Microbes and Infection 9 (2007) 400e407www.elsevier.com/locate/micinf

Original article

Citrobacter rodentium virulence in mice associates with bacterial loadand the type III effector NleE

Mark E. Wickham a,1, Claudia Lupp a, Alejandra Vazquez b, Mariola Mascarenhas c,Bryan Coburn a, Brian K. Coombes a,d, Mohamed A. Karmali c, Jose L. Puente a,b,

Wanyin Deng a, B. Brett Finlay a,*

a Michael Smith Laboratories, The University of British Columbia, #301 2185 East Mall, Vancouver BC V6T 1Z4, Canadab Departamento de Microbiologıa Molecular, Instituto de Biotecnologıa, Universidad Nacional Autonoma de Mexico, Cuernavaca, Mexico

c Laboratory for Foodborne Zoonoses, Population and Public Health Branch, Health Canada, Ontario, Canadad McMaster University, Hamilton, Ontario, Canada

Received 25 September 2006; accepted 15 December 2006

Available online 13 January 2007

Abstract

Severe disease caused by Shiga toxin-producing Escherichia coli (STEC) has been associated with a pathogenicity island, O-Island 122,which encodes the type III secretion system-effector NleE. Here we show that full virulence of the related attaching and effacing mouse path-ogen Citrobacter rodentium requires NleE. Relative to wild-type bacteria, nleE-mutant C. rodentium are attenuated for colonisation in mice inboth single and mixed infections. Examination of the ability of nleE-mutant bacteria to induce pathologic change in vivo revealed that nleE-mutant bacteria induce significantly less pathologic change than wild-type bacteria in susceptible mice. Consistent with these results, miceinfected with nleE-mutant bacteria exhibit delayed mortality. These results suggested that pathologic change during attaching and effacingpathogen infection may associate with the degree of pathogen colonisation. Using mutants of 23 type III secretion genes, including the typeIII effectors nleC, nleD, nleE and nleF, the association of pathologic change with the ability of these mutants to colonise mice was examined.The induction of in vivo disease correlates strongly with the degree of colonisation, suggesting that the colonisation advantage type III secretiongenes afford the bacteria, contribute to, and are required for, full virulence.� 2007 Published by Elsevier Masson SAS.

Keywords: Virulence; Bacteria; Bacterial infections

1. Introduction

Haemolytic uraemic syndrome (HUS) causes acute renalfailure in children and adults, and follows infection withShiga-toxin producing Escherichia coli (STEC) [1e4]. HUSand human outbreaks caused by Shiga-toxin producingE. coli infection have been associated with a pathogenicity

* Corresponding author. Tel.: þ1 604 822 2493; fax: þ1 604 822 2114.

E-mail address: bfinlay@interchange.ubc.ca (B. Brett Finlay).1 Present address: The Walter and Eliza Hall Institute of Medical Research,

1G Royal Parade, Melbourne, Victoria 3050, Australia. Tel.: +61-3-93452348;

fax: +61-3-93470852. wickham@wehi.edu.au; http://www.wehi.edu.au.

1286-4579/$ - see front matter � 2007 Published by Elsevier Masson SAS.

doi:10.1016/j.micinf.2006.12.016

island of the enterohemorrhagic E. coli (EHEC) O157:H7 ge-nome called O-Island 122 [5,6]. O-Island 122 is absent fromnon-pathogenic E. coli and contains putative virulence deter-minants including homologues of putative Salmonella and Shi-gella virulence-associated factors.

Work using the attaching and effacing mouse pathogen Cit-robacter rodentium identified a family of type III secretedeffectors encoded outside of the locus of enterocyte efface-ment (LEE)-pathogenicity island [7], a region which containsthe virulence-associated type III secretion system. This family,the non-LEE encoded effectors (nle), includes nleE and nleB,which are encoded within the disease-associated pathogenicityisland, O-Island 122 [6,8].

401M.E. Wickham et al. / Microbes and Infection 9 (2007) 400e407

While O157:H7 is considered the most virulent STEC sero-type, many non-O157 strains are as virulent as O157:H7STEC. Examination of the distribution of O-Island 122 genesin human STEC strains has demonstrated that both nleB andnleE associate strongly with non-O157 STEC virulence in hu-mans [8]. For instance, non-O157 STEC strains positive fornleE are four times as likely to be associated with the abilityto cause HUS, and 28 times more likely to be associated withoutbreaks of non-O157 STEC [8]. That these type III effectors(nleB and nleE ) associate with non-O157 STEC virulence, sug-gests that these molecules (present in all O157:H7 STEC strainsexamined) are pivotal to STEC virulence. Examination of thecontribution of these molecules in vivo disease is important tounderstand the pathogenesis of pathogenic E. coli infection. Pre-vious work has demonstrated a pivotal role for NleB in modulat-ing infectious dose and the outcome of infection [8].

Here we show that the translocated effector NleE is requiredfor full colonisation of mice, and full induction of disease. Sincethe reduction in virulence of nleE-mutant bacteria appeared re-lated to its lessened ability to colonise, we sought to examinewhether the in vivo virulence of other C. rodentium mutantsassociate with bacterial load. Using defined mutations in 23 C.rodentium type III secretion-associated genes, we show that hostmorbidity correlates with the colonisation ability of C. rodentium.Interestingly, the in vivo virulence defects of bacteria mutant fortype III components necessary for secretion depend on the hoststrain examined, suggesting that bacteria deficient for majorvirulence pathways may persist in populations, something pro-posed for STEC in cattle [9]. The acquisition of virulence genesby such a reservoir of bacteria of lessened pathogenic potential(and acquisition of the fitness advantage these genes afford thebacteria) may facilitate the emergence of virulence.

2. Materials and methods

2.1. Infection analysis of C. rodentium andmutants in mice

Non-polar mutations in nleC, nleD, nleE and nleF inC. rodentium DBS100 [10] were generated using lambda Redrecombinase based method [11]. Non-polar mutations in thetype III secretion-associated genes rorf3, grlR, grlA, cesD,rorf6, escJ, rorf8, sepZ, orf12, escV, escN, orf15, orf16, sepQ,espH, cesF, cesT, sepL, and cesD2 have been described previ-ously [7]. Infection of mice with wild-type C. rodentium andnleE-mutant bacteria, and pathological analyses were performedusing standard procedures [7]. To check for bacterial colonisation,colonic tissues plus fecal pellets were homogenized in PBS,using a Polytron tissue homogenizer, and serially diluted beforebeing plated onto MacConkey agar (Difco) or LB agar. Colonictissue and the fecal pellets were combined to determine the totalbacterial burden in the mouse colon at the time of sacrifice.

2.2. Histopathology

Distal colons of infected C3H/HeJ mice removed 6 dayspost-infection, C57BL/6 mice 6 or 12 days post-infection,

and NIH Swiss mice 10 days post-infection, were fixed in3% formalin for 18 h followed by 18 h in 70% ethanol priorto being embedded in paraffin, sectioned and stained with hae-matoxylin and eosin (H&E). Pathological scores were deter-mined by blind scoring (by MEW and BC) of H&E sectionsas per Ref. [8].

2.3. Statistical analysis

Competitive index scores were examined using one samplet-test. Survival curves were examined using Logrank test, col-onisation was examined using ManneWhitney test, and cryptheight was examined using one-way ANOVA and NewmaneKeuls multiple comparison test. Quantitative histopathologicalanalysis was examined using ManneWhitney test. To examinethe association between C. rodentium colonisation and in vivodisease, both the Pearson correlation coefficient (r) and the P-value of the correlation were calculated. For display purposes,non-linear regression was used to plot a curve of the relation-ship, and r2 was calculated.

3. Results

3.1. NleE is conserved across pathogenic bacteria

In addition to an association with EHEC O157:H7 viru-lence, O-Island 122 has been demonstrated to associate withthe ability of non-O157 STEC to cause severe disease inhumans [6]. Since the presence of nleE in non-O157 STECstrains associates with the ability of non-O157 strains to causesevere disease in humans [8], we sought to examine the impor-tance of this molecule to in vivo disease. A survey of NleEamino acid sequences (Fig. 1) revealed NleE, which is ex-pressed by the attaching and effacing mouse pathogens C. ro-dentium, enteropathogenic E. coli (EPEC), Shiga-toxinproducing E. coli (STEC) and the Shigella species, Shigelladysenteriae (and Shigella sonnei, data not shown) [7,12,13]is highly conserved. NleE of Shigella spp. shares 76e77%amino acid sequence identity and 88e89% similarity withthat of EHEC O157:H7. Similarly, NleE of STEC strains(O157:H7, O103:H21 and O26:H�) shares 86% identity and94% similarity with C. rodentium, suggesting that the functionof NleE is conserved across these species. nleE of the C.rodentium strains ICC168 and DBS100 is 100% identical.Consequently, we sought to examine the role of nleE in viru-lence using the attaching and effacing mouse pathogenC. rodentium (DBS100).

3.2. NleE is a virulence determinant in mice

To examine the contribution of nleE to virulence, we per-formed competitive index experiments (Fig. 2A). Mice wereinfected with a mixed inoculum of wild-type C. rodentiumDBS100 and nleE-mutant bacteria, and the bacterial load ofeach strain was examined 7 days post-infection in C57BL/6mice, and 4 days post-infection in C3H/HeJ mice. In mixedinfections with wild-type bacteria, the nleE-mutant was

402 M.E. Wickham et al. / Microbes and Infection 9 (2007) 400e407

1 10 20 30 40 50 60 70

C.rodentium MI P N INT E V NIYP H YFNE PN YDKKYISG TR A LK E F NEK RR N IT TQ L K A V K EIE V V E Q G I A N LS H Y N P GV LEHEC O157:H7 MI P N INT E V NIYP H YFNE PN YDKKYISG TR A LK E F NEK RR N VT TQ V K A V K EIK I I E Q E V A G SP Y H D S GV FSTEC O103:H2 MI P N INT E V NIYP H YFNE PN YDKKYISG TR A LK E F NEK RR N VT TQ V K A V K EIK I I E Q E V A G SP Y H D S GV FSTEC O26:H- MI P N INT E V NIYP H YFNE PN YDKKYISG TR A LK E F NEK RR N VT TQ V K A V K EIK I I E Q E V A G SP Y H D S GV FEPEC MI P N INT E V NIYP H YFNE PN YDKKYISG TR A LK E F NEK RR N VT TQ V K A V K KIK I I E Q E V A G SP Y H D S GV FS.dysenteriae MI P N INT E V NIYP H YFNE PN YDKKYISG TR A LK E F NEK RR S IK IK V A T I R RVE I I Q I E I S N FP N Y G S SM LS.flexneri MI P N INT E V NIYP H YFNE PN YDKKYISG TR A LK E F NEK RR S IK IK V A T I R RVE I I Q I E I S N FP N Y G S SM L

80 90 100 110 120 130 140

C.rodentium YM MYS CPE F PISR EASTPEGSWLTVISGKRPMGQFSVDSLY PDLHALCELP I CKIF KENA QA K N D V E H C P EHEC O157:H7 YM MYS CPE F PISR EASTPEGSWLTVISGKRPMGQFSVDSLY PDLHALCELP I CKIF KENS KT E N D V A N C P STEC O103:H2 YM MYS CPE F PISR EASTPEGSWLTVISGKRPMGQFSVDSLY PDLHALCELP I CKIF KENS KT E N D V A N C P STEC O26:H- YM MYS CPE F PISR EASTPEGSWLTVISGKRPMGQFSVDSLY PDLHALCELP I CKIF KENS KT E N D V A N C P EPEC YM MYS CPE F PISR EASTPEGSWLTVISGKRPMGQFSVDSLY PDLHALCELP I CKIF KENS KT E N D V A N C P S.dysenteriae YM MYS CPE F PISR EASTPEGSWLTVISGKRPMGQFSVDSLY PDLHALCELP I CKIF KENN KT Q D E P D H S P S.flexneri YM MYS CPE F PISR EASTPEGSWLTVISGKRPMGQFSVDSLY PDLHALCELP I CKIF KENN KT Q A E P D H S S

150 160 170 180 190 200 210

C.rodentium DFLYI V RNDSP GE RANRFIELY IKR IMQ L E PE K N VI Y L Q D E N S L AV N Y KSEMIIAREMGEIFGYVLEEIDEHEC O157:H7 DFLYI V RNDSP GE RANRFIELY IKR IMQ L E PE K N VV Y L Q D E N L L AV N Y KSEMIIAREMGEIFSYMPGEIDSTEC O103:H2 DFLYI V RNDSP GE RANRFIELY IKR IMQ L E PE K N VV Y L Q D E N L L AV N Y KSEMIIAREMGEIFSYMPGEIDSTEC O26:H- DFLYI V RNDSP GE RANRFIELY IKR IMQ L E PE K N VV Y L Q D E N L L AV N Y KSEMIIAREMGEIFSYMPGEIDEPEC DFLYI V RNDSP GE RANRFIELY IKR IMQ L E PE K N VV Y L Q D E N L L AV N Y KSEMIIAREMGEIFSYMPGEIDS.dysenteriae DFLYI V RNDSP GE RANRFIELY IKR IMQ L E PE K S IV F Q L E V R S L SI D D KSEIIIAREMGELFSYASEEIDS.flexneri DFLYI V RNDSP GE RANRFIELY IKR IMQ L E PE K S IV F Q L E V R S I VY D D

220 230

C.rodentium SYMKNMNNKLAQIEARRSAT STEC O157:H7 SYMKYINNKLSKIE STEC O103:H2 SYMKYINNKLSKIE STEC O26:H- SYMKYINNKLSKIE EPEC SYMKYINNKLSKIE S.dysenteriae SYIKQMNDRLSQIKARMPVT

Fig. 1. NleE is conserved across pathogenic bacteria. Sequence alignment of NleE using the complete or near-complete genome sequences from EHEC [5],

C. rodentium and EPEC strain O127:H6 (EPEC1 strain E2348/69) [37], STEC strains O103:H2 (GenBank accession no. AJ303141) and O26:H� (GenBank

accession no. AJ277443), and Shigella dysenteriae M131649 (M131) [37] and Shigella flexneri strain 5a (GenBank accession no. AF348706).

significantly attenuated, with a competitive index of 0.45�0.062 (P< 0.0001) in C57BL/6 mice, and 0.49� 0.052(P< 0.0001) in C3H/HeJ mice. This defect is consistent withrecent work that identified a competitive disadvantage fornleE-mutant bacteria only in mixed infections with C. rodentiumICC169 [14].

3.3. NleE is required for full induction andprogression of disease

While many type III effectors appear to be absolutelyrequired for mortality in vivo (tir, sepZ, nleA, and nleB)[7,8,15], delayed mortality has been demonstrated for feweffector mutants (espF and ent/espL2), with many (includingespG, espH, and map) exhibiting no significant delay in death[7,8,15]. To assess the role of nleE in causing mortality inmice most susceptible to C. rodentium infection, C3H/HeJmice were infected with the nleE-mutant and survival of theinfected mice was examined. C3H/HeJ mice are more suscep-tible to infection with C. rodentium than C57BL/6 mice andsuffer high mortality rates. Examination of survival time indi-cated that mice infected with the nleE-mutant bacteria survivesignificantly longer than mice infected with wild-type bacteria(P¼ 0.0429) (Fig. 2B). This result suggests that nleE playsa role in the progression of disease.

To examine the contribution of nleE to colonisation, resis-tant C57BL/6 mice, in which virulence defects are morepronounced [7,16,17], were infected with nleE-mutant orwild-type C. rodentium (DBS100). Colonisation of mice bythe nleE-mutant was significantly reduced (P¼ 0.0207)

relative to colonisation by wild-type C. rodentium at day 6post-infection (5.85� 0.97� 107 and 2.48� 0.81� 108 bac-teria per colon, respectively) (Fig. 2C), indicating that NleEcontributes to the ability of C. rodentium to colonise mice,consistent with the observed defect in competitive experimentsrelative to wild-type C. rodentium (Fig. 1A). This result is incontrast to recent work suggesting that nleE plays no role inthe colonisation or disease [14]. In this work, Kelly et al.[14] generated an nleE mutation in the C. rodentium isolateICC169, an isolate for which differences in virulence relativeto the C. rodentium isolate DBS100 have been highlightedpreviously [18]. Similarly, earlier work has suggested thatdifferent C. rodentium isolates may possess a spectrum ofvirulence [19].

To examine the ability of the nleE-mutant to cause diseasein vivo, we measured crypt height as a marker of epithelial cellhyperplasia (a hallmark of C. rodentium disease) [7] in the de-scending colon at days 6 and 12 post-infection [7] (Fig. 2D).At day 6 post-infection, the crypt heights of mice infectedwith the nleE-mutant was significantly less (P< 0.01) relativeto mice infected with wild-type C. rodentium (0.146� 0.0071,and 0.201� 0.0078 mm, respectively), indicating that nleE-mutant bacteria are defective for the induction of hyperplasia.At day 12 post-infection, the crypt heights of mice infectedwith the nleE-mutant was again significantly less relative tomice infected with wild-type C. rodentium (0.289� 0.011,and 0.329� 0.011 mm, respectively). However, the meandifferences in crypt height between mice infected with nleE-mutant and wild-type bacteria are diminished at day 12 rela-tive to day 6 (0.055 and 0.040 mm, respectively), indicating

403M.E. Wickham et al. / Microbes and Infection 9 (2007) 400e407

C57BL/6 C3H/HeJ0.01

0.1

1

A

Co

mp

etitive In

dex ** **

nleE

WT

Day

B

Percen

t su

rvival

0 2 4 6 8 100

20

40

60

80

100

*

WT

1.0x1009

1.0x1008

1.0x1007

nleE

C

cfu

/co

lo

n

*

WT nleE WT nleE

D

Cryp

t h

eig

ht (m

m)

*

0

0.05

0.10

0.15

0.20

0.25

0.30

0.35

Day 6 Day 12

*

Fig. 2. nleE is a virulence determinant in mice. (A) Competitive indices of wild-type bacteria relative to isogenic nleE-mutant bacteria in the mouse colon indicates

nleE is attenuated relative to wild-type in mixed infections in C3H/HeJ and C57BL/6 mice at days 4 and 7 post-infection, respectively (P< 0.0001, one sample t-

test). Mean competitive index of chromosomally marked wild-type versus wild-type C. rodentium shown as dotted line (wild-type C. rodentium versus chromo-

somally marked wild-type C. rodentium give a competitive index not significantly different from 1). Each data point represents one animal and horizontal bars

indicate means. (B) Survival plots of susceptible mice infected with wild-type bacteria (inverted triangles, red) and nleE-mutant bacteria (squares, green) indicate

mice infected with nleE-mutant bacteria exhibit delayed mortality relative to mice infected with wild-type bacteria (P< 0.05, Logrank test). The percentages of

mice surviving after infection are shown. (C) nleE-mutant bacteria are attenuated for colonisation relative to wild-type (WT) bacteria (P¼ 0.0207, ManneWhitney

test). Living bacteria were quantified in colon homogenates collected from infected C57BL/6 mice at day 6 post-infection. Mean (�SE) colony forming units (cfu)

per colon are presented. (D) Crypt heights [mean (�SE)] of infected C57BL/6 mice at days 6 and 12 post-infection indicates nleE-mutant bacteria are attenuated

for virulence at both time-points [P< 0.01, one-way analysis of variance (ANOVA) with NewmaneKeuls multiple comparison test]. Single asterisk (*) and double

asterisk (**) denote statistical significance (P< 0.05 and P< 0.01, respectively).

that nleE-mutant bacteria are more attenuated early ininfection. This result is consistent with nleE-mutant bacteriacausing delayed mortality in mice.

3.4. NleE contributes to pathologic change in micesusceptible to C. rodentium infection

To assess the role of nleE in disease pathology, histopatho-logical change was examined in both C57BL/6 mice which areresistant to lethal infection with C. rodentium, and C3H/HeJmice which are susceptible to C. rodentium infection. In resis-tant C57BL/6 mice at 6 days post-infection (Fig. 3A and B),mild, patchy regenerative change in the epithelium withlimited but apparent edema in the lamina propria of crypt api-ces was evident in wild-type but not nleE-mutant infection;however, this difference was not significantly different(P¼ 0.2345, ManneWhitney test). Importantly, while

pathologic change induced by wild-type C. rodentium can bedistinguished from uninfected mice (P< 0.0451, ManneWhitney test) inflammatory pathology at this time-point wasminimal in both strains tested.

Pathologic change was examined 12 days post-infection inC57BL/6 mice: a time-point at which C. rodentium-inducedinflammation is maximal. Mice infected with wild-type andnleE-mutant bacteria exhibit significant histopathologicalchange relative to uninfected mice (Fig. 3C and D). Miceinfected with nleE-mutant bacteria exhibited decreased patho-logic change relative to change induced during infection withwild-type bacteria, although this did not reach statistical sig-nificance (P< 0.54, ManneWhitney test).

Pathologic change was also examined 6 days post-infectionin C3H/HeJ mice which are susceptible to C. rodentium infec-tion. C3H/HeJ mice are Toll-like receptor 4 (TLR4)-negativeand usually succumb to lethal infection with C. rodentium

404 M.E. Wickham et al. / Microbes and Infection 9 (2007) 400e407

LumenSurface EpitheliumMucosaLamina Propria

LumenSurface EpitheliumMucosaLamina Propria

LumenSurface EpitheliumMucosaLamina Propria

D

nleEWT Uninfected

nleEWT UninfectednleEWT Uninfected

nleEWT Uninfected

A

0

5

10

15

20

C57BL/6 - Day 12

C57BL/6 - Day 6

Path

olo

gy S

co

re

0

5

10

15

20

Path

olo

gy S

co

re

0

5

10

15

20

Path

olo

gy S

co

re

B

C

WT nleE Uninfected

UninfectednleEWT

FE

-0.0079

P value

vs WTvs Uninfected

0.01590.01590.0159 -

C3H/HeJ - Day 6

Fig. 3. nleE-mutant bacteria are attenuated for the ability to induce pathologic change in susceptible mice. (A) Microscopic examination of colons of C. rodentium-

resistant C57BL/6 mice 6 days post-infection shows that pathologic change is most severe in the colon of mice infected with wild-type bacteria, while mice in-

fected with the nleE-mutant exhibit less pathologic change relative to wild-type C. rodentium DBS100 infection. This pathologic change was not significantly

different from wild-type C. rodentium DBS100 infection (ManneWhitney test). Representative colons of infected mice at day 6 post-infection are shown in

(B). (C) Microscopic examination of colons of C57BL/6 mice 12 days post-infection showed that quantitative pathologic change is most severe in the colon

of mice infected with wild-type bacteria, while mice infected with the nleE-mutant exhibit varying pathologic change not significantly different from wild-

type C. rodentium DBS100 infection (ManneWhitney test). Representative colons of infected mice at day 12 post-infection are shown in (D). (E) In C. roden-

tium-susceptible C3H/HeJ mice, microscopic examination of colons 6 days post-infection showed that mice infected with nleE-mutant bacteria display significantly

lessened pathologic change relative to infection with wild-type C. rodentium DBS100 (P¼ 0.0159, ManneWhitney test). Representative colons of infected C3H/

HeJ mice at day 6 post-infection are shown in (F). Each bar in panels A, C and E represents an individual mouse. Images are shown at a magnification of 100�.

by day 7 post-infection and display a strong inflammatory re-sponse. Mice infected with wild-type and nleE-mutant bacteriaexhibit significant histopathological change relative to unin-fected mice (Fig. 3E and F). Importantly, susceptible C3H/HeJ mice infected with nleE-mutant bacteria exhibited signif-icantly reduced pathologic change relative to the changeinduced during infection with wild-type bacteria (P¼0.0159, ManneWhitney test), indicating that NleE playsa role in virulence in susceptible mice.

3.5. In vivo disease associates with bacterialload during C. rodentium infection

The decrease in the induction of hyperplasia and pathologyin mice by nleE-mutant bacteria, which are also attenuated forcolonisation (Fig. 1C and D), suggested that the degree of hy-perplasia observed in mice may relate to bacterial load duringinfection. To examine this possibility, the degree of hyperpla-sia induced and bacterial load were examined for C. rodentium

405M.E. Wickham et al. / Microbes and Infection 9 (2007) 400e407

mutants of 23 type III secretion-associated genes [7] in NIHSwiss mice at day 10 post-infection (Fig. 4). NIH Swissmice were chosen, since these mice get the most extensive co-lonic hyperplasia during C. rodentium infection [20,21]. Thisexamination showed that as bacterial load increases, there isa concomitant increase in the induction of hyperplasia (Pear-son r¼ 0.7732; P< 0.0001). This correlation is also signifi-cant in C57BL/6 mice at day 10 post-infection (data notshown). That some mutants defective for type III secretion(such as escV, escN, sepQ, grlA, grlR, and sepD) [7] areable to colonise mice and induce some pathologic change issuggestive that the contribution to virulence that these genesmake is via the colonisation advantage they afford the bacteriaduring infection.

4. Discussion

In this work we have identified a role for NleE in the induc-tion and progression of in vivo disease. Both the absence ofnleE from many non-O157 STEC strains, in addition to thepresence of nleE homologues in C. rodentium, STEC and Shi-gella spp. suggests that like other non-LEE encoded effectors,nleE appears to have been acquired not only independently ofthe LEE, but across pathogenic bacteria. Obviously, the fitnessbenefit (in the case of nleE, colonisation advantage) conferredupon pathogens harbouring nleE is dependent on the presenceof the type III secretion system, and appropriate transcriptionalregulation of nle gene expression. Since type III effectors

0.9

1.1

1.3

1.5

1.7

1.9

2.1

Bacterial load (cfu)

r=0.7732; P<0.0001

1.0x1

04

1.0x1

09

1.0x1

08

1.0x1

07

1.0x1

06

1.0x1

05

1.0x1

03

Fo

ld

w

eig

ht in

crease

Fig. 4. In vivo disease associates with bacterial load during C. rodentium in-

fection. Plot of bacterial load (cfu, colony forming units) versus fold change

in colon weight (relative to uninfected mice) for wild-type C. rodentium

DBS100 (inverted triangle), nleE-mutant C. rodentium (unfilled circle), and

23 type III secretion-associated gene mutants (filled circles). Fold-increase

in colon weight correlates significantly with increased bacterial load (Pearson

r¼ 0.7732; P< 0.0001). Also shown is the non-linear regression curve of

fold-increase in colon weight as a function of bacterial load (r2¼ 0.862).

Type III secretion-associated mutants examined are rorf3, grlR, grlA, cesD,

rorf6, escJ, rorf8, sepZ, orf12, escV, escN, orf15, orf16, sepQ, espH, cesF,

cesT, sepL, cesD2, nleC, nleD, nleE, and nleF. Each data point represents

the means of data from 10 NIH Swiss mice sacrificed at day 10 post-infection

(n¼ 240 mice, total).

expressed under the control of exogenous promoters are capa-ble of being secreted in a type III dependent fashion, it islikely that when new effectors such as nleE are first acquiredby pathogens, they are able to be expressed and secreted, ashas been demonstrated in Salmonella [22,23]. The subsequentfine-tuning of virulence gene expression (and expression in re-sponse to appropriate environmental cues) is important for thefine-tuning of fitness, and consequently virulence [24,25]. Fur-ther work is underway to examine how transcriptional regula-tion both effectors and type III secretion system componentsimpacts upon in vivo virulence.

Early work [19] characterising the mouse attaching andeffacing pathogen C. rodentium identified potential differencesin the virulence of three C. rodentium strains: DBS100[10,26], mouse-pathogenic E. coli (MPEC) [27], andCDC1843-73T [28]. This work found that MPEC colonisedmice to a significantly lower level than DBS100 andCDC1843-73T strains of C. rodentium. DBS100-infectedmice exhibited a significant amount of necrosis, multifocalerosion and ulceration not apparent in mice infected with theother two C. rodentium strains, and strain CDC1843-73T didnot appear to cause the morbidity and mortality observedwith DBS100 [19]. Consequently it remains possible thatwhile C. rodentium strains appear clonal [19], there are differ-ences in the virulence of individual isolates, much like theapparent variability in the virulence of EHEC O157:H7[29e31] which is also considered clonal [32]. These potentialdifferences in the virulence may be attributable to differentvirulence gene content [9,33,34] as has been suggested forEHEC O157:H7, or possibly differences in the regulation ofvirulence gene expression. The differences between the workpresented here and that of others [14] may be due to suggesteddifferences in the virulence of these strains [18]. It alsoremains possible that the regulation of virulence in DBS100,the C. rodentium strain in which the nle family of effectorswas originally characterised, differs from other C. rodentiumisolates.

To rule out the possibility that the apparent differences invirulence of C. rodentium strains are attributable to sequencedifferences in nleE, we examined the sequence of this viru-lence factor in the C. rodentium strains DBS100 andICC168. nleE was 100% identical between these strains sug-gesting that it is indeed the genetic background of the strainsthat is responsible for these differences. It is possible that thehigh virulence of C. rodentium DBS100 allows the examina-tion of attenuation across a broader spectrum of disease. None-theless, based on the high degree of identity and similarityobserved between NleE of different pathogens (Fig. 1), it isreasonable to expect that like NleE of C. rodentium, theNleE homologues of Shigella spp. and STEC are importantfor the progression of in vivo disease.

It is apparent that type III secretion-associated mutants areattenuated to different degrees in different host strains; forinstance C. rodentium mutant for the LEE-encoded regulator(ler) or the translocated intimin-receptor (tir) while unable tocolonise the outbred mouse strain NIH Swiss, are able tocolonise the inbred mouse line C57BL/6 [7]. That some type

406 M.E. Wickham et al. / Microbes and Infection 9 (2007) 400e407

III-associated genes contribute to virulence to differentdegrees in different hosts suggests that host diversity mayhave contributed to the evolution of type III secretion systemsable to deal with diversity in pathogen resistance. Similarly,that the type III effector nleE appears to play a greater rolein the progression of disease in susceptible hosts suggeststhat the targets of virulence factors (and virulence strategies)may differ across hosts, as has been recently demonstratedin Salmonella [35].

It is interesting that nleE appears to play a role in patho-logic change in the Toll-like receptor 4 (TLR4)-deficientmouse strain C3H/HeJ. Previous work has suggested thatTLR4, which triggers an inflammatory response to lipopoly-saccharide of Gram-negative bacteria, mediates much of thepathology and inflammation observed during infection [36].However, it is clear that wild-type C. rodentium is able to in-duce colitis and pathologic change in TLR-deficient mice([36], this work). The decreased pathologic change inducedby nleE-mutant bacteria in TLR4-deficient mice may suggesta role for this molecule in the induction or progression ofTLR4-independent pathologic change. It has been shown pre-viously that Shigella spp. induce inflammatory responses toaid in dissemination throughout the host; it is possible thatthe family of molecules presented in this work contributes tothis induction.

The differences in contribution to in vivo fitness by type IIIsecretion genes observed has important implications for theway in which we consider the evolution of type III secretion.It is possible that these differences in contribution to bacterialfitness have facilitated the assembly and evolution of type IIIsecretion systems, with ancestral core components allowingsome degree of virulence, and the subsequent accumulationof additional components that modulate the core function ofthe system (with regard to regulation, efficiency, timing,etc.), and effectors (such as the nles which modulate host func-tion), increasing fitness of the bacteria. It is suggestive thatirrespective of molecular function, the contribution to viru-lence that these genes make is via the colonisation advantagethey afford the bacteria during infection.

References

[1] M.A. Karmali, B.T. Steele, M. Petric, C. Lim, Sporadic cases of haemo-

lyticeuraemic syndrome associated with faecal cytotoxin and cytotoxin-

producing Escherichia coli in stools, Lancet 1 (1983) 619e620.

[2] J. Konowalchuk, J.I. Speirs, S. Stavric, Vero response to a cytotoxin of

Escherichia coli, Infect. Immun. 18 (1977) 775e779.

[3] L.W. Riley, R.S. Remis, S.D. Helgerson, H.B. McGee, J.G. Wells,

B.R. Davis, R.J. Hebert, E.S. Olcott, L.M. Johnson, N.T. Hargrett,

P.A. Blake, M.L. Cohen, Hemorrhagic colitis associated with a rare

Escherichia coli serotype, N. Engl. J. Med. 308 (1983) 681e685.

[4] A.O. O’Brien, T.A. Lively, M.E. Chen, S.W. Rothman, S.B. Formal, Es-

cherichia coli O157:H7 strains associated with haemorrhagic colitis in

the United States produce a Shigella dysenteriae 1 (SHIGA) like cyto-

toxin, Lancet 1 (1983) 702.

[5] N.T. Perna, G. Plunkett 3rd., V. Burland, B. Mau, J.D. Glasner, D.J. Rose,

G.F. Mayhew, P.S. Evans, J. Gregor, H.A. Kirkpatrick, G. Posfai,

J. Hackett, S. Klink, A. Boutin, Y. Shao, L. Miller, E.J. Grotbeck,

N.W. Davis, A. Lim, E.T. Dimalanta, K.D. Potamousis, J. Apodaca,

T.S. Anantharaman, J. Lin, G. Yen, D.C. Schwartz, R.A. Welch,

F.R. Blattner, Genome sequence of enterohaemorrhagic Escherichia

coli O157:H7, Nature 409 (2001) 529e533.

[6] M.A. Karmali, M. Mascarenhas, S. Shen, K. Ziebell, S. Johnson, R. Reid-

Smith, J. Isaac-Renton, C. Clark, K. Rahn, J.B. Kaper, Association of

genomic O island 122 of Escherichia coli EDL 933 with verocytotoxin-

producing Escherichia coli seropathotypes that are linked to epidemic

and/or serious disease, J. Clin. Microbiol. 41 (2003) 4930e4940.

[7] W. Deng, J.L. Puente, S. Gruenheid, Y. Li, B.A. Vallance, A. Vazquez,

J. Barba, J.A. Ibarra, P. O’Donnell, P. Metalnikov, K. Ashman, S. Lee,

D. Goode, T. Pawson, B.B. Finlay, Dissecting virulence: systematic

and functional analyses of a pathogenicity island, Proc. Natl. Acad.

Sci. U.S.A. 101 (2004) 3597e3602.

[8] M.E. Wickham, C. Lupp, M. Mascarenhas, A. Vazquez, B.K. Coombes,

N.F. Brown, B.A. Coburn, W. Deng, J.L. Puente, M.A. Karmali,

B.B. Finlay, Bacterial genetic determinants of non-O157 STEC outbreaks

and hemolyticeuremic syndrome after infection, J. Infect. Dis. 194

(2006) 819e827.

[9] J. Kim, J. Nietfeldt, A.K. Benson, Octamer-based genome scanning dis-

tinguishes a unique subpopulation of Escherichia coli O157:H7 strains in

cattle, Proc. Natl. Acad. Sci. U.S.A. 96 (1999) 13288e13293.

[10] D.B. Schauer, B.A. Zabel, I.F. Pedraza, C.M. O’Hara, A.G. Steigerwalt,

D.J. Brenner, Genetic and biochemical characterization of Citrobacterrodentium sp. nov, J. Clin. Microbiol. 33 (1995) 2064e2068.

[11] K.A. Datsenko, B.L. Wanner, One-step inactivation of chromosomal

genes in Escherichia coli K-12 using PCR products, Proc. Natl. Acad.

Sci. U.S.A. 97 (2000) 6640e6645.

[12] T. Le Gall, M. Mavris, M.C. Martino, M.L. Bernardini, E. Denamur,

C. Parsot, Analysis of virulence plasmid gene expression defines three

classes of effectors in the type III secretion system of Shigella flexneri,Microbiology 151 (2005) 951e962.

[13] T. Tobe, S.A. Beatson, H. Taniguchi, H. Abe, C.M. Bailey, A. Fivian,

R. Younis, S. Matthews, O. Marches, G. Frankel, T. Hayashi,

M.J. Pallen, An extensive repertoire of type III secretion effectors in

Escherichia coli O157 and the role of lambdoid phages in their dissem-

ination, Proc. Natl. Acad. Sci. U.S.A. 103 (2006) 14941e14946.

[14] M. Kelly, E. Hart, R. Mundy, O. Marches, S. Wiles, L. Badea, S. Luck,

M. Tauschek, G. Frankel, R.M. Robins-Browne, E.L. Hartland, Essential

role of the Type III secretion system effector NleB in colonization of

mice by Citrobacter rodentium, Infect. Immun. 74 (2006) 2328e2337.

[15] S. Gruenheid, I. Sekirov, N.A. Thomas, W. Deng, P. O’Donnell,

D. Goode, Y. Li, E.A. Frey, N.F. Brown, P. Metalnikov, T. Pawson,

K. Ashman, B.B. Finlay, Identification and characterization of NleA,

a non-LEE-encoded type III translocated virulence factor of enterohae-

morrhagic Escherichia coli O157:H7, Mol. Microbiol. 51 (2004)

1233e1249.

[16] P.R. Hardwidge, W. Deng, B.A. Vallance, I. Rodriguez-Escudero,

V.J. Cid, M. Molina, B.B. Finlay, Modulation of host cytoskeleton func-

tion by the enteropathogenic Escherichia coli and Citrobacter rodentiumeffector protein EspG, Infect. Immun. 73 (2005) 2586e2594.

[17] C. Ma, M.E. Wickham, J.A. Guttman, W. Deng, J. Walker, K.L. Madsen,

K. Jacobson, W.A. Vogl, B.B. Finlay, B.A. Vallance, Citrobacter roden-

tium infection causes both mitochondrial dysfunction and intestinal

epithelial barrier disruption in vivo: role of mitochondrial associated pro-

tein (Map), Cell. Microbiol. 8 (2006) 1669e1686.

[18] S. Wiles, S. Clare, J. Harker, A. Huett, D. Young, G. Dougan, G. Frankel,

Corrigendum: organ specificity, colonization and clearance dynamics in

vivo following oral challenges with the murine pathogen Citrobacter

rodentium, Cell. Microbiol. 7 (2005) 459.

[19] S.A. Luperchio, J.V. Newman, C.A. Dangler, M.D. Schrenzel,

D.J. Brenner, A.G. Steigerwalt, D.B. Schauer, Citrobacter rodentium,

the causative agent of transmissible murine colonic hyperplasia, exhibits

clonality: synonymy of C. rodentium and mouse-pathogenic Escherichia

coli, J. Clin. Microbiol. 38 (2000) 4343e4350.

[20] W. Deng, B.A. Vallance, Y. Li, J.L. Puente, B.B. Finlay, Citrobacter

rodentium translocated intimin receptor (Tir) is an essential virulence

factor needed for actin condensation, intestinal colonization and colonic

hyperplasia in mice, Mol. Microbiol. 48 (2003) 95e115.

407M.E. Wickham et al. / Microbes and Infection 9 (2007) 400e407

[21] D.B. Schauer, Murine colonic hyperplasia, in: V.L. Miller, J.B. Kaper,

D.A. Portnoy, R.R. Isberg (Eds.), Molecular Genetics of Bacterial Path-

ogenesis, American Society for Microbiology, Washington, DC, 1994,

pp. 197e208.

[22] S. Mirold, W. Rabsch, H. Tschape, W.D. Hardt, Transfer of the Salmo-nella type III effector sopE between unrelated phage families, J. Mol.

Biol. 312 (2001) 7e16.

[23] B.K. Coombes, M.E. Wickham, N.F. Brown, S. Lemire, L. Bossi,

W.W. Hsiao, F.S. Brinkman, B.B. Finlay, Genetic and molecular analysis

of GogB, a phage-encoded type III-secreted substrate in Salmonella en-

terica serovar typhimurium with autonomous expression from its associ-

ated phage, J. Mol. Biol. 348 (2005) 817e830.

[24] B.K. Coombes, M.E. Wickham, M.J. Lowden, N.F. Brown, B.B. Finlay,

Negative regulation of Salmonella pathogenicity island 2 is required for

contextual control of virulence during typhoid, Proc. Natl. Acad. Sci.

U.S.A. 102 (2005) 17460e17465.

[25] N.F. Brown, M.E. Wickham, B.K. Coombes, B.B. Finlay, Crossing the line:

selection and evolution of virulence traits, PLoS Pathog. 2 (2006) e42.

[26] S.W. Barthold, G.L. Coleman, P.N. Bhatt, G.W. Osbaldiston, A.M. Jonas,

The etiology of transmissible murine colonic hyperplasia, Lab. Anim.

Sci. 26 (1976) 889e894.

[27] K. Itoh, T. Matsui, K. Tsuji, T. Mitsuoka, K. Ueda, Genetic control in the

susceptibility of germfree inbred mice to infection by Escherichia coliO115a,c:K(B), Infect. Immun. 56 (1988) 930e935.

[28] D.J. Brenner, P.A. Grimont, A.G. Steigerwalt, G.R. Fanning, E. Ageron,

C.F. Riddle, Classification of citrobacteria by DNA hybridization: desig-

nation of Citrobacter farmeri sp. nov., Citrobacter youngae sp. nov., Cit-robacter braakii sp. nov., Citrobacter werkmanii sp. nov., Citrobacter

sedlakii sp. nov., and three unnamed Citrobacter genomospecies, Int. J.

Syst. Bacteriol. 43 (1993) 645e658.

[29] M.L. McKee, A.R. Melton-Celsa, R.A. Moxley, D.H. Francis,

A.D. O’Brien, Enterohemorrhagic Escherichia coli O157:H7 requires

intimin to colonize the gnotobiotic pig intestine and to adhere to HEp-

2 cells, Infect. Immun. 63 (1995) 3739e3744.

[30] E.A. Dean-Nystrom, B.T. Bosworth, W.C. Cray Jr., H.W. Moon, Patho-

genicity of Escherichia coli O157:H7 in the intestines of neonatal calves,

Infect. Immun. 65 (1997) 1842e1848.

[31] E.A. Dean-Nystrom, A.R. Melton-Celsa, J.F. Pohlenz, H.W. Moon,

A.D. O’Brien, Comparative pathogenicity of Escherichia coli O157

and intimin-negative non-O157 Shiga toxin-producing E. coli strains in

neonatal pigs, Infect. Immun. 71 (2003) 6526e6533.

[32] T.S. Whittam, I.K. Wachsmuth, R.A. Wilson, Genetic evidence of clonal

descent of Escherichia coli O157:H7 associated with hemorrhagic colitis

and hemolytic uremic syndrome, J. Infect. Dis. 157 (1988) 1124e1133.

[33] M. Ohnishi, J. Terajima, K. Kurokawa, K. Nakayama, T. Murata,

K. Tamura, Y. Ogura, H. Watanabe, T. Hayashi, Genomic diversity of

enterohemorrhagic Escherichia coli O157 revealed by whole genome

PCR scanning, Proc. Natl. Acad. Sci. U.S.A. 99 (2002) 17043e17048.

[34] L.M. Wick, W. Qi, D.W. Lacher, T.S. Whittam, Evolution of genomic

content in the stepwise emergence of Escherichia coli O157:H7, J. Bac-

teriol. 187 (2005) 1783e1791.

[35] K.B. Hisert, M. MacCoss, M.U. Shiloh, K.H. Darwin, S. Singh,

R.A. Jones, S. Ehrt, Z. Zhang, B.L. Gaffney, S. Gandotra,

D.W. Holden, D. Murray, C. Nathan, A glutamateealanineeleucine

(EAL) domain protein of Salmonella controls bacterial survival in

mice, antioxidant defence and killing of macrophages: role of cyclic

diGMP, Mol. Microbiol. 56 (2005) 1234e1245.

[36] M.A. Khan, C. Ma, L.A. Knodler, Y. Valdez, C.M. Rosenberger,

W. Deng, B.B. Finlay, B.A. Vallance, Toll-like receptor 4 contributes

to colitis development but not to host defense during Citrobacter roden-

tium infection in mice, Infect. Immun. 74 (2006) 2522e2536.

[37] R.R. Chaudhuri, A.M. Khan, M.J. Pallen, coliBASE: an online database

for Escherichia coli, Shigella and Salmonella comparative genomics,

Nucleic Acids Res. 32 (2004) D296eD299.