CbrAB-dependent regulation of pcnB, a poly(A) polymerase gene involved in polyadenylation of RNA in Pseudomonas fluorescens

Authors


E-mail x.x.zhang1@massey.ac.nz; Tel. (+64) 9 414 0800; Fax (+64) 9 441 8142.

Summary

CbrB is a global σ54-dependent regulator required for nutrient acquisition in Pseudomonas. Located downstream of cbrB on the Pseudomonas fluorescens SBW25 chromosome is pcnB, a putative poly(A) polymerase gene. Presence of a σ54 promoter in the intergenic region of cbrB and pcnB led to the hypothesis that CbrB regulates pcnB expression in a σ54-dependent manner. Here we show that transcription of pcnB is CbrB dependent. However, 5′-RACE analysis of the pcnB transcript using primers located in the pcnB coding region shows that transcription starts immediately upstream of the putative ATG site at a σ70-like promoter. Deletion of pcnB caused ∼80% decrease of ployadenylated 23S rRNA; growth of the pcnB mutant was compromised in a range of laboratory media and on sugar beet seedlings. Further 5′-RACE analysis confirmed the existence of the predicted σ54 promoter. Genetic analysis showed that the σ54 promoter drives expression of crcZ, a homologue of the recently described small RNA from Pseudomonas aeruginosa, in a CbrB-dependent manner. Taken together, our data show that both pcnB and crcZ are part of the CbrB regulon. Moreover, the data draw further attention to the central regulatory role of CbrB and provides a link between mRNA degradation and cellular catabolism.

Introduction

Bacteria belonging to the genus of Pseudomonas are metabolically versatile and colonize both marine and terrestrial environments and also eukaryotic hosts (plants and animals). The ecological success of Pseudomonas is largely determined by their rapid and accurate detection of – and response to – environmental change. Key components in this regard are the so-named two-component regulatory systems, which are particularly numerous in the genomes of Pseudomonas (Kiil et al., 2005).

In terms of regulation of carbon and nitrogen nutrition two signal transduction systems are of special note. The first is NtrBC – a global activator of genes involved in nitrogen assimilation (Cases et al., 2003; Ninfa and Jiang, 2005); the second, CbrAB, plays a global regulatory role in the regulation of both carbon and nitrogen catabolism (Nishijyo et al., 2001; Zhang and Rainey, 2008). As shown by Biolog phenotypic analysis in Pseudomonas fluorescens SBW25, inactivation of CbrAB resulted in growth defects on 20 carbon sources including seven amino acids (histidine, proline, leucine, isoleucine, valine, arginine and tyrosine) and four carbohydrates (d-xylose, d-mannose, d-ribose and l-arabitol) (Zhang and Rainey, 2008). A similar defect was earlier found with cbrAB mutants of the opportunistic human pathogen Pseudomonas aeruginosa PAO1 (Nishijyo et al., 2001). The cbrB mutant of P. fluorescens SBW25 was also compromised in its ability to grow on eight nitrogen sources (Zhang and Rainey, 2008).

Interestingly, both response regulators (CbrB and NtrC) of the two two-component regulation systems are enhancer-binding proteins (EBPs) for σ54 recruitment, suggesting a key role of σ54 in catabolic control in Pseudomonas. The alternative σ54 factor recognizes promoter elements located in the conserved −24 and −12 regions (TGGCAC N5 TTGCW), which causes the formation of a stable closed holoenzyme complex with RNA polymerase (Reitzer and Schneider, 2001; Xu and Hoover, 2001). However, transcription occurs only in the presence of an EBP in its activated form, which binds to site at a distance of usually 70–150 bp upstream of the promoter and contacts the σ54 holoenzyme by DNA looping (Buck et al., 2000). Like most EBPs, CbrB and NtrC possess a highly conserved σ54-interaction ATPase domain and a helix–turn–helix DNA-binding domain – both located at the C-terminal end. Additionally, the N-terminal region of CbrB and NtrC each contains a signal receiver domain, which is phosphorylated by the cognate sensor kinase, CbrA and NtrB respectively.

The unique features of σ54 promoters and EBPs have been used in a genome-wide prediction of σ54-regulatedgenes in Pseudomonas putida (Cases et al., 2003). A total of 46 σ54 promoters and 22 EBPs were predicted in the P. putida KT2440 genome. Of particular interest is a σ54 promoter located downstream of an EBP (CbrB) and upstream of pcnB, which encodes a putative poly(A) polymerase responsible for adding poly(A) tails to RNA molecules. A similar bioinformatic analysis of P. fluorescens SBW25 also identified a σ54 promoter in the intergenic region between cbrB and pcnB, suggesting the potential involvement of σ54 in regulation of mRNA degradation (Jones et al., 2007). The cbrBpcnB locus including the putative σ54 binding site is highly conserved among different species of Pseudomonas, e.g. in P. aeruginosa PAO1, and Pseudomonas syringae pv. tomato DC3000 (http://xbase.bham.ac.uk/pseudodb/). The presence of a σ54 binding site and an EBP (CbrB) upstream of pcnB suggests that CbrB regulates the level of PcnB expression in a σ54-dependent manner.

pcnB was first identified in Escherichia coli as a gene required for ColE1 plasmid copy number control (Lopilato et al., 1986), but it was later found to be the principal poly(A) polymerase (PAP I) (Cao and Sarkar, 1992). PcnB is responsible for the polyadenylation of over 90% of cellular mRNAs (O'Hara et al., 1995). Poly(A) tails of bacterial mRNA are usually shorter than that of eukaryotic mRNA (14–60 versus 80–200 nucleotides in length) (Sarkar, 1997). Polyadenylation in bacteria is believed to accelerate mRNA degradation, which is contrary to the well-established role of poly(A) tail in transcript stabilization in eukaryotes (O'Hara et al., 1995; Xu and Cohen, 1995). Recently, Mohanty and Kushner (2006) showed in E. coli that most mRNA transcripts in exponentially growing cells are subject to poly(A) modification and there is mounting evidence that suggests that PcnB is involved in specific gene regulation. For example, glmS, a gene encoding glucosamine-6-phosphate synthase, is regulated by polyadenylation (Reichenbach et al., 2008).

Regulation of PcnB expression is complex and not been fully understood. When E. coli is grown in the complete (LB) medium, transcription starts from a single σ70-type promoter (Jasiecki and Wegrzyn, 2006). However, when grown in minimal medium with glucose as the carbon source, pcnB transcription starts from at least three sites (Jasiecki and Wegrzyn, 2006) with the level of transcription being inversely proportional to growth rate (Jasiecki and Wegrzyn, 2003). Interestingly, PcnB translation initiation occurs from an inefficient AUU start codon, which may be important in minimizing toxic effects arising from overproduction of PcnB (Binns and Masters, 2002). As in E. coli, mRNA polyadenylation is also prevalent in Pseudomonas. For example, Saravanamuthu and colleagues (2004) showed that almost half of the transcripts in P. aeruginosa PAO1 are polyadenylated. However, apart from this single report, little information is available on the role of mRNA polyadenylation in regulation of gene expression in Pseudomonas.

Here we describe genetic analysis of pcnB in P. fluorescens SBW25, a plant growth-promoting bacterium originally isolated from the rhizosphere of field-grown sugar beets (Bailey et al., 1995). Transcription of pcnB under different growth conditions was investigated using a chromosomally integrated lacZ fusion. We show that pcnB expression is regulated by CbrB, but that transcription occurs from a putative σ70 type promoter, not the σ54 promoter as previously thought. Although the predicted σ54-type promoter in the front of pcnB is regulated by the adjacent EBP (CbrB), it in fact controls the expression of crcZ– a recently identified non-coding small RNA in P. aeruginosa (Sonnleitner et al., 2009). A mutant lacking pcnB displayed a slow growth phenotype in laboratory media and was compromised in competitive colonization in sugar beet seedlings. Our data implicate the significance of polyadenylation for bacterial acclimation and reveal a link between CbrAB, mRNA degradation and carbon catabolism.

Results

Transcriptional analysis of pcnB in P. fluorescens SBW25

To investigate the regulatory role of CbrB in pcnB expression, chromosomally integrated pcnB–lacZ transcriptional fusions were constructed in wild-type P. fluorescens SBW25 and mutant ΔcbrB (strains PBR847 and PBR848 respectively). Activity of β-galactosidase was monitored hourly for cells growing in M9 medium starting from OD600∼0.1. As shown in Fig. 1, in the wild-type background, the level of pcnB expression gradually decreased during the first 7 h, began to increase from 9 to 10 h and then stabilized at an elevated level. A comparison with absorbance data of the assayed cultures showed that pcnB displayed the highest level of expression from late exponential phase through to stationary phase (Fig. 1). The decrease of β-galactosidase activities during the first 7 h was likely due to the dilution of β-galactosidase already synthesized in the inoculants cells (in stationary phase) via cell division. Significantly, pcnB expression was almost undetectable in the ΔcbrB background during the whole growth period, implicating a regulatory role of CbrB in pcnB expression.

Figure 1.

Expression of pcnB–lacZ in wild-type and ΔcbrB backgrounds. β-Galactosidase activity (µM 4MU OD600−1 min−1) was measured for cells grown in M9 broth (open symbols). Bacterial growth was monitored by measuring absorbance at 600 nm (filled symbols). Data are means and standard errors of six independent cultures: PBR847 (wild type, triangle) and PBR848 (ΔcbrB, circle).

A previous study of pcnB promoter activity in E. coli showed that pcnB expression is inversely correlated with growth rate (Jasiecki and Wegrzyn, 2003). To test whether this holds for P. fluorescens SBW25 the correlation between maximum growth rate (µmax) and pcnB–lacZ activity (strain PBR847) for cells in late exponential phase was determined. Results shown in Fig. 2 indicate a strong positive correlation between growth rate and the level of pcnB transcription (correlation coefficient R = 0.892, P = 0.008).

Figure 2.

Correlation between pcnB transcription and growth rate. Maximum growth rate (µmax) was determined on 10 different media (complete LB medium and minimal M9 salt medium with ammonium as nitrogen source and the following sole carbon sources: glucose, glutamine, glycerol, alanine, succinate, histidine, proline and isoleucine). Bacteria were grown in a 96-well microtitre plate and transcription of pcnB was measured using a lacZ reporter at late exponential phase where expression is maximal (see Fig. 1). Data are means and standard errors of six independent cultures. Bivariate analysis of β-galactosidase activity (µM 4MU OD600−1 min−1) and growth rate (h−1) reveals a correlation coefficient (R) of 0.892 with a P-value of 0.008. When the two data points (growth on LB and glutamine) with very high µmax are removed, the difference remains highly significant (R = 0.75, P < 0.001).

Determination of the pcnB transcriptional start site

To identify the 5′ end of the pcnB transcript, two nested primers (pcnB6 and pcnB5) were designed for 5′-RACE analysis. The inner primer (pcnB5) matches the 5′ end of the putative pcnB coding region. If transcription were to start from the −12 position of the putative σ54 promoter, then a ∼540 bp DNA fragment would result. Total mRNA was prepared from wild-type cells grown in M9 salt medium with histidine as the sole source of carbon and nitrogen, an environment in which CbrB is known to be active and thus the predicted σ54 promoter is likely to also be active. Surprisingly, 5′-RACE analysis produced a very short DNA fragment (∼80 bp). This small fragment was cloned into pCR8/GW/TOPO. Sequencing of six random clones showed that transcription started at an adenine residue, the 53rd nucleotide upstream of the putative ATG start site of pcnB (Fig. 3). Immediately upstream of the pcnB transcriptional start site is a sequence (CCTACA N19 TAGAAT) that is similar to the −10 element of the σ70 promoter consensus (TTGaca N16–18 TAtaaT). Lack of the critical −35 consensus sequence suggests that a positive activator is required for efficient pcnB transcription.

Figure 3.

Genetic organization of the pcnB locus from P. fluorescens SBW25 with details of the cbrBpcnB intergenic region. Transcriptional start sites as determined by 5′-RACE analysis are indicated by ‘+1’ and arrowed lines. The predicted σ54- and σ70-type promoters are underlined. Palindromic sequences indicated by inverted arrow lines might function as a transcriptional terminator. Sequence in bold letters denote a putative non-coding small RNA identified in P. aeruginosa PAO1 (PA4726.2), which was transcribed from the opposite direction of the cbrB and pcnB (Livny et al., 2006).

Having shown that pcnB is not transcribed from the predicted σ54 promoter, we proceeded to test the hypothesis that it controls expression of a small orf or non-coding RNA between cbrB and pcnB. To do this, we designed two new nested primers (pcnB9 and pcnB10) for 5′-RACE analysis of the putative σ54 promoter. Unlike primers pcnB5/pcnB6 described above, primers pcnB9/pcnB10 were located ∼200 nt downstream of the promoter. Significantly, 5′-RACE analysis produced a DNA fragment with the predicted size. Subsequent cloning and sequencing of the PCR product revealed a transcript starting from an adenine residue, which is the 13th nucleotide downstream from the conserved C residue (underlined) of the predicted −12 element (TGGCAC N5 CTGCT, Fig. 3). These data thus confirm the activity of the σ54 promoter.

To test whether CbrB is required for σ54 promoter activity, a lacZ fusion plasmid pUIC3-65 was constructed with a promoterless lacZ gene fused to the 106th nucleotide of the transcript. The fusion plasmid pUIC3-65 was then introduced into wild-type SBW25 and the ΔcbrB mutant. The resulting fusion strains (PBR849 for wild type and PBR850 for ΔcbrB) were assayed for β-galactosidase activity using cells grown in M9 broth. Results showed that the promoter activities weregreatly reduced in ΔcbrB mutant compared with wild type (Fig. 4).

Figure 4.

Activities of the σ54 promoter in wild-type SBW25 and ΔcbrB mutant. β-Galactosidase activity (µM 4MU OD600−1 min−1) was measured for cells grown in M9 broth. Data are means and standard errors of six independent cultures. Two-way anova analysis revealed a significant effect of both genotype (wild type versus ΔcbrB: F1,83 = 316.43, P < 0.0001) and time (F6,83 = 7.31, P < 0.0001).

Interrogation of the intergenic region of cbrB and pcnB revealed two GC-rich palindromic sequences (Fig. 3): one is likely to be a Rho-independent transcriptional terminator for cbrB whereas the other might be a terminator for the σ54-driven transcript. This σ54-driven transcript shows 77% sequence identity with a recently identified small RNA named CrcZ in P. aeruginosa PAO1 (Sonnleitner et al., 2009); expression of the P. aeruginosa crcZ is also regulated by the adjacent cbrB gene (Sonnleitner et al., 2009). Accordingly, in P. fluorescens SBW25, this transcript is likely to be a small RNA and does not encode a short peptide.

Functional characterization of pcnB

In silico analysis of the pcnB locus suggested that pcnB forms a dicistronic operon with the downstream folK, a putative gene involved in folic acid biosynthesis (Fig. 3). A pcnB in-frame deletion mutant (PBR845) was thus constructed in order to minimize the possible polar effect of pcnB mutation. To test the predicted role of PcnB in RNA polyadenylation, mutant PBR845 was subjected to assays for polyadenylated RNAs in comparison with the wild-type strain. To do this, total RNAs were prepared from cells grown in LB and reverse-transcribed using the biotin-labelled oligo(T) primer (T20). However, detection of the resulting cDNAs revealed no significant difference between wild type and the ΔpcnB mutant (data not shown), possibly reflecting non-specific reverse transcription. We therefore focused attention on detection of specific RNA transcripts, specifically the 23S rRNA given previous work that 23S rRNAs were predominantly polyadenylated in E. coli (Mohanty and Kushner, 1999) and P. aeruginosa (Saravanamuthu et al., 2004).

The first-strand cDNAs were prepared by reverse transcription from total RNAs of three independent cultures using the oligo(T) adapter primer AP (Table 1). Specific amplification of 23S cDNA was then performed using forward primer 23S1 and reverse primer AAP (Abridged Adapter Primer, biotin-labelled). The resulting PCR products were firstly analysed by electrophoresis through a 1.5% agarose gel. Results showed a weak band of ∼350 bp in the wild-type strain, but no bands were visible in the pcnB mutant (data not shown). Further detection of the biotin-labelled PCR products, as described in Experimental Procedures, revealed the presence of multiple smeared bands (Fig. 5). Significantly, the amplified 23S cDNAs were greatly reduced (∼80%) in the pcnB mutant compared with wild type (Fig. 5).

Table 1.  Results of phenotypic characterization of P. fluorescens SBW25ΔpcnB.
Growth environmentaGrowth rate (h−1)bLag time (h)bRelative fitness (SRC)c
Wild typeΔpcnBWild typeΔpcnB
  • a. 

    Bacteria were grown in complete medium LB or M9 salt medium (MSM) supplemented with ammonium and the respective carbon substrate at a final concentration of 15 mM.

  • b. 

    Data are means and standard errors of six independent cultures. One-way anova revealed significant differences among means (F17,119 = 48.32, P < 0.0001 for µmax; F17,119 = 571.33, P < 0.0001 for lag time). The growth rate (µmax) and lag time identified by different letters in parenthesis are significantly different (P < 0.05) by Tukey's HSD.

  • c. 

    Fitness of the ΔpcnB mutant relative to wild-type SBW25 was expressed as selection rate constant (SRC). Data are means and standard errors of six independent cultures or sugar beet seedlings. One-sample t-tests were performed to test whether the average values are significantly (S) or not significantly (NS) different from zero (P < 0.05).

LB0.48 ± 0.03 (a)0.5 ± 0.05 (a)2.88 ± 0.19 (j)3.58 ± 0.65 (j)−1.33 ± 0.11 (S)
Glutamine0.34 ± 0.009 (b)0.27 ± 0.012 (bc)6.95 ± 0.18 (h)9.45 ± 0.2 (fg)−2.51 ± 0.06 (S)
Glycerol0.26 ± 0.007 (bcd)0.18 ± 0.007 (ef)11.1 ± 0.22 (e)15.18 ± 0.14 (d)−3.97 ± 0.20 (S)
Alanine0.25 ± 0.008 (cde)0.23 ± 0.006 (cdef)9.62 ± 0.17 (fg)11.13 ± 0.13 (e)−1.23 ± 0.12 (S)
Glucose0.22 ± 0.011 (cdef)0.2 ± 0.005 (def)5.42 ± 0.15 (i)5.64 ± 0.17 (i)−0.29 ± 0.13 (NS)
Succinate0.21 ± 0.003 (cdef)0.22 ± 0.008 (cdef)7.71 ± 0.19 (h)9.15 ± 0.21 (g) 
Histidine0.21 ± 0.004 (cdef)0.16 ± 0.003 (f)10 ± 0.11 (f)14.51 ± 0.24 (d)−4.19 ± 0.38 (S)
Proline0.18 ± 0.007 (ef)0.21 ± 0.009 (cdef)11.47 ± 0.28 (e)16.26 ± 0.31 (c)−2.09 ± 0.20 (S)
Isoleucine0.17 ± 0.012 (f)0.03 ± 0.004 (g)23.54 ± 0.41 (b)28.01 ± 0.52 (a)−1.9 ± 0.17 (S)
Shoot    −0.85 ± 0.18 (S)
Rhizosphere    −0.86 ± 0.08 (S)
Figure 5.

RT-PCR analysis of the 3′ ends of polyadenylated 23S rRNA. Total RNAs were prepared from three independent cultures of wild-type SBW25 and mutant PBR845 (ΔpcnB). Reverse transcription was performed using the oligo(T) adaptor primer AP and subsequent amplification of 23S cDNA was performed using primers 23S1 and AAP (the abridged adaptor primer, biotin-labelled). The purified amplification product (15 µl) was then separated on a 6% polyacrylamide gel and transferred to nylon membrane for detection of biotin-labelled DNA fragment (see details in Experimental procedures). The multiple smeared bands reflect polyadenylation at various sites.

To identify the polyadenylation target sites and the sequence of the poly(A) tails, 23S cDNAs were prepared from total RNAs following the procedure described above using the non-biotin-labelled AP primer. The PCR products were cloned into vector pCR8/GW/TOPO after purification via Qiagen columns. Twenty-six clones were randomly selected from wild-type SBW25. Results of DNA sequencing (Fig. S1) showed that polyadenylation occurred at 15 different locations; the length of the poly(A) tails varied from 13 to 66 nt. Analysis of nine 23S rDNA clones from mutant PBR845 (ΔpcnB) revealed seven target sites, three of which were also identified in the wild type (Fig. S2). The multiple ployadenylation sites are consistent with the observed smeared bands shown in Fig. 5. Both homopolymeric poly(A) tails and adenosine rich heteropolymers were identified in wild type and mutant PBR845. However, mutant PBR845 had a higher proportion of heteropolymeric tails (45%) compared with the wild type (30%). The longest homopolymeric tail identified in wild type had 38 contiguous adenosine residues (Fig. S1), whereas the longest heteropolymer was 92 nt with 66.1% adenosine found in mutant PBR845 (Fig. S2).

Phenotypic characterization of the ΔpcnB mutant

Mutant PBR845 (ΔpcnB) was characterized by determining its ability to grow in complete medium (LB) and minimal medium with eight different sole carbon sources. The results are summarized in Table 1. In general, the mutant grew more slowly than wild type with either lower maximum growth rate (µmax) or longer lag time. The differences are significant (P < 0.05) in terms of µmax for two carbon sources (glycerol and isoleucine) and lag time for all growth media except LB and M9 (glucose). As an example, growth curves of wild-type SBW25 and mutant PBR845 on histidine as the sole source of carbon are shown in Fig. 6. Significantly, the growth defect of PBR845 was restored with a cloned copy of pcnB in the shuttle vector pME6032 (Fig. 6). Moreover, growth of the ΔpcnB mutant, together with the wild-type strain, was examined in a Biolog GN2 microplate, which assessed the utilization of a panel of 95 carbon sources. A significant growth defect was found with 14 substrates, including glycerol, succinate and histidine (see detail in Table S1). Together the growth data show that the primary effect of pcnB deletion is to reduce the rate of carbon source utilization.

Figure 6.

Growth curves of wild-type P. fluorescens SBW25, mutant PBR845 (ΔpcnB) and the pcnB complementation strain PBR854 (ΔpcnB, pME6032-pcnB). Absorbance at 450 nm was measured for cultures grown in M9 salt medium with histidine (15 mM) as the sole carbon source and ammonium as nitrogen source. IPTG (500 µM) was supplemented in all samples to induce expression of pcnB. Data are means and standard errors of six independent cultures.

To measure the growth effects more precisely, competitive growth of the ΔpcnB mutant was assessed relative to a neutral lacZ-marked strain of wild-type SBW25 (Zhang and Rainey, 2007b). Relative fitness was expressed as selection rate constant (SRC: a SRC of zero indicates that the two competitors are equally fit). The relative fitness results listed in Table 1 showed that the ΔpcnB mutant was less competitive than the wild type in all the tested media and the differences are significant (P < 0.05) except in M9 (glucose). Notably, the fitness data are poorly correlated with the levels of pcnB expression detected in these media (correlation efficient R = 0.27). The competitive ability of ΔpcnB relative to SBW25–lacZ was also determined over the course of colonization of sugar beet seedlings. Bacteria were inoculated onto the coated seeds and recovered from the shoot and rhizosphere after 2-week competitive growth in planta. Results showed that the ΔpcnB mutant was significantly impaired in the plant environment (Table 1).

Discussion

Bacteria acclimate to environmental change by altering levels of gene expression, which is achieved via a coordinated process of mRNA synthesis and degradation. While considerable efforts have been taken to understand the regulation of transcription and mRNA degradation, the link between the two processes remains unclear. In this work we demonstrate that in P. fluorescens SBW25 pcnB expression is controlled by the two-component signal transduction system CbrAB. Given that PcnB poly(A) polymerase regulates mRNA stability and that CbrB is a global transcriptional regulator for nutrient utilization, our data implicate an important role of CbrAB in coordinating transcription of catabolic genes and the rate of transcript degradation. Such a role is consistent with the positive correlation between levels of pcnB expression and bacterial growth rate as revealed in this work. Further support is provided by the co-identification of cbrAB and pcnB in microarray study of stress responsive genes in P. putida KT2440 (Reva et al., 2006).

Expression of pcnB is not mediated by the σ54 promoter as previously predicted from bioinformatic studies (Cases et al., 2003; Jones et al., 2007): rather, pcnB is transcribed from a σ70-like promoter. The deduced amino acid sequence of cbrB contains the highly conserved helix–turn–helix DNA-binding domain and it is predicted to be a σ54-dependent transcriptional activator, thus the regulatory role of CbrB in pcnB expression appears to be indirect. This is similar to our previous finding that CbrB activates the histidine utilization (hut) operon from a putative σ70-type promoter in cells grown on histidine as the sole nitrogen source (Zhang and Rainey, 2008). The results from this study suggest the presence of σ70-type activator(s) under the control of CbrAB, but such a σ70-type regulator(s) remains unidentified. The possibility that it might be a homologue of the NAC protein, which couples regulation of σ70-dependent genes to the σ54-dependent Ntr system in enteric bacteria (Bender, 1991), has been rejected (Zhang and Rainey, 2008).

Molecular mechanisms of mRNA degradation in bacteria have been extensively studied only in E. coli. Given the phylogenetic closeness of E. coli and Pseudomonas (both are members of the gamma subdivision of the Proteobacteria), it is surprising to notice the functional differences between the major mRNA decay enzymes from the two organisms. Oligoribonuclase (Orn) is involved in the final step of mRNA degradation to hydrolyse small oligoribonucleotides to monoribonucleotides. Inactivation of orn is lethal in E. coli (Ghosh and Deutscher, 1999), whereas the orn mutant of P. putida has a reduced growth, but is nevertheless viable (Zhang et al., 2004). Inactivation of pcnB in E. coli has little effect on growth (Masters et al., 1993) and the levels of pcnB expression are higher in slow-growing cells than in fast-growing cells (Jasiecki and Wegrzyn, 2003). However, in P. fluorescens SBW25, we found the level of pcnB transcription to be positively correlated with bacterial growth rate. Moreover, deletion of pcnB resulted in a slow growth phenotype in laboratory media; the mutant was also comprised in competitive colonization of sugar beet plants. These differences between key mRNA decay enzymes (Orn and PcnB) in E. coli and Pseudomonas may reflect differences in the environments that the two organisms inhabit.

While PcnB is the principal poly(A) polymerase in E. coli, another mRNA decay enzyme PNPase (polynucleotide phosphorylase) – a 3′ to 5′ exonuclease – also plays an important role in mRNA polyadenylation (Mohanty and Kushner, 2000). The two kinds of poly(A) tails differ in nucleotide composition. PcnB synthesizes homopolymeric poly(A) tails, whereas the PNPase-synthesized tails are primarily adenosine rich heteropolymers (Mohanty and Kushner, 2000; 2006). Notably, our preliminary analysis of poly(A) tails of the 23S rRNA in P. fluorescens SBW25 indicates that pcnB deletion results in an increase of heteropolymeric tails, while fewer 23S rRNAs were polyadenylated in the mutant compared with wild type. Additionally, an almost random location of polyadenylation sites in Pseudomonas suggests that the poly(A) tails in Pseudomonas function in a similar manner to those in E. coli, i.e. to facilitate efficient RNA degradation.

Our analysis of the pcnB promoter activity led to identification of a σ54-driven transcript under the control of CbrB. This result was initially surprising because the σ54-driven transcript is encoded by the complementary strand of a previous identified small RNA (PA4726.2) in P. aeruginosa PAO1 (Fig. 3) that is highly conserved across different species of Pseudomonas (Livny et al., 2006). However, our finding is consistent with a recent report from P. aeruginosa PAO1 that a homologous small RNA (termed CrcZ) is transcribed in a CbrB- and σ54-dependent manner (Sonnleitner et al., 2009). Significantly, in PAO1, CrcZ was shown to be a global regulator of carbon catabolite repression, which is capable of binding with high affinity to the Crc (catabolite repression control) protein (Sonnleitner et al., 2009). Crc is responsible for the succinate-provoked catabolite repression of several catabolic pathways in P. aeruginosa and in P. putida. It has been shown that repression is achieved as a result of decreased translation via specific Crc binding to the mRNA transcripts (Moreno et al., 2007; 2009; Moreno and Rojo, 2008). Sonnleitner and colleagues (2009) proposed that the Crc-mediated catabolite repression could be relieved by high affinity binding of CrcZ to Crc. Notably, the CrcZ homologue in P. fluorescens SBW25 possesses the five CA-motifs required for Crc binding (Moreno et al., 2009; Sonnleitner et al., 2009), implicating a similar role of SBW25 CrcZ in catabolite repression.

The precise function of Crc protein is currently unknown. Despite a high level of sequence similarity with nucleases, it does not bind DNA and exhibits no endo- or exo-nuclease activities (MacGregor et al., 1996). It is thus possible that Crc not only binds RNA, but also functions as an RNA degradation enzyme involved in the turnover of certain mRNAs. The fact that both PcnB and CrcZ are part of the CbrAB regulon leads us to suggest that CbrAB-dependent PcnB expression is an integral mechanism of catabolite repression: PcnB either stimulates the rate of degradation of catabolic gene transcripts in cooperation with Crc, or controls the ratio of CrcZ to Crc via polyadenylation modification of CrcZ or Crc mRNA.

Experimental procedures

Bacterial strains, plasmids and growth conditions

A summary of bacterial strains and plasmids used in this study is provided in Table 2. Both P. fluorescens and E. coli strains were routinely grown in Luria–Bertani (LB) medium (Sambrook et al., 1989) at 28°C and 37°C respectively. Escherichia coli strain DH5αλpir was used as a recipient for gene cloning and a donor for the subsequent conjugation into Pseudomonas. Pseudomonas fluorescens SBW25 and the derived strains were also grown in minimal medium M9 (Sambrook et al., 1989): M9 salts medium (MSM) supplemented with glucose (0.4% or 22.2 mM) and ammonium chloride (1 mg ml−1). Where glucose was replaced by other carbon sources, the final substrate concentration was 15 mM.

Table 2.  Bacterial strains, plasmids and oligonucleotide primers used in this work.
Strain or plasmidRelevant propertiesReference or application
  • a. 

    Primer sequences are shown 5′ to 3′. Restriction sites incorporated into the primers are underlined. Complementary sequences designed for the SOE-PCR are shown in lowercase type.

Pseudomonas fluorescens  
SBW25Wild-type strain isolated from sugar beetBailey et al. (1995)
PBR810SBW25ΔcbrBZhang and Rainey (2007a)
PBR845ΔpcnB, SBW25 with deletion of pflu5239This work
PBR847SBW25 carrying pcnB–lacZ fusion, SBW25::pUIC3-59This work
PBR848SBW25ΔcbrB carrying pcnB–lacZ fusion, SBW25ΔcbrB::pUIC3-59This work
PBR849SBW25 carrying crcZ–lacZ fusion, SBW25::pUIC3-65This work
PBR850SBW25ΔcbrB carrying crcZ–lacZ fusion, SBW25ΔcbrB::pUIC3-65This work
PBR854pcnB complementation strain, ΔpcnB, pME6032-pcnBThis work
Plasmid
pRK2013Helper plasmid, Tra+, KmrDitta et al. (1980)
pUIC3Integration vector with promoterless lacZ, oriR6K, Mob+, TcrRainey (1999)
pUIC3-59pUIC3 carrying pcnB–lacZ fusion, oriR6K, TcrThis work
pUIC3-60pUIC3 carrying 1.6 kb fragment with pcnB deletion, TcrThis work
pUIC3-65pUIC3 containing crcZ–lacZ fusion, oriR6K, TcrThis work
pME6032-pcnBRecombinant plasmid for pcnB complementation, TcrThis work
Primera
pcnB1GAGATCTGGCGCGCAAATTGGGCGTpcnB deletion, pcnB– & crcZ–lacZ
pcnB2cagcatgcggatccgttgacggaCATGGGATGCACTGTTTGAApcnB deletion, pcnB–lacZ
pcnB3tccgtcaacggatccgcatgctgCGAATAAACGTGGGACGCATCpcnB deletion
pcnB4GAGATCTGGCATGAAAAGGGTCCTCpcnB deletion
pcnB5CATGGGATGCACTGTTTGAA5′-RACE for pcnB promoter
pcnB6GCTGCGTTTGTGTTGCGTAC5′-RACE for pcnB promoter
pcnB7AGGATCCGGGATTTGTCTGGCATGAATCpcnB complementation
pcnB8AGGATCCAGCGCTTCAATAGCACTGpcnB complementation
pcnB9TCCCAATCGGGAGCCAAGAGCA5′-RACE for crcZ promoter
pcnB10GTCGCTGTGTTGCTGATCTAC5′-RACE for crcZ promoter
pcnB11GAGATCTCCAAAAGAATCAGTTAGCTcrcZ–lacZ
23S1AGAACGTCGTGAGACAGTTC23S rDNA amplification
T20TTTTTTTTTTTTTTTTTTTT 
APGATGGTACCTCTAGAGCTCTTTTTTTTTTTTTTTTTTAdaptor primer
AAPGATGGTACCTCTAGAGCTCTAbridged adaptor primer

Mutational analysis and gene complementation

Gene cloning and PCR amplification were performed using standard DNA techniques. Deletion of bacterial genes was achieved by SOE-PCR (splicing by overlapping extension using the polymerase chain reaction) (Horton et al., 1989) in conjugation with a two-step allelic exchange strategy as previously described (Zhang and Rainey, 2007a). For the construction of each mutant four oligonucleotide primers were designed to amplify two fragments of similar sizes flanking the deleted region. The two DNA fragments were joined together via an additional step of PCR reaction because of the complementary sequences incorporated into the primers. A summary of the oligonucleotide primers is provided in Table 2. The PCR product was first cloned into pCR8/GW/TOPO (Invitrogen) and then to the integration vector pUIC3 (Rainey, 1999). The pUIC3 recombinant plasmid was mobilized into Pseudomonas by conjugation with the help of pRK2013 (Tra+) and selection of allelic exchange mutants was performed by a modified procedure of d-cycloserine enrichment (Zhang and Rainey, 2007a). Complementation of pcnB was performed by cloning the PCR-amplified coding region of pcnB into the broad-host-range plasmid pME6032 under the control of an IPTG-inducible Ptac promoter (Heeb et al., 2002).

Construction of lacZ fusions and assay for β-galactosidase activity

To generate a gene fusion to the promoterless lacZ, ∼800 bp DNA fragment was PCR-amplified and cloned into pCR8/GW/TOPO (Invitrogen). The error-free DNA fragment was then retrieved and cloned into the multi-cloning site of pUIC3 (Rainey, 1999). The resulting pUIC3 plasmid was mobilized into Pseudomonas and integrated into the chromosome via a homogenous recombination event of insertion-duplication (Zhang et al., 2006). β-Galactosidase activity was assayed using a standard protocol with 4-methylumbelliferyl-β-d-galactoside (4MUG) as the enzymatic substrate. The fluorescent product, 7-hydroxy-4-methylcoumarin (4MU), was detected at 460 nm after excision at 365 nm using a Synergy 2 plate reader (BioTek Instruments).

RNA techniques

Total RNAs were prepared from P. fluorescens cells grown in stationary phase using the TRIzol RNA extraction reagent purchased from Invitrogen (Auckland, New Zealand). Transcriptional start sites were determined with the rapid amplification of cDNA 5′ ends (5′-RACE) system (Invitrogen) following the manufacturer's recommendation. The resulting PCR fragments were cloned into pCR8/GW/TOPO and ∼10 colonies were randomly picked up for DNA sequencing.

To examine the 3′ ends of the ployadenylated 23S rRNA, total RNAs were treated with RNase-free DNase I (Invitrogen) and 2 µg was then used for first-strand cDNA synthesis using the adaptor primer (AP). Reverse transcription was catalysed by Superscript II reverse transcriptase from Invitrogen in a total volume of 20 µl. Next 2 µl was used as template for amplification of 23S cDNA using primers 23S1 and the abridged adaptor primer AAP. Amplification was performed using standard protocol with an annealing temperature of 57°C and 30 thermal cycles. The resulting PCR products were purified using the Qiagen DNA purification columns and 15 µl for each sample was analysed in 6% polyacrylamide gel. The DNA fragments were then transferred by electroblotting to a BiotransTM nylon membrane, which was subjected to probe detection using Biotin Chromogenic Detection Kit (Roche, Auckland). For sequence analysis of 23S rRNA ploy(A) tails, the 23S cDNAs were amplified using primer 23S1 and non-biotin-labelled AAP primer. The resultant amplification products were directly cloned into pCR8/GW/TOPO (Invitrogen) for DNA sequencing analysis.

Assessment of bacterial growth in laboratory media and in planta

Growth kinetics of P. fluorescens SBW25 and the derived mutants in laboratory media were determined in 96-well plates using a VersaMax microtitre plate reader with SOFTmax PRO software (Molecular Devices) and absorbance at 450 nm was measured every 5 min. Growth rate (µmax) was calculated by plotting the logarithm of absorbance date against time. The maximum slope of the resulting line is defined as the maximum specific growth rate under the tested condition. GN2 Biolog assays were performed as previously described (Zhang and Rainey, 2008) and optical density (A660) was measured at three time points (0, 24 and 48 h).

Fitness of a mutant strain relative to the wild-type ancestor was measured by direct competition with a lacZ-marked strain of P. fluorescens SBW25 (Zhang and Rainey, 2007b). Relative fitness was calculated as SRC, which produces a zero value when the mutant and wild type are equally fit (a negative value indicates a reduction of fitness relative to wild type) (Zhang and Rainey, 2007b).

Acknowledgements

We thank Gail Preston for drawing our attention to the existence of a putative σ54 promoter between cbrB and pcnB. This work was supported by the Marsden Fund Council from government funding administered by the Royal Society of New Zealand.

Ancillary