|Home | About | Journals | Submit | Contact Us | Français|
Riboregulation has a major role in the fine-tuning of multiple bacterial processes. Among the RNA players, trans-encoded untranslated small RNAs (sRNAs) regulate complex metabolic networks by tuning expression from multiple target genes in response to numerous signals. In Sinorhizobium meliloti, over 400 sRNAs are expressed under different stimuli. The sRNA MmgR (standing for Makes more granules Regulator) has been of particular interest to us since its sequence and structure are highly conserved among the alphaproteobacteria and its expression is regulated by the amount and quality of the bacterium's available nitrogen source. In this work, we explored the biological role of MmgR in S. meliloti 2011 by characterizing the effect of a deletion of the internal conserved core of mmgR (mmgRΔ33–51). This mutation resulted in larger amounts of polyhydroxybutyrate (PHB) distributed into more intracellular granules than are found in the wild-type strain. This phenotype was expressed upon cessation of balanced growth owing to nitrogen depletion in the presence of surplus carbon (i.e., at a carbon/nitrogen molar ratio greater than 10). The normal PHB accumulation was complemented with a wild-type mmgR copy but not with unrelated sRNA genes. Furthermore, the expression of mmgR limited PHB accumulation in the wild type, regardless of the magnitude of the C surplus. Quantitative proteomic profiling and quantitative reverse transcription-PCR (qRT-PCR) revealed that the absence of MmgR results in a posttranscriptional overexpression of both PHB phasin proteins (PhaP1 and PhaP2). Together, our results indicate that the widely conserved alphaproteobacterial MmgR sRNA fine-tunes the regulation of PHB storage in S. meliloti.
IMPORTANCE High-throughput RNA sequencing has recently uncovered an overwhelming number of trans-encoded small RNAs (sRNAs) in diverse prokaryotes. In the nitrogen-fixing alphaproteobacterial symbiont of alfalfa root nodules Sinorhizobium meliloti, only four out of hundreds of identified sRNA genes have been functionally characterized. Thus, uncovering the biological role of sRNAs currently represents a major issue and one that is particularly challenging because of the usually subtle quantitative regulation contributed by most characterized sRNAs. Here, we have characterized the function of the broadly conserved alphaproteobacterial sRNA gene mmgR in S. meliloti. Our results strongly suggest that mmgR encodes a negative regulator of the accumulation of polyhydroxybutyrate, the major carbon and reducing power storage polymer in S. meliloti cells growing under conditions of C/N overbalance.
Rhizobia are alpha- and betaproteobacteria capable of establishing symbiotic interactions with leguminous plant roots, where they induce a de novo formation of root organs called nodules and after colonization subsequently differentiate into bacteroid cells devoted to carrying out biological nitrogen fixation (BNF) (1, 2). Sinorhizobium meliloti is a soil-dwelling alphaproteobacterium able to establish N2-fixing root nodule symbiosis with legumes from the genera Medicago, Melilotus, and Trigonella (3, 4). The beneficial association between S. meliloti and alfalfa (Medicago sativa) plant roots represents a reference model for studying the interactions between symbiotic bacteria and the plant host (5). This partnership involves a complex signal exchange between the symbionts that commits them to a joint differentiation process involving biochemical and morphological changes that terminate in the formation of the mature N2-fixing root nodule (5,–7). This remarkably complex interplay has been shown to be tightly regulated at multiple levels in both partners (5, 6).
Extracellular carbon polymers like succinoglycan and galactoglucan are well-known critical elements involved in the root hair infection process, whereas the surface lipopolysaccharide (LPS) is required for proper differentiation of S. meliloti into nitrogen-fixing bacteroids within root nodule cells (5). S. meliloti can also accumulate intracellular carbon polymers such as glycogen and polyhydroxybutyrate (PHB), notably under growth-limiting conditions (8). PHB appears to be critical for N2 fixation in young developing M. sativa nodules, and glycogen synthesis is essential for N2 fixation in mature M. sativa nodules (9). In free-living S. meliloti cells, PHB accumulates when C availability exceeds that of other major nutrients such as N (i.e., at a C/N molar ratio greater than 10) (8) and/or when the cells have excess reducing capacity and need to regenerate NAD(P)H (10). Thus, PHB is a storage compound for both C and reducing power that also confers long-term persistence under starvation conditions (11). In contrast, the role of glycogen in free-living cells remains obscure. The carbon flux between the two intracellular C storage compounds of S. meliloti seems to be under transcriptional control by the product of the aniA gene (12).
Gene expression in bacteria has been historically associated almost exclusively with the activity of protein regulators that switch on or off transcription; however, during the last 2 decades extensive research has highlighted the roles that untranslated trans-encoded small RNA molecules (sRNAs) play as regulators of diverse physiological processes, e.g., cell cycle control, amino acid uptake and metabolism, quorum sensing, iron metabolism and the osmotic stress response, among others, mostly at the posttranscriptional level (13, 14).
S. meliloti has been shown to express more than 400 different sRNAs under different growth conditions and lifestyles, i.e., both free-living in the soil and symbiotically associated in a plant (15,–21). In addition, S. meliloti expresses a functional homolog of the Escherichia coli Hfq protein, a chaperone of RNA-RNA interactions and a modulator of RNA activity (22,–29). Nevertheless, the biological implications of the large S. meliloti RNome remain largely unexplored (20), since only four small RNAs have been functionally characterized so far, i.e., the cell cycle regulator sRNA EcpR1 (30), the quorum-sensing regulator sRNA RcsR1 (31), and the tandemly encoded orthologs AbcR1 and AbcR2 (32). We are tempted to speculate that the large regulatory network is yet to be revealed.
In this work, we aimed at elucidating the biological role of a particular trans-encoded sRNA, which we have originally identified as a transcript of unknown function encoded in the SMc04042-SMc04043 intergenic region of S. meliloti 2011 (19). For reasons that will become clear below, we have renamed this sRNA gene mmgR (it was formerly designated sm8). The encoded transcript MmgR belongs to the subfamily of orthologous sRNAs αr8s1, which is widely distributed within the alphaproteobacteria and whose members bear a highly conserved sequence and structure in the internal nucleotide core and share microsynteny with their flanking genes (33). The conservation implies there may be a shared, conserved function for mmgR homologs; however, to date none have been characterized in other alphaproteobacteria. MmgR is a noncoding 77-nucleotide (nt) transcript (16, 17, 19) that binds to and is stabilized by Hfq (25, 27). Under symbiotic conditions, MmgR transcripts have been found to be more abundant in the N2-fixing mature bacteroids of nodule zone III (21). In free-living S. meliloti cells, MmgR expression is modulated in response to the quality and amount of the available N source, reaching the highest intracellular level with nitrate as the N source or upon starvation of the organic N sources (34). Here, we report that S. meliloti MmgR sRNA acts as a fine-tuning negative regulator of polyhydroxybutyrate (PHB) storage under conditions of N starvation and C surplus.
Our objective was to generate an S. meliloti 2011 isogenic mutant strain lacking the MmgR transcript. For that purpose, the region carrying mmgR was first inspected to prevent introducing nucleotide changes that could affect the expression of the flanking genes, namely, SMc04042 and SMc04043. The recent annotation of transcriptional start sites in the S. meliloti 2011 genome (16) allowed us to delimit a safe region in which the risk of introducing polar mutations was minimized. Unfortunately, the complete deletion of the mmgR gene was not possible because of its partial overlap with the farthest promoter elements of the upstream and divergently transcribed gene SMc04042; we instead targeted an internal and conserved 18-nt core sequence of mmgR in order to perform an allelic exchange with a 12-nt unrelated sequence (Fig. 1a). The mutant transcript MmgRΔ33–51 was predicted to have a modified RNA secondary structure but to retain the Rho-independent terminator to ensure transcription termination (Fig. 1b and andc).c). Expression of the mutant sRNA was analyzed during the stationary phase of growth in Rhizobium defined medium (RDM), a condition previously reported to induce the highest intracellular MmgR levels in the wild-type strain (19, 25). The cellular level of MmgRΔ33–51 RNA in the mmgRΔ33–51 mutant was below the detection limit of a Northern blot assay targeting the 5′ sequence shared in common with the wild-type sRNA (see Fig. 3c). Quantitative reverse transcription-PCR (qRT-PCR)-based expression analysis of the flanking gene SMc04042 demonstrated that the introduction of the mmgRΔ33–51 mutation did not significantly influence the abundance of SMc04042 mRNA [log2 (abundance in mmgRΔ33–51/abundance in wild type) = −0.4 ± 0.1], thus eliminating the possibility of a polarity in the mmgRΔ33–51 allele over the expression of SMc04042. Together, these results validated the usefulness of the S. meliloti 2011 isogenic mmgRΔ33–51 strain in studying the effect caused by lowering the MmgR transcript level.
Initially, we compared the growth behaviors (optical density at 600 nm [OD600]) of the wild-type and mmgRΔ33–51 mutant strains in RDM (C/N molar ratio = 30:1). Although the strains displayed a similar growth performance during the exponential phase, the mutant strain was associated with a higher OD600 than the wild type after leaving the balanced-growth phase (Fig. 2a). However, we detected no differences in terms of viable cell counts between the wild-type and mmgRΔ33–51 strains (Fig. 2a). As the OD600 measurements were done after washing and resuspending the cells in saline solution, these results ruled out the possibility of interference by optically active substances in the culture supernatant and furthermore suggested that the mutation in mmgR might cause changes in cell biomass that would be reflected in the OD600 values of the mmgRΔ33–51 culture (35). As expected for a phenotype resulting primarily from changes in intracellular MmgR activity, the differential behavior in terms of OD600 progression between the strains became evident at the point in the growth curve when the mmgR promoter was naturally turned on in the wild-type background (i.e., at the end of the exponential phase of growth), which in turn leads to a large increase in the cellular sRNA transcript level, as we have reported elsewhere (34).
In order to explore the cellular basis of the observed differential OD600 yield after exit from the exponential growth phase (Fig. 2a), we determined the cell dry weight of the wild-type S. meliloti 2011 and mmgRΔ33–51 strains in the stationary phase of growth in RDM. The mmgRΔ33–51 mutant cells were, on an average, 20% heavier than those of the wild-type bacteria (Fig. 3a), whereas no significant differences in the protein content per cell of either strain were observed under the same physiological condition (Fig. 3a). This difference suggested that the higher biomass of the mmgRΔ33–51 mutant might be related to a disproportionate accumulation of some storage compound that was not accompanied by a balanced synthesis of the rest of the cellular components.
Considering that the cessation of the exponential phase of growth represents a metabolic status in which synthesis of the C storage polymer PHB had been described to be turned on in rhizobia (8), we explored whether or not a differential accumulation of PHB could be responsible for the observed differences in the average cellular biomass of the two strains under that growth condition. The results indicated that the mmgRΔ33–51 strain did, in fact, accumulate over 20% more PHB than the wild-type strain (Fig. 3a). Furthermore, despite the difference between the geometric mean values for the cellular PHB content of each population, both distributions were unimodal, thus eliminating the possibility of a phenotypic heterogeneity (Fig. 3d; see Fig. S2 in the supplemental material). No difference in the amount of intracellular glycogen as a form of C storage was detected as a result of the mutation of mmgR (Fig. 3a).
A genetic complementation of the mmgRΔ33–51 mutant strain through expression of the MmgR sRNA from an inducible copy carried in plasmid pSRK-MmgR (Fig. 3c; see Fig. S1 in the supplemental material) fully restored the wild-type behavior (Fig. 2b and and3b).3b). In order to mimic the expression pattern of the mmgR gene in the wild-type strain (34) during the mutant complementation assay, a plasmid-borne copy of the mmgR locus was induced with 500 μM isopropyl-β-d-thiogalactopyranoside (IPTG) once the cultures reached an OD600 of 1.6 (Fig. 2b). As expected, neither overexpression of either the Hfq-independent unrelated antisense RNA asRNA812 or the Hfq-dependent sRNA Sm84 (27) nor induction of the mutant allele mmgRΔ33–51 restored the wild-type levels of PHB during the stationary phase within the mmgRΔ33–51 genetic background (Fig. 3b).
To test the hypothesis that the role of MmgR is to control utilization of the C surplus upon the cessation of balanced growth, we evaluated the effect of doubling the C availability in the culture medium on biomass production by the wild-type S. meliloti 2011 and the isogenic mmgRΔ33–51 mutant. In the wild-type strain, the higher C availability in the culture medium (C/N ratio, 60:1) did not result in an increased accumulation of biomass, whereas the mmgRΔ33–51 cultures exhibited an even higher biomass production than had been observed during growth in RDM (C/N ratio, 30:1), as revealed by their higher OD600 in the stationary phase (Fig. 4a). Measurements of the amounts of intracellular PHB in stationary wild-type and mmgRΔ33–51 bacteria under these growth conditions thus further supported the existence of an uncontrolled accumulation of PHB with an increased C availability in the absence of the MmgR gene product (Fig. 4b).
The observation of stationary-phase wild-type and mmgRΔ33–51 cells growing in RDM by transmission electronic microscopy (TEM) revealed profound morphological and ultrastructural differences between the strains. First, mmgRΔ33–51 cells presented a statistically significant 20% higher length/width ratio (2.33 ± 0.12; n = 122; P < 0.05) than the wild-type bacteria (1.91 ± 0.05; n = 82) (Fig. 5). In addition, the mutant cells displayed a higher number of smaller and morphologically heterogeneous cytoplasmic PHB granules per cell than the wild type (with 3 to 6 well-defined granules per cell on average) (Fig. 5). As expected, exponential-phase cells growing in RDM did not show any such PHB granules in their cytoplasm (Fig. 5).
With an aim of shedding light on the molecular changes underlying the physiological phenotype of the mmgRΔ33–51 mutant, we initially performed a semiquantitative differential proteomic profiling of the wild-type and mmgRΔ33–51 strains at stationary phase in RDM. Total protein extracts from each strain were prepared, and the proteins were separated by 15% (wt/vol) SDS-PAGE. A single band corresponding to low-molecular-weight proteins was clearly overrepresented in the mutant proteome and was therefore excised for protein identification (Fig. 6). The PHB granule-associated protein PhaP2 (SMc02111) was identified as the major protein component of the material extracted from the gel (36).
In order to further test the molecular findings reported above, we carried out quantitative and comparative proteomic profiling of the wild-type and mutant strains during the stationary phase of growth in RDM by mass spectrometry after differential metabolic labeling of cellular proteins with a heavy (15N) or light (14N) isotope during growth of the bacteria on medium containing either 15NH4Cl or 14NH4Cl, respectively, as the sole nitrogen source. This approach allowed us to confirm in quantitative terms the overaccumulation of the PHB phasin PhaP2 in the mutant strain [i.e., log2 (mmgRΔ33–51/wild type) = 1.85 ± 0.77] and to further detect similar changes in the relative abundance of the other chromosomally encoded PHB-phasin, PhaP1 (SMc00777) [i.e., log2 (mmgRΔ33–51/wild type) = 2.10 ± 0.28]. Both proteins are much more abundant in the mmgRΔ33–51 mutant than in the wild-type strain (Fig. 7). The inability to detect PhaP1 by SDS-PAGE might be because this phasin possibly reaches lower cellular amounts than the PhaP2. No other proteomic changes of the same relevance and with an evident functional relationship to PHB metabolism were detected by this quantitative approach. Other polypeptides that accumulated slightly in the mmgR mutant [i.e., log2 (mmgRΔ33–51/wild type) > 1.0] were SMb21630 (a conserved hypothetical protein), SoxA2 (a putative sarcosine-oxidase-alpha subunit), FbpA (an Fe3+ ABC transporter), and SMc01140 (a probable σ54 modulation protein), whereas the single repressed protein in the mmgR mutant strain was NrtA [the periplasmic nitrate binding protein of a nitrate transporter; log2 (mmgRΔ33–51/wild type) = −1.27].
In order to determine whether the differential accumulation of phasins was the result of primary changes at the transcriptional or translational level, the corresponding mRNA concentrations were measured by qRT-PCR assays under the same growth conditions in which PhaP1 and PhaP2 were upregulated in the mutant strain. There were no significant differences between the two strains in the relative abundance of the two mRNAs (Fig. 7). These overall results confirm that MmgR negatively controls, in a direct or indirect manner, the expression of the phasin genes at a posttranscriptional level.
Under conditions of aerobic growth with glucose as the sole C source, S. meliloti diverts approximately equal amounts of C to the biosynthesis of balanced cellular biomass in anabolism and to oxidation to CO2 for energy in catabolism (37). Since the average minimal formula of bacterial biomass is approximated by C5H9O2.5N (38), a total C/N molar ratio of 10:1 in a culture medium is expected to be enough to support balanced growth and respiration (i.e., 5 mol of C out of a total of 10 for each mole of N to fuel biomass synthesis, plus another 5 mol for respiration and energy production). In contrast, under an overbalanced C/N ratio (>10:1), the C surplus that is present at the end of the exponential and balanced growth would be expected to be utilized in the synthesis of internal storage compounds, e.g., PHB or glycogen, or of exopolysaccharides that are excreted into the medium (8). The fate of C under such conditions is determined mainly by the current physiological state of the bacterium and as such is subjected to stringent regulation in order to maximize bacterial fitness and well-being (39). Our results demonstrated that when the mmgRΔ33–51 strain (lacking the MmgR transcript) (Fig. 1) grows in RDM with a C/N overbalance of 30:1, that strain produces more biomass and accumulates higher levels of PHB than does the wild type (Fig. 3). This difference suggests that the MmgR transcript negatively controls PHB storage. That the mmgR promoter is sharply induced upon cessation of balanced growth because of N depletion regardless of the presence of a C surplus (see Fig. S3 in the supplemental material) suggests that the physiological stimulus promoting mmgR expression does not require active PHB synthesis; rather, once mmgR is turned on and MmgR accumulates, the fine-tuning capacity of the sRNA becomes manifest under conditions of C/N overbalance so as to regulate PHB synthesis (Fig. 2 to to4).4). Moreover, we observed that MmgR was able to limit PHB accumulation in the wild-type strain, irrespective of the magnitude of the excess of C in the growth medium (Fig. 4), implying that the MmgR constitutes a tight regulator of PHB storage that serves to set a maximum level of C channeling into the biosynthesis of PHB. Accordingly, the physiological findings were confirmed through direct observation of cells by transmission electron microscopy. The sole absence of MmgR activity in S. meliloti had a profound impact on cell morphology during the stationary phase owing to the accumulation of an abnormally high number of irregularly shaped PHB granules (Fig. 5). The higher surface/volume ratio associated with this change in granule number and size distribution suggests that the elevated content of PHB polymer is directly correlated with an increased concentration of phasins that mediate stabilization of the granule-cytoplasm interphase (40).
Complementation with the wild-type mmgR allele in trans fully restored the growth phenotype (OD600) and PHB content of the mmgRΔ33–51 mutant strain (Fig. 2 and and3).3). In contrast, induction of a plasmid-borne mmgRΔ33–51 allele resulted in undetectable levels of the MmgRΔ33–51 transcript (Fig. 3c), thus suggesting that the internal conserved sequence core of MmgR (Fig. 1) is critical for its stability and therefore for its biological activity. As described elsewhere (41), Hfq is a limiting factor for riboregulatory circuits in bacteria because of competition among sRNAs for accessing Hfq binding sites. To examine the possibility of a side effect of ectopic sRNA induction on the availability of Hfq, we effected overexpression of another Hfq binding sRNA transcript (i.e., Sm84 [25, 27]) and of a small Hfq-independent transcript (asRNA812) (17). Induction of either the Sm84 or the asRNA812 sRNA, however, failed to restore the wild-type ability to limit PHB storage in the mmgRΔ33–51 mutant background under the conditions assayed (Fig. 3b). In conclusion, the inability of all RNAs other than the wild-type mmgR to complement the mmgRΔ33–51 mutation demonstrated that the phenotypes associated with the artificial expression of MmgR are specific for this riboregulator and that nonspecific Hfq titration effects could be ruled out. In terms of the symbiotic interaction with the host plant Medicago sativa (alfalfa), the mmgRΔ33–51 mutant was Nod+ Fix+ and the symbiotic phenotype was indistinguishable from that of the wild-type parental strain (data not shown). This phenotypic equivalence implies that the loss of negative control over PHB production associated with the lack of MmgR sRNA does not introduce major changes in the symbiotic performance of S. meliloti.
S. meliloti has the complete enzymatic machinery to synthesize and degrade intracellular PHB (42). Upon N starvation over a background of C surplus, the production of PHB is turned on as a way to store C and reducing power (42). The production of PHB after growth arrest under saprophytic conditions has been shown to improve the fitness and survival of rhizobial populations during prolonged starvation (43). In addition to the role of PHB as a C storage compound, the possibility of storing reducing power in the form of PHB has been proposed to serve as a buffer to balance the cell's redox state (42). The regulation of PHB accumulation in bacteria has been reported to take place at the transcriptional or enzymatic level or at a combination of both, depending on the species (44). In S. meliloti, in addition to the biosynthetic genes phbAB and phbC and to phaZ, the locus encoding the depolymerizing enzyme (42), the formation of PHB granules requires the specific synthesis of the granule-associated phasins PhaP1 and PhaP2 (36), whereas the regulatory protein AniA (PhaR) has been reported to control the flux of C into the alternative polymeric by-products PHB, glycogen, and extracellular polysaccharide (EPS) (12). We have not detected significant differences in the production of extracellular polysaccharide or glycogen (Fig. 3a) in combination with the overproduction of PHB in the mmgRΔ33–51 mutant. A quantitative profiling of the proteomic changes that underlie the mmgRΔ33–51 mutant phenotype during the stationary phase of growth in RDM revealed higher cellular contents of the proteins PhaP1 and PhaP2 in the mmgRΔ33–51 strain (Fig. 6 and and7).7). The increased abundance of the PhaP1/P2 polypeptides was not, however, accompanied by a comparable increase in their mRNA transcripts (Fig. 7). This finding strongly suggests that an MmgR-dependent posttranscriptional regulatory mechanism operates, either directly or indirectly, on the phaP1/P2 genes. In silico analyses predicted energetically favorable RNA-RNA interactions around the ribosome binding site of both phasin mRNAs and the highly conserved core of the MmgR sRNA (not shown); nevertheless, we cannot rule out the possibility that MmgR targets other mRNAs involved in controlling the flux of C into the synthesis of PHB.
Despite the reported positive role of PHB in multiple metabolic aspects related to the cellular needs for C and energy, it seems reasonable that the accumulation of the polymer and its associated granule proteins has to be limited. Nevertheless, neither the mechanisms nor the regulatory signals by which free-living wild-type S. meliloti cells manage to carefully accumulate controlled amounts of PHB within a defined granule architecture have been elucidated. Our results tempt us to speculate that MmgR might be part of a regulatory system that operates to maintain a proper structure and amount of PHB granules through a fine-tuning of the intracellular levels of phasins and polymer, on the basis of the availability of N and C (the latter relationship is supported by recent and as-yet-unpublished findings from our group). As a result of the mutation in mmgR, the loss of that fine adjustment might impair general rhizobial fitness. The negative effects of such a mutation should be investigated under conditions that simulate the natural environment that rhizobia inhabit, which may impose a natural form of selective pressure to preserve this posttranscriptional regulatory mechanism. The results detailed here create new possibilities for a better understanding of the biological significance of PHB dynamics in rhizobia (both in the free-living state and in symbiosis).
The wide distribution of the capacity to synthesize polyhydroxyalkanoates (PHA) observed among eubacterial and archaeal species, along with comparative sequence analyses, has provided extensive evidence for a horizontal genetic flow of structural PHA genes and their corresponding transcriptional regulators (45, 46). The posttranscriptional regulation of PHA (including PHB), however, has been reported in only a single species. In the gammaproteobacterium Azotobacter vinelandii, PHB synthesis was previously demonstrated to be under the control of the posttranscriptional regulatory cascade Gac/Rsm (47) and also negatively regulated by the activity of the iron-responsive small RNA ArrF (48,–50). The S. meliloti genomes, however, lack genetic elements of the Gac/Rsm type (51), and there is no evidence for sequence or structural homology between A. vinelandii ArrR and S. meliloti MmgR RNAs or their flanking genes. These observations suggest that these two unrelated sRNAs have evolutionarily converged into a common regulation of the same cellular process.
To summarize, our results have demonstrated that the transcript encoded in the SMc04042-SMc04043 intergenic region of S. meliloti 2011 (originally referred to as sm8) (Fig. 1) negatively regulates the net metabolic flux of C into the accumulation of the storage polymer PHB. For this reason, we renamed the sm8 sRNA gene mmgR, i.e., a regulatory RNA whose mutation results in a strain that makes more PHB granules. In most instances, these regulatory RNAs act to fine-tune cellular processes and so exhibit mild phenotypes after mutation, thus hindering their functional characterization by reverse genetics (52). In this regard, we would like to highlight the unusual observation that a clear-cut phenotype could be identified following the mutation of the S. meliloti mmgR gene and that complementation experiments conclusively proved that the changed phenotype resulted from an alteration in the activity of the MmgR transcript.
As we have reported recently (33), the occurrence of mmgR alleles in alphaproteobacteria correlates well with the capacity to synthesize and store PHB within this bacterial group, thus suggesting a fundamental coevolutionary relationship between the mmgR locus and PHB metabolism. Further studies on the functional adaptation of each mmgR ortholog to the biology of its corresponding species will help us to understand more comprehensively the biological role within the phylogeny of one of the most ancient and widely distributed sRNA genes in alphaproteobacteria.
The bacterial strains used in this work are listed in Table 1. Escherichia coli strains were cultured at 37°C either in nutrient yeast broth (NYB) (nutrient broth, 25 g per liter; yeast extract, 5 g per liter) or in nutritive agar (blood agar base, 40 g per liter; yeast extract, 5 g per liter). S. meliloti strains were cultured at 28°C either in tryptone yeast complex medium (TY) (tryptone, 5 g per liter; yeast extract, 3 g per liter; CaCl2 · 2H2O, 0.7 g per liter ) or in Rhizobium defined medium (RDM) (sucrose, 5 g per liter; MgSO4 · 7H2O, 0.25 g per liter; NH4Cl, 0.32 g per liter; CaCl2 · 2H2O, 100 mg per liter; anhydrous FeCl3, 6 mg per liter; H3BO3, 3 mg per liter; MnSO4 · H2O, 1.7 mg per liter; ZnSO4 · 7H2O, 0.3 mg per liter; NaMoO4 · 2H2O, 0.12 mg per liter; CoCl2 · 6H2O, 0.065 mg per liter; K2HPO4, 1 g per liter; KH2PO4, 1 g per liter; biotin, 1 mg per liter; thiamine, 10 mg per liter [adapted from reference 54]). When appropriate, antibiotics were added to culture media at the following concentrations: for E. coli, kanamycin at 25 μg per ml and tetracycline at 10 μg per ml and for S. meliloti, streptomycin at 400 μg per ml, neomycin at 120 μg per ml, and tetracycline at 5 μg per ml. Bacterial growth was estimated by monitoring the optical density at 600 nm of culture dilutions made after washing and resuspending cells in an appropriate volume of saline solution (SS) (NaCl, 0.9% [wt/vol]) in order to reach an OD of 0.2 to 0.8. At certain time points during the generation of the curves, viable cell counts were determined (55, 56). The growth assays were repeated at least twice.
Table 1 summarizes the oligonucleotides and plasmids used in this study. The S. meliloti 2011 isogenic mmgR mutant (mmgRΔ33–51) was generated by introduction of a sequence replacement within the conserved core of the mmgR gene. Eighteen nucleotides (positions 3046287 to 3046363 of the S. meliloti strain 2011 chromosome) were replaced by a DNA fragment containing KpnI and SmaI restriction site sequences (5′-GGTACCCCCGGG-3′) via homologous recombination and allelic exchange (Fig. 1). To this end, a construct carrying the mmgR mutant allele flanked by the corresponding 600-nt DNA fragments located upstream and downstream in the chromosomal mmgR target sequence (GenScript, USA) was synthesized and cloned into the pK18mob:sacB vector. The plasmid was transformed into E. coli S17-1 and transferred to strain 2011 by conjugation. Single-crossover neomycin-resistant mutants were first selected and further plated onto TY agar supplemented with 5% sucrose. Double-crossover strains were selected by checking the loss of sucrose sensitivity and neomycin resistance. The correct allelic exchange in the mmgRΔ33–51 candidate mutants was verified by PCR amplification of the mmgR locus with the extHRmmgRFw/extHRmmgRRv primer pair (listed in Table 1), followed by detection of the corresponding restriction products upon amplicon treatment with KpnI. The mmgRΔ33–51 mutant clones were finally confirmed by sequencing the mmgR locus (Macrogen Inc., South Korea).
MmgR, MmgRΔ33–51, the Hfq-independent asRNA812, or the Hfq-dependent Sm84 (17, 27) sRNA was induced in the mmgRΔ33–51 background by an isopropyl-β-d-thiogalactopyranoside (IPTG)-inducible plasmid expression system based on the sinR-sinI regulation that has been previously introduced elsewhere (see Fig. S1 in the supplemental material) (30, 57). In brief, a PCR-generated DNA fragment with the complete sequence of the S. meliloti 2011 sinR gene followed by the sinR-sinI inter-open reading frame (ORF) region was fused to a PCR-amplified DNA fragment containing the sequence of MmgR, MmgRΔ33–51, or Sm84. Each resulting fragment was cloned into the pSRK-Km vector to obtain the respective plasmids pSRK-MmgR, pSRK-MmgRΔ33–51, and pSRK-Sm84. Plasmid pSRK-asRNA812 (30) was constructed following the same protocol described above. The sRNA expression plasmids were transferred to the S. meliloti double mutant strain 2011RI (sinR sinI) carrying the mmgRΔ33–51 allele in order to reduce background expression from the chromosomal sinI copy upon overexpression of the plasmid-borne sinR copy. The resulting strains were named 2011RI mmgRΔ33–51 p-MmgR, 2011RI mmgRΔ33–51 p-MmgRΔ33–51, 2011RI mmgRΔ33–51 p-as812, and 2011RI mmgRΔ33–51 p-Sm84 (Table 1).
Total RNA from the bacterial cells was extracted with acid phenol-guanidinium isothiocyanate (TRIzol, Life Sciences) and chloroform, following the manufacturer's instructions. The RNA was then purified by precipitation with isopropanol. Before reverse transcription, the RNA was treated with DNase I for 1 h at 37°C (Thermo Scientific; 1 U DNase I per μg RNA). DNase I was then inactivated by incubation at 65°C after the addition of 0.1 volume of 50 mM EDTA. The purified RNA was then quantified by UV absorbance (NanoDrop; Thermo Scientific, USA) and the quality of the preparation further assessed by denaturing agarose gel electrophoresis (58).
Northern blot analyses were performed as reported elsewhere (58). Two to five micrograms of total RNA from each sample was initially electrophoresed for 45 min at a constant current (15 mA) in polyacrylamide gels (8.3 M urea, 8% [wt/vol] acrylamide, 0.2% [wt/vol] bisacrylamide in 1× Tris-borate-EDTA [TBE] buffer). With the low-range RNA ladder (Thermo Scientific, USA) serving as a molecular weight marker, the corresponding lane was cut and stained separately with ethidium bromide and the image registered with a UV transilluminator. The remaining gel was electroblotted at 150 mA for 30 min onto a Hybond-N membrane in 1× TBE buffer. After a two washes of the membrane with 2× SSC solution (30 mM sodium citrate, 0.3 M NaCl), the RNA was cross-linked to the membrane by exposure to UV light for 5 min. The membranes were then blocked with prehybridization buffer (50% [wt/vol] formamide, 5× SSC, 50 mM phosphate buffer [pH 7.0], 2% [wt/vol] blocking reagent, 0.1% [wt/vol] N-laurylsarcosine, 7% [wt/vol] sodium dodecyl sulfate [SDS]) for 1 h at 43°C in a hybridization oven and then incubated overnight at 43°C with the hybridization buffer containing the specific anti-mmgR1–32 digoxigenin-labeled double-stranded DNA (dsDNA) probe (previously generated by amplification of the corresponding genomic locus with the mmgR1–32Fw and mmgR1–32Rv primers [Table 1]). The hybridized membranes were washed under standard stringent conditions, incubated with an alkaline phosphatase-coupled antidigoxigenin antibody solution, washed with the same buffer, and covered with the Lumiphos chemiluminescent reagent (Lumigen, USA) in the dark at room temperature for 5 min. The membranes were exposed for 5 to 180 min to photographic films and then further developed. The membranes were stripped by two incubations with a boiling 0.1% (wt/vol) SDS solution for 30 min. The prehybridization, hybridization, and developing steps were repeated using an anti-5S rRNA probe in order to provide an indication of total RNA load.
Real-time quantitative reverse transcription-PCR (qRT-PCR) was performed with the Kapa SYBR Fast one-step qRT-PCR Universal kit (Kapa Biosystems, USA). Briefly, 50 ng of purified total RNA was retrotranscribed and amplified by means of a gene-specific primer pair at a final concentration of 200 mM in a reaction volume of 8 μl. Table 1 lists the primers used for the reactions. The transcript abundance for each gene of interest was calculated through the use of three biological replicates for each strain or culture condition (each with three technical replicates). qPCRs were carried out with a Bioer 9600 thermal cycler (Bioer, China). To confirm the presence of a single amplification product, a melting curve was determined after each cycling. The uniformly expressed gene SMc01852 was used to normalize gene expression (59). The baseline of each amplification curve, the mean amplification efficiency, the threshold fluorescence value for each primer pair, and the quantification cycle (Cq) corresponding to each reaction were determined with the LinRegPCR software (60).
The cellular PHB content was determined by flow cytometry after staining cells with the fluorescent dye Nile red. Approximately 5 × 106 cells were centrifuged from aliquots of three independent cultures, washed once with phosphate-buffered saline (PBS), and then permeabilized by treatment with 35% (vol/vol) aqueous ethanol for 15 min. The permeabilized cells were then collected by centrifugation, washed again with PBS, and resuspended in 500 μl PBS with 0.04% (wt/vol) Nile red. After 30 min of incubation, the cells were analyzed in a FACSCalibur (Becton Dickinson, USA) flow cytometer with excitation by a 488-nm-wavelength argon laser. The fluorescence emitted by Nile red was recorded in channel FL2 (containing a filter with bandwidth of 585 ± 42 nm). The background fluorescence from the Nile-red-stained bacteria was minimal (61). Forward (FSC)- and side (SSC)-scattered light were also registered for a total number of 100,000 events. Cytometric data analysis was performed with the FlowJo software (FlowJo, USA). The geometric mean and standard deviation (SD) of the arbitrary fluorescence values for each cell population distribution were determined. Tukey's multiple-comparison test was carried out to determine whether or not the mean amounts of intracellular PHB in the populations subjected to analysis were statistically different. Flow cytometry assays were carried out twice.
Aliquots of stationary-phase cultures at a defined OD600 were centrifuged at 14,000 × g for 10 min. After two washes with SS, the pellets were dried at 105°C for 24 h and weighed with an analytical balance. After relativization to viable cell counts that had been previously determined, values were expressed as picograms per cell. Cell pellets obtained from three independent cultures of each strain were analyzed, and the results are expressed as the average bacterial dry weight ± one standard deviation.
Cells of S. meliloti 2011 and its isogenic mmgRΔ33–51 mutant from stationary-phase cultures in RDM were negatively stained with a solution of 2% (wt/vol) uranyl acetate for 5 min and then directly observed at a magnification of ×20,000 with a JEOL/JEM 1200 EX II transmission electron microscope (TEM) at the Microscopy Service of the School of Veterinary Sciences (National University of La Plata, La Plata, Argentina). The dimensions of at least 100 cells were analyzed in digitalized images of each sample.
Cell pellets were collected by centrifuging 1 ml from each of three independent cultures of the wild-type and mmgRΔ33–51 strains in RDM in the stationary phase of growth. The pellets were resuspended in 200 μl of Laemmli buffer (62) and incubated at 100°C for 10 min. After clearing the lysates by centrifugation at 12,000 × g for 5 min, 10 μl of each supernatant was loaded into a 15% (wt/vol) SDS-polyacrylamide gel and the proteins separated by electrophoresis for 3 h (62). The proteins were stained with Coomassie brilliant blue R-250 (63) and digitalized. The proteins from the excised gel bands were reduced, alkylated, and digested with trypsin; the peptides were then separated by reverse-phase nanoscale liquid chromatography (NanoLC) and subjected to mass spectrometry (MS) analysis at the CEQUIBIEM facility (Center of Chemical and Biochemical Studies by Mass Spectrometry, Buenos Aires University) as described elsewhere (64). The protein identity was assigned by tandem MS (MS/MS) ion search of MS/MS signals obtained after higher-energy collisional-dissociation fragmentation of the most intense MS peaks.
The Sinorhizobium meliloti 2011 wild-type and mmgRΔ33–51 strains were cultured up to the stationary phase of growth in RDM containing 14NH4Cl or 15NH4Cl as the sole nitrogen source. Three independent cultures of each bacterial population were processed as biological replicates. At the moment of the cell harvesting, aliquots of cultures containing 40 OD units each were mixed with their respective pair for proteomic comparison and cooled on ice before centrifuging to collect cell pellets. Proteins were extracted and separated into the cytosolic and membrane subcellular fractions following the protocol described elsewhere (24). The protein concentration in each fraction was determined by the Bradford colorimetric assay (Bradford protein assay Ready-to-Use; Bio-Rad, USA) (65) with a bovine serum albumin solution as a reference, and the quality was verified by SDS-PAGE (62). The proteins from both fractions were precipitated overnight with 6 volumes of ice-cold acetone and stored as dry pellets. An on-pellet trypsin digestion was performed after pellet resuspension in 50 mM Tris-HCl (pH 8.5) buffer, followed by peptide reduction with dithiothreitol and alkylation with iodoacetamide, as described previously (66). The mass spectrometric analysis of the samples was performed with an Orbitrap Velos Pro mass spectrometer (Thermo Scientific, Germany). An Ultimate nanoscale rapid-separation LC–high-pressure liquid chromatography (nanoRSLC-HPLC) system (Thermo Scientific, Germany), equipped with a custom 20-cm by 75-μm column filled with 1.7-μm C18 beads, was connected on line to the mass spectrometer through a Thermo Scientific Nanospray Flex ion source. One microliter of the tryptic digest was injected onto a C18 micro-preconcentration column (300 μm [inner diameter] by 5 mm). The automated trapping and desalting of the sample, the separation of the tryptic peptides, and the mass spectrometric analysis were performed as described previously (67). Spectra were loaded into the QuPE server (68) at Bielefeld University (Germany). An initial preprocessing of spectra was performed at the server default settings. The peptide identity was assigned through the use of the S. meliloti decoy database at the Mascot search engine (Matrix Science, UK) with the following parameters: enzyme, trypsin with 2 uncleaved sites allowed; peptide tolerance, 10 ppm; allowed precursor mass, 12C; 15N metabolic labeling; tolerance threshold, 0.05; allowed charge states, +2 and + 3; instrument, ESI-TRAP; fixed modifications, carbamidomethyl (C); variable modifications, oxidation (M). A false-discovery-rate (FDR) analysis was performed, and only peptides with an FDR lower than 0.05 were considered further. The intensities of the light and heavy isotopes of each peptide were quantified by the RelEx linear exclusion algorithm (69). mmgRΔ33–51/wild-type protein ratios were calculated as the median of the corresponding peptide ratios. A Student t test was performed to assess statistical significance.
A.L. and C.V. thank Susana Jurado for assistance with the transmission electron microscopy, Valeria Segatori for facilitation of the FACS assays, and Tina Krieg (Marburg University, Germany) and Pia Valacco and Silvia Moreno (CEQUIBIEM, University of Buenos Aires, Argentina) for their technical assistance with the mass spectrometer. We thank Donald F. Haggerty for his valuable help in editing the final version of the manuscript.
We declare no conflicts of interest.
This work was supported by the Argentinean National Scientific and Technical Research Council (CONICET), the National Agency for Promotion of Science and Technology (ANPCyT), and Universidad Nacional de Quilmes. A.L. was supported by CONICET and DAAD (German Academic Exchange Service) fellowships. G.C.B. was supported by ANPCyT and CONICET fellowships. C.V. is a researcher of CONICET. A.B. was supported by CRC 987 (German Research Foundation). We acknowledge technical assistance and access to resources supported by BMBF grant FKZ 031A533 within the de.NBI network. The Orbitrap mass analyzer was funded by the Deutsche Forschungsgemeinschaft (DFG grant INST 160/503-1 FUGG).
Supplemental material for this article may be found at https://doi.org/10.1128/JB.00776-16.