Abstract
Polyribonucleotide phosphorylase (PNPase) is a phosphorolytic RNA exonuclease highly conserved throughout evolution. In Escherichia coli, PNPase controls complex phenotypic traits like biofilm formation and growth at low temperature. In human cells, PNPase is located in mitochondria, where it is implicated in the RNA import from the cytoplasm, the mitochondrial RNA degradation and the processing of R-loops, namely stable RNA–DNA hybrids displacing a DNA strand. In this work, we show that the human PNPase (hPNPase) expressed in E. coli causes oxidative stress, SOS response activation and R-loops accumulation. Hundreds of E. coli RNAs are stabilized in presence of hPNPase, whereas only few transcripts are destabilized. Moreover, phenotypic traits typical of E. coli strains lacking PNPase are strengthened in presence of the human enzyme. We discuss the hypothesis that hPNPase expressed in E. coli may bind, but not degrade, the RNA, in agreement with previous in vitro data showing that phosphate concentrations in the range of those found in the bacterial cytoplasm and, more relevant, in the mitochondria, inhibit its activity.
Similar content being viewed by others
Introduction
Polyribonucleotide phosphorylase (PNPase) is an evolutionarily conserved phosphorolytic exonuclease that degrades the RNA from 3′- to 5′-end in presence of phosphate (Pi) and Mg2+, producing NDPs. In vitro, it also catalyses the reverse template-independent NDP polymerization reaction leading to de novo synthesis of RNA molecules or 3′-end tailing of pre-existing RNAs1,2,3. Escherichia coli PNPase (hereafter EcPNPase) in vitro activity on RNA and its cellular function in RNA turnover and in the control of complex phenotypic traits, like biofilm formation and adaptation to the growth at sub-optimal temperature, have been extensively studied by different groups over the many years since bacterial PNPase discovery1,4,5. Global mRNA profiling of E. coli mutants lacking EcPNPase by RNA-Seq showed that the enzyme controls—directly or indirectly—the level of around 200 transcripts, with the 59% of RNAs increased and the others decreased in EcPNPase absence6. In a previous genome-wide analysis of E. coli transcription profile performed by microarray, an even higher number of RNAs whose level was modulated by EcPNPase was found (i.e. more than 500)7. Auto-regulation of the EcPNPase pnp gene is considered a paradigmatic example of post-transcriptional regulation. EcPNPase promotes pnp mRNA degradation by digesting a short non-coding RNA that protects it from RNase E-dependent endonucleolytic cleavage8. EcPNPase regulates its own expression also via other pathways not requiring the enzyme phosphorolytic activity, namely by binding the pnp mRNA and eliciting premature transcription termination and probably, translational repression9,10,11,12. Far less understood aspects related to EcPNPase are how (and if) small molecules reported to bind and modulate EcPNPase in vitro impact the enzyme activity in the bacterial cells13,14,15,16, and the EcPNPase role in DNA repair and oxidative stress response. In vitro, the enzyme degrades ssDNA molecules in presence of Pi and Mn2+ producing dNDPs. It can also catalyse template-independent polymerization of dNDPs and 3′-end tailing of DNA molecules17,18. In vivo, EcPNPase enhances homologous recombination, and pnp mutants are sensitive to UV radiation and defective in DNA repair of double-stranded breaks17,19,20. The mechanistic function of EcPNPase in DNA repair is unknown. It is still unclear also its role in the protection against oxidative stress, which was imputed to EcPNPase-dependent removal of oxidized RNA, for which such enzyme has high affinity21.
The studies on the human orthologous protein (hereafter hPNPase) began much more recently, when the human PNPase (hPNPase) cDNA was cloned and the protein shown to localize in mitochondria22,23. hPNPase is essential in mice and human cells, and severe pathological conditions with a wide spectrum of symptoms are caused by mutations in the hPNPase PNPT1 gene24,25,26,27. RNA import from cytoplasm into mitochondria and degradation of mitochondrial RNAs are the main activities credited to hPNPase fractions located in the intermembrane space and in the matrix, respectively28. Interestingly, also EcPNPase can switch from an RNA degrading to an RNA carrier (and stabilizing) mode depending on the RNA substrate and the interaction with other proteins like the DEAD-box RNA helicase RhlB, which can unfold RNA secondary structures inhibiting EcPNPase activity6,7,29,30. Similarly, in the mitochondrial matrix hPNPase forms a complex with the DExH-box RNA helicase SUV3 able to efficiently degrade dsRNA31,32,33. The complex is involved in mitochondrial DNA homeostasis, a function that has been linked to its ability of removing R-loops, i.e. stable RNA–DNA hybrids displacing a DNA strand and potentially interfering with DNA replication34. Whether the human enzyme can bind and degrade DNA like its bacterial counterpart is unknown. Conversely, it is known that hPNPase can bind, but not degrade, oxidized RNA35,36. hPNPase is 40% identical to EcPNPase (Supplementary Fig. S1), and orthologues of the main EcPNPase interactors, namely RhlB and RNase E1, are absent in human cells. Human and E. coli PNPases differ as for their optimal Pi concentration for in vitro RNA phosphorolysis, which is around a 100-fold higher for the bacterial than the human enzyme (i.e. 10 vs. 0.1 mM). Indeed, hPNPase phosphorolytic activity was reported to be inhibited in vitro at phosphate concentrations in the range of those found in the E. coli cytoplasm37,38,39,40.
In this work, to get insight into hPNPase activity in a cellular context simpler than that of the human cell and containing high Pi38, we analysed the phenotype and global transcription profile of an E. coli strain expressing the human enzyme instead of the natural orthologue. We found that hPNPase cannot complement, and actually worsens, several phenotypes due to the lack of EcPNPase. hPNPase deeply impacts E. coli gene expression and triggers the SOS response, most likely by eliciting oxidative stress and accumulation of R-loops.
Results
Expression in E. coli of the PNPT1 Ec gene
We constructed by λ RED recombination41 an E. coli strain expressing human PNPase instead of the natural EcPNPase from the pnp chromosomal locus (strain C-6001, carrying the mutation Δpnp::PNPT1Ec; Table 1 and Fig. 1A). The recombinant strain was designed so that transcription of the Δpnp::PNPT1Ec region (hereafter PNPT1Ec) from the pnp promoter produced an mRNA with the same 5′-UTR as the pnp transcript and the hPNPase open reading frame (ORF) with the E. coli codon usage15.
We found that PNPT1Ec cultures had longer generation time (g) in exponential phase and reached lower optical density after 24 h than both pnp+ and Δpnp cultures (Table 2 and Fig. 1B), showing that the presence of hPNPase affected growth more than the simple lack of EcPNPase.
PNPT1Ec gene expression analysis by Northern blotting and RT-PCR showed that the gene was transcribed during exponential growth at 37 °C (Fig. 2A,B). The main transcripts were a ca. 2.2 kb long mRNA, potentially covering the whole PNPT1Ec gene, and a shorter RNA, less than 0.5 kb long. Overall, the PNPT1Ec transcription profile was similar to that of the pnp locus in the pnp+ strain, although the short RNA was more abundant in PNPT1Ec than in the pnp+ (Fig. 2A). Western blotting with a hPNPase-specific antibody showed a single signal in PNPT1Ec corresponding to a protein migrating between the 70 and the 100 kDa molecular weight (MW) markers, compatible with the predicted MW of hPNPase lacking the mitochondrial localization signal (i.e. 81 kDa; Fig. 3A). The half-life of the protein at 37 °C was 1.3 ± 0.5 h (Fig. 3B), indicating that like EcPNPase44, the human protein is relatively stable in E. coli exponential cultures.
To compare the level of hPNPase in the PNPT1Ec strain with that of EcPNPase in the pnp+, we constructed two strains expressing histidine-tagged variants of either EcPNPase or hPNPase (C-6009 and C-6011, respectively; Fig. 1A) from the pnp locus. As shown in Fig. 3C, the expression of His-tagged EcPNPase was around 30-fold higher than that of His-tagged hPNPase. However, the untagged hPNPase in PNPT1Ec was five-fold more abundant than its His-tagged version, whereas the tagged and untagged variants of EcPNPase were equally expressed (Fig. 3C). Based on these data, we evaluated that in mid-exponential cultures at 37 °C, hPNPase should be around sixfold less abundant in the PNPT1Ec than EcPNPase in the pnp+ strain. Since the amount of EcPNPase has been estimated to be ca. 5700 molecules/cell45, hPNPase should be at 900–1000 molecules/cell, more abundant than the median E. coli protein (i. e. 526 molecules/cell)45. We did not investigate the mechanism determining low abundance of the tagged hPNPase, which could be due to decreased mRNA or protein stability.
Phenotypes due to the lack of E. coli PNPase are strengthened in presence of hPNPase
We tested whether hPNPase could compensate the lack of EcPNPase and complement phenotypes caused by pnp null mutations in E. coli. In particular, we assayed (i) aggregation of cultures growing in minimal medium with glucose as sole carbon source, detectable as a drop in the optical density of late exponential cultures46. This phenotype was tested also at low Pi concentration, which should increase and decrease, respectively, hPNPase and EcPNPase phosphorolytic activity28; (ii) grow inhibition at temperatures below 18–20°C11,47; (iii) increased sensitivity to H2O2, causing oxidative damage, and to mutagenic agents like UV rays and zeocin, an antibiotic inducing DNA double-strand breaks (DSBs)17,19,20,21; and iv) reduced frequency of spontaneous mutations19,20. As shown in Fig. 4A, the turbidity of PNPT1Ec cultures growing in minimal medium started decreasing when the cultures were at OD600 = 0.2–0.3, whereas the Δpnp strain aggregated at higher cell density. Interestingly, we observed that low Pi concentration promoted the aggregation of the pnp+ strain, whereas it prevented PNPT1Ec aggregation and had a modest effect, if any, on the Δpnp strain (Fig. 4B).
The PNPT1Ec strain was more sensitive to H2O2, UV rays and zeocin than the Δpnp (Fig. 4C) and cold sensitive like the Δpnp (Fig. 4D). Moreover, its spontaneous mutation frequency was even lower than that of the Δpnp (Fig. 4E). Cold sensitivity and increased adhesion and aggregation were exhibited also by a Δpnp mutant overexpressing hPNPase from a multicopy plasmid (Supplemental Fig. S2), implying that lack of compensation of these Δpnp-linked phenotypes by hPNPase was not due to the lower expression of the PNPT1Ec gene with respect to pnp.
hPNPase deeply affects the E. coli transcriptional profile
To assess the impact of hPNPase on E. coli physiology, we analysed by RNA-Seq the global transcription profile of the PNPT1Ec strain in comparison with that of the pnp+ and the Δpnp. RNA was extracted from exponential cultures grown in broth at 37 °C and the transcriptomes of the three strains were compared to identify differentially expressed genes (DEGs) with FDR < 0.01 and |log2(FoldChange)|≥ 1 as thresholds (Fig. 5A; Supplemental Table S1). DEGs found in each pairwise comparison are listed in Supplemental Table S1. 307 and 202 genes were down- and up-regulated in the Δpnp mutant versus the pnp+. These genes are directly or indirectly regulated by PNPase. 421 and 331 genes were down- or up-regulated, respectively, in PNPT1Ec versus pnp+, showing that replacing the natural pnp gene with the human ortholog alters the transcriptional profile more than deleting the bacterial gene. Accordingly, 237 genes had different expression in the Δpnp with respect to the PNPT1Ec strain, with 135 and 102 down- and up-regulated genes, respectively, in PNPT1Ec. The genes with the largest decrease in expression levels in presence of hPNPase (log2(FoldChange) < − 3) were genes belonging to tna and tdc operons controlling tryptophan and threonine degradation, and the glpABC genes encoding the anaerobic glycerol-3-phosphate dehydrogenase48,49. Most of the genes with highest relative expression in PNPT1Ec (log2(FoldChange) > 3), like tisB, umuD, sulA, dinI, recN and recA, belonged to the SOS response, a regulatory mechanism that allows the coordinated induction of tens of genes, mainly involved in DNA repair, in presence of DNA damage50.
To better understand how hPNPase impacts the bacterial cell physiology, we looked for regulons moving in the pairwise comparisons between the PNPT1Ec and the Δpnp strains and maintaining similar expression in presence or absence of the EcPNPase (i.e. in the pnp+ vs. Δpnp comparison) (Fig. 5B). These regulons should respond to the human, and not to the bacterial, enzyme. We found that the LexA and CpxR regulons were activated in presence of the human enzyme, as confirmed also by Northern blotting showing enhanced expression of recA and cpxQ genes in PNPT1Ec (Fig. 5B,C). LexA is the repressor of the SOS response50, whereas CpxR is the regulator of the Cpx pathway mediating adaptation to envelope stress at the inner membrane52.
Overall, these data indicate that hPNPase has differential activity with respect to the bacterial enzyme and that may cause genotoxic and membrane stress.
hPNPase causes SOS response induction
The SOS response is induced by LexA repressor self-cleavage stimulated by activated RecA (RecA*), which is a RecA filament assembled on a ssDNA scaffold50. Constitutive induction of the SOS response in PNPT1Ec indicates that this strain may have high levels of ssDNA, which typically accumulates when cells attempt to replicate damaged DNA.
The SOS response can be induced in E. coli by Reactive oxygen species (ROS), which are the major source of endogenous DNA damage53,54. Considering that hPNPase induces ROS when overexpressed in human cells55, we speculated that hPNPase may either induce ROS or interfere with ROS detoxification. In fact, ROS levels measured using the fluorescent probe DCFH-DA resulted significantly higher in PNPT1Ec than in pnp+ (Fig. 6A), and the soxS gene, a transcriptional activator of the oxidative stress genes54, was induced in presence of hPNPase (Fig. 5C; Supplemental Fig. S3A,B). To check whether oxidative damage could be the cause of constitutive SOS induction, we tested recA gene expression, taken as a measure of SOS response induction, in anaerobic cultures. In these conditions, the extent of oxidative damage should be reduced54. As shown in Fig. 6B, recA expression was diminished in anaerobiosis, although it remained higher in the PNPT1Ec than in the other strains. Thus, oxidative damage may contribute to recA induction and SOS response activation but it is not the only cause of this phenomenon.
hPNPase causes the accumulation of R-loops
hPNPase was reported to participate in R-loop removal from mitochondrial genome together with the SUV3 RNA helicase34. Since the SOS response is induced in mutants lacking RNase HI, which processes R-loops, and ssDNA, which is the SOS response inducing stimulus50, is exposed in R-loops56,57, we tested whether the R-loop level was enhanced in presence of hPNPase. Dot blot analysis of genomic DNA with the S9.6 antibody, which primarily recognizes DNA-RNA hybrids though having some affinity also for dsRNA58, showed sixfold stronger signals with the PNPT1Ec DNA than with the DNA of the other pnp+ and Δpnp strains (Fig. 6C). Digestion of genomic DNA with the DNA-RNA hybrid-specific nuclease RNase H before the dot blot analysis eliminated the difference in signal intensity among the genomic DNAs extracted from the three strains, difference which was instead maintained after DNA treatment with the dsRNA-specific RNase III (Fig. 6C). Thus, hPNPase stimulates R-loops in E. coli.
In mammals, DNA damage and replication slowing induced by ROS causes R-loop formation59,60,61. To assess whether oxidative damage could be involved in hPNPase-dependent R-loop accumulation, we measured R-loops in bacteria grown anaerobically. We found that also in these condidtions, the amount of R-loops was strongly increased in presence of hPNPase (Fig. 6D), making unlikely that R-loops accumulate in presence of hPNPase only as a consequence of oxidative damage.
Another mechanism that could cause R-loop increase may be low expression of the R-loop processing enzyme RNase HI in presence of hPNPase. However, the rnhA gene, encoding RNase HI, was not differentially expressed in the PNPT1Ec (Supplemental Table S1). Consistent with it, the PNPT1Ec growth was not impaired in anaerobiosis (Fig. 6E), a condition in which rnhA mutants cannot grow57.
hPNPase is essentially devoid of RNA degradation activity in E. coli
To assess whether the human enzyme had catalytic activity in the bacterial context, we measured the overall stability of E. coli RNAs in presence or absence of the human enzyme. This was done by analysing by RNA-Seq transcript levels (measured in counts per million, CPM) of the PNPT1Ec, pnp+ and Δpnp strains immediately before (T0) and 4 min after blocking transcription initiation with rifampicin (T4). The residual RNA amounts of the different E. coli RNA species at T4 should depend on the stability of each of them in the three strains. Stability was thus estimated for each gene according to the percentage of its RNA found remaining at T4, expressed as the percentage of the CPM values at T0 maintained at T4 (Supplementary Fig. S3C). For simplicity, genes were classified in four categories from stable RNAs to unstable RNAs, as follows:
-
1.
RNA remaining > 75%;
-
2.
50% < RNA remaining < 75%;
-
3.
25% < RNA remaining < 50%;
-
4.
RNA remaining < 25%.
As shown in Fig. 7 and Supplemental Table S2, hPNPase caused a drastic redistribution of the mRNA levels among the categories, with a strong decrease of the most unstable mRNAs (i.e. those in cat. 4). In particular, only 27 RNAs where less stable in presence of hPNPase than in Δpnp, whereas 521 RNAs were less stable in presence of EcPNPase than in Δpnp (Fig. 7, right panel, pnp+ vs. Δpnp). Conversely, the transcripts more stable in the presence of hPNPase were more than double those stabilized in the presence of EcPNPase (1738 and 786 in PNPT1Ec and pnp+, respectively, in comparison with Δpnp). These data are consistent with a very limited, if any, RNA degrading activity of hPNPase in E. coli cells.
Discussion
In this work, we characterized an E. coli strain that carries the PNPT1Ec gene encoding the human PNPase inserted into the pnp locus. Our data clearly show that human PNPase cannot compensate the phenotypic effects due to the lack of the natural EcPNPase. On the contrary, the PNPase of spinach chloroplast complemented the cold sensitive phenotype of an E. coli pnp mutant62. This difference may be due to higher similarity of chloroplast PNPase to EcPNPase, as both the RPH2 and the KH-S1 domains of EcPNPase are more similar to those of spinach PNPase than to hPNPase ones. Indeed, only spinach PNPase, and not hPNPase, can form heterotrimers with EcPNPase subunits37,62. Since residues 327–331 of EcPNPase, with which RNase E forms hydrogen-bonding interactions63, are not conserved in the human enzyme, it is very unlikely that hPNPase could be recruited in the RNase E-based RNA degradosome. Thus, it seems sound that hPNPase cannot replace the endogenous enzyme in all those activities in which EcPNPase participates as a component of the multisubunit RNA degradosome.
The comparison of total RNA stability between the Δpnp and the PNPT1Ec strains showed that the bacterial enzyme caused the destabilization of more than 500 transcripts, whereas only 27 were modestly destabilized in PNPT1Ec, suggesting that the human enzyme may be substantially devoid of degradative activity in the bacterial context. Previous in vitro analyses showed that the phosphorolytic activity of hPNPase is strongly impaired when Pi concentration is greater than 2.5 mM37. Since in the cytoplasm of E. coli growing cells Pi concentration is around 10 mM38, this would explain why hPNPase has no degradation activity in the bacterial cell. Consistent with this hypothesis, at low Pi, which should improve hPNPase phosphorolytic activity and worsen that of EcPNPase, the strong aggregative phenotype shown in M9 by PNPT1Ec is suppressed, whereas the pnp+ strain aggregates. Although these data should be taken with some caution, giving the pleiotropic effects that Pi concentration modulation may have, they support the hypothesis that hPNPase may be inactive in E. coli because of high Pi level. Interestingly, Pi should be at about 10 mM in the mammalian cell cytoplasm64,65 and, as a consequence, also in the mitochondrial intermembrane space, as this compartment has similar composition to the cytosol with respect to ions and small molecules to which the mitochondrial outer membrane is highly permeable66,67. Indeed, phosphate concentration in whole mitochondria has been estimated to be greater than 2.5 mM64,68,69, implying that hPNPase phosphorolytic activity should be essentially inhibited or limited to sub-mitochondrial regions in which Pi concentration is locally and/or temporarily low. In agreement with this hypothesis, mitochondrial RNA (mtRNA) degradation by the hPNPase-hSUV3 complex, namely the mitochondrial degradosome, has been reported to take place in relatively few spots inside the mitochondrial matrix32.
Lack of catalytic activity and impaired interaction with other components of the bacterial RNA degradation machinery explain why hPNPase cannot compensate for the lack of EcPNPase. However, the PNPT1Ec strain does not simply recapitulate the phenotype of strains with loss of function mutations in the pnp gene. Phenotypic traits due to EcPNPase lack, like the fitness decrease, the aggregation in minimal medium, the sensitivity to agents causing—directly or indirectly—DNA damage like UV, zeocin, and H2O2, the low frequency of spontaneous mutations and the ROS accumulation are all exacerbated by hPNPase, and more than 200 genes are differentially expressed between the PNPT1Ec and the Δpnp strains. These findings are consistent with hPNPase retaining some activity in E. coli cells, and we speculate that it may retain RNA binding activity, which is not necessarily linked to catalysis28. Binding by hPNPase could interfere with RNA function and RNA degradation pathways active in the absence of EcPNPase worsening the effect of the Δpnp mutation. A limitation of this work is that we have not assayed the effect of a PNPT1Ec mutation eliminating the RNA binding domains of hPNPase. This experiment, which we plan to do, will allow to test our hypothesis that at least some PNPT1Ec phenotypic traits may be due to ectopic RNA binding.
Even assuming that indirect effects linked to the PNPT1Ec slow growth play a major role in RNA stabilization and only a fraction of the RNAs responsive to hPNPase are actually direct targets of the enzyme, the effect of hPNPase on overall RNA stability remains striking, with many RNAs (i.e. about the 42%) more stable in presence of hPNPase in PNPT1Ec than in Δpnp mutant. Paradoxical RNA stabilizing effect exerted by E. coli exonucleases, among which EcPNPase, was already reported1,6,7,70 and it has been confirmed by our data, as around 700 RNAs (i.e. about the 19%) were more stable in the presence of EcPNPase in the pnp+ than in its absence in the Δpnp strain.
Intriguingly, overexpressed hPNPase promotes ROS in Hela cells through an unknown mechanism55. Our data suggest that hPNPase may exert the same effect in E. coli, making unlikely that hPNPase-dependent ROS induction may require the interaction with specific human proteins or RNAs.
In the PNPT1Ec strain, the SOS response is constitutively active. Since the SOS response is elicited by RecA*71, i.e. a filament of RecA protein assembled on ssDNA, this indicates that hPNPase stimulates ssDNA accumulation. A relevant source of ssDNA in the PNPT1Ec are most likely the R-loops that are abundant in such strain. Accordingly, some phenotypic traits of the PNPT1Ec, and in particular SOS induction and slower growth in aerobiosis, are also shown by rnhA mutants lacking RNase HI56,57. In principle, hPNPase may indirectly affect R-loop level by down-regulating the expression of the RNase HI rnhA gene or of recG and topA genes encoding RecG and topoisomerase I, respectively, which have also been implicated in R-loop homeostasis72. However, transcriptomic data show that rnhA, recG and topA genes are all equally expressed in presence or absence of hPNPase, ruling out their down-regulation by hPNPase, at least at the level of mRNA synthesis and/or degradation. Moreover, rnhA mutants do not grow in anaerobiosis57, whereas the PNPT1Ec does, suggesting that the presence of hPNPase does not abolish RNase HI activity. hPNPase may indirectly promote R-loops because it causes ROS accumulation, which is known to induce R-loops in human cells34,60,61,73,74,75. However, R-loops are abundant also in anaerobiosis, suggesting that other mechanisms can contribute to their formation. Factors binding the nascent RNA co-transcriptionally, like Rho, have been shown to play a pivotal role in avoiding excessive R-loop formation76,77,78. hPNPase ectopic binding of RNAs emerging from transcriptional bubbles may interfere with the anti-R-loop activity of such factors. On the contrary, in human cells, hPNPase loading onto the nascent transcript would prevent the formation of R-loops or destabilize them because hPNPase would bring to the transcript the SUV3 helicase, which is able to denature RNA–DNA hybrids33,34,79.
Despite the comparable level of the PNPT1Ec and pnp long mRNAs in the PNPT1Ec and pnp+ strains, respectively, hPNPase was less abundant than EcPNPase. EcPNPase negatively regulates its own expression through complex post-transcriptional mechanisms modulating pnp mRNA translation and stability, which mainly occur at the level of the mRNA 5′-UTR1. Thus, while it is not surprising that replacing EcPNPase with hPNPase impacts autoregulation, given the dissimilar activity of the two proteins, we would have expected mRNA stabilization and increased translation in presence of hPNPase, as shown by strains expressing mutated EcPNPases8,12,39,80,81,82. It is possible that despite codon optimization for E. coli, translation elongation of the PNPT1Ec mRNA is not efficient and impairs hPNPase expression even though autoregulation is not stringent. The abundance of the short PNPT1Ec RNA in the PNPT1Ec (Fig. 2A), which could derive from either premature transcription termination and/or degradation of untranslated transcripts83,84, is consistent with this hypothesis.
Methods
Bacterial strains and plasmids
E. coli strains and plasmids used in this work are listed in Table 1. Coordinates of E. coli genes refer to MG1655 reference genome (GenBank accession number U00096.3) unless differently stated. C-6001 was obtained by replacing in C-1a42 the pnp ORF (region 3309033–3311168) with a PNPT1Ec:kanR cassette by λRED recombination41. PNPT1Ec:kanR was constructed by assembling amplicons deriving from PCR amplification with ad hoc primers of PNPT1Ec and kanR cassette on pET28b-H6-hPNPase and pKD13, respectively14,41. PNPT1Ec encodes hPNPase ORF with the E. coli codon usage and without the first 45 amino acids corresponding to the mitochondrial localization signal. All other strains were similarly constructed by assembling in vitro, by PCR amplification, recombinant cassettes and inserting them into the E. coli chromosome by λRED recombination. FLP recombination was exploited to remove antibiotic resistance genes in all strains41. The PNPases expressed by C-6009 and C-6011 have 6 residues long His-tag at the C-terminus. For pETD-hPNP construction, PNPT1Ec was amplified by PCR on pET28b-H6-hPNPase with proper primers and cloned in pETDuet-1 digested with NcoI-EcoRI, by NEBuilder Hi-Fi DNA Assembly (New England Biolabs).
Bacterial cultures were grown in LD broth, M9 supp, i.e. M9 minimal medium with 0.4% (w/v) glucose and 2.5% LD85 or M9/2.5 i.e. M9 with a ca. 40-fold reduction of the phosphate content (from 57 to 1.3 mM) and 0.4% glucose. When needed, media were supplemented with 100 µg/ml ampicillin, 50 µg/ml kanamycin, 0.1 mM IPTG. Growth at low oxygen in liquid cultures was performed by inoculating bacteria in 15 ml conical Falcon tubes filled to the brim with LD medium with screwed caps and further sealed with parafilm. Incubation was performed static at 37 °C. In this condition, given low solubility of oxygen in culture medium and poor diffusion, bacteria below about 1 mm grow anaerobically86. Growth in anaerobiosis in solid media was performed by incubating the plates at 37 °C in an airtight jar in presence of an Aerogen (Oxoid) sachet.
Western blotting and hPNPase protein stability
Proteins were extracted by sonication from mid-exponential cultures grown in LD at 37 °C as described87. For hPNPase stability determination, overnight cultures were diluted to OD600 = 0.02 in LD broth and grown at 37 °C, aerated. At OD600 = 0.6, 25 µg/ml streptomycin was added, and 2 ml samples were collected every 20 min for 1 h. Samples were pelleted and resuspended in protein loading sample buffer (50 mM Tris–HCl pH 6.8, 2% SDS, 0.1% bromophenol blue, and 10% glycerol) to OD600 = 25. After 3 min boiling, protein samples (15 µl, corresponding to 0.375 OD600) were loaded onto 10 or 12% SDS–polyacrylamide gel, as stated in Figure legends. After the run, the gels were blotted onto PROTRAN nitrocellulose membranes (GE Healthcare) and the membranes stained with Ponceau S solution (Sigma). Immunodecoration was performed with one of the following antibodies diluted in Blotto87 as indicated in figure legends: (i) monoclonal anti-hPNPase (Anti-hPNP: PNPase (G-11) sc-365049; Santa Cruz Biotechnology); (ii) polyclonal anti-hPNPase (PA5-22397; Invitrogen); (iii) polyclonal anti-L4 or S3 ribosomal proteins (kindly provided by C. Gualerzi); (iv) polyclonal anti-EcPNPase88 (v) anti-His tag antibody (MA-1–21315; Invitrogen).
RNA extraction and analysis by Northern blot and RT-qPCR
Procedures for RNA extraction, electrophoresis on denaturing polyacrylamide and agarose gels, Northern blotting, staining of the RNA on filters with methylene blue and 5′-end labelling of oligonucleotides with [γ32P]-ATP and T4 polynucleotide kinase were previously described89,90. Oligonucleotide probes used for Northern blotting were CPXQ (4416706–4416727); RECA (1059766–1059783); SOXS (4245855–4245873); 1842 (5S-specific primer)91. RT-PCR (Reverse Transcription- PCR) was performed on RNA extracted as for the RNA-Seq. 2 μg were reverse-transcribed with Superscript III Reverse Transcriptase (Invitrogen). 1:40 cDNA dilution was used for PCR with primers specific for either the PNPT1Ec mRNA or 16S rRNA (used as reference gene87), respectively. RT-qPCR (Reverse Transcription- quantitative PCR) was performed on RNA extracted from three independent cultures for each strain. 1 μg of RNA was reverse-transcribed with Superscript III Reverse Transcriptase (Invitrogen) and 1:10 cDNA dilution was used for RT-qPCR with SYBR Premix Ex Taq (Takara) with primers specific for the PNPT1Ec mRNA. Two technical duplicates were performed for each biological replicate and reactions were run in a CFX Connect Real-Time PCR Detection System (Bio-Rad). 16S rRNA was used as reference gene to normalize the results and calculate the relative fold change in gene expression using the CFXmaestro Software (Bio-Rad).
RNA-Seq analysis of global transcript levels and stability
Global transcript levels were determined by extracting total RNA as described92 from three replicate cultures of C-1a, C-5691 and C-6001 grown in LD at 37 °C up to OD600 = 0.8. Each RNA sample was qualitatively and quantitatively checked on Agilent 2100 Bioanalyzer RNA Pico Chip (Agilent, Santa Clara, CA, USA). Directional RNA-Seq libraries were prepared from 1 µg of total RNA using the TruSeq Stranded Total RNA with Illumina Ribo-Zero Plus Sample Prep Kit (Illumina, San Diego, CA) according to the manufacturer’s protocol. Sequencing was performed on an Illumina NextSeq 500 platform (Illumina, San Diego, CA) with 12 M 2 × 75 paired end reads for each sample. For RNA stability analysis, three replicate cultures of C-1a, C-5691 and C-6001 were grown in LD at 37 °C up to OD600 = 0.8 and rifampicin (0.4 mg/ml) was added to the cultures. Samples for RNA extraction were taken immediately before (time 0) and 4 min after the addition of the antibiotic. The RNA was processed for library preparation and sequencing on Illumina NextSeq 500 platform as described before. For all datasets, sequencing reads were mapped on the E. coli K12 MG1655 RefSeq genome assembly (NCBI accession number NC_000913.3) with bowtie293 in “very sensitive” mode allowing a maximum insert size of 1000 bps. Expression quantification was performed with respect to the latest gene annotation retrieved from NCBI (GCF_000005845.2_ASM584v2). Transcript levels of each gene were defined as the number of reads overlapping its transcribed region, computed with bedtools94. Differential expression analysis was performed with edgeR95. Initial read counts were normalized by trimmed mean of M values (TMM), with default parameters. Differentially expressed genes were identified by the quasi-likelihood (QL) F-test of edgeR (glmQLFfit and glmQLFtest functions, with default parameters). We selected as differentially expressed all genes with an adjusted p-value (FDR) < 0.01 and a log fold change higher than 1 or lower than -1 (that is, showing at least a two-fold change of expression). For RNA stability analysis, raw reads counts were normalized by library size and counts per million employed in subsequent comparisons.
ROS assay
200 µl of exponential cultures (OD600 = 0.8) grown in LD at 37 °C were harvested, resuspended in 200 µl of PBS containing 1.0 mM DCFH-DA (Sigma-Aldrich) and incubated 30 min at 37 °C. The samples were then washed threefold with PBS and transferred into a 96-well plate. The OD600 and fluorescence at 488/530 (excitation/emission) nm were measured with an EnSight (PerkinElmer) plate reader. Fluorescence was normalized by the OD600 of the tested culture sample.
Detection of R-loops
Total nucleic acids extraction from 10 ml overnight cultures grown in LD at 37 °C and R-loop detection were performed essentially as described78. 10 μg samples were immobilized on Hybond-N+ nylon membrane (PerkinElmer) using a Dot-blot apparatus (Hybri-Dot Manifold). RNase treatment was performed before immobilization on the membrane by incubating the samples with 20 U of Ribonuclease H (NEB) 3 h at 37 °C or 1U of RNase III (Epicentre) for 1 h at 37 °C. In this condition, RNase III visibly degrades 8 µg of total RNA as determined in preliminary experiments (not shown). After UV crosslinking, the filters were immunodecorated with the RNA–DNA hybrid specific S9.6 antibody (EDM Millipore) diluted 1:5000 in blocking buffer (5% milk in TBST). Incubation was performed either 3 h at room temperature or overnight at 4 °C with similar results. After washing, incubation with the secondary anti-mouse IgG antibody (Thermo Fisher Scientific) was performed 1 h at room temperature. Signal detection was carried out via chemiluminescence (PDS standard ECL GeneSpin) using ChemiDoc Touch Imaging System (Bio-Rad) for membrane imaging. After the acquisition, the membrane was washed in water, soaked in 5% acetic acid for 15 min with shaking, and the nucleic acids stained with 0.05% methylene blue dissolved in 0.5 M sodium acetate buffer (pH 5.2).
Fluctuation test
A fluctuation test was performed to evaluate streptomycin resistance mutation rate (µ, i.e. the probability of mutation per cell per division). Independent cultures of C-1a, C-5691 and C-6001 (N = 12) were incubated overnight at 37 °C and titers were measured by plating proper dilutions on LD-Agar plates. The undiluted cultures were also plated on LD-Agar plates containing 0.025 mg/ml streptomycin. Mutations per culture value (m) was calculated as described96 from the median of the titer obtained in LD plate (N) and the median of the number of mutated colonies obtained on streptomycin plates (r), by resolving the equation: \(r - 1.24m - m*{\text{ln}}\left( m \right) = 0\). µ was calculated as the ratio between m and the median number of colonies obtained in permissive conditions (without streptomycin).
Statistical analysis
The replicate number for each sample are indicated in the figure legends, when appropriate. Statistical significance was determined by two-tailed Student's t-test when comparing two groups or ANOVA with post-hoc Tuckey test when comparing three or more groups. Pearson’s Chi-square test was used to compare distributions. The differences in means were considered statistically significant at P < 0.05.
Data availability
All relevant data are shown. Original images of gels and filters are shown in Supplementary Fig. S4. RNA-Seq raw data are available at GEO repository (accession n. GSE211960, token axujemsonzcpdov, and GSE221825, token ejgtwucwrxyzrkr).
References
Briani, F., Carzaniga, T. & Dehò, G. Regulation and functions of bacterial PNPase. Wiley Interdiscip. Rev. RNA 7, 241–258 (2016).
Leszczyniecka, M., DeSalle, R., Kang, D. C. & Fisher, P. B. The origin of polynucleotide phosphorylase domains. Mol. Phylogenet. Evol. 31, 123–130 (2004).
Lin-Chao, S., Chiou, N. T. & Schuster, G. The PNPase, exosome and RNA helicases as the building components of evolutionarily-conserved RNA degradation machines. J. Biomed. Sci. 14, 523–532 (2007).
Grunberg-Manago, M., Ortiz, P. J. & Ochoa, S. Enzymic synthesis of polynucleotides I. Polynucleotide phosphorylase of Azotobacter vinelandii. BBA Gen. Subj. 20, 269–285 (1956).
Mohanty, B. K. & Kushner, S. R. Enzymes involved in posttranscriptional RNA metabolism in gram-negative bacteria. Microbiol. Spectr. 6, 19–35 (2018).
Pobre, V. & Arraiano, C. M. Next generation sequencing analysis reveals that the ribonucleases RNase II, RNase R and PNPase affect bacterial motility and biofilm formation in E. coli. BMC Genomics 16, 72 (2015).
Mohanty, B. K. & Kushner, S. R. Genomic analysis in Escherichia coli demonstrates differential roles for polynucleotide phosphorylase and RNase II in mRNA abundance and decay. Mol. Microbiol. 50, 645–658 (2003).
Jarrige, A. C., Mathy, N. & Portier, C. PNPase autocontrols its expression by degrading a double-stranded structure in the pnp mRNA leader. EMBO J. 20, 6845–6855 (2001).
Robert-le Meur, M. & Portier, C. Polynucleotide phosphorylase of Escherichia coli induces the degradation of its RNase III processed messenger by preventing its translation. Nucleic Acids Res. 22, 397–403 (1994).
Carzaniga, T., Dehò, G. & Briani, F. RNase III-independent autogenous regulation of Escherichia coli polynucleotide phosphorylase via translational repression. J. Bacteriol. 197, 1931–1938 (2015).
Zangrossi, S. et al. Transcriptional and post-transcriptional control of polynucleotide phosphorylase during cold acclimation in Escherichia coli. Mol. Microbiol. 36, 1470–1480 (2000).
Marchi, P. et al. Autogenous regulation of Escherichia coli polynucleotide phosphorylase during cold acclimation by transcription termination and antitermination. Mol. Genet. Genomics 278, 75–84 (2007).
Carzaniga, T. et al. A conserved loop in polynucleotide phosphorylase (PNPase) essential for both RNA and ADP/phosphate binding. Biochimie 97, 49–59 (2014).
Stone, C. M. et al. Inhibition of homologous phosphorolytic ribonucleases by citrate may represent an evolutionarily conserved communicative link between RNA degradation and central metabolism. Nucleic Acids Res. 45, 4655–4666 (2017).
Nurmohamed, S. et al. Polynucleotide phosphorylase activity may be modulated by metabolites in Escherichia coli. J. Biol. Chem. 286, 14315–14323 (2011).
Tuckerman, J. R., Gonzalez, G. & Gilles-Gonzalez, M. A. Cyclic di-GMP activation of polynucleotide phosphorylase signal-dependent RNA processing. J. Mol. Biol. 407, 633–639 (2011).
Cardenas, P. P. et al. Polynucleotide phosphorylase exonuclease and polymerase activities on single-stranded DNA ends are modulated by RecN, SsbA and RecA proteins. Nucleic Acids Res. 39, 9250–9261 (2011).
Kaufmann, G. & Littauer, U. Z. Deoxyadenosine diphosphate as substrate for polynucleotide phosphorylase from Escherichia coli. FEBS Lett. 4, 79–83 (1969).
Carzaniga, T., Sbarufatti, G., Briani, F. & Dehò, G. Polynucleotide phosphorylase is implicated in homologous recombination and DNA repair in Escherichia coli. BMC Microbiol. 17, 81 (2017).
Rath, D., Mangoli, S. H., Pagedar, A. R. & Jawali, N. Involvement of pnp in survival of UV radiation in Escherichia coli K-12. Microbiology (Reading) 158, 1196–1205 (2012).
Wu, J. et al. Polynucleotide phosphorylase protects Escherichia coli against oxidative stress. Biochemistry 48, 2012–2020 (2009).
Leszczyniecka, M. et al. Identification and cloning of human polynucleotide phosphorylase, hPNPase old-35, in the context of terminal differentiation and cellular senescence. Proc. Natl. Acad. Sci. USA 99, 16636–16641 (2002).
Piwowarski, J. et al. Human polynucleotide phosphorylase, hPNPase, is localized in mitochondria. J. Mol. Biol. 329, 853–857 (2003).
Rius, R. et al. Clinical spectrum and functional consequences associated with bi-allelic pathogenic Pnpt1 variants. J. Clin. Med. 8, 2020 (2019).
Wang, G. et al. PNPASE regulates RNA import into mitochondria. Cell 142, 456–467 (2010).
Shimada, E. et al. PNPase knockout results in mtDNA loss and an altered metabolic gene expression program. PLoS ONE 13, e0200925 (2018).
Chen, H.-W. et al. Mammalian polynucleotide phosphorylase is an intermembrane space rnase that maintains mitochondrial homeostasis. Mol. Cell. Biol. 26, 8475–8487 (2006).
Falchi, F. A., Pizzoccheri, R. & Briani, F. Activity and function in human cells of the evolutionary conserved exonuclease polynucleotide phosphorylase. Int. J. Mol. Sci. 23, 1652 (2022).
Dendooven, T. et al. A cooperative PNPase-Hfq-RNA carrier complex facilitates bacterial riboregulation. Mol. Cell 81, 2901-2913.e5 (2021).
Bernstein, J. A., Lin, P. H., Cohen, S. N. & Lin-Chao, S. Global analysis of Escherichia coli RNA degradosome function using DNA microarrays. Proc. Natl. Acad. Sci. U. S. A. 101, 2758–2763 (2004).
Szczesny, R. J. et al. Human mitochondrial RNA turnover caught in flagranti: Involvement of hSuv3p helicase in RNA surveillance. Nucleic Acids Res. 38, 279–298 (2009).
Borowski, L. S., Dziembowski, A., Hejnowicz, M. S., Stepien, P. P. & Szczesny, R. J. Human mitochondrial RNA decay mediated by PNPase-hSuv3 complex takes place in distinct foci. Nucleic Acids Res. 41, 1223–1240 (2013).
Wang, D. D. H., Shu, Z., Lieser, S. A., Chen, P. L. & Lee, W. H. Human mitochondrial SUV3 and polynucleotide phosphorylase form a 330-kDa heteropentamer to cooperatively degrade double-stranded RNA with a 3′-to-5′ directionality. J. Biol. Chem. 284, 20812–20821 (2009).
Silva, S., Camino, L. P. & Aguilera, A. Human mitochondrial degradosome prevents harmful mitochondrial R loops and mitochondrial genome instability. Proc. Natl. Acad. Sci. 115, 11024–11029 (2018).
Hayakawa, H. & Sekiguchi, M. Human polynucleotide phosphorylase protein in response to oxidative stress. Biochemistry 45, 6749–6755 (2006).
Hayakawa, H., Kuwano, M. & Sekiguchi, M. Specific binding of 8-oxoguanine-containing RNA to polynucleotide phosphorylase protein. Biochemistry 40, 9977–9982 (2001).
Portnoy, V., Palnizky, G., Yehudai-Resheff, S., Glaser, F. & Schuster, G. Analysis of the human polynucleotide phosphorylase (PNPase) reveals differences in RNA binding and response to phosphate compared to its bacterial and chloroplast counterparts. RNA 14, 297–309 (2008).
Amin, N. & Peterkofsky, A. A dual mechanism for regulating cAMP levels in Escherichia coli. J. Biol. Chem. 270, 11803–11805 (1995).
Jarrige, A.-C., Bréchemier-Baey, D., Mathy, N., Duché, O. & Portier, C. Mutational analysis of polynucleotide phosphorylase from Escherichia coli. J. Mol. Biol. 321, 397–409 (2002).
Littauer, U. Z. & Grunberg-Manago, M. Polynucleotide phosphorylase. In Encyclopedia of Molecular Biology (ed. Creighton, T. E.) 1911–1918 (Wiley, 2002). https://doi.org/10.1002/047120918X.emb1169.
Datsenko, K. A. & Wanner, B. L. One-step inactivation of chromosomal genes in Escherichia coli K-12 using PCR products. Proc. Natl. Acad. Sci. U. S. A. 97, 6640–6645 (2000).
Sasaki, I. & Bertani, G. Growth abnormalities in Hfr derivatives of Escherichia coli strain C. J. Gen. Microbiol. 40, 365–376 (1965).
Regonesi, M. E. et al. Analysis of the Escherichia coli RNA degradosome composition by a proteomic approach. Biochimie 88, 151–161 (2006).
Chen, G. & Deutscher, M. P. RNase R is a highly unstable protein regulated by growth phase and stress. RNA 16, 667–672 (2010).
Ishihama, Y. et al. Protein abundance profiling of the Escherichia coli cytosol. BMC Genomics 9, 102 (2008).
Carzaniga, T., Antoniani, D., Dehò, G., Briani, F. & Landini, P. The RNA processing enzyme polynucleotide phosphorylase negatively controls biofilm formation by repressing poly-N-acetylglucosamine (PNAG) production in Escherichia coli C. BMC Microbiol. 12, 270 (2012).
Luttinger, A., Hahn, J. & Dubnau, D. Polynucleotide phosphorylase is necessary for competence development in Bacillus subtilis. Mol. Microbiol. 19, 343–356 (1996).
Iuchi, S., Cole, S. T. & Lin, E. C. C. Multiple regulatory elements for the glpA operon encoding anaerobic glycerol-3-phosphate dehydrogenase and the glpD operon encoding aerobic glycerol-3-phosphate dehydrogenase in Escherichia coli: Further characterization of respiratory control. J. Bacteriol. 172, 179–184 (1990).
Cole, S. T. et al. Nucleotide sequence and gene-polypeptide relationships of the glpABC operon encoding the anaerobic sn-glycerol-3-phosphate dehydrogenase of Escherichia coli K-12. J. Bacteriol. 170, 2448–2456 (1988).
Maslowska, K. H., Makiela-Dzbenska, K. & Fijalkowska, I. J. The SOS system: A complex and tightly regulated response to DNA damage. Environ. Mol. Mutagen 60, 368–384 (2019).
Keseler, I. M. et al. The EcoCyc database in 2021. Front. Microbiol. 12, 711077 (2021).
Raivio, T. L. Everything old is new again: An update on current research on the Cpx envelope stress response. Biochimica et Biophysica Acta (BBA) Mol. Cell Res. 1843, 1529–1541 (2014).
Gupta, A. & Imlay, J. A. Escherichia coli induces DNA repair enzymes to protect itself from low-grade hydrogen peroxide stress. Mol. Microbiol. 117, 754–769 (2022).
Imlay, J. A. The molecular mechanisms and physiological consequences of oxidative stress: Lessons from a model bacterium. Nat. Rev. Microbiol. 11, 443–454 (2013).
Sarkar, D. et al. Human polynucleotide phosphorylase (hPNPaseold-35): A potential link between aging and inflammation. Cancer Res. 64, 7473–7478 (2004).
Kogoma, T., Hong, X., Cadwell, G. W., Barnard, K. G. & Asai, T. Requirement of homologous recombination functions for viability of the Escherichia coli cell that lacks RNase HI and exonuclease V activities. Biochimie 75, 89–99 (1993).
Kouzminova, E. A., Kadyrov, F. F. & Kuzminov, A. RNase HII saves rnhA mutant Escherichia coli from R-loop-associated chromosomal fragmentation. J. Mol. Biol. 429, 2873–2894 (2017).
Vanoosthuyse, V. Strengths and weaknesses of the current strategies to map and characterize R-loops. Noncoding RNA 4, 9 (2018).
Petermann, E., Lan, L. & Zou, L. Sources, resolution and physiological relevance of R-loops and RNA–DNA hybrids. Nat. Rev. Mol. Cell Biol. 23, 521–540 (2022).
Renaudin, X., Lee, M., Shehata, M., Surmann, E. M. & Venkitaraman, A. R. BRCA2 deficiency reveals that oxidative stress impairs RNaseH1 function to cripple mitochondrial DNA maintenance. Cell Rep. 36, 109478 (2021).
Andrs, M. et al. Excessive reactive oxygen species induce transcription-dependent replication stress. Nat. Commun. 14, 1791 (2023).
Yehudai-Resheff, S., Portnoy, V., Yogev, S., Adir, N. & Schuster, G. Domain analysis of the chloroplast polynucleotide phosphorylase reveals discrete functions in RNA degradation, polyadenylation, and sequence homology with exosome proteins. Plant Cell 15, 2003–2019 (2003).
Nurmohamed, S., Vaidialingam, B., Callaghan, A. J. & Luisi, B. F. Crystal structure of Escherichia coli polynucleotide phosphorylase core bound to RNase E, RNA and manganese: Implications for catalytic mechanism and RNA degradosome assembly. J. Mol. Biol. 389, 17–33 (2009).
Soboll, S., Sholz, R. & Heldt, H. W. Subcellular metabolite concentrations: Dependence of mitochondrial and cytosolic ATP systems on the metabolic state of perfuse rat liver. Eur. J. Biochem. 87, 377–390 (1978).
Sarpel, G., Barp, A. N., Lubansky, H. J. & Omachi, A. Erythrocyte phosphate content in Huntington’s disease. Neurosci. Lett. 31, 91–96 (1982).
Lemasters, J. J. Modulation of mitochondrial membrane permeability in pathogenesis, autophagy and control of metabolism. J. Gastroenterol. Hepatol. 22 Suppl 1, S31–S37 (2007).
Colombini, M. The VDAC channel: Molecular basis for selectivity. Biochim. Biophys. Acta 1863, 2498–2502 (2016).
Corkey, B. E., Duszynski, J., Rich, T. L., Matschinsky, B. & Williamson, J. R. Regulation of free and bound magnesium in rat hepatocytes and isolated mitochondria. J. Biol. Chem. 261, 2567–2574 (1986).
Saleet Jafri, M. & Kotulska, M. Modeling the mechanism of metabolic oscillations in ischemic cardiac myocytes. J. Theor. Biol. 242, 801–817 (2006).
Hajnsdorf, E., Steier, O., Coscoy, L., Teysset, L. & Regnier, P. Roles of RNase E, RNase II and PNPase in the degradation of the rpsO transcripts of Escherichia coli: Stabilizing function of RNase II and evidence for efficient degradation in an ams pnp rnb mutant. EMBO J. 13, 3368–3377 (1994).
Sassanfar, M. & Roberts, J. W. Nature of the SOS-inducing signal in Escherichia coli. The involvement of DNA replication. J. Mol. Biol. 212, 79–96 (1990).
Aguilera, A. & García-Muse, T. R Loops: From transcription byproducts to threats to genome stability. Mol. Cell 46, 115–124 (2012).
Teng, Y. et al. ROS-induced R loops trigger a transcription-coupled but BRCA1/2-independent homologous recombination pathway through CSB. Nat. Commun. 9, 4115 (2018).
Demple, B. Radical ideas: Genetic responses to oxidative stress. Clin. Exp. Pharmacol. Physiol. 26, 64–68 (1999).
Renaudin, X. & Venkitaraman, A. R. A mitochondrial response to oxidative stress mediated by unscheduled RNA-DNA hybrids (R-loops). Mol. Cell. Oncol. 8, 2007028 (2021).
Leela, J. K., Syeda, A. H., Anupama, K. & Gowrishankar, J. Rho-dependent transcription termination is essential to prevent excessive genome-wide R-loops in Escherichia coli. Proc. Natl. Acad. Sci. U. S. A. 110, 258–263 (2013).
Gowrishankar, J., KrishnaLeela, J. & Anupama, K. R-loops in bacterial transcription: Their causes and consequences. Transcription 4, 153–157 (2013).
Raghunathan, N. et al. Genome-wide relationship between R-loop formation and antisense transcription in Escherichia coli. Nucleic Acids Res. 46, 3400–3411 (2018).
Shu, Z., Vijayakumar, S., Chen, C. F., Chen, P. L. & Lee, W. H. Purified human SUV3p exhibits multiple-substrate unwinding activity upon conformational change. Biochemistry 43, 4781–4790 (2004).
Carzaniga, T. et al. Autogenous regulation of Escherichia coli polynucleotide phosphorylase expression revisited. J. Bacteriol. 191, 1738–1748 (2009).
Regonesi, M. E. et al. A mutation in polynucleotide phosphorylase from Escherichia coli impairing RNA binding and degradosome stability. Nucleic Acids Res. 32, 1006–1017 (2004).
Briani, F. et al. Genetic analysis of polynucleotide phosphorylase structure and functions. Biochimie 89, 145–157 (2007).
Washburn, R. S. & Gottesman, M. E. Regulation of transcription elongation and termination. Biomolecules 5, 1063–1078 (2015).
Hui, M. P., Foley, P. L. & Belasco, J. G. Messenger RNA degradation in bacterial cells. Annu. Rev. Genet. 48, 537–559 (2014).
Raneri, M., Sciandrone, B. & Briani, F. A whole-cell assay for specific inhibitors of translation initiation in bacteria. J. Biomol. Screen. 20, 627–633 (2015).
Somerville, G. A. & Proctor, R. A. Cultivation conditions and the diffusion of oxygen into culture media: The rationale for the flask-to-medium ratio in microbiology. BMC Microbiol. 13, 9 (2013).
Sciandrone, B., Forti, F., Perego, S., Falchi, F. & Briani, F. Temperature-dependent regulation of the Escherichia coli lpxT gene. Biochim. Biophys. Acta Gene Regul. Mech. 1862, 786–795 (2019).
Fontanella, L. et al. Photometric assay for polynucleotide phosphorylase. Anal. Biochem. 269, 353–358 (1999).
Delvillani, F. et al. Tet-Trap, a genetic approach to the identification of bacterial RNA thermometers: Application to Pseudomonas aeruginosa. RNA 20, 1963–1976 (2014).
Briani, F., Zangrossi, S., Ghisotti, D. & Dehò, G. A Rho-dependent transcription termination site regulated by bacteriophage P4 RNA immunity factor. Virology 223, 57–67 (1996).
Briani, F. et al. Polynucleotide phosphorylase hinders mRNA degradation upon ribosomal protein S1 overexpression in Escherichia coli. RNA 14, 2417–2429 (2008).
Dehò, G., Zangrossi, S., Sabbattini, P., Sironi, G. & Ghisotti, D. Bacteriophage P4 immunity controlled by small RNAs via transcription termination. Mol. Microbiol. 6, 3415–3425 (1992).
Langmead, B. & Salzberg, S. L. Fast gapped-read alignment with Bowtie 2. Nat. Methods 9, 357–359 (2012).
Quinlan, A. R. & Hall, I. M. BEDTools: A flexible suite of utilities for comparing genomic features. Bioinformatics 26, 841–842 (2010).
Robinson, M. D., McCarthy, D. J. & Smyth, G. K. edgeR: A Bioconductor package for differential expression analysis of digital gene expression data. Bioinformatics 26, 139–140 (2010).
Lea, D. E. & Coulson, C. A. The distribution of the numbers of mutants in bacterial populations. J. Genet. 49, 264–285 (1949).
Acknowledgements
We thank A. Callaghan (University of Portsmouth) for pET28-hPNP plasmid and I. Dalle Donne (University of Milan) for helping us with the ROS assay set-up. This work was supported by the Telethon Foundation Grant GGP20001-POLYVAR to FB.
Author information
Authors and Affiliations
Contributions
F.A.F., F.F., C.C., A.G., R.P. and C. M. performed the experiments; G.P. and F.B. analyzed data; F.B. conceived the project and wrote the manuscript. All authors reviewed the manuscript.
Corresponding authors
Ethics declarations
Competing interests
The authors declare no competing interests.
Additional information
Publisher's note
Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.
Supplementary Information
Rights and permissions
Open Access This article is licensed under a Creative Commons Attribution 4.0 International License, which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons licence, and indicate if changes were made. The images or other third party material in this article are included in the article's Creative Commons licence, unless indicated otherwise in a credit line to the material. If material is not included in the article's Creative Commons licence and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this licence, visit http://creativecommons.org/licenses/by/4.0/.
About this article
Cite this article
Falchi, F.A., Forti, F., Carnelli, C. et al. Human PNPase causes RNA stabilization and accumulation of R-loops in the Escherichia coli model system. Sci Rep 13, 11771 (2023). https://doi.org/10.1038/s41598-023-38924-x
Received:
Accepted:
Published:
DOI: https://doi.org/10.1038/s41598-023-38924-x
Comments
By submitting a comment you agree to abide by our Terms and Community Guidelines. If you find something abusive or that does not comply with our terms or guidelines please flag it as inappropriate.