The small RNA Aar in Acinetobacter baylyi: a putative regulator of amino acid metabolism
- First Online:
- Cite this article as:
- Schilling, D., Findeiß, S., Richter, A.S. et al. Arch Microbiol (2010) 192: 691. doi:10.1007/s00203-010-0592-6
- 180 Views
Small non-coding RNAs (sRNAs) are key players in prokaryotic metabolic circuits, allowing the cell to adapt to changing environmental conditions. Regulatory interference by sRNAs in cellular metabolism is often facilitated by the Sm-like protein Hfq. A search for novel sRNAs in A. baylyi intergenic regions was performed by a biocomputational screening. One candidate, Aar, encoded between trpS and sucD showed Hfq dependency in Northern blot analysis. Aar was expressed strongly during stationary growth phase in minimal medium; in contrast, in complex medium, strongest expression was in the exponential growth phase. Whereas over-expression of Aar in trans did not affect bacterial growth, seven mRNA targets predicted by two in silico approaches were upregulated in stationary growth phase. All seven mRNAs are involved in A. baylyi amino acid metabolism. A putative binding site for Lrp, the global regulator of branched-chain amino acids in E. coli, was observed within the aar gene. Both facts imply an Aar participation in amino acid metabolism.
Traditionally, RNA is thought of as ribosomal RNA (rRNA), transfer RNA (tRNA), and messenger RNA (mRNA) that function in the assembly of proteins. More than 20 years ago, the fortuitous identification of MicF RNA as a post-transcriptional regulator of mRNAs encoding outer membrane proteins changed this perception of RNA function. MicF causes translational inhibition by forming an imperfect duplex upon base-pairing with the translation-initiation region of the trans-encoded ompF mRNA (Mizuno et al. 1984; Delihas and Forst 2001; Vogel and Papenfort 2006). RNA-mediated regulation is also known to occur at the protein level. For example, 6S RNA is able to form a complex with the σ70 RNA polymerase holoenzyme, which leads to downregulation of σ70-dependent transcription (Wassarman 2007). Another regulatory RNA, CsrB, sequesters the carbon storage regulator protein CsrA, which is a negative regulator of glycogen biosynthesis, gluconeogenesis, and glycogen catabolism, and antagonizes the CsrA ability of translation repression (Liu et al. 1997). Bacterial small regulatory RNAs (sRNAs) typically range from 50 to 250 nt in length (Hershberg et al. 2003; Altuvia 2007), with some exceptions, e.g., CsrB and RNAIII, which are 366 nt and 514 nt in length (Liu et al. 1997; Boisset et al. 2007). To detect novel bacterial sRNAs, systematic bioinformatic searches have been performed and have identified numerous sRNAs expressed within intergenic DNA regions (Argaman et al. 2001; Axmann et al. 2005; Livny et al. 2006). Whereas these bioinformatic predictions focused strictly on intergenic regions, experimental approaches have been applied to clone all RNAs of 30–65 nt or 50–500 nt and screen for novel sRNAs (Kawano et al. 2005; Willkomm et al. 2005; Sonnleitner et al. 2008). Half of all validated sRNAs to date (over 150 sRNAs) have been identified in the model organism Escherichia coli and in closely related pathogenic Enterobacteriaceae including Salmonella, Klebsiella, or Yersinia (Hershberg et al. 2003; Livny and Waldor 2007; Gardner et al. 2009). Although the majority of known sRNAs have yet to be assigned to cellular functions, some prominent examples demonstrate that sRNAs are key players in various cellular processes. During iron starvation, RyhB from E. coli functions in iron metabolism by downregulating genes encoding iron-containing proteins (Masse and Gottesman 2002). The Vibrio harvey sRNAs Qrr 1–5 affect quorum sensing by duplex formation at the ribosome-binding site of luxR mRNA, which encodes the master regulator of quorum-sensing genes (Tu and Bassler 2007). Several sRNAs like Spot42, SgrS, GlmY, and Z are regulators of sugar metabolism (Gorke and Vogel 2008). Furthermore, GcvB inhibits translation of seven periplasmic substrate-binding proteins of the ABC uptake system (Sharma et al. 2007).
Altering accessibility of the ribosome-binding site by sRNA–mRNA duplex formation is the predominant regulatory mechanism of sRNAs. Such interference by trans-encoded sRNAs is due to incomplete base pairing often stabilized by the conserved RNA chaperone Hfq, as is the case for ompA-MicA and sodB-RyhB (Moll et al. 2003; Geissmann and Touati 2004; Udekwu et al. 2005). Hfq also mediates RNA–RNA interactions by accelerating duplex formation between sRNA and mRNA (Kawamoto et al. 2006) and interferes in polyadenylation, translation, and mRNA degradation (Brescia et al. 2003; Valentin-Hansen et al. 2004; Maki et al. 2008; Regnier and Hajnsdorf 2008). In a previous study, we could show that Acinetobacter baylyi hfq encodes an unusual large protein compared to other Hfqs. Its absence results in severe reduction in growth and an abnormal cell phenotype of A. baylyi. However, despite its abnormal size, A. baylyi hfq is able to complement an E. coli hfq mutant in vivo (Schilling and Gerischer 2009). In this report, we describe Aar (Acinetobacter amino acid regulator), a trans-encoded sRNA involved in amino acid regulation in Acinetobacter baylyi, a Gram-negative, strictly aerobic soil bacterium (Barbe et al. 2004; Vaneechoutte et al. 2006). One of two detected Aar transcripts is affected by hfq deletion. We show that an increased Aar level results in upregulation of fadA, ilvI, ppC, glnA, serC, leuC, and gcvH mRNAs, which all function in amino acid metabolism.
Prediction of small RNAs in two Acinetobacter species
Properties of putative small RNA genes from Acinetobacter baylyi
Genomic location of putative sRNA
Predicted φ-independent termination signal next to the sRNA gene
Small RNA Aar is encoded as an independent Acinetobacter gene
The mapped 3′-end aligns with the predicted Aar Rho-independent transcription termination stem (2,812,382–2,812,408). Putative elements of a σ70 promoter with a perfect Pribnow box containing the consensus sequence TATAAT and a 50% match to the σ70 −35 consensus region were found upstream of the aar transcription start site (Fig. 2a). BLASTN analysis was performed with the mapped aar sequence of A. baylyi ADP1 and the NCBI database (nucleotide collection). Aar homologs were found only in Acinetobacter baumannii strains SDF, AB307-0294, AB0057, ACICU, AYE, and ATCC17978 with sequence identities between 69 and 71% and a 92% query coverage (missing the first nine and last five bases of ADP1 aar). These findings indicated that Aar is unique to Acinetobacter. A multiple sequence alignment with all Acinetobacter aar homologs is shown in Fig. 2b. While the Acinetobacter baumannii aar sequences share an identity of 98%, A. baylyi ADP1 aar showed only a strictly conserved 3′-end compared to the respective A. baumannii homologs (Fig. 2b). The genomic localization of aar was found to be conserved as well. At least either trpS or sucD is encoded at a position flanking the A. baumannii aar homologs (i.e., A. baumannii SDF aar location: upstream of sucD on the negative strand; A. baumannii AB307-0294 aar location: downstream of trpS on the same strand; A. baumannii ACICU aar location: between trpS and sucD).
Expression profile of aar
Effects of Aar modification
Prediction of putative Aar targets
Predicted Aar–mRNA targets that have been selected for further experimental analysis
Interaction site mRNA (relative to AUG)a
Interaction site Aar
Interaction energy (kcal/mol)
Branch-chained amino acid synthesis
Branch-chained amino acid degradation
Branch-chained amino acid synthesis
Nitrogen fixation into glutamine
Possible involvement of Lrp
Gene expression of many sRNAs is regulated by specific transcriptional regulators like Fur-RyhB, SgrR-SgrS, or LuxO-Qrr1-5 (Bejerano-Sagie and Xavier 2007; Masse et al. 2007; Vanderpool 2007). To find potential regulator-binding sites, we checked the A. baylyi aar gene (position 2,812,231–2,812,412) with the Prokaryotic Database of Gene Regulation PRODORIC (Munch et al. 2003). Within the aar sequence, 59 bases downstream of the transcription start site (position 2,812,290), we found an 80% sequence match to a confirmed 15-base pair-long binding site for the leucine-responsive protein (Lrp, Fig. 4c). Lrp is a global regulator of operons involved in amino acid biosynthesis and degradation (Calvo and Matthews 1994; Cui et al. 1995). The presence of the putative Lrp-binding site within aar and the accumulation of a set of mRNAs related to amino acid metabolism upon Aar over-expression might indicate that these metabolically related genes are regulated by a common mechanism, which might include the known global regulator Lrp and the newly discovered sRNA Aar.
In this study, we report the discovery of Aar (Acinetobacter amino acid regulator), the first sRNA of the genus Acinetobacter with no homology to already known prokaryotic sRNAs. Aar was predicted by an in silico approach, and BLASTN analysis indicated that Aar is present only within the genus Acinetobacter. Within these species, the aar genomic location is conserved. A few examples of species-specific sRNAs like Aar are known, including Qrr 1–5 in Vibrio harvey and V. cholerae (Bejerano-Sagie and Xavier 2007) and InvR in Salmonella typhimurium and S. bongori (Pfeiffer et al. 2007).
The size of ADP1 Aar was determined to be 181 bases, but the presence of two bands in Northern blot analysis indicates that this sRNA is processed after transcription or an alternative promoter might exist. The absence of the second, smaller Northern blot signal in the Δhfq mutant suggests that Hfq, as global RNA chaperone, may play a role in Aar processing. However, recent results indicate that Hfq affects also the transcription efficiency in addition to mRNA degradation (Le Derout et al. 2010), but neither CR-RT-PCR—nor 5′-RACE—experiments detected different aar-promoters that would explain the absence of the smaller Aar transcript in the Δhfq mutant. Thus, instead of an Hfq impact upon aar transcription, it is more likely that Hfq—directly or indirectly—affects Aar stability. In general, Hfq is essential for processing, polyadenylation, and degradation of RNAs (Hajnsdorf and Regnier 2000; Le Derout et al. 2003; Masse et al. 2003; Valentin-Hansen et al. 2004; Folichon et al. 2005). Furthermore, Hfq is involved in sRNA-dependent mRNA inactivation by promoting RNA structure refolding (Brescia et al. 2003; Geissmann and Touati 2004). Thereby, the Hfq interference often causes repression of mRNA translation and coupled sRNA and mRNA degradation by ribonucleases (Masse et al. 2003; Udekwu et al. 2005; Kawamoto et al. 2006). Thus, it is conceivable that the loss of Hfq may result in the accumulation of unprocessed Aar transcripts, which was seen in the Northern blot experiments with the A. baylyi Δhfq mutant.
Under normal conditions, Aar is present in high amounts during stationary growth, which coincides with the fact that many sRNAs are induced under stress or when nutrients become scarce. For example, sRNAs identified in E. coli that are induced after environmental changes are RyhB under iron depletion, OxyS after oxidative stress, DsrA at low temperatures, and SgrS after the accumulation of glucose-P (Gottesman 2005). Interestingly, Aar is influenced by changing salt conditions but not by the temperature variations tested. Sodium chloride excess as well as iron depletion resulted in earlier expression of Aar during growth.
As seen by Northern blot analysis, Aar over-expression led to the accumulation of seven mRNAs that encode proteins involved in amino acid metabolism. Of these, glnA encodes glutamine synthetase, which assimilates nitrogen by converting glutamate and nitrogen into glutamine (Calvo and Matthews 1994). In the Aar over-expression mutant, glnA mRNA accumulated during stationary growth in minimal medium but the mRNA was barely detectable under normal Aar expression conditions. No differences were seen during exponential growth. Aar also affected the mRNAs leuC, serC, ilvI, ppc, gcvH, and fadA similarly.
In contrast to a direct base pair interaction, it is also conceivable that another regulator of amino acid metabolism in A. baylyi is involved in the regulation. In fact, we found a putative binding site with 80% identity to the 15-nucleotide E. coli Lrp consensus sequence (59 bases downstream of the mapped Aar transcription start site) (Cui et al. 1995). In E. coli, Lrp acts as global transcriptional regulator of amino acid metabolism, transport, and pili formation (Newman and Lin 1995). The ADP1 and E. coli Lrp proteins contain almost the same size and show 58% sequence identity and 75% sequence similarity regarding the Lrp helix-turn-helix domain. In ADP1, the putative Lrp-binding site in the aar gene might indicate negative regulation of the Aar expression by Lrp. We observed that Aar is highly abundant during exponential and stationary growth in complex medium, but only during stationary growth in minimal medium. This agrees with the fact that the lrp gene is self-repressed, primarily during growth in complex medium (Lin et al. 1992). During stationary growth in both minimal and complex media, amino acids become scarce and therefore Lrp synthesis might be reduced, which in turn could result in increased Aar levels. An inverse effect is also possible, since Lrp was predicted as Aar target (RNAz: −3.768 kcal/mol, interaction site: Fig. 5).
The identification, experimental verification, and initial investigation of a novel and unique sRNA from the soil bacterium A. baylyi contribute to the ongoing effort to characterize and understand the extent and function of these exciting new molecules. Obviously, more experimental effort will reveal explanations to the exciting observation of the involvement of Hfq and the direct or indirect connection of Aar to amino acid metabolism.
Bacterial strains and growth conditions
Strains, plasmids, and oligonucleotides used in this study
Source or reference
Strain or plasmid
Acinetobacter wild type (strain BD413, ATCC 33305)
ADP1 pRK415 aar
Acinetobacter expressing sRNA Aar in trans via natural promoter, tetr
ADP1 pRK415 plac aar
Acinetobacter expressing sRNA Aar in trans via lac promoter, tetr
Acinetobacter containing an Ω-interposon instead of hfq ORF, sper
Schilling and Gerischer (2009)
ADP1 pRK415 hfq
Acinetobacter expressing Hfq in trans via natural promoter, tetr
Schilling and Gerischer (2009)
Escherichia coli general cloning strain
ColE1 replicon, ampr, lacZα, 2,958 bp
RK2 replicon, tetr, 10,500 bp
Keen et al. (1988)
ColE1 replicon, kanr, 4,800 bp
Figurski and Helinski (1979)
aar region (2,812,113–2,812,606) cloned with EcoRI/PstI in pRK415
pRK415 aar Plac
aar region (2,812,113–2,812,606) cloned with HindIII/PstI in pRK415
E. coli strains were grown in LB medium with aeration at 37°C and were supplied with antibiotics when appropriate (tetracycline 12 μg/μl, ampicillin 100 μg/μl, kanamycin 50 μg/μl).
Plasmid and strain construction
Standard methods were used for plasmid isolation, DNA purification, restriction endonuclease cleavage, ligation, and transformation (Sambrook and Russel 2001). Over-expression of Aar was achieved by cloning aar into the broad-host-range plasmid pRK415 (Keen et al. 1988). To obtain the DNA region of interest from the A. baylyi genome, a PCR using Pfu-DNA polymerase, primers 368/369 (Table 3), and 10 ng A. baylyi chromosomal DNA was carried out. The PCR product was gel purified and digested with either PstI/HindIII or PstI/EcoRI. These products were ligated with pRK415 that had been digested with the same enzyme pair, yielding plasmids pRK415 aar and pRK415 plac aar, respectively. After transformation into E. coli DH5α cells (Hanahan 1983), both plasmids were established via tetracycline resistance. These cloning steps resulted in plasmid pRK415 aar expressing Aar from its own promoter and pRK415 plac aar expressing it from the lac promoter. Plac functions in A. baylyi and cannot be controlled using isopropyl β-d-1-thiogalactopyranoside, since the bacterium does not have a lac operon and the plasmid does not contain lacI either. Plasmids were recovered from positive E. coli clones and verified by digestion with PstI/HindIII, PstI/EcoRI, and agarose gel electrophoresis. Furthermore, plasmids were sequenced; those without mutations were conjugated by means of plasmid pRK2013 (Figurski and Helinski 1979) from E. coli DH5α into A. baylyi ADP1. Positive ADP1 clones were selected by growth in minimal medium with 6 μg/μl tetracycline and 5 mM quinate as sole carbon source.
Northern blot analysis
Total RNA was isolated by a previously described procedure (Oelmüller et al. 1990). The RNA quality and concentration were determined by the OD260/280 ratio. Fifteen micrograms of purified RNA was combined with RNA loading dye [1 ml 5× RNA loading dye = 8 μl 500 mM ethylenediaminetetraacetic acid, 200 μl 100% glycerol, 72 μl 37% formaldehyde, 308 μl formamide, 400 μl 10× running buffer (200 mM 3-(N-morpholino)propanesulfonic acid, 50 mM sodium acetate, 5 mM ethylenediaminetetraacetic acid, pH 7), 2 μl saturated bromphenol blue, 10 μl deionized water] and then heat-denatured for 5 min at 65°C. Afterward, RNA was separated on a 1.2% formaldehyde–agarose gel and transferred to a Hybond-N+ nylon membrane by capillary blotting overnight with 10× SSC [1.5 M sodium chloride, 0.15 mM tri-sodium citrate]. The membrane was dried at 25°C, and the RNA was covalently bound to the membrane by UV cross-linking for 90 s at 120,000 μJ/cm2. Specific RNA detection was performed by the DIG labeling and detection system (Roche Applied Science, Mannheim, Germany) with specific PCR probes and primers 315/316 (sRNA2), 331/332 (Aar), 396/397 (IlvI), 398/399 (FadA), 402/403 (Ppc), 404/405 (MinE), 419/420 (GlnA), 421/422 (GlnT), 423/424 (LeuC), 425/426 (SerC), and 431/432 (GcvH) (Table 3).
RNA end mapping
RNA ends were mapped by a modified protocol of the circularized RT-PCR method described previously (Forner et al. 2007). Here, 5 μg of RNA (treated with 25 U tobacco acid pyrophosphatase for 1 h at 37°C) was denatured at 65°C for 10 min and quickly cooled on ice. RNA self-ligation was performed with 40 U T4 RNA ligase I, 10 U RNase inhibitor, 1 U RNase-free DNase I, and 1× T4 RNA ligase buffer (50 mM Tris, 10 mM magnesium chloride, 1 mM adenosine triphosphate, 10 mM dithiothreitol, pH 7.8) in a total volume of 25 μl at 37°C for 2 h. The volume was adjusted with water to 500 μl, and the enzymes were removed by phenol treatment. Self-ligated RNA was precipitated overnight at −20°C with 250 mM sodium acetate pH 5.2 and 100% ethanol. Finally, cDNA synthesis was performed with 0.5 mM dNTP’s, 1 pmol gene-specific primer, 355 for Aar and 347 for sRNA2 (Table 3), 5 μg self-ligated RNA, 200 U reverse transcriptase (RNase H minus mutant) in 1× RT buffer (50 mM Tris, 75 mM potassium chloride, 3 mM magnesium chloride, 10 mM dithiothreitol, pH 8.3) for 2 h at 45°C. PCR was carried out with 2.5 μl of heat-treated cDNA reaction (70°C, 10 min) and primers 355/346 (Aar), 347/348 (sRNA2), in a standard PCR mixture. Alternatively, 5′-RACE experiments (Roche Applied Science, Mannheim, Germany) were performed for aar as described in the manual with primer 377 and nested primer 378 (Table 3).
Primer sequences (Table 3) were selected to have a melting temperature of at least 60°C. The PCR conditions including Taq- or Pfu-DNA polymerase were 95°C for 5 min, followed by 30 cycles of denaturation at 95°C for 45 s, annealing at 54°C for 1 min, and extension at 72°C for at least 30 s (longer for products over 500 bases in length).
sRNA gene and mRNA target predictions
Comparative intergenic genome analysis of A. baylyi ADP1 (NCBI Refseq Id NC_005966) and A. baumannii ATCC17978 (NCBI Refseq Id NC_009085) in combination with the detection of thermodynamically stable putative non-coding RNAs using RNAz (Washietl et al. 2005) was kindly performed by Voss et al. (2009). This initial computer-aided search resulted in 481 non-coding RNA candidates. Inspection of these predicted loci revealed among others Aar, with a Z-score of −3.09 and a P-value of 0.99, to be a good candidate for further analysis.
Putative mRNA targets of Aar were predicted by two bioinformatic programs: IntaRNA (Busch et al. 2008) and RNAup (Mückstein et al. 2006). Both programs utilize a model based on the hybridization energy of the two interacting RNAs as well as the energy required to make the interaction sites in both molecules accessible (based on all possible conformations). The main difference between the two programs is that IntaRNA enforces a region of continuous pairing (seed region) as a hybridization start. Here, a minimum seed length of eight base pairs was required. The search for interactions was performed in a region of 250 nt upstream and 150 nt downstream of each A. baylyi annotated mRNA. As part of the IntaRNA prediction algorithm, only interactions that involve the highly conserved regions of Aar (positions 10–53 and 136–149) RNA were considered. The most predominant mechanism of post-transcriptional gene regulation by trans-encoded sRNAs is interference with ribosome binding to the Shine Dalgarno (SD) sequence followed by degradation of the mRNA–sRNA duplex. Therefore, SD sequence locations for all mRNAs were predicted by simulating hybridization with the single-stranded 3′-tail of the 16S rRNA (Starmer et al. 2006). Targets for which Aar RNA was predicted to bind at or in the immediate vicinity of the SD sequence were selected for further analysis. Additional targets were predicted by RNAup and a pipeline based on sequence as well as interaction site conservation of both molecules. Using a reciprocal best BLAST approach (http://www.bioinf.uni-leipzig.de/Software/proteinOrtho/), all annotated protein coding genes of A. baylyi orthologs within the six completely sequenced and annotated A. baumanni strains (NCBI Refseq Ids NC_010611, NC_011586, NC_009085 NC_010410, NC_011595 and NC_010400) were identified. Then, optimal interactions for each Aar homolog were predicted in the corresponding species. Only interactions conserved in at least five Acinetobacter strains were assumed to be functional.
We would like to thank Björn Voss for his predictions of sRNA genes in Acinetobacter and Véronique de Berardinis for her effort to delete aar. Furthermore, we would like to thank Iris Steiner for her contribution. This work was supported by the German Federal Ministry of Education and Research (BMBF grant 0313921 FRISYS to A.S.R.); the German Research Foundation (DFG grant SPP1258 to S.F. STA850/7-1 and A.S.R. BA2168/2-1); and the state of Baden-Württemberg, Germany (personal LGFG grant to D.S.).