OUP user menu

The role of the small regulatory RNA GcvB in GcvB/mRNA posttranscriptional regulation of oppA and dppA in Escherichia coli

Sarah C. Pulvermacher , Lorraine T. Stauffer , George V. Stauffer
DOI: http://dx.doi.org/10.1111/j.1574-6968.2008.01068.x 42-50 First published online: 1 April 2008

Abstract

The gcvB gene encodes two small, nontranslated RNAs that regulate OppA and DppA, periplasmic binding proteins for the oligopeptide and dipeptide transport systems. Analysis of the gcvB sequence identified a region of complementarity near the ribosome-binding sites of dppA and oppA mRNAs. Several changes in gcvB predicted to reduce complementarity of GcvB with dppA-lacZ and oppA-phoA reduced the ability of GcvB to repress the target RNAs while other changes had no effect or resulted in stronger repression of the target mRNAs. Mutations in dppA-lacZ and oppA-phoA that restored complementarity to GcvB restored the ability of GcvB to repress dppA-lacZ but not oppA-phoA. Additionally, a change that reduced complementarity of GcvB to dppA-lacZ reduced GcvB repression of dppA-lacZ with no effect on oppA-phoA. The results suggest that different regions of GcvB have different roles in regulating dppA and oppA mRNA, and although pairing between GcvB and dppA mRNA is likely part of the regulatory mechanism, the results do not support a simple base pairing interaction between GcvB and its target mRNAs as the complete mechanism of repression.

Keywords
  • GcvB
  • sRNA
  • dppA-lacZ
  • oppA-phoA

Introduction

The Escherichia coli gcvTHP (gcv) operon encodes the glycine cleavage enzyme complex that provides one-carbon units for cellular methylation reactions (Kikuchi, 1973). The gcvB gene encodes two small, nontranslated RNAs (sRNAs) of 130 and 206 nucleotides (nts) (referred to as GcvB) that negatively regulate OppA and DppA, the oligopeptide and dipeptide periplasmic binding proteins (Urbanowski et al., 2000). Both gcv and gcvB are activated by GcvA in response to glycine and repressed by GcvA+GcvR in the absence of glycine (Wilson et al., 1993; Stauffer & Stauffer, 1994; Urbanowski et al., 2000). Because the expression pattern of gcvB is qualitatively similar to the expression pattern of the gcv operon, it is hypothesized that GcvB plays a role in an integrated cellular response to regulate other genes in conjunction with the glycine cleavage enzyme system.

The E. coli chromosome encodes 50–100 sRNAs that function in different cellular processes (Wassarman et al., 1999, 2001; Argaman et al., 2001; Wagner & Flardh, 2002; Wassarman, 2002; Gottesman, 2004). Many sRNAs posttranscriptionally regulate expression of target genes positively or negatively by forming base pairs (bps) with target mRNAs (Storz et al., 2004). Often sRNAs, including MicF, DicF, OxyS, and Spot42, negatively regulate target mRNAs by base pairing and preventing the ribosome accessibility with the target mRNA (Chen et al., 2004). A region in GcvB was identified complementary to both dppA and oppA mRNAs near their ribosome-binding sites. The nucleotides in GcvB predicted to base pair with dppA and oppA mRNAs were changed and their effects on dppA-lacZ and oppA-phoA regulation were determined. Compensatory mutations in either dppA-lacZ or oppA-phoA were then created to restore base pairing to determine whether regulation was restored. These results suggest that although complementary pairing of GcvB and its target sequences is important, GcvB-mediated regulation of the DppA and OppA proteins is more complex than simply pairing of GcvB to the mRNAs.

Materials and methods

Bacterial strains, plasmids, and phage

The E. coli strains and plasmids used are listed in Table 1. The λdppA-lacZ translational fusion includes 300 bp upstream of the translation initiation site and the first 14 codons of dppA fused in frame with the eighth codon of the lacZYA genes in phage λgt2 (Urbanowski et al., 2000). The λoppA-phoA translational fusion includes 210 bp upstream of the translation initiation site and the first 37 codons of oppA, including the leader peptide, fused in frame to the leaderless phoA gene in phage λgt2 (Brickman & Beckwith, 1975). Plasmid pGS571 carries the gcvB gene on a 406-bp EcoRI–HindIII fragment in plasmid pGS272 (Jourdan & Stauffer, 1999). Plasmids pGS571 (gcvB), pA55 (dppA-lacZ translational fusion) (Urbanowski et al., 2000), and pA32 (oppA-phoA translational fusion) (Urbanowski et al., 2000) were used as templates to create nucleotide changes in gcvB (Fig. 1a and b), dppA-lacZ (Fig. 1a), and oppA-phoA (Fig. 1b) by PCR ‘megaprimer’ mutagenesis (Sarkar & Sommer, 1990) or by QuikChange® II site-directed mutagenesis (Stratagene, Cedar Creek, TX). Mutations were verified by DNA sequence analysis at the DNA Core Facility of the University of Iowa. The gcvB mutations were subcloned into the single-copy plasmid pGS341 (Jourdan & Stauffer, 1998), replacing the wild-type (WT) gcvA gene. The 5947-bp EcoRI–MfeI fragments carrying the dppA-lacZ mutations and the 3768-bp EcoRI–MfeI fragments carrying the oppA-phoA mutations were ligated into the EcoRI site of phage λgt2 (Panasenko et al., 1977). These phages were used to lysogenize various E. coli strains in single copy as described (Shimada et al., 1972; Urbanowski & Stauffer, 1986). All lysogens were grown at 30 °C because all fusion phages carry the λcI857 mutation, resulting in a temperature-sensitive λcI repressor (Panasenko et al., 1977).

View this table:
1

Strains and plasmids

Strain or plasmidRelevant genotypeSources or references
Strain
 GS162WTThis lab
 GS1144GS162ΔgcvB∷ΩCMrThis study
Plasmid
 pGS311Single-copy vectorThis lab
 pGS341Single-copy vector+WT gcvAJourdan & Stauffer (1998)
 pGS571Multicopy vector+WT gcvBThis study
 pGS594Single-copy vector+WT gcvBThis lab
 pGS595pGS594 with a 7 bp change in gcvB Con (gcvBCon–7)This study
 pGS596pGS594 with a –TGT– to –CCC– change of +71 to +73 bp in gcvB (gcvB+71CCC)This study
 pGS597pGS594 with a deletion of –G– at +75 bp in gcvB (gcvB+75ΔG)This study
 pGS602pGS594 with a –TGT– to –AAA– change of +76 to +78 bp in gcvB (gcvB+76AAA)This study
 pGS603pGS594 with a deletion of –GTGTTTGCAAT– from +83 to +93 bp in gcvB (gcvB+83Δ11bp)This study
 pGS605pGS594 with a –TT– to –AA– change of +65 and +66 bp in gcvB (gcvB+65AA)This study
 pGS618pGS594 with a –T– to –C– change at +73 bp in gcvB (gcvB+73C)This study
 pGS619pGS594 with a –TG– to –CC– change of +71 and +72 bp in gcvB (gcvB+71CC)This study
  • * All strains also carry the pheA905 thi araD129 rpsL150 relA1 deoC1 flbB5301 ptsF25 rbsR mutations.

1

Comparison of GcvB between +60 and +100 nt with the dppA (a) and oppA (b) mRNAs. The AUG translation initiation sites for dppA and oppA mRNAs are underlined and potential Shine–Dalgarno sequences are overlined and labeled SD (Igarashi et al., 1997; Sharma et al., 2007). A 13-base region of GcvB from +79 to +91 nt (Con) conserved in GcvB homologs from the genera Escherichia, Salmonella, Yersinia, Haemophilus, Vibrio, Pasteurella, Shigella, Klebsiella, and Photorhabdus is underlined (see Fig. 3). Regions of complementarity are indicated with dots between the sequences. GU base pairs are indicated by a line. Changes made to GcvB are indicated below the sequence. A 1 and an 11 bp deletion are indicated with Δs. A 7 bp substitution in the Con region is shown in brown. The –AAA– mutation made to GcvB, GcvB+76AAA, is shown in pink, the –CCC– mutation, GcvB+71CCC, is shown in red, the –CC– mutation, GcvB+71CC, is shown in blue, the –C– mutation, GcvB+73C, is shown in purple, and the –AA– mutation, GcvB+65AA, is shown in green. Mutations made in dppA-lacZ and oppA-phoA that restore complementarity to the GcvB mutations are shown above the sequences and are color coded to correspond with the mutations made in GcvB. The –A– nucleotide of the translation initiation codon in dppA and oppA mRNAs is designated +1. Mutations upstream of the translation initiation sites are defined using negative numbers and mutations downstream are defined with positive numbers. (c) Proposed secondary structure of Escherichia coli GcvB using m-fold (Zuker, 2003). The nucleotides changed in GcvB are shown to the left of the WT GcvB secondary structure in the respective colors designated in (a) and (b). The single nucleotide deletion is designated by Δ. The bases between the arrows designate the nucleotides deleted by the 11 bp deletion. Conserved predicted stem-loop structures among E. coli, Shigella dysenteriae, Salmonella typhimurium, Yersinia pestis, Haemophilus influenzae, and Vibrio cholerae are designated by letters A–F.

Media

The complex medium used was Luria–Bertani broth (LB) (Miller, 1972). Agar was added at 1.5% (w/v) to make solid media. Ampicillin was added at 50 μg mL−1.

Enzyme assays

Cells were grown in LB or LB+ampicillin to mid-log phase (OD600 nmc. 0.5) and assayed for either β-galactosidase activity using the chloroform/sodium dodecylsulfate (SDS) lysis procedure (Miller, 1972) or alkaline phosphatase activity (Brickman & Beckwith, 1975). Results are the averages of two or more assays with each sample carried out in triplicate.

RNA extraction and Northern blot analysis

Escherichia coli strains were grown in 10 mL LB or LB+ampicillin to an OD600 nmc. 0.5. Rifampicin (250 μg mL−1) was added to each sample, and cells were immediately added to 5% (v/v) acidic phenol/95% (v/v) ethanol, centrifuged for 5 min at 2000 g, and pellets were frozen at −80 °C. RNA was isolated by phenol extraction (Ledeboer et al., 2006). An 8 M urea 8% polyacrylamide gel was run with 10 μg of each RNA sample and was electroblotted to a positively charged nylon membrane (Roche, Mannheim, Germany). The blot was hybridized with a DNA probe specific to +1 to +134 of the E. coli gcvB gene labeled using the PCR-digoxigenin Probe synthesis kit (Roche, Mannheim, Germany). Hybridization was performed at 42 °C as described (Engler-Blum et al., 1993), and the membrane was exposed to a film and imaged using the FUJIFILM LAS-1000 camera and Intelligent Dark Box. Quantification of RNA was performed using image gauge 3.12 software. The membrane was subsequently stripped using 0.1% (w/v) SDS/2x SSC heated to c. 95 °C and rehybridized using a digoxigenin-labeled DNA probe specific for 5S RNA gene from +2 to +112. The relative amount of GcvB expressed from each mutant gcvB allele was determined by taking the ratio of GcvB to 5S RNA gene detected for each sample.

DNA manipulation

The procedures for plasmid DNA purification, restriction enzyme digestion, etc. were as described (Sambrook et al., 1989). PCR amplifications were carried out under standard reaction conditions using Vent DNA polymerase (New England BioLabs Inc., Beverly, MA). Restriction enzymes and other DNA-modifying enzymes were from New England Biolabs Inc.

Results and discussion

Effects of mutations in gcvB on dppA-lacZ and oppA-phoA expression

A G–U rich region was identified in GcvB complementary to dppA and oppA mRNAs near the translation start sites (Fig. 1a and b). Sharma (2007). report that this same region of Salmonella's GcvB base pairs with dppA and oppA mRNA in Salmonella. In E. coli changes were made in gcvB in the region of complementarity with dppA and oppA mRNAs and their effects on dppA-lacZ and oppA-phoA expression were tested. All changes are predicted by the m-fold program to leave the GcvB secondary structure intact (Mathews et al., 1999; Zuker, 2003). Mutations made include a –CCC– change of +71 to +73 bp and an –AAA– change of +76 to +78 bp that disrupt complementarity between GcvB and the dppA and oppA mRNAs, an –AA– change of +65 and +66 bp that primarily decreases complementarity to the dppA mRNA, and a deletion of +75 bp that increases complementarity with the dppA mRNA but decreases complementarity with oppA mRNA (Fig. 1a and b). Single-copy plasmids carrying either WT gcvB or a gcvB mutation were transformed into the ΔgcvB strain, GS1144, lysogenized with either λdppA-lacZ or λoppA-phoA, and assayed for either β-galactosidase or alkaline phosphatase activity. As expected, expression of dppA-lacZ and oppA-phoA in the WT lysogens were nine- and fourfold lower than in the ΔgcvB lysogens (Fig. 2a and b). Transformation of either ΔgcvB lysogen with WT gcvB restored repression to WT levels, and transformation with the empty vector alone (pGS311) showed expression levels similar to the untransformed ΔgcvB lysogens (Fig. 2a and b). To determine the effects of mutations in gcvB, activity from ΔgcvB lysogens transformed with gcvB mutant alleles was always compared with WT and ΔgcvB lysogens transformed with WT gcvB. When the ΔgcvBλdppA-lacZ or ΔgcvBλoppA-phoA strains carried the gcvB+71CCC allele, β-galactosidase activity was 4.5-fold higher (Fig. 2a) and alkaline phosphatase activity was twofold higher (Fig. 2b) compared with the controls. The gcvB+76AAA allele did not significantly affect repression of dppA-lacZ or oppA-phoA in the ΔgcvB strain (Fig. 2a and b), suggesting that these nucleotides are not necessary for GcvB regulation of either target. The gcvB+65AA allele resulted in a reduced ability of GcvB to repress dppA-lacZ, as the β-galactosidase level in ΔgcvBλdppA-lacZ carrying gcvB+65AA was more than threefold higher (Fig. 2a). In contrast, the gcvB+65AA allele resulted in an increased ability of GcvB to repress oppA-phoA, as the alkaline phosphatase level in ΔgcvBλoppA-phoA carrying gcvB+65AA was twofold lower (Fig. 2b). The –AA– change disrupts complementarity between GcvB and dppA mRNA, and the reduced repression of dppA-lacZ is possibly due to the lack of complementary base pairing between GcvB and dppA-lacZ mRNA. However, the changes do not significantly disrupt complementarity between GcvB and oppA mRNA (Fig. 1a and 1b), it is unclear why these changes result in increased repression of oppA-phoA. The gcvB+75ΔG allele, which results in a better match of GcvB to dppA mRNA but reduces complementarity to oppA mRNA (Fig. 1a and 1b), did not affect expression of either fusion (Fig. 2a and b). These results show that the region in GcvB between base +65 and +78 is important for GcvB regulation of both dppA-lacZ and oppA-phoA. However, certain changes affect regulation of the two targets differently. In addition, changing bases that decrease or increase complementarity with the target mRNAs does not necessarily correlate with decreased or increased repression.

2

β-Galactosidase (a) or alkaline phosphatase (b and c) levels of λdppA-lacZ or λoppA-phoA from WT, ΔgcvB, or the ΔgcvB strain transformed with the single copy empty vector (e.v.) or a single copy plasmid carrying either the WT gcvB or a mutated gcvB allele. The respective gcvB allele in the ΔgcvB background strain is designated below the bar graph after the slash mark. The lysogens and transformants were grown in LB or LB+ampicillin to an OD600 nmc. 0.5 and assayed for either β-galactosidase activity (Miller, 1972) (a) or alkaline phosphatase activity (Brickman & Beckwith, 1975) (b and c). Results are the averages of two or more assays, with each assay performed in triplicate.

A comparison of the gcvB gene from E. coli with presumptive gcvB genes in the genera Shigella, Salmonella, Klebsiella, Photorhabdus, Yersinia, Haemophilus, Pasterurella, and Vibrio shows extensive homology (Fig. 3). One 13 bp region of gcvB, designated Con, is highly conserved (Fig. 3) and forms part of the predicted stem loop B in GcvB (Fig. 1c). It is not entirely clear whether the predicted stem loop B forms as part of the secondary structure because the E. coli and S. typhimurium GcvB sRNAs are c. 95% identical and stem loop B was not observed in Salmonella by in vitro structure mapping (Sharma et al., 2007). It was tested whether the Con sequence or the stem loop B is required for GcvB regulation of either dppA-lacZ or oppA-phoA expression. Seven base pairs in the gcvB Con sequence were changed and an 11 bp deletion from +83 to +93 was also made that removes the last 9 bp of the Con sequence and the predicted stem loop B, designated pGS595 and pGS603 (Fig. 1a and b). The deletion of predicted stem loop B disrupts significant complementarity between GcvB and oppA mRNA, with six predicted base pairing interactions destroyed compared with GcvB and dppA mRNA, with only one predicted base pairing interaction destroyed (Fig. 1). The gcvBCon–7 allele resulted in an inability of GcvB to fully repress dppA-lacZ or oppA-phoA, as the β-galactosidase level was fourfold higher (Fig. 2a) and the alkaline phosphatase level was twofold higher (Fig. 2b) compared with the controls. Thus, the conserved region is important for GcvB regulation of both dppA and oppA mRNA. Sharma (2007) also found that a deletion of +66 to +89 nt in gcvB in Salmonella results in an inability to repress either dppA-gfp or oppA-gfp. Because of the extensive number of changes in the gcvBCon–7 allele, it is unknown if the effect of the mutation is due to a decrease in complementarity, or whether it plays some additional role in GcvB regulation. In contrast, the gcvB+83Δ11 bp allele showed increased repression of both dppA-lacZ and oppA-phoA (Fig. 2a and b) despite the loss of a region of GcvB predicted to be involved in base pairing. Thus, part of the Con sequence or stem-loop B normally functions to prevent repression of dppA and oppA mRNA by GcvB.

3

Comparison of the Escherichia coli gcvB sequence with homologs from other organisms. Regions of high consensus are boxed in gray. The consensus is shown below the sequence with high consensus bases in capital letters. One 13 bp region conserved in these homologs is overlined and designated Con. Numbering is based on the E. coli sequence. The region of complementarity between GcvB and the target mRNAs dppA and oppA occurs between +60 and +100 nt in GcvB. Ec, E. coli; Sd, Shigella dysenteriae; St, Salmonella typhimurium; Kp, Klebsiella pneumoniae; Pl, Photorhabdus luminescens; Yp, Yersinia pestis; Hi, Haemophilus influenzae; Pm, Pasteurella multocida; Vc, Vibrio cholerae; Con, conserved sequence.

Effects of mutations made to gcvB on GcvB RNA production

It is possible the increased repression and derepression of dppA-lacZ and oppA-phoA observed for the different gcvB alleles is a result of over- or underexpression of the mutant RNAs. To determine whether each gcvB allele produces comparable amounts of GcvB, Northern blots were performed. Although there are small changes in the levels of GcvB observed in the Northern blot (less than twofold) (Fig. 4), these differences do not appear to be responsible for the observed effects on regulation. The gcvB+76AAA and gcvB+83Δ11 bp alleles, which show the lowest levels of GcvB (Fig. 4), result in essentially normal or increased repression of dppA-lacZ and oppA-phoA rather than a loss of repression (Fig. 2a and b). The gcvBCon–7 allele, which has the highest level of RNA, shows a loss of repression of both dppA-lacZ and oppA-phoA (Fig. 2a and b). These results suggest that the loss of or increased repression observed for dppA-lacZ and oppA-phoA is likely not due to changes in the synthesis or stability of GcvB, but are direct effects on GcvB's role in regulation of dppA and oppA mRNA.

4

Northern blot analysis of GcvB expressed from plasmids carrying either a WT or a mutant gcvB allele. Total cell RNA was isolated from each strain, run on an 8 M urea–8% polyacrylamide gel, and probed with either a digoxigenin-labeled GcvB or a 5S-specific DNA probe. Total RNA extracts are as follows: lane 1, WT (GS162); lane 2, gcvB+ (GS1144 pGS594); lane 3, gcvB+71CCC (GS1144 pGS596); lane 4, gcvB+76AAA (GS1144 pGS602); lane 5, gcvB+65AA (GS1144 pGS605); lane 6, gcvB+75ΔG (GS1144 pGS597); lane 7, gcvBCon–7 (GS1144 pGS595); lane 8, gcvB+83Δ11 bp (GS1144 pGS603); lane 9, gcvB+73C (GS1144 pGS618); lane 10, gcvB+71CC (GS1144 pGS619). RNA was quantified using image gauge 3.12 software, and the counts determined for each sample are shown below each lane. The bottom row shows the ratio of GcvB RNA compared with the 5S RNA gene for each sample. The ratio observed for lane 1 (WT GcvB expressed in WT cells) was considered to be the WT amount of GcvB expressed in LB.

Effects of mutations in dppA-lacZ or oppA-phoA that restore base pairing to GcvB on GcvB repression

The above results identified a region of GcvB complementary to dppA and oppA mRNAs required for normal regulation. To determine whether GcvB regulates dppA and oppA mRNA by base pairing with the mRNAs, three nucleotides were changed in either dppA-lacZ or oppA-phoA to restore base pairing with the gcvB+71CCC mutation. However, these mutations in dppA-lacZ and oppA-phoA had dramatic effects on β-galactosidase and alkaline phosphatase activity (data not shown). Instead, a –GG– change at −20 and −21 nts and a –G– change at −22 nts in dppA-lacZ and a –GG– change at +9 and +10 nts and a –G– change at +8 nts in oppA-phoA were made (Fig. 1a and b). The activity levels of the ΔgcvB strain lysogenized with these four fusions were high enough compared with the WT fusions to test whether complementary pairing between GcvB and the mRNAs is part of the regulatory mechanism (Fig. 5a and c). Thus, we constructed a –C– change at +73 nts and a –CC– change at +71 and +72 nts in gcvB, designated pGS618 and pGS619 (Fig. 1), transformed the ΔgcvB strain lysogenized with λdppA-lacZ, λdppA−22G-lacZ, and λdppA−20GG-lacZ, and assayed for β-galactosidase activity. The β-galactosidase levels in ΔgcvBλdppA-lacZ carrying the gcvB+73C or gcvB+71CC allele were 1.3- and 4.5-fold higher than the levels in the WT λdppA-lacZ lysogen, demonstrating that the gcvB+73C and gcvB+71CC alleles fail to repress fully when complementarity is disrupted (Fig. 2a). The gcvB+73C and gcvB+71CC alleles produce essentially equivalent amounts of RNA compared with WT gcvB (Fig. 4), suggesting that their failure to repress dppA-lacZ is not due to low levels of GcvB. The β-galactosidase levels in ΔgcvBλdppA−22G-lacZ carrying WT gcvB or the gcvB+73C allele were reduced 4- and 5.2-fold compared with ΔgcvBλdppA−22G-lacZ (Fig. 5a, compare bar 3 with bars 4 and 5). The β-galactosidase levels in ΔgcvBλdppA−20GG-lacZ carrying WT gcvB or the gcvB+71CC allele were reduced 2.9- and 9-fold (Fig. 5a, compare bar 6 with bars 7 and 8). Thus, compensatory mutations in gcvB increased the ability of GcvB to repress the dppA−22G-lacZ and dppA−20GG-lacZ lysogens compared with WT GcvB, consistent with base pairing between GcvB and dppA-lacZ mRNA as part of the regulatory mechanism.

5

β-Galactosidase levels of WT λdppA-lacZ, λdppA−22G-lacZ, or λdppA−20GG-lacZ (a) or alkaline phosphatase activity of WT λoppA-phoA, λoppA+8G-phoA, or λoppA+9GG-phoA (c) from either WT, ΔgcvB, or the ΔgcvB strain transformed with WT gcvB or the mutant gcvB+73C or gcvB+71CC alleles. β-Galactosidase levels of either λdppA-lacZ or λdppA−14UU-lacZ (b) from WT, ΔgcvB, or the ΔgcvB strain transformed either WT gcvB or the mutant gcvB+65AA allele. The fusion assayed in each strain is designated below the bar graph. The lysogens and transformants were grown in LB or LB+ampicillin to an OD600 nmc. 0.5 and assayed for either β-galactosidase activity (Miller, 1972) (a and b) or alkaline phosphatase activity (Brickman & Beckwith, 1975) (c). Results are the average of two or more assays with each assay performed in triplicate.

To test the effect of gcvB+73C and gcvB+71CC alleles on oppA-phoA expression, plasmids pGS618 and pGS619 were transformed into ΔgcvB lysogenized with λoppA-phoA, λoppA+8G-phoA, and λoppA+9GG-phoA, and assayed for alkaline phosphatase activity. These strains were assayed with a different batch of reagents compared with the other alkaline phosphatase results, and probably explain the slightly lower expression levels for the WT and ΔgcvBλoppA-phoA control strains in Fig. 2c. The alkaline phosphatase level in ΔgcvBλoppA-phoA carrying the gcvB+73C allele was essentially the same as the WT λoppA-phoA lysogen, demonstrating that the gcvB+73C allele is able to repress when complementarity is disrupted at this position (Fig. 2c). The alkaline phosphatase level in ΔgcvBλoppA-phoA carrying the gcvB+71CC allele was twofold higher than WT λoppA-phoA, demonstrating that the gcvB+71CC allele fails to fully repress when complementarity is disrupted (Fig. 2c). The alkaline phosphatase level in ΔgcvBλoppA+8G-phoA carrying WT gcvB or the gcvB+73 allele was reduced 1.5- and 2.4-fold compared with ΔgcvBλoppA+8G-phoA (Fig. 5c, compare bar 3 with bars 4 and 5). The alkaline phosphatase level in ΔgcvBλoppA+9GG-phoA carrying WT gcvB or the gcvB+71CC allele was reduced 1.7- and 1.1-fold compared with ΔgcvBλoppA+9GG-phoA (Fig. 5c, compare bar 6 with bars 7 and 8). Surprisingly, there is better repression of oppA+9GG-phoA with the WT gcvB allele than with the complementary gcvB+71CC allele. It is unclear why restoring complementarity between the gcvB+73C allele and oppA+8G-phoA represses, but restoring complementarity between the gcvB+71CC allele and oppA+9GG-phoA does not. It is possible that regulation of oppA-phoA by GcvB occurs by a different mechanism than for dppA-lacZ. Previous results suggested that GcvB-mediated repression of dppA-lacZ and oppA-phoA is different (Urbanowski et al., 2000). The targetrna program (Tjaden et al., 2006) predicts other small regions of possible base pairing between GcvB and oppA mRNA; therefore, it is possible that other nucleotides in oppA mRNA are more essential for regulation by GcvB and base pairing is part of the regulatory mechanism.

Additionally, it was tested whether loss of gcvB+65AA repression of dppA-lacZ is due to loss of base pairing. Nucleotides were changed in dppA-lacZ to –UU– at −14 and −15 bps to restore pairing with the gcvB+65AA allele. Expression of β-galactosidase activity from ΔgcvBλdppA−14UU-lacZ was 5.3-fold higher than WT λdppA-lacZ (Fig. 5b, compare bar 3 with bar 1). The β-galactosidase level in ΔgcvBλdppA−14UU-lacZ carrying WT gcvB or the gcvB+65AA allele was 2.8-or 3.4-fold lower compared with the control ΔgcvBλdppA−14UU-lacZ (Fig. 5b, compare bar 3 with bars 4 and 5). Thus, the compensatory mutation in gcvB increased the ability of GcvB+65AA to repress dppA−14UU-lacZ better than WT GcvB. Because the gcvB+65AA mutation showed a threefold loss of repression compared with WT λdppA-lacZ (Fig. 2a), these results are consistent with RNA/RNA pairing as part of the mechanism for GcvB regulation of dppA-lacZ. Additionally, the gcvB+65AA allele shows that certain sequences of GcvB play different roles in regulating dppA-lacZ and oppA-phoA (Fig. 2a and b).

The largest family of sRNAs acts by base pairing with target mRNAs to alter mRNA translation or stability (Majdalani et al., 2005). Although these results suggest that complementary RNA/RNA pairing between GcvB and dppA-lacZ mRNA is likely part of the regulatory mechanism, the regulation for both dppA and oppA mRNA is more complex than simply base pairing between GcvB and the target mRNAs. It was predicted that the binding of GcvB with dppA or oppA mRNA blocks ribosomal binding, preventing translation of both dppA and oppA mRNAs. Sharma et al. reported that WT GcvB in Salmonella blocks translation of both dppA and oppA mRNAs, and when +66 to +89 nts were deleted, the nucleotides that are predicted to be involved in base pairing in E. coli, translation was no longer blocked (Sharma et al., 2007). Additionally, all E. coli sRNAs that regulate by RNA/RNA base pairing with target mRNAs bind the RNA chaperone protein Hfq (Chen et al., 2004), and Hfq has been shown to interact directly with some mRNAs (Geissmann & Touati, 2004; Vytvytska et al., 1998). Recently, it was shown GcvB is a target for Hfq binding (Zhang et al., 2003), and it has been found that Hfq is required for the ability of GcvB to regulate both dppA-lacZ and oppA-phoA negatively (unpublished results). It is possible one or more of the gcvB mutant alleles that fail to fully repress dppA-lacZ or oppA-phoA or that result in increased repression fails to interact with Hfq, or binds Hfq more tightly, possibly explaining the altered regulation. We are in the process of determining how Hfq is involved in GcvB regulation of dppA and oppA mRNA as well as where Hfq binds with GcvB. These studies should clarify whether RNA/RNA pairing is necessary for GcvB regulation of dppA and oppA mRNA.

Acknowledgements

This work was supported by Public Health Service Grant GM069506 from the National Institute of General Medical Sciences.

Footnotes

  • Editor: Roger Buxton

References

View Abstract