Introduction

The free-living bacteria face constant environmental changes; for survival, they require the adaptation of their metabolism through a flexible genic regulation. One of the signal transduction systems involved in such adaptation is the Gac-Rsm system, which is composed of the two-component system (TCS) GacS/A and the Rsm post-transcriptional regulatory system [1]. The TCS GacS/A has been extensively studied in Pseudomonas spp. and in several enterobacteria, in which TCS controls transcription of small regulatory RNAs (sRNAs) from the Rsm regulatory system (also called Csr, outside the Pseudomonadaceae). Generally, the Rsm system has two or more sRNAs, and a protein that acts as a post-transcriptional repressor (RsmA/CsrA). The Rsm-sRNAs have stem-loop structures containing GGA motifs counteracting the repressor activity of the RsmA protein. When Rsm-sRNAs are absent, the RsmA protein binds over or near the ribosome binding site of its mRNA target blocking translation and promoting its degradation [1].

In Erwinia carotovora and Pseudomonas fluorescens, the Rsm (Csr) system modulates the production of extracellular enzymes, secondary metabolites and motility [2, 3]. In P. aeruginosa it controls the quorum sensing system, exoproduct synthesis, motility and biofilm formation [4]. Moreover, in Escherichia coli the system controls the expression of genes involved in carbon metabolism, motility, biofilm formation and the production of the exopolysaccharide PGA (poly-β-1,6-N-acetyl-d-glucosamine) [5,6,7,8], while in Salmonella it controls the expression of invasion genes [9].

Rsm-sRNAs have been classified into three families, RsmX, RsmY and RsmZ [10]. The presence of two or more sRNAs of the Rsm family has been reported in some bacterial species. In bacteria of the Pseudomonadaceae family, the number of Rsm-sRNAs varies from two in P. aeruginosa (RsmZ and RsmY) to five in Pseudomonas syringae (five isoforms of RsmX) [10,11,12,13]. A. vinelandii is the bacterium with the highest number of Rsm-sRNAs reported to date, with seven sRNAs belonging to the RsmZ family (RsmZ1–RsmZ7) and one of the RsmY family [14, 15]. Several authors have proposed that such sRNAs act synergically to counteract the negative effect of RsmA [1, 16, 17].

The TCS GacS/A controls the expression of thousands of genes; however the only identified genes directly regulated by this TCS are those encoding the Rsm/Csr-sRNAs [1, 18, 19]. Generally, genes encoding the Rsm/Csr-sRNAs have GacA regulatory boxes located upstream from the transcription start site [13, 18]. Thus, conditions activating the TCS GacS/GacA are commonly identified by their effect on the rsm/csr sRNA gene transcription. In E. coli, csrB expression responds to short-chain organic acids such as acetate or formate [20] and the csrB and csrC expressions are up-regulated during growth in a nutrient-poor medium [21]. In P. fluorescens, Krebs cycle intermediaries are related to the stimulus triggering Rsm-sRNAs transcription [22]. In this bacterium, transcriptional activation of rsmZ during transition from exponential to stationary growth phase also responds to an extracellular compound, whose nature is still unknown [23].

Global regulators such as IHF, H–NS and CRP are also involved in the control of the rsm (csr) gene transcription [18, 19]. IHF in P. fluorescens and MvaT/MvaU (H–NS homologues) in P. aeruginosa are involved in the rsmZ regulation.

Azotobacter vinelandii is a nitrogen-fixing soil bacterium that belongs to the Pseudomonadaceae family and has the capacity to form desiccation-resistant cysts [24]. In this bacterium, the GacS/A-Rsm pathway controls the alginate biosynthesis, a linear co-polymer of ß-d-mannuronic and its C-5 epimer α-l-guluronic acid [25]. This polymer has a broad range of applications as gelling, stabilizing and thickening agents in many industries. Alginate is produced under vegetative growing conditions and during the differentiation process leading to the formation of cysts. The GacS/A-Rsm systems control the algD gene expression encoding a GDP-mannose dehydrogenase, the key enzyme of the alginate biosynthetic pathway [14]. RsmA recognizes the algD mRNA leader blocking translation, while GacA is necessary for the Rsm-sRNAs expression, which in turn counteracts the RsmA translational repression effect. In A. vinelandii, the roles of sRNAs RsmZ1 and RsmZ2 in alginate production have been studied; individual mutations in both genes significantly diminished alginate production and rsmZ1 over-expression restored to alginate synthesis in a gacA mutant [14].

Other metabolites controlled by GacA and RsmZ1 are the alkylresorcinols (ARs), a family of phenolic lipids, which are important components of the membrane of A. vinelandii cysts. Mutations in gacA and rsmZ1 impaired the synthesis of ARs [26]. The genes necessary to produce ARs are encoded by the arsABCD operon. This operon is activated by ArpR, a LysR-type transcriptional regulator. RsmA represses arpR expression by binding to its mRNA; thus, the positive role of GacA and RsmZ in the biosynthesis of ARs is explained by their negative effect on RsmA activity.

As in other bacteria, transcription of all the rsmZ-Y sRNAs genes in A. vinelandii is GacA-dependent [14, 15]. Interestingly, the location of putative GacA-binding boxes varies among the different rsmZ-sRNAs genes, suggesting different modes of regulation. While the GacA-binding box for rsmZ1 and rsmZ2 was located at positions ranging from − 175 to − 155 and − 181 to − 163, relative to the transcription start site, the GacA boxes in the rsmZ3, rsmZ4, rsmZ5, rsmZ6 and rsmZ7 genes were found at positions − 50 to − 80 bp [27]. However, the conditions promoting transcription of A. vinelandii rsm-sRNAs genes are still unknown. In this work, we studied the effect of different types of carbon sources on the expression of the rsmZ1-7 genes and the possible relationship between the expression of these genes with the production of alginate and ARs.

Materials and Methods

Microbiological Procedures

Bacterial strains and plasmids used in this study are listed in Table 1. The A. vinelandii wild-type E strain (also named AEIV) was used in this study [28]. A. vinelandii was grown at 30 °C in Burk’s nitrogen-free salts medium [29] supplemented with 20 g/L of sucrose (BS), fructose (BFru), glucose (BGlu) or succinate (BSucc). E. coli strain DH5-α was grown in Luria–Bertani medium (LB) [30] at 37 °C. Antibiotic concentrations used (in µg/ml) for A. vinelandii and E. coli, respectively, were as follows: tetracycline (Tc), 40 and 20; kanamycin (Km), 4 and 20; gentamicin (Gm), 1.5 and 10; streptomycin (Sm) 2 and 20; ampicillin (Ap), not used and 100; nalidixic acid (Nal), 10 and 10. A. vinelandii transformation was carried out as previously described [31] with some modifications as reported previously [32].

Table 1 Strains and plasmids used in this study
Full size table

Nucleic Acid Procedures

DNA isolation and cloning procedures were carried out as described previously [33]. The A. vinelandii DJ genome sequence was used to design the oligonucleotides used for PCR amplifications. The DreamTaq polymerase (Thermo Fisher Scientific) and Vent DNA polymerase (NEB) were used for PCR amplifications.

Cloning the A. vinelandii rsmZ3, rsmZ4, rsmZ5, rsmZ6 and rsmZ7 Genes

Fragments of approximately 2 kb containing the rsmZ3, rsmZ4, rsmZ5, rsmZ6 and rsmZ7 loci were amplified by PCR using the primers specified in Table S1; chromosomal DNA from the wild-type mucoid strain E was used as DNA template. The amplified fragments were individually cloned into pGEM-T Easy (Promega), which generated pGEMrsmZ3, pGEMrsmZ4, pGEMrsmZ5, pGEMrsmZ6 and pGEMrsmZ7 plasmids. These plasmids were used to determine the nucleotide sequence of the rsmZ3, rsmZ4, rsmZ5, rsmZ6 and rsmZ7 loci of the A. vinelandii E strain.

Generation of rsmZ3, rsmZ4, rsmZ5, rsmZ6 and rsmZ7 Mutants Derived from the E Strain

A schematic representation of the procedure to generate deletion mutants in each of the rsmZ3–7 genes is depicted in Fig. S1. Individual rsmZ-sRNA deletions were generated by inverse PCR using the primers specified in Table S1. These primers contain either XhoI or KpnI sites used in subsequent steps to insert a selection marker (an antibiotic-resistance cassette). Inverse PCRs were carried out using pGEMrsmZ3, pGEMrsmZ4, pGEMrsmZ5, pGEMrsmZ6 or pGEMrsmZ7 plasmids as DNA template. The resulting PCR fragments were digested with XhoI or KpnI endonucleases and were re-ligated generating pGEMΔrsmZ3, pGEMΔrsmZ4, pGEMΔrsmZ5, pGEMΔrsmZ6 and pGEMΔrsmZ7 plasmids, respectively. These plasmids were linearized with KpnI (pGEMΔrsmZ4, pGEMΔrsmZ5 and pGEMΔrsmZ6) or XhoI (pGEMΔrsmZ3 and pGEMΔrsmZ7) and were ligated to Gm-, Km- or Sm-resistance cassettes [34] excised with the corresponding endonuclease. The generated plasmids were named pGEMΔrsmZ3-Gm, pGEMΔrsmZ4-Km, pGEMΔrsmZ5-Sm, pGEMΔrsmZ6-Gm and pGEMΔrsmZ7-Gm (Table 1). These plasmids, unable to replicate in A. vinelandii, were linearized by excision with ScaI restriction enzyme to ensure double homologous recombination event and allelic exchange. Afterwards, they were introduced into E strain by transformation. Double recombinants were selected on plates of BS medium amended with the corresponding antibiotic. The presence of the desired ΔrsmZ mutation and the absence of the corresponding wild-type rsmZ allele were verified by PCR (data not shown). The E derivative mutants carrying individual deletions of rsmZ3–7 genes were named EZ3 (ΔrsmZ3::Gm), EZ4 (ΔrsmZ4::Km), EZ5 (ΔrsmZ5::Sm), EZ6 (ΔrsmZ6::Gm) and EZ7 (ΔrsmZ7::Gm) (Table 1).

Construction of PrsmZ-gusA Transcriptional Fusions

Fragments containing promoter regions from rsmZ1, rsmZ2, rsmZ4, rsmZ5, rsmZ6, rsmZ7 (PrsmZ) were obtained from pSAHFUSTs-Z1–7 plasmids [15], as EcoR1 fragments, except for in rsmZ3, which was obtained by excision with XbaI. PrsmZ1–7 were individually cloned into the integrative vector pUMATcgusAT [35], generating PrsmZ-gusA transcriptional fusions (Tcr) for each gene. This vector directs the integration of the gusA fusion into the neutral melA locus [35]. The resulting plasmids (named pUMATcgusAT-rsmZ1 to -rsmZ7) were linearized with ScaI endonuclease and introduced by transformation into E strains. Double recombinants Tcr carrying the individual PrsmZ-gusA fusion integrated into the chromosome were isolated and confirmed by PCR analysis. The resulting strains were named EZ1T (PrsmZ1-gusA), EZ2T (PrsmZ2-gusA), EZ3T (PrsmZ3-gusA), EZ4T (PrsmZ4-gusA), EZ5T (PrsmZ5-gusA), EZ6T (PrsmZ6-gusA), EZ7T (PrsmZ7-gusA) (Table 1).

Construction of gyrA-gusA Transcriptional Fusion

A fragment of 130 pb containing the regulatory region of the A. vinelandii gyrA gene (PgyrA) (Avin_RS07245) was PCR amplified using primers DgyrXbaI and RgyrEcoRI, which carry XbaI and EcoRI artificial sites, respectively. The PgyrA was cloned into the pUMATcgusAT vector as an EcoRI-XbaI fragment, generating a PgyrA-gusA transcriptional fusion. This plasmid was made linear with ScaI endonuclease and was transformed into the E strain. Double recombinants Tcr were isolated and confirmed by PCR to carry the PgyrA-gusA transcriptional fusion into the chromosome. One representative transformant was chosen for further studies and was named EgyrAT (Table 1).

Analytical Methods

Alginate production was determined using spectrophotometric determination of the uronic acids with carbazole [36]. The synthesis of ARs was measure as reported previously [37]. β-glucuronidase activity was measured as reported by Wilson et al. [38]; 1U corresponds to 1 nmol of p-nitrophenyl-β-d-glucuronide hydrolyzed per min per mg of protein. Protein was determined by the Lowry method [39].

Results and Discussion

Transcriptional Profile of the rsmZ1-7 sRNAs Genes

In Pseudomonadaceae, it is common to find two or more Rsm-sRNAs of the RsmZ, RsmY or RsmX family. However, the presence in a bacterium of reiterated alleles of the same family is uncommon. The presence of multiple alleles of rsmX has been reported in P. syringae [10]. A. vinelandii is the only bacterium having multiple alleles of the rsmZ family; however, the physiological significance of such reiteration has not been explored. rsmZ1–7 genes are located in different regions of the chromosome; in silico analysis suggests that they are contained in monocistronic operons. The genetic arrangement of the rsmZ1–7 loci is shown in Figure S2.

In contrast to the DJ strain, the E strain (also named AEIV) is a wild-type isolate of A. vinelandii [24, 28]. In order to study the transcriptional regulation of the A. vinelandii rsmZ1–7 genes, we constructed a series of PrsmZ1–7-gusA transcriptional fusions, as described in Materials and Methods, and they were tested in the genetic background of the wild-type E strain. For this purpose, DNA fragments of approximately 2 kb containing the regulatory and structural region of each one of the rsmZ1–7 genes from the wild-type mucoid strain E were amplified and sequenced. The DNA sequence of PCR products showed from 98 to 100% identity with the corresponding chromosomal region of the DJ strain already sequenced [24]. For this reason, we used the rsmZ1–7-sRNAs regulatory regions derived from the DJ strain [15] to generate the PrsmZ-gusA transcriptional fusions to be tested in the E strain.

Strains carrying the PrsmZ1–7-gusA transcriptional fusions were cultivated in Burk’s Sucrose medium (BS) and showed similar growth kinetics (Fig. 1A), which allowed the comparative analysis among the different PrsmZ-gusA fusions. The transcription of every rsmZ-sRNA gene, indirectly estimated by the activities of the PrsmZ1–7-gusA fusion, showed a similar temporal pattern (Fig. 1b,c), reaching its maximum expression at the stationary growth phase (36 h). A similar fact has been reported in P. fluorescens and P. aeruginosa, in which rsmZ expression was delayed showing its greatest point at the stationary phase [3, 4]. Interestingly, in Pseudomonas spp. the expression of rsmY and rsmX correlates with the cell growth [4, 12].

Fig. 1
figure 1

Expression profiles of rsmZ1–7 sRNAs genes in A. vinelandii. (a). Growth kinetics of the strains carrying the individual rsmZ1–7-gusA transcriptional fusions (EZ1T–EZ7T). Promoter activity of the rsmZ1 (Z1), rsmZ3 (Z3) and rsmZ5 (Z5) (b) or rsmZ2 (Z2), rsmZ4 (Z4), rsmZ6 (Z6) and rsmZ7 (Z7) (c) genes, measured along the growth curve, using the rsmZ1–7-gusA transcriptional fusions. Cells were grown in Burk’s sucrose medium. Bars of standard deviation from three independent experiments (biological replicates) are shown

Full size image

Although all the rsmZ alleles showed a similar transcriptional pattern throughout the growth curve, the activity of the different rsmZ promoters was very variable. Those of the rsmZ1, rsmZ3 and rsmZ5 showed lower expression levels (Fig. 1B) when compared to the activity derived from rsmZ2, rsmZ4, rsmZ6 and rsmZ7 promoters (Fig. 1C). rsmZ1 showed the lowest expression pattern while rsmZ2 presented the highest expression throughout the growth curve.

Differences in the expression levels of the rsmZ alleles may reflect differences in their regulatory regions, including the location of the GacA-binding box. The regulatory regions of rsmZ alleles could be grouped in two sets: those with large regulatory regions (rsmZ1 and rsmZ2) and those with short regulatory regions (rsmZ3–7). Although the regulatory regions from rsmZ1 and rsmZ2 have similar lengths, their sequences are not conserved which may explain the observed differences in their expression (Fig. S3). In both cases, however, the sequences in between the GacA-binding box and the transcription start site have an unusual high AT content (A. vinelandii has a rich GC genome) suggesting the binding of regulators with affinity for AT-rich sequences, such as H–NS and IHF. Indeed, using the bioinformatics tools, Virtual Footprint-Prodoric (http://prodoric.tu-bs.de/vfp/vfp_promoter.php) and Softberry-BPROM (http://www.softberry.com/berry.phtml?topic=bprom&group=programs&subgroup=gfindb), we found putative IHF and H–NS binding sites in the regulatory region of rsmZ2 and rsmZ1, respectively (Fig. S3). Similar regulatory regions have been reported for the homologous rsmZ genes from Pseudomonas spp [13, 18]. In P. fluorescens, IHF binds to the regulatory region of rsmZ with high affinity [18], while in P. aeruginosa, H–NS binds to an AT-rich sequence in the rsmZ regulatory region repressing its expression [13]. Another factor that may influence the differential expression is the conservation degree of the palindromic GacA-binding box (TGTAAGNNATNNCTTACA) [27]. As shown in Figure S3, except for rsmZ1, the regulatory regions of all rsmZ genes showed conserved GacA boxes. Since rsmZ1 showed the lowest activity, it suggests that the conservation of the GacA box determines the transcriptional level of the individual rsmZ genes. However, as the expression of the rsmZ2–7 transcriptional fusions was highly variable, it also implies that GacA is necessary but not sufficient to establish a suitable expression of the rsmZ1–7-sRNAs genes.

The Type of the Carbon Source Influences the Transcription of the rsmZ-sRNAs

In E. coli, the Csr system controls carbon metabolism, and different types of carbon sources (glycolytic or gluconeogenic) modify the expression of csrB and csrC [40]. In P. fluorescens, the presence in the culture medium of Krebs cycle intermediaries, such as succinate or malate, promotes the rsm-sRNAs expression [22]. Therefore, we explored the effect of different types of carbon sources on the A. vinelandii rsmZ-sRNAs expression. For this purpose, the strains carrying rsmZ1–7-gusA transcriptional fusions were grown in Burk’s minimal medium amended with glucose (BGluc) or fructose (BFru), as glycolytic carbon sources, or with succinate (BSucc), as a gluconeogenic carbon source (Fig. 2). Interestingly, the expression of all the rsmZ genes was higher in the presence of succinate than in the glycolytic condition. Regardless of the tested carbon source, rsmZ2, rsmZ6 and rsmZ7 showed the highest expression levels while rsmZ1 presented the lowest expression (Fig. 2), a pattern previously observed in the presence of sucrose (Fig. 1b, c). In order to rule out a possible effect of the different carbon sources on the expression of the gusA reporter gene, a transcriptional PgyrA-gusA fusion was constructed as described in Materials and Methods. In previous reports, the expression of gyrA in A. vinelandii has been shown to be constitutive under different growing conditions [15, 41]. As expected, the expression of the PgyrA-gusA transcriptional fusion was constitutive and was not affected by the type of carbon source (Fig. S4). This result clearly indicated that the activity of the PrsmZ1–7-gusA constructions reflects the activity of the fused promoters.

Fig. 2
figure 2

Effect of glycolytic or gluconeogenic carbon sources on the rsmZ1–7 sRNAs expression. Promoter activity of the rsmZ1–7 genes was assessed using the transcriptional fusions of Fig. 1. Cells were grown in, Burk’s minimal medium amended with fructose (BFru), glucose (BGlu) or succinate (BSucc). Cells were grown to the stationary phase (48 h). Bars of standard deviation from three independent experiments (biological replicates) are shown

Full size image

The rsmZ-sRNA expression levels were different when the cells were cultivated in fructose or glucose, even though these glycolytic substrates could produce similar amounts of Krebs cycle metabolites. This result suggests the existence of an additional signal responsible for the increased rsmZ1–7 expression levels in the presence of fructose (Fig. 2). In A. vinelandii, there are clear differences in the uptake and metabolism of these hexoses; glucose uptake is carried out by the GluP transporter and is oxidized through the Entner–Doudoroff pathway [42, 43]. In the case of fructose, the presence of fruA (Avin_RS05545), fruB (Avin_RS05535) and fruK (Avin_RS05530) genes in the genome of several strains of A. vinelandii suggests that, as in Pseudomonas putida, fructose is assimilated through the PTSfru system that, upon conversion to fructose 1–6 bisphosphate, following the Embden–Meyerhof glycolytic pathway [44]. Hence, the differences in the RsmZ-sRNAs expression between fructose and glucose could be attributed to differences in the assimilation and metabolism of these sugars. Besides controlling sugar assimilation, the PTSFru system is involved in other regulatory processes [45, 46]. In P. putida, the PTSFru system is related to the PTSNtr (nitrogen-related PTS), a non-canonical PTS system reported in several Pseudomonadaceae that lack the subunit needed for carbohydrate transport, so this PTS system is not involved in carbohydrate assimilation. Early studies suggested that the PTSNtr system had a role in nitrogen metabolism; however, recent data indicated that it has several regulatory functions [35, 47,48,49]. In P. putida the two PTS branches are connected establishing a cross-talk process. A. vinelandii has a PTSNtr homologue system that controls the production of secondary metabolites such as PHB and ARs [50]. Therefore, fructose is likely to trigger a regulatory pathway that affects the rsmZ-sRNAs expression. On the other hand, fructose assimilation by the PTSFru generates fructose 1-P, which is the metabolic negative effector of FruR/Cra regulator. In E. coli, FruR/Cra controls a large number of genes related to carbon metabolism [51], a similar fact might occur in P. putida [52]. The A. vinelandii genome has a putative FruR/Cra homologue, which could be involved in the regulation of rsmZ-sRNAs genes in response to the metabolic state of the cell. To support this idea, a regulatory link between FruR and CsrA has been recently reported [53].

Mutations in rsmZ-sRNAs Genes Differentially Affect Alginate Production

The importance of the sRNAs RsmZ1 and RsmZ2 in the regulation of alginate production in A. vinelandii has been previously established [14]. In the absence of RsmZ1 or RsmZ2, alginate production diminished 70%. Therefore, we investigated the role of the RsmZ-sRNAs in alginate production in cells grown in the gluconeogenic and glycolytic carbon sources previously used. To do this, we used the individual rsmZ1 and rsmZ2 mutants previously constructed [14], and generated independent deletions of the sRNAs rsmZ3, rsmZ4, rsmZ5, rsmZ6 and rsmZ7 genes, as described in Materials and Methods. These mutants were constructed by deleting the rsmZ3–7 genes and inserting an antibiotic-resistance cassette. Given the monocistronic nature of the predicted operons containing the rsmZ1–7 genes (Fig. S2), it is unlikely that this strategy generates polar effects on downstream genes.

Mutants carrying individual deletions of the rsmZ1–7 genes were grown in Burk’s minimum medium in the presence of the different types of carbon sources and the production of alginate was determined (Fig. 3). Absence of any of these sRNAs did not abrogate alginate production. In glucose-grown cells, the sRNAs RsmZ1–5 were necessary for maximum alginate production. Interestingly, in cells grown in fructose the individual positive effect of these sRNAs was more accentuated than in glucose-grown cells (Fig. 3a,b), which might be associated with their observed enhanced expression in fructose (Fig. 2). The lack of RsmZ1 and RsmZ4 almost completely inhibited alginate production, whereas the absence of the remaining sRNAs reduced the levels of this polymer from 3 to 5 times.

Fig. 3
figure 3

Alginate production in rsmZ1-7 sRNAs mutants. The production of alginate was determined in cells of the wild-type strain E and in its derivatives carrying individual deletions of the rsmZ1-7 sRNAs (EZ1–EZ7). Cells were grown for 48 h (stationary phase) in Burk’s minimal media with fructose (a), glucose (b) or succinate (c) as the sole carbon source. Bars of standard deviation from three independent experiments (biological replicates) are shown

Full size image

The effect of the rsmZ-sRNAs mutations in alginate production was also tested in succinate as carbon source (Fig. 3c). Under this condition, the lack of RsmZ2 totally impaired alginate synthesis, which correlated with its high expression level observed in this carbon source (Fig. 2). Absence of RsmZ1 and RsmZ4 diminished the production of this polymer by 70 and 80%, respectively, which also implies an important role of these two sRNAs under this condition. Finally, the absence of RsmZ3, Z5, Z6 or Z7 only reduced the synthesis of alginate about 25% suggesting a minor role of these sRNAs in succinate.

Our results indicated that only RsmZ1, RsmZ2 and RsmZ4 had a consistent positive effect on alginate production in the presence of the three carbon sources tested (Fig. 3). However, it was not related to the extent of expression of these sRNAs. While rsmZ2 showed the highest expression levels, the expression of rsmZ1 was low in either glycolytic or gluconeogenic carbon sources (Fig. 2). Furthermore, expression of rsmZ6 and rsmZ7 was always higher than that of rsmZ1 but their effects on alginate synthesis were marked only in the presence of fructose (Fig. 3a), but moderated in succinate (Fig. 3c). Collectively, our data indicated the absence of a direct correlation between alginate production and the expression of the rsmZ-sRNAs. It suggests the existence of additional factors involved in the regulatory functions of the RsmZ-sRNAs. It is possible that the sRNAs have different affinities for the RsmA protein. In E. coli, the RsmZ functional homologues, CsrB and CrsC have important differences in their affinities for CsrA [16]. Stability of each rsmZ-sRNA is another factor to consider for understanding their functional differences. Additionally, the predicted ΔG values of the different A. vinelandii rsmZ-sRNAs rank from − 50.36 to 68.04 kcal / mol [27]. Enzymatic degradation is another factor that affects the sRNAs turnover. The RNase E cleavage sites present in the RsmZ-sRNAs from Pseudomonas spp. are conserved in the RsmZ1–7-sRNAs of A. vinelandii, suggesting that their accumulation is influenced by this process (Fig. S3). In P. fluorescens, the interaction of the Rsm-sRNA with the RsmA protein prevents its enzymatic degradation [54]. Similarly, in P. aeruginosa, the interaction of Hfq with RsmY, but not with RsmZ or RsmX, protects against RNase E [55]. The existence of these mechanisms controlling the turnover of the RsmZ-sRNA in A. vinelandii deserves further investigation.

Effect of the RsmZ1–7 sRNAs on the Production Of Alkylresorcinols

In A. vinelandii, ARs are other metabolites controlled by the Gac–Rsm system; these aromatic lipids are produced during the differentiation process leading to the formation of cysts resistant to desiccation. Cyst biogenesis could be induced in laboratory conditions with n-butanol as the sole carbon source. In strain SW, an rsmZ1 mutant showed a strong reduction in the transcription of arsA, a biosynthetic ARs gene, suppressing the production of these lipids. The strain SW is a derivative of the non-mucoid strain DJ, which carries a natural IS insertion within algU, encoding a stress sigma factor essential for alginate production [24, 56]. The strain SW was generated by genetic complementation of the DJ strain with a wild-type copy of algU integrated into its chromosome [57].

Since n-butanol triggered the accumulation of ARs in an RsmZ1-dependent manner, the possible link between this carbon source and the expression of the RsmZ1–7 sRNAs was evaluated in the wild-type strain E. For this reason, expression levels of each sRNA were determined in cells grown in liquid Burk’s medium with n-butanol as the sole carbon source, for 5 days (Fig. 4a). As in other carbon sources, the expression of the RsmZ-sRNAs varied greatly, rsmZ2 showed the maximum expression and rsmZ1 and rsmZ5 showed the lowest expression levels. The transcriptional levels of rsmZ3, rsmZ4, rsmZ6 and rsmZ7 were somewhat similar.

Fig. 4
figure 4

Expression of the rsmZ1-7 sRNAs genes in the presence of n-butanol and their effect on alkylresorcinols production. a The promoter activity of rsmZ1-7genes was assessed using the transcriptional fusions of Fig. 1. Cells were grown for 5 days in Burk’s minimal media amended with n-butanol as the sole carbon source. b The effect of each of the RsmZ-sRNAs on alkylresorcinols (ARs) accumulation was determined after 5 days of growth in liquid Burk’s-butanol media. Bars of standard deviation from three independent experiments (biological replicates) are shown

Full size image

Later, the effect of each sRNA in ARs production was determined. In contrast to the pronounced effect shown by RsmZ1 on ARs production in the SW strain, absence of RsmZ1 in the E strain caused a reduction of 70%. Interestingly, absence of RsmZ2, RsmZ4 or RsmZ5 resulted in a marked reduction of about 83 to 95% in ARs accumulation. It is worth noting that although rsmZ6 and rsmZ7 expressions were higher than that of rsmZ1, absence of RsmZ6 or RsmZ7 did not reduce the synthesis of ARs and a significant increase was observed in mutant rsmZ6 (Fig. 4b).

Altogether, our results indicated that the expression levels of the rsmZ1-7-sRNAs in ARs accumulating conditions did not correlate with their individual effect on the synthesis of these lipids.

Concluding Remarks

A. vinelandii has the uncommon characteristic of harbouring large numbers of highly similar carbohydrate metabolism homologues, which are proposed to confer adaptive benefits with respect to certain environmental factors and life styles [58]. The presence of multiple sRNAs of the RsmZ family in A. vinelandii, might be related to this trait, which may provide a versatile adaptability through a highly flexible Rsm system.

We found no correlation between the extent of the transcription of the individual RsmZ-sRNA and its corresponding effect on the production of either alginate or ARs. Our data revealed that additional factors, besides to the transcriptional control, influence the different regulatory roles of the RsmZ-sRNAs. These factors could respond to diverse stimuli, signals or conditions generating a better and versatile response of the Rsm system.