Alegado Cell Microbiol 2008

8/6/2019 Alegado Cell Microbiol 2008

1/15

Resistance to antimicrobial peptides contributes topersistence of Salmonella typhimurium in theC. elegans intestine

Rosanna A. Alegado1 and Man-Wah Tan1,2*

Departments of 1Microbiology and Immunology and2Genetics, School of Medicine, Stanford University,

Stanford, CA 94305, USA.

Summary

The human pathogen Salmonella typhimurium can

colonize, proliferate and persist in the intestine

causing enteritis in mammals and mortality in the

nematode Caenorhabditis elegans. Using C. elegans

as a model, we determined that the Salmonellapatho-

genicity islands-1 and -2 (SPI-1 and SPI-2), PhoP and

the virulence plasmid are required for the establish-

ment of a persistent infection. We observed that the

PhoP regulon, SPI-1, SPI-2 and spvR are induced in

C. elegans and isogenic strains lacking these viru-

lence factors exhibited significant defects in the ability

to persist in the worm intestine. Salmonella infection

also leads to induction of two C. elegansantimicrobial

genes, abf-2 and spp-1, which act to limit bacterial

proliferation. The SPI-2, phoPand DpSLT mutants are

more sensitive to the cationic peptide polymyxin B,suggesting that resistance to worms antimicrobial

peptides might be necessary for Salmonellato persist

in the C. elegans intestine. Importantly, we showed

that the persistence defects of the SPI-2, phoP and

DpSLTmutants could be rescued in vivowhen expres-

sion of C. elegans spp-1 was reduced by RNAi.

Together, our data suggest that resistance to host

antimicrobials in the intestinal lumen is a key mecha-

nism for Salmonellapersistence.

Introduction

Salmonella typhimurium is a Gram-negative pathogen

that causes enteritis in humans and livestock (Baumler

et al., 1998; Kingsley and Baumler, 2000). Salmonella

evolved to exist in the alimentary tract of the host

(Baumler et al., 1998) in the presence of a number of host

imposed stresses, including low pH in the stomach, bile

and antimicrobial peptides (AMPs) in the small intestine,

and an aerophilic to microaerophilic shift (Foster and

Spector, 1995; Rychlik and Barrow, 2005). Although a

number of S. typhimurium virulence determinants essen-

tial for infecting the mammalian intestinal tract have been

described (reviewed in Darwin and Miller, 1999; Wallis

and Galyov, 2000), their interactions with host-derived

factors in vivo are largely unknown.

Specific virulence factors have been shown to act atdiscrete phases of infection (Galan, 2001). Within the

terminal ileum of the small intestine, attachment and inva-

sion are mediated by activation of SPI-1 and SPI-4 (Finlay

and Falkow, 1997; Morgan et al., 2004). Following trans-

location through the intestinal epithelia, SPI-2 and PhoPQ

are critical for survival in phagocytes (Fields et al., 1989;

Hensel et al., 1998). PhoQ directly senses AMPs (Bader

et al., 2005), acidic pH, changes in cation concentration

(Bearson et al., 1998), as well as membrane damage

brought about by several classes of AMPs that result in

modification of bacterial lipopolysaccharide (Groisman

et al., 1989; Groisman et al., 1992). SPI-2 is also thoughtto respond to acidic pH and cation depletion within the

phagolysosome (Kim and Falkow, 2004).

Recently, components of the PhoP regulon (Merighi

et al., 2005) and SPI-2 (Brown et al., 2005) were reported

to be expressed prior to invasion of murine intestinal

enterocytes. These findings have expanded the activities

of PhoP and SPI-2 beyond the intracellular stage of

infection. In addition, SPI-2 appears to be required for

pathogenesis in a murine colitis model, in which bacteria

remain predominantly luminal (Coburn et al., 2005). Fur-

thermore, mutants lacking pmrH, the first gene within the

PhoP-regulated operon required for LPS modification in

response to AMPs, are attenuated by oral infection (Gunn

et al., 2000). The exact roles that PhoP and SPI-2 play

during the intestinal phase of infection are not clear. The

molecular cues of acidic pH and AMPs present within the

intracellular environment of the phagolysosome are also

present in the gastrointestinal tract and may have similar

roles in inducing virulence gene expression.

In the current study, we used infection of C. elegansby

Salmonellaas the experimental system to explore the role

Received 12 October, 2007; revised 8 January, 2008; accepted 8January, 2008. *For correspondence. E-mail [email protected];Tel. (+1) 650 736 1688; Fax (+1) 650 725 1534. Present address:Department of Molecular and Cell Biology, University of California,Berkeley, Berkeley, CA 94720, USA.

Cellular Microbiology (2008) 10(6), 12591273 doi:10.1111/j.1462-5822.2008.01124.xFirst published online 15 February 2008

2008 The AuthorsJournal compilation 2008 Blackwell Publishing Ltd
mailto:[email protected]:[email protected]


2/15

of bacterial virulence factors and host intestinal AMPs

during infection. C. elegans has proven to be an ame-

nable infection model for a number of bacterial patho-

gens, including S. typhimurium (reviewed in Alegado

et al., 2003; Kurz and Ewbank, 2007). One striking feature

of S. typhimurium pathogenesis in C. elegans is its ability

to colonize and establish a persistent intestinal infection,

even after a limited exposure (Aballay et al., 2000;

Labrousse et al., 2000). Worms feeding on the laboratory

food source Escherichia coliOP50 can limit bacterial pro-

liferation in their gut. In contrast, worms feeding on patho-

genic S. typhimurium rapidly accumulate the pathogen in

the intestine concomitant with bacterial proliferation and

distention of the intestinal lumen. However, the bacterial

factors required for persistence have yet to be fully

investigated.

A large number of putative AMP genes encoded in

the C. elegans genome are expressed in the pharynx

and intestine, sites of contact with intestinal microbes

(Alegado et al., 2003; Kurz and Tan, 2004), suggesting

that AMPs may play a significant role in hostpathogeninteractions. Two of these, ABF-2 (Kato et al., 2002) and

SPP-1 (Banyai and Patthy, 1998), have demonstrated

antimicrobial activity. ABF-2 is homologous to insect and

mollusk defensins and recombinant ABF-2 has broad

activity against a number of yeast, Gram-positive and

Gram-negative bacteria. Under normal growth conditions,

abf-2 is constitutively expressed in the pharynx (Kato

et al., 2002). SPP-1 is a member of the saposin-like

protein family, which includes mammalian NK-lysin and

granulysin. SPP-1 is active against E. coli (Banyai and

Patthy, 1998) and is expressed in the intestine (Alper

et al., 2007). While the in vitro activity of these proteinsimplicates their role in host defence, the immunological

significance of ABF-2 and SPP-1 has not yet been dem-

onstrated at the organismal level.

Here, we show that SPI-1, SPI-2, PhoP and the viru-

lence plasmid are required for optimal establishment of a

persistent intestinal infection in C. elegans. We observe

that in vivo induction of bacterial virulence genes coin-

cides with the induction of host innate defence factors,

ABF-2 and SPP-1. Moreover, both worm antimicrobials

appear to be instrumental in controlling bacterial prolifera-

tion in the intestine.

Results

Salmonella virulence genes are specifically expressed

in vivo during infection of C. elegans

Although C. elegans is efficient at limiting proliferation of

E. coli, its laboratory food source, worms are less effec-

tive at preventing colonization of S. typhimurium. Salmo-

nella infection appears to be restricted to the intestinal

lumen and a number of bacterial genes that contribute to

mortality have been identified (Aballay et al., 2000;

Labrousse et al., 2000; Tenor et al., 2004). While PhoP,

SPI-1 and SPI-2 are required for killing over a course of

9 days (Aballay et al., 2000), the mechanisms by which

these virulence factors operate in the worm is unknown.

We sought to define the context in which each of these

virulence factors may act in vivo during C. elegans

infection. Wild-type worms were infected with S. typhi-

murium (SL1344) bearing individual promotergfp

fusions: prgH (Hautefort et al., 2003), ssaG (Hautefort

et al., 2003) and mig-14 (Brodsky et al., 2002), reporters

for transcriptional activity of SPI-1, SPI-2 and the PhoP

regulon respectively. PrgH and SsaG are structural com-

ponents of the type III secretion apparatus and Mig-14 is

a protein involved in resistance to AMPs that is dependent

on PhoP for expression (Valdivia et al., 2000; Brodsky

et al., 2005). The spatiotemporal expression of each of

these reporters within C. elegans was determined every

24 h. GFP expression was detected within the intestinal

tract of C. elegans 48 h after initial exposure (Fig. 1AF).

Importantly, under our assay conditions, these promoterfusions were expressed specifically in vivo and not when

grown on solid media alone. As no intracellular bacteria

were detected within C. elegans intestinal cells over the

course of the experiment, either by fluorescence or elec-

tron microscopy (data not shown), our observations indi-

cate that these genes are expressed during extracellular

infection of the worm intestine.

To directly assess the transcript abundance of the

reporter genes during early infection, we monitored gene

expression in vivo using quantitative real-time reverse

transcription polymerase chain reaction (qRT-PCR).

Because the PhoP regulon is diverse and complex, weincluded other genes in our analysis in addition to mig-14.

PhoP regulates its own transcription as well as SlyA, a

transcription factor that activates a distinct set of genes

(Norte et al., 2003) and, indirectly, PmrA that activates the

LPS modification pathway in response to AMPs (Gunn and

Miller, 1996). The mig-14 and pagC genes appear to be

regulated by both SlyA and PhoP, by a mechanism that is

not yet understood (Navarre et al., 2005). We chose the

transcription factors SlyA and PmrA, as well as down-

stream targets mig-14, pagC and pagD to represent the

PhoP regulon during infection. We also quantified expres-

sion of spvR, the regulator of the spvlocus on the virulence

plasmid (pSLT) that has been shown to be required for full

virulence in mammals and mortality in worms (Tenor et al.,

2004). The transcript levels of these bacterial genes within

infected worms were determined 1, 24, 48 and 72 h after

initial exposure andnormalized to RNA levelsfrom bacteria

grown on solid media from matched time points.

At 48 h of infection, the transcript levels of bacterial

prgH, ssaG and mig-14 within C. elegans intestine was

approximately 10-fold higher than in bacteria grown on

1260 R. A. Alegado and M.-W. Tan

2008 The AuthorsJournal compilation 2008 Blackwell Publishing Ltd, Cellular Microbiology, 10, 12591273


3/15

solid media (Fig. 1I), thus confirming the reporter assays

that the SPI-1, SPI-2 and PhoP regulons are expressed

during intestinal colonization of C. elegans (Fig. 1AF).

In addition, two distinct transcriptional patterns were

observed. The first set, composed of the genes within the

PhoP regulon, was highly expressed in vivo within 1 h of

exposure relative to bacteria grown on plates (Fig. 1G). At

24 h, these transcripts were no longer induced relative to

external bacteria (Fig. 1H). However, later during infec-

tion, at 48 and 72 h, expression of the phoP regulon was

once again induced (Fig. 1I and J). The reason for this

dynamic change is currently not understood. In contrast,

in vivo levels of the second set of transcripts, prgH, ssaG

and spvR, were initially indistinguishable from external

bacteria at the first hour of infection (Fig. 1G) but were

significantly higher after 24 h of infection and was sus-

tained until at least 72 h of infection (Fig. 1HJ). The

varied expression of these virulence gene sets implies

that the worm gut may exert a number of stresses that

Salmonella must respond to and suggests that these

Salmonella virulence pathways may play a critical role

during extracellular infection and persistence.

Several Salmonella virulence factors are required for

colonization of the worm intestine

Caenorhabditis elegans exposed to Salmonella begin to

die at day 4 of infection (Aballay et al., 2000), yet the

expression pattern of genes within the PhoP regulon,

SPI-1, SPI-2 and pSLT during C. elegans infection

Fig. 1. In vivo expression of Salmonellavirulence genes. GFP expression in wild-type worms infected for 48 h with S. typhimurium bearinggfp fused to the promoter of prgH (A, B), ssaG (C, D) or mig-14 (E, F). Representative images showing DIC (A, C, E), and merge image from

I3 (GFP, green) and A4 (auto-fluorescence, blue) filters. B, D and F are at 400 magnification. Note: Bacteria carrying PprgH::gfpandPmig-14::gfpconsistently displayed GFP signal localized as aggregates adjacent to intestinal cells (B and F). Fluorescence fromPssaG::gfp-expressing bacteria was observed in cells not associated with intestinal cells (D, arrowheads). (G and H) Relative levels ofSalmonellagene transcripts during C. elegans infection as determined by qRT-PCR. Levels of Salmonellatranscripts obtained from infectedworms were normalized to levels of transcripts from bacteria grown on solid media at matched time points, 1 h (G) 24 h (H), 48 h (I) and 72 h(J). Shown is the mean s. e. of five independent experiments. Dotted line indicates normalized transcript levels of each gene in solid mediacontrols. Unpaired t-test, * P< 0.05,** P< 0.01, ***P< 0.001.

Persistence of Salmonella in the C. elegans intestine 1261

2008 The AuthorsJournal compilation 2008 Blackwell Publishing Ltd, Cellular Microbiology, 10, 12591273


4/15

suggest that these virulence factors might be required

early, perhaps during the establishment of an intestinal

infection. We therefore ascertained the consequence of

lacking phoP, orgA (SPI-1) or ssaV (SPI-2) or the viru-

lence plasmid (DpSLT) on the ability of S. typhimurium to

colonize the C. elegans. We accomplished this by follow-

ing the kinetics of bacterial accumulation in the nematode

intestine over time. The population of intestinal Salmo-nellareaches between 104 and 105 bacteria/worm the first

3 days of infection (data not shown, Aballay et al., 2000).

To take advantage of the ability to visualize bacteria within

living worms over the course of infection, we infected

animals with SM022, a derivative of SL1344 harbouring a

single chromosomal copy of gfp constitutively driven by

the rpsM promoter (Vazquez-Torres et al., 1999). SM022

is phenotypically identical to wild-type SL1344 under all

conditions tested (data not shown). Worms were visually

scored for severity of colonization based on the extent of

luminal distention and gfpsignal in the intestine (Fig. 2A).

As worms have a number of mechanical and chemical

mechanisms for restricting bacteria in the gut, individual

animals were colonized at different rates (Fig. 2B). Com-

pared with SM022, the phoPgfp, DpSLTgfp, orgAgfp

and ssaVgfpisogenic mutants colonized C. elegansto a

similar degree during the first 48 h of infection. After 72 h,

however, these mutants colonized C. eleganssignificantly

less than SM022 (Fig. 2B). Specifically, almost 90% of

animals feeding on SM022 scored in the full colonization

category whereas animals constantly exposed to any of

the mutants had a score of 50% or less. Differences

between SM022 and mutants were also notable when we

compared the change in the severity of colonization

between 48 and 72 h (Fig. 2B, chi-squared test, P

Documents

Alegado Cell Microbiol 2008