Elisabeth Kay, Christophe Dubuis, and Dieter Haas *

Département de Microbiologie Fondamentale, Université de Lausanne, CH-1015 Lausanne, Switzerland

Author contributions: E.K., C.D., and D.H. designed research; E.K. and C.D. performed research; and D.H. wrote the paper.

*To whom correspondence should be addressed.
Dieter Haas, E-mail: dieter.haas@unil.ch

In many Gram-negative bacteria, the GacS/GacA two-component system positively controls the expression of extracellular products or storage compounds. In the plant-beneficial rhizosphere bacterium Pseudomonas fluorescens CHA0, the GacS/GacA system is essential for the production of antibiotic compounds and hence for biological control of root-pathogenic fungi. The small (119-nt) RNA RsmX discovered in this study, together with RsmY and RsmZ, forms a triad of GacA-dependent small RNAs, which sequester the RNA-binding proteins RsmA and RsmE and thereby antagonize translational repression exerted by these proteins in strain CHA0. This small RNA triad was found to be both necessary and sufficient for posttranscriptional derepression of biocontrol factors and for protection of cucumber from Pythium ultimum. The same three small RNAs also positively regulated swarming motility and the synthesis of a quorum-sensing signal, which is unrelated to N-acyl-homoserine lactones, and which autoinduces the Gac/Rsm cascade. Expression of RsmX and RsmY increased in parallel throughout cell growth, whereas RsmZ was produced during the late growth phase. This differential expression is assumed to facilitate fine tuning of GacS/A-controlled cell population density-dependent regulation in P. fluorescens.

Bacterial small RNAs of another type act by sequestering RNA-binding proteins belonging to the CsrA (carbon storage regulator) family (7), which regulate translation initiation by binding to mRNA sequences near the ribosome-binding site. In E. coli, CsrA regulates the utilization of carbon sources, glycogen synthesis, biofilm formation, and motility (7–11), whereas in Erwinia carotovora, the CsrA homolog RsmA (repressor of
secondary metabolism) controls the expression of extracellular enzymes and type III secretion (12). The similar RsmA and RsmE proteins of the plant-beneficial root-colonizing biocontrol strain CHA0 of Pseudomonas fluorescens negatively regulate the synthesis of extracellular antifungal secondary metabolites (13–15). Small noncoding RNAs, such as CsrB and CsrC of E. coli (8, 16), RsmB of E. carotovora (17), or RsmZ and RsmY of P. fluorescens (14, 18), bind multiple CsrA/RsmA molecules with high affinity and thereby allow translation of mRNAs, which are repressed by CsrA or RsmA. Among the factors that influence the expression of these small RNAs, the DNA-binding protein GacA stands out as a major activator (19, 20). The response regulator GacA is activated by phosphorylation from the cognate membrane-bond sensor GacS (21–24).

The GacS/GacA two-component system is conserved in many Gram-negative bacteria. Whereas in plant- and animal-pathogenic species GacS/GacA is important for virulence (20, 25–27), the same system is required for biocontrol in plant-beneficial strains (19, 20, 24, 28, 29). We have previously reported that, in P. fluorescens CHA0, GacS/GacA positively controls transcription initiation of two small RNA genes, rsmZ and rsmY (14, 18). However, RsmZ and RsmY alone cannot fully explain how the GacS/GacA system determines biocontrol
activity, because an rsmY rsmZ double mutant retains partial expression of biocontrol traits (18). Here, we report the discovery of a third GacA-controlled small RNA, RsmX, in P. fluorescens and show that the simultaneous absence of RsmX, RsmY, and RsmZ RNAs mimicks the biocontrol-negative phenotype of gacS and gacA mutants. This triad of small RNAs was also found to control swarming motility and the synthesis of a
low-molecular-weight quorum-sensing signal that induces the Gac/Rsm cascade.

Materials and Methods

Bacterial Strains and Culture Conditions.

P. fluorescens CHA0 (wild type), CHA19 (DgacS), CHA89 (gacA::Km), CHA810 (DrsmZ), CHA822 (DrsmY), CHA825 (DrsmY DrsmZ), CHA1003 (rsmE::W-Hg), CHA1008 (DgacS rsmE::W-Hg rsmA::W-Km), CHA1009 (rsmE::W-Hg rsmA::W-Km), and CHA1076 (rsmA::W-Km) and their derivatives carrying chromosomal
hcnA'-'lacZ or aprA'-'lacZ fusions have been described (14, 15, 18, 19, 24). Strains carrying a chromosomal rsmE'-'lacZ fusion were constructed by using pME7545 for delivery (29). Growth conditions, antibiotic concentrations, and conditions for b-galactosidase assays were the same as those previously used
(14, 15, 18).

RNA Extraction and Northern Blot Analysis.

These were performed as described (14, 18). Hybridizations were done with a digoxigenin- labeled DNA probe generated by PCR covering the entire rsmX structural gene with primers PTX3hind and PTX4eco (Table 3, which is published as supporting information on the PNAS web site).

Isolation of rsmX.

His-tagged RsmA and RsmE (RsmA6H and RsmE6H) were overexpressed in strain CHA825 carrying
pME6078 or pME7011 and purified by Ni-nitrilotriacetic acid affinity chromatography, as described (14). RNA was separated from the purified proteins by two extractions with phenol/chloroform (1:1) and once with chloroform. RNA was ethanol-precipitated, dissolved in diethylpyrocarbonate-treated water, and electrophoresed on a denaturing urea-polyacrylamide gel. An RNA band of ~100 nucleotides was extracted with the ElutaTube RNA Extraction Kit (Fermentas, Hanover, MD). cDNA synthesis was performed as described (8) by using random hexameric primers and the Universal Riboclone cDNA Synthesis System (Promega). The resulting blunt-ended cDNA fragments were cloned into SmaI-digested phosphatase-treated pBluescript KS (Stratagene). Ten independent clones were sequenced by using primers T3 and T7; the inserts of two clones contained an rsmX fragment. The rsmX gene was located between the genes PFL-4113 and PFL-4112 on the chromosome of P. fluorescens Pf-5 (ref. 30; http://www.tigr.org ) by using the program FUZZNUC ( http://bioweb.pasteur.fr/seqanal/interfaces/fuzznuc.html ).

The rsmX gene and f lanking sequences were obtained by PCR amplification of P. fluorescens CHA0 chromosomal DNA with primers PRX1hind and PRX2eco (Table 3), which were designed according to the genome sequence of P. fluorescens Pf-5. The 1.6-kb PCR product was cloned into pBluescript KS (giving
pME7312) and sequenced. A 64-bp deletion in the chromosomal rsmX gene (extending from nucleotide + 17 to nucleotide + 81) was constructed as follows. Two sequences f lanking rsmX were PCR-amplified from CHA0 chromosomal DNA with primers PRX1hind + PMUBAM2 and PRX2eco + PMUBAM1 (Table 3). The resulting 635-bp upstream and 955-bp downstream fragments were cleaved with HindIII + BamHI, and BamHI + EcoRI, respectively, and inserted into the suicide plasmid pME3087 (14) cut with HindIII + EcoRI. The resulting plasmid pME7315 was introduced into strains CHA0, CHA207, CHA805, CHA825, CHA826, and CHA827 (13, 14, 18) to generate, via homologous recombination, the corresponding mutants CHA1141 (DrsmX), CHA1142 (DrsmX, hcnA'-'lacZ), CHA1143 (DrsmX, aprA'-'lacZ), CHA1144 (DrsmX DrsmY
DrsmZ), CHA1145 (DrsmX DrsmY DrsmZ, hcnA'-'lacZ), and CHA1146 (DrsmX DrsmY DrsmZ, aprA'-'lacZ). An rsmX-lacZ fusion was constructed by PCR amplification of the 312-bp rsmX promoter region with primers PROF and PROR1 (Table 3). The product was digested by EcoRI and BamHI and introduced into pME6916 (18) digested with the same enzymes to produce pME7317, in which the +1 of the lacZ gene is fused to the +4 site of the rsmX promoter. An rsmX overexpression plasmid was constructed as follows. A PCR product containing rsmX was obtained with pME7312 and primers PTX4eco + POVEX (Table 3), digested with EcoRI, and cloned into pME6552 [=pUK21-Ptac (13)] cut with KpnI (blunted) and EcoRI, resulting in pME7319. A 0.9-kb MluI-EcoRI fragment of pME7319 carrying the tac promoter fused at the +1 site to rsmX was inserted into pME6032 (14), producing pME7320. This construct was introduced into strains CHA89–207 (13), CHA806 (14), CHA1145, and CHA1146.

In Vitro Transcription of rsmX and Gel Mobility-Shift Assays.

The rsmX gene was PCR-amplified with primers PTZhind and PTZeco (Table 3) and cloned under the control of the T7 promoter in pTZ19R (Fermentas) to produce pME7318. The radioactively labeled RsmX transcript was synthesized from linearized pME7318 with the T7 transcription kit (Fermentas) in the presence of
[^a-33P]UTP and purified (18). RsmA6H and RsmE6H were overexpressed in E. coli DH5a/pME6078 and
BL21/pME7013, respectively; purified; dialyzed against 10 mM Tris-acetate, pH 7.5, at 4°C; concentrated on a Centricon membrane (Millipore); and stored at -20°C. Binding reactions were carried out with labeled RsmX and RsmA6H or RsmE6H, as described (15, 18).

Detection of Exoproducts and Swarming Motility.

Hydrogen cyanide production was measured as described (31). Exoprotease activity was detected on skim-milk agar (0.4% blood agar base/0.05% yeast extract/15% skim milk/1.35% agar) by spotting 10⁷ cells in 10 ml of 0.9% NaCl. 2,4-Diacetylphloroglucinol was extracted with ethyl acetate from cell cultures grown to OD₆₀₀ ~7 in 100 ml of a medium containing 0.75% malt extract, 1% proteose peptone, 0.5% glycerol, 0.075% MgSO4, and 0.075% K₂HPO₄; incubation was for 40 h. The extract was dried under vacuum, dissolved in 200 ml of methanol, and analyzed by HPLC, as described (32). The quorum-sensing signal was extracted with dichloromethane from culture supernatants of strains CHA0, CHA89, and CHA1144, as described (14). For semiquantitative determination of signal molecules produced by these strains, different amounts of extracted supernatants were added to the reporter strain CHA0/pME6530 (hcnA'-/lacZ) (13). The amount of culture extract resulting in half-maximal induction of the reporter gene was calculated from Hanes plots ([S] v^-1 +V_max ^-1 [S] + K_m V_max ^-1 , where [S] is the amount of extract, v is the b-galactosidase-specific activity of the reporter, V_max is maximal specific activity, and K_m the amount of extract giving half-maximal induction). Swarming motility was assessed on semisolid medium (0.8% nutrient broth/0.5% glucose/0.5% agar) (33).

Biocontrol Assay in Natural Soil.

Ten f lasks containing three plants each were planted with cucumber and treated with Pythium ultimum and P. fluorescens, as described (18).

Results

Discovery of the GacA-Controlled Small RNA RsmX.

In P. fluorescens CHA0, the RNA-binding protein RsmA, together with its homolog RsmE, represses translation of biocontrol genes in the Gac/Rsm cascade (15). Histidine-tagged RsmA and RsmE were
separately overexpressed as a bait in the rsmY rsmZ double mutant CHA825, purified by affinity chromatography, and stripped of RNA molecules, which were converted to cDNA clones (see Materials and Methods). Among these cDNAs, two fragments belonged to a small RNA gene that was termed rsmX.
Based on the genomic sequence of P. fluorescens Pf-5, which is closely related to strain CHA0 (30, 34), primers were designed to amplify the entire rsmX gene, including f lanking sequences. The 1.6-kb rsmX region thus cloned was 99% identical to that of strain Pf-5, except for an insert of 311 nucleotides, which occurred only in the CHA0 chromosome downstream of rsmX (Fig. 1A).

Fig. 1. The rsmX gene is not linked to gacS and gacA, depends on gacA function for expression, and shows characteristic unpaired GGA motifs.

(A) Organization of the rsmX region of P. fluorescens CHA0. The -35 and -10 promoter sites are indicated with gray boxes. The palindromic sequence boxed from -70 to -53 denotes the upstream activating sequence (UAS), the putative GacA regulatory site. A sequence of 311 nucleotides (rhomboid box), which is absent from strain Pf-5, does not present any homology with a known sequence.

(B) Northern blot showing the differential temporal accumulation of RsmX. Total RNA was extracted from CHA0 (wild type; lanes 1–3), CHA89 (gacA; lanes 4–6). OD₆₀₀ values at the time of harvesting were, in lanes 1 and 4, 0.5; in lanes 2 and 5, 1.9; and in lanes 3 and 6, 2.5.

(C) Predicted secondary structure of RsmX at 30°C using the MFOLD program
( http://www.bioinfo.rpi.edu/applications/mfold ).

Expression of RsmX increased with increasing cell population densities in the wild type and totally depended on GacA function (Fig. 1B). The rsmX transcription start was mapped by RACE (described in Supporting Text and Fig. 6, which are published as supporting information on the PNAS web site). The deduced RsmX RNA (~119 nt) is predicted to have an elaborate secondary structure with a p-independent terminator and four potentially unpaired GGA motifs (Fig. 1C). These motifs are characteristic of RNAs that bind CsrA/RsmA-type proteins (7, 8, 11, 16–18). Homologs of rsmX were found in the database for P. fluorescens strains Pf-5 and Pf0–1, as well as for P. syringae pv. tomato DC3000, but not for P. aeruginosa (data not
shown).

RsmX Specifically Binds RsmA and RsmE Proteins.

In vitro, a 4-to 5-fold molar excess of RsmA (Fig. 2A) or RsmE (Fig. 2C) over RsmX completely converted the free RNA molecule to complexes of higher molecular weights.

Fig. 2. RsmA and RsmE bind to RsmX, as demonstrated by RNA gel mobility-shift analysis.

Fig. 2. RsmA and RsmE bind to RsmX, as demonstrated by RNA gel mobility-shift analysis.

RsmX was synthesized in vitro by T7 RNA polymerase in the presence of [ ^a-33 P]-UTP. Labeled RsmX (80 nM) was incubated with the concentrations of RsmA or RsmE indicated in A and C, respectively. The positions
of free (F) and bound (B) RNA species are indicated. Unlabeled competitor RNAs (X = RsmX, Y = RsmY, and Z = RsmZ and CarA leader) were synthesized following the same protocol but with unlabeled UTP. Labeled RsmX (80 nM) with either specific (RsmX, RsmY, or RsmZ) or nonspecific (CarA leader) non-labeled
competitors (300 nM) was incubated with 800 nM RsmA and 300 nM RsmE in B and D, respectively.

The specificity of the RsmX–RsmA (Fig. 2B) and RsmX–RsmE (Fig. 2D) interactions was verified: unlabeled RsmX, RsmY, or RsmZ competed effectively for the RNA-binding proteins, whereas an unrelated mRNA
fragment (carA) did not. From these and published observations (15, 18), we conclude that RsmX, RsmY, and RsmZ all have similar high affinities for RsmA and RsmE.

Regulation of rsmX.

The rsmX promoter was fused, at the fourth nucleotide downstream of the transcription start site, to a
transcriptional lacZ reporter in pME7317. This construct gave increasing b-galactosidase expression with increasing cell population densities (Fig. 3A).

Fig. 3. The expression of the rsmX gene is controlled by GacA, RsmA, RsmE, and the CHA0 signal in P. fluorescens CHA0.

Fig. 3. The expression of the rsmX gene is controlled by GacA, RsmA, RsmE, and the CHA0 signal in P. fluorescens CHA0.

(A) Activation of rsmX expression by RsmA and RsmE. b-Galactosidase activities of a transcriptional rsmX-lacZ fusion carried by pME7317 were determined in the wild type (CHA0; squares), a gacA mutant (CHA89; open diamonds), an rsmE mutant (CHA1003; diamonds), an rsmA mutant (CHA1076; triangles), an rsmAE double mutant (CHA1009; open circles), and a gacS rsmAE mutant (CHA1008; open triangles).

(B) Activation of rsmX expression by the signal from CHA0. b-Galactosidase activities of an rsmX-lacZ fusion
were determined in the wild-type strain without (squares) or with extract (open squares) and the gacA mutant without (triangles) or with extract (open triangles).

(C) Differential temporal expression of rsmX, rsmY, and rsmZ. b-Galactosidase activities of the transcriptional rsmX-lacZ (pME7317; squares), rsmY-lacZ [pME6916 (18), diamonds], and rsmZ-lacZ [pME6091 (14), triangles] were determined in strain CHA0 (wild-type; filled symbols) and in the gacA mutant CHA89 (empty symbols).

Like rsmY and rsmZ (14, 15, 18), rsmX was silent in a gacA mutant and very poorly expressed in an rsmA
rsmE double mutant; in rsmA or rsmE single mutants, rsmX was transcribed almost normally (Fig. 3A). A crude preparation of a low-molecular-weight quorum-sensing signal, which is produced by strain CHA0 in late exponential phase and which induces the expression of rsmY (18) and rsmZ (14), also induced rsmX
transcription ~3-fold in strain CHA0 (Fig. 3B). By contrast, in a gacA (Fig. 3B)or rsmA rsmE mutant (data not shown), no induction was observed, suggesting that GacA, RsmA, and RsmE together are involved in the transduction of this signal [which are unrelated to N-acyl-homoserine lactones (24) and has not yet been identified chemically]. GacA-dependent expression of rsmX-lacZ and rsmY-lacZ occurred in parallel during the
entire growth cycle, whereas rsmZ-lacZ expression was delayed (Fig. 3C).

Regulation of Target Genes in the Gac/Rsm Cascade.

An rsmX rsmY rsmZ triple mutant, CHA1144, was constructed, by introducing a 64-nt deletion (Fig. 1A) into the rsmX gene in the DrsmYDrsmZ mutant CHA825 (Materials and Methods). In vitro growth of strain CHA1144 was indistinguishable from that of the wild-type CHA0 (Fig. 4A).

Fig. 4. Deletion of rsmX, rsmY, and rsmZ results in low expression of target genes similar to that found in a gacA mutant.

Fig. 4. Deletion of rsmX, rsmY, and rsmZ results in low expression of target genes similar to that found in a gacA mutant.

(A) Expression of a chromosomal hcnA'-'lacZ translational fusion and growth of the wild type (CHA207; squares), a gacA mutant (CHA89.207; open diamonds), an rsmX mutant (CHA1142; diamonds), an rsmY mutant (CHA823; circles), an rsmZ mutant (CHA811; triangles), an rsmY rsmZ mutant (CHA826; open squares), and an rsmX rsmY rsmZ mutant (CHA1145; open triangles).

(B) Expression of a chromosomal aprA'-'lacZ translational fusion and growth in the wild-type (CHA805; squares), a gacS mutant (CHA806; open diamonds), an rsmX mutant (CHA1143; diamonds), an rsmY mutant (CHA824; circles), an rsmZ mutant (CHA812; triangles), an rsmY rsmZ mutant (CHA827; open squares), and an rsmX rsmY rsmZ mutant (CHA1146; open triangles).

(C) Expression of a chromosomal rsmE'-'lacZ fusion in the wild type (CHA1134; squares), a gacA mutant (CHA1136; diamonds), and an rsmX rsmY rsmZ mutant (CHA1164; triangles). Growth is indicated by dotted lines.

Strain CHA1144 and the gacA mutant CHA89 showed equally low basal expression levels of typical biocontrol genes, hcnA and aprA (Fig. 4 A and B), which are needed for the production of hydrogen cyanide (HCN) and
the major exoprotease, respectively (35, 36). Similarly, both strains CHA1144 and CHA89 gave strongly reduced expression of the rsmE gene, by comparison with wild-type CHA0 (Fig. 4C). This finding was confirmed by monitoring the RsmE protein by Western blotting; in the rsmX rsmY rsmZ triple mutant, RsmE
levels were as strongly decreased (Fig. 7, which is published as supporting information on the PNAS web site, and Supporting Text), as they are in a gacA mutant (15). In the same experiment, we verified that the RsmA protein levels were not influenced by rsmX rsmY rsmZ expression. We also measured the GacA-controlled extracellular products HCN, AprA exoprotease, and 2,4-diacetylphloroglucinol (32) in culture supernatants; both strains CHA89 and CHA1144 were equally deficient, whereas mutations in rsmX alone or rsmY rsmZ had less drastic consequences (Table 1).

Overexpression of rsmX Can Compensate for Lack of rsmY and rsmZ.

It has been shown that artificial overexpression of rsmY or rsmZ suppresses the negative effects of gacS and gacA mutations on target gene expression (14, 18, 38). When rsmX was overexpressed from the tac promoter (on pME7320) in an rsmX rsmY rsmZ mutant background, the expression of an hcnA'-'lacZ fusion was restored to 170%, and that of an aprA'-'lacZ fusion to 30%, relative to the expression of these fusions in the wild type (set at 100%). Similarly, overexpression of rsmX suppressed gacS and gacA mutations in P. fluorescens (data not shown). These results indicate that RsmX, RsmY, and RsmZ can functionally replace each other to a considerable extent. Whether these small RNAs display some target specificity is still a matter of speculation (39).

An rsmX rsmY rsmZ Mutant Lacks Biocontrol Activity.

Strain CHA0 protects cucumber from disease caused by the oomycete P. ultimum in natural soil (18). The rsmX rsmY rsmZ mutant was devoid of such biocontrol activity (in terms of emerged plants and plant fresh weight) and was slightly impaired in root colonization in natural soil. The properties of a gacS mutant were
not significantly different in this experiment (Table 2).

Three Small RNAs Jointly Execute Posttranscriptional Regulation of Biocontrol Factor Synthesis.

Mutation of either gacS or gacA results in a 30- to 100-fold reduction of biocontrol factor expression in P. fluorescens CHA0 (13, 19, 24). This strong global effect on genes involved in exoproduct synthesis occurs
essentially at a posttranscriptional level and is superimposed on and epistatic over various types of transcriptional control of individual target genes (40). As we have shown here, a triple rsmX rsmY rsmZ mutation was needed to cause the same phenotypic effects on exoproduct formation and biocontrol as those observed in gacS or gacA mutants. It therefore appears unlikely that additional small RNAs would participate in the activation of exoproduct synthesis via the Gac/Rsm cascade. However, this does not exclude the possibility that GacS/GacA might control the expression of additional small RNAs having different functions.

Potential Benefits of Triple RNA Control.

RsmX, RsmY, and RsmZ of P. fluorescens have similar affinities for RsmA and RsmE (ref. 15; Fig. 2) and, in this respect, they appear to be functionally redundant. By contrast, in plant-pathogenic Erwinia spp., there appears to be only one GacA-controlled regulatory RNA termed RsmB, which has 479 nucleotides, thus about the size of RsmX, RsmY, and RsmZ together (17). The Gac/Rsm cascade controls the expression of pathogenicity factors in Erwinia spp.; both gacA and rsmB mutants are less virulent (17, 41). In E. coli, two GacA(=UvrY)-dependent small RNAs, CsrB and CsrC, are needed for regulation of central carbon metabolism and carbon storage (7, 11, 16). Only one RNA-binding protein (RsmA/CsrA) is sequestered by these small RNAs in Erwinia spp. as well as in E. coli. The question then arises what benefits P. fluorescens may obtain from having three GacA-controlled small RNAs (RsmX, RsmY, and RsmZ) as antagonists of two RNA-binding proteins (RsmA and RsmE). In principle, this arrangement may allow a more efficient regulatory response via a gene dosage effect. However, when any one of the three small RNA genes is deleted, the phenotypic consequences are minor (Table 1; Fig. 4 A and B; refs. 14 and 18). In addition, overexpression of rsmX, rsmY,or rsmZ alone from a strong vector promoter to a large extent overrides the negative effects of gacS and gacA mutations. Alternatively, differential expression of the rsmX, rsmY, and rsmZ genes may be advantageous, because it may allow fine tuning of the Gac/Rsm system in response to different environmental stimuli. Whereas rsmX and rsmY were expressed similarly under our in vitro conditions, expression of rsmZ was clearly distinct (Fig. 3C). Functional redundancy of small RNAs has also been observed in P. aeruginosa, where two Fur-controlled RNAs, PrrF1 and PrrF2, regulate the expression of genes involved in iron storage and resistance to oxidative stress. By contrast, the same function is carried out by a single homolog, RyhB, in E. coli (42).

Signaling Creates a Positive Feedback Loop in the Gac/Rsm Cascade of P. fluorescens.

This study provides evidence for an important feedback mechanism operating in the Gac/Rsm signal transduction pathway of P. fluorescens. The synthesis of the signal molecules that lead to activation of target gene expression in the Gac/Rsm cascade (14, 24) depends on GacA and on the three small RNAs (Table 1). The signal activates transcription of rsmX (Fig. 3B), rsmY (18), and rsmZ (14). Perception of the signal needs functional GacS (24). Biocontrol factors and signal are produced in parallel at the end of exponential growth (data not shown). By this mechanism, the Gac/Rsm cascade (summarized in Fig. 5) positively autoregulates its
activity as a function of increasing cell population densities.

Fig. 5. Model for gene regulation in the Gac/Rsm system of P. fluorescens.

Fig. 5. Model for gene regulation in the Gac/Rsm system of P. fluorescens.

—, direct or indirect regulation; =, physical interaction; ->, positive effect; -] , negative effect. A physical interaction of the signal with GacS has not been demonstrated experimentally.

Similar autoinduction patterns are common in quorum-sensing regulation depending on N-acyl-homoserine lactone signaling (43). As Vibrio spp. (6), P. fluorescens CHA0 has a set of apparently redundant small RNAs that have a key function in cell–cell communication and in the regulation of extracellular products.

We thank Patrick Michaux and Christoph Keel for help with plant experiments and Lionel Guy for help with statistics. We also thank Cornelia Reimmann (Department of Fundamental Microbiology, University
of Lausanne, Lausanne, Switzerland) for providing plasmid pME7011 and for critically reading the manuscript. Support from the Swiss National Science Foundation (project 3100A0-100180) and European Union project Nanofoldex (QLK3-CT-2002-0286) is gratefully acknowledged.

http://www.pnas.org/cgi/content/full/0505673102/DC1

Fig. 6. Determination of the 5'-end of the rsmX transcript by RACE.

rsmX cDNA was obtained after reverse transcription of CHA0 total RNA with primer PRACE1 (Table 3), 3'-anchoring with oligonucleotide DT88 and PCR with primers DT89 and PRACE2 (underlined). DNA sequences of three independent rsmX-cDNA clones are shown. Nucleotides in bold correspond to the 62-nt rsmX cDNA obtained by RACE. The 5'-end of RsmX is deduced from the sequence after subtraction of the DT89 sequence (gray background). pGEM-T, sequence of the cloning vector pGEM-T (Promega)

Fig. 7. Cellular levels of the RsmA and RsmE proteins in CHA0 and CHA1144.

The samples for immunoblot analysis were taken at increasing cell densities. The protein load in each well was similar (see Supporting Text).

 

Table 3. Oligonucleotides used in this study.
 

Oligonucleotides (5’-3’)	Description
PTX3hind	GAACAAGCTTTTTGCGTAAGTGCGGTATAG, HindIII site underlined
PTX4eco	TGTTGAATTCCATGAAGTATCAAAGTGGCC, EcoRI site underlined
POVEX	TTTCCCACAACTGCAGCGCAGGAAGCGC
T3	AATTAACCCTCACTAAAGGG
T7	GTAATACGACTCACTATAGGGC
PRX1hind	CATAAAGCTTGTGCCACCCTGCAACAGTTG, HindIII site underlined
PRX2eco	CGCTGAATTCTTCCGGGCTGTCGAAAAGCG, EcoRI site underlined
PMUBAM1	GTCAGGATCCTCTTGGAAAAAACCCGCTTC, BamHI site underlined
PMUBAM2	GCTTGGATCCCTGCAGTTGTGGGAAAGAAG, BamHI site underlined
PROF	TGCCGAATTCCACCGACCCTGTGCGCTTTG, EcoRI site underlined
PROR1	CAGTGGATCCAAAGAAGATTAAACACAAAG, BamHI site underlined
PTZhind	TAATAAGCTTTCCCACAACTGCAGCGCAGG, HindIII site underlined
PTZeco	ATGCGAATTCAAAAAAAACCCGCCGAAGCG, EcoRI site underlined
DT88	GAAGAGAAGGTGGAAATGGCGTTTTGG
DT89	CCAAAACGCCATTTCCACCTTCTCTTC
PRACE1	CCAAGACCATTATGACTTCCTGTCTGCAGC
PRACE2	ATCCGTGCTAATGGTCGACC

Determination of 5'-end. The 5'-end of the rsmX transcript was mapped by RACE as previously (1). A reverse transcription reaction was performed using total RNA from strain CHA0 isolated in late stationary phase and the Omniscript RT kit (Qiagen, Valencia, CA). After RNA removal by alkaline hydrolysis and neutralization, cDNA was precipitated and dissolved in sterile water. The 5'-phosphorylated, 3'-end ddATP-blocked oligonucleotide DT88 was ligated to the single-stranded cDNA, and the resulting ligation mixture was used directly in the subsequent seminested PCRs. The anchor-ligated cDNA was amplified first with primers DT89 (anchor-specific primer) and PRACE1 (rsmX-specific primer) and then with the internal primer PRACE 2 (Table 3). An 82-bp band was purified from an agarose gel, made blunt-ended, and cloned into the cloning vector pGEM-T, as recommended by the manufacturer (Promega). Three independent clones were sequenced using primer T7 and reverse primer (Fig. 6).

Detection of RsmA and RsmE by Western Blotting. Cells were grown in 20 ml of nutrient yeast broth (NYB) amended with 0.05% Triton X-100 at 30°C with shaking and aliquots equivalent to an optical density at 600 nm (OD₆₀₀) of 0.4 units per ml were harvested at several time points after inoculation. Cells were washed with 0.9% NaCl and dissolved in 20 ml of loading buffer (50 mM Tris·HCl, pH 6.8/2% SDS/0.1% bromophenol blue/15% glycerol/5% 2-mercaptoethanol) and immediately treated at 100°C for 10 min. Fifteen microliters of each sample was loaded onto a 16% acrylamide-bisacrylamide gel containing tricine and SDS, then electrophoresis and electrotransfer of proteins were performed as described (2). The RsmA and RsmE proteins were detected with polyclonal antibodies raised against purified Yersinia enterocolitica RsmA (kindly provided by Paul Williams, University of Nottingham, Nottingham, U.K.) and secondary antibody coupled to peroxidase. Membranes were developed with the enhanced chemiluminescence Western blotting analysis system (Amersham Pharmacia) following the manufacturer’s instructions (Fig. 7).

1. Tillett, D., Burns, B. P. & Neilan, B. A. (2000) Biotechniques 28, 448-456.

2. Reimmann, C., Valverde, C., Kay, E. & Haas, D. (2005) J. Bacteriol. 187, 276-285.

1. Song X, Sun Y, and Garen A, "Roles of PSF protein and VL30 RNA in reversible gene regulation".

2. Kuwabara T, Hsieh J, Nakashima K, Warashina M, Taira K, and Gage FH, "The NRSE smRNA specifies the fate of adult hippocampal neural stem cells".

3. Andolfatto P, "Adaptive evolution of non-coding DNA in Drosophila".

4. Ling J, Baibakov B, Pi W, Emerson BM, and Tuan D, "The HS2 Enhancer of the b-globin Locus Control Region Initiates Synthesis of Non-coding, Polyadenylated RNAs Independent of a cis-linked Globin Promoter".

5. Hovsepian JA, and Frenster JH, "Sense and Antisense during RNA Initiation of the DNA Transcription Bubble".

Links to RNA and Biological Causality:

Links to Euchromatin Activator RNA Reviews:
Links to Euchromatin Activator RNA Research:
Links to Ultrastructural Probes of DNase I-Sensitive Sites:
Links to RNA as a Therapeutic Agent:
Links to Hodgkin Lymphoma Immuno-Pathology:
Links to Activated T-Lymphocyte Immunotherapy:
Links to Medical Systems Biology:
Links to Selective Gene Transcription:
Links to RNA-Induced Epigenetics:
Links to RNA-Induced Embryogenesis:
Links to RNA and Biological Causality:
Links to Reprogramming and Neoplasia:

A Brief History of Activator RNA:

"Ultrastructural Probes of Active DNA Sites, and the RNA Activators of DNA". (PowerPoint Presentation).