• We are sorry, but NCBI web applications do not support your browser and may not function properly. More information
Logo of nihpaAbout Author manuscriptsSubmit a manuscriptNIH Public Access; Author Manuscript; Accepted for publication in peer reviewed journal;
J Mol Biol. Author manuscript; available in PMC Feb 11, 2012.
Published in final edited form as:
PMCID: PMC3030618

RNase III Participates in GadY-Dependent Cleavage of the gadX-gadW mRNA


The adjacent gadX and gadW genes encode transcription regulators that are part of a complex regulatory circuit controlling the Escherichia coli response to acid stress. We previously showed that the small RNA GadY positively regulates gadX mRNA levels. The gadY gene is located directly downstream of the gadX coding sequence on the opposite strand of the chromosome. We now report that gadX is transcribed in an operon with gadW, though this full-length mRNA does not accumulate. Base pairing of the GadY small RNA with the intergenic region of the gadX-gadW mRNA results in directed processing events within the region of complementarity. The resulting two halves of the cleaved mRNA accumulate to much higher levels than the unprocessed mRNA. We examined the ribonucleases required for this processing, and found multiple enzymes are involved in the GadY-directed cleavage including the double strand RNA specific endoribonuclease RNase III.

Keywords: antisense RNA, acid response, ribonuclease, OOP RNA


Regulatory RNAs in organisms from all kingdoms act by a wide array of mechanisms that enable them to play major roles in controlling gene expression. In bacteria, these regulatory RNAs are typically small in size (50-300 nucleotides) and therefore have been termed small RNAs (sRNAs). Much of what is currently known about sRNA function in bacteria has arisen from studies of the Gram negative model organism Escherichia coli in which around 80 sRNAs have been identified.1 While not all of these sRNAs have been characterized, great strides have been made in understanding their mechanisms of action (reviewed in 2). The largest class of sRNAs elicits regulatory effects through the formation of base pairing interactions with target mRNAs. Many of these sRNAs are encoded in trans with respect to their target mRNAs and contain only limited complementarity with the targets. In these cases, the RNA binding protein Hfq is usually required for sRNA function. Approximately one third of all E. coli sRNAs bind to Hfq, and it is thought that each of these sRNAs acts by base pairing with trans-encoded mRNA targets. Four major regulatory outcomes have been established for the Hfq dependent RNAs: inhibition or activation of translation or increased or decreased degradation of target mRNAs.

A limited number of base pairing sRNAs that are encoded on the strand opposite annotated genes on bacterial chromosomes, and thus in cis to their targets, have also been characterized.3 Due to this genetic arrangement, the sRNA and the target mRNA have the potential for extensive base pairing across the region of overlap. In these cases, the Hfq protein generally has not been found to be required for base pairing. Most of the cis-encoded sRNAs with known functions negatively regulate translation and promote the degradation of the complementary mRNA.

We have been characterizing the 105 nucleotide E. coli GadY RNA, which is encoded in cis to the 3′ end of the gadX gene in the ~371 nucleotide intergenic region between gadX and gadW (Fig. 1a). Intriguingly, this sRNA was found to positively regulate gadX mRNA levels.4 The gadX mRNA encodes a transcription regulator that is involved in a highly complex regulatory circuit controlling the response to acid stress.5;6 The gadW gene located immediately downstream from gadX also encodes a regulator of the acid response and can be transcribed with gadX or from its own promoter as an independent transcript.5; 7; 8 We confirm here that gadX is transcribed as a two gene mRNA with gadW. When the GadY sRNA is present, the gadX-gadW mRNA is processed to give rise to products with ends within the region complementary to GadY. Part of this processing is due to the GadY-dependent RNase III cleavage. However, GadY pairing with the gadX-gadW mRNA also modulates cleavage by other ribonucleases in mutants lacking RNase III.

Fig. 1
GadY RNA-directed mRNA cleavage. (a) Diagram of the gadXWY region. The gadY promoter mutation is indicated by the X. (b) Levels of the gadX (~1.0 kb) and gadW (~1.1 kb) mRNAs in cells without a plasmid, carrying the vector control (pRI), ...


Region of complementarity confers GadY-dependent regulation to a heterologous gene

We previously showed that the GadY RNA positively regulates gadX and that this regulation requires sequences in the 3′ untranslated region (UTR) of the gadX mRNA4 (see also Fig. 1b). The 3′-UTR sequences of gadX needed for regulation are perfectly complementary to GadY due to the cis-encoded nature of these two RNAs, and it was proposed that base pairing within this region is necessary for regulation to occur.

We further investigated the mechanism of regulation by testing if the GadY RNA was capable of regulating a heterologous gene tagged with the 3′-UTR of gadX. A hybrid gene was constructed by replacing the gadX coding sequence with the coding sequence of the chloramphenicol resistance gene cat such that the sequences downstream of the gadX coding sequence will be expressed as part of the cat mRNA (Fig. 1c). In addition, the gadX promoter was replaced with the cat promoter to eliminate the wild type transcriptional regulation. To abolish expression of the chromosomally-encoded GadY RNA, the gadY promoter mutation described previously (TATATT -10 sequence replaced by GGGGGG) was also introduced into the strain. We then examined the levels of cat mRNA without and with GadY overexpression from a plasmid (Fig. 1d). The Northern blots showed that the gadX 3′ UTR is sufficient to confer GadY-dependent regulation on the heterologous cat gene. The strain carrying the pRI control vector had low levels of a band corresponding to the expected ~0.8 kb size of the cat mRNA. Multiple faint bands corresponding to longer transcripts that hybridized with the cat gene probe were also observed for this control strain. In contrast, only the ~0.8 kb cat mRNA was detected in the strain overexpressing GadY, and this transcript was present at much higher levels.

GadY RNA directs processing

The longer cat transcripts observed in the absence of GadY suggested that transcription of the reporter was proceeding through cat to downstream sequences. In the presence of the GadY RNA these possible read-through transcripts were no longer detected. To confirm the read-through transcription, we made an additional modification to the cat reporter construct. A promoterless gfp gene encoding the green fluorescent protein was inserted into the chromosome downstream from cat and separated from the cat gene by the sequences complementary to GadY such that the native promoters for gadW and gadY were eliminated. The 3′ UTR of gadX, encompassing the region of base pairing with GadY, remained intact. The gfp gene was also engineered to contain rrn terminators downstream from the coding sequence so that any transcripts reading through gfp would be forced to terminate (Fig. 1e). We assayed for gfp expression in the absence or presence of GadY by examining cells under a fluorescent microscope. Unexpectedly, the reporter strain was fluorescent under both conditions indicating that transcription initiating at the cat promoter was proceeding through gfp regardless of GadY expression (data not shown). We then examined each gene in the reporter operon by Northern blot analysis in strains carrying the control vector or overexpressing GadY. When GadY was absent, a single hybridizing mRNA was detected when the Northern blots were probed with the cat specific probe (Fig. 1f). The size of this mRNA (~1.7 kb) was consistent with a two-gene transcript encompassing both cat and gfp (Fig. 1e). When GadY was overexpressed a single hybridizing mRNA was detected, but now the size of the transcript (~0.8 kb) was consistent with a single gene mRNA encompassing only the cat gene (Fig. 1e). Similar results were seen when the same total RNA samples were probed with a gfp-specific oligonucleotide. In the absence of the GadY RNA, a longer mRNA encompassing both cat and gfp was detected while expression of GadY resulted in a smaller transcript (~0.9 kb) encompassing only gfp (Fig. 1f). These results suggest that expression of GadY causes a processing event in the mRNA intergenic region between the two reporter genes.

To test whether the GadY sRNA also was directing processing of the wild type gadX-gadW mRNA, we performed Northern analysis with a probe to detect gadW in a strain with the chromosomal gadY promoter mutation transformed with the control vector or the GadY overexpression plasmid (Fig. 1b). In the absence of GadY overexpression, we saw low levels of a gadW mRNA (lower band) whose size (~1.1 kb) corresponds to a transcript originating from the previously identified gadW specific P1 promoter7, which we confirmed by primer extension analysis (data not shown). When GadY was overexpressed, higher levels of the first transcript were detected together with a second slightly larger transcript. The size of the slightly larger mRNA (~ 1.5 kb) is consistent with the gadW half of a gadX-gadW transcript that could be obtained upon GadY-dependent processing. The increase in the lower band could also be the result of increased gadW transcription due to increased GadX levels or could correspond to other gadX-gadW fragments derived from processing. We note that we did not detect a clear band corresponding to the full-length gadX-gadW mRNA (~2.3 kb) in the presence or absence of the GadY RNA indicating that the two fragments resulting from processing are capable of accumulating to higher levels than the unprocessed mRNA.

To establish an assay in which the effects of GadY could easily be detected, we also constructed a reporter strain in which the 105-nucleotide gadX 3′ UTR sequence complementary to GadY was placed in-frame in the amino-terminal portion of the lacZ coding sequence in the E. coli chromosome (Fig. 2a). The gadX 3′ UTR sequence, which is not normally translated, contained at least one nonsense codon in each of the three possible frames. Therefore a stop codon was eliminated from one of the frames by introduction of an A→G transition. This single mutation maintains base pairing potential since the introduced G residue is capable of pairing with the corresponding U nucleotide in wild type GadY. Our expectation was that in the absence of GadY, the full-length lacZ mRNA would be generated allowing the production of β-galactosidase, while in the presence of GadY, the mRNA would be cleaved preventing the production of β-galactosidase. The GadY sRNA was previously shown to exist in three stable forms, a full-length 105-nucleotide form and two processed fragments of 90 and 59 nucleotides.4 We tested the ability of all three forms of GadY to regulate this reporter gene by examining β-galactosidase activity on indicator plates containing X-gal (Fig. 2b). Colonies of this reporter strain were blue in the absence of GadY, but were white when full-length GadY was overexpressed and slightly blue when GadY(90) or GadY(59) were overexpressed. These results are consistent with the conclusion that all three forms of GadY direct processing of the lacZ reporter mRNA.

Fig. 2
(a) Map of the lacZgadX reporter. (b) Plate with lacZgadX reporter strain (GSO404) harboring pRI, pRI-GadY, pRI-GadY(90), pRI-GadY(59). The strains were streaked on an LB plate containing 100 μg/ml of X-gal. At OD600 = 0.7, the levels of β-galactosidase ...

Processing sites are within region of complementarity

GadY-dependent processing was examined more precisely by carrying out primer extension analysis to identify the 5′ ends of the gadX-gadW cleavage products in a strain carrying the gadY promoter mutation (Fig. 3a). We also mapped the 5′ ends of the cleavage products in the lacZ reporter strain (Fig. 3b). Although a few bands were different between the two strains, the majority of the 5′ ends resulting from GadY-dependent processing were identical between gadX-gadW and the lacZ reporter. The strains constitutively overexpressing the full-length GadY RNA displayed four of the same products. The relative intensity of the bands varied somewhat between experiments but all corresponded to 5′ ends within the region of base pairing with GadY. These 5′ ends are within 1-3 nucleotides of the gadW ends denoted P3 and P4 by Tramonti et al.7 We also tested the ability of the two shorter forms of GadY to affect processing. Both the 90 and 59 nucleotide forms of GadY led to processing of the gadX-gadW mRNA and the lacZ reporter at multiple nucleotides, but interestingly, the pattern of primer extension products observed is different for each form of GadY. In each case, most of the 5′ ends of the products correspond to cleavage sites within the region of base pairing near the 5′ end of that particular form of GadY.

Fig. 3
All three forms of GadY direct processing of its complementary sequence. (a) Primer extension analysis to map 5′ ends of the processed gadX-gadW mRNA. (b) Primer extension analysis to map 5′ ends of the processed lacZgadX reporter mRNA. ...

Processing is not abolished in endoribonuclease mutants

Given that the same cleavage products were observed for the lacZgadX construct and the native gadX-gadW and that the lacZ products were easier to detect, we utilized the reporter strain to test the roles of five endoribonucleases in E. coli (reviewed in 9) in GadY-dependent processing. We first assayed strains carrying single mutations in the genes encoding RNase III (rnc), RNase E (rne), RNase G (rng), RNase BN (elaC) and RNase P (rnpA). Strains lacking the non-essential ribonucleases, rnc (RNase III), rng (RNase G) and elaC (RNase BN), were grown under conditions that allowed constitutive expression of the lacZ reporter. Expression of the three different forms of GadY was induced for 20 min from the arabinose-inducible PBAD promoter, whereupon the 5′ ends of the lacZ reporter mRNA were mapped by primer extension (Fig. 4 and data not shown). GadY-dependent processing in the rng and elaC mutants was identical to the wild strain for all three versions of GadY, though an additional band was observed for all samples, including the vector control strain, for the rng mutant. Although processing was not eliminated in the rnc mutant, the level of one band was decreased for strains overexpressing full-length GadY (indicated by the asterick). We used a modified protocol to test the effects of temperature-sensitive mutations in the essential rne and rnpA genes encoding RNase E and the protein component of the RNase P complex, respectively. In this assay the temperature sensitive mutant strains were grown at the permissive temperature under conditions that promoted constitutive expression of the reporter gene. The strains were then shifted to the nonpermissive temperature for 30 min to deplete the cells of active RNase E or RNase P. GadY was then induced for 20 min while the cells remained at the nonpermissive temperature, and total RNA isolated from these samples was examined by primer extension. Neither of these mutations had an affect on the ability of GadY to direct processing of our reporter gene (data not shown). In addition, we examined the effects of GadY-dependent processing in strains carrying different combinations of these mutations including several double mutants and the triple mutant strains rnc rne rng, rnc rne elaC, rnc rng elaC and rne rng elaC. The rnc mutation was also combined with a mutation in the rnlA gene encoding the RNase LS endonuclease. We did not detect any effects on GadY-dependent processing beyond the band decreased in rnc mutants (Fig. 4, Fig. S1 and data not shown).

Fig. 4
GadY-directed processing in RNase mutant strains. Total RNA was isolated from wild type and RNase mutant (rnc, rnc rng, rnc rne, rnc rnpA) derivative of the lacZgadX reporter strain harboring pBAD-RI, pBAD-GadY, pBAD-GadY(90) or pBAD-GadY(59) 20 min following ...

In vitro processing of gadX

Since processing was not abolished in the endonuclease mutant strains, we established an in vitro assay to determine whether GadY itself was catalytic or whether a protein was required for the GadY-dependent cleavage. For this assay, a 300-nucleotide gadX mRNA fragment containing the entire sequence complementary to GadY was synthesized in vitro and labeled throughout its length with 32P (Fig. 5a). This gadX mRNA fragment, pre-incubated with or without unlabeled in vitro-synthesized GadY RNA, was mixed with a protein lysate. When GadY was added in the absence of lysate, no cleavage was detected (Fig. 5b, lane 2) suggesting that GadY itself was not catalytic and that an additional factor was needed. When only the lysate was added to the reaction, the gadX transcript was almost completely digested (Fig 5b, lane 3). However, when both GadY and the lysate were added, a specific processed fragment of gadX was detected (Fig 5b, lane 4). The size of this GadY-dependent fragment (~140 nucleotides) suggested it could correspond to the 5′ portion of the gadX mRNA fragment. Although only the 140 nucleotide fragment was visible in Fig. 5b, additional GadY-dependent fragments centered around 100 nucleotides in size corresponding to the 3′ portion of the gadX mRNA were detected when the reaction products were analyzed by primer extension analysis (data not shown). The fragments were not generated when the protein lysate was heat-treated prior to incubation in the reaction indicating that the factor required for GadY-dependent processing is likely a protein (Fig. 5b, lane 6).

Fig. 5
GadY can direct processing of gadX in vitro. (a) Diagram of the gadX and GadY constructs used for the in vitro assay. Black arrows indicate the in vitro synthesized gadX and GadY RNAs. Gray bars indicate the 140 nt 5′ and 100 nt 3′ fragments ...

RNase III is partially responsible for GadY-dependent processing in vitro

We then set out to purify the processing activity from an E. coli lysate. A six-step purification process was established with the activity monitored at each purification step (see Materials and Methods). After the third purification step, we detected a second major RNA fragment of 100 nucleotides in the in vitro processing reaction, in addition to the 140-nucleotide fragment detected when the assay was carried out with the lysate. The size of the smaller fragment corresponds with the expected size of the 3′ gadX fragment detected by the primer extension assay carried out for the lysate sample (see above). We attribute the appearance of this second fragment to increased stability resulting from the elimination of exonucleases during purification. The activity assays and protein profile for the column fractions for the last purification step are shown in Fig. 6. The protein band that co-purified with the activity was excised from the gel and analyzed by mass spectrometry. This analysis revealed that the co-purifying protein was the double strand specific-endoribonuclease RNase III.

Fig. 6
RNase III is partially responsible for GadY-directed processing. (a) Processing activity from column fractions from final purification step. The 32P-labeled gadX mRNA fragment was incubated with GadY and 5 μl of Mono S column fractions for 1 hr ...

Fragments yielded by GadY-dependent RNase III processing

The in vitro processing assay with our partially purified RNase III protein yielded two fragments of 140 and 100 nucleotides, which do not add up to the size of the full-length 300 nucleotide gadX mRNA template. Primer extension analysis showed that the 100-nucleotide fragment corresponds to the 3′ portion of the processed RNA with the most prominent primer extension product corresponding to the rnc-dependent band observed for in vivo cleavage products (Fig. S2). Other bands detected in our primer extension assays carried out on the products generated by the purified fraction correspond to other 5′ ends mapped in vivo, as well as three nearby nucleotides. To more accurately define the 140-nucleotide fragment, we also carried out 3′ RACE analysis (Fig. S2). The gadX mRNA that was treated with our purified RNase III in the absence of GadY had a 3′ end that mapped to the terminal 3′ end of the full-length in vitro synthesized gadX RNA. However, when full-length GadY was added to the reaction two different 3′ ends separated by 4 nucleotides were identified. These ends were centered around 140 nucleotides downstream from the terminal 5′ end of the in vitro synthesized gadX RNA in agreement with the 140 nucleotide fragment detected. Other fragments that correspond to the central region of the gadX fragment, that may or may not be degraded, were not detected by our primer extension and 3′ RACE analysis.

Fig. 3 showed that all three forms of GadY could direct processing, although the 5′ ends generated are different. To test whether our in vitro assay mimics the in vivo results, we also mapped the 5′ and 3′ ends of the fragments generated by our purified RNase III fraction when GadY90 and GadY59 were added to the reaction. The inclusion of the two shorter forms of GadY in the in vitro assay resulted in essentially the same GadY(90)- and GadY(59)-specific 5′ ends detected by primer extension for the in vivo products, with some additional nearby ends (Fig. S2). The 3′ ends corresponding to the 140-nucleotide fragment generated by GadY90 and GadY59 were identical to those observed for full-length GadY (Fig. S2). Taken together these results indicate that RNase III in conjunction with any of the three forms of GadY processes the gadX mRNA. The 140-nucleotide 5′ fragment generated is common for all three forms of GadY, while the 3′ fragments were specific to each form of the GadY RNA.

Processing detected in rnc mutant strain

In Fig. 4, we noted that the rnc (RNase III) mutant strain had reduced levels of one of the full-length GadY-dependent bands. However, this mutant did not abolish processing indicating redundancy with respect to enzymes capable of catalyzing GadY-directed processing. We further confirmed these results by testing for processing of the lacZ reporter in an rnc mutant strain on X-gal indicator plates (Fig. 7a). As also seen in Fig. 2b, colonies of the wild type strain overexpressing GadY were white on the plates demonstrating efficient processing of lacZ. In the absence of GadY, the colonies were dark blue consistent with a lack of processing. Overexpression of GadY in the rnc mutant strain resulted in colonies that were pale blue indicating that processing still occurs, although slightly less efficiently than in the RNase III proficient strain.

Fig. 7
GadY-directed processing in the absence of RNase III. (a) Plate with lacZ reporter strain harboring pRI, pRI-GadY in wild type and rnc mutant backgrounds (GSO404 and GSO405, respectively). (b) Levels of the cat (~0.8 kb, 1.7 kb) and gfp (~0.9 ...

We also examined the processing of the cat-gfp reporter construct by Northern analysis (Fig. 7b). Again, GadY-dependent processing was still observed. However, some differences were noted. While the full-length cat-gfp transcript was completely eliminated when GadY was overexpressed in the wild type strain, approximately equal amounts of the cat-gfp transcript (~1.7 kb) were detected in the presence and absence of GadY overexpression in the rnc mutant. In addition, the cat transcript resulting from GadY-dependent processing migrated at a higher molecular weight (~0.85 kb) for RNA extracted from the rnc mutant strain. Using 3′ RACE we determined that the cat transcript was ~20-50 nucleotides longer in the rnc mutant strain compared to the wild type strain (Fig. S3). Primer extension analysis also showed that the 5′ end of the gfp transcript was ~10 nucleotides shorter in the rnc mutant when the full-length GadY was overexpressed but remained the same when GadY(59) was overexpressed (Fig. 7c and Supplemental Fig. S3). Thus, although processing of the cat-gfp transcript is still GadY-dependent and occurs in the region of base pairing in an rnc mutant strain, the 5′ and 3′ ends are different indicating that the processing sites have shifted.

Finally, we prepared extracts from the rnc mutant strain and examined GadY-directed processing in vitro. Again, heat-sensitive GadY-dependent processing was observed in the protein extract (Fig. 7d). As for the in vivo assays, the products observed for the rnc mutant extracts were of somewhat different size than the products observed for the wild type extracts. A different subset of 5′ ends was also mapped by primer extension analysis (Fig. S4). Taken together, these results indicate that a nuclease in addition to RNase III is responsible for GadY-directed gadX-gadW dependent cleavage and that this enzyme cleaves at different nucleotides. We have not yet been able to identify this endonuclease despite attempts at genetic screens as well as biochemical purification.


GadY is a member of a class of regulatory RNAs encoded on the DNA strand opposite their target RNAs. Initially, these cis-encoded RNAs seemed most abundant on extra chromosomal elements such as plasmids and bacteriophage, however an increasing number of antisense sRNAs are being found to be encoded on bacterial chromosomes. These cis-encoded RNAs can regulate gene expression by numerous mechanisms including transcription inhibition and attenuation and modulating mRNA stability and translation (reviewed in 3; 10). In many cases the sRNA is encoded in a region the overlaps the 5′ end of the regulatory target gene. Three examples of this are symE-SymR of E. coli, copT-CopA from plasmid R1 and tnp (RNA-IN)-RNA-OUT from the IS10 transposon. In these examples, base pairing between the antisense RNA and the 5′ end of the target mRNA results in translational repression of symE and tnp as well as RNase III-dependent destabilization of copT and tnp.11; 12; 13 Unlike SymR, CopA and RNA-OUT, the E. coli GadY RNA is encoded such that the gene overlaps the intergenic region of a two-gene operon and is one of only a few known cases where the sRNA is encoded opposite the middle part of a larger mRNA. Two other examples where a cis-encoded small RNA base pairs internal to an mRNA are the OOP RNA from phage λ and RNAα from Vibrio anguillarum pJM1 plasmid.14;15

The genetic arrangement of GadY and gadX-gadW is most reminiscent of the OOP RNA and the cII-O operon of phage λ. The OOP RNA is transcribed from sequences opposite the intergenic region of the cII-O operon and extends to sequences opposite the 3′ end of cII. The OOP RNA was shown to be responsible for discoordinate regulation of the cII and O mRNAs.14; 16 This regulation is achieved by an OOP-dependent processing event mediated in part by the double stranded endonuclease RNase III. This initial cleavage is followed by degradation of the upstream cII fragment while the downstream O mRNA remains stable. We found that GadY acts similarly; base pairing of GadY with the gadX-gadW mRNA initiates processing mediated in part by RNase III. Processing directed by both OOP and GadY was shown to occur in regions of the target gene that are complementary to the sRNA.

The OOP and GadY RNAs also are similar in that they can direct processing of the target mRNA by a mechanism that is RNase III independent. In both cases, the alternative processing sites are detected adjacent to the major RNase III processing sites in backgrounds lacking RNase III. This secondary processing mechanism has not been elucidated for the OOP RNA. We speculate that the mechanism of RNase III-independent processing directed by OOP will be similar to the RNase III-independent processing directed by GadY. Our mutant analysis suggests that this additional ribonuclease may be an enzyme that has not yet been identified since mutations affecting RNase E, RNase G, RNase BN and RNase P, the putative YbeY RNase17, as well as five of the endonucleases that are components of the type II toxin-antitoxin systems18, did not affect GadY-dependent cleavage (Fig. 4, Fig. S1, and data not shown). Given that the levels of the RNase III-dependent band are increased in the rne mutant strain (Fig. S1), it is possible that the GadY-dependent cleavage involves a cascade of events. It is also conceivable that some of the observed GadY-dependent products arise when GadY base pairing protects gadX 3′ ends against the action of exonucleases. An important direction for future studies will be to identify the additional ribonuclease(s) involved in OOP- and GadY-dependent, RNase III-independent target mRNA cleavage.

GadY and OOP do differ in some respects. The OOP RNA partially overlaps the cII coding sequence at its 3′end whereas the overlap between GadY and gadX-gadW is contained completely within the intergenic region of the mRNA. This difference results in alternative outcomes for the target mRNA. Because OOP overlaps the cII coding sequence and processing was shown to occur in this region, the net outcome of OOP regulation is to downregulate cII expression while O expression levels remain relatively stable. Negative regulation is also seen for the target of RNAα from Vibrio anguillarum where the sRNA overlaps a large portion of the coding sequence of fatB.15 In contrast, GadY directs processing downstream from the gadX stop codon, most likely leading to increased stability of the both the gadX half and gadW half of the mRNA.

We currently do not understand how the GadY-dependent processing event achieves this positive regulation but some of our observations could be providing clues as to how the proposed stabilization might occur. Unprocessed gadX-gadW mRNA was nearly undetectable by Northern blot analysis and is likely to be unstable. The determinants for this instability are likely contained within both gadX and gadW; the cat-gadW transcripts were more easily detectable by Northern blot analysis than the gadX-gadW mRNA, while the cat-gfp band was even more distinct (Figures 1d and 1f and data not shown). GadY could remain base paired at the processed 3′ end of gadX and block recognition of instability determinants present either directly at or some distance from the 3′ end, thus increasing the stability of the mRNA. Alternatively, GadY directed cleavage could be leading to the removal of the instability determinants. Finally, the full-length gadX-gadW mRNA and the transcripts produced by GadY-dependent cleavage could fold into different secondary structures that have different susceptibility to degradation.

We predict that the GadY RNA will not be the only cis-encoded RNA that acts to increase the stability of a message. Most cis-encoded sRNAs that have been examined thus far either block translation of their target and/or lead to the destabilization of the target. However, only a few chromosomal cis-encoded RNAs have been studied in detail, and numerous putative cis-encoded sRNAs are being discovered using deep sequencing techniques.3 While further experimentation is needed, it is possible that some of these antisense RNAs will act similar to the GadY or OOP RNAs to direct target mRNA cleavage.

Materials and Methods


All oligonucleotides and plasmids used in this study are given in Supplemental Tables S1 and S2, respectively. To generate pRI-GadY(90) and pRI-GadY(59), the corresponding regions of the gadY gene were amplified using PCR (using primers GadY-S1-gadY-A1 and JK117-JK118, respectively), digested with EcoRI and HindIII and cloned into the corresponding restriction sites of pRI.4 A second EcoRI restriction site at the +1 position of the PBAD promoter of pBAD18-amp19 was introduced by Quick-Change mutagenesis (Stratagene) (using primers AZ784 and AZ785) generating pBAD-RI. The gadY regions corresponding to the full-length sRNA as well as the 90 and 59 nucleotide fragments were amplified by PCR (using primers JK116-JK118, JK117-JK118 and GadY-S1-JK118, respectively), digested with EcoRI and HindIII and cloned into the corresponding restriction sites of pBAD-RI to create pBAD-GadY, pBAD-GadY(90) and pBAD-GadY(59).

The lacZ gene was PCR amplified (using primers lacZ-S3 and lacZ-A3) and cloned into the BamHI and HindIII restriction site of pACYC18420 to create plasmid pACYC-lacZ. To introduce the region complementary to GadY, the entire pACYC-lacZ plasmid was amplified with primers lacZ-S10 and lacZ-A6. These primers hybridize directly adjacent to each other in the lacZ coding sequence and carry 6-nucleotide in frame extensions corresponding to the BglII and XhoI restriction sites, respectively. The 105 nt region corresponding to the gadY sequence was amplified and a stop codon was eliminated by first amplifying two halves of the sequence overlapping at the position of the point mutation. The 5′ half of the sequence was PCR amplified with primers gadY-QC2 and gadX-S1 and the 3′ half of the sequence was PCR amplified with primers gadY-QC1 and gadW-A1. These two PCR fragments were fused by overlapping extension as described21 using primers IG-S1 and IG-A2. The modified complementary GadY fragment was cloned into the BglII and XhoI restriction sites of the pACYC-lacZ PCR fragment, creating a new plasmid named pACYC-lacZgadX.


All of the strains used in the study are listed in Supplemental Table S3. All of the reporter strains were generated using the mini-λ-red recombination method22. In all cases, the mini-λ-TetR located on the chromosome of NM500 (kind gift of N. Majdalani) was P1 transduced into the relevant strain.

For the cat substitution of gadX the cat gene from pACYC18420 was amplified by PCR (using primers Cat-S2 and PxCat-S1) and recombined into the gadX region of GSO109 carrying mini-λ-red. Chloramphenicol-resistant recombinants that no longer contained the mini-λ or that still retained the mini-λ were denoted GSO129 and GSO129-mini-λ-TetR, respectively. For the gfp substitution of gadW the gfpmut3.1 gene from plasmid pMG4723 was amplified with primers gfp-S3 and gfp-A2. The kan gene from plasmid pKD424 was amplified with primers pKD4-S2 and pKD4-A2. The gfp DNA fragment and the kan DNA fragment were fused by splicing by overlapping extension as described21 with primers gfp-S2 and PS2-X-A1 and inserted into strain GSO129-mini-λ-TetR creating GSO403.

For the lacZ reporter strain, the modified lacZ gene from pACYC-lacZgadX was PCR amplified with primers lacZ-S3 and lacZ-234-R. The kan resistance gene was amplified from pKD4 as above. The lacZ PCR fragment and kan PCR fragment were fused by splicing by overlapping extension as described21 with primers GadX-LacUP and PS2-X-A1 and recombined into the gadX locus in strain NM700 (kind gift of N. Majdalani). The kan gene was removed by using the plasmid pCP2025 resulting in strain GSO404.

Strains carrying mutations in single genes encoding RNase activity were constructed by P1 transduction as follows. The rnc::Tn10 allele from HT11526 was transduced into GSO403 to give GSO432 and the rnc::cat allele from NB47827 was transduced into GSO404 to give GSO405. To obtain the rne mutant strain, zce-726::Tn10 linked to the rne3071 allele from AC2228 was transduced into GSO404. A transductant capable of growing at 43.5°C, indicating it retained the wild type rne gene, was named GSO406. A transductant not able to grow at 43.5°C, indicating that it received the linked rne3071 temperature sensitive mutation, was named GSO407. For the rnpA mutant strain, zci-501::Tn10 linked to rnpA79 allele from strain NHY32229 was transduced into GSO404. A transductant capable of growing at 43.5°C, indicating that it retained the wild type rnpA gene, was named GSO408. A transductant not able to grow at 43.5°C, indicating that it received the linked rnpA79 temperature sensitive mutation, was named GSO409. For the rng, rnlA and elaC mutants, the kanamycin resistance gene was amplified from pKD424 (using primers PS1-rng-PS2-rng, PS1-rnlA-PS2-rnlA and mh686-mh687, respectively) and inserted into strain NM400 using the miniλ-red recombination method22. The corresponding Δrng::kan ΔrnlA::kan and ΔelaC::kan alleles were then P1 transduced into GSO404 to generate GSO410, GSO411 and GSO412, respectively.

The strains carrying mutations in multiple RNase genes were constructed as follows. The rnc::cat and Δrng::kan mutations were transduced into GSO406 and GSO407 to generate GSO413 (rnc::cat zce-726::Tn10), GSO414 (rnc::cat zce-726::Tn10 rne-3071), GSO415 (Δrng::kan zce-726::Tn10) and GSO416 (Δrng::kan zce-726::Tn10 rne-3071). The rnc:cat mutation was also transduced into GSO408 and GSO409 to generate GSO417 (rnc:cat zci-501::Tn10) and GSO418 (rnc:cat zci-501::Tn10 rnpA79), respectively. The Δrng::kan and ΔrnlA::kan alleles were transduced into GSO405 to generate GSO419 (rnc:cat Δrng::kan) and GSO420 (rnc:cat ΔrnlA::kan), respectively. The Δrnc::Tn10 allele was transduced into GSO404 to generate GSO421. Subsequently, the rnhA::cat allele from MIC300930 were transduced into GSO421 to give GSO422 (rnc::Tn10 rnhA::cat).

Several triple mutant strains were also constructed for this study. An rnc rne rng null strain was generated by tranducing rng::kan allele into GSO413 and GSO414 to give GSO423 (rnc::cat Δrng::kan zce-726::Tn10) and GSO424 (rnc::cat Δrng::kan zce-726::Tn10 rne-3071). To generate the rne rng elaC strain, we first had to construct a ΔelaC::cat allele by amplifying the chloramphenicol resistance gene from pKD324 (using primers mh686-mh687) and inserting this into NM400 using the miniλ-red recombination method22. The ΔelaC::cat allele was then moved into GSO415 and GSO416 to generate GSO425 (Δrng::kan ΔelaC::cat zce-726::Tn10) and GSO426 (Δrng::kan ΔelaC::cat zce-726::Tn10 rne-3071). A two-step process was needed to generate an rne elaC rnc deletion strain. In the first step, the ΔelaC::kan allele was transduced into GSO406 and GSO407. Following this, the Δrnc::cat allele was moved into the two strains giving rise to GSO427 (zce-726::Tn10 ΔelaC::kan Δrnc::cat) and GSO428 (zce-726::Tn10 rne-3071 ΔelaC::kan Δrnc::cat). Finally, to generate the elaC rnc rng deletion strain, the kanamycin cassette was removed from GSO412 (ΔelaC::kan) by using the plasmid pCP2025. The rnc::cat allele was then transduced into the resulting unmarked strain followed by the introduction of Δrng::kan allele also via P1 transduction giving rise to GSO429 (ΔelaC rnc::cat Δrng::kan).

The strains used for the biochemical studies were constructed as follows: GSO430 was generated by transducing the ΔgadXW::kan allele from strain EK4425 into MG1655, and GSO431 was generated by transducing the rnc::cat allele from NB47827 into GSO430.

RNA isolation

Total RNA was isolated as described31. Briefly, 700 μl of cell culture was added directly into 500 μl of 65°C acid-phenol containing 100 μl of 8% SDS, 320 mM sodium acetate, 16 mM EDTA. The cell lysate was extracted 2× with 65°C acid-phenol (Ambion) and 1× with phenol-chloroform-IAA. RNA was ethanol precipitated and washed with 70% ethanol. RNA pellets were resuspended in DEPC-water.

RNA analysis

Primer extension and Northern blot analyses were performed as described4. 5 μg of total RNA were used in all cases except the mapping of the processed ends of the native gadX-gadW mRNA where 30 μg of total RNA were used. Primers gadX-A1 and GadW-A2 were used to detect the wildtype gadX-gadW mRNA; primers cat-A1 and Gfp-R were used to detect the reporter cat-gfp mRNA. Primers GadW-A2, lacZ-R2 and Gfp-R were used for mapping the 5′ ends in the context of the wild type, lacZ reporter and cat-gfp reporter, respectively. 3′ RACE was carried out as described32 using the following gene specific primers: gadX-T7 to amplify the gadX mRNA, EF584 to amplify the cat mRNA and EF607 to amplify the gfp mRNA. The cDNA products were cloned into pCRII TOPO (Invitrogen) and sequenced.

In vivo assay for GadY-dependent processing

Non-temperature sensitive strains carrying pBAD-RI, pBAD-GadY, pBAD-GadY(90) and pBAD-GadY(59) were grown in LB medium containing 100 μg/ml amp and 1 mM IPTG at 37°C until OD600 = 0.7. GadY sRNA was induced by the addition of arabinose to a final concentration of 0.2%, and cells were incubated for an additional 20 min at 37°C. For the temperature sensitive mutant strains carrying the same plasmids, the cells were grown at 30°C until OD600 = 0.7. The cells were then shifted to 43.5°C for 30 min to inactivate the mutant protein. GadY sRNA was induced by the addition of arabinose to a final concentration of 0.2% and incubated for anadditional 20 min at 43.5°C. In all cases, induction was stopped by addition of 700 μl of cell culture to 500 μl of 65°C acid-phenol (Ambion) and total RNA was purified as described above. The 5′ ends of the processed RNAs were analyzed by primer extension analysis with 32P-labeled primer lacZ-R2.

In vitro processing of gadX

The DNA template for the in vitro synthesis of full-length GadY was generated by PCR with primers GadY-T7 and GadY-T7-R. The DNA templates for the in vitro synthesis of GadY(90) and GadY(59) were generated with primers GadY90-T7 + GadY-T7-R and GadY59-T7 + GadY-T7-R, respectively. The DNA template for the in vitro synthesis of the gadX mRNA fragment was generated with primers GadX-T7 and GadX-T7-R. The three forms of GadY sRNA were synthesized by in vitro transcription using the PCR templates described above together with 100 U (2 μl) T7 RNA polymerase (New England Biolabs) and unlabeled nucleotides (each at a 1.6 mM final concentration) in a 50 μl volume. The 32P-labeled gadX mRNA fragment was similarly synthesized using the above PCR template, T7 RNA polymerase, unlabeled ATP, CTP and GTP and 32P-UTP (Perkin-Elmer).

Protein lysate was generated by growing strain GSO430 or GSO431 to OD600 = 1.0. Cells were harvested by centrifugation and resuspended in 50 mM Tris pH7.5, 5% glycerol, 10 mM MgCl2, 1 mM DTT, 200 mM KCl and lysed by a high pressure using a cell disruptor (Constant Systems LTD) at 27, 000 psi. After lysis, the insoluble material was removed by centrifugation at 20,000× g.

1 μl of 32P-labeled gadX RNA, and 5 μg of soluble protein lysate were incubated with or without 1 μl of GadY sRNA in 50 mM Tris pH 7.5, 5% glycerol, 10 mM MgCl2, 1 mM DTT, 200 mM KCl (10 μl total volume) at 37°C for 60 min. The reactions were stopped by the addition of phenol-chloroform-IAA and sodium acetate to a final concentration of 0.3 M. RNA samples were extracted and ethanol precipitated. RNA pellets were resuspended in formamide gel loading buffer. RNA samples were separated on a 6% acrylamide gel alongside a Perfect Marker RNA ladder (Novagen) and visualized by autoradiography. Processing of unlabeled gadX mRNA was performed identically except the treated RNA was resuspended in DEPC-treated water after ethanol precipitation.

Purification of processing activity (RNase III)

All purification samples were tested for processing activity as described above. Cells from 2 L of GSO430 grown in 2 L of LB broth at 37°C to OD600 =1.0 were harvested by centrifugation and resuspended in 20 ml of 50 mM Tris pH 7.5, 5% glycerol, 10 mM MgCl2, 1 mM DTT, 20 mM KCl and then lysed by high pressure using a cell disruptor (Constant Systems LTD) at 27,000 psi. The insoluble material was removed by centrifugation at 20,000× g for 20 min. The protein lysate was subtractively purified on a HiTrap Q anion exchange column (GE Healthcare) by collecting the flow-through lysate. Ammonium sulfate was added to the protein lysate to a final concentration of 40% and incubated on ice for 60 min. The precipitated protein was removed by centrifugation at 20,000× g and supernatant was saved for further analysis. Ammonium sulfate was added to the remaining supernatant to a final concentration of 60% and incubated on ice for 60 min. The precipitated protein was harvested by centrifugation at 20,000 × g. The protein pellet was resuspended in 4 ml of 50 mM Tris pH 7.5, 5% glycerol, 10 mM MgCl2, 1 mM DTT, 20 mM KCl. The resuspended pellet was further purified on a 26/60 Hi Prep Sephacryl S-300 sizing column (GE Healthcare). Active fractions from the sizing column were pooled and purified on a HiTrap Heparin column (GE Healthcare) using a linear 20 - 300 mM KCl gradient. The active fractions from the Heparin column were pooled and further purified on 26/60 Hi Prep Sephacryl S-300 sizing column (GE Healthcare). The active fractions from the sizing column were pooled and purified on a Mono S cation exchange column (GE Healthcare). At this final step, proteins were eluted with a linear 20 - 300 mM KCl gradient. Proteins from the active fractions were TCA precipitated and separated on a 10-20% Tris Glycine SDS PAGE gel and visualized with GelCode Blue stain reagent (Thermo Scientific). Protein bands were excised from the gel and analyzed by mass spectrometry.

Supplementary Material



We thank A. Zhang for technical assistance, N. Majdalani for strains, P. Backlund and A. Yergey for performing the mass spectrometric analysis, J. Belasco, S. Durand, S. Gottesman and M. Thomason for helpful discussions and comments. This research was supported by the Intramural Research Program of the Eunice Kennedy Shriver National Institute of Child Health and Human Development, a Research Associateship from the National Research Council (E.M.F.) and a postdoctoral fellowship from the Life Sciences Foundation (M.R.H.).


  • GadY small RNA directs cleavage of the gadX-gadW mRNA
  • GadY-directed cleavage is partially dependent on RNase III
  • Additional unidentified RNase is responsible for GadY-directed cleavage

Publisher's Disclaimer: This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final citable form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.


1. Altuvia S. Identification of bacterial small non-coding RNAs: experimental approaches. Curr Opin Microbiol. 2007;10:257–261. [PubMed]
2. Waters LS, Storz G. Regulatory RNAs in bacteria. Cell. 2009;136:615–628. [PMC free article] [PubMed]
3. Thomason MK, Storz G. Bacterial antisense RNAs: How many are there and what are they doing? Annu Rev Genet. 2010;44 [PMC free article] [PubMed]
4. Opdyke JA, Kang JG, Storz G. GadY, a small-RNA regulator of acid response genes in Escherichia coli. J Bacteriol. 2004;186:6698–6705. [PMC free article] [PubMed]
5. Ma Z, Richard H, Tucker DL, Conway T, Foster JW. Collaborative regulation of Escherichia coli glutamate-dependent acid resistance by two AraC-like regulators, GadX and GadW (YhiW) J Bacteriol. 2002;184:7001–7012. [PMC free article] [PubMed]
6. Tramonti A, Visca P, De Canio M, Falconi M, De Biase D. Functional characterization and regulation of gadX, a gene encoding an AraC/XylS-like transcriptional activator of the Escherichia coli glutamic acid decarboxylase system. J Bacteriol. 2002;184:2603–2613. [PMC free article] [PubMed]
7. Tramonti A, De Canio M, De Biase D. GadX/GadW-dependent regulation of the Escherichia coli acid fitness island: transcriptional control at the gadY-gadW divergent promoters and identification of four novel 42 bp GadX/GadW-specific binding sites. Mol Microbiol. 2008;70:965–982. [PubMed]
8. Tramonti A, De Canio M, Delany I, Scarlato V, De Biase D. Mechanisms of transcription activation exerted by GadX and GadW at the gadA and gadBC gene promoters of the glutamate-based acid resistance system in Escherichia coli. J Bacteriol. 2006;188:8118–8127. [PMC free article] [PubMed]
9. Carpousis AJ, Luisi BF, McDowall KJ. Endonucleolytic initiation of mRNA decay in Escherichia coli. Prog Mol Biol Transl Sci. 2009;85:91–135. [PubMed]
10. Brantl S. Regulatory mechanisms employed by cis-encoded antisense RNAs. Curr Opin Microbiol. 2007;10:102–109. [PubMed]
11. Ma C, Simons RW. The IS10 antisense RNA blocks ribosome binding at the transposase translation initiation site. EMBO J. 1990;9:1267–1274. [PMC free article] [PubMed]
12. Case CC, Simons EL, Simons RW. The IS10 transposase mRNA is destabilized during antisense RNA control. EMBO J. 1990;9:1259–1266. [PMC free article] [PubMed]
13. Blomberg P, Wagner EGH, Nordström K. Control of replication of plasmid R1: the duplex between the antisense RNA, CopA, and its target, CopT, is processed specifically in vivo and in vitro by RNase III. EMBO J. 1990;9:2331–2340. [PMC free article] [PubMed]
14. Krinke L, Wulff DL. OOP RNA, produced from multicopy plasmids, inhibits lambda cII gene expression through an RNase III-dependent mechanism. Genes Dev. 1987;1:1005–1013. [PubMed]
15. Waldbeser LS, Chen Q, Crosa JH. Antisense RNA regulation of the fatB iron transport protein gene in Vibrio anguillarum. Mol Microbiol. 1995;17:747–756. [PubMed]
16. Krinke L, Wulff DL. RNase III-dependent hydrolysis of lcII-O gene mRNA mediated by l OOP antisense RNA. Genes Dev. 1990;4:2223–2233. [PubMed]
17. Davies BW, Köhrer C, Jacob AI, Simmons LA, Zhu J, Aleman LM, Rajbhandary UL, Walker GC. Role of Escherichia coli YbeY, a highly conserved protein, in rRNA processing. Mol Microbiol. 2010;78:506–518. [PMC free article] [PubMed]
18. Tsilibaris V, Maenhaut-Michel G, Mine N, Van Melderen L. What is the benefit to Escherichia coli of having multiple toxin-antitoxin systems in its genome? J Bacteriol. 2007;189:6101–6108. [PMC free article] [PubMed]
19. Guzman LM, Belin D, Carson MJ, Beckwith J. Tight regulation, modulation, and high-level expression by vectors containing the arabinose PBAD promoter. J Bacteriol. 1995;177:4121–4130. [PMC free article] [PubMed]
20. Chang ACY, Cohen SN. Construction and characterization of amplifiable multicopy DNA cloning vehicles derived from the P15A cryptic miniplasmid. J Bacteriol. 1978;134:1141–1156. [PMC free article] [PubMed]
21. Ho SN, Hunt HD, Horton RM, Pullen JK, Pease LR. Site-directed mutagenesis by overlap extension using the polymerase chain reaction. Gene. 1989;77:51–59. [PubMed]
22. Court D, Swaminathan S, Yu D, Wilson H, Baker T, Bubunenko M, Sawitzke J, Sharan SK. Mini-lambda: a tractable system for chromosome and BAC engineering. Gene. 2003;315:63–69. [PubMed]
23. Batchelor E, Goulian M. Imaging OmpR localization in Escherichia coli. Mol Microbiol. 2006;59:1767–1778. [PubMed]
24. Datsenko KA, Wanner BL. One-step inactivation of chromosomal genes in Escherichia coli K-12 using PCR products. Proc Natl Acad Sci USA. 2000;97:6640–6645. [PMC free article] [PubMed]
25. Cherepanov PP, Wackernagel W. Gene disruption in Escherichia coli: TcR and KmR cassettes with the option of Flp-catalyzed excision of the antibiotic-resistance determinant. Gene. 1995;158:9–14. [PubMed]
26. Takiff HE, Chen SM, Court DL. Genetic analysis of the rnc operon of Escherichia coli. J Bacteriol. 1989;171:2581–2590. [PMC free article] [PubMed]
27. Yu D, Ellis HM, Lee EC, Jenkins NA, Copeland NG, Court DL. An efficient recombination system for chromosome engineering in Escherichia coli. Proc Natl Acad Sci USA. 2000;97:5978–5983. [PMC free article] [PubMed]
28. Carpousis AJ, Van Houwe G, Ehretsmann C, Krisch HM. Copurification of E. coli RNAase E and PNPase: evidence for a specific association between two enzymes important in RNA processing and degradation. Cell. 1994;76:889–900. [PubMed]
29. Kirsebom LA, Baer MF, Altman S. Differential effects of mutations in the protein and RNA moieties of RNase P on the efficiency of suppression by various tRNA suppressors. J Mol Biol. 1988;204:879–888. [PubMed]
30. Itaya M, Crouch RJ. A combination of RNase H (rnh) and recBCD or sbcB mutations in Escherichia coli K12 adversely affects growth. Mol Gen Genet. 1991;227:424–432. [PubMed]
31. Massé E, Escorcia FE, Gottesman S. Coupled degradation of a small regulatory RNA and its mRNA targets in Escherichia coli. Genes Dev. 2003;17:2374–2383. [PMC free article] [PubMed]
32. Argaman L, Hershberg R, Vogel J, Bejerano G, Wagner EGH, Margalit H, Altuvia S. Novel small RNA-encoding genes in the intergenic regions of Escherichia coli. Curr Biol. 2001;11:941–950. [PubMed]
PubReader format: click here to try


Related citations in PubMed

See reviews...See all...

Cited by other articles in PMC

See all...


Recent Activity

Your browsing activity is empty.

Activity recording is turned off.

Turn recording back on

See more...