8 μg/ml, and gentamicin at 40 μg/ml. E. coli cells were grown to stationary phase in LB medium, and then were exposed to different antibiotics, where undiluted cultures were used for incubation without shaking at 37°C for various times (17). The number of CFUs per milliliter was determined by plating dilutions of the bacterial cells on LB plates without antibiotics.
Persister assay for various stresses
For heat stress, E. coli cells from stationary phase cultures were treated at 52 °C for 30 minutes. The CFUs were determined after serial dilutions. For acid stress, E. coli cells were incubated with acid pH 3 for 4 days at 37°C. For hypertonic saline stress, cultures were grown in LB medium containing 3M NaCl at 37°C for 6 days without shaking. E. coli cells were also exposed to 80 mM H2O2 at 37°C for 5 days without shaking. The number of CFUs under acid, hypertonic saline, and H2O2 was determined daily.
RNA extraction and sequencing
The wild type W3110 cells and the pnp mutants were cultivated at 37°C for 6.5 hours to stationary phase. Total RNA was isolated from 1 ml culture, using the RNeasy Mini kit (Qiagen, USA), according to the manufacturer’s instructions. All procedures for RNA sequencing and alignment of the transcriptome were conducted by Oebiotech (Shanghai, China). RNA sequencing was performed using Illumina HiSeq™2000. Raw reads were filtered to remove low quality sequences, adapter sequences, and reads with poly N. The clean reads were subjected to BLAST search by tophat/bowtie2. Differential expression analysis of two samples was performed using the software DEseq. David Bioinformatics Resource 6.7 was used to perform gene ontology and KEGG pathway analysis.
Detection of bacterial internal redox status
The concentration of intracellular ATP was detected by BacTiter-Glo Microbial Cell Viability Assay Kit (Promega, USA). The intracellular NADH/NAD+ ratio was measured as described (18). Carbonyl cyanide m-chlorophenylhydrazone (CCCP), an oxidative phosphorylation inhibitor, was purchased from Sigma-Aldrich (St Louis, MO, USA). CCCP (100 μM) and antibiotics were added to LB as ATP inhibitors in stationary phase. The number of CFUs per milliliter was determined by plating serial dilutions of bacteria on LB plates without antibiotics after 24h incubation at 37°C with 210 rpm shaking.
Detection of transcriptional activity of cyclic adenosine monophosphate (cAMP) receptor protein (CRP)
The β-galactosidase gene lacZ was inserted into the polycloning site of pET-28 vector, and the original T7 promoter was replaced with the promoter region of the crp gene or promoter region of the crp gene plus 5’-UTR region. The recombinant plasmid constructs pET-lacZ-Pcrp/pET-lacZ-Pcrp+5U were transformed into E. coli W3110 and the pnp mutant (Δpnp W3110) by electroporation. The activity of β-galactosidase was measured by the β-Gal assay kit (Invitrogen, USA) according to the manufacturer’s instructions. Enhanced BCA protein assay kit (Beyotime Biotechnology, China) was used to determine the concentration of proteins.
RESULTS
pnp mutant has defect in persistence to various antibiotics
The stationary phase cultures of the pnp mutants and the wild-type strain W3110 as a control were exposed to various antibiotics, including ampicillin (200 μg/ml), norfloxacin (8 μg/ml), and gentamicin (40 μg/ml). The results showed that the Δpnp mutant was more susceptible than the parent strain W3110 to all the three antibiotics. Especially upon treatment with gentamicin, the persister levels of the Δpnp mutant decreased significantly after 24h exposure. The effect of ampicillin and norfloxacin on sterilization was similar. The Δpnp mutant was killed by ampicillin and norfloxacin significantly from the second day. Complementation of the Δpnp mutant with the functional pnp gene conferred increased persistence to the three antibiotics to the wild-type levels. However, there was no significant change in the persister level between the pnp gene overexpression strain and wild-type strain W3110 (see Figure 1).
The pnp mutant is more susceptible to various stresses
Since the persister bacteria are not only tolerant to antibiotics, they can be tolerant to certain stress conditions (3, 19). Thus, we also tested the survival of the Δpnp mutant under several stress conditions, including heat, acid pH, hydrogen peroxide, and hypertonic saline stress. As shown in Figure 2, the stationary phase culture of the Δpnp mutant was more sensitive to the four stress conditions, than that of wild-type strain W3110. Complementation of the Δpnp mutant with the functional pnp gene mostly restored the level of persistence of the wild-type strain. It is worth noting that the persistence level was significantly higher in the pnp overexpression strains than the wild-type strains under heat stress, but significantly lower in the pnp overexpression strains than the wild-type strains under hypertonic saline stress.
RNA-seq analysis reveals a higher metabolic status of the pnp deletion mutant strain
Figure 3 shows the difference in gene expression profiling between the W3110 wild type strain and the Δpnp mutant. Altogether, 242 genes showed significant differences in the Δpnp mutant compared with the parent strain W3110, where 166 genes were up-regulated and 76 genes down-regulated.
Pathway analysis of the differentially expressed genes was performed using the KEGG database, and the significance of the differentially expressed gene enrichment in each pathway was calculated by the hypergeometric distribution test. The top 20 pathways after KEGG analysis are shown in Figure 4. Flagellar assembly (PATH: ecj02040), Ribosome (PATH: ecj03010), ABC transporters (PATH: 02010), Sulfur metabolism (PATH: ecj00920), Two-component system (PATH: ecj02020), and carbon metabolism (PATH: ecj01200) are the main enrichment pathways of significant differences in gene transcripts, in which only the genes in ribosome pathway are down-regulated. We selected 41 genes whose mRNA levels increased more than 3 fold in the pnp mutant for validation by Real-time quantitative PCR. The results showed that the expression levels of 32 genes increased more than 4-fold (Table 2).
Detection of internal redox status of bacteria
Since the RNA-seq results showed that genes involved in metabolic and virulence-related pathways were expressed at a higher level in the Δpnp mutant than the parent strain, we measured the intracellular ATP levels and NADH/NAD+ ratios to estimate whether the Δpnp mutant was in a state of higher metabolism than the parent strain W3110. Since the persister assay was performed with stationary phase bacteria, we determined the ATP level in Δpnp mutant and the wild-type strain during the stationary phase. As shown in Figure 5, the ATP level of Δpnp mutant was higher than that of the wild type strain W3110 in stationary phase, suggesting that the higher metabolic status in the Δpnp mutant produces excessive ATP and renders the Δpnp mutant less able to form persisters and thus become more susceptible to antibiotics and stresses. Consistent with the above observation, the NADH/NAD+ ratio which reflects the redox status of the microbial cells in the Δpnp mutant was higher than that in the control parent strain in stationary phase, suggesting that the Δpnp mutant was indeed in a high metabolic state (Figure 5).
Antibiotic exposure assays of metabolism related gene knockout strains
Since the RNA-seq analysis indicated that the impact on bacterial metabolism after pnp gene deletion is extensive, this suggests that maintenance of the normal metabolism of bacteria by PNPase may be mediated via global regulators. To address this possiblity, we performed real-time quantitative PCR analysis of several major regulators (ArcA, ArcB, Cra, Crp, CyaA, Fnr, and RpoS) responsible for global regulation of diverse aspects of metabolism in E. coli. Significantly, we found that in the early stationary phase, expression ofthe cyaA encoding adenylate cyclase and crp encoding cAMP receptor protein (CRP) increased 3.13 and 3.22 fold, respectively. therefore, we knocked out the cyaA and crp and assessed their effect on persister levels in drug exposure experiments. In the killing curve experiments (Figure 6), persister levels of the Δcrp deletion strain and ΔcyaA deletion strain were significantly higher than the wild-type W3110 under gentamicin treatment. The persister levels of the ΔpnpΔcrp double deletion strain and ΔpnpΔcyaA deletion strain under gentamicin and norfloxacin treatment were much higher than that of the Δpnp mutant.
Effect of energy inhibitor CCCP on persister level
In order to further verify the increased sensitivity to antibiotics was caused by high level of metabolism in the bacteria, we performed ATP inhibition experiments. The results showed that addition of CCCP (100 μM) caused significantly higher tolerance to antibiotics. The results are shown in Figure 7.
Transcriptional activity of crp gene in W3110 and pnp deletion strains
Previous results suggested that PNPase is likely to maintain the downstream metabolic enzymes at the normal level by inhibiting CRP. The 5’-UTR region of prokaryotes is one of the important components of post-transcriptional regulation and can affect the initiation of mRNA translation. It has been found that C-1a PNPase in E. coli can bind to 5’-UTR region of the operon pgaABCD and inhibit the formation of biofilm by inhibiting the expression of acetylglucosamine (14). Therefore, we hypothesized that PNPase may also bind to the 5’-UTR region of crp mRNA to inhibit CRP protein translation. To address this, we made crp-lacZ reporter construct and transformed into the Δpnp mutant and the parent strain W3110. As shown in Figure 8, the results of β-galactosidase activity assay showed that the β-galactosidase activity was 8.3-fold higher in the Δpnp/PET-lacZ-Pcrp+5u than that in the parent strain W3110/PET-lacZ-Pcrp+5u at early stationary phase, and 3.6-fold higher at the end of stationary phase. The activity of β-galactosidase in W3110/PET-lacZ-Pcrp was 5.1-fold that in W3110/PET-lacZ-Pcrp+5u at early stationary phase, and 4.1-fold at the end of stationary phase. At any stage, the activity of β-galactosidase in the Δpnp/PET-lacZ-Pcrp was the highest. These findings suggest that PNPase may inhibit the expression of crp mRNA or CRP protein translation, leading to higher metabolic status of the Δpnp mutant.
DISCUSSION
In this study, we identified a new mechanism of persistence mediated by PNPase, part of the RNA degradasome. PNPase catalyzes the polymerization of nucleoside diphosphate and the phosphorylation of polynucleotides in vitro. But in vivo, PNPase is not only the main component of RNA degradasome, which catalyzes decomposition of RNA and promote metabolism and stability of mRNA, but also is independently involved in regulation of bacterial pathogenicity. For example, in Salmonella enterica, PNPase acts as a global regulator of virulence gene expression (20).
We found that, compared to wild-type, the Δpnp deletion strain was highly susceptible to three different antibiotics and stresses, indicating that pnp plays an important role in the formation or maintenance of persisters. In order to confirm that PNPase was the direct cause of persister formation, we constructed pnp complementation strain and found that the persistence phenotype was restored in the pnp complementation strain. These in vitro experiments have shown that PNPase participates in persister formation or survival. It is worth noting that phenotype of the Δpnp deletion strain mainly appeared in early stationary phase and mid-stationary phase. When the cultures were incubated to late stationary phase, these differences became less obvious, which indicates that the role of PNPase in maintaining the persistence state likely occurs in the early stage.
To shed new insight on the mechanism of by which PNPase mediates persistence, we performed RNA-seq analysis of the Δpnp deletion strain compared with the wild type W3110 strain, and found a number of genes belonging to metabolism or virulence pathway are up-regulated in the Δpnp deletion strain. These finding suggest that elevated metabolism in the Δpnp deletion strain is likely present. This is confirmed by measurement of the ATP levels and NADH/NAD+, two indicators of redox state in bacteria, and indicate that the Δpnp deletion strain had significantly higher metabolism than that of the wild type W3110 strain. The Escherichia coli CRP is an important transcription factor that its DNA-binding can result in positive or negative regulation of more than 100 genes mainly involved in catabolism of carbon sources other than glucose. In our study, the results of validation of differentially expressed genes showed that the mRNA levels of genes encoding carbohydrate transport systems (mglBAC, lamB, malK), TCA cycle (sdhA, sdhD, fumA, mdh, and acnB), Glycerolipid metabolism (glpD, glpK, glpT), etc., were significantly up-regulated. Interestingly, the metabolic master regulator CRP has a positive regulatory effect on these genes (21, 22). In addition, the cAMP-CRP complex is known to be involved in regulation of biofilm formation, quorum sensing systems, and the transcription of the nitrogen regulatory system (23–26). When the culture environment lacks glucose, CRP binds to the effector cAMP, and the activated CRP binds to the TGTGAnnnnnnTCACA sequence near the promoter of the regulatory gene, thereby recruiting RNA polymerase to initiate the transcription of the downstream gene (27). CRP was found to play an important role in the metabolic process in E. coli, and the perturbation of cAMP-CRP can significantly diminish the metabolic capabilities of persisters, and can prevent drugs like aminoglycosides from getting into the bacterial cells, thus weakening the ability of such drugs to kill persisters (28).
We knocked out crp and cyaA gene in Δpnp deletion strain, respectively, and found that the ability of persister formation was significantly higher in the double ΔcrpΔpnp deletion strain and ΔcyaAΔpnp deletion strain than that in the Δpnp deletion strain. PNPase can bind to its own 5’-UTR region to realize the autologous expression regulation by promoting degradation of pnp mRNA in the presence of RNase III (29–31). We inferred that PNPase in E. coli can inhibit the translation of the crp mRNA, which controls the CRP level in the normal range, to maintain the balance of intracellular metabolism. To confirm this inference, we measured the activity of β-galactosidase to reflect the transcriptional activity of crp in the wild type W3110 strain and the Δpnp deletion strain. We found that the transcriptional activity of crp in the Δpnp deletion strain had 8.3-fold increase in the early stationary phase, indicating that PNPase can inhibit the expression of CRP protein at the transcriptional level.
The metabolic status is the critical basis in the formation of persisters in bacteria (32–34). Although some researchers believe that dormancy is not a necessary condition for the formation of persisters (35), the idea that “persistence can be considered a termination of a metabolic procedure, that is bacteria enters a dormant state” remains the mainstream thinking. Based on our findings, we propose that PNPase in E. coli can bind to the 5’-UTR of the crp mRNA to cause a negative regulation and maintain a constant level of CRP protein, and thatdeletion of PNPase could cause overexpression of CRP, the downstream regulation of a series of metabolic-related gene expression levels increased abnormally, leading to failure to enter the dormant state, causing decreased persistence capacity in the Δpnp deletion strain and its inability to tolerate antibiotics and stress conditions.
In summary, our study established a new mechanism of persister formation mediated by PNPase regulation in E. coli (Figure 9). We found that PNPase controls cellular metabolism by negatively regulating the global regulator cyclic AMP receptor protein (CRP) operon at post-transcriptional level through targeting the 5’-UTR region of the crp transcript to regulate persister formation. The results of the in vitro studies need to be further tested in animal models to determine if PNPase has any role in persistence and virulence in vivo in the future. PNPase, as a regulator of persistence and virulence, is a promising new drug target for developing improved treatment of persistent bacterial infections.
ACKNOWLEDGEMENTS
This work was supported in part by the National Natural Science Foundation of China (81572046, 81772231).