|Home | About | Journals | Submit | Contact Us | Français|
RamA is a transcription factor involved in regulating multidrug resistance in Salmonella enterica serovar Typhimurium SL1344. Green fluorescent protein (GFP) reporter fusions were exploited to investigate the regulation of RamA expression by RamR. We show that RamR represses the ramA promoter by binding to a palindromic sequence and describe a superrepressor RamR mutant that binds to the ramA promoter sequence more efficiently, thus exhibiting a ramA inactivated phenotype.
In most members of the family Enterobacteriaceae, but not Escherichia coli or Shigella spp., the RamA transcription activator plays a role in the regulation of the AcrAB-TolC efflux pump in response to environmental conditions, and its overexpression confers multidrug resistance (8, 13). Studies of Salmonella enterica serovar Typhimurium (1, 9, 10, 14) have identified a gene, ramR, located upstream of ramA, that encodes a repressor of ramA transcription. Point mutations and insertions have been identified in multidrug-resistant clinical and veterinary isolates of S. Typhimurium and other S. enterica serovars, and it has been hypothesized that these mutations and insertions inactivate RamR and cause increased expression of RamA (1, 6).
In this study, we used green fluorescent protein (GFP) reporter gene technology to investigate RamR-dependent regulation of the ramA promoter and to further understand the regulation of ramA and multidrug resistance. To do this, a DNA fragment carrying the ramA promoter region, previously described by Baucheron et al. (3) flanked by BamHI and XbaI restriction sites (VR1 [Fig. 1]) was cloned into the pMW82 GFP reporter vector as described previously (4), transformed into Salmonella enterica serovar Typhimurium SL1344 or a ramR aph derivative. GFP expression was measured under standard conditions, and the values are an empirical measurement of ramA promoter activity. Before the fluorescence assays were performed, growth kinetics in LB broth were determined for each strain (data not shown). There were no significant differences in growth kinetics for the strains, suggesting that the cultures would grow at the same rate in the experiment measuring fluorescence. Thus, from overnight cultures of test and control strains, a 4% inoculum was transferred to prewarmed LB broth supplemented with 50 μg/ml ampicillin and/or kanamycin, as required, and incubated at 37°C with agitation until the cultures reached late logarithmic growth phase (optical density [OD] at a wavelength of 600 nm of approximately 0.9). Aliquots of cultures (100-μl aliquots) were loaded into the wells on 96-well plates (Corning), and fluorescence from each construct was measured every 3 min for 5 h at an excitation wavelength of 492 nm and an emission wavelength of 520 nm, using a FLUOStar Optima (BMG Labtech) plate reader. Each experiment was performed on three separate occasions, and two biological and three technical repeats were included in each experiment. The GFP expression data shown in Table 1 are the means of readings taken every 3 min. GFP expression in the SL1344ramR::aph strain was 5.6-fold higher than GFP expression in strain SL1344 (Table 1). Deletion mutants, strains with mutations in the ramA promoter region, postulated to contain the RamR binding site, were constructed (Fig. 1 and Table 1). The high level of GFP expression, seen when ramR was inactivated, was reduced by over 90% by the VR1Δ5 deletion that removed the ramA promoter −10 and −35 elements (Fig. 1B and Table 1). However, the VR1Δ4 deletion of downstream sequences hardly affected GFP expression (Fig. 1C). This suggests that the −10 and −35 elements identified in the ramA promoter region by Abouzeed et al. (1) are essential elements required for promoter activity.
To check that the high level of GFP expression in the ramR::aph strain was due to the lack of RamR, the ramR::aph mutation was complemented with the low-copy-number pBR1MCS2 plasmid carrying ramR (pBR1MCS2ramR) (1); GFP expression fell to the level found with S. enterica serovar Typhimurium SL1344 (Table 1). In parallel to this study, the binding site of RamR and the ramA promoter has been experimentally identified recently by Baucheron et al. (3). To confirm this, we constructed a derivative of the VR1 fragment carrying the G140T and C151T mutations in the proposed DNA target for RamR. GFP expression from pMW82 carrying the wild-type ramA promoter region (VR1) or mutant VR1 fragment in the SL1344ramR::aph strain carrying pBR1MCS2ramR was determined. The two mutations in the palindrome cause a 14.6-fold increase in GFP expression (Table 1), confirming that the palindromic sequence located downstream of the ramA promoter −10 element is the DNA target for RamR.
To investigate which region(s) of the RamR protein interacts with the DNA target, we used the plasmids described above that carry the ramA promoter fragments and the ramR gene. Error-prone PCR (used as described in reference 5) was used to create a library of random mutations throughout the ramR gene cloned in pBR1MCS2ramR. The library was transformed into S. enterica serovar Typhimurium SL1344ramR::aph carrying pMW82-VR1 with the G140T and C151T mutations [pMW82-VR1(G140T/C151T)]. We hypothesized that recovery of repression in this background could identify regions of the RamR protein important in binding its DNA target. Therefore, we searched for a RamR mutant that had recovered the ability to repress expression from the mutated VR1 fragment. After screening 5,000 colonies using fluorescence-activated cell sorting (FACS), we identified one mutant pBR1MCS2ramR derivative carrying a single-base change that causes a substitution of valine for glutamic acid at position 118 in RamR (EV118). To quantify repression by the mutant RamR, we retransformed the pBR1MCS2ramR EV118 derivative into strain SL1344ramR::aph carrying pMW82-VR1 or pMW82-VR1(G140T/C151T). EV118 substitution in RamR improved repression of the ramA promoter carried on both the starting VR1 fragment and the VR1 derivative carrying the G140T and C151T mutations (Fig. 2). RamR is a member of the TetR family of transcriptional repressors, most of which contain an N-terminal DNA-binding domain and a C-terminal ligand-binding regulatory domain, separated by a linker (3, 7, 11, 12). The EV118 substitution is located within the linker region, which may affect RamR-mediated repression.
To investigate the effect of the EV118 substitution in RamR, we made use of the previous observation (2) that inactivation of ramA in S. enterica serovar Typhimurium SL1344 conferred better growth in the presence of subinhibitory concentrations of 53 compounds. Since RamR represses ramA expression, we reasoned that a RamR superrepressor would produce the same phenotype as a ramA mutant would. We performed growth experiments (Fig. 3) by growing bacterial cells in minimal medium at 37°C for 16 h and measuring absorbance values every 10 min at an emission wavelength of 600 nm, using a FLUOStar Optima (BMG Labtech) plate reader. To ensure that the effect we observe is due to the mutation and not just a plasmid effect, we included S. enterica serovar Typhimurium SL1344(pMW82-VR1) and SL1344(pMW82-VR1, pBR1MCS2ramR) as controls. Growth of SL1344 in minimal medium with subinhibitory concentrations of fusidic acid (500 μg/ml), nafcillin (500 μg/ml), and oxacillin (500 μg/ml) showed that the superrepressor had a similar phenotype to inactivation of ramA.
In summary, we have exploited GFP reporter fusions to measure RamR-dependent regulation of the ramA promoter, which confirmed the identification of the DNA target for RamR as described by Baucheron et al. (3) and to find a RamR superrepressor. Although our data strongly suggest binding of RamR to the promoter region of ramA, direct evidence by performing electrophoretic mobility shift assays (EMSA) is required to support this. The behavior of the superrepressor mutant suggests the importance of association and dissociation of the wild-type repressor to its promoter sequence and how this would affect promoter activity. Repressors such as RamR may dissociate rapidly from its operator, as this would be more beneficial in terms of responding to environmental stimuli.
We thank Tim Overton and Lorraine Wallace for help with FACS and Dave Lee for advice.
This work was supported by the MRC Programme grant GO501415.
Published ahead of print 4 September 2012