Salmonella expresses foreign genes during infection by degrading their silencer

Significance

Foreign genes confer new properties upon organisms. However, their expression often requires overcoming the silencing effects of the heat-stable nucleoid structuring (H-NS, also referred to as histone-like nucleoid structuring) protein, which binds to AT-rich foreign DNA in enteric bacteria. We establish that Salmonella enterica degrades H-NS when inside macrophages and that the resulting decrease in H-NS abundance de-represses foreign genes, even those not bound by antisilencing DNA-binding proteins. Conservation in the amino acid sequences of both the protease degrading H-NS and H-NS itself suggests that enteric bacteria share the uncovered strategy to express foreign genes silenced by H-NS.

Abstract

The heat-stable nucleoid structuring (H-NS, also referred to as histone-like nucleoid structuring) protein silences transcription of foreign genes in a variety of Gram-negative bacterial species. To take advantage of the products encoded in foreign genes, bacteria must overcome the silencing effects of H-NS. Because H-NS amounts are believed to remain constant, overcoming gene silencing has largely been ascribed to proteins that outcompete H-NS for binding to AT-rich foreign DNA. However, we report here that the facultative intracellular pathogen Salmonella enterica serovar Typhimurium decreases H-NS amounts 16-fold when inside macrophages. This decrease requires both the protease Lon and the DNA-binding virulence regulator PhoP. The decrease in H-NS abundance reduces H-NS binding to foreign DNA, allowing transcription of foreign genes, including those required for intramacrophage survival. The purified Lon protease degraded free H-NS but not DNA-bound H-NS. By displacing H-NS from DNA, the PhoP protein promoted H-NS proteolysis, thereby de-repressing foreign genes—even those whose regulatory sequences are not bound by PhoP. The uncovered mechanism enables a pathogen to express foreign virulence genes during infection without the need to evolve binding sites for antisilencing proteins at each foreign gene.

Horizontal gene transfer plays a critical role in bacterial evolution (1, 2). Horizontally transferred genes (HTGs; also referred to as foreign genes) can provide new properties to an organism, such as the utilization of new carbon sources, resistance to antibiotics, and virulence (1–3). However, HTGs are transcriptionally silenced in bacterial genomes by specific DNA-binding proteins in a process termed xenogeneic silencing (4). How, then, do bacteria overcome gene silencing so they can access the benefits of HTGs? Here, we establish that the facultative intracellular pathogen Salmonella enterica serovar Typhimurium promotes expression of HTGs during infection by degrading its major xenogeneic silencer.

The ability to survive inside macrophages is required for Salmonella virulence (5). This ability requires many HTGs (6–10) as well as the ancestral gene phoP, which specifies a DNA-binding transcriptional regulator necessary for expression of HTGs (11–13), some of which enable intramacrophage survival (7–10). Salmonella experiences a mildly acidic pH inside macrophage phagosomes (14, 15). Phagosome acidification is necessary for Salmonella survival inside macrophages (15). This is due, in part, to a mildly acidic pH activating virulence regulatory proteins necessary to overcome gene silencing by the heat-stable nucleoid structuring (H-NS) protein (14, 16–20), the dominant xenogeneic silencer in Salmonella (4, 10, 21–23).

H-NS amounts are reported to stay constant when Salmonella is grown in different laboratory media (22, 24–26). To our knowledge, however, whether H-NS amounts change during infection of a mammalian host has not been examined. We now report that Salmonella reduces H-NS amounts when inside macrophages, allowing expression of HTGs. We establish that the reduction in H-NS amounts is due to H-NS proteolysis by the Lon protease and determine that H-NS is protected from Lon when bound to DNA. We establish that the PhoP protein promotes H-NS degradation by displacing H-NS from DNA. PhoP’s action enables transcription of HTGs, even those lacking a PhoP-binding site in their promoter regions. The uncovered antisilencing strategy enables Salmonella to express foreign genes during infection and is likely to operate in other enteric pathogens.

Results

Salmonella Reduces H-NS Amounts Inside Macrophages, Overcoming Silencing of HTGs.

To understand how Salmonella expresses HTGs during infection, we examined H-NS amounts at different times following Salmonella internalization by the macrophage-like cell line J774A.1. H-NS amounts decreased within 1 h and reached levels that were 16-fold lower by 6 h post infection (Fig. 1A). The marked decrease in H-NS amounts appears to result from H-NS degradation because H-NS amounts did not decrease in a lon mutant (Fig. 1A), which lacks a cytoplasmic protease (27, 28) required for Salmonella virulence (29).

Fig. 1.

Salmonella overcomes H-NS silencing of HTGs inside macrophages by decreasing H-NS abundance. (A) Western blot analysis of crude extracts prepared from wild-type (JC805), phoP (JC837), and lon (JC864) Salmonella harvested from the macrophage-like cell line J774A.1 at the indicated times using antibodies recognizing the FLAG epitope or the loading control AtpB. Bacterial cells grown overnight in LB were used for the macrophage infection. Numbers above the blots correspond to the normalized relative amounts of H-NS protein. A representative of at least three independent experiments is shown. (Right) A densitometry graph for replicates. (B) mRNA abundance of the STM14_1977, ycjE, pagC, and ssaG genes produced by wild-type (JC805), phoP (JC837), and lon (JC864) Salmonella harvested from the macrophage-like cell line J774A.1 at the indicated times. The three strains exhibited similar survival inside macrophages at the investigated times. The mean and SD from three independent experiments are shown. *P < 0.05, **P < 0.01, ***P < 0.001, two-tailed t test with each mutant vs. wild type at each time point. The absence of a * indicates no significant difference between mutant and wild type.

The decrease in H-NS amounts taking place when Salmonella is inside macrophages is accompanied by an increase in the messenger RNA (mRNA) abundance of multiple HTGs (Fig. 1B) bound by H-NS during bacterial growth in laboratory media (10, 30) but transcriptionally induced inside macrophages (9). These include ssaG, which specifies a component of a type III secretion system necessary for intramacrophage survival (6). The mRNA amounts of the investigated HTGs were lower in the lon mutant than in the wild-type strain (Fig. 1B), suggesting that the Lon-dependent decrease in H-NS abundance is necessary for transcription of foreign genes. These results suggest that, when Salmonella is inside macrophages, it expresses genes silenced by H-NS by degrading H-NS.

Purified Lon Degrades Purified H-NS When H-NS Is Not Bound to DNA.

The Lon protease is directly responsible for the decrease in H-NS abundance taking place inside macrophages (Fig. 1A) (rather than Lon altering the abundance of a protein that reduces H-NS amounts) because the purified Lon protein proteolyzed the purified H-NS protein (Fig. 2A). By contrast, H-NS was stable in the absence of Lon (Fig. 2A). H-NS proteolysis was prevented if the reaction also contained pHTG plasmid DNA, which harbors HTGs silenced by H-NS and activated by the PhoP protein (Fig. 2A). Plasmid pHTG DNA specifically inhibited H-NS proteolysis [as opposed to sequestering the Lon protease, which has the ability to bind DNA (31)] because pHTG did not hinder proteolysis of β-casein (SI Appendix, Fig. S1), a Lon substrate used as a control because it does not bind DNA. We conclude that Lon degrades free H-NS but not H-NS bound to DNA.

Fig. 2.

PhoP displaces H-NS from DNA, rendering it susceptible to degradation by Lon. (A) In vitro degradation of purified H-NS protein by purified Lon protease in the presence or absence of a plasmid harboring a foreign DNA fragment containing PhoP-activated genes targeted by H-NS (pHTG) and (where indicated) purified wild-type PhoP or variant PhoP (V193M) proteins. Pyruvate kinase (PK) is part of the ATP generation mix. t_1/2, the half-life of H-NS. A representative of two independent experiments is shown. (Bottom) A densitometry graph for replicates. (B) Schematic of H-NS and the H-NS^1−50aa-GFP fusion proteins. AA, amino acids. (C) Western blot analysis of crude extracts prepared from wild-type (JC805), phoP (JC837), and lon (JC864) Salmonella harboring plasmids expressing the H-NS^1−50aa-GFP fusion protein using antibodies recognizing GFP or the loading control AtpB. Bacteria were grown to midlog phase in N-minimal media with 1 mM Mg²⁺ at pH 4.9 (acidic pH) with 20 µM of IPTG. Numbers above the blots correspond to the normalized relative amounts of H-NS^1−50aa-GFP protein. A representative of at least three independent experiments is shown.

We reasoned that DNA-binding antisilencing proteins promote H-NS proteolysis by displacing H-NS from DNA. As proposed, the purified PhoP protein restored H-NS proteolysis to the pHTG-containing reaction (Fig. 2A). This restoration required PhoP’s DNA-binding ability because a PhoP variant defective in DNA binding (SI Appendix, Fig. S2) failed to restore H-NS proteolysis (Fig. 2A). The purified PhoP did not promote H-NS proteolysis in reactions lacking pHTG DNA (SI Appendix, Fig. S3). In addition, H-NS was degraded similarly when PhoP was preincubated with pHTG DNA as when H-NS was preincubated with pHTG DNA (SI Appendix, Fig. S4A). Furthermore, PhoP dislodged H-NS from DNA (SI Appendix, Fig. S4B). Cumulatively, the results presented above argue that PhoP promotes H-NS degradation by displacing H-NS from DNA.

In agreement with the results of the biochemical experiments presented above, H-NS amounts were lower in wild-type Salmonella than in the phoP mutant at 6 h post internalization by macrophages (Fig. 1A). The higher H-NS amounts present in the phoP mutant resulted in a lower mRNA abundance of HTGs than in wild-type Salmonella (Fig. 1B). That wild-type and mutant Salmonella differ in both H-NS amounts and mRNA abundance of HTGs cannot be ascribed to differential intramacrophage survival because the numbers of wild-type, lon, and phoP Salmonella were similar at 6 h inside macrophages (SI Appendix, Fig. S5A). [Note that the defective intracellular survival of lon and phoP Salmonella was observed starting at 20 h (29, 32)]. Moreover, because PhoP is hardly active at 1 h inside macrophages, wild-type and phoP Salmonella exhibit a similar decrease in H-NS amounts at this time (Fig. 1A). For unknown reasons, H-NS basal amounts were slightly lower in the phoP and lon mutants than in the wild-type strain in some of the biological replicates (Fig. 1A, −0.5 h; see Mendeley: doi: 10.17632/m4vt7hwrgc.1). In addition, chromosomal copy numbers of wild-type, lon, and phoP Salmonella were also similar at the investigated time points (SI Appendix, Fig. S5B). Cumulatively, these data argue that a Lon- and PhoP-dependent decrease in H-NS amounts is necessary for expression of HTGs inside macrophages.

Growth in a Mildly Acidic pH Decreases H-NS Amounts in a Lon-Dependent Manner.

We hypothesized that a mildly acidic pH is a key signal that triggers H-NS proteolysis inside macrophages because a mildly acidic pH is critical for both Salmonella survival inside macrophages (15) and activation of several antisilencing proteins, including PhoP (14, 16–18). As hypothesized, H-NS abundance was 6.6-fold lower in wild-type Salmonella grown in laboratory media buffered to pH 4.9 than in media buffered to pH 7.6 (Fig. 3A). Thus, a mildly acidic pH is sufficient to promote a reduction in H-NS abundance, albeit one not as pronounced as that observed inside macrophages.

Fig. 3.

Degradation of H-NS by the Lon protease decreases H-NS binding to HTGs, promoting their expression. (A and B) Western blot analysis of crude extracts prepared from (A) wild-type (JC805), phoP (JC837), and lon (JC864) Salmonella grown to midlog phase in N-minimal media with 1 mM Mg²⁺ at pH 4.9 (acidic pH) or pH 7.6 (neutral pH) and (B) wild-type (JC805), lon (JC864), clpX (JC865), and clpA (JC867) Salmonella grown to midlog phase in N-minimal media with 1 mM Mg²⁺ at pH 4.9 (acidic pH) using antibodies recognizing the FLAG epitope or the loading control AtpB. Numbers above the blots correspond to the normalized relative amounts of H-NS protein. (C) Stability of H-NS was determined in wild-type (JC805), phoP (JC837), and lon (JC864) Salmonella grown to midlog phase in N-minimal media with 1 mM Mg²⁺ at pH 4.9 (acidic pH). Protein synthesis was inhibited with chloramphenicol (1 mg⋅ml⁻¹). Samples were removed at the indicated times and analyzed by Western blotting using antibodies recognizing the FLAG epitope or the loading control GroEL. t_1/2, half-life of H-NS. (D) Western blot analysis of crude extracts prepared from wild-type (JC805) and lon (JC864) Salmonella with or without plasmids expressing wild-type Lon or a variant defective in DNA binding (Lonmu; R306E K308E K310E K311E) using antibodies recognizing the FLAG epitope, Lon, or the loading control AtpB. Bacteria were grown in N-minimal media with 1 mM Mg²⁺ at pH 4.9 (acidic pH) to midlog phase. L-Arabinose was added at the denoted final concentrations. Numbers above the blots correspond to the normalized relative amounts of H-NS protein. Representatives of at least three independent experiments are shown (A–D). (E) mRNA abundance of the pagC, STM14_1977, and ycjE genes produced by wild-type (JC805), phoP (JC837), and lon (JC864) Salmonella grown to midlog phase in N-minimal media with 1 mM Mg²⁺ at pH 4.9 (acidic pH). (F) In vivo binding of H-NS to the promoter regions of the pagC, STM14_1977, and ycjE genes was determined in wild-type (JC805) and lon (JC864) Salmonella grown to midlog phase in N-minimal media with 1 mM Mg²⁺ at pH 4.9 (acidic pH). (G) In vivo binding of PhoP to the promoter regions of the pagC, STM14_1977, and ycjE genes was determined in wild-type (JC805) Salmonella grown to midlog phase in N-minimal media with 1 mM Mg²⁺ at pH 7.6 (neutral pH) or pH 4.9 (acidic pH). The mean and SD from three independent experiments are shown. *P < 0.05, **P < 0.01, ****P < 0.0001, two-tailed t test with each mutant vs. wild type; ns, not significant.

H-NS abundance was 6.6- to 8.5-fold higher in the lon mutant than in wild-type Salmonella following growth in laboratory media at pH 4.9 (Fig. 3 A and B), mimicking the differences observed inside macrophages (Fig. 1A). The H-NS half-life was <30 min in wild-type Salmonella but >2 h in the lon mutant (Fig. 3C), further supporting the notion that Lon degrades H-NS in vivo. By contrast, mutants defective in the clpA and clpX genes, which specify proteins that, together with ClpP, form the ATP-dependent proteases ClpAP and ClpXP, respectively (28), had ∼50% more H-NS than the wild-type strain (Fig. 3B). As stated above, Lon is a DNA-binding protein (31). However, Lon’s ability to bind DNA is dispensable for H-NS degradation because wild-type H-NS abundance was restored to the lon mutant both by a plasmid expressing the wild-type lon gene and by one specifying a variant defective in DNA binding (33) (Fig. 3D). The Lon-dependent decrease in H-NS abundance is specific to a mildly acidic pH because wild-type and lon Salmonella exhibited the same high H-NS abundance when grown in media buffered to pH 7.6 (Fig. 3A).

Specific DNA-Binding Proteins Activated in Mildly Acidic pH Decrease H-NS Amounts.

When bacteria were grown in a mildly acidic pH, the phoP mutant had higher H-NS abundance than wild-type Salmonella (Fig. 3A), behaving like the lon mutant (Fig. 3A). By contrast, H-NS abundance (Fig. 3A) and stability (SI Appendix, Fig. S6) were similar in the three strains when bacteria were grown in neutral pH.

The Lon and PhoP proteins are part of the same H-NS degradation pathway for the following reasons. First, the same high H-NS amounts were present in a phoP lon double mutant as in the lon and phoP single mutants (Fig. 4A). Second, a phoP-expressing plasmid restored wild-type H-NS abundance to the phoP mutant but failed to do so in the lon mutant (Fig. 4 B and C). And third, the lon-expressing plasmid restored wild-type H-NS abundance to the lon mutant but not to the phoP mutant (Fig. 4 B and C).

Fig. 4.

Lon and PhoP are responsible for decreasing H-NS amounts in acidic pH. (A–C and E) Western blot analysis of crude extracts prepared from (A) wild-type (JC805), phoP (JC837), lon (JC864), and phoP lon (JC1223) Salmonella grown to midlog phase in N-minimal media with 1 mM Mg²⁺ at pH 4.9 (acidic pH). (B and C) Wild-type (JC805), phoP (JC837), and lon (JC864) Salmonella with the plasmid vector (pVec) or plasmids expressing PhoP (pPhoP) or Lon (pLon). (E) Wild-type (JC805) and phoP (JC837) Salmonella grown to midlog phase in N-minimal media with 1 mM Mg²⁺ at pH 4.9 (acidic pH). When necessary, IPTG and L-arabinose were added at the denoted final concentrations. Samples were analyzed using antibodies recognizing the FLAG epitope, Lon or the loading control AtpB. Numbers above the blots correspond to the normalized relative amounts of H-NS protein. Representatives of at least three independent experiments are shown (A–C and E). (D) mRNA abundance of the lon gene produced by wild-type (JC805) and phoP (JC837) Salmonella grown to midlog phase in N-minimal media with 1 mM of Mg²⁺ at pH 4.9 (acidic pH). The mean and SD from three independent experiments are shown. Two-tailed t test with each mutant vs. wild type; ns, not significant.

We considered the possibility of the DNA-binding proteins SsrB and SlyA reducing H-NS abundance because, like PhoP, they are activated in a mildly acidic pH, required for virulence and necessary to overcome silencing of HTGs by displacing H-NS from DNA (14, 16–20). In addition, PhoP is a direct transcriptional activator of the ssrB (8) and slyA (18, 34) genes. As hypothesized, the ssrB and slyA null mutants displayed higher H-NS amounts than wild-type Salmonella when grown in a mildly acidic pH (SI Appendix, Fig. S7 A and B), mimicking the behavior of the phoP and lon mutants (Fig. 3A).

PhoP plays a dominant role in decreasing H-NS abundance because the phoP-expressing plasmid restored wild-type H-NS abundance to both the phoP mutant (Fig. 4B) and a phoP ssrB double mutant (SI Appendix, Fig. S7C). By contrast, an ssrB-expressing plasmid rescued the ssrB single mutant (SI Appendix, Fig. S7A) but not the phoP ssrB double mutant (SI Appendix, Fig. S7C).

Expression of the slyA gene from a heterologous promoter reduced H-NS abundance in the slyA ssrB double mutant. By contrast, the ssrB-expressing plasmid did not (SI Appendix, Fig. S7D). These results suggest that SsrB’s role in altering H-NS abundance is dependent on the presence of a functional slyA gene. Given that SlyA is necessary for transcription of ugtL (34, 35), a gene that specifies a product necessary to activate the S. enterica PhoP protein under mildly acidic pH (34, 35), SlyA may decrease H-NS abundance by increasing the amount of active PhoP protein.

Strains defective in the PhoP-activated rstA and pmrA genes retained wild-type H-NS amounts (SI Appendix, Fig. S7E), in contrast to the increased H-NS abundance displayed by the ssrB and slyA mutants (SI Appendix, Fig. S7 A and B). Because the rstA and pmrA genes encode DNA-binding proteins that are also activated in a mildly acidic pH (36, 37), these results indicate that specific DNA-binding proteins promote a decrease in H-NS abundance when Salmonella experiences a mildly acidic pH.

Transcription can displace H-NS from DNA (38). Therefore, PhoP may render H-NS a Lon substrate not only by competing with H-NS for binding to a particular DNA sequence but also by recruiting RNA polymerase (RNAP) to DNA regions bound by H-NS. We addressed the latter possibility in two independent experiments. First, the transcription inhibitor rifampicin slightly increased H-NS half-life in wild-type Salmonella (SI Appendix, Fig. S8A). This effect is phoP-dependent because rifampicin addition did not alter the H-NS half-life in the phoP mutant (SI Appendix, Fig. S8A).

Second, H-NS abundance was slightly higher in a Salmonella strain expressing the PhoP S171P variant, which is defective in recruiting RNAP, than in the isogenic strain expressing the wild-type PhoP protein (SI Appendix, Fig. S8B). Control experiments demonstrated that a Salmonella strain expressing the PhoP V193M variant, which is defective in DNA binding, had similarly large H-NS amounts as the phoP mutant (SI Appendix, Fig. S8B), which agrees with the results of the H-NS degradation experiments in vitro (Fig. 2A).

We ruled out the possibility of PhoP reducing H-NS amounts by altering Lon abundance or activity by showing the following. First, wild-type and phoP Salmonella harbored similar amounts of lon mRNA (Fig. 4D) and Lon protein (Fig. 4E). Second, wild-type and phoP Salmonella contained similar amounts of the Lon substrate SulA (SI Appendix, Fig. S9A), which accumulated in the lon mutant to higher amounts than in the wild-type strain when Salmonella was grown in defined media buffered to pH 4.9 or 7.6 (SI Appendix, Fig. S9B).

A mildly acidic pH furthers H-NS proteolysis by a mechanism that is independent of PhoP activation. This is because of the following. First, H-NS abundance hardly decreased when Salmonella was exposed to the antimicrobial peptide C18G or low Mg²⁺ (SI Appendix, Fig. S10 A and B), even though both of these conditions activate the PhoP protein as much as a mildly acidic pH (39, 40). Second, when Salmonella was grown in mildly acidic pH, a plasmid expressing the phoP gene from a heterologous promoter decreased H-NS amounts in the phoP null mutant to those displayed by wild-type Salmonella harboring the plasmid vector (SI Appendix, Fig. S10C). By contrast, H-NS amounts were uniformly high when the same strains were grown in low Mg²⁺ (SI Appendix, Fig. S10C). Cumulatively, the results in this section indicate that specific DNA-binding proteins promote H-NS degradation during growth in mildly acidic pH.

A Reduction in H-NS Amounts Reduces H-NS Binding to Foreign Genes, Resulting in Their Expression.

The PhoP- and Lon-dependent decrease in H-NS amounts is necessary for transcription of HTGs because the phoP and lon single mutants had similarly lower mRNA amounts of HTGs than wild-type Salmonella when bacteria were grown in defined media at pH 4.9 (Fig. 3E). Chromatin immunoprecipitation experiments demonstrated that more H-NS bound to HTGs in the lon mutant than in wild-type Salmonella (Fig. 3F), reflecting the higher H-NS amounts present in the former than the latter strain (Fig. 3A). Because a decrease of H-NS abundance de-represses transcription of the H-NS silenced gene csgB (41), we reasoned that a reduction in H-NS amounts should result in transcription of HTGs even when Salmonella experiences a neutral pH and is missing the proteins required to overcome silencing by H-NS.

To test the prediction made above, we examined the mRNA abundance of the HTGs STM14_1977 and ycjE in a strain engineered to produce low H-NS amounts. The engineered strain lacked the phoP and hns genes and harbored a plasmid expressing hns from a heterologous inducible promoter (Fig. 5A). [Note that inactivation of the hns gene is lethal in Salmonella but that a phoP hns double mutant is viable (10).] The STM14_1977 and ycjE genes were de-repressed despite the media having a neutral pH (Fig. 5B). Given that this growth condition fails to activate the SsrB and SlyA proteins (14, 16–20), our results argue that a reduction in H-NS abundance is sufficient to de-express HTGs silenced by H-NS.

Fig. 5.

A decrease in H-NS abundance promotes expression of HTGs in neutral pH. (A) Western blot analysis of crude extracts prepared from phoP hns (JC1397) Salmonella with a plasmid expressing H-NS-FLAG from an IPTG-inducible heterologous promoter grown to midlog phase in N-minimal media with 1 mM Mg²⁺ at pH 7.6 (neutral pH) with or without 200 µM of IPTG (+ or −, respectively) using antibodies recognizing the FLAG epitope or the loading control AtpB. Numbers above the blots correspond to the normalized relative amounts of H-NS protein. A representative of at least three independent experiments is shown. (B) mRNA abundance of the ycjE and STM14_1977 genes produced by phoP hns (JC1397) Salmonella with a plasmid expressing H-NS from IPTG-inducible heterologous promoter grown to midlog phase in N-minimal media with 1 mM Mg²⁺ at pH 7.6 (neutral pH) with or without 200 µM of IPTG (+ or −, respectively). The mean and SD from three independent experiments are shown. **P < 0.01, two-tailed t test with strains with IPTG vs. those without IPTG.

The PhoP protein promotes transcription of the STM14_1977 and ycjE genes in mildly acidic pH by virtue of decreasing H-NS amounts (as opposed to displacing H-NS from these foreign genes) because the promoter regions of these two HTGs lack sequences resembling a PhoP-binding site (13) and also because the PhoP protein did not bind to these DNA regions in vivo (Fig. 3G). The STM14_1977 and ycjE genes were not expressed during growth in low Mg²⁺ (SI Appendix, Fig. S11), despite this condition activating the PhoP protein as much as growth in a mildly acidic pH (39). Control experiments demonstrated that growth in low Mg²⁺ resulted in PhoP binding to the pagC promoter region (Fig. 3G) and transcription of the pagC gene (SI Appendix, Fig. S11), in agreement with previous results (11, 19). Altogether, the results in this section argue that conditions that reduce H-NS amounts can de-repress HTGs.

The Abundance of an H-NS Variant Lacking Its DNA-Binding Domain Is Lon Dependent but PhoP Independent.

The 137-amino-acid-long H-NS consists of an N-terminal oligomerization domain (residues 1 to 83) and a C-terminal DNA-binding domain (residues 91 to –137) joined by a short linker (23) (Fig. 2B). If PhoP renders H-NS a Lon substrate solely by displacing H-NS from DNA (Fig. 2A and SI Appendix, Fig. S8), then, a phoP mutant should display wild-type abundance of an H-NS variant defective in DNA binding. We tested this notion by investigating the abundance of a chimera consisting of the 50 N-terminal amino acids of H-NS fused to the full-length GFP protein (Fig. 2B) specified in a plasmid and expressed from a heterologous promoter. This approach enabled testing of the hypothesis in a strain retaining a wild-type copy of the essential hns gene in the chromosome.

We determined that the H-NS-GFP chimera is a bona fide Lon substrate because its abundance was 11-fold higher in the lon mutant than in wild-type Salmonella (Fig. 2C), behaving like the full-length H-NS protein (Fig. 1A). By contrast, the PhoP protein is dispensable for degradation of the H-NS-GFP chimera because the abundance of the chimera was similarly low in wild-type and phoP Salmonella (Fig. 2C). These results argue that PhoP reduces H-NS amounts by displacing H-NS from DNA, rendering H-NS a substrate of the Lon protease.

Discussion

We uncovered a singular mechanism that Salmonella uses to overcome silencing of HTGs: the degradation of a xenogeneic silencing protein (Fig. 6). We establish that the Lon protease degrades the xenogeneic silencer H-NS when H-NS is not bound to DNA (Figs. 2A and 3C). Antisilencing DNA-binding proteins, such as PhoP, render H-NS susceptible to proteolysis by Lon by displacing H-NS from DNA (Fig. 2), and to a lower extent, by promoting gene transcription (SI Appendix, Fig. S8), which can dislodge H-NS from DNA (38). The Lon- and PhoP-dependent decrease in H-NS amounts allows expression of HTGs when Salmonella is inside macrophages (Fig. 1B). These results challenge the notion that H-NS autoregulation (42) maintains H-NS amounts within a narrow range (22, 24–26). Moreover, they demonstrate how DNA-binding antisilencing proteins can promote expression of HTGs without actually binding to the corresponding DNAs (Fig. 6 and SI Appendix, Fig. S12).

Fig. 6.

Salmonella expresses horizontally transferred genes by degrading their silencer, H-NS. HTGs are transcriptionally repressed by H-NS, which binds to their regulatory regions. (Left) Under conditions resulting in constant high amounts of H-NS (such as neutral pH conditions or outside macrophages), only foreign genes targeted by antisilencing protein(s) are expressed. H-NS still binds to HTGs not bound by antisilencing proteins and silences their expression. (Right) When Salmonella is inside an acidic macrophage phagosome, DNA-binding antisilencing proteins (e.g., PhoP, SsrB, and SlyA) bind to the promoters of HTGs and displace H-NS from DNA. Lon degrades displaced H-NS. The resulting decrease in H-NS amounts de-represses HTGs, even those not bound by antisilencing proteins.

The number of H-NS molecules per cell is ∼33,000 when Salmonella is grown in Luria-Bertani (LB) or defined media of neutral pH (26, 43). This number is sufficient to cover 14% of the genome (44). We have determined that the number of H-NS molecules per cell is reduced to 2,000 to 3,000 by 6 h inside macrophages. This number is anticipated to cover ∼1% of the Salmonella genome. Thus, in addition to alleviating gene silencing by H-NS, a reduction in H-NS abundance may further expression of HTGs silenced by the nucleoid-associated StpA and Hha proteins, providing an explanation for the pleiotropic phenotypes resulting from a decrease in H-NS abundance (10, 21, 30). This is because H-NS protects its sequalog StpA from degradation by Lon when forming heterodimers with it (45). Likewise, H-NS forms hetero-oligomeric filaments with Hha (4), a sequalog of YmoA, a nucleoid-associated protein from Yersinia pestis degraded by the Lon and ClpXP proteases (46).

Our findings reveal why a mildly acidic pH is critical for Salmonella survival inside macrophages (15) and virulence in mice (47): the need to overcome silencing of foreign virulence genes during infection (Fig. 1B). In support of this notion, H-NS abundance is lower in bacteria grown in a mildly acidic pH than in neutral pH (Fig. 3A). Moreover, phoP and lon single mutants harbored more H-NS than wild-type Salmonella when grown in a mildly acidic pH (Fig. 3A) or harvested from macrophages (Fig. 1A), but the same amount when grown in neutral pH (Fig. 3A) or under a different PhoP-inducing condition (SI Appendix, Fig. S10). Furthermore, phoP and lon mutants are attenuated for virulence (5, 11, 29).

The virulence regulatory protein PhoP is not fully activated inside macrophages until 5 h post internalization (14, 48). This is why wild-type and phoP Salmonella displayed a similar H-NS abundance at 1 h inside macrophages (Fig. 1A) but an eightfold difference at 6 h (Fig. 1A). Although phoP and lon mutants had similar H-NS amounts at the latter time (Fig. 1A), they exhibited differences in the mRNA amounts of HTGs (Fig. 1B). Given that the mRNA amounts of these genes were similar in the phoP and lon mutants when grown in acidic pH (Fig. 3E), PhoP appears to utilize additional mechanisms to de-repress HTGs when inside macrophages.

That Salmonella achieves a larger reduction in H-NS amounts inside macrophages than during growth in mildly acidic pH may result from the phagosome being an oxidizing environment (49), which enhances Lon’s proteolytic activity (50), and/or from Salmonella reducing the amount of LoiA, a transcriptional repressor of the lon gene, when inside macrophages (51).

The virulence regulator PhoP renders H-NS a Lon substrate by displacing H-NS from DNA (Fig. 2). The fact that H-NS is protected from Lon when bound to DNA parallels the decreased susceptibility of TFAM (human mitochondrial transcription factor A) and Abf2 (the primary yeast mitochondrial DNA packaging protein) to the mitochondrial Lon protease when these DNA-binding proteins are bound to DNA (52, 53).

The high conservation in the deduced amino acid sequences of the hns (SI Appendix, Fig. S13) and lon (54) genes suggests that the mechanism reported here for Salmonella likely operates in other bacterial species. In agreement with this notion, expression of the HTG bgl was threefold lower in a lon mutant than in wild-type Escherichia coli (55). Although bgl expression was ascribed to an unknown Lon target operating on H-NS (55), we suggest an alternative explanation: by decreasing H-NS amounts below a certain threshold, Lon promotes bgl expression. Whereas the E. coli experiments were performed in media of neutral pH (55), Salmonella expresses HTGs by Lon-mediated degradation of H-NS specifically in a mildly acidic pH (Fig. 3 A and C). Therefore, E. coli and Salmonella may de-repress HTG under distinct conditions, reflecting their different lifestyles.

We propose that bacteria utilize two general mechanisms to overcome silencing of HTGs by xenogeneic silencers: 1) removal of a silencer from HTGs by DNA-binding antisilencing proteins (4, 19, 20, 56, 57) and 2) reduction in the amounts of the active form of a xenogeneic silencer. The latter mechanism, in turn, can be achieved by proteolysis of the silencer (Fig. 6) or upon overexpression of antisilencing proteins that form nonfunctional heterooligomers with the silencer (58, 59). These two mechanisms differ in that removal of a xenogeneic silencer from a specific foreign gene promotes transcription of that gene, whereas a decrease in silencer abundance impacts expression of multiple foreign genes. However, the former mechanism impacts the latter because silencer displacement from foreign DNA by a DNA-binding antisilencing protein renders the silencer susceptible to proteolysis.

Finally, silencer degradation may also take place in the absence of a DNA binding antisilencing protein under conditions that favor silencer dissociation from DNA, such as high osmolarity (60, 61) or high temperature (61, 62). Silencer degradation allows bacteria to express foreign genes under infection-relevant conditions without the need to evolve DNA sequences for specific DNA-binding antisilencing proteins at each foreign gene.

Materials and Methods

Bacterial Strains, Plasmids, Oligodeoxynucleotides, and Growth Conditions.

Bacterial strains and plasmids used in this study are listed in SI Appendix, Table S1. All S. enterica serovar Typhimurium strains were derived from the wild-type strain 14028s (32) and constructed by phage P22-mediated transductions as described (63). Bacteria were grown at 37 °C in LB broth or N-minimal media (64) supplemented with 0.1% casamino acids, 38 mM glycerol, and the indicated pH (pH 7.6 or 4.9) and 1 mM of MgCl₂ unless specified. E. coli DH5α was used as the host for preparation of plasmid DNA (65). To induce plasmid expression, isopropyl β-d-1-thiogalactopyranoside (IPTG) and L-arabinose were added at the indicated concentrations (0 to 1 mM). When necessary to select for plasmid maintenance, appropriate antibiotics were added at the following final concentrations of ampicillin at 50 µg⋅mL⁻¹, chloramphenicol at 20 µg⋅mL⁻¹, kanamycin at 50 µg⋅mL⁻¹, and tetracycline at 10 µg⋅mL⁻¹. For the protein stability assay, chloramphenicol was used at 1 mg⋅mL⁻¹. DNA oligonucleotides used in this study are listed in SI Appendix, Table S2.

qRT-PCR.

Total RNA was isolated using RNeasy Kit (Qiagen) according to the manufacturer’s instructions. The purified RNA was quantified using a Nanodrop machine (NanoDrop Technologies). Complementary DNA (cDNA) was synthesized using High Capacity RNA-to-cDNA Master Mix (Applied Biosystems). The mRNA amount of the STM14_1977, ycjE, pagC, ssaG, and lon genes was determined by quantification of cDNA using Fast SYBR Green PCR Master Mix (Applied Biosystems) and appropriate primers (STM14_1977: 15973/15974; ycjE: 16675/16676; pagC: 6492/6493; ssaG: 13098/13099; lon: 15854/15855) and monitored using a QuantStudio 6 machine (Applied Biosystems). Data were normalized to the levels of 16S ribosomal RNA amplified with primers 3203 and 3204.

Chromatin Immunoprecipitation.

Bacterial cells were cross-linked with 1% formaldehyde for 15 min at room temperature, quenched with 200 mM glycine for 10 min at room temperature, and washed three times with cold phosphate-buffered saline (PBS). Then cells were lysed in cell lysis solution A (10 mM Tris, pH 8.0, 50 mM NaCl, 10 mM EDTA, 20% sucrose, 10 mg⋅mL⁻¹ lysozyme) and 10× RIPA solution (Millipore). DNA was fragmented to an average size of 500 bp by sonication (VirTis) and a 50-µL aliquot was taken as “input DNA.” Immunoprecipitation of PhoP-DNA and H-NS-DNA complexes were performed using antibodies recognizing Influenza hemagglutinin (HA) and FLAG (DYKDDDDK) respectively, and using MagnaChip protein A/G magnetic beads (Millipore). Samples were then washed twice with 1× RIPA solution, twice with LiCl immune complex wash buffer (Millipore), twice with TE buffer (20 mM Tris [pH 8.0], 1 mM EDTA) and eluted in elution buffer (50 mM Tris (pH 8.0), 10 mM EDTA, 1% sodium dodecyl sulfate [SDS]) with incubation at 65 °C for 15 min. Both immunoprecipitated (IP) and input DNA samples were incubated at 65 °C for 9 h to reverse cross-links, purified using a Qiagen PCR purification column, and quantified using qRT-PCR using primers (STM14_1977: 15973/15974; ycjE: 16675/16676; pagC: 6964/6965; rpoD: 4149/4150). Binding of PhoP and H-NS proteins to DNA were calculated as a ratio of “IP DNA/input DNA” and normalized to that of the rpoD gene.

Western Blot Analysis.

Bacterial cells were grown as described and crude extracts were prepared in B-PER bacterial protein extraction reagent (Pierce) with 100 µg⋅mL⁻¹ lysozyme and EDTA-free protease inhibitor (Roche). Samples were separated in 4 to 12% NuPAGE gels (Life Technologies). Then samples were analyzed by Western blotting using antibodies recognizing FLAG (Sigma; 1:2,000), Lon (Biorbyt; 1:5,000), GFP (Invitrogen, 1:3,000), GroEL (Abcam; 1:5,000), or AtpB (Abcam; 1:5,000). Secondary horseradish-peroxidase–conjugated antisera recognizing rabbit or mouse antibodies (GE Healthcare) were used at 1:5,000 dilution. The blots were developed with the Amersham ECL Western Blotting Detection Reagents (GE Healthcare) or SuperSignal West Femto Chemiluminescent system (Pierce) and were visualized using LAS-4000 (Fuji Film). The density of protein bands was determined by quantification using ImageJ software version 1.48 (NIH).

Determination of Salmonella Protein Amounts Inside Macrophages.

The murine-derived macrophage cell line J774A.1 was cultured in Dulbecco modified Eagle medium (DMEM; Life Technologies) supplemented with 10% fetal bovine serum (Life Technologies) at 37 °C under 5% CO₂. Macrophages were seeded in 12-well tissue culture plates with 10⁶ cells per well on the day before infection with Salmonella. Confluent monolayers were inoculated with bacterial cells that had been grown overnight in LB broth, washed with PBS, and resuspended in 0.1 mL of prewarmed DMEM at a multiplicity of infection of 10. Following a 30-min incubation, the wells were washed three times with prewarmed PBS to remove extracellular bacteria and then incubated with prewarmed medium supplemented with 100 mg⋅mL⁻¹ gentamicin for 1 h to kill extracellular bacteria. Next, the wells were washed three times with PBS and then incubated with prewarmed medium supplemented with 10 mg⋅mL⁻¹ gentamicin. At the desired times, cells were washed with PBS and lysed with 1 mL of cell lysis solution B (0.1% [wt/vol] SDS, 1% [vol/vol] acidic phenol and 19% [vol/vol] ethanol in double-distilled water) for 30 min. The cell lysates from two plates were pooled and centrifuged at 5,000 × g for 20 min. Pellets were washed twice with PBS and resuspended in 100 µL of 100 mM NH₄ HCO₃, pH 8.4. The recovered Salmonella cells were subsequently analyzed by Western blot.

Determination of Salmonella mRNA Abundance Inside Macrophages.

J774A.1 macrophages were cultured and infected as described above. At the desired times, samples were harvested using TRIzol reagent (Invitrogen) solution. Total RNA was isolated, and cDNA was synthesized as described above. mRNA levels were measured by qRT-PCR as described above.

In Vivo Protein Degradation Assay.

Bacterial cells were grown as described, and protein synthesis was blocked by addition of 1 mg⋅mL⁻¹ chloramphenicol. Samples were taken at the times denoted in Fig. 3C, SI Appendix Fig. S6, and SI Appendix Fig. S8A and analyzed by Western blot. When necessary, 0.25 mg⋅mL⁻¹ rifampicin was added to stop transcription.

In Vitro Protein Degradation Assay.

Protein degradation assays were performed in a buffer containing 25 mM Tris (pH 8.0), 100 mM KCl, 12 mM MgCl₂, 2 mM ATP, 4 mM phosphoenol pyruvate (Roche), and 20 µg⋅mL⁻¹ pyruvate kinase (Sigma) at 30 °C. Purified H-NS, PhoP (wild-type or V193M variant), and Lon were used at 5, 10, and 1 µM, respectively. β-Casein (Sigma, C6905) was used at 5 µM. When denoted, plasmid DNA harboring a PhoP-activated foreign gene targeted by H-NS was added to the reaction at a final concentration of 20 nM, and H-NS and PhoP proteins were incubated with DNA for 5 min before adding other components. Samples were taken at desired time points, quenched with SDS loading dye, and frozen in dry ice. Samples were separated in 4 to 12% NuPAGE gels (Life Technologies), and proteins were detected by Coomassie blue G-250 staining. In agreement with results obtained with the purified E. coli Lon protein (66), the Salmonella Lon protein degrades H-NS in vitro in buffers at pH 7.5 and 8.0 but not when the pH is 6.5 (SI Appendix, Fig. S14).

Statistical Analyses.

Sample sizes (biological replicates) for each experimental group or condition are described each figure legend. For comparisons of two groups, t tests were applied. Two-sided analysis provides P values for each comparison.

Other Experimental Procedures.

Detailed descriptions of constructing strains and plasmids and determination of Salmonella number inside macrophages, purification of proteins, electrophoretic mobility shift assay, and amino acid sequence comparisons are in SI Appendix, SI Materials and Methods.

Data Availability.

All data original to this article are available in Mendeley; V1, doi: 10.17632/m4vt7hwrgc.1.