Cooperation of the multidrug efflux pump and lipopolysaccharides in the intrinsic antibiotic resistance of Salmonella enterica serovar Typhimurium.

OBJECTIVES: In Gram-negative bacteria, drug susceptibility is associated with multidrug efflux systems and an outer membrane (OM) barrier. Previous studies revealed that Salmonella enterica serovar Typhimurium has 10 functional drug efflux pumps. Among them, AcrB is a major factor to maintain the intrinsic drug resistance in this organism. The lipopolysaccharide (LPS) content of OM is also important for resistance to lipophilic drugs; however, the interplay between the multidrug efflux pump and LPS in the intrinsic antibiotic resistance of Salmonella remains to be studied in detail. The aim of this study was to investigate the relationship between AcrB and LPS in the intrinsic drug resistance of this organism.
METHODS: The genes encoding LPS core biosynthetic proteins and AcrB were disrupted from the wild-type S. enterica strain ATCC 14028s. The plasmid carrying acrB was transformed into these mutants and then the drug susceptibilities of the mutants and transformants were determined.
RESULTS: Our results showed that the levels of Salmonella intrinsic antibiotic resistance were decreased when the length and branches of core oligosaccharide were lost. Furthermore, the deletion of acrB reduced multidrug resistance of all LPS mutants and AcrB production from the plasmid complemented this phenotype. However, AcrB production could not completely compensate for LPS function in intrinsic resistance.
CONCLUSIONS: Both pump inactivation and shortened LPS enhanced drug susceptibility, although the maximum susceptibility was achieved when the two were combined. Hence, these results indicated that the multidrug efflux system and OM barrier are both essential for maintaining intrinsic antibiotic resistance in Salmonella.

Introduction

Multidrug efflux pumps cause serious problems in cancer chemotherapy and in the treatment of bacterial infections. In bacteria, drug resistance is often associated with multidrug efflux pumps, which decrease cellular drug accumulation.[1,2] In Gram-negative bacteria, pumps belonging to the resistance–nodulation–division family are particularly effective in generating resistance, because they form a tripartite complex with the periplasmic proteins of the membrane fusion protein family and the outer membrane (OM) channels, ensuring that drugs are pumped out directly to the external medium.[3] High-level fluoroquinolone resistance in Salmonella enterica serovar Typhimurium phage type DT204 has been shown to be primarily due to multiple target gene mutations and active efflux by the AcrAB-TolC efflux system belonging to the resistance–nodulation–division family.[4]

S. enterica is a pathogen that causes a variety of diseases in humans ranging from gastroenteritis to bacteraemia and typhoid fever. Previous studies have shown that S. enterica serovar Typhimurium has 10 functional drug efflux pumps: AcrAB, AcrD, AcrEF, MdtABC, MdsAB, EmrAB, MdfA, MdtK, MacAB and SmvA.[5,6] Among these, AcrAB is constitutively expressed and is the most effective in intrinsic drug resistance in Salmonella.

In addition to drug efflux pumps, OM is also important for intrinsic antibiotic resistance. Gram-negative bacteria, which have an OM barrier, are usually much more resistant than Gram-positive bacteria to a wide range of drugs.[7] In particular, lipopolysaccharides (LPS), located exclusively in the outer leaflet of OM, prevent the easy entry of lipophilic agents.[8] The LPS molecule comprises three parts: lipid A, core oligosaccharides and the O-antigen (Figure 1). Lipid A anchors the LPS molecule into the bacterial OM. The core oligosaccharides and O-antigen are located in the outer domain of the LPS molecule (Figure 1). LPS is only found in the OM of Gram-negative bacteria and many genes required for its synthesis and modification have been identified.[8]

Figure 1.

LPS in S. enterica serovar Typhimurium. Genes encoding LPS biosynthetic proteins are listed for each synthetic route. This figure has been modified from EcoSal with permission.[21]

LPS is important for intrinsic antibiotic resistance[9-11] and previous studies have shown that the AcrB efflux pump is related to both the intrinsic and the acquired multidrug resistance of Salmonella.[4,5,12] In Francisella sp., another Gram-negative bacterium, it has been reported that both the LPS and the AcrAB efflux pump system play a role in azithromycin susceptibility.[13] However, the synergistic interplay between AcrB and LPS of Salmonella remains to be studied in detail. In the present report, we examined the interplay between the AcrB efflux pump and LPS by determining the drug susceptibilities of mutants with varying LPS lengths and by investigating the effect of the acrB deletion in LPS mutants.

Materials and methods

Bacterial strains, plasmids and growth conditions

The bacterial strains and plasmids used in this study are listed in Table S1 (available as Supplementary data at . The S. enterica serovar Typhimurium strains were derived from the wild-type strain ATCC 14028s.[14] Bacterial strains were grown at 37°C in lysogeny broth (1.0% tryptone, 0.5% yeast extract, 1.0% NaCl).[15]

Construction of gene deletion mutants

To construct all mutants, gene disruption was performed as described by Datsenko and Wanner.[16] Primers used for the construction of the mutants are listed in Table S2 (available as Supplementary data at . The chloramphenicol resistance gene cat or the kanamycin resistance gene aph, flanked by Flp recognition sites, was PCR amplified and the resulting products were used to transform the recipient ATCC 14028s strain harbouring plasmid pKD46, which expresses Red recombinase. The chromosomal structure of the mutated loci was verified by PCR, as described previously.[16] Both cat and aph were eliminated using plasmid pCP20, as previously described.[16]

Plasmid construction

acrB was PCR amplified from ATCC 14028s genomic DNA with primers acrB-F and acrB-R (Table S2, available as Supplementary data at ) which introduced restriction enzyme sites of BamHI and XbaI at both ends of the amplified fragments. The PCR fragments were cloned into the corresponding sites of the vector pTrcHis2B (Invitrogen) to produce the plasmid pacrB.

LPS analysis

LPS was purified as described previously.[17] Culture samples were adjusted to an optical density of 1.0 at 600 nm in a final volume of 100 μL, and LPS samples normalized to the number of cells were separated on 12% acrylamide gels using Tris–Glycine/SDS buffer systems and stained using a modification of the conventional silver staining method.[18]

Determination of MICs of toxic compounds

The antibacterial activities of various agents were determined on lysogeny broth agar plates containing oxacillin, cloxacillin, nafcillin, erythromycin, rhodamine 6G, crystal violet, ethidium bromide, novobiocin, benzalkonium chloride, SDS or deoxycholic acid (Sigma, St Louis, MO, USA) at various concentrations. Agar plates were prepared by the 2-fold agar dilution technique, as described previously.[19] To determine MICs, bacteria were grown in lysogeny broth at 37°C overnight, diluted with the same medium and then tested at a final inoculum of 105 cfu/μL using a multipoint inoculator (Sakuma Seisakusyo, Tokyo, Japan) after incubation at 37°C for 20 h. The MIC was the lowest compound concentration to inhibit cellular growth.

Results and discussion

Effects of the length and branches of LPS core oligosaccharides on intrinsic antibiotic resistance

To investigate whether the length and branches of LPS core oligosaccharides affect intrinsic antibiotic resistance in S. enterica serovar Typhimurium, genes encoding LPS core biosynthetic proteins (Figure 1) were deleted (Table S1, available as Supplementary data at ). In addition to the mutants of the genes presented in Figure 1, the deletion mutants of the wzz gene, encoding the O-chain length determinant, were constructed (Table S1, available as Supplementary data at ). The deletion mutant of the rfaY gene, which is necessary for phosphorylating the Hep(II) heptose in the core region of the LPS, was also constructed (Table S1, available as Supplementary data at ). To confirm the effects of deletions of these genes on the LPS structure, we analysed the LPS profiles in silver-stained polyacrylamide gels. LPS profiles of the deletion mutants were all different from that of the wild-type strain (Figure 2). For the MIC measurement, the AcrB substrates were chosen to compare the effect of deletion of the genes involved in the LPS biosynthesis with the effect of deletion of the acrB gene on drug susceptibilities. Compared with the wild-type strain, the ΔrfaK, Δwzz and ΔrfaJ strains maintained intrinsic resistance to all antimicrobial agents and chemical compounds tested; however, the ΔrfaI, ΔrfaB, ΔrfaY, ΔrfaP, ΔrfaG, ΔrfaF and ΔrfaC strains showed increased susceptibility to almost all drugs as the length and branches of LPS core oligosaccharides were lost (Table 1). Interestingly, deletion of rfaB, which encodes a protein that adds a galactose moiety to produce one branch of the LPS core oligosaccharide, had no impact on novobiocin resistance; however, the strains that lost a core oligosaccharide phosphorylation gene (i.e. rfaY or rfaP) were more susceptible to novobiocin than ΔrfaC. The electric charge produced by the phosphate group seems to be effective in inhibiting the entry of aminocoumarin antibiotics. These data indicate that the length and branches of LPS core oligosaccharides play a role in the maintenance of intrinsic resistance of S. enterica against multiple drugs.

Figure 2.

SDS-PAGE analysis of LPS. LPSs were isolated from the wild-type strain (ATCC 14028s), ΔrfaI (NKS363), ΔrfaB (NKS365), ΔrfaC (NKS366), ΔrfaF (NKS367), ΔrfaG (NKS368), ΔrfaJ (NKS369), ΔrfaP (NKS371), ΔrfaY (NKS372), ΔrfaK (NKS375) and Δwzz (NKS877) strains.

Table 1.

Susceptibility of S. enterica serovar Typhimurium acrB and/or LPS mutants to toxic compounds

Strain	MIC (mg/L)
	OXA	CLO	NAF	ERY	R6G	CV	EB	NOV	BENZ	SDS	DOC
Wild-type	1024	1024	2048	512	4096	256	>2048	512	512	>32 768	>32 768
ΔacrB	4	4	16	8	16	4	128	4	8	1024	>32 768
ΔrfaK	1024	1024	2048	512	4096	256	>2048	512	512	>32 768	>32 768
ΔrfaKΔacrB	4	4	16	8	16	4	128	16	8	128	>32 768
Δwzz	1024	1024	2048	512	4096	256	>2048	512	512	>32 768	>32 768
ΔwzzΔacrB	4	4	16	8	16	4	128	8	8	2048	>32 768
ΔrfaJ	1024	1024	2048	512	4096	256	>2048	512	512	>32 768	>32 768
ΔrfaJΔacrB	4	4	16	8	16	4	128	16	8	128	>32 768
ΔrfaI	512	512	1024	512	4096	64	>2048	64	32	>32 768	>32 768
ΔrfaIΔacrB	4	4	16	4	16	4	64	4	4	16 384	>32 768
ΔrfaB	512	512	1024	512	4096	64	>2048	512	32	>32 768	>32 768
ΔrfaBΔacrB	4	4	16	4	16	4	64	8	4	64	32 768
ΔrfaY	512	512	1024	256	4096	32	>2048	16	64	>32 768	>32 768
ΔrfaYΔacrB	2	4	8	4	8	2	64	0.5	4	128	8192
ΔrfaP	256	256	512	128	4096	8	2048	4	8	2048	>32 768
ΔrfaPΔacrB	2	2	8	2	8	1	32	0.125	4	32	512
ΔrfaG	256	256	256	64	4096	8	2048	32	4	256	>32 768
ΔrfaGΔacrB	1	2	4	1	8	1	32	0.5	2	64	512
ΔrfaF	128	128	256	16	1024	8	2048	32	8	256	>32 768
ΔrfaFΔacrB	4	8	32	8	16	4	128	16	8	256	>32 768
ΔrfaC	128	64	128	16	128	4	1024	32	4	128	2048
ΔrfaCΔacrB	1	1	2	< 0.5	2	< 0.25	32	0.5	2	16	128

OXA, oxacillin; CLO, cloxacillin; NAF, nafcillin; ERY, erythromycin; R6G, rhodamine 6G; CV, crystal violet; EB, ethidium bromide; NOV, novobiocin; BENZ, benzalkonium chloride; DOC, deoxycholic acid.

MIC determinations were repeated at least three times.

Effect of acrB deletion on multidrug resistance of the LPS mutants

In Salmonella, the AcrAB-TolC efflux system is constitutively expressed and effective in intrinsic drug resistance.[5] To investigate the function of multidrug efflux systems in LPS mutants, we disrupted acrB from the genomic DNA (Table S1, available as Supplementary data at ). The ΔacrB strain was more susceptible to oxacillin, cloxacillin, nafcillin, erythromycin, rhodamine 6G and ethidium bromide than any other single LPS mutant and almost all drug susceptibilities of the ΔrfaKΔacrB, ΔwzzΔacrB, ΔrfaJΔacrB, ΔrfaIΔacrB and ΔrfaBΔacrB double mutants were comparable to that of ΔacrB. The ΔrfaFΔacrB strain was more susceptible to SDS than ΔacrB. The ΔrfaYΔacrB strain was more susceptible to novobiocin, SDS and deoxycholic acid than ΔacrB. The ΔrfaPΔacrB was more susceptible to erythromycin, crystal violet, ethidium bromide, novobiocin, SDS and deoxycholic acid than ΔacrB. The ΔrfaGΔacrB and ΔrfaCΔacrB strains were more susceptible to almost all drugs than the ΔacrB strain (Table 1). These results indicate that AcrAB-TolC plays a role in drug resistance even if the LPS function is weakened.

Effect of acrB overexpression on drug susceptibilities of the LPS mutants

To investigate the effect of acrB overexpression on drug susceptibilities of the LPS mutants, acrB in S. enterica ATCC 14028s was cloned into the vector pTrcHis2B and then the constructed plasmid was transformed into LPS mutants lacking acrB. Overexpression of acrB conferred multidrug resistance to all of the mutants (Table 2). These data indicate that the AcrB efflux pump can function in bacteria with imperfect LPS. However, overexpressed acrB did not completely restore multidrug resistance of some LPS mutants to the wild-type level, e.g. ΔrfaFΔacrB and ΔrfaCΔacrB (Table 2). These results indicate that the multidrug efflux system cannot recover the loss of LPS required for maintenance of intrinsic Salmonella resistance.

Table 2.

Susceptibility of S. enterica serovar Typhimurium strains to toxic compounds

Strain	MIC (mg/L)
	ERY	R6G	CV	EB	NOV	BENZ	SDS	DOC
Wild-type	512	4096	256	>2048	256	512	>32 768	>32 768
ΔacrB/pTrcHis2B	4	8	2	32	4	4	256	32 786
ΔacrB/pacrB	256	4096	128	2048	256	64	>32 786	>32 786
ΔrfaKΔacrB/pTrcHis2B	4	16	2	64	8	4	128	>32 786
ΔrfaKΔacrB/pacrB	256	4096	128	4096	512	128	>32 786	>32 786
ΔwzzΔacrB/pTrcHis2B	4	8	2	32	8	4	256	>32 786
ΔwzzΔacrB/pacrB	128	4096	128	4096	512	128	>32 786	>32 786
ΔrfaJΔacrB/pTrcHis2B	4	16	2	64	8	4	128	32 786
ΔrfaJΔacrB/pacrB	256	4096	128	4096	512	64	>32 786	>32 786
ΔrfaIΔacrB/pTrcHis2B	4	8	2	32	4	4	256	>32 786
ΔrfaIΔacrB/pacrB	128	4096	128	4096	128	64	>32 786	>32 786
ΔrfaBΔacrB/pTrcHis2B	4	16	1	64	4	4	64	32 786
ΔrfaBΔacrB/pacrB	128	4096	32	4096	256	16	>32 786	>32 786
ΔrfaYΔacrB/pTrcHis2B	2	4	1	32	1	4	128	4096
ΔrfaYΔacrB/pacrB	128	4096	64	2048	32	32	>32 786	>32 786
ΔrfaPΔacrB/pTrcHis2B	1	4	1	32	0.25	4	32	512
ΔrfaPΔacrB/pacrB	64	2048	16	2048	8	8	256	>32 786
ΔrfaGΔacrB/pTrcHis2B	1	8	1	32	4	2	32	256
ΔrfaGΔacrB/pacrB	32	2048	8	1024	16	4	128	>32 786
ΔrfaFΔacrB/pTrcHis2B	4	16	2	64	16	4	128	>32 786
ΔrfaFΔacrB/pacrB	16	256	4	512	16	4	128	>32 786
ΔrfaCΔacrB/pTrcHis2B	0.25	2	0.5	64	0.5	4	16	128
ΔrfaCΔacrB/pacrB	8	64	2	1024	8	4	128	2048

ERY, erythromycin; R6G, rhodamine 6G; CV, crystal violet; EB, ethidium bromide; NOV, novobiocin; BENZ, benzalkonium chloride; DOC, deoxycholic acid.

MIC determinations were repeated at least three times.

Values in bold are larger than those of the corresponding strains harbouring the vector only.

Concluding remarks

Herein, we investigated the interplay between the multidrug efflux system and the OM barrier in intrinsic Salmonella antibiotic resistance at the genetic level. The results showed that the length and branches of LPS core oligosaccharides and the AcrB efflux pump are necessary for the maintenance of the intrinsic resistance of S. enterica serovar Typhimurium. The maximal susceptibility was achieved when deletions of acrB and genes related to LPS synthesis were combined. Additive synergistic effects were especially observed in the ΔrfaGΔacrB and ΔrfaCΔacrB mutants. Compared with ΔacrB, ΔrfaGΔacrB was susceptible to oxacillin (4-fold), nafcillin (4-fold), erythromycin (8-fold), crystal violet (4-fold), ethidium bromide (4-fold), novobiocin (8-fold), benzalkonium chloride (4-fold), SDS (16-fold) and deoxycholic acid (>64-fold). ΔrfaCΔacrB was also susceptible to oxacillin (4-fold), cloxacillin (4-fold), nafcillin (8-fold), erythromycin (>16-fold), rhodamine 6G (8-fold), crystal violet (>16-fold), ethidium bromide (4-fold), novobiocin (8-fold), benzalkonium chloride (4-fold), SDS (64-fold) and deoxycholic acid (>256-fold) when compared with ΔacrB. These data indicate that the AcrAB-TolC efflux system is important for maintaining the intrinsic antibiotic resistance even when most of the core region of LPS is lost. The overexpression of acrB cannot completely compensate the function of LPS in the maintenance of intrinsic resistance, although functional AcrB was present in all of the LPS mutants. Interestingly, Giraud et al.[20] reported that there was an increased density of the long O-polysaccharide chains and an increased level of the AcrAB pump in in vitro-selected quinolone-resistant mutants of S. enterica serovar Typhimurium, suggesting that both efflux pump and LPS are also important for the acquired resistance. Our results support the notion that both pump inhibition and OM disruption could constitute an effective approach to increasing the drug susceptibility of multidrug-resistant strains.[9,10] In summary, we genetically determined that the AcrB multidrug efflux pump and bulkiness of LPS core oligosaccharides are essential for intrinsic antibiotic resistance in S. enterica.

Funding

This study was supported by a research grant from the Uehara Memorial Foundation (to K. N.), the Institute for Fermentation (to M. H.-N.), Grants-in-Aid for Young Scientists and Bilateral Joint Research Projects from the Japan Society for the Promotion of Science (to M. H.-N.) and a grant NEXT Program (LS080) from the Cabinet Office, Government of Japan (to K. N.). S. Y. was supported by a research fellowship from the Japan Society for the Promotion of Science for Young Scientists.

Transparency declarations

None to declare.

Supplementary data

Tables S1 and S2 are available as Supplementary data at