Figure 1.
Effect of hfq and gcvB deletions on the expression of a yifK-lacZ fusion.
Deleting gcvB causes yifK expression to increase approximately five-fold in exponential cultures (OD600≈0.4) and less than three-fold in stationary overnight cultures (OD600≈2.0). A greater increase is observed in the hfq deletion mutant, suggesting the involvement of a separate Hfq-dependent step in yifK regulation. Strains used were MA8020 (wt), MA8021 (Δhfq), MA10377 (ΔgcvB) and MA10403 (Δhfq ΔgcvB). All strains carry the yifK::MudK lac fusion. Their full genotypes are listed in Table S1.
Figure 2.
Characterization and mutational analysis of the leader region of the yifK gene.
The promoter-proximal region of the yifK gene was randomly mutagenized by error-prone PCR using primers ppF45 and ppF47 and chromosomal DNA from a strain carrying a cat cassette 78 bp upstream from the yifK promoter (in opposite orientation; strain MA11780) as template. The mutagenized fragment was introduced into strain MA10280 (yifK-lacZY/pKD46) and recombinants were selected as described in the text. Three mutants expressing higher ß-galactosidase activity were identified: one carrying a yifK promoter change that causes the -35 box of yifK to match the consensus sequence (TTGACA) (A, top); the other two isolates carrying mutations in the 5′UTR (U to C at +21 and G to A at +27) (A, bottom). Grey and green underlining denote a putative Hfq binding site and the initiating AUG, respectively. Brown overlining denotes the Shine-Dalgarno sequence. A sequence stretch complementary to GcvB is boxed in light green. B. Primer extension of yifK mRNA. Total RNA extracted from wild-type and mutant strains was used to map the 5′ end of the yifK mRNA by reverse transcription (primer ppF49)(B). This analysis identifies the 5′ end of yifK mRNA and shows that all three mutations lead to higher mRNA levels. Measurements of ß-galactosidase activity (C) show that U21C, but not G27A, does not cause any further increase in yifK-lacZ expression in the Δhfq background, suggesting that U21C affects Hfq binding to yifK mRNA. ß-galactosidase activity was measured in exponentially growing cultures (OD600≈0.4).
Figure 3.
Northern blot analysis of yifK transcription.
All strains used as source of RNA carried -33 the promoter “up” mutation. The RNAse E mutant carried the temperature sensitive (ts) rne-3071 allele [28]. In experiments involving this strain, bacteria were grown at 30°C and shifted to 43°C 15 min prior to RNA extraction. RNA was separated on a 1% agarose-formaldehyde gel and probed with 32P-labeled DNA oligonucleotides complementary to a sequence near the 5′ end of yifK mRNA (ppF16; probe 1 above) or to a sequence in the argX-hisR intercistronic region (ppH27, probe 2). The blot in A and B was initially probed with ppF16 (A), then stripped and re-probed with ppH27 (B). Probing for the SsrA RNA (pp813) served as loading control. The blot in C was probed simultaneously with ppF16 and pp813. “SD−” denotes a G to C change in the Shine-Dalgarno sequence (+59), which causes an about 10-fold reduction in yifK expression (construct n. 2 in Figure 9).
Figure 4.
Lrp control of yifK-lacZY expression.
The yifK gene is expressed at very low levels in minimal medium. As a result, strains carrying the yifK-lacZY fusion are phenotypically Lac− in this medium, regardless of the gcvB allele (A, left panel). Inactivation of Lrp confers a Lac+ phenotype (A, left panel). This suggests that low yifK expression results from Lrp repression. The Lac+ phenotype is also restored upon addition of leucine (0.3 mM) (A, right panel), indicating that leucine relieves Lrp repression. ß-galactosidase measurements (B) confirm that GcvB plays no significant role in yifK regulation in minimal medium and provide a quantitative estimate of the Lrp effects. NCE medium [43] supplemented with 0.2% lactose (A) or 0.2% glycerol (B) was used as minimal medium.
Figure 5.
Differential effects of target sequence mutations on yifK-lacZ expression and on the response to compensatory changes in GcvB.
Adjacent portions of the sequence presumed to base-pair with GcvB in yifK mRNA were randomly mutagenized by a “scarless” λ red recombineering procedure (see Materials and Methods). Briefly, DNA fragments amplified by “reciprocal priming” with oligonucleotide pairs ppG48/ppG49 (mutagenesis of +46 to +48) and ppH12/ppH13 (mutagenesis of +49, +50) were introduced into strain MA11526 and tetracycline-sensitive recombinants selected one plates supplemented with fusaric acid (12 µg/ml). Two of the mutants obtained were chosen for further study. Compensatory changes in GcvB were obtained by standard recombineering using fragments amplified from the chromosome of strain MA11779 with primer pairs ppG63/ppF18 (mutagenesis of the +86 to +88) and ppH61/ppF18 (mutagenesis of +84, +85). Results above show that changing yifK mRNA sequence from positions +46 to +48, or making the opposite change in GcvB, both relieve repression (A). Repression is restored in a strain carrying the two sets of changes combined, showing that base-pairing is required for repression (A). In contrast, replacing the CA doublet at +49, +50 by UC causes a reduction, rather than an increase, of yifK-lacZ expression; introduction of the compensatory changes in GcvB does not accentuate this trend (B). Thus, the CA to UC conversion impairs yifK expression and renders it insensitive to GcvB repression.
Figure 6.
Toeprinting analysis of yifK mRNA.
30S ribosomal toeprinting was carried out in the absence or in the presence of GcvB RNA or of RyhB RNA as described in the Materials and Methods. “+” and “−” signs denote the presence of absence of indicated components. The decline and disappearance of the toeprint (shown by arrows) at increasing GcvB concentrations (50 nM, 100 nM and 500 nM), is indicative of interference with the 30S subunit binding to yifK mRNA. Failure of RyhB to do so at the concentration of 6.0 µM (lane “R”) confirms the specificity of the effect.
Figure 7.
Effect of yifK 5′ UTR's changes on mRNA translation.
The CA49,50 to UC49,50 change in yifK mRNA affects expression levels in vivo (A) and mRNA translation in vitro (B). In vitro translation was performed using the coupled transcription/translation PURExpress kit (see Materials and Methods). DNA templates were from plasmids carrying the entire yifK 5′ UTR from wild-type, or from the UC49,50 mutant, fused to the coding sequence of a 3×FLAG epitope-tagged version of the cat gene. Fusions were initially obtained as chromosomal constructs using DNA fragments amplified from strain MA7224 with primer pairs ppL50/ppL52 (wt) and ppL51/ppL52 (UC49,50). Subsequently, the fusions were cloned into plasmid DHFR following amplification (ppM29/ppM30) and double Xba I/Pst I digestion. Transcription/translation reactions were carried out at a template DNA concentration of 0.5 pM for 90 min or 5 pM for 30 min and products analyzed by Western blotting using anti-FLAG monoclonal antibodies [53]. Under both conditions, higher amounts of cat-3×FLAG protein were synthesized from the construct with the wild-type yifK sequence than from the construct harboring the CA49,50 to UC49,50 change.
Figure 8.
Mutational analysis of the enhancer element.
Each base of an 8-nt sequence spanning positions +48 to +55 in yifK mRNA was randomized and changes were introduced in the chromosome of strain MA11594 (yifK-lacZY ΔgcvB) as described in the legend to Figure 5 (primers used to generate the set of mutagenized fragments are listed in Table S2). Mutants obtained were screened by PCR (primers ppF45/ppF62) and DNA sequencing. All possible replacements (a total of 24 mutants) were identified. These strains were assayed for ß-galactosidase activity. Typically, duplicate or triplicate ß-galactosidase measurements were carried out in parallel for all variants of any given position and the wild-type strain, whose value was set to 100. Standard deviations were less than 5% of the mean in all cases. Representative examples of triple substitutions in the upstream ACA are also shown.
Figure 9.
Increasing the distance between the enhancer and the translation initiation region.
Constructs were made as described in the legend to Figure 5 using strain MA11594 (yifK-lacZY ΔgcvB) as recipient and fragments amplified by reciprocal priming of oligonucleotides described in Table S2. A 7-nt segment (boxed in grey), duplicating the SD sequence (boxed in green), was inserted between the ACA triplets (underlined in red) and the SD sequence (construct n. 3). A G to C change was then introduced in either copy of the SD (constructs n. 4 and 5; construct n. 2 shows the effect of this change in a strain with a single SD). The upstream ACA was converted to GGG in the constructs carrying either SD sequence mutated (constructs n. 6 and 7). The same procedure was used to replace the SD-AUG interval of yifK with the corresponding segment from the chiP gene (constructs n. 8 and 9). ß-galactosidase activity was measured as described in the legend to Figure 8. The activity of the wild-type strain (construct n. 1) was set to 100. Standard deviations were less than 5% of the mean in all cases. The data (see also strains' phenotypes on MacConkey-lactose plates) show that the upstream ACA maintains its enhancer effect when placed further upstream from the initiation region, independent of the spacing between the SD sequence and the starting AUG.
Figure 10.
Randomizing ACA motifs in the ribosome binding site of dppA gene.
A dppA-lacZ translational fusion was constructed by converting a KanR insertion derived from plasmid pKD13 [47] to a lac fusion using plasmid pCE40 [56]. A tetAR insertion deleting the 15 bp ACA-encoding segment was constructed using a fragment amplified with primer pair ppN82/ppN83 (Tables S2 and S3). Next, the tetAR insert was replaced with a PCR-amplified fragment (reciprocal priming of ppN85/ppN86) containing a randomized sequence in the ACA-encoding portion. Tetracycline-sensitive recombinants were selected as described ([52], see Materials and Methods) and subsequently screened on MacConkey-lactose indicator plates. A number of isolates were characterized by DNA sequence analysis and ß-galactosidase assays. The activity of the wild-type strain was set to 100.