Mg²⁺ Transporters and F₁F_O ATP Synthase Inhibitor Repress rRNA Transcription Initiation

PhoP appears to repress rrs transcription indirectly, via a PhoP-regulated gene(s). This is because the purified PhoP protein did not alter the mobility of DNA fragments corresponding to the rrsC P1 and P2 promoters (Figure S1C), which lack sequences resembling a PhoP box (Kato et al., 1999; Zwir et al., 2012). The PhoP protein shifted a DNA fragment corresponding to the phoP promoter (Figure S1C), which was used as a positive control (Soncini et al., 1995; Shin and Groisman, 2005), but not one harboring the PhoP-independent hisL promoter, which was used as a negative control (Figure S1C).

We hypothesized that the PhoP-activated mgtA gene and mgtCBR operon might be responsible for the regulation of rrs transcription taking place in low Mg²⁺. The mgtA and mgtB genes encode distinct Mg²⁺ transporters (Snavely et al., 1989a and 1989b), and mgtC encodes an inhibitor of the F₁F_O ATP synthase (Lee et al., 2013) (MgtR is a regulator of MgtC proteolysis; Alix and Blanc-Potard, 2008). Transcription of the mgtA and mgtCBR coding regions is promoted when cytosolic Mg²⁺ levels decrease below a certain threshold (Cromie et al., 2006; Spinelli et al., 2008). Together, MgtA, MgtB and MgtC maintain cytosolic Mg²⁺ homeostasis by importing Mg²⁺ into the cytosol and by freeing Mg²⁺ from high affinity complexes with ATP (Groisman et al., 2013; Pontes et al., 2015a).

Both an mgtC mutant and an mgtA mgtB double mutant exhibited elevated rrs transcription during growth in low Mg²⁺ (Figure 2B). This was also true for mgtA mgtC and mgtB mgtC double mutants (Figure 2B), but not for mgtA or mgtB single mutants, which displayed minor or no alterations in rrs transcription (Figure 2B). The observed effects are specific to rrs because ompA mRNA levels were the same in all examined strains (Figure 2B). The elevated rrs transcription in the mgtC and mgtA mgtB mutants was reduced to the level observed in wild-type cells by plasmids expressing MgtC and MgtA, respectively, but not by the plasmid vector (Figures S1D and S1E). Moreover, the joint expression of the mgtC and mgtB genes from a heterologous promoter lowered rrs transcription in the phoP mutant to near wild-type levels (Figure S1F and S7C), but isogenic plasmids expressing mgtC or mgtB, or the plasmid vector did not (Figure S1F and S7C).

In rapidly growing cells, the majority of rRNA is transcribed from the rrs P1 promoters (Sarmientos and Cashel, 1983). To determine whether the MgtA, MgtB and MgtC proteins exert their effects at the transcription initiation level, we constructed a transcriptional fusion between the Salmonella rrsC P1 promoter and a promoterless gfp gene, and measured the fluorescence produced in three isogenic strains. rrsC P1 promoter activity was 12- and 7-fold higher in mgtC and mgtA mgtB mutants than in wild-type Salmonella, respectively (Figure S2A and S2B). Cumulatively, the data presented in this section indicate that a decrease in cytosolic Mg²⁺ triggers the production of the MgtA, MgtB and MgtC proteins, which, in turn, repress rrs transcription initiation.

MgtC Represses rRNA Transcription by Reducing ATP Levels and by Increasing Guanosine Tetraphosphate (ppGpp) Amounts

Transcription initiation at the rrs P1 promoter is regulated by the concentration of the initiating nucleotide triphosphate (iNTP) (Gourse, 1988; Gaal et al., 1997; Schneider et al., 2002) and by the levels of (p)ppGpp, a second messenger mediating the response to nutrient limitation often referred to as the stringent response (Cashel and Gallant, 1969; Sarmientos and Cashel, 1983; Sarmientos et al., 1983; Xiao et al., 1991; Ross et al., 2013). The iNTP and (p)ppGpp regulate P1 promoters by controlling the stability of the RNA polymerase promoter open complex (Gaal et al., 1997; Barker et al., 2001).

ATP is the iNTP in all Salmonella rrs P1 promoters (Figure S2C), and MgtC lowers intracellular ATP levels (Lee et al., 2013). We therefore reasoned that the abnormal rRNA accumulation displayed by the mgtC mutant might be due to heightened ATP levels stimulating rrs P1 transcription. As proposed, ATP hydrolysis promoted by the ATPase-expressing plasmid pATPase (Koebmann et al., 2002; Pontes et al., 2015b) drastically reduced rrs levels in the mgtC mutant, whereas the plasmid vector had no effect (Figure 3A and Figure S7A). By contrast, plasmid pATPase did not alter the mRNA levels of the control ompA gene (Figure 3A). Given that pATPase did not lower rrs transcription of the mgtC mutant to the levels displayed by wild-type Salmonella (Figure 3A and Figure S7A), we hypothesized that MgtC might also increase (p)ppGpp amounts. Such a dual mode of rrs regulation by ATP and (p)ppGpp would be analogous to that exerted by the FbaA protein (Schneider and Gourse, 2003b).

Open in a separate window

Figure 3

MgtC Represses rrs Transcription by Reducing ATP Levels and Promoting ppGpp Accumulation During Cytosolic Mg²⁺ Starvation

(A) Primer extension analyses of wild-type (14028s), mgtC (EL4) and mgtC (EL4) Salmonella harboring the ATPase expressing plasmid pATPase or the plasmid vector (pCP44) during growth in low (10 μM) Mg²⁺. For rrs primer extension: lanes 1 and 2, and 3 and 4 correspond to lanes 1 and 2, and 7 and 8 of the gel image displayed on Figure S7A, respectively. (B) Quantification of ppGpp levels in wild-type (14028s), mgtC (EL4) and mgtC (EL4) strains of Salmonella harboring the mgtC-expressing plasmid (pMgtC) or the plasmid vector (pUHE-21-2-lacI^q) during growth in low Mg²⁺. Error bars represent standard deviations. Graph shows the average of four independent experiments. (C) Primer extension analyses of wild-type (14028s), mgtC (EL4), relA spoT (MP342) and mgtC relA spoT (MP343) Salmonella growth in MOPS glucose medium containing low (10 μM) Mg²⁺. For rrs primer extension: lanes 1 and 2, and 3 and 4 correspond to lanes 1 and 2, and 5 and 6 of the gel image displayed on Figure S7B, respectively. (D) Primer extension analyses of wild-type (14028s), mgtC (EL4), and mgtC relA (MP335) Salmonella grown in low Mg²⁺, in the presence/absence of serine hydroxamate (Shx, 50 μg/ml). (E) Primer extension analyses of wild-type (14028s), mgtC (EL4), mgtC rpoB_L533P (MP571) and mgtC rpoB_Δ532A (MP570) Salmonella during growth in low Mg²⁺. (F) Primer extension analyses of wild-type (14028s), mgtC (EL4) and mgtC (EL4) Salmonella with pATPase or the plasmid vector (pCP44) during growth in low (10 μM) Mg²⁺, in the presence/absence of 50 μg/ml Shx. Unless stated otherwise, strains were grown in N-minimal medium. All measurements were carried following 6 h of growth in low Mg²⁺ (OD₆₀₀ ≈ 0.4–0.6). Primer extension analyses are representative of 3–4 independent experiments. See also Figures S1, S2, S3 and S7.

The mgtC mutant displayed lower ppGpp levels than wild-type Salmonella (Figures 3B and S3A). This defect was corrected by heterologous expression of the mgtC gene, but not by the plasmid vector (Figures 3B and S3A). A number of experiments indicated that the MgtC-mediated repression of rRNA transcription results from its effects on the cellular pools of both ppGpp and ATP. First, a relA spoT double mutant, which is unable to synthesize ppGpp (Xiao et al., 1991) (Figure S3A) but retains normal ATP levels during cytosolic Mg²⁺ starvation (Figure S3B), displayed more rrs transcription than wild-type Salmonella (Figure 3C and Figure S7B). Furthermore, rrs transcription was similar in mgtC single and mgtC relA spoT triple mutants, and higher than those in the relA spoT double mutant (Figure 3C and Figure S7B). Second, serine hydroxamate (Shx), a serine analog that triggers RelA mediated (p)ppGpp production (Tosa and Pizer, 1971; Haseltine and Block, 1973; Figure S3A), decreased rrs transcription in the mgtC mutant but not to wild-type levels (Figure 3D). Third, alleles of the RNA polymerase beta-subunit gene rpoB that shorten the half-life of promoter open complexes (Zhou and Jin, 1998) decreased rrs transcription in the mgtC mutant (Figure 3E). Despite making RNA polymerase behave stringently (Zhou and Jin, 1998), these rpoB alleles did not normalize rrs transcription to wild-type levels (Figure 3E). Fourth, Shx treatment restored rrs transcription to wild-type levels if the mgtC mutant carried plasmid pATPase but not when harboring the vector control (Figure 3F). As expected ompA mRNA levels were not affected by pATPase (Figure 3A), Shx treatment (Figure 3D) or the rpoB allele (Figure 3E). Together, these results demonstrated that MgtC represses rrs transcription both by reducing ATP levels and by accumulating ppGpp.

MgtA and MgtB Repress rRNA Transcription by Preventing ATP Accumulation

During growth in low Mg²⁺, PhoP promotes transcription of the mgtA and mgtB genes (Soncini et al., 1996) and muzzles the activity of the Mg²⁺ channel CorA (Alteri et al., 2011). This response indicates that Salmonella relies on MgtA and MgtB to maintain physiological cytosolic Mg²⁺ levels. As suggested, when experiencing low Mg²⁺ conditions, the concentration of free cytosolic Mg²⁺ was lower in the mgtA mgtB double mutant than in wild-type Salmonella (Figures 2C and S4A–C). The MgtA-expressing plasmid restored cytosolic Mg²⁺ to wild-type levels, an isogenic plasmid expressing mgtB increased cytosolic Mg²⁺ only partially, and the vector control had no effect (Figure 2C). These results indicate that MgtA functions as the main Mg²⁺ transporter under the investigated conditions.

Mg²⁺ is essential for ribosome assembly (Gesteland, 1966). Because most cellular ATP is consumed in translation (Stouthamer, 1973), a sudden decrease in protein synthesis resulting from insufficient cytosolic Mg²⁺ would increase ATP levels, thus stimulating rrs P1 promoters (Gaal et al., 1997; Schneider et al., 2002). In agreement with this notion, we established that ATP levels were ~10-fold higher in the mgtA mgtB double mutant than in the wild-type strain (Figure 2D), and ~4-fold higher in the mgtA single mutant than in the mgtB single mutant, which retained wild-type ATP levels (Figure 2D). ATP levels were ~50-fold higher in the mgtC mutant than in wild-type Salmonella (Figure 2D). Whereas the mgtA-expressing plasmid restored ATP levels of the mgtA mgtB double mutant to those of the wild-type strain, the isogenic mgtB-expressing plasmid lowered ATP levels of the mgtA mgtB double mutant to those of the mgtA single mutant, and the plasmid vector had no effect (Figure 2D). Notably, we established that a mgtC and mgtA mgtB mutations had an additive effect on ATP levels of cells grown in low Mg²⁺ (Figure S1G), indicating that MgtC and MgtA MgtB decrease ATP levels through distinct pathways. Thus, the rrs transcription profile (Figure 2B) mirrors the ATP levels of the Mg²⁺ transporter mutants.

Plasmid pATPase decreased rrs transcription in the mgtA mgtB double mutant to that of wild-type Salmonella (Figure 2E), whereas the vector control had no effect (Figure 2E). As expected, neither pATPase nor the plasmid vector altered ompA mRNA levels, which were the same in all strains (Figure 2E). Taken together, these results indicate that MgtA and MgtB inhibit rrs transcription by preventing a rise in ATP levels.

RNA/Protein Ratio, Growth Rate and the Efficiency of Translation

Ribosome synthesis is normally coupled to growth rate, such that rapidly growing cells have more ribosomes (Schaechter et al., 1958; Bremer and Dennis, 1996). The total RNA/protein ratio is also proportional to the amount of ribosomes and, thus, to the growth rate (Schaechter et al., 1958; Bremer and Dennis, 1996). Surprisingly, the RNA/protein ratio was higher in the phoP mutant than in wild-type Salmonella (Figure 2F) even though the phoP mutant grows slower than wild-type Salmonella in low Mg²⁺ (Soncini et at., 1996). Such uncoupling between the RNA/protein ratio and growth rate has been observed in mutants with slow translation rates, and in cells whose protein synthesis capability was compromised by pharmacological treatment (Nomura and Watson, 1959; Mikkola and Kurland, 1991). These findings suggested that the phoP mutant has decreased efficiency of protein synthesis during cytosolic Mg²⁺ starvation.

Cytosolic Mg²⁺ Homeostasis is Required for Ribosome Assembly

A decrease in the efficiency of protein synthesis in the phoP mutant could reflect a lower rate of peptide chain elongation, and/or a decrease in the fraction of ribosomes actively participating in protein synthesis. Because PhoP regulates rrs transcription via MgtA, MgtB and MgtC, and the mgtC single and mgtA mgtB double mutants also displayed elevated RNA/protein ratios (Figure 2F), we asked whether mutants lacking these proteins exhibited altered polysome profiles. Both the mgtC mutant and the mgtA mgtB double mutant accumulated 30S and 50S ribosomal subunits that did not efficiently assemble into 70S ribosomes (Figure 4A). This is in contrast to wild-type Salmonella, which assembled ribosomal subunits into 70S ribosomes that partake into mRNA translation as polysomes both in high (Figure S5A) and low Mg²⁺ (Figure 4A). The mgtC-expressing plasmid lowered the amounts of 30S and 50S subunits and increased polysome amounts in the mgtC mutant (Figure 4A). And the same was true for complementation of the mgtA mgtB double mutant with the mgtA-expressing plasmid (Figure 4A). The vector control rescued neither the mgtC nor the mgtA mgtB mutants (Figure 4A). These data established that MgtA, MgtB and MgtC are necessary for the assembly of functional ribosomes when bacteria experience low cytoplasmic Mg²⁺.

Open in a separate window

Figure 4

MgtA, MgtB and MgtC Promote Ribosome Assembly, Thereby Enabling Translation During Cytosolic Mg²⁺ Starvation

(A) Polysome analyses of wild-type (14028s), mgtC (EL4), and mgtC (EL4) Salmonella harboring the mgtC-expressing plasmid (pMgtC), or the plasmid vector (pUHE-21-2-lacI^q), mgtA mgtB (EG16741) and mgtA mgtB (EG16741) Salmonella harboring the mgtA-expressing plasmid (pMgtA) or the plasmid vector (pUHE-21-2-lacI^q) during growth in low (10 μM) Mg²⁺. Cells were collected for polysome analysis following 6 h (low Mg²⁺) of growth (OD₆₀₀ ≈ 0.4–0.6). Polysome profiles are representative of four independent experiments. (B) Quantification of L-azidohomoalanine (AHA) labeling of wild-type (14028s), mgtC (EL4) and mgtC (EL4) Salmonella harboring pMgtC or the plasmid vector (pUHE-21-2-lacI^q) following growth in low (10 μM) Mg²⁺. (C) Quantification of AHA labeling of wild-type (14028s), mgtA mgtB (EG16741) and mgtA mgtB (EG16741) Salmonella harboring pMgtA or the plasmid vector (pUHE-21-2-lacI^q) during growth in low Mg²⁺. Measurements were carried following 6 h growth in low Mg²⁺ N-minimal medium lacking methionine (OD₆₀₀ ≈ 0.4–0.6). Error bars represent standard deviations. Graphs are the average of 4 independent experiments. See also Figures S5 and S6.

Cytosolic Mg²⁺ Homeostasis Enables Translation

The data presented above suggest a model whereby maintenance of a free cytosolic Mg²⁺ pool supports ribosomal assembly, which makes translation possible (Figures 1A and ​and1B).1B). As the cytosolic concentration of free Mg²⁺ decreases, the assembly of 70S ribosomes is compromised and translation slows down (Figures 1C and ​and1D).1D). Wild-type Salmonella responds by increasing Mg²⁺ intake and by reducing ATP levels. The latter action alleviates low cytoplasmic Mg²⁺ stress both by releasing Mg²⁺ ions trapped in complexes with negatively charged ATP molecules and by reducing rRNA transcription (Figures 1E and ​and1F).1F). The inability to mount such a response results in a self-perpetuating cycle in which increased ATP levels stimulate the synthesis of ribosomal subunits not competent for translation (Figure 4A).

To test the proposed model, we examined protein synthesis in vivo by measuring the incorporation of the non-toxic methionine analog L-azidohomoalanine (AHA) into nascent proteins (Dieterich et al., 2007). The mgtC mutant displayed much lower AHA levels than wild-type Salmonella (Figures 4B and S5B). The protein synthesis defect of the mgtC mutant appears to result primarily from excess ATP because the mgtC mutant was rescued not only by the mgtC-expressing plasmid (Figure 4B and S5B) but also by pATPase (Figure S6A). By contrast, neither the mgtA-expressing plasmid nor the vector control corrected the mgtC mutant defect (Figures S6A and S6B). These results indicate that by inhibiting ATP synthesis, MgtC frees up Mg²⁺ ions required for translation.

The mgtA mgtB double mutant also incorporated less AHA than the wild-type strain (Figures 4C and S5C). The defect of the mgtA mgtB mutant was corrected by the increased cytosolic Mg²⁺ conferred by the mgtA-expressing plasmid, but not by the vector control (Figure 4C and S5C). In contrast to the rescue of the mgtC mutant (Figure S6A), pATPase-mediated ATP hydrolysis did not correct protein synthesis in the mgtA mgtB mutant (Figure S6C). These results likely reflect that the mgtA mgtB double mutant fails to import (as opposed to re-distribute) Mg²⁺ ions required for translation.

Wild-Type Salmonella Reduces the Rate of Protein Synthesis during Growth in Low Cytoplasmic Mg²⁺

Growth of wild-type Salmonella in low Mg²⁺ media is characterized by a logarithmic phase taking place in the first 4 h (OD₆₀₀ ≤ 0.4), followed by a linear, slow-growing phase (Figure 5A). The onset of the slow-growing phase coincides with a decrease in cytosolic Mg²⁺ that triggers the synthesis of MgtA, MgtB and MgtC (Figure 5A) (Cromie et al. 2006; Cromie and Groisman, 2010; Alix and Blanc-Potard, 2008). By reducing the synthesis of ribosomal components (Figures 2B and ​and5A),5A), MgtA, MgtB and MgtC are expected to reduce the translation rate of the wild-type strain. As predicted, AHA incorporation decreased two-fold between 2.5 and 5.5 h of growth in low Mg²⁺ (Figures 5B and ​and5C),5C), which correspond to times before and after the synthesis of MgtA, MgtB and MgtC, respectively. A reduced translation rate may help Salmonella enter a semi-quiescent state characteristic of chronic infections and increased tolerance to antimicrobial agents. Notably, the MgtC protein, which plays a key role in this transition (Figure 5A), is required for Salmonella survival in a mouse model of chronic infection (Lawley et al., 2006).

Open in a separate window

Figure 5

MgtA, MgtB and MgtC Readjust Cellular Translation Rates During Cytosolic Mg²⁺ Starvation

(A) Growth curve of wild-type Salmonella (14028s) in low (10 μM) Mg²⁺. The arrow indicates the time when MgtA, MgtB and MgtC are expressed. The color intensity of the legend depicts the concentration of Mg²⁺ in the cytosol. (B) Quantification of L-azidohomoalanine (AHA) labeling of wild-type Salmonella (14028s) prior to (2.5 h) and after (5.5 h) cytosolic Mg²⁺ starvation. Error bars represent standard deviations. Graphs show the average of four independent experiments. (C) SDS-PAGE analysis of AHA labeling of wild-type (14028s) prior to (2.5 h) and after (5.5 h) cytosolic Mg²⁺ starvation. The gel is representative of four independent experiments. All experiments were carried out in low Mg²⁺ (10 μM) Mg²⁺ N-minimal medium lacking methionine.

E. coli Buffers Translation During Cytosolic Mg²⁺ Starvation by Importing Mg²⁺ and Lowering ATP Levels

E. coli has homologs of the phoP, phoQ and mgtA genes (Kato et al., 1999) but lacks mgtC and mgtB (Blanc-Potard et al., 1999). We thus wondered whether E. coli is capable of a response analogous to that described above for Salmonella when facing low cytosolic Mg²⁺.

We determined that when Mg²⁺ is removed from the growth medium, wild-type E. coli experienced a decrease in both cytosolic Mg²⁺ (Figure 6A) and ATP levels (Figure 6B). By contrast, ATP levels actually increased in an E. coli phoP mutant, and to a lesser degree in an mgtA mutant (Figure 6B). In E. coli, PhoP promotes transcription of the yhiD gene, which specifies a product with low sequence similarity to MgtC and hypothesized to function as an Mg²⁺ transporter (Alteri et al., 2011; Hattori et al., 2009). However, in contrast to MgtC, YhiD does not interact with and inhibit the F₁F_O ATP synthase (Lee et al., 2013). In agreement with this notion, a yhiD mutant displayed wild-type ATP levels (Figure 6B).

Open in a separate window

Figure 6

PhoP/PhoQ Regulates Ribosome Synthesis and Translation in E. coli During Cytosolic Mg²⁺ Starvation

(A) Time-course cytosolic Mg²⁺ concentrations of wild-type E. coli grown in medium containing 0 or 10 mM Mg²⁺. Cytosolic Mg²⁺ concentrations reflect the normalized β-galactosidase activity values obtained with a genetic reporter for cytosolic Mg²⁺ and are expressed in arbitrary units (A.U.) (see Experimental Procedures). (B) Time-course quantification of ATP levels in wild-type (MG1655), phoP (EG12976), mgtA (EG16771), yhiD (MP751) and mgtA yhiD (MP752) E. coli grown in medium lacking Mg²⁺. (C) qPCR quantification of rrs transcriptional activity in wild-type (MG1655), phoP (EG12976), mgtA (EG16771), yhiD (MP751) and mgtA yhiD (MP752) E. coli 2 h following removal of Mg²⁺ from the growth medium. (D) Quantification of L-azidohomoalanine labeling of wild-type (MG1655), phoP (EG12976), mgtA (EG16771), yhiD (MP751) and mgtA yhiD (MP752) strains of E. coli 2 h following removal of Mg²⁺ from the growth medium. All experiments were carried out in MOPS glucose medium. Error bars represent standard deviations. Graphs show the average of four independent experiments.

We established that under cytosolic Mg²⁺ starvation, translation efficiency is inversely related to rrs transcription and cytosolic ATP concentrations. That is, wild-type and yhiD E. coli displayed low ATP (Figure 6B) and rrs transcription (Figure 6C), and synthesized proteins efficiently (Figure 6D). By contrast, mutants defective in phoP, mgtA or both mgtA and yhiD had elevated ATP (Figure 6B) and rrs transcription (Figure 6C), and were defective for translation (Figure 6D). The phoP mutant displayed the highest ATP levels (Figure 6B) and rrs transcriptional activity (Figure 6C), and exhibited the greatest translation defect (Figure 6D). Taken together, the results presented in this section indicate that E. coli utilizes the PhoP-activated MgtA and yet an unidentified gene(s) to maintain translation homeostasis when faced with low cytosolic Mg²⁺.

Construction of mgtC rpoB_L533P and mgtC rpoB_δ532a Strains

Introduction of point mutation in the rpoB gene was carried out via λ-red-mediated recombination (Datsenko and Wanner, 2002), as described in the Supplemental Experimental Procedures.

Construction of Plasmid Harboring rrsC P1-gfp

Wild-type Salmonella genomic DNA was used as template in a PCR reaction with primers pair W1480 and W1482. PCR product corresponding to the promoter region of rrsC P1 (14028s chromosomal base pairs 4,113,424–605) was digested with BamHI and EcoRI and ligated into pFPV25_AAV(Choi and Groisman, 2013) digested with the same restriction enzymes. Ligation reactions were transformed into DH5α cells, and the identity of the inserts was verified by DNA sequencing.

Estimation of ATP Levels from Bacterial Samples

ATP measurements were carried out as described (Pontes et al., 2015b).

RNA Extraction and Primer Extension

RNA extraction and primer extension reactions were carried out using primers W2596 (rrs leader) and W2595 (ompA) as described (Prévost et al., 2007), with the exception that 20 μg of RNA were used in reverse transcription reactions.

RNA Extraction and Quantitative PCR

RNA extraction and quantification was carried out as described (Park et al., 2010), with the exception that cDNA was synthesized from RNA samples using the Superscript VILO cDNA Synthesis Kit (Thermo Fisher Scientific) according to the manufacturer’s instructions. Relative amounts of cDNA were determined using a standard curve obtained from qPCR with serially diluted wild-type E. coli genomic DNA, wild-type Salmonella genomic DNA or pFPV25_AAV DNA. For quantification of the reporter fusion, gfp mRNA levels were normalized by mRNA levels derived from plasmid-borne bla gene. The following strategy was used to quantify rrs transcriptional activity: following rrs transcription, the structural portion of the 16S RNA is assembled into a stable 30S subunit whereas the leader region is quickly degraded (Shajani et al., 2011; Zhi et al., 2003). Thus, a decline in transcription initiation at the rrs promoters will result in a decrease in the ratio of the unstable leader to its corresponding stable structural 16S rRNA.

L-azidohomoalanine (AHA) Labeling and Quantification

Labeling of nascent protein synthesis was carried out using Click-iT Protein Reaction Buffer Kit (Thermo Fisher Scientific) according to the manufacturer’s instructions. Detailed procedure is described in Supplemental Experimental Procedures.

Polysome Analysis

Following overnight growth in high Mg²⁺ N-minimal medium, cells were washed 3x in Mg²⁺-free medium and used to inoculate (1:50) fresh N-minimal medium containing high or low Mg²⁺. After 3 (high Mg²⁺) or 6 h (low Mg²⁺) of growth (OD₆₀₀ ≈ 0.4–0.6), cells were collected, lysed and their polysomes were separated in a sucrose gradient as described (Qin and Fredrick, 2013). Sucrose gradients were separated into 30 fractions using a Gradient Station IP fractionator (BioComp), and the first 27 fractions were analyzed. The absorbance at 260 nm of the fractions was determined using a NanoDrop 8000 (Thermo Scientific). The identity of the 260 nm peaks were established by extracting fractions with phenol:chloroform, ethanol precipitating, and separating them by denaturing agarose gel electrophoresis.

Estimation of Protein/RNA Ratios

Following overnight growth in high Mg²⁺ N-minimal medium, Salmonella strains were washed 3x in Mg²⁺-free medium and used to inoculate (1:50) fresh low Mg²⁺ N-minimal medium. After 6 h of growth (OD₆₀₀ ≈ 0.4–0.6), cells were collected, resuspended in dH₂O and lysed by sonication. Crude cell extracts were used for protein and RNA quantification using a Pierce BCA Protein Assay Kit (Thermo Fisher Scientific) and orcinol assay, respectively.

Estimation of Cytosolic Mg²⁺

Quantification of cytosolic Mg²⁺ concentration was carried out using a genetic reporter for intracellular Mg²⁺ (Cromie et al., 2006) and is detailed in Supplemental Experimental Procedures.

Electrophoretic mobility shift assay

Electrophoretic mobility shift assays were performed as described (Choi and Groisman, 2013), with the minor modifications described in the Supplemental Experimental Procedures.

The authors would like to thank Dr. Peter R. Jensen (Technical University of Denmark) for kindly providing plasmids pCP44 and pCP41-AtpAGD, Anastasia Sevostiyanova for kindly providing purified PhoP*-His, Hubert Salvail for assistance with primer extension analyses, Anastasia Sevostiyanova, Richard L. Gourse and Michael Ibba for critical reading of this manuscript. This research was supported by grant AI49561 from the National Institutes of Health to EAG.