An Essential Role for DNA Adenine Methylation in Bacterial Virulence

See allHide authors and affiliations

Science  07 May 1999:
Vol. 284, Issue 5416, pp. 967-970
DOI: 10.1126/science.284.5416.967


Salmonella typhimurium lacking DNA adenine methylase (Dam) were fully proficient in colonization of mucosal sites but showed severe defects in colonization of deeper tissue sites. These Dam mutants were totally avirulent and were effective as live vaccines against murine typhoid fever. Dam regulated the expression of at least 20 genes known to be induced during infection; a subset of these genes are among those activated by the PhoP global virulence regulator. PhoP, in turn, affected Dam methylation at specific genomic sites, as evidenced by alterations in DNA methylation patterns. Dam inhibitors are likely to have broad antimicrobial action, and Dam derivatives of these pathogens may serve as live attenuated vaccines.

Methylation at adenine residues by Dam controls the timing and targeting of important biological processes such as DNA replication, methyl-directed mismatch repair, and transposition (1). In addition, Dam regulates the expression of operons such as pyelonephritis-associated pili (pap), which are an important virulence determinant in upper urinary tract infections (2, 3). The latter regulatory mechanism involves formation of heritable DNA methylation patterns, which control gene expression by modulating the binding of regulatory proteins. Although Dam regulates pili gene expression, its role in microbial pathogenesis has never been tested.

To determine whether Dam plays a role in the pathogenesis ofSalmonella typhimurium, we assessed the effect of an insertion in the dam gene (Mud-Cm). The oral lethal dose required to kill 50% of the animals (LD50) of this Dam mutant was more than 10,000 times the LD50 for the wild type, and the intraperitoneal LD50 for the mutant was more than 1000 times that for the wild type (Table 1). Because thedam insertion could decrease the expression of downstream genes (polar effects), an in-frame, nonpolar dam deletion was constructed (4) and shown to have the same reduced virulence as the dam insertion. Thus, the attenuation was specifically attributable to the lack of Dam. Moreover, intraperitoneal inoculation of mice with a mixture of equal numbers of Dam+ and Dam Salmonella showed that Dam mutants were completely eliminated during growth in the mouse (competitive index assay). Similar results were obtained with a strain that overproduces Dam from a recombinant plasmid, which suggested that precise amounts of the Dam methylase are required for full virulence. These results show that the Dam methylase is essential for bacterial pathogenesis.

Table 1

Dam is required for Salmonella virulence. ND, not determined.

View this table:

Dam plays an essential role in methyl-directed mismatch repair (MDMR) because it allows discrimination between parental and daughter DNA strands (1). Thus, in the absence of Dam, bacteria show an increased mutation rate. To test the hypothesis that the reduction in virulence of Dam Salmonella was due to a high mutation rate, we measured the virulence of mutS Salmonella, which lack MDMR and also have a high mutation rate. Table 1 shows that in both the oral LD50 and the competitive index virulence assays, mutS Salmonella were identical to the wild type, indicating that Dam does not affect pathogenesis via an increased mutation rate. Because more DNA exchange between species occurs in MutS strains than in MutS+ strains, they more readily acquire new virulence determinants (1). The fact that MutS strains are fully virulent could explain the high frequency at which mutS Escherichia coli andSalmonella mutants are found among clinical isolates (5).

Dam controls the expression of Pap pili by modulating the binding of leucine-responsive regulatory protein (Lrp) to papregulatory DNA sequences (3). Lrp is a global regulator of at least 35 genes in E. coli that include operons involved in metabolism, transport, and adhesion (6). To determine whether Dam affects Salmonella virulence through an Lrp-mediated pathway, we analyzed Lrp Salmonella (Table 1). Salmonella lacking Lrp were fully virulent, as assessed by the LD50 and competitive index assays. These data show that Lrp is not required for virulence in a mouse model of typhoid fever.

The results discussed above show that adenine methylation is critical for Salmonella pathogenesis. DNA methylation of cytosine residues appears to be important for the regulation of biological processes in both plants and animals. Although Salmonellacontain a DNA cytosine methylase (Dcm), the role of cytosine methylation in this organism is unclear. Thedcm mutant was virulent in the LD50 and competitive index assays (Table 1). These results demonstrate that methylation of adenine but not cytosine residues is required for Salmonella pathogenesis.

DNA adenine methylation has been shown to directly control virulence gene expression (7). Therefore, we determined whether Dam regulates Salmonella genes that are preferentially expressed in the mouse [designated as in vivo–induced (ivi) genes (8–11)]. Dam significantly repressed the expression of more than 20 ivi genes (by a factor of 2 to 18) when grown in rich medium (Fig. 1). Four of the eight fusions in Fig. 1 are in known genes, all of which have been shown to be involved or have been implicated in virulence:spvB resides on the Salmonella virulence plasmid and functions to facilitate growth at systemic sites of infection (12); pmrB is involved in resistance to antibacterial peptides termed defensins (13); andmgtA and entF are involved in the transport of magnesium and iron, respectively (14, 15). Additional ivi genes of unknown function were also Dam-regulated. These results indicate that Dam is a global regulator ofSalmonella gene expression and that thedam-regulated ivi genes constitute adam regulon (1).

Figure 1

Dam regulates in vivo induced genes. β-galactosidase expression from S. typhimurium ivi fusions in Dam+ and Dam strains, grown to saturation in LB medium as described (27), was measured. The vertical axis shows β-galactosidase activities [micromoles ofonitrophenol (ONP) formed per minute perA 600 unit per milliliter of cell suspension × 103]. The β-galactosidase activities were assayed as described (28).

Salmonella pathogenesis is known to be controlled by PhoP, a DNA binding protein that acts as both an inducer and repressor of specific virulence genes [reviewed in (16)]. To determine whether the Dam and PhoP regulatory pathways share common genes, we tested the effect of Dam on seven PhoP-activatedivi genes, including spvB, pmrB, andmgtA. Dam repressed the expression of these three genes by a factor of 2 to 19 (Fig. 2), and this repression was not dependent on the PhoP protein. Dam did not significantly affect the expression of the remaining four PhoP-activated genes (17). These results indicate that Dam and PhoP constitute an overlapping global regulatory network controllingSalmonella virulence.

Figure 2

Dam represses PhoP-activated genes. β-galactosidase expression from S. typhimurium ivi fusion strains, grown in minimal medium (pH 5.5, 50 μM Mg2+) as described (27), was measured. The vertical axis shows β-galactosidase activities (calculated as in Fig. 1). The β-galactosidase activities were assayed as described (28).

Binding of regulatory proteins to DNA can form DNA methylation patterns by blocking the methylation of specific Dam target sites (GATC sequences) (18). Therefore, we further investigated the interactions between Dam and PhoP by determining whether the binding of PhoP (or a PhoP-regulated protein) to specific DNA sites blocks methylation of these sites by Dam, resulting in an alteration in the DNA methylation pattern. Analysis of PhoP+ and PhoP Salmonella showed distinct differences in DNA methylation patterns. Digestion of genomic DNA from PhoP bacteria with Mbo I (which cleaves only at nonmethylated GATC sites) resulted in the appearance of DNA fragments that were not present in DNA from PhoP+ bacteria, indicating that the PhoP protein (or a PhoP-regulated gene product) blocks Dam methylation at specific GATC-containing sites in theSalmonella genome (Fig. 3, arrows). Recent data have shown that although catabolite gene activator protein binds to a DNA sequence containing GATC, it does not protect this site from methylation (18). Thus, not every protein that binds to a Dam target site protects the GATC sequence from methylation. It is also possible that PhoP+ and PhoP strains have different amounts of Dam activity, which in turn could affect DNA methylation patterns. However, this regulation does not occur at the transcriptional level because Dam does not alter PhoP expression, nor does PhoP alter Dam expression (17). Further analysis will determine whether these PhoP-protected sites are within regulatory regions of virulence genes, and whether DNA methylation directly affects the PhoP regulon by altering DNA-PhoP interactions.

Figure 3

PhoP affects the formation ofSalmonella DNA methylation patterns. DNA methylation patterns formed in PhoP+ and from PhoPstrains in minimal medium. Genomic DNA prepared from PhoP+and from PhoP strains embedded in agarose was cleaved with Mbo I (which cleaves nonmethylated Dam-target sites) and subjected to pulsed-field gel electrophoresis (29). The arrows indicate two DNA fragments that were present in PhoP Salmonella but were absent in PhoP+ Salmonella.

In E. coli, almost all GATC sites protected from methylation are in 5′ noncoding DNA regions presumably involved in the control of gene expression (19, 20). Thus, it is likely that the DNA methylation patterns identified in Salmonella (Fig. 3) are also within gene regulatory regions. Methylation of specific GATC sites in the regulatory regions of virulence genes could affect the binding of regulatory proteins to DNA. Such altered protein-DNA interactions can affect gene expression, as has been shown for the papvirulence operon in E. coli (7, 18). Similarly, Dam methylation could directly or indirectly affect the expression of PhoPQ-regulated genes in S. typhimurium.

Because Dam mutants were highly attenuated, we determined whether Dam Salmonella could serve as a live attenuated vaccine. Table 2 shows that all (17/17) mice immunized with a S. typhimuriumDam insertion strain survived a wild-type challenge of 104 above the LD50, whereas all nonimmunized mice (12/12) died after challenge. Moreover, because all (8/8) mice immunized with Salmonella containing the damdeletion survived challenge, these data indicate that protection was specifically due to the absence of Dam methylase. Preliminary experiments indicate that mice immunized with Dam S. typhimurium showed cross-protection against another pathogenic strain of Salmonella (17). The virulence attenuation and effectiveness of Dam mutants as a vaccine (Tables 1 and 2) could be due to the ectopic expression of virulence determinants (Figs. 1 and 2), which would likely be deleterious to the growth (or survival) of Salmonella during infection.

Table 2

Dam Salmonella serve as effective live attenuated vaccines.

View this table:

Dam Salmonella could have been avirulent as a result of multiple defects in basic cellular processes that reduced viability. This hypothesis was tested by comparing the survival of Dam+ and Dam Salmonella in mouse tissues. As shown in Fig. 4, Dam bacteria were fully proficient in colonization of a mucosal site (Peyer's patches) but showed severe defects in colonization of deeper tissue sites. Five days after infection, we observed a reduction of three orders of magnitude in numbers of Dam Salmonella in the mesenteric lymph nodes (relative to numbers of Dam+ bacteria) and a reduction of eight orders of magnitude in numbers of Dam Salmonella in the liver and spleen. These data show that Dam Salmonella survive in Peyer's patches of the mouse small intestine for at least 5 days, providing an opportunity for elicitation of a host immune response. Dam Salmonella, however, were unable to cause disease; they either were unable to invade systemic tissues or were able to invade but could not survive.

Figure 4

Colonization of mouse tissue sites by Dam Salmonella. BALB/c mice were infected via gastrointubation at a dose of 109 Dam+ (open boxes) or Dam (closed boxes) S. typhimurium. After 1 day or 5 days after infection, mice were killed and bacteria were recovered from the host tissues indicated. PP, Peyer's patches (the four patches proximal to the ileal-cecal junction); MLN, mesenteric lymph nodes; CFU, colony-forming units.

DNA adenine methylases are potentially excellent targets for both vaccines and antimicrobials. They are highly conserved in many pathogenic bacteria that cause significant morbidity and mortality, such as Vibrio cholerae (21), Salmonella typhi (22), pathogenic E. coli(23), Yersinia pestis (22),Haemophilus influenzae (24), andTreponema pallidum (25). In addition, because Dam is a global regulator of genes expressed during infection (Fig. 1), Dam mutants may ectopically express multiple immunogens that are processed and presented to the immune system. Such ectopic expression could elicit a cross-protective immune response between related bacterial strains that share common epitopes. Finally, because the Dam methylase is essential for bacterial virulence, Dam inhibitors are likely to have broad antimicrobial action, hence Dam is a promising target for antimicrobial drug development.


View Abstract

Stay Connected to Science

Navigate This Article